Effects of Atorvastatin Versus Other Statins on Fasting and Postprandial C-Reactive Protein and Lipoprotein-Associated Phospholipase A2 in Patients With Coronary Heart Disease Versus Control Subjects • ARTICLE Pages 1025-1032 Ernst J. Schaefer, Judith R. McNamara, Bela F. Asztalos, Timothy Tayler, Jennifer A. Daly, Joi L. Gleason, Leo J. Seman, Andrea Ferrari and Joel J. Rubenstein
3.
Usefulness of Real-Time Myocardial Perfusion Imaging to Evaluate Tissue Level Reperfusion in Patients With Non–ST-Elevation Myocardial Infarction • ARTICLE Pages 1033-1038 Grigorios Korosoglou, Nina Labadze, Evangelos Giannitsis, Raffi Bekeredjian, Alexander Hansen, Stefan E. Hardt, Christiane Selter, Roger Kranzhoefer, Hugo Katus and Helmut Kuecherer
4.
Usefulness of Tissue Doppler Imaging in the Diagnosis and Prognosis of Acute Right Ventricular Infarction With Inferior Wall Acute Left Ventricular Infarction • ARTICLE Pages 1039-1042 Hisham Dokainish, Heather Abbey, Kenneth Gin, Krishnan Ramanathan, Pui-Kee Lee and John Jue
5.
Influence of Stent Length to Lesion Length Ratio on Angiographic and Clinical Outcomes After Implantation of Bare Metal and Drug-Eluting Stents (the TAXUS-IV Study) • ARTICLE Pages 1043-1048 S. Chiu Wong, Mun K. Hong, Stephen G. Ellis, Maurice Buchbinder, Mark Reisman, Augustin DeLago, Mirle Kellett, Jeffrey J. Popma, Mary E. Russell, Roxana Mehran et al.
Preventive Cardiology
6.
Comparison of the Relation of Triglyceride-Rich Lipoproteins and Muscular Artery Compliance in Healthy Women Versus Healthy Men • ARTICLE Pages 1049-1054 Ngoc-Anh Le, W. Virgil Brown, Warren W. Davis, David M. Herrington, Lori Mosca, Shunichi Homma, Barry Eggleston, Howard J. Willens and Jeffrey K. Raines
Arrhythmias and Conduction Disturbances
7.
Comparison of Autotriggered Memory Loop Recorders Versus Standard Loop Recorders Versus 24-Hour Holter Monitors for Arrhythmia Detection • ARTICLE Pages 1055-1059 James A. Reiffel, Robert Schwarzberg and Maria Murry
Systemic Hypertension
8.
Long-Term Effects of Stress Reduction on Mortality in Persons ≥55 Years of Age With Systemic Hypertension • ARTICLE Pages 1060-1064 Robert H. Schneider, Charles N. Alexander, Frank Staggers, Maxwell Rainforth, John W. Salerno, Arthur Hartz, Stephen Arndt, Vernon A. Barnes and Sanford I. Nidich
Miscellaneous
9.
Matrix Metalloproteinases and Their Inhibitors in Malignant and Autoreactive Pericardial Effusion • ARTICLE Pages 1065-1069 Steffen Lamparter, Michael Schoppet, Michael Christ, Sabine Pankuweit and Bernhard Maisch
Editorial
10.
Prevalence of T-Wave Inversion Beyond V1 in Young Normal Individuals and Usefulness for the Diagnosis of Arrhythmogenic Right Ventricular Cardiomyopathy/dysplasia • DISCUSSION Pages 1070-1071 Frank I. Marcus
Brief Reports
11.
Comparison of All-Cause Mortality in Women With Known or Suspected Coronary Artery Disease Referred for Dobutamine Stress Echocardiography With Normal Versus Abnormal Test Results • SHORT COMMUNICATION Pages 1072-1075 Elena Biagini, Abdou Elhendy, Arend F.L. Schinkel, Vittoria Rizzello, Ron T. van Domburg, Boudewijn J. Krenning, Olaf Schouten, Fabiola B. Sozzi, Angelo Branzi, Guido Rocchi et al.
12.
Ability for Visualization, Reasons for Nonassessable Image, and Diagnostic Accuracy of 16-Slice Multidetector Row Helical Computed Tomography for the Assessment of the Entire Coronary Arteries • SHORT COMMUNICATION Pages 1076-1079 Toshiro Kitagawa, Takashi Fujii, Yasuyuki Tomohiro, Kouji Maeda, Masakazu Kobayashi, Eiji Kunita, Yoshitaka Sekiguchi, Takayuki Suzuki, Kouichi Fujikawa and Hiroshi Yamaguchi
13.
Effectiveness of High-Intensity Interval Training for the Rehabilitation of Patients With Coronary Artery Disease • SHORT COMMUNICATION Pages 1080-1084 Darren E.R. Warburton, Donald C. McKenzie, Mark J. Haykowsky, Arlana Taylor, Paula Shoemaker, Andrew P. Ignaszewski and Sammy Y. Chan
14.
Effectiveness of Inhibition of Cholesteryl Ester Transfer Protein by JTT-705 in Combination With Pravastatin in Type II Dyslipidemia • SHORT COMMUNICATION Pages 1085-1088 Jan Albert Kuivenhoven, Greetje J. de Grooth, Hitoshi Kawamura, Anke H. Klerkx, Francois Wilhelm, Mieke D. Trip and John J.P. Kastelein
15.
Effect of Chronic Total Coronary Occlusion on Treatment Strategy • SHORT COMMUNICATION Pages 1088-1091 Ryan D. Christofferson, Kenneth G. Lehmann, Gary V. Martin, Nathan Every, James H. Caldwell and Samir R. Kapadia
16.
Outcomes Following Elective Percutaneous Coronary Intervention Without On-Site Surgical Backup in a Community Hospital • SHORT COMMUNICATION Pages 1091-1093 Alexander Paraschos, Dwayne Callwood, Marilyn B. Wightman, James E. Tcheng, Harry R. Phillips, Gary L. Stiles, John M. Daniel and Michael H. Sketch, Jr
17.
Follow-up of Coronary Artery Bypass Graft Patency by Multislice Computed Tomography • SHORT COMMUNICATION Pages 1094-1097 Emilio Chiurlia, Mila Menozzi, Carlo Ratti, Renato Romagnoli and Maria Grazia Modena
18.
Do Statins Increase the Risk of Idiopathic Polyneuropathy? • SHORT COMMUNICATION Pages 1097-1099 Jeffrey L. Anderson, Joseph B. Muhlestein, Tami L. Bair, Steven Morris, Aaron N. Weaver, Donald L. Lappé, Dale G. Renlund, Robert R. Pearson, Kurt R. Jensen and Benjamin D. Horne
19.
Relation of Echocardiographic Wall Motion Score Index and Response to Dobutamine Stress to Defibrillation Threshold at the Time of Implantation of a Cardiac Defibrillator • SHORT COMMUNICATION Pages 1099-1101 Ali M. Kizilbash, Paul A. Grayburn, Prasad Anand, Robert C. Kowal, Richard L. Page, Kim Smith and Mohamed H. Hamdan
20.
Effect of Syncope-Related Traumatic Injuries on the Diagnostic Evaluation and Syncope Recurrence of Patients With Syncope and Apparently Normal Hearts • SHORT COMMUNICATION Pages 1101-1103 Andreas Schuchert, Renke Maas, Kai Mortensen, Muhammet Ali Aydin, Christina Kretzschmar and Thomas Meinertz
21.
Monitoring of Physical Activity and Heart Rate Variability in Patients With Chronic Heart Failure Using Cardiac Resynchronization Devices • SHORT COMMUNICATION Pages 1104-1107 Frieder Braunschweig, Peter T. Mortensen, Daniel Gras, Wolfgang Reiser, Thomas Lawo, Hassan Mansour, Peter Sogaard, Berthold Stegemann, Hans-Jürgen Bruns, Cecilia Linde and InSync III Study Investigators
22.
Echocardiographic Examination of Atrioventricular and Interventricular Delay Optimization in Cardiac Resynchronization Therapy • SHORT COMMUNICATION Pages 1108-1110 Maria Cristina Porciani, Cristina Dondina, Roberto Macioce, Gabriele Demarchi, Paolo Pieragnoli, Nicola Musilli, Andrea Colella, Giuseppe Ricciardi, Antonio Michelucci and Luigi Padeletti
23.
Effect of Cardiac Resynchronization Therapy on Inducibility of Ventricular Tachyarrhythmias in Cardiac Arrest Survivors With Either Ischemic or Idiopathic Dilated Cardiomyopathy • SHORT COMMUNICATION Pages 1111-1114 Philippine Kiès, Jeroen J. Bax, Sander G. Molhoek, Gabe B. Bleeker, Katja Zeppenfeld, Marianne Bootsma, Lieselot van Erven, Paul Steendijk, Ernst E. van der Wall and Martin J. Schalij
24.
Long-Term Effects of Carvedilol or Metoprolol on Left Ventricular Function in Ischemic and Nonischemic Cardiomyopathy • SHORT COMMUNICATION Pages 1114-1116 Philip Green, Michael Anshelevich, Ashok Talreja, Joyce L. Burcham, Srinivas M. Ravi, Jamshid Shirani and Thierry H. Le Jemtel
25.
Prognostic Impact of Diabetes Mellitus in Patients With Acute Decompensated Heart Failure • SHORT COMMUNICATION Pages 1117-1119 Andrew J. Burger, Lana Tsao and Doron Aronson
26.
Effect of a Seated Exercise Program to Improve Physical Function and Health Status in Frail Patients ≥70 Years of Age With Heart Failure • SHORT COMMUNICATION Pages 1120-1124 Miles D. Witham, Joan M. Gray, Ishbel S. Argo, Derek W. Johnston, Allan D. Struthers and Marion E.T. McMurdo
27.
Usefulness of Enalapril Versus Propranolol or Atenolol for Prevention of Aortic Dilation in Patients With the Marfan Syndrome • SHORT COMMUNICATION Pages 1125-1127 Anji T. Yetman, Renee A. Bornemeier and Brian W. McCrindle
28.
Mechanism of Spaceflight-Induced Changes in Left Ventricular Mass • SHORT COMMUNICATION Pages 1128-1130 Richard L. Summers, David S. Martin, Janice V. Meck and Thomas G. Coleman
29.
Usefulness of Contrast Echocardiography for Diagnosis of Left Ventricular Noncompaction • SHORT COMMUNICATION Pages 1131-1134 Lotte E. de Groot-de Laat, Boudewijn J. Krenning, Folkert J. ten Cate and Jos R.T.C. Roelandt
EDITOR IN CHIEF
William C. Roberts, MDBaylor Heart & Vascular HospitalBaylor University Medical CenterWadley Tower No. 4573600 Gaston AvenueDallas, Texas 75246(214)826-8252Fax: (214)826-2855
ASSOCIATE EDITORSThomas C. AndrewsPaul A. GrayburnASSISTANT EDITORSCalixto A. RomeroRobert L. RosenthalCarlos E. Velasco
EDITORIAL BOARDCARDIOVASCULARMEDICINEIn AdultsJonathan AbramsRobert J. AdolphJoseph S. AlpertMartin A. AlpertEzra A. AmsterdamJeffrey AndersonWilliam F. ArmstrongRichard W. AsingerDonald S. BaimGary John BaladyThomas M. BashoreEric BatesGeorge A. BellerDavid G. BendittPeter C. BlockWilliam E. BodenMonty M. BodenheimerRobert O. BonowJeffrey S. BorerHarisios BoudoulasMartial G. BourassaEugene BraunwaldJeffrey A. BrinkerBruce R. BrodieAlfred E. BuxtonMichael E. CainRichard O. Cannon IIISamuel Ward CasscellsAgustin CastellanosBernard R. ChaitmanKanu ChatterjeeJohn S. ChildRobert J. CodyLawrence S. CohenMarc CohenC. Richard ContiMichael H. CrawfordRobert F. DeBuskPrakash C. DeedwaniaGregory J. DehmerEfthymios N. DeliargyrisJames A. de LemosAnthony N. DeMariaPablo DenesNicholas L. DePaceRichard B. DevereuxGeorge A. DiamondJohn P. DiMarcoMichael J. DomanskiGerald DorrosJohn S. Douglas, Jr.Pamela S. DouglasEric J. EichhornMark J. EisenbergUri Elkayam
Kenneth A. EllenbogenMyrvin H. EllestadStephen G. EllisToby R. EngelAndrew E. EpsteinN. A. Mark Estes, IIIMichael EzekowitzRodney H. FalkJohn A. FarmerDavid P. FaxonRobert L. FeldmanTed FeldmanJack FerlinzJerome L. FlegGerald F. FletcherNancy C. FlowersJames S. ForresterJoseph A. FranciosaGary S. FrancisVictor F. Froelicher, Jr.W. Bruce FyeWilliam H. GaaschWilliam GanzJulius M. GardinBernard J. GershMihai GheorghiadeRaymond GibbonsD. Luke GlancyStephen P. GlasserMichael R. GoldSamuel Z. GoldhaberRobert E. GoldsteinSidney GoldsteinSteven A. GoldsteinJ. Anthony GomesAntonio M. Gotto, Jr.K. Lance GouldDonald C. HarrisonRichard H. HelfantGary V. HellerPhilip D. HenryL. David HillisDavid R. Holmes, Jr.William G. HundleyAbdulmassih S. IskandrianAllan S. JaffeWilliam B. KannelNorman M. KaplanJoel S. KarlinerJohn A. KastorSanjiv KaulHarold L. KennedyKenneth M. KentRichard E. KerberDean J. KereiakesMorton J. KernSpencer B. King IIIRobert E. KleigerGeorge J. Klein
Lloyd W. KleinPaul KligfieldRobert A. KlonerJohn B. KostisMorris N. KotlerCharles LandauRichard L. LangeCarl J. LavieCarl V. LeierJoseph Lindsay, Jr.Gregory Y.H. LipJoseph LoscalzoG.B. John ManciniFrancis E. MarchlinskiFrank I. MarcusBarry J. MaronRandolph P. MartinAttilo MaseriDean T. MasonMichael D. McGoonRaymond G. McKayJawahar L. MehtaRichard S. MeltzerFranz H. MesserliEric L. MichelsonAlan B. MillerGary S. MintzFred MoradyArthur J. MossJames E. MullerRobert J. MyerburgGerald B. NaccarelliNavin C. NandaChristopher O’ConnorRobert A. O’RourkeErik Magnus OhmanAntonio PacificoRichard L. PageEugene R. PassamaniRichard C. PasternakAlan S. PearlmanCarl J. PepineJoseph K. PerloffIleana PinaBertram PittPhilip J. PodridArshed A. QuyyumiCharles E. RackleyC. Venkata RamNathaniel ReichekRobert RobertsWilliam J. RogersMelvin M. ScheinmanNelson B. SchillerDavid J. SchneiderJohn S. SchroederPravin M. ShahPrediman K. ShahJamshid Shirani
Robert J. SiegelMarc A. SilverMark E. SilvermanRoss J. Simpson, Jr.Steven N. SinghSidney C. Smith, Jr.Burton E. SobelJohn C. SombergPeter C. SpittellDavid H. SpodickLynne W. StevensonJohn R. StrattonJonathan M. TobisEric J. TopolByron F. VandenbergHector O. VenturaGeorge W. VetrovecRonald G. VictorRobert A. VogelFrans J. WackersDavid D. WatersNanette K. WengerWilliam B. WhiteRobert WilenskyJames T. WillersonJohn R. WilsonMiguel ZabalgoitiaBarry L. ZaretDouglas P. ZipesIn Infants and ChildrenHugh D. AllenBruce S. AlpertArthur Garson, Jr.Stanley J. GoldbergThomas P. Graham, Jr.Warren G. GuntherothHoward P. GutgesellJohn D. KuglerJames E. LockJohn W. MooreLowell W. PerryDavid J. SahnRichard M. SchiekenCARDIOVASCULAR SURGERYEugene H. BlackstoneLawrence I. BonchekLawrence H. CohnJohn A. ElefteriadesThomas L. SprayRELATED SPECIALISTSL. Maximilian BujaBarry A. FranklinCharles B. HigginsJeffrey E. SaffitzRenu VirmaniRedford B. Williams
A2
Effects of Atorvastatin Versus OtherStatins on Fasting and PostprandialC-Reactive Protein and Lipoprotein-
Associated Phospholipase A2 in PatientsWith Coronary Heart Disease Versus
Control SubjectsErnst J. Schaefer, MD, Judith R. McNamara, MT, Bela F. Asztalos, PhD,
Timothy Tayler, PhD, Jennifer A. Daly, MD, Joi L. Gleason, MS, RD,Leo J. Seman, MD, PhD, Andrea Ferrari, RN, and Joel J. Rubenstein, MD
The effects of atorvastatin (40 mg/day) versus placeboon fasting and postprandial plasma levels of high-sen-sitivity C-reactive protein (hs-CRP) and lipoprotein-asso-ciated phospholipase A2 (Lp-PLA2) were examined over36 weeks in 84 patients who had coronary heart dis-ease and low-density lipoprotein cholesterol levels>130 mg/dl and compared directly with the effects offluvastatin, lovastatin, pravastatin, and simvastatin. Re-sults were also compared with those obtained in age-and gender-matched control subjects (n � 84). Feedingincreased median hs-CRP levels by 2% in patients (p �NS) and 22% in controls (p <0.01) and increased meanLp-PLA2 values by 9% in patients (p � NS) but de-creased values by 21% in controls (p <0.0001). Patientshad 51% higher median hs-CRP values and 29% highermean Lp-PLA2 values than did controls (p <0.05 forhs-CRP and Lp-PLA2) in the fasting state; however, Lp-
PLA2 values were 62% higher (p <0.0001) in the fedstate in patients compared with controls. Atorvastatindecreased median hs-CRP levels by 32% (p <0.01) andmean Lp-PLA2 values by 26% in patients (p <0.0001),with similar decreases in the fed state, and none of theother statins had any significant effect on these param-eters. Change in Lp-PLA2 was significantly related tochange in low-density lipoprotein cholesterol (p <0.01),with no significant relations with change in hs-CRP. Ourdata indicate greater differences in patients with coro-nary heart disease compared with controls in Lp-PLA2 inthe fed state than in the fasting state and that atorva-statin is more effective than fluvastatin, lovastatin, prav-astatin, or simvastatin for decreasing not only low-density lipoprotein cholesterol but also hs-CRP andLp-PLA2. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1025–1032)
The focus of this study was the examination of theeffects of atorvastatin versus those of other statins
(fluvastatin, lovastatin, pravastatin, and simvastatin)on high-sensitivity C-reactive protein (hs-CRP), li-poprotein-associated phospholipase A2 (Lp-PLA2),and plasma lipoprotein subspecies in the fasting andpostprandial states (4 hours after a meal rich in satu-rated fat and cholesterol) in patients who had coronaryheart disease (CHD) and a comparison of their param-eters on and off therapy with those of age- and gender-matched normal control subjects. High plasma levels
of Lp-PLA2 and CRP are markers of inflammationand increased CHD risk. Lp-PLA2 is produced bymacrophages, whereas CRP is primarily synthesizedin the liver. We previously documented that patientswho have CHD with low-density lipoprotein choles-terol levels �130 mg/dl have high plasma levels oftriglycerides and remnant lipoprotein cholesterol andtriglyceride in the fasting and fed states comparedwith controls and that a fat challenge is not necessaryto detect these abnormalities.1 In these studies, wealso documented that fasting was not essential todetect abnormalities in direct low-density lipoproteincholesterol, non–high-density lipoprotein cholesterol,and high-density lipoprotein cholesterol.1 In addition,we reported that atorvastatin therapy at 40 mg/daynormalizes levels of all classes of triglyceride-richlipoproteins and low-density lipoproteins (large, me-dium, and small low-density lipoproteins) in the fast-ing and fed states in patients who have CHD com-pared with control subjects2 and that atorvastatin ismore effective than other tested statins in this regard.3
METHODSStudy subjects: The 84 patients in this study who
had CHD were recruited from the clinic or by adver-
From the Cardiovascular Research and Lipid Metabolism Laboratories,Tufts University School of Medicine, Friedman School of NutritionScience and Policy at Tufts University, and Jean Mayer USDA HumanNutrition Research Center on Aging at Tufts University, Boston,Massachusetts; and the Cardiology Division, Newton-Wellesley Hos-pital, Newton, Massachusetts. This study was supported by investiga-tor-initiated research contracts from Parke-Davis/Pfizer, New York,New York; diaDexus, Inc., South San Francisco, California; andOtsuka America Pharmaceuticals, Inc., Rockville, Maryland. Manu-script received September 22, 2004; revised manuscript receivedand accepted January 3, 2005.
Address for reprints: Ernst J. Schaefer, MD, Tufts University, 711Washington Street, Boston, Massachusetts 02111. E-mail: [email protected].
tising and were required to have evidence of estab-lished CHD (coronary artery bypass grafting, angio-plasty, documented myocardial infarction, significantcoronary artery stenosis �50% as assessed by angiog-raphy, or significantly decreased cardiac perfusionbased on cardiac imaging with and without exercise).All subjects were studied after providing informedconsent according to protocols approved by the humaninvestigation review committee at the New EnglandMedical Center (Boston, Massachusetts) and Newton-Wellesley Hospital (Newton, Massachusetts). Patientswere excluded if they had evidence of renal impair-ment, hypothyroidism, or liver dysfunction based onclinical chemistry testing or had previous adversereactions to statins. In addition, patients were requiredto have a low-density lipoprotein cholesterol serumconcentration �130 mg/dl while off lipid-loweringmedication for �6 weeks, including anion exchangeresins, statins, fibric acid derivatives, fish oil, or nia-cin-containing products. Patients were maintained onother medications throughout the study, with nochange, including calcium channel blockers, � block-ers, and diuretics. Patients had previously receivedinstruction on a National Cholesterol Education Pro-gram diet that contained �30% of calories as fat,�7% of calories as saturated fat, and �200 mg/day ofcholesterol. Patients had a lead-in baseline period of 4weeks on this diet and did not use lipid-loweringmedication.
At the end of the 4-week lead-in period, patientswere sampled after a 12-hour overnight fast and thenrandomized to receive atorvastatin or 1 of 4 otherstatins (fluvastatin, n � 26; lovastatin, n � 20; prav-astatin, n � 17; or simvastatin, n � 16) for a protocolin which they were sampled in the fasting state after 4weeks on 20 mg/day, after another 4 weeks on 40 mg/day, after another 4 weeks on 80 mg/day or the max-imal dose (40 mg for pravastatin and simvastatin), andthen at 4 and 8 weeks during an 8-week placeboperiod. Thereafter, patients went on a second 12-weekstatin period in which they received 20 mg/day for 4weeks, 40 mg/day for 4 weeks, and 80 mg/day or themaximal dose for 4 weeks with the sampling schedule.If they had not received atorvastatin in the first statinperiod, they received it during the second period. Ifthey had received atorvastatin during the first statinperiod, they were randomized to 1 of the other 4statins during the second statin period of 12 weeks.
At the end of the 4-week lead-in period, at the endof the 40 mg/day dose of each statin, and at the end of8-week period on placebo, most subjects (n � 61)were also sampled 4 hours after a fat-rich meal (ob-tained from a McDonald’s restaurant; 2 eggs, 2 sau-sages, and 2 muffins; 880 total calories, 500 calories(57%) from fat, 180 calories (20%) from saturated fat,and 510 mg of cholesterol).
We also recruited an equal number of age- andgender-matched control subjects (n � 84) using thesame exclusion criteria as for patients with CHD,except that they were also required to be clinically freeof CHD by history, and there were no exclusioncriteria concerning low-density lipoprotein choles-
terol. These subjects were studied only 1 time in thefasting state and 4 hours after the same fat-rich meal.Data on lipid results in the fasting and postprandialstates at the end of the baseline period in patients whohad CHD are not provided because they were notsignificantly different from values at the end of the8-week placebo period.
Lipoprotein analyses: Blood was drawn from eachsubject after a 12-hour overnight fast into tubes thatcontained ethylenediaminetetraacetic acid (final concen-tration 0.15%). Plasma was separated by centrifugation(2,500 rpm at 4°C for 20 minutes). Plasma concentra-tions of total cholesterol, triglycerides, and high-densitylipoprotein cholesterol were measured fresh according tostandard enzymatic methods and reagents obtained fromAbbott Diagnostics (Irving, Texas) as described previ-ously.4 Low-density lipoprotein cholesterol was mea-sured directly with kits from Equal Diagnostics (Exton,Pennsylvania). Apolipoprotein-AI and apolipoprotein-Bwere measured by immunoturbidimetric assays fromWako Chemicals (Richmond, Virginia), and lipopro-tein(a) was measured by mass with Macra kits5
(Wampole Laboratories, Cranbury, New Jersey) and bycholesterol content using kits from Genzyme Diagnos-tics (Cambridge, Massachusetts).6 Data on lipoprotein(a)mass are not reported because none of the statins inducedstatistically significant effects on this parameter com-pared with placebo or each other. Among-run precisionstudies indicated coefficients of variation of �2% fortotal cholesterol, �3% for triglycerides, �4% for low-density lipoprotein cholesterol, �5% for high-densitylipoprotein cholesterol, apolipoprotein-AI, and lipopro-tein(a), and �10% for apolipoprotein-B and lipopro-tein(a) cholesterol. Lipid assays are standardized throughthe Centers for Disease Control Lipid StandardizationProgram (Atlanta, Georgia).
Remnant lipoprotein analyses: Isolation of remnantlipoprotein was based on removal of apolipoprotein-AI–containing particles (high-density lipoprotein) andmost apolipoprotein-B–containing particles (low-den-sity lipoprotein, nascent very low-density lipoprotein,and nascent chylomicrons) using a previously de-scribed immunoseparation technique (JIMRO, JapanImmunoresearch Laboratories, Takasaki, Japan) thathas been shown to leave particles characteristic ofvery low-density lipoprotein remnants and chylomi-cron remnants in the unbound fraction.7 Briefly,monoclonal antibodies to apolipoprotein-AI and spe-cific monoclonal antibodies to apolipoprotein-B (JI-H),which do not recognize partially hydrolyzed, apoli-poprotein-E–enriched lipoprotein remnants, were im-mobilized on agarose gel. Remnant lipoprotein choles-terol and remnant lipoprotein triglyceride concentrationswere measured in plasma aliquots and stored at�80°C until analysis. Thawed plasma was incubatedwith the gel for 2 hours, after which the gel, whichcontained the bound (nonremnant lipoprotein) li-poproteins, was precipitated with low-speed centrifu-gation (5 minutes at 135g). Remnant lipoprotein cho-lesterol and remnant lipoprotein triglyceride levelswere then measured in the unbound supernates on anAbbott Spectrum CCx chemistry analyzer (Abbott Di-
1026 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
TABLE 1 Fasting Concentrations for Controls and Patients on Placebo
Normally distributed values are reported as mean � SD; other values are reported as medians (interquartile ranges). All p values were derived from 2-tailed, pairedt-test analysis.
*Values for hs-CRP, triglycerides, remnant lipoprotein cholesterol, and remnant lipoprotein triglycerides were log-transformed for statistical analysis.†p �0.05; ‡p�0.01; §p �0.001;#p �0.0001 for comparisons with placebo; the dataset comprised 12 women and 72 men; mean percentage � SD or median percent changefrom placebo.
Normally distributed values are reported as mean � SD; other values are reported as medians (interquartile ranges). All p values were derived from 2-tailed, pairedt-test analysis.
*Values for hs-CRP, creatine phosphokinase, aspartate aminotransferase, arginine aminotransferase, triglycerides, lipoprotein(a) cholesterol, lipoprotein(a) mass,remnant lipoprotein cholesterol, and remnant lipoprotein triglycerides were log-transformed for statistical analysis.
†p �0.05; ‡p �0.01; §p �0.001; #p �0.0001 for comparisons with placebo; the dataset comprised 12 women and 72 men; mean percentage � SD or medianpercent change from placebo.Abbreviations as in Table 1.
CORONARY ARTERY DISEASE/DIRECT COMPARISONS OF STATIN EFFECTS ON CRP AND LP-PLA2 1027
agnostics) that used 2-reagent enzymatic, colorimetricassays that contained a sensitive chromophore(Kyowa Medex, Tokyo, Japan). Precision studiesyielded among-run remnant lipoprotein cholesterolimprecision for 2 levels of remnant lipoprotein controlover 20 runs, i.e., 9.1% at 7 mg/dl and 7.3% at 24mg/dl. Among-run remnant lipoprotein triglycerideimprecision levels for the same controls were 8.3% at22 mg/dl and 5.0% at 109 mg/dl.7
Nuclear magnetic resonance analysis: Lipoproteinsubclass distributions were determined by proton nu-clear magnetic resonance spectroscopy.8 Subclasseswere separated as chylomicrons (�200 nm); large (60to 200 nm), medium (35 to 60 nm), and small (27 to35 nm) very low-density lipoproteins; intermediate-density lipoproteins (23 to 27 nm); large (21 to 23nm), medium (20 to 21 nm), and small (18 to 20 nm)low-density lipoproteins; and large (9 to 13 nm), me-dium (8 to 9 nm), and small (7 to 8 nm) high-densitylipoproteins. Very low-density lipoprotein subclassconcentrations are expressed as milligrams per decili-ter of triglycerides, and other lipoprotein subclassconcentrations are expressed as milligrams per decili-ter of cholesterol. Data on the chylomicron and inter-mediate-density lipoprotein fractions were omitted be-cause the levels of lipids within the fractions werelow, with no statistically significant differences acrossgroups.
Clinical chemistry analyses: Standardized serum mea-surements of liver transaminases and creatine phos-phokinase were carried out at each 4-week time pointin the study. Measurements of hs-CRP were per-formed on a Hitachi 911 (Roche Diagnostics,Indianapolis, Indiana) using the CRP-UL kit fromWako Chemicals. Within- and between-run coeffi-cients of variation were �5%. Measurements of Lp-PLA2 mass were performed by an enzyme-linkedimmunosorbent assay (diaDexus, Inc., South SanFrancisco, California) at the diaDexus facility in ablinded fashion. Within- and between-run coefficientsof variation were �7.5%.
Statistical analysis: Statistical analyses were per-formed with patients who had CHD by comparing theeffects of each statin at 40 mg/day with placebo, and itseffects were compared with those observed on the samedose of atorvastatin in the same patients. The effects ofstatins at 40 mg/day in the fasting and fed states werecompared with placebo and with values obtained on thesame dose of atorvastatin. Values in control subjects whodid not have CHD are provided for comparison. Stu-dent’s t test was used to compare mean levels of contin-uous measurements, and a chi-square statistic was cal-culated for categorical factors. For continuousmeasurements that were highly skewed, we performedcomparisons using log-transformed values, althoughwe report the untransformed means and SDs for eachgroup. Analyses were also conducted separately bygender, but we report only pooled results because menand women had similar responses to statins. Similarly,we do not report baseline lipid data because theseresults were not statistically different from valuesobtained on placebo. For the postprandial studies,
statistical analyses were done to compare fasting val-ues with postprandial values, values on medicationwith those on placebo, and values in patients who hadCHD and used placebo and at 40 mg/day of statinswith values on atorvastatin. We also carried out cor-relations for change in hs-CRP and Lp-PLA2 versuschange in other parameters influenced by atorvastatin.One outlier that had a 8,531% increase in hs-CRP onatorvastatin was omitted from the correlation analysis.
RESULTSBaseline demographic information for subjects who
had CHD (n � 84, 87% men and 13% women) and forage- and gender-matched controls (n � 84) were meanage (62 � 9 years), diabetes mellitus (0% of controls and5% of patients), hypertension (31% of controls and 56%of patients), smoking (20% of controls and 18% ofpatients), and body mass index (26 � 4 kg/m2 for con-trols and 28 � 4 kg/m2 for patients). Mean values � SDfor control subjects and subjects who had CHD and wereusing placebo with respect to fasting plasma levels ofhs-CRP, Lp-PLA2, total cholesterol, triglycerides, low-density lipoprotein cholesterol, high-density lipoproteincholesterol, non–high-density lipoprotein cholesterol,apolipoprotein-AI, apolipoprotein-B, and lipoprotein(a)or lipoprotein(a) cholesterol levels and to levels of rem-nant lipoprotein cholesterol and triglyceride, small low-density lipoprotein, low-density lipoprotein particlenumber, and large high-density lipoprotein are presentedin Table 1. In addition, values for the ratio of totalcholesterol to high-density lipoprotein cholesterol andpercentage of patients who had hs-CRP levels �3.0 and�2.0 mg/L, low-density lipoprotein cholesterol levels�100 and �70 mg/dl, non–high-density lipoproteincholesterol levels �130 mg/dl, and triglyceride levels�150 mg/dl are presented.
Patients had significantly (p �0.05) higher levels ofhs-CRP, Lp-PLA2, total cholesterol, low-density li-poprotein cholesterol, triglycerides, remnant lipoproteincholesterol, remnant lipoprotein triglycerides, total cho-lesterol/high-density lipoprotein cholesterol ratio, smalllow-density lipoprotein, and total low-density lipoproteinparticle number and significantly (p �0.05) lower levelsof high-density lipoprotein cholesterol and large high-density lipoprotein than did controls (Table 1). Medianhs-CRP and mean Lp-PLA2 values were 51% and 29%higher in patients than in controls, respectively (p �0.05).
The effects of atorvastatin at 40 mg/day versus pla-cebo on various parameters in 84 patients who had CHDare presented in Table 2. Median decreases in hs-CRPwere 32% (p �0.01) and mean decreases in Lp-PLA2were 26% (p �0.0001). There was a high degree ofvariability in the hs-CRP response to atorvastatin butsomewhat less variability in the Lp-PLA2 response (Fig-ure 1). For these parameters, there were increases insome patients (5 patients developed dramatic increases inhs-CRP and were excluded from the data shown inFigure 1 but not from the data analyses), but, in general,atorvastatin decreased these values substantially, withalmost normal values for low-density lipoprotein choles-terol and triglycerides (Table 2). These increases mayhave been due to concurrent infections. Atorvastatin also
1028 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
resulted in modest median increases in creatine phos-phokinase, aspartate aminotransferase, and alanine ami-notransferase of 8%, 9%, and 19%, respectively, withonly the latter 2 changes being significant (p �0.0001).The variability in response to atorvastatin by creatinephosphokinase and alanine aminotransferase is shown inFigure 1. As we previously reported, atorvastatin is veryeffective in decreasing levels of total cholesterol, low-density lipoprotein cholesterol, non–high-density li-poprotein cholesterol, triglycerides, remnant lipoproteincholesterol, remnant lipoprotein triglycerides, low-den-sity lipoprotein particle number, and small low-densitylipoprotein and in increasing levels of large high-densitylipoprotein as assessed by nuclear magnetic resonancespectroscopy.2,3
Data that examined the effects of atorvastatin onfasting and fed levels of hs-CRP, Lp-PLA2, low-densitylipoprotein cholesterol, triglyceride, and remnant li-poprotein cholesterol in 61 patients who had CHD arelisted in Table 3. Decreases induced by atorvastatin forall these parameters were very similar in the fasting andfed states, and all decreases were highly significant(p �0.0001), except for fasting hs-CRP level (p �0.05),
which was moderately significant, andnonfasting hs-CRP level, which, al-though decreased by 29%, was not sig-nificant. Of note, for all these parame-ters, there were highly significantdifferences (p �0.0001) between pa-tients on placebo and controls exceptfor Lp-PLA2. However, differenceswere very significant in the fed statebecause Lp-PLA2 was decreased by21% in the fed state in controls but didnot change significantly in patients.Moreover, although atorvastatin cameclose to normalizing Lp-PLA2, low-density lipoprotein cholesterol, andremnant lipoprotein cholesterol in pa-tients versus controls, it did not do sofor hs-CRP (especially in the fed state)or triglycerides.
Table 4 presents a comparison ofthe effects of the different statins onfasting and postprandial levels of hs-CRP and Lp-PLA2 and of low-den-sity lipoprotein cholesterol. Atorva-statin was not only significantlymore efficacious than other statins indecreasing low-density lipoproteincholesterol but also the only statinthat significantly decreased levels ofhs-CRP and Lp-PLA2 in the fastingor fed state.
Because of these effects, we re-lated changes in fasting levels of hs-CRP and Lp-PLA2 induced by ator-vastatin to other variables based oncorrelation coefficient analysis. ForLp-PLA2, the change was signifi-cantly correlated (p �0.01) withchanges in total cholesterol (r �
0.49), low-density lipoprotein cholesterol (r � 0.43),and ratio of total cholesterol to high-density lipopro-tein cholesterol (r � 0.44). For hs-CRP, the changewas not significantly correlated with change in anyparameter.
DISCUSSIONWe previously documented that fasting values of
low-density lipoprotein cholesterol, high-density li-poprotein cholesterol, triglycerides, and remnant li-poprotein cholesterol are correlated with postprandialvalues and that differences between patients who haveCHD and controls with respect to these parameterscan readily be detected in the fasting state.1 Therefore,a fat challenge is not necessary. Moreover, we foundthat low-density lipoprotein cholesterol and high-den-sity lipoprotein cholesterol can readily be measured inthe fed state, provided that newer direct methods areused and that these values are just slightly lower in thefed state than in the fasting state.1
In addition, we documented that atorvastatin isvery effective in normalizing many of the lipid abnor-malities found in patients who have CHD compared
FIGURE 1. Variability (percent change from values on placebo) in response to atorva-statin (40 mg/day) for hs-CRP, Lp-PLA2, creatine phosphokinase (CPK), and aspartateaminotransferase (AST) are shown. For hs-CRP, 5 extreme outliers have been omittedfrom the figure (�8,531%, �495%, �316%, �295%, and �242%); for CPK, 2 outli-ers have been omitted from the figure (�592% and �499%).
CORONARY ARTERY DISEASE/DIRECT COMPARISONS OF STATIN EFFECTS ON CRP AND LP-PLA2 1029
TABLE 3 Fasting and Postprandial Concentrations on Placebo and 40 mg/day Atorvastatin Versus Matched Controls
9.6 � 4.5� 19.5 � 15.1� 11.6 � 7.1� �100% (30 to 171)†† �10% (�25 to 100) �41% (�52 to �24)††
Median postprandial change �31% (13 to 50) �42% (34 to 52) �43% (29 to 54)
Normally distributed values are reported as mean � SD; other values are reported as medians (interquartile ranges). All p values were derived from 2-tailed, paired t-test analysis.*Values for hs-CRP, triglycerides, and remnant lipoprotein cholesterol were log-transformed for statistical analysis.†p �0.05; ‡p �0.01; §p �0.001; �p �0.0001 for comparisons of fasting and postprandial values.¶p �0.05; #p �0.01; **p �0.001; ††p �0.0001 for paired t-test comparison among treatment groups; the dataset comprised 9 women and 52 men; controls were matched by age and gender to patients mean percentage � SD or medianpercent change from placebo.Abbreviation as in Table 1.
1030TH
EA
MERIC
AN
JOU
RNA
LO
FC
ARD
IOLO
GY�
VOL.
95M
AY
1,2005
with controls, including increased low-density li-poprotein and non–high-density lipoprotein choles-terol, cholesterol in small dense low-density lipopro-tein, total low-density lipoprotein particle number,triglycerides, triglycerides in different very low-den-sity lipoprotein classes, postprandial triglycerides, andfasting and postprandial remnant lipoprotein choles-terol and triglycerides.2 Atorvastatin was found todecrease lipoprotein(a) cholesterol but not total li-poprotein(a) mass.2 Moreover, we noted that, althoughatorvastatin had a modest effect in increasing high-density lipoprotein cholesterol, it significantly in-creased, but did not normalize, cholesterol in largehigh-density lipoprotein particles.2 In addition, wedocumented a wide range of patient variability inresponse to statins with regard to lipid modification.2,9
For this reason, statin comparisons should be done inthe same subjects to remove genetic confounders instudying the efficacy of these statins. Other investiga-tors have examined the effects of statins on lipopro-teins, including remnants, in the fasting and postpran-dial states, but there have been no direct comparisonsof all of these statins in the same subjects with respectto hs-CRP and Lp-PLA2 in the fasting and fed states.In addition, we documented that atorvastatin is moreeffective than other statins tested (fluvastatin, lova-statin, pravastatin, and simvastatin) in normalizinglipid abnormalities in patients who have CHD.3
The focus of this examination has been on hs-CRPand Lp-PLA2 as measured in the fasted and fed statesin patients who have CHD and age- and gender-matched controls and examining the effects of atorva-statin and other statins on these parameters in patients
who have CHD. Other investigators have documentedthat hs-CRP and Lp-PLA2 are important inflammatorymarkers and markers of CHD risk.10–13 Moreover, ithas been documented that Lp-PLA2 is carried onlipoproteins, especially low-density lipoprotein.14,15 Ithas also been reported that statins can decrease thelevels of these parameters.16–18
A novel feature of our study is the finding thatfeeding a meal rich in saturated fat and cholesterolcaused significant increases of 22% in hs-CRP andsignificant decreases of 21% in Lp-PLA2 in controlsbut not in patients. Therefore, differences in hs-CRPare somewhat less in patients than in controls in thefed state versus the fasting state (median difference114% vs 168%), but differences were greater for Lp-PLA2 (mean difference 62% vs 23%). These findingsneed to be confirmed in larger studies. We have alsodocumented that atorvastatin results in similar de-creases in hs-CRP (34% and 29%) and Lp-PLA2(23% and 25%) in patients who have CHD in thefasting and fed states.
We have also documented that atorvastatin is muchmore effective than other statins tested (fluvastatin, lo-vastatin, pravastatin, and simvastatin) in decreasing notonly lipid parameters in the fasting and fed states but alsohs-CRP and Lp-PLA2. Other statins did not significantlydecrease hs-CRP levels in our study in contrast to someother studies.19 These differences may be related to dif-ferences in sample size, duration of therapy, and baselinecharacteristics of the study subjects. A 17% decrease inLp-PLA2 levels was reported with pravastatin in theWest of Scotland Study, which was greater than what weobserved.15 A recent intravascular ultrasound study that
TABLE 4 Percentage Change in Fasting Lipoprotein-associated Phospholipase A2 Concentrations From Placebo After 40 mg/dayTreament With Fluvastatin, Lovastatin, Pravastatin, or Simvastatin Versus Atorvastatin
Normally distributed values are reported as mean � SDs; other values are reported as median and interquartile range. *Values for hs-CRP were log transformedbefore statistical analysis.
All p values were derived from unpaired and paired 2-tailed t-test analyses. †p �0.05; ‡p �0.01; §p �0.001; �p �0.0001 for paired comparisons of treatmentwith placebo; ¶p �0.05; #p �0.01; **p �0.001; ††p �0.0001 for paired comparisons between same doses of statins. ‡‡p �0.05; §§p �0.01; ��p �0.001; ¶¶p�0.0001 for unpaired comparisons of controls with placebo; the data sets comprised 1 woman, 25 men (F), 2 women, 18 men (L), 6 women, 11 men (P), 2 women,14 men (S); mean � SD% or median percent change from placebo.
Abbreviation as in Table 1.
CORONARY ARTERY DISEASE/DIRECT COMPARISONS OF STATIN EFFECTS ON CRP AND LP-PLA2 1031
involved �500 evaluable patients who took atorvastatin(80 mg/day) for 18 months reported that hs-CRP de-creased by 36% versus baseline, whereas pravastatindecreased this parameter by only 5%, which is consistentwith our data.20
We tested for correlations of atorvastatin-inducedchanges in these parameters, with changes in lipid andother parameters. Changes in Lp-PLA2 were corre-lated with changes in total cholesterol, low-densitylipoprotein cholesterol, and the total cholesterol/high-density lipoprotein cholesterol ratio, consistent withLp-PLA2 being carried on low-density lipoprotein,and these particles were substantially decreased byatorvastatin. For changes in hs-CRP, no significantcorrelations with atorvastatin-induced lipid changeswere noted. In our data, hs-CRP increased with fatfeeding as triglycerides and remnants increased incontrols but not in patients.
The mechanisms responsible for these interactionsrequire further elucidation.21–25 The focus of currentguidelines for lipid modification in patients who haveCHD is for drastic decreases in low-density lipopro-tein cholesterol.26 In our view, in the future, the em-phasis may shift toward optimizing other lipoproteinparticles, namely remnants, high-density lipoprotein,and lipoprotein(a), improving the ratio of total choles-terol to high-density lipoprotein cholesterol, and de-creasing levels of inflammatory markers, such as hs-CRP and LpPLA2, to decrease the risk for CHD.
1. Schaefer EJ, Audelin MC, McNamara JR, Shah PK, Tayler T, Daly JA,Augustin JL, Seman LJ, Rubenstein JJ. Comparison of fasting and postprandialplasma lipoproteins in subjects with and without coronary heart disease. Am JCardiol 2001;88:1129–1133.2. Schaefer EJ, McNamara JR, Tayler T, Daly JA, Gleason JA, Seman LJ, FerrariA, Rubenstein JJ. Effects of atorvastatin on fasting and postprandial lipoproteinsubclasses in coronary disease patients versus control subjects. Am J Cardiol2002;90:689–696.3. Schaefer EJ, McNamara JR, Tayler T, Daly JA, Gleason JA, Seman LJ, FerrariA, Rubenstein JJ. Direct comparisons of statin effects on fasting and postprandiallipoproteins in coronary heart disease patients versus control subjects. Am JCardiol 2004;93:31–39.4. McNamara JR, Schaefer EJ. Automated enzymatic standardized lipid analysesfor plasma and lipoprotein fractions. Clin Chim Acta 1987;166:1–8.5. Jenner JL, Ordovas JM, Lamon-Fava S, Schaefer MM, Wilson PW, Castelli WP,Schaefer EJ. Effects of age, sex, and menopausal status on plasma lipoprotein(a)levels. The Framingham Offspring Study. Circulation 1993;87:1135–1141.6. Seman LJ, Jenner JL, McNamara JR, Schaefer EJ. Quantitation of plasmalipoprotein(a) by cholesterol assay of lectin-bound lipoprotein(a). Clin Chem1994;40:400–403.7. McNamara JR, Shah PK, Nakajima K, Cupples LA, Wilson PWF, OrdovasJM, Schaefer EJ. Remnant lipoprotein cholesterol and triglyceride: referenceranges from the Framingham Heart Study. Clin Chem 1998;44:1224–1232.8. Otvos JD, Jeyarajah EJ, Bennett DW, Krauss RM. Development of a protonnuclear magnetic resonance spectroscopic method for determining plasma li-poprotein concentrations and subspecies distributions from a single, rapid mea-surement. Clin Chem 1992;38:1632–1638.
9. Pedro-Botet J, Schaefer EJ, Bakker-Arkema RG, Black DM, Stein FM, CorellaD, Ordovas JM. Apolipoprotein E genotype affects plasma lipid response toatorvastatin in a gender-specific manner. Atherosclerosis 2001;158:183–194.10. Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Comparison of C-reactiveprotein and low-density lipoprotein cholesterol levels in the prediction of firstcardiovascular events. N Engl J Med 2002;347:1557–1565.11. Sattar N, Gaw A, Scherbakova O, Ford I, O’Reilly DS, Haffner SM, Isles C,Macfarlane PW, Packard CJ, Cobbe SM, Shepherd J. Metabolic syndrome withand without C-reactive protein as a predictor of coronary heart disease anddiabetes in the West of Scotland Coronary Prevention Study. Circulation 2003;108:414–419.12. Packard CJ, O’Reilly DS, Caslake MJ, McMahon AD, Ford I, Cooney J,Macphee CH, Suckling KE, Krishna M, Wilkinson FE, et al. Lipoprotein-associated phospholipase A2 as an independent predictor of coronary heartdisease. West of Scotland Coronary Prevention Study Group. N Engl J Med2000;343:1148–1155.13. Ballantyne CM, Hoogeveen RC, Bang H, Coresh J, Folsom AR, Heiss G,Sharrett AR. Lipoprotein-associated phospholipase A2, high-sensitivity C-reac-tive protein, and risk for incident coronary heart disease in middle-aged men andwomen in the Atherosclerosis Risk in Communities (ARIC) study. Circulation2004;109:837–842.14. MacPhee CH, Moores KE, Boyd HF, Dhanak D, Ife RJ, Leach CA, LeakeDS, Milliner KJ, Patterson RA, Suckling KE, et al. Lipoprotein-associatedphospholipase A2, platelet-activating factor acetylhydrolase, generates two bio-active products during the oxidation of low-density lipoprotein: use of a novelinhibitor. Biochem J 1999;338:479–487.15. Caslake MJ, Packard CJ. Lipoprotein-associated phospholipase A2 (platelet-activating factor acetylhydrolase) and cardiovascular disease. Curr Opin Lipidol2003;14:347–352.16. Horne BD, Muhlestein JB, Carlquist JF, Bair TL, Madsen TE, Hart NI,Anderson JL, Intermountain Heart Collaborative (IHC) Study Group. Statintherapy interacts with cytomegalovirus seropositivity and high C-reactive proteinin reducing mortality among patients with angiographically significant coronarydisease. Circulation 2003;107:258–263.17. Marz W, Winkler K, Nauck M, Bohm BO, Winkelmann BR. Effects of statinson C-reactive protein and interleukin-6 (the Ludwigshafen Risk and Cardiovas-cular Health study). Am J Cardiol 2003;92:305–308.18. Wang TD, Chen WJ, Lin JW, Cheng CC, Chen MF, Lee YT. Efficacy offenofibrate and simvastatin on endothelial function and inflammatory markers inpatients with combined hyperlipidemia: relations with baseline lipid profiles.Atherosclerosis 2003;170:315–323.19. Jialal I, Stein D, Balis D, Grundy SM, Adams-Huet B, Devaraj S. Effect ofhydroxymethyl glutaryl coenzyme a reductase inhibitor therapy on high sensitiveC-reactive protein levels. Circulation 2001;103:1933–1935.20. Nissen SE, Tuzcu EM, Schoenhagan P, Brown BG, Ganz P, Vogel RA,Crowe T, Howard G, Cooper CJ, Brodie B, et al. Effect of intensive comparedwith moderate lipid-lowering therapy on progression of coronary atherosclerosis:a randomized controlled trial. JAMA 2004;291:1071–1080.21. Koenig W, Lowel H, Baumert J, Meisinger C. C-reactive protein modulatesrisk prediction based on the Framingham score: implications for future riskassessment: results from a large cohort study in southern Germany. Circulation2004;109:1349–1353.22. Saito M, Ishimitsu T, Minami J, Ono H, Ohrui M, Matsuoka H. Relations ofplasma high-sensitivity C-reactive protein to traditional cardiovascular risk fac-tors. Atherosclerosis 2003;167:73–79.23. Luc G, Bard JM, Juhan-Vague I, Ferrieres J, Evans A, Amouyel P, ArveilerD, Fruchart JC, Ducimetiere P, PRIME Study Group. C-reactive protein, inter-leukin-6, and fibrinogen as predictors of coronary heart disease: the PRIMEStudy. Arterioscler Thromb Vasc Biol 2003;23:1255–1261.24. van der Meer IM, de Maat MP, Kiliaan AJ, van der Kuip DA, Hofman A,Witteman JC. The value of C-reactive protein in cardiovascular risk prediction:the Rotterdam Study. Arch Intern Med 2003;163:1323–1328.25. St-Pierre AC, Bergeron J, Pirro M, Cantin B, Dagenais GR, Despres JP,Lamarche B. Quebec Cardiovascular Study. Effect of plasma C-reactive proteinlevels in modulating the risk of coronary heart disease associated with small,dense, low-density lipoproteins in men (The Quebec Cardiovascular Study). Am JCardiol 2003;91:555–558.26. Grundy SM, Cleeman JI, Merz CN, Brewer HB Jr, Clark LT, HunninghakeDB, Pasternak RC, Smith SC Jr, Stone NJ. National Heart, Lung, and BloodInstitute; American College of Cardiology Foundation; American Heart Associ-ation. Implications of recent clinical trials for the National Cholesterol EducationProgram Adult Treatment Panel III guidelines. Circulation 2004;110:227–239.
1032 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Usefulness of Real-Time MyocardialPerfusion Imaging to Evaluate TissueLevel Reperfusion in Patients With
Raffi Bekeredjian, MD, Alexander Hansen, MD, Stefan E. Hardt, MD,Christiane Selter, RN, Roger Kranzhoefer, MD, Hugo Katus, MD,
and Helmut Kuecherer, MD
Microvascular integrity is a prequisite for functional re-covery in patients who have myocardial infarction afterrecanalization of the infarct-related coronary artery. Inthis study, we investigated whether impaired myocar-dial perfusion is present in patients who have non–ST-elevation myocardial infarction and whether the extentand time course of myocardial tissue reperfusion asassessed by myocardial contrast echocardiography(MCE) are related to functional recovery. Consecutivepatients (n � 32) who presented with a first non–ST-elevation myocardial infarction were included in ourstudy. MCE was performed on admission, 1 to 4 hoursafter angioplasty, and at 24 hours, 4 days, and 4 weeksof follow-up. Contrast images were analyzed visuallyand quantitatively. Myocardial blood flow was esti-mated by calculating the product of peak signal intensityand the slope of signal intensity increase. Improvementof wall motion on follow-up echocardiograms after 4weeks served as a reference for functional recovery ofimpaired left ventricular function. Of 496 segmentsavailable for analysis, 128 (26%) were initially dys-
functional and 96 (75%) recovered at 4 weeks offollow-up. Myocardial tissue reperfusion occurredgradually, expanding over the first 24 hours after per-cutaneous coronary intervention (myocardial bloodflow of 0.4 � 0.3 initially, 0.6 � 0.4 at 24 hours, and1.6 � 0.7 dB/s at 4 weeks of follow-up, p <0.001).Extent of tissue reperfusion was closely related tograde of improvement of global ejection fraction (r2 �0.76, p <0.001). MCE predicted functional recoverywith a sensitivity of 81%, a specificity of 88%, andaccuracy of 83% on a segmental level. Thus, impairedmicrovascular integrity is suggested by MCE in pat-ients who present with non–ST-elevation myocardialinfarction. Improvement of regional tissue perfusionafter revascularization is closely related to functionalrecovery. This information may aid risk stratificationand allow monitoring of the effectiveness of reperfu-sion therapy in these patients. �2005 by ExcerptaMedica Inc.
(Am J Cardiol 2005;95:1033–1038)
Treatment of patients who have acute coronary syn-dromes aims at preservation of flow in the infarct-
related artery and the corresponding myocardial tis-sue.1 However, restoration of epicardial blood flowmay not necessarily guarantee functional recoverywhen impaired microvascular integrity is present.1,2
Distal embolization of plaque material3 and vascularreperfusion injury4 may compromise myocardial per-fusion. Extent of tissue reperfusion is related to func-tional recovery and is associated with clinical out-comes.5 Impaired microvascular integrity (no reflow)has been observed by myocardial contrast echocardi-ography (MCE), nuclear scintigraphy, and magneticresonance imaging in patients who have ST-elevationmyocardial infarction (STEMI).6–9 However, there areno data about the functional implications of this
phenomenon in patients who have non-STEMI(NSTEMI). Further, most previous studies have used acategorical definition of microvascular integrity. Be-cause no-reflow is a complex phenomenon that canexpand over time,9 serial estimation of tissue levelreperfusion is necessary. We and others have shownthat MCE can be used to assess microvascular integ-rity.10–12 In this study, we investigated whether im-paired myocardial perfusion is present in patients whohave NSTEMI and whether the extent and the timecourse of tissue reperfusion are related to functionalrecovery.
METHODSPatient population: In this prospective study, we
included consecutive patients who presented with afirst NSTEMI. Inclusion criteria were chest pain last-ing �20 minutes compatible with myocardial isch-emia �12 hours before presentation and high levels oftroponin T (�0.03 �g/L) at initial presentation or at 4to 8 hours of follow-up. Exclusion criteria were STelevation and electrocardiographic signs or history of
From the Department of Cardiology, University of Heidelberg, Heidelberg,Germany. Manuscript received October 6, 2004; revised manuscriptreceived and accepted December 16, 2004.
Address for reprints: Grigorios Korosoglou, MD, Department of Car-diology, Im Neuenheimer Feld 410, 69115 Heidelberg, Germany.E-mail: [email protected].
infarction. The study protocol was approved by thelocal ethics committee, and all patients gave writteninformed consent.
Two-dimensional echocardiography: Imaging wasperformed from standard apical 2-, 3-, and 4-chamberviews using an ATL HDI 5000 system (Philips Med-ical System, Bothell, Washington). To avoid bias bysimultaneous interpretation of regional wall motionand perfusion, wall motion was assessed before con-trast administration using a semi-quantitative gradescale of 1 to represent normal wall motion, 2 torepresent hypokinesia, 3 to represent akinesia, and 4 torepresent dyskinesia.12,13 Wall motion and myocardialopacification were assessed initially, at 1 to 4 hoursafter angioplasty, at 24 hours, at 4 days, and at 4weeks of follow-up. Recovery of regional contractilefunction was defined as an improvement of �1 gradein wall motion at 4-week follow-up.12,14 The wallmotion score was serially calculated for initially dys-functional segments, and ejection fraction was seriallymeasured using the biplane Simpson’s method.13 Be-cause the mean � SD of differences in ejection frac-tion judged by 2 independent observers was 0.04points (4%), we defined a 2 SD value of 8% as arelevant improvement of ejection fraction at 4-weekfollow-up.
Myocardial contrast echocardiography: Myocardialcontrast echocardiographic perfusion imaging wasperformed as previously described11,12,15 using a lowmechanical index (0.14 to 0.18) in harmonic powerpulse inversion mode. SonoVue (Bracco, Byk-Gulden,Konstanz, Germany) was applied as a slow bolus in-jection (1.0 to 1.5 ml/bolus) to obtain optimal visual-ization of the left ventricle. When attenuation wasminimized, a brief pulse of higher mechanical indexwas transmitted to “clear” the myocardium of micro-bubbles. Returning immediately to low power imag-
ing, replenishment was visualized over 10 to 15 car-diac cycles. Images were analyzed visually andquantitatively off-line with a commercially availablesoftware tool (HDI Laboratory, Advanced Technol-ogy Laboratories, Bothell, Washington).
Visual analysis of myocardial opacification: Myocar-dial opacification was analyzed by 2 independent ob-servers. A third observer proved the variability of the2 first observers and resolved differences in opinionby consensus of all 3. Myocardial opacification wasgraded as 3 to represent homogeneous opacification, 2to represent mildly decreased or patchy opacification,1 to represent severely decreased opacification, and 0to represent absent opacification.11,12 We previouslyreported that segments with absent (score 0) and se-verely decreased (score 1) perfusion have a low prob-ability of recovering contractile function in patientswho have ischemic heart disease.12 Therefore, in thisstudy, improvement of regional perfusion in dysfunc-tional segments was defined as an increase from ab-sent (score 0) to patchy (score 2) or to normal (score3) or as an increase from severely decreased (score 1)to patchy (score 2) or to normal (score 3). An increasefrom absent (score 0) to severely decreased (score 1)was not considered an indication of improved myo-cardial perfusion.
Quantitative analysis: Regions of interest wereplaced in each segment from the epicardium to theendocardium to analyze replenishment kinetics fromend-systolic frames.16 Plots of contrast intensity ver-sus time were constructed and fit to an exponentialfunction, y � A*(1 � e��t), as described by Wei etal.17 The plateau of signal intensity (A) and the slopeof maximal intensity increase (�) were measured, andthe product of A*� was calculated to estimate myo-cardial blood flow. Perfusion defect size was mea-sured by planimetry of end-systolic frames, whenmaximal intensity was reached, usually 10 to 15 car-diac cycles after a “flash.”
Coronary angiography: Selective coronary angiog-raphy was performed �12 hours after admission. An-giograms were analyzed and quantitated by an inde-pendent cardiologist. Degree of stenosis wasexpressed as percent decrease in internal luminal di-ameter in relation to the normal reference. Coronarystenosis was defined as �75% narrowing of the ref-erence lumen diameter. Thrombolysis In MyocardialInfarction (TIMI) flow grade was assessed visually.18
Statistical analysis: Data are presented as mean �SD. Agreement between observers was assessed with �statistics.19 Inter- and intraobserver variabilities for theestimation of the myocardial blood flow (A*�) wereobtained by double-blinded observers by repeating anal-ysis of 20 representative myocardial contrast echocardio-grams. Repeated measures analysis of variance withBonferroni’s adjustment for multiple comparisons wasused to compare wall motion and perfusion parameters.Temporal changes in perfusion defect size were com-pared with ejection fraction changes using linear regres-sion analysis. Statistical significance of differences indiagnostic value was evaluated by McNemar’s chi-square
TABLE 1 Demographic and Clinical Characteristics (n � 32)
Age (yrs) 65 � 12Men 22Interval between onset of chest pain and
admission (min)312 � 220
TIMI flow grade 0 before PCI 3TIMI flow grade 1 or 2 before PCI 10TIMI flow grade 3 before PCI 19Multivessel CAD 17Advanced age (men �45 yrs old, women
�55 yrs old)30
Hypertension 21Elevated cholesterol level (LDL �130 mg/dl) 20Family history of CAD 16Tobacco use 20Diabetes mellitus 9Peak CK (U/L) 377 � 61296-h troponin T level (�g/L) 0.87 � 1.20Antihypertensive therapy 13Diuretic therapy 11Aspirin therapy 7
Values are numbers of patients or mean � SD, unless otherwise indicated.CAD � coronary artery disease; CK � creatine kinase; LDL � low-density
1034 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
test.11,12,15 Differences were considered significant at a pvalue �0.05.
RESULTSClinical, angiographic, and safety data: Forty con-
secutive patients fulfilled the inclusion criteria, but 8were excluded from analysis for the following reasons: 4patients were scheduled for urgent coronary artery by-pass grafting so that serial myocardial contrast echocar-diograms were missed, and perfusion defects could notbe detected by MCE in 4 patients, including 2 who hadextremely poor echocardiographic windows. Of the 32remaining patients who serially underwent myocardialcontrast echocardiographic studies, 19 showed high-grade flow-limiting stenotic lesions (TIMI flow 1 or 2)and 10 showed non–flow-limiting stenotic lesions(�75%) before angioplasty. Three patients showed com-
plete occlusion of the infarct-relatedartery (left circumflex coronary arteryin 2 patients and right coronary arteryin 1 patient), resulting in TIMI flowgrade 0 before angioplasty. All 32 pa-tients underwent successful mechani-cal reperfusion, including stent place-ment, achieving TIMI grade flow 3and �50% residual stenosis. TIMIflow before angioplasty was not signif-icantly related to initial perfusion de-fect size by MCE, to peak creatinekinase values, or to follow-up ejectionfraction (r2 �0.1, p � NS). Twenty of32 patients had high levels of troponinT on admission (0.14 � 0.26 �g/L),and all patients had high levels of tro-ponin T at follow-up (0.87 � 1.2 �g/Lat 96 hours after admission). Clinical,biochemical, and angiographic dataare listed in Table 1. No effects ofSonoVue were noted on rhythmand blood pressure, and no allergicreactions were observed during and30 minutes after contrast agentadministration.
Temporal course of myocardialperfusion and wall motion: Wall mo-tion analysis was feasible in 496 of512 segments (97%), and abnormalwall motion was detected initially in128 of 496 segments (26%), includ-ing 83 hypokinetic and 45 akineticand dyskinetic segments. Wall mo-tion score improved initially at 4days after angioplasty and showedfurther improvement at 4-week fol-low-up (Figure 1). Similarly, globalejection fraction started to recover at4 days and further improved at4-week follow-up (Figure 1). Con-versely, perfusion score improved insegments with functional recovery at24 hours and showed a further in-
crease at 4 days and at 4-week follow-up (Figure 1).Thus, recovery of wall motion at rest followed myo-cardial tissue reperfusion. Improved perfusion andlack of wall motion recovery, a condition suggestiveof myocardial “stunning,” was seen in 62 of 128initially dysfunctional segments (48%) at 24-hour fol-low-up and in 52 of 128 segments (41%) at 4-dayfollow-up (Table 2).
Visual assessment of perfusion was feasible in 476of 512 segments (93%), and quantification was feasi-ble in 449 segments (87%). Estimation of myocardialblood flow showed similar temporal patterns with thevisually assessed perfusion score. Thus, A*� signifi-cantly increased at 24 hours and showed further im-provement at 4-day and 4-week follow-up, reachingvalues similar to those of segments with normal wallmotion (Figure 2). Conversely, A*� remained un-changed in segments without functional recovery.
FIGURE 1. (A) Wall motion started to improve at 4-day follow-up in segments thatshowed tissue reperfusion. (B) Similarly, global ejection fraction initially occurred 4days after angioplasty. (C) Improvement of myocardial perfusion occurred sooner, at24 hours, and improved more at 4-day and 4-week follow-up. *p <0.05, initiallyversus 24 hours; #p <0.05, 24 hours versus 4 days; ‡p <0.05, 4 days versus 4weeks. PCI � percutaneous coronary intervention.
CORONARY ARTERY DISEASE/MCE IN NON–ST-ELEVATION MYOCARDIAL INFARCTION 1035
Prediction of regional functional recovery: Becausemyocardial tissue reperfusion occurred continuously,expanding over the first 24 hours, MCE at 4 days wasmore sensitive (81% vs 60%, p �0.05) and accurate
(83% vs 63%, p �0.05) for prediction of functionalrecovery compared with MCE at 24 hours (Table 3).Figure 3 shows a patient whose reperfusion spontane-ously occurred 24 hours to 4 days after angioplasty.Thus, correct prediction of functional recovery wasachieved by MCE at 4-day follow-up.
Analysis by patients: Change of perfusion defectarea at 4 days correlated closely (r2 � 0.76, p�0.0001) with changes in ejection fraction after 4weeks (Figure 4). Twenty-four patients showed anincreased ejection fraction of �8% at 4 weeks (42 �5% initially vs 58 � 8% at 4 weeks, p �0.001). Inthese patients, perfusion defect size significantly de-creased at 24 hours and showed further decreases at 4days and at 4 weeks (Figure 2). In patients who did notshow increased ejection fraction (n � 8), perfusiondefect size remained unchanged.
Inter- and intraobserver variabilities: Agreementsbetween observers were 92% (� � 0.81) for interpret-ing myocardial contrast echocardiograms and 89% (�� 0.78) for interpreting wall motion. Intra- and inter-observer variabilities were 13% and 17%, respec-tively, for assessment of A*� and 12% and 7%, re-spectively, for planimetric quantification of theperfusion defect area.
DISCUSSIONThis study demonstrates that impaired microvascu-
lar integrity is present in patients who have NSTEMIand that its extent and temporal course are related torecovery of wall motion. Real-time MCE can accu-rately predict functional recovery in these patients.
Temporal course of myocardial reperfusion in NSTEMI:Most patients (60% to 80%) who present with acutecoronary syndromes in the emergency departmentshow no ST elevation on their electrocardiogram.20 Inthis study, we have described for the first time thetemporal course of myocardial reperfusion in patientswho have NSTEMI and undergo successful angio-plasty. Myocardial reperfusion is a complex phenom-enon that expands over the first 24 hours after revas-
FIGURE 2. (A) Myocardial blood flow significantly increased at24 hours and showed a further improvement at 4 days and at 4weeks in segments that showed functional recovery. (B) Simi-larly, perfusion defect area significantly decreased at 24 hoursand showed a further decrease at 4 days and at 4 weeks in pa-tients who showed improvement of global ejection fraction(>8%) at 4-week follow-up. *p <0.05, initially versus 24 hours;#p <0.05, 24 hours versus 4 days; ‡p <0.05, 4 days versus 4weeks. Abbreviation as in Figure 1.
TABLE 2 Temporal Course of Myocardial Wall Motion and Perfusion in 128 Initially Dysfunctional Segments by Visual MyocardialContrast Echocardiography
WM � wall motion. Other abbreviation as in Table 1.
TABLE 3 Detection of Functional Recovery by Real-time Myocardial Contrast Echocardiography
Sensitivity Specificity PPV NPV Accuracy
MCE at 24 hours 60% 69% 85% 37% 63%MCE at 4 days 81%* 88% 95% 61%* 83%*
*p �0.05 for MCE at 24 hours versus 4 days.NPV � negative predictive value; PPV � positive predictive value.
1036 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
cularization, and its temporal patterns are closelyrelated to functional recovery. Galiuto et al21 andBronchet et al22 reported that reperfusion can sponta-neously occur between 24 hours and 6 to 9 days afterSTEMI. In accord with these results, in the presentstudy, myocardial perfusion improved spontaneouslyin 30% of dysfunctional segments between 24 hoursand 4 weeks. Quantitative estimation of myocardialblood flow confirmed visual findings, showing im-provement (1.5-fold to 2-fold) after 24 hours and afurther increase (3-fold) after 4 days. These findingsmay be explained by coronary spasm of the infarct-related artery or through microvascular embolizationafter successful angioplasty that resolves in laterstages of reperfusion.23 Because myocardial tissuereperfusion was not limited to the first 24 hours,prediction of functional recovery was more accurateby performance of MCE at 4-day follow-up. We pre-viously reported that myocardial perfusion imagingcan detect myocardial viability in patients who have
stable ischemic heart disease.12 Inthis study, MCE demonstratedgood accuracy for prediction offunctional recovery in patients whohave acute NSTEMI.
The phenomenon of myocardialstunning, defined as the presence ofpostischemic dysfunction withoutnecrosis, is well established in thesetting of acute infarction.5,6 In thisstudy, a considerable number of seg-ments (41% to 48%) demonstratedimproved perfusion but abnormalwall motion at 24-hour and 4-dayfollow-up, a condition compatiblewith stunning. On a patient level,contractile function started improv-ing after 4 days, whereas myocardialtissue reperfusion was seen 24 hoursafter angioplasty. Thus, these dataprove the principle of myocardialstunning in patients who haveNSTEMI and suggest that MCE canaccurately identify stunned tissuewith high probability for functionalrecovery.
Previous studies that investigatedmyocardial reperfusion after myo-cardial infarction focused on predic-tion of functional recovery on a seg-mental level.5,6,24 However, becauseresidual left ventricular function af-ter infarction is a primary determi-nant of long-term survival, predic-tion of global improvement is moreimportant for patient risk stratifica-tion. In this study, we have shownfor the first time that temporalchanges of myocardial reperfusion inpatients who have NSTEMI can beassessed quantitatively by real-timeMCE and are closely related to ex-
tent of improvement in global ejection fraction.Study limitations: The number of patients studied
was relative small. Because most patients studiedshowed functional recovery at 4-week follow-up, thesensitivity of MCE may have been overestimated.However, the aim of this study was to investigatefunctional recovery, which is more probable in pa-tients who have NSTEMI than in those who haveSTEMI. We used slow bolus injections for adminis-tration of contrast agent, where delivery of micro-bubbles is not constant and may confound calculationsof the peak rate of increase in contrast intensity. Theuse of continuous infusion may have provided moreaccurate results for estimation of myocardial bloodflow. However, we and others have shown that calcu-lation of replenishment kinetics is still feasible byslow bolus injections during optimal myocardialopacification,12,15,25 which may, therefore, represent amore practical approach in routine clinical settings.
FIGURE 3. Myocardial reperfusion of the apex occurred spontaneously in a pa-tient who had acute anterior NSTEMI 24 hours to 4 days after revascularizationof the left anterior descending artery. Assessment of myocardial perfusion 4 daysafter revascularization correctly predicted functional recovery at 4-week follow-up.
CORONARY ARTERY DISEASE/MCE IN NON–ST-ELEVATION MYOCARDIAL INFARCTION 1037
1. Anderson JL, Karagounis LA, Becker LC, Sorensen SG, Menlove RL. TIMIperfusion grade 3 but not grade 2 results in improved outcome after thrombolysisfor myocardial infarction. Ventriculographic, enzymatic, and electrocardio-graphic evidence from the TEAM-3 Study. Circulation 1993;87:1829–1839.2. Lepper W, Hoffmann R, Kamp O, Franke A, de Cock CC, Kuhl HP, SieswerdaGT, Dahl J, Janssens U, Voci P, et al. Assessment of myocardial reperfusion byintravenous myocardial contrast echocardiography and coronary flow reserveafter primary percutaneous transluminal coronary angioplasty (correction ofangiography) in patients with acute myocardial infarction. Circulation 200023;101:2368–2374.3. Davies MJ. A macro and micro view of coronary vascular insult in ischemicheart disease. Circulation 1990;82:38–46.4. Neumann FJ, Ott I, Gawaz M, Richardt G, Holzapfel H, Jochum M, SchomigA. Cardiac release of cytokines and inflammatory responses in acute myocardialinfarction. Circulation 1995;92:748–755.5. Ito H, Maruyama A, Iwakura K, Takiuchi S, Masuyama T, Hori M, HigashinoY, Fujii K, Minamino T. Clinical implications of the ‘no reflow’ phenomenon. Apredictor of complications and left ventricular remodeling in reperfused anteriorwall myocardial infarction. Circulation 1996;93:223–238.6. Ito H, Tomooka T, Sakai N, Yu H, Higashino Y, Fujii K, Masuyama T,Kitabatake A, Minamino T. Lack of myocardial perfusion immediately aftersuccessful thrombolysis. A predictor of poor recovery of left ventricular functionin anterior myocardial infarction. Circulation 1992;1699–1705.7. Porter TR, Li S, Oster R, Deligonul U. The clinical implications of no reflowdemonstrated with intravenous perfluorocarbon containing microbubbles follow-
ing restoration of Thrombolysis In Myocardial Infarction (TIMI) 3 flow inpatients with acute myocardial infarction. Am J Cardiol 1998;82:1173–1177.8. Schofer J, Montz R, Mathey DG. Scintigraphic evidence of the “no reflow”phenomenon in human beings after coronary thrombolysis. J Am Coll Cardiol1985;5:593–598.9. Wu KC, Kim RJ, Bluemke DA, Rochitte CE, Zerhouni EA, Becker LC, LimaJA. Quantification and time course of microvascular obstruction by contrast-enhanced echocardiography and magnetic resonance imaging following acutemyocardial infarction and reperfusion. J Am Coll Cardiol 1998;32:1756–1764.10. Kaul S, Senior R, Dittrich H, Raval U, Khattar R, Lahiri A. Detection ofcoronary artery disease with myocardial contrast echocardiography: comparisonwith 99mTc-sestamibi single-photon emission computed tomography. Circulation1997;96:785–792.11. Korosoglou G, Da Silva KG, Labadze N, Dubart AE, Hansen A, RosenbergM, Zehelein J, Kuecherer H. Real-time myocardial contrast echocardiography forpharmacological stress testing. Is quantitative estimation of myocardial bloodflow reserve necessary? J Am Soc Echocardiogr 2004;17:1–9.12. Korosoglou G, Hansen A, Hoffend J, Gavrilovic G, David W, Zehelein J,Haberkorn U, Kuecherer H. Comparison of real-time myocardial contrast echo-cardiography for the assessment of myocardial viability with 18fluordeoxyglucosepositron emission tomography and dobutamine stress echocardiography. Am JCardiol 2004;94:570–576.13. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, FeigenbaumH, Gutgesell H, Reichek N, Sahn D, Schnittger I. Recommendations for quanti-fication of the left ventricle by two-dimensional echocardiography. J Am SocEchocardiogr 1989;2:358–367.14. Pagano D, Bonser RS, Townend JN, Ordoubadi F, Lorenzoni R, Camici PG.Predictive value of dobutamine echocardiography and positron emission tomog-raphy in identifying hibernating myocardium in patients with postischaemic heartfailure. Heart 1998;79:281–288.15. Korosoglou G, Labadze N, Hansen A, Selter C, Giannitsis E, Katus H, KuechererH. Usefulness of real-time myocardial perfusion imaging in the evaluation of patientswith first time chest pain. Am J Cardiol 2004;94:1225–1231.16. Iwanaga S, Ewing SG, Husseini WK, Hoffman JI. Changes in contractilityand afterload have only slight effects on subendocardial systolic flow impedi-ment. Am J Physiol 1995;269:H1202–1212.17. Wei K, Jayaweera AR, Firoozan S, Linka A, Skyba DM, Kaul S. Quantificationof myocardial blood flow with ultrasound-induced destruction of microbubblesadministered as a constant venous infusion. Circulation 1998;97:473–483.18. Gibson CM, Cannon CP, Murphy SA, Marble SJ, Barron HV, Braunwald E.TIMI Study Group. Relationship of the TIMI myocardial perfusion grades, flowgrades, frame count, and percutaneous coronary intervention to long-term out-comes after thrombolytic administration in acute myocardial infarction. Circula-tion 2002;105:1909–1913.19. Kramer MS, Feinstein AR. Clinical biostatistics: the biostatistics of concor-dance. Clin Pharmacol Ther 1981;29:111–123.20. Steg PG, Goldberg RJ, Gore JM, Fox KA, Eagle KA, Flather MD, Sadiq I,Kasper R, Rushton-Mellor SK, Anderson FA. GRACE Investigators. Baselinecharacteristics, management practices, and in-hospital outcomes of patients hos-pitalized with acute coronary syndromes in the Global Registry of Acute Coro-nary Events (GRACE). Am J Cardiol 2002 15;90:358–363.21. Galiuto L, Lombardo A, Maseri A, Santoro L, Porto I, Cianflone D, RebuzziAG, Crea F. Temporal evolution and functional outcome of no reflow: sustainedand spontaneously reversible patterns following successful coronary recanalisa-tion. Heart 2003;89:731–737.22. Brochet E, Czitrom D, Karila-Cohen D, Seknadji P, Faraggi M, Benamer H,Aubry P, Steg PG, Assayag P. Early changes in myocardial perfusion patternsafter myocardial infarction: relation with contractile reserve and functional re-covery. J Am Coll Cardiol 1998;32:2011–2017.23. Villanueva FS, Camarano G, Ismail S, Goodman NC, Sklenar J, Kaul S.Coronary reserve abnormalities in the infarcted myocardium. Assessment ofmyocardial viability immediately versus late after reflow by contrast echocardi-ography. Circulation 1996;94:748–754.24. Kereiakes DJ. Adjunctive pharmacotherapy before percutaneous coronaryintervention in non–ST-elevation acute coronary syndromes: the role of modu-lating inflammation. Circulation 2003 21;108:III22–III27.25. Lindner JR, Villanueva FS, Dent JM, Wei K, Sklenar J, Kaul S. Assessmentof resting perfusion with myocardial contrast echocardiography: theoretical andpractical considerations. Am Heart J 2000;139:231–240.
FIGURE 4. (B) Decrease in perfusion defect area at 4 days corre-lated strongly with improvement of global ejection fraction at4-week follow-up (r2 � 0.76, p <0.001). (A) A weaker correla-tion (r2 � 0.22, p <0.01) was found at 24-hour follow-up.
1038 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Usefulness of Tissue Doppler Imaging inthe Diagnosis and Prognosis of Acute
Right Ventricular Infarction With InferiorWall Acute Left Ventricular Infarction
Hisham Dokainish, MD, Heather Abbey, RN, Kenneth Gin, MD,Krishnan Ramanathan, MBBS, Pui-Kee Lee, MBBS, and John Jue, MD
Tissue Doppler (TD) imaging is a novel echocardio-graphic technique that measures myocardial velocities.However, there are sparse data on TD imaging of theright ventricular (RV) free wall in the diagnosis andprognosis of RV myocardial infarction (MI) in inferiorwall left ventricular MI. Fifty patients who had left ven-tricular MI underwent TD echocardiography and an-giography within 48 hours of MI. For diagnosis, theability of RV TD imaging to detect RV MI was assessedusing coronary angiography as the reference standard.For prognosis, the ability of TD detection of RV dysfunc-tion to predict cardiac death or rehospitalization at 1year was determined. For diagnosis, the univariate pre-dictors of RV MI included RV diastolic dimension (p �0.001), TD imaging of tricuspid annular systolic velocity(p � 0.001), and early diastolic velocity (p � 0.002). On
multivariate analysis, systolic annular velocity (p �0.04) and RV dimension (p � 0.05) predicted RV MI. Forprognosis, nonculprit coronary artery disease (p �0.003), TD imaging of RV systolic annular velocity(p � 0.005), and early diastolic velocity (p � 0.01) wereamong the univariate predictors of cardiac death orrehospitalization. On multivariate analysis, nonculpritcoronary artery disease (p � 0.02) and TD imaging ofsystolic annular velocity (p � 0.04) were independentpredictors of outcome. Decreased RV systolic annularvelocity on TD images detects RV MI in first left ventric-ular acute inferior MI and predicts cardiac death orrehospitalization at 1 year. �2005 by Excerpta Med-ica Inc.
(Am J Cardiol 2005;95:1039–1042)
T issue Doppler (TD) imaging is a novel echocar-diographic technique that measures myocardial ve-
locities and has been demonstrated to be an indicatorof global left ventricular (LV) function1 and rightventricular (RV) function.2 Although there are prelim-inary data on the role of TD imaging of the RV freewall in the demonstration of RV myocardial infarction(MI) in LV MI,3–5 there are no published data on itsprognostic role. We hypothesized that TD imaging ofthe RV free wall would detect RV MI and that de-pressed TD images of RV velocities would connote aworse outcome in LV MI. Therefore, this study as-sessed the role of TD imaging of the right ventricle inthe diagnosis and prognosis of RV MI in a first LVinferior wall MI.
METHODSInclusion/exclusion criteria: Patients admitted to our
institution with first, acute, ST-elevated LV inferiorwall MI were eligible. Inclusion criteria were charac-teristic chest pain, �1-mm ST-segment elevation in 2of 3 of leads II, III, and aVF, and diagnostic angio-graphic and comprehensive echocardiographic studies
with TD imaging performed within 48 hours of symp-tom onset of MI. Exclusion criteria were refusal ofconsent, previous MI, previous heart failure, previousabnormal LV or RV function, documented significantvalve disease (any stenosis or greater than mild regur-gitation), pulmonary hypertension, atrial fibrillation,and paced rhythm.
Clinical definitions: Baseline characteristic and clin-ical data were gathered prospectively on enrolled pa-tients. Heart failure was defined as systemic conges-tion (high jugular venous pressure and peripheraledema) and/or pulmonary vascular congestion (thirdheart sound and pulmonary rales). Heart block wasdefined as Mobitz class II or higher.
Echocardiographic assessment: Echocardiographywas performed on a General Electric Vingmed SystemFive ultrasound machine (General Electric, Milwaukee,Wisconsin). All 2-dimensional and conventional Dopp-ler echocardiographic measurements were made accord-ing to guidelines of the American Society of Echocardi-ography,6 except LV ejection fraction, which was madeby visual estimation. All echocardiographic analyseswere made at a later date by a level III echocardiog-rapher who was blinded to all clinical, angiographic,and outcome data. For RV assessment by TD imaging,the sample volume was placed 1 cm above the tricus-pid annulus at the RV free wall in the apical 4-cham-ber view. Mean systolic tricuspid annular velocity andearly and late tricuspid diastolic velocities were mea-sured as centimeters per second.
From the Department of Medicine, Division of Cardiology, University ofBritish Columbia, Vancouver, British Columbia, Canada; and the BaylorCollege of Medicine, Houston, Texas. Manuscript received October21, 2004; revised manuscript received and accepted December 16,2004.
Address for reprints: Hisham Dokainish, MD, Baylor College ofMedicine, Ben Taub General Hospital, 6550 Fannin, SM1901,Houston, Texas 77030. E-mail: [email protected].
Coronary angiography: An interventional cardiolo-gist who did not perform any of the diagnostic angio-grams and was blinded to all clinical and echocardio-graphic data assessed angiograms to determine theinfarct culprit lesion and its location in relation to thefirst RV branch of the right coronary artery. Signifi-cant coronary artery disease was defined as �70%stenosis. Nonculprit coronary artery disease was de-fined as significant disease in a major coronary arteryother than the infarct-related artery.
Diagnosis of RV MI in LV MI: RV MI was determinedto have occurred angiographically if the infarct culpritlesion was proximal to the first RV branch of the rightcoronary artery.4,7–9
Prognosis of patients who had LV MI: The primaryend point was defined as cardiac death or cardiacrehospitalization at 1-year follow-up.
Statistical analysis: All data are expressed as mean� SD, except discrete variables, which are reported asnumbers and percentages. Comparisons were madewith chi-square test for categorical variables and with2-tailed Student’s t test for continuous variables. Foroutcome measurements, univariate and multivariateCox’s proportional hazards analysis were used to ad-just for variable time to event. Receiver-operatingcharacteristic curves were constructed to determinethe optimal accuracy for TD imaging variables indetecting RV MI and in predicting patient outcome. Ap value �0.05 was statistically significant. All analy-ses were performed with GB Stat 6.5 (Dynamic Mi-crosystems, Inc., London, United Kingdom) and Sig-maStat 3.0 (SPSS, Inc., Chicago, Illinois).
RESULTSSixty-one consecutive patients who had a first,
acute LV inferior wall MI were admitted to our insti-tution during the study period; 2 were excluded be-cause they did not undergo angiography, 3 were ex-cluded because of a previous infarction, and 6 wereexcluded because they did not undergo echocardiog-raphy or angiography within 48 hours of symptomonset of MI (late presenters and transfers from otherinstitutions). Thus, 50 patients were included in thestudy (43 men and 7 women).
Diagnosis of RV MI in acute LV MI: Mean time be-tween onset of chest pain and echocardiography was18 � 12 hours, and mean interval between echocar-diography and coronary angiography was 9 � 10hours. Angiographically, 22 patients (44%) had theinfarct culprit lesion proximal to the first RV branch ofthe right coronary artery (RV MI) and 28 patients(56%) had the infarct culprit lesion distal to the firstRV branch of the right coronary artery or in thecircumflex artery (no RV MI). Clinical characteristicsof the 2 groups stratified to the presence of RV MI andto outcome are presented in Table 1.
Of the conventional echocardiographic variables,RV diastolic dimension (p � 0.001) and visual assess-ment of RV function (p � 0.05) were significantunivariate predictors (Table 2). Of the TD imagingvariables, mean systolic annular velocity (7.3 � 1.8 vs9.5 � 1.8 cm/s, p � 0.001) was significantly lower inpatients who had RV involvement than in those whodid not, as was early diastolic velocity (6.1 � 1.7 vs8.1 � 1.6 cm/s, p � 0.002); late diastolic velocity wasnot significantly different between groups.
*p �0.05 on univariate analysis between patients with and without RV infarction.†p �0.05 on univariate analysis between patients with and without an event at 1 year.TIMI � Thrombolysis In Myocardial Infarction.
1040 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Multivariate regression analysis included signifi-cant univariate predictors. When early diastolic veloc-ity was excluded from the model (significant overlapwith systolic annular velocity r � 0.56, p �0.001),systolic annular velocity (p � 0.04) and RV dimen-sion (p � 0.05) were independent predictors of RVMI. When systolic annular velocity was excluded fromthe model, RV dimension was a significant independentpredictor of RV MI, and early diastolic velocity (p �0.08) demonstrated a trend toward significance.
We then assessed the ability of different values ofsystolic annular velocity to detect RV MI. Using re-ceiver-operating characteristic analysis, a systolic an-nular velocity �8 cm/s had 78% sensitivity and 86%specificity (area under the curve 0.80, p �0.001) fordetecting RV MI, whereas a systolic annular velocity�9 cm/s had 86% sensitivity and 75% specificity(area under the curve 0.78, p �0.001; Table 3). Toassess the relation between TD assessment and visualassessment of RV function, patients whose systolicannular velocity was �9 cm/s (n � 25) were catego-rized as those who had marked impairment of RVfunction (systolic annular velocity �6 cm/s) and thosewho had modest impairment (systolic annular velocity6 to 8 cm/s). Of the 10 patients who had a systolicannular velocity �6 cm/s, the right ventricle wasdepressed by visual assessment in 9 of 10 (90%,chi-square 11.3, p � 0.0008) compared with 7 of 15patients (47%, chi-square 0.1, p � 0.76) who had asystolic annular velocity of 6 to 8 cm/s.
RV TD imaging in the prognosis of LV MI: One-yearfollow-up was completed in 46 of 50 patients (92%).There were 4 cardiac deaths (3 due to pump failureand 1 due to sudden cardiac death) and 14 cardiacrehospitalizations (8 for heart failure and 6 for MI orunstable angina that required coronary intervention).Univariate Cox’s proportional hazards analysisshowed that heart failure status at presentation (chi-square 8.9, p � 0.003), nonculprit coronary arterydisease on the angiogram (chi-square 9.0, p � 0.003),and location of the infarct culprit lesion in relation tothe first RV branch of the right coronary artery (chi-square 4.1, p � 0.05) were among the clinical predic-tors of outcome (Table 1). Of the echocardiographicvariables, LV ejection fraction (chi-square 9.9, p �0.002), systolic annular velocity (chi-square 8.1, p �0.005), and early diastolic velocity (chi-square 6.0, p� 0.01) were among the significant predictors of out-come (Table 2). The RV TD velocities according toevent status are displayed in Figure 1.
Multivariate Cox’s proportional hazards analysisof the univariate predictors, after excluding early di-astolic velocity from the model (due to significantoverlap with systolic annular velocity r � 0.56,p �0.001), showed that nonculprit coronary arterydisease (p � 0.03) and systolic annular velocity (p �0.04) were significant predictors of an event at 1 year.
*p �0.05 on univariate analysis between patients with and without RV infarction.†p �0.05 on univariate analysis between patients with and without an event at 1 year.
TABLE 3 Diagnostic Accuracy of Tissue Doppler TricuspidSystolic Annular Velocity in the Diagnosis of RightVentricular Infarction
Sa Value Sensitivity (%) Specificity (%) PPV (%) NPV (%)
FIGURE 1. RV TD velocities and patient outcome. Aa � TD tricuspidannular late diastolic velocity; Ea � TD tricuspid annular early dia-stolic velocity; Sa � TD tricuspid annular systolic velocity.
CORONARY ARTERY DISEASE/TD OF RV INFARCTION WITH ACUTE LV INFARCTION 1041
Receiver-operating characteristic analysis showed that asystolic annular velocity �8 cm/s was the optimal cutoffto predict the primary end point (sensitivity 85%, spec-ificity 77%, area under the curve 0.82, p �0.001).Kaplan-Meier survival curves for systolic annular veloc-ities �8 cm/s and �8 cm/s are shown in Figure 2.
DISCUSSIONIn this study, as determined by angiographic site of
the infarct culprit lesion, TD systolic annular velocityof the RV free wall was an independent predictor ofRV MI in acute LV inferior wall MI. This study hasalso demonstrated, for the first time, that a depressedRV systolic annular velocity by TD imaging predictscardiac death or rehospitalization at 1 year.
Recent studies have assessed RV MI with TD imagesof the tricuspid annulus in the setting of LV MI anddemonstrated a decreased systolic annular velocity in theRV MI group. Alam et al3 associated depressed RV TDvelocities with ST-segment elevation in lead V4R. Oz-demir et al4 and Oguzhan et al5 reported similar resultsaccording to electrocardiographic criteria. Our study isconsistent with these investigations in the detection ofRV MI in acute LV MI, demonstrating that decreasedRV myocardial systolic velocities are associated withculprit infarcted lesions in the proximal right coronaryartery. However, this study, to our knowledge, is the firstto use coronary angiography as the reference standard forRV MI and to describe the prognostic significance ofdecreased TD myocardial velocities.
TD measurement of RV performance in the settingof acute LV MI has several putative advantages. First,TD detection of decreased myocardial velocities hasbeen demonstrated in the left ventricle to precedeglobal impairment by ejection fraction assessment.10
By corollary, depressed RV TD velocities would beexpected to precede global impairment by RV fractionalarea change or by visual assessment. In the present study,visual estimation of RV dysfunction correlated betterwith a marked decrease in systolic annular velocity (�6cm/s) than with modest decreases (6 to 8 cm/s). Thus,TD measurements may be a more sensitive indicator ofRV MI in acute LV MI than conventional echocardio-
graphic variables. Second, TD is a simple, reproduciblemeasurement that is quantitative. Further, only the lateraltricuspid annulus needs to be visualized for measurementof RV velocities by TD imaging as opposed to visualassessment of RV function or RV fractional area change,where the entire RV free wall needs to be adequatelyvisualized.11
One limitation of this study is the small sample.Nevertheless, we were able to reach several significantconclusions. Another limitation is the lack of a true“gold standard” for RV MI. Although angiographicsite of occlusion has been widely used, it remains animperfect determinant of RV MI. Ideally, all echocar-diographic studies with TD imaging would be per-formed immediately after presentation with acute MIand preferably before thrombolysis or primary angio-plasty due to the potential early recovery of RV func-tion. Because 44% of our study patients had trans-ferred from other institutions and because of thevariable times from onset of symptoms to presenta-tion, this was not possible. Previous studies havedemonstrated that RV MI in acute LV MI can increasemortality due to arrhythmias or heart failure.12,13 Inthe present study, cardiac deaths and rehospitaliza-tions due to heart failure could be attributed, at least inpart, to RV dysfunction. However, readmissions forMI or unstable angina that required revascularizationwere more likely secondary to the other independentpredictor of outcome in this study, namely significantcoronary disease in the noninfarct-related artery.
1. Gulati VK, Katz WE, Follansbee WP, Goresan J III. Mitral annular descentvelocity by tissue Doppler echocardiography as an index of global left ventricularfunction. Am J Cardiol 1996;77:979–984.2. Meluzin J, Spinarova L, Bakala J, Toman J, Krejci J, Hude P, Kara T, Soucek M.Pulsed Doppler tissue imaging of the velocity of tricuspid annular systolic motion. Anew, rapid, and non-invasive method of evaluating right ventricular systolic function.Eur Heart J 2001;22:340–348.3. Alam M, Wardell J, Andersson E, Samad BA, Nordlander R. Right ventricularfunction in patients with first inferior myocardial infarction: assessment by tricuspidannular motion and tricuspid annular velocity. Am Heart J 2000;139:710–715.4. Ozdemir K, Altunkeser BB, Icli A, Ozdil H, Gok H. New parameters inidentification of right ventricular myocardial infarction and proximal right coro-nary artery lesion. Chest 2003;124:219–226.5. Oguzhan A, Abaci A, Eryol NK, Topsakal R, Seyfeli E. Color tissue Dopplerechocardiographic evaluation of right ventricular function in patients with rightventricular infarction. Cardiology 2003;100:41–46.6. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, Feigenbaum H,Gutgesell H, Reichek N, Sahn D, Schnittger I, et al. Recommendations forquantification of the left ventricle by 2-dimensional echocardiography. J Am SocEchocardiogr 1989;2:358–367.7. Andersen RA, Falk E, Nielsen D. Right ventricular infarction: frequency, size andtopography in coronary heart disease: a prospective study comprising 107 consecu-tive autopsies from a coronary care unit. J Am Coll Cardiol 1987;10:1223–1232.8. Goldstein JA. Pathophysiology and management of right heart ischemia. J AmColl Cardiol 2002;40:841–853.9. Kosuge M, Kimura K, Ishikawa T, Hungo Y, Mochida Y, Sugiyama M,Tochikubo O. New electrocardiographic criteria for predicting the site of coro-nary artery occlusion in inferior wall myocardial infarction. Am J Cardiol1998;82:1318–1322.10. Yu C-M, Lin H, Yang H, Kong S-L, Zhang Q, Lee SW-L. Progression ofsystolic abnormalities in patients with “isolated” diastolic heart failure anddiastolic dysfunction. Circulation 2002;105:1195–1201.11. Kozakova M, Palombo C, Distante A. Right ventricular infarction: the role ofechocardiography. Echocardiography 2001;18:701–707.12. Zehender M, Kasper W, Kauder E, Schonthaler M, Geibel A, Olschewski M,Just H. Right ventricular infarction as an independent predictor of prognosis afteracute inferior myocardial infarction. N Engl J Med 1993;328:981–988.13. Mehta SR, Eikelboom JW, Natarajan MK, Diaz R, Yi C, Gibbons RJ, YusufS. Impact of right ventricular involvement on mortality and morbidity in patientswith inferior myocardial infarction. J Am Coll Cardiol 2001;37:37–43.
FIGURE 2. Event-free survival curves according to RV TD systolicvelocity. Abbreviation as in Figure 1.
1042 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Influence of Stent Length to LesionLength Ratio on Angiographic and
Clinical Outcomes After Implantationof Bare Metal and Drug-Eluting Stents
(the TAXUS-IV Study)S. Chiu Wong, MD, Mun K. Hong, MD, Stephen G. Ellis, MD, Maurice Buchbinder, MD,
Mark Reisman, MD, Augustin DeLago, MD, Mirle Kellett, MD, Jeffrey J. Popma, MD,Mary E. Russell, MD, Roxana Mehran, MD, Jeffrey W. Moses, MD, and
Gregg W. Stone, MD
Longer bare metal stent lengths have been associatedwith greater restenosis. However, the effect of the ratioof stent length to lesion length on clinical and angio-graphic restenosis after implantation of bare metal anddrug-eluting stents has not been clearly defined. Patientsin the TAXUS-IV study who underwent single-study stentplacement were categorized into tertiles based on ratiosof stent length to lesion length. Clinical results at 1 yearand angiographic outcomes at 9 months were comparedacross the 3 groups. The median ratios of stent length tolesion length were 1.20, 1.58, and 2.27 in the 3 tertiles.Analysis segment restenosis rates at 9 months weresimilar across the 3 tertiles with bare metal stents (24.7%vs 26.7% vs 23.8%, respectively, p � 0.90 for trend)and paclitaxel-eluting stents (11.7% vs 6.5% vs 5.4%,respectively, p � 0.24). Similarly, there were no differ-
ences in 1-year rates of target lesion revascularizationacross the 3 tertiles for bare metal stents (14.6% vs14.8% vs 13.7%, respectively, p � 0.91) or paclitaxel-eluting stents (6.1% vs 3.6% vs 4.0%, respectively, p �0.38). By multivariate analysis, the ratio of stent lengthto lesion length was an independent predictor of neither9-month angiographic restenosis nor 1-year target le-sion revascularization in the bare metal stent arm (oddsratio 1.21, p � 0.36, and hazard ratio 0.80, p � 0.31,respectively) or in the paclitaxel-eluting stent arm (oddsratio 0.86, p � 0.76, and hazard ratio 0.58, p � 0.21,respectively). These data do not support the arbitraryuse of larger ratios of stent length to lesion length inpatients who undergo implantation of drug-elutingstents. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1043–1048)
I t has been hypothesized that restenosis outside theconfines of the stent may be caused by incomplete
lesion coverage or by balloon injury to the stent mar-gins. Because there is little incremental restenosiswith drug-eluting stents (DESs) as a function of stentlength, it has been suggested that implantation oflonger DESs may minimize restenosis. Using theTAXUS-IV database, we examined whether implan-tation of paclitaxel-eluting stents with larger ratios ofstent length to lesion length may improve clinical andangiographic outcomes.
METHODSThe TAXUS-IV study has been described in detail
elsewhere.1 In brief, TAXUS-IV was a multicenter,prospective randomized trial that compared clinicalwith angiographic outcomes in patients who under-went implantation of the bare metal stent Express(Boston Scientific Corp., Natick, Massachusetts) orthe slow-release, polymer-based, paclitaxel-elutingstent TAXUS (Boston Scientific Corp.). Eligibilitycriteria included control of a single, nonostial de novonative coronary lesion that was visually estimated as10 to 28 mm in length with a reference vessel diameterof 2.5 to 3.75 mm. Dilatation of the target lesion witha balloon angioplasty catheter before stent placementwas mandatory. Study stents were available from 2.5to 3.5 mm in diameter and in lengths of 16, 24, and 32mm (although no stents that were 2.5 mm in diameterand 32 mm long were supplied). Stent deploymentwith a balloon-to-artery diameter ratio of 1:1 to 1.1:1,with at least 2 to 4 mm of stent beyond the margins ofthe target lesion, was recommended. Postdilatation ofthe deployed stent was left to the discretion of theinvestigator to ensure full stent to vessel apposition,with a final residual stenosis of �10% by visualestimation. Clinical follow-up was performed at 1, 4,9, and 12 months and then yearly for 5 years. The
From the Weill Medical College of Cornell University, New York,New York; the Cleveland Clinic Foundation, Cleveland, Ohio; theScripps Memorial Hospital, La Jolla, California; the Swedish MedicalCenter, Seattle, Washington; the Albany Medical Center Hospital,Albany, New York; the Maine Medical Center, Portland, Maine; theBrigham and Women’s Hospital, Boston, Massachusetts; the BostonScientific Corp., Natick, Massachusetts; and the Cardiovascular Re-search Foundation and College of Physicians and Surgeons, Colum-bia University, New York, New York. Manuscript received November3, 2004; revised manuscript received and accepted January 4,2005.
Address for reprints: S. Chiu Wong, MD, 520 East 70th Street,Starr Pavilion, 4th Floor, New York, New York 10021. E-mail:[email protected].
TABLE 1 Selected Baseline Clinical and Angiographic Characteristics Stratified by Stent Length to Lesion Length Ratio in Tertiles in the Pooled Study Population
*Median (interquartile range).ACC/AHA � American College of Cardiology/American Heart Association.
1044TH
EA
MERIC
AN
JOU
RNA
LO
FC
ARD
IOLO
GY�
VOL.
95M
AY
1,2005
clinical end points of the trial have been previouslydefined.1 At the time of this report, follow-up wascomplete to 1 year.
Quantitative coronary angiography: Angiographicfollow-up at 9 months (or sooner for recurrent symp-toms) was prespecified in a cohort of 732 patients. Allbaseline and follow-up angiograms were analyzed atan independent core laboratory as previously de-scribed.1 Lesion length was defined as the axial extentof the lesion that contained a shoulder-to-shoulderluminal decrease by �20%. A 5-mm segment of ref-erence diameter proximal and distal to the stenosiswas used to calculate the average reference vesseldiameter at baseline, after stent implantation, and atfollow-up. The reference vessel diameter and mini-mum luminal diameter were measured within the stent
(in-stent measurement), within the 5-mm proximaland distal margins outside the stent, at the worst edge(the worst result at the proximal or distal edge), overthe entire stent length, and the 5-mm proximal anddistal margins (analysis segment measurement). Fromthese measurements, diameter stenosis, acute gain,late loss, and loss index were calculated at each pe-riod, as previously described.2
Analysis plan and statistical analysis: To avoid po-tential confounding effects, we limited our presentanalysis to patients who received a single bare metalor a TAXUS stent in the target lesion. As a result, theexact implanted stent length was known. The ratio ofstent length to lesion length was then calculated usingthe known stent length implanted and the measuredbaseline lesion length. Patients were grouped into
TABLE 2 Immediate Postprocedural Angiographic Measurements in the Pooled Patient Population
CORONARY ARTERY DISEASE/IMPACT OF STENT LENGTH TO LESION LENGTH RATIO 1045
tertiles based on ratio of stent length to lesion length.Baseline clinical characteristics, procedural findings,1-year clinical outcomes, and 9-month angiographicfollow-up results were compared across the 3 groups.Continuous variables are presented as mean � SD ormedian and interquartile range and were comparedwith the nonparametric Kruskal-Wallis test. Categor-ical variables were compared with Fisher’s exact testfor pairwise comparisons or chi-square test for trendof 3-way comparisons. Survival estimates were deter-mined with Kaplan-Meier curves, and intergroupcomparisons were determined with log-rank test.Stepwise multiple logistic regression analysis was per-formed to determine the independent effect of ratio ofstent length to lesion length on 9-month analysis seg-ment binary restenosis using an � value �0.1 for entryand removal, except for ratio of stent length to lesionlength, which was forced into the model. Cox’s re-gression analysis with similar entry and removal cri-teria was used to identify predictors of 1-year targetlesion revascularization (TLR). For the 2 models, 22different clinical variables were examined, namely,age, gender, smoking status, diabetes, hypertension,hyperlipidemia, unstable angina, epicardial vessel,ejection fraction, balloon-to-artery ratio, ostial loca-tion, bend �45°, tortuosity, calcification, baseline ref-erence vessel diameter, baseline minimal luminaldiameter, in-stent acute gain, randomization arm, le-sion length, maximum device pressure, maximumstent diameter, and ratio of stent length to lesionlength. All p values are 2-sided, and an � value �0.05was considered statistically significant.
RESULTSBaseline features and procedural results: Of the
1,314 patients who were enrolled in TAXUS-IV,1,190 patients (91%) met the criteria of a single-studystent implanted and constituted the study population.Angiography at 9 months was completed in 524 of667 patients (79%) within this subset who met theenrollment criteria and were prespecified for fol-low-up angiography. The median ratios of stent lengthto lesion length were 1.20 in the shortest tertile, 1.58in the middle tertile, and 2.27 in the longest tertile(Table 1). Median lesion length was progressively
shorter across the 3 tertiles, whereasmedian stent length (16 mm) wasidentical across all groups (Table 1).Similar findings were present foreach of the 3 tertiles in the controlsand study patients. Other than diabe-tes, which was less commonlypresent in patients in the third tertile,baseline clinical characteristics werewell matched. In addition, patients inthe third tertile also had smaller ref-erence vessel diameters, less severeand less complex lesions, and weretreated with a slightly greater bal-loon-to-artery ratio (Table 1). Afterstent implantation, minimal luminaldiameter was similar in all 3 tertiles
in the analysis segment, proximal edge, and in thestent but larger at the distal edge and the worst edge intertile 1 (Table 2). Although the in-stent acute gainwas also similar in all 3 tertiles, the increment waslarger in all other measured parameters in the firsttertile. However, the analysis segment and in-stentdiameter stenosis were lowest in tertile 3. Similarfindings were present for the bare metal and paclitax-el-eluting stents.
Clinical outcomes at 1 year: As presented in Table 3,all primary clinical outcome measurements were sim-ilar in the 3 groups. Stent thrombotic rates were notincreased with larger ratios of stent length to lesionlength. Most importantly, however, no decrease in therates of TLR was present with larger ratios of stentlength to lesion length among the bare metal and DEScohorts, although the paclitaxel-eluting stent signifi-cantly decreased TLR rates compared with the controlstent at all ratios of stent length to lesion length. Bymultivariate analysis, neither lesion length nor ratio ofstent length to lesion length was an independent de-terminate of 1-year TLR in the bare metal and pacli-taxel-eluting stent cohorts (Table 4).
Angiographic results at 9 months: In patients whoreceived bare metal stents, there was no effect of ratioof stent length to lesion length on late loss, follow-updiameter stenosis, or rates of binary restenosis withinthe stent at either edge or over the entire analysissegment. In patients who received paclitaxel-elutingstents, late loss (0.21 � 0.46 vs 0.06 � 0.30 vs 0.13 �0.46 mm in tertiles 1, 2, and 3, respectively, p � 0.07)and follow-up diameter stenosis (16.2 � 16.7% vs11.2 � 10.8% vs 11.4 � 11.2% in tertiles 1, 2, and 3,respectively, p � 0.02) at the proximal edge and theworse edge (0.36 � 0.46 vs 0.21 � 0.30 vs 0.28 �0.38 mm in tertiles 1, 2, and 3, respectively, p � 0.04and 19.6 � 15.2% vs 14.8 � 10.9% vs 14.8 � 11.3%in tertiles 1, 2, and 3, respectively, p � 0.01) weresignificantly lower with larger ratios of stent length tolesion length. However, late loss and follow-up diam-eter stenosis over the entire analysis segment of thestent were not significantly influenced by ratio of stentlength to lesion length. Moreover, binary restenosiswithin the analysis segment, in the stent, and at theedges did not vary with ratio of stent length to lesion
TABLE 4 Multivariate Correlates of One-year Target Lesion Revascularization
Hazard Ratio (95% CI) p Value
All patients (n)Randomized to TAXUS 0.38 (0.22–0.64) 0.0003In-stent acute gain 0.40 (0.21–0.77) 0.006Baseline reference vessel diameter 0.55 (0.31–0.97) 0.04Stent length to lesion length ratio 0.80 (0.55–1.16) 0.24
Bare metal stent groupBaseline reference vessel diameter 0.43 (0.21–0.88) 0.02In-stent acute gain 0.33 (0.15–0.74) 0.007Stent length to lesion length ratio 0.80 (0.52–1.23) 0.31
Paclitaxel-eluting stent groupStent length to lesion length ratio 0.58 (0.24–1.38) 0.21
CI � confidence interval.
1046 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
length. Ratio of stent length to lesion length was not asignificant predictor of analysis segment angiographicrestenosis in the bare metal and paclitaxel-elutingstent groups (Table 5).
DISCUSSIONIn the era of bare metal stents, studies were con-
flicting as to whether stent length or lesion length wasa stronger correlate of restenosis.3–10 These studieswere flawed by their retrospective nature, frequentinclusion of different stent configurations and strutthicknesses, evolving stent implantation techniques,and overlapping stents. Moreover, the effect of theratio of stent length to lesion length has rarely beenexamined. Goldberg et al11 retrospectively examined456 patients who received different designs of baremetal stent and associated a larger ratio of stent lengthto lesion length with more aggressive angiographicrestenosis. The optimal ratio of DES to lesion lengthhas not been investigated, although it has been hy-pothesized that use of longer DESs may overcomefocal edge recurrence and improve clinical and angio-graphic outcomes.
In the present study, no significant relation wasfound between ratio of stent length to lesion lengthand clinical or angiographic outcome. Further, byrestricting the analysis to a single stent in our presentanalysis, the exact stent length was known and theconfounding effects of multiple overlapping stentswere avoided. After implantation of bare metal andpaclitaxel-eluting stents, patients in whom stent lengthwas closely matched to the lesion (tertile 1) had sim-ilar long-term clinical outcomes as patients in whomthe stent was �2 times as long as the measured lesion(tertile 3). Slight advantages were noted in terms ofless angiographic late loss and a decreased follow-updiameter stenosis at the edges in patients who receivedthe paclitaxel-eluting stent in whom the ratios of stentlength to lesion length ratios were larger (these find-ings were not present in patients who received baremetal stents). However, these benefits were not ofsufficient magnitude to translate into improvements
over the entire analysis segment or inbinary restenotic rates, which are theprimary determinates of recurrentangina and the need for repeat TLR.
The ideal clinical trial designed toassess the true effect of the ratio ofstent length to lesion length wouldrandomly assign patients who hadsimilar lesion lengths to receive dif-ferent stent lengths. In the presentretrospective analysis of a prospec-tive dataset, the converse actuallyoccurred: stent length was constantacross the 3 studied groups, whereaslesion length was progressively shorter.Moreover, fewer patients in the thirdtertile with the largest ratios of stentlength to lesion length had diabetes,which may have predisposed to lessrestenosis. Conversely, the reference
diameters were smaller in this group, an offsettingfactor that is associated with increased restenosis.When the effects of these and other covariates wereexamined in multivariate models, ratio of stent lengthto lesion length was not a predictor of clinical reste-nosis or angiographic restenosis in the bare metalgroup or paclitaxel-eluting group. However, lesionlength was a strong predictor of angiographic reste-nosis with bare metal stents and a weak correlate withpaclitaxel-eluting stents. Differences in safety or effi-cacy across the different ratios of stent length to lesionlength might have become evident had more patientsbeen studied. Because only the TAXUS stent wasevaluated in the present study, our conclusions do notnecessarily apply to other DESs, which may havedifferent release kinetics, drug diffusion characteris-tics, and delivery systems.
In the TAXUS-IV trial, operators were instructedto choose stent lengths to cover the lesion by 2 to 4mm at each margin. This visually guided strategyresulted in a very low rate of 3.0% for TLR at 9months. Capriciously increasing the length of stentsbeyond that used in the present study carries thetheoretical risk of greater periprocedural myocardialinfarction (from distal embolization or side branchloss) and stent thrombosis. Thus, on the basis of thepresent analysis, arbitrarily larger ratios of stent lengthto lesion length than required to cover the lesion witha several-millimeter margin at each edge cannot becurrently recommended. Unanswered by this study isthe optimal DES length strategy to manage diffuselydiseased vessels, an important issue that requires fur-ther study as these potent bioactive devices are appliedin more and more complex anatomic scenarios.
1. Stone GW, Ellis SG, Cox DA, Hermiller J, O‘Shaughnessy C, Mann JT, TurcoM, Caputo R, Bergin P, Greenberg J, et al, the TAXUS-IV Investigators. Apolymer-based paclitaxel-eluting stent in patients with coronary artery disease.N Engl J Med 2004;350:221–231.2. Lansky A, Popma J. Qualitative and quantitative angiography. In: Topol EJ, ed.Textbook of Interventional Cardiology. Philadelphia: WB Saunders, 1999:725–747.
TABLE 5 Multivariate Correlates of Nine-month Analysis Segment Restenosis
Odds Ratio (95% CI) p Value
All patients (n)Randomized to TAXUS 0.23 (0.13–0.41) �0.0001Lesion length 1.08 (1.03–1.13) 0.002Baseline reference vessel diameter 0.37 (0.19–0.71) 0.003Bend �45° 0.22 (0.05–0.96) 0.04Stent length to lesion length ratio 1.10 (0.75–1.62) 0.62
Bare metal stent groupUnstable angina 2.01 (1.05–3.85) 0.04Lesion length 1.08 (1.02–1.15) 0.008Baseline reference vessel diameter 0.23 (0.11–0.51) 0.0003Stent length to lesion length ratio 1.21 (0.81–1.82) 0.36
CORONARY ARTERY DISEASE/IMPACT OF STENT LENGTH TO LESION LENGTH RATIO 1047
3. Kasaoka S, Tobis JM, Akiyama T, Reimers B, Di Mario C, Wong ND,Colombo A. Angiographic and intravascular ultrasound predictors of in-stentrestenosis. J Am Coll Cardiol 1998;32:1630–1635.4. Kobayashi Y, De Gregorio J, Kobayashi N, Akiyama T, Reimers B, Finci L,Di Mario C, Colombo A. Stent segment length as an independent predictor ofrestenosis. J Am Coll Cardiol 1999;34:651–659.5. Escaned J, Goicolea J, Alfonso F, Perez-Vizcayno MJ, Hernandez R, Fernandez-Ortiz A, Banuelos C, Macaya C. Propensity and mechanisms of restenosis in differentcoronary stent designs. J Am Coll Cardiol 1999;34:1490–1497.6. Kereiakes D, Linnemeier TJ, Baim DS, Kuntz R, O’Shaughnessy C, HermillerJ, Fink S, Lansky A, Nishimura N, Broderick TM, Popma J. Usefulness of stentlength in predicting in-stent restenosis (the Multi-Link stent trials). Am J Cardiol2000;86:336–341.7. Cutlip DE, Chauhan MS, Baim DS, Ho KK, Popma JJ, Carrozza JP, Cohen DJ,Kuntz RE. Clinical restenosis after coronary stenting: perspectives from multi-center clinical trials. J Am Coll Cardiol 2002;40:2082–2089.
8. Hausleiter J, Kastrati A, Mehilli J, Schuhlen H, Pache J, Dotzer F, DirschingerJ, Schomig A. Predictive factor for early cardiac events and angiographic reste-nosis after coronary stent placement in small coronary arteries. J Am Coll Cardiol2002;40:882–889.9. Kastrati A, Ellezi S, Dirschinger J, Hadamitzky M, Neumann FJ, Schomig A.Influence of lesion length on restenosis after coronary stent placement. Am JCardiol 1999;83:1617–1622.10. Ormiston JA, Webster MWI, Ruygrok PN, Meredith IT, Ardill JP, Buller CE,Ricci DR, Chan C, Devlin GP, Stewart JT, et al, Stella Trial Investigators.Six-month angiographic and 12-month clinical follow-up of Multilink long (25 to35 mm) stents for long coronary narrowings in patients with angina pectoris. Am JCardiol 2002;90:222–226.11. Goldberg SL, Loussararina A, DeGregorio J, Di Mario C, Albiero R,Colombo A. Predictors in a native coronary artery. N Engl J Med 2003;349:1315–1323.
1048 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Comparison of the Relation ofTriglyceride-Rich Lipoproteins and
Muscular Artery Compliance inHealthy Women Versus Healthy Men
Ngoc-Anh Le, PhD, W. Virgil Brown, MD, Warren W. Davis, MD,David M. Herrington, MD, MHS, Lori Mosca, MD, PhD, Shunichi Homma, MD,
Barry Eggleston, Howard J. Willens, MD, and Jeffrey K. Raines, PhD
To determine whether structural features or concentra-tions of plasma lipoproteins are predictive of arterialcompliance in healthy women versus healthy men, co-horts of 111 men and 112 women with a wide range of10-year risks for coronary artery disease were selectedusing assessments based on the Framingham HeartStudy. Age ranges were restricted to 35 to 69 years formen and 45 to 79 years for women. Lipid-loweringdrugs or any evidence of vascular disease was cause forexclusion. Fasting lipoprotein analysis and arterial com-pliance measurements in thigh and calf were completedin all patients. Plasma triglyceride levels were the most
powerful predictor of compliance in women. Weakerbut significant relations were observed between plasmanon–high-density lipoprotein cholesterol, apolipopro-tein-B, and apolipoprotein-CIII. In contrast, the only sig-nificant predictor of compliance in men was bodyweight. Thus, the major lipid predictors of arterial stiff-ness in women are concentrations of triglyceride-richlipoproteins. These results are consistent with previousfindings that triglyceride measurements are more stronglyrelated to clinical vascular events in women than inmen. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1049–1054)
We recently demonstrated that compliance of thelarge muscular arteries in the thigh and calf
correlates with the number of stenotic coronary ves-sels1 and the wall area of the abdominal aorta asmeasured by magnetic resonance imaging.2 We alsodocumented that the Framingham risk score3 corre-lates with this measurement of arterial stiffness in thethighs and calves of men and women.2,4 In this study,we investigated the relations between compliance inthigh and calf arteries with a series of quantitativeplasma lipoprotein measurements, including apoli-poproteins. For this purpose, we recruited a cohort ofhealthy men and women who were documented not tohave clinical vascular disease.
METHODSPatient cohort: The protocol for the study was ap-
proved by the institutional review boards at each studysite. The intent of this study was to relate potentiallipoprotein predictors in healthy men and women tomeasurements of the early (subclinical) phase of vas-cular disease manifest by arterial compliance abnor-
malities. Men (35 to 69 years old) and women (45 to79 years old) were recruited into 3 groups of �40patients per group with predicted 10-year risks of�10%, 10% to 20%, or �20% according to themethod of Wilson et al4 as derived from the Framing-ham Heart Study cohort. Patients who had type 2diabetes mellitus were assigned to group 3.5 Evidenceof vascular disease in any arterial distribution or use oflipid-lowering drugs was cause for exclusion. Of theinitial 262 patients who were recruited, 33 were ex-cluded due to recent use of lipid-lowering drugs. Sixsubjects who had abnormalities discovered duringtreadmill stress testing for electrocardiographic orventricular wall motion abnormalities (on echocardi-ography) were also excluded. These exclusions left111 women and 112 men for data analysis.
Laboratory analysis: At the initial visit, medical his-tory, physical examination, and blood analyses werecompleted for 269 patients after an 8-hour fast. TheEmory Lipid Research Laboratory (Atlanta, Georgia)performed all plasma lipoprotein analyses on a BeckmanCX7 chemistry autoanalyzer (Beckman Coulter Diag-nostics, Fullerton, California). Levels of total triglycer-ides and cholesterol were determined by enzymaticmethods (Beckman Coulter Diagnostics) and direct high-density lipoprotein (HDL) and direct low-density li-poprotein cholesterol by homogeneous assays (EqualDiagnostics, Exton, Pennsylvania). Plasma concentra-tions of apolipoprotein-B (apo-B) and lipoprotein(a)were determined by immunoturbidimetry (DiaSorin,Stillwater, Minnesota) and that of apolipoprotein-C-IIIwas determined by another immunoturbidimetricmethod (Wako Chemicals). Triglyceride-rich lipopro-teins were isolated by preparative ultracentrifugation at a
From the Atlanta Veterans Affairs Hospital and Emory University Schoolof Medicine, Atlanta, Georgia; the Wake Forest University MedicalCenter, Winston-Salem, North Carolina; the Columbia UniversityMedical Center, New York, New York; Rho, Inc., Chapel Hill, NorthCarolina; and the University of Miami Medical Center, Miami,Florida. This study was funded in part by grants from Vasocor, Inc.,Boston, Massachusetts, and Credit First Suisse, New York, New York.Manuscript received September 29, 2004; revised manuscript re-ceived and accepted December 27, 2004.
Address for reprints: W. Virgil Brown, MD, Atlanta VAMC MailCode 111, 1670 Clairmont Road, Decatur, Georgia 30033. E-mail:[email protected].
density of 1.020 g/ml (Beckman Coulter). Plasmalevels of high-sensitivity C-reactive protein were mea-sured with an ULTRA-CRP assay (Polymadco, Inc.,Cortlandt Manor, New York), and homocysteine wasdetermined by fluorescence polarization immunoassay(Abbott IMx Analyzer).
Exercise electrocardiography/echocardiography: Allsubjects underwent transthoracic echocardiographyperformed according to guidelines of the AmericanSociety of Echocardiography6 before and after tread-
mill exercise using Bruce’s protocoland were analyzed for segmentalwall motion abnormalities at the Col-lege of Physicians and Surgeons,Columbia University (New York,New York). Significant electrocar-diographic changes (�1 mm of STdepression) or segmental wall mo-tion abnormalities were cause forexclusion.
Measurement of arterial compli-ance: Peripheral arterial compliancewas assessed with a fully automatedcomputer-controlled air plethysmo-graph (Vasogram). The device con-sists of an air pump, a calibrationchamber, and a high-resolution pres-sure transducer.1–3 Special bloodpressure cuffs are placed at the thighand calf, and measurements at theselevels are taken independently.1–3 Acontinuous tracing of the pulse vol-ume is recorded (Figure 1) as thecuff pressure is increased. The max-imum volume change under the cuffis calculated and normalized to apulse pressure of 50 mm Hg(MaxV50) and measured in millili-ters. Higher MaxV50 measure-ments correspond to more compli-ant arteries. Three measurements ofcompliance were made �1 weekapart, and the results were aver-aged. The correlation between pairedmeasurements of calf and thighMaxV50 values obtained during thefirst and second visits were 0.77 ml(p �0.0001) and 0.79 ml (p�0.0001), respectively.1
Statistical analysis: Simple de-scriptive statistics were used to de-scribe the characteristics of the studypopulation. Wilcoxon’s 2-sampletest was used to compare genderswith respect to study variables.Spearman’s correlation coefficients,scatterplots scatteroverlaid with sim-ple linear regression lines, and 95%confidence intervals of the meanwere used to measure the linear re-lations between anthropometric andlipid variables and mean arterial
compliance. Gender-specific multivariable regressionmodeling was used to identify lipid and other hema-tologic predictors that might explain significantamounts of additional variation in mean arterial com-pliance after accounting for statistically significantanthropometric and traditional lipid predictors. First,the statistically significant anthropometric and tradi-tional lipid predictors of mean arterial compliancewithin each gender were found using multivariableregression and forward selections (the criterion for
FIGURE 1. To determine peripheral arterial compliance, cuff pressure (Pc) is measuredon an instantaneous basis. During early diastole the system rapidly extracts and rein-jects the cuff volume by 0.65 ml of air, which produces a step decrease (Pcal) in in-stantaneous cuff pressure. The maximum pressure change in the cuff for the cardiaccycle is also measured (Pm). By knowing Pcal, Pm, and the expansion volume, it ispossible to calculate the maximum volume change (Vm) associated with the cardiaccycle.
TABLE 1 Clinical Characteristics of Subjects by Gender
*Wilcoxon’s 2-sample t test.LDL � low-density lipoprotein.
1050 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
entry was a p value of 0.05). The result was a set of 4gender-specific minimal models for calf and thighmean arterial compliance measurements that includedonly anthropometric and traditional lipid predictorswith a significance level �0.05. The possible predic-tor set was expanded to include advanced lipid andother hematologic variables. With this expanded pre-dictor set, stepwise model selection was used to iden-tify any additional predictors of mean arterial com-pliance after adjusting for the variables contained inthe minimal models. The p value criterion for entryinto the stepwise selection process was 0.10 andthat for staying in the model was 0.05. To assess theimportance of any additional predictors that re-mained after stepwise model selection, minimalmodel R2 values, change in R2 in relation to mini-mal model R2 values, and accompanying p valueswere computed.
All statistical analyses were performed with SAS8.02 (SAS Institute, Cary, North Carolina). Data weretransferred electronically from each clinical site,stored, and analyzed by staff at Rho, Inc. (Chapel Hill,North Carolina).
RESULTSThe clinical characteristics of men and women are
presented in Table 1. The women were older (6 years)by design to help compensate for their delayed onsetof coronary artery disease. However, the Framinghamrisk score in women was 9.6 � 6.7% per decade,which was significantly lower than the score in men(12.9 � 8.8% per decade). The men were heavier thanthe women, but no difference in body mass index wasfound. Systolic blood pressure was higher but dia-stolic pressure was lower in women.
Women had significantly higher levels of totalcholesterol due entirely to the expected higher level ofHDL cholesterol (Table 2). A higher level of apoli-poprotein-AI was consistent with this finding (Table 3).Plasma levels of triglycerides did not differ, but thecontent of apolipoprotein-CIII per apolipoprotein-B(apolipoprotein-CIII/apolipoprotein-B ratio) in triglyc-eride-rich lipoprotein particles (�1.020 g/ml) waslower in women. Levels of low-density lipoproteincholesterol and total apolipoprotein-B and low-densitylipoprotein size were not significantly different be-tween men and women. Levels of high-sensitivityC-reactive protein were higher in women but those ofhomocysteine were similar to those in men.
Patients’ measurements of MaxV50 in calvesand thighs versus weight are shown in Figure 2. Inmen, the correlation coefficient for MaxV50 in thecalf (r � 0.45, p �0.001) was 2 times that in women(r � 0.21, p � 0.03). The MaxV50 value as mea-sured in the thigh also correlated with weight inmen (r � 0.23, p � 0.02) but not in women.Because MaxV50 is designed to be an index ofabsolute volume change with each pulse wave, it isnot surprising that the larger legs of larger subjectswould manifest a larger pulse volume wave. Ac-cordingly, multivariate analysis conducted for other
TABL
E3
Add
ition
alBl
ood
Plas
ma
Ana
lyse
sin
Subj
ects
byG
ende
r
Gen
der
Apo
-CIII
(mg/
dl)
Apo
-CIII
/apo
-B*
Apo
-B(m
g/dl
)A
po-A
I(m
g/dl
)Lp
(a)(
mg/
dl)
LDL
Dia
met
er(n
m)
Hcy
(�m
/L)
CRP
(mg/
L)
Wom
en22
.3�
6.0
0.25
�0.
1613
1.3
�20
.014
1.0
�31
.529
.2�
30.3
25.3
�0.
88.
9�
3.7
4.3
�5.
4M
en21
.0�
5.8
0.31
�0.
1912
7.3
�20
.912
4.4
�27
.829
.8�
30.0
25.2
�1.
19.
1�
2.8
2.0
�2.
1G
ende
rco
mpa
rison
(pva
lue)
†0.
050.
030.
08�
0.00
10.
670.
100.
150.
002
*Rat
ioin
trigl
ycer
ide-
rich
lipop
rote
ins
(d�
1.01
9g/
ml).
†W
ilcox
on’s
2-sa
mpl
ett
est.
CRP
�C
-reac
tive
prot
ein;
Hcy
�ho
moc
yste
ine;
Lp(a
)�lip
opro
tein
(a).
Oth
erab
brev
iatio
nas
inTa
ble
2.
PREVENTIVE CARDIOLOGY/TRIGLYCERIDE-RICH LIPOPROTEINS AND MUSCULAR ARTERY COMPLIANCE BY GENDER 1051
variables was done after adjusting for body weight,except for female thigh (see the following).
In univariate analysis, highly significant decreasesin the MaxV50 measured at the calf (r � �0.30, p �0.001) and thigh (r � �0.39, p �0.001) was associ-ated with increasing levels of triglycerides in women(Figure 3). In women, apolipoprotein-B and non-HDLalso correlated negatively with MaxV50 in the thighand calf (Table 4). Total apolipoprotein-CIII relatedinversely to MaxV50 in the thigh (r � �0.21, p �0.03), but no significant correlation was found in thecalf. The relations for women with correlation co-efficients �0.15 and a p value �0.05 at the thigh orcalf are shown in Figure 4. Total cholesterol metthese criteria but must be considered of marginalsignificance because multiple comparisons weremade in this analysis. In men, there were no signif-icant relations between triglyceride levels (Figure3) and other lipoprotein measurements. High-sensi-tivity C-reactive protein was inversely but weaklyrelated (r � �0.21, p � 0.02) to compliance only atthe calf and only in men.
Results of multivariate analysis of the relations be-tween MaxV50 and the various anthropometric and lipidparameters as predictors are presented in Table 5. Theminimal model adjusted for the correlation betweenbody weight and the compliance measurements at thethigh and calf in men and only in the calf in women.With only weight adjustment as the minimal model inmen, adding apolipoproten-CIII provided slight im-provement in predicting MaxV50 at the calf (but not at
the thigh). No other parameters added predictive value inthe men. In women, the minimal model for the calf basedon physical and traditional lipid parameters includedweight and triglyceride levels, the strongest of the cor-relates in univariate analysis. With this model, levels ofapolipoprotein-B, apolipoprotein-CIII/apolipoprotein-Bratio (see Discussion) in triglyceride-rich particles, andhomocysteine added predictive value. These additionalparameters increased the R2 value from 0.16 to 0.31 at
FIGURE 2. Plot of patients’ compliance measurements (MaxV50)versus subjects’ weight in kilograms for the (A) thigh and (B) calfin women and for the (C) thigh and (D) calf in men. Standardregression lines and 95% confidence intervals are shown. Onlythe thigh data are significant in women (r2 � 0.044), whereasthe thigh (r2 � 0.20) and calf (r2 � 0.05) are significant in men.
FIGURE 3. Patients’ compliance measurements (MaxV50) in the(A) thigh and (B) calf in women and the (C) thigh and (D) calf inmen versus fasting triglyceride levels (milligrams per deciliter).Standard regression lines and 95% confidence intervals areshown.
FIGURE 4. Spearman’s correlation coefficients for relations be-tween various lipoprotein measurements and Max V50 at thethigh (black bars) and calf (gray bars) in women. Only thosemeasurements with a value >0.15 are included. The p value foreach relation is presented in Table 4. Chol � cholesterol; CIII �apolipoprotein-CIII; NonHdl � non-HDL; TG � triglycerides.
1052 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
the calf. At the thigh, only the apolipoprotein-CIII/apo-lipoprotein-B ratio added predictive value to the minimalmodel, which included only triglycerides. Thus, apoli-poprotein-CIII/apolipoprotein-B ratio was a significantpredictor of decreased compliance at the thigh and calf
only after adjusting for total triglycer-ides. Neither non-HDL cholesterol nortotal apolipoprotein-B added to thepredictive value when triglycerideswere considered, which is consistentwith their known relation to eachother, probably as components ofvery low-density lipoproteins and in-termediate-density lipoproteins.
DISCUSSIONThe method of measuring com-
pliance at the thigh and calf used inthis study has been shown to pre-dict the extent of coronary arterydisease measured by angiography2
and the wall area of the abdominalaorta as assessed by magnetic res-onance imaging.3 In 2 previousstudies that included �350 patientseach, the relations between the Fra-mingham risk calculations (usingclassic risk factors) and MaxV50measurements were found to behighly significant, with partial R2
values of 0.19 for the thigh and0.17 for the calf.3,5 These studiesincluded patients who had provedvascular disease and variable risksover a wide age range. Previousinvestigators have also found thatvarious measurements of arterio-sclerotic disease correlate with de-creased compliance as assessed byother methods, such as ultrasoundtechniques, pulse velocity, andpulse-wave analysis.7–13 It is note-worthy that these techniques tendto depend primarily on propertiesof the thoracic aorta rather than onthose of muscular arteries. Theseare research techniques that requirehighly trained personnel. Thepresent device is applicable in anoutpatient setting and can be usedby staff members who have mini-mal training.
The work described in this studywas done to determine whether pa-tients who had a range of classicrisk factors but no clinical evidenceof vascular disease would show re-lations between lipoprotein param-eters and the stiffness of thigh orcalf arteries as measured by this airplethysmographic device. The studydesign allowed comparison of co-
horts of women and men of equivalent size and com-parable risk as calculated by the Framingham riskfunctions. The most striking finding in this study wasthe very strong and independent relation between fast-ing triglyceride concentrations and decreased compli-
TABLE 4 Univariate Correlations With Compliance
Variable n
Calf MaxV50* Thigh MaxV50*
R2 p Value R2 p Value
WomenApo AI 109 �0.04 0.65 �0.02 0.80Apo B 110† �0.24† 0.01† �0.27† 0.005†
*Spearman’s correlation coefficient is shown with p values for compliance at the thigh and calf. Thismethod provides for assessment of p values in nonparametric data.
†Parameters with correlations �0.2 and p values �0.05 at the thigh and calf.‡Ratio in isolated triglyceride-rich lipoproteins (d �1.019 g/ml).apo � apolipoprotein; Lp(a) � lipoprotein (a); other abbreviations as in Tables 2 and 3.
TABLE 5 Relation of Compliance and Various Parameters in MultivariateAnalysis*
MaxV50 in womenThigh Weight/BMI 0.10 NoneMaxV50 in menThigh TG 0.15 Apo-CIII/apo-B 0.06 0.009MaxV50 in women
*The minimal model in women for MaxV50 recorded at the thigh included plasma TG only, and thatfor Max V50 at the calf included body weight and plasma TG. In men, the minimal model includedweight at the calf and weight and BMI at the thigh. Those additional measurements that addedindependently to the strength of the correlation with vascular compliance are shown for men and womenand at the 2 anatomic locations.
BMI � body mass index; TG � triglycerides. Other abbreviation as in Tables 3 and 4.
PREVENTIVE CARDIOLOGY/TRIGLYCERIDE-RICH LIPOPROTEINS AND MUSCULAR ARTERY COMPLIANCE BY GENDER 1053
ance in the arteries of the thigh and that separatelymeasured in the calf of women. This relation inwomen appears to be confirmed by significant corre-lations of MaxV50 in the thigh and calf with non-HDLcholesterol and with total plasma apolipoprotein-B.Plasma apolipoprotein-CIII, another indicator of in-creased triglyceride-rich lipoproteins,14 also corre-lated with decreased compliance in thighs of women.None of these relations were detected in men. Otherstudies have reported stronger relations in women thanin men between plasma triglycerides and the incidence(or prevalence) of vascular disease. This has beenfound in the Framingham Heart Study,15 thePROCAM study in Europe,16 the Evans County Studyin Georgia,17 the Lipid Research Clinic Follow-upStudy,18 a meta-analysis of community studies andclinical trials,19 and, more recently, the Third NationalHealth and Nutrition Examination Survey.20 The weakcorrelation between HDL cholesterol and low-densitylipoprotein size was surprising because these measure-ments tend to be inversely related to triglyceride con-centrations. This finding may be due to the smallnumbers of subjects in this study compared with thecommunity-based datasets that have defined these re-lations. Enrichment of apolipoprotein-C III in apoli-poprotein-B–containing lipoproteins has been foundto be a risk factor for vascular disease in cohorts ofpatients in clinical trials whether treated with placeboor cholesterol-lowering drugs.21–23
Our data differ from previous measurements ofvascular compliance, because subjects’ ages did nothave a strong correlation with MaxV50 at the calf orthigh. In previous measurements of compliance, therelation between wall stiffness and increased pulsevelocity has been the most common assessment.7–10
The length of the thoracic aorta is the major contrib-utor to the distance over which this velocity is mea-sured. The alteration of this very elastic tissue withage may be a strong determinant of these results,whereas the specific assessment of muscular arteries(MaxV50), such as those in the thigh and calf, mayrespond differently to aging.
1. Herrington DM, Kesler K, Reiber JHC, Davis W, Brown WV, Helms R, MallonS, Raines JK. Arterial compliance adds to conventional risk factors for prediction ofangiographic coronary artery disease. Am Heart J 2003;146:662–667.2. Herrington DM, Brown WV, Mosca L, Davis W, Helms R, Mallon SM, RainesJK. Relationship between arterial stiffnes and subclinical aortic atherosclerosis.Circulation 2004;110:432–437.3. Wilson PWF, D‘Agostino RB, Levy D, Belanger AM, Silbershatz H, KannelWB. Prediction of coronary heart disease using risk factor categories. Circulation1998;97:1837–1847.
4. Willens HJ, Davis W, Herrington DM, Wade K, Kesler K, Mallon S, Brown WV,Reiber JHC, Raines JK. The relationship of peripheral arterial compliance to standardcardiovascular risk factors. Vasc Endovasc Surg 2003;37:197–206.5. Expert Panel on Detection, Evaluation, and Treatment of High BloodCholesterol in Adults (Adult Treatment Panel III). Third report of the Na-tional Cholesterol Education Program (NCEP) expert panel on detection,evaluation, and treatment of high blood cholesterol in adults. Circulation2002:106:3144 –3421.6. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, Feigenbaum H,Gutfesell H, Reichek N, Sahn D, Schnittinger I, et al, and the American Societyof Echocardiography Committee on Standards, Subcommittee on Quantitation ofTwo-Dimensional Echocardiograms. Recommendations for quantitation of leftventricle by two-dimensional echocardiography. J Am Soc Echocardiogr 1989;2:358–367.7. Toto-Moukouo JJ, Achimastos A, Asmar RG, Hugues CJ, Safar ME. Pulse wavevelocity in patients with obesity and hypertension. Am Heart J 1986;112:136–140.8. Asmar RG, Pannier B, Santoni JP, Laurent S, London GM, Levy BI, Safar ME.Reversion of cardiac hypertrophy and reduced arterial compliance after convert-ing enzyme inhibition in essential hypertension. Circulation 1988;78:941–950.9. Isnard RN, Pannier BM, Laurent S, London GM, Diebold B, Safar ME. Pulsatilediameter and elastic modulus of the aortic arch in essential hypertension: a nonin-vasive study. J Am Coll Cardiol 1989;13:399–405.10. Hirai T, Sasayama S, Kawasaki T, Yagi S. Stiffness of systemic arteries inpatients with myocardial infarction. Circulation 1989;80:78–86.11. Bogren HG, Mohiaddin RH, Klipstein RK, Firmin DN, Underwood RS, ReesSR, Longmore DB. The function of the aorta in ischaemic heart disease: amagnetic resonance and angiographic study of aortic compliance and blood flowpatterns. Am Heart J 1989;118:234–247.12. London GM, Marchais SJ, Safar ME, Genest AF, Guerin AP, Metivier F,Chedid K, London AM. Aortic and large artery compliance in end-stage renalfailure. Kidney Int 1990;37:137–142.13. Lehmann ED, Riley WA, Clarkson P, Gosling RG. Non-invasive assessmentof cardiovascular disease in diabetes mellitus. Lancet 1997;350(suppl 1):14–19.14. Ginsberg HN, Le NA, Goldberg IJ, Gibson JC, Rubinstein A, Wang-IversonP, Norum R, Brown WV. Apolipoprotein B metabolism in subjects with defi-ciency of apolipoprotein C-III and A-I: evidence that apolipoprotein C-III inhibitslipoprotein lipase in vivo. J Clin Invest 1986;78:1287–1295.15. Castelli WP. The triglyceride issue: a view from Framingham. Am Heart J1986;112:432–437.16. Assmann MD, Schulte H. Relation of high-density lipoprotein cholesterol andtriglycerides to incidence of atherosclerotic coronary artery disease (the PROCAMexperience). Am J Cardiol 1992;70:733–737.17. Heyden S, Heiss G, Hames CG, Bartel AG. Fasting triglycerides as predictorsof total and CHD mortality in Evans County, Georgia. J Chron Dis 1980;33:275–282.18. Criqui MH, Heiss G, Cohn R, Cowan LD, Suchindran CM, Bangdiwala S,Kritchevsky S, Jacobs DR Jr, O’Grady HK, Davis CE. Plasma triglyceride levelsand mortality from coronary heart disease. N Engl J Med 1993;328:1220–1225.19. Hokanson JE, Austin MA. Plasma triglyceride level is a risk factor forcardiovascular disease independent of high-density lipoprotein cholesterol level:a meta-analysis of population-based prospective studies. J Cardiovasc Risk1996;3:213–219.20. Ninomiya JK, Italien GL, Criqui MH, Whyte JL, Gamst A, Chen RS.Association of the metabolic syndrome with history of myocardial infarction andstroke in the Third National Health and Nutrition Examination Survey. Circula-tion 2004;109:42–46.21. Hodis HN, Mack WJ, Azen SP, Alaupovic P, Pogoda JM, LaBree L,Hemphill LC, Kramsch DM, Blankenhorn DH. Triglyceride- and cholesterol-richlipoproteins have a differential effect on mild/moderate and severe lesion pro-gression as assessed by quantitative coronary angiography in a controlled trial oflovastatin. Circulation 1994;90:42–49.22. Sacks FM, Alaupovic P, Moye LA, Cole TG, Sussex B, Stampfer MJ, PfefferMA, Braunwald E. VLDL, apolipoproteins B, CIII, and E, and risk of recurrentcoronary events in the Cholesterol and Recurrent Events (CARE) trial. Circula-tion 2000;102:1886–1892.23. Luc G, Fievet C, Arveiler D, Evans AE, Bard JM, Cambien F, Fruchart JC,Ducimetiere P. Apolipoproteins C-III and E in apoB- and non-apoB-containinglipoproteins in two populations at contrasting risk for myocardial infarction: theECTIM study. J Lipid Res 1996;37:508–517.
1054 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Comparison of Autotriggered MemoryLoop Recorders Versus Standard Loop
Recorders Versus 24-Hour HolterMonitors for Arrhythmia Detection
James A. Reiffel, MD, Robert Schwarzberg, MD, and Maria Murry, RN
To determine the relative yields of Holter monitoring(HM), memory loop recording (MLR), and autotriggeredMLR (AT-MLR), we retrospectively interrogated the verylarge database of Lifewatch (a Card Guard companyand a commercial monitoring company) and comparedthe results obtained by each method. From among atotal database of �100,000 patients, records of 1,800patients from 2003 were randomly selected and exam-ined, 600 from each of the 3 different monitoringgroups. Each session of MLR and AT-MLR was appliedfor 30 days. For each patient we determined the symp-tomatic and asymptomatic events that were docu-mented, including those that met predefined immediatephysician notification criteria and the time to first notifi-cation event. The groups were identical in age andsymptoms that necessitated monitoring; fewer womenhad HM. Information on the type of underlying structuralheart disease, if present, and medications taken, if any,was not available to us in this database. The AT-MLRapproach provided a higher yield of diagnostic events(e.g., 37, 108, and 216 total patients who had events;37, 212, and 524 total events; and 6.2%, 17%, and 36%
with a diagnostic yield for HM, MLR, and AT-MLR, re-spectively) and an earlier diagnosis. AT-MLR was alsothe most effective technique for capturing asymptomaticsignificant events, such as atrial fibrillation (52 withAT-MLR vs 1 for standard MLR). AT-MLR detected morethan half as many asymptomatic episodes of atrialfibrillation (n � 52) as the total number of symptomaticepisodes detected by patient activated recording (n �94), thus confirming the common presence of asymp-tomatic atrial fibrillation. AT-MLR provided electro-cardiographic documentation of tachyarrhythmias(n � 392) more often than MLR (n � 47) or HM (n � 44)and bradyarrhythmias/pauses/atrioventricular block(n � 38) more often than MLR (n � 13) or HM (n � 18).Thus, MLR and AT-MLR provide a diagnosis more oftenthan does HM, thus confirming the benefit of prolongedmonitoring. Further, the higher yield of AT-MLR versusMLR demonstrates the significantly enhanced benefit ofautotriggered programmable recording. �2005 byExcerpta Medica Inc.
(Am J Cardiol 2005;95:1055–1059)
A lthough standard 24-hour Holter monitoring(HM), memory loop recording (MLR), and re-
cently, autotriggered MLR (AT-MLR)1–9 are availablefor use in clinical practice and are frequently em-ployed, the relative yield of each method for symp-tomatic and asymptomatic events10–12 has not beenadequately assessed. We performed this study toquantitate the relative yield of HM, MLR, and AT-MLR in detecting clinically significant dysrhythmias.
METHODSTo determine the relative yield of HM, MLR, and
AT-MLR, we retrospectively examined the large da-tabase of Lifewatch, a commercial monitoring com-pany, which is part of the Card Guard group (BuffaloGrove, Illinois). Specifically, we compared the resultsobtained by 24-hour continuous HM, standard 30-dayMLR, and 30-day AT-MLR in the documentation of
clinically significant symptomatic and asymptomaticrhythm disturbances in patients who were referred formonitoring for known or suspected dysrhythmias.From among a database of �100,000 patients, therecords of 1,800 patients were randomly selected fromrecordings made during 2003 using each of the 3different monitoring techniques, i.e., HM, MLR, andAT-MLR (600 patients from each). For each patientwe determined the symptomatic and asymptomaticevents that were documented, including those that metpredefined immediate physician notification criteria(Figure 1) and the time to first physician notificationevent. The groups were identical in age and symptomsthat necessitated monitoring; fewer women had HM(Table 1). Information on the type of underlying struc-tural heart disease, if present, and medications taken,if any, was not available to us in this database. Spe-cifically, the following data were collected: total trans-missions in each group, total events recorded in eachgroup, total patients in each group with physiciannotification, total events meeting physician notifica-tion criteria, patient’s age and gender, primary diag-nosis (reason for monitoring as provided by the pa-tient’s physician), days from baseline to physiciannotification, physician notification classification, symp-tomatic episodes with physician notification, asymptom-
From the Division of Cardiology, Department of Medicine, ColumbiaUniversity, New York, New York; the Division of Cardiology, Depart-ment of Medicine, Florida Atlantic University, Boca Raton, Florida; andLifewatch, A Card Guard Company, Buffalo Grove, Illinois. Manu-script received November 4, 2004; revised manuscript received andaccepted January 3, 2005.
Address for reprints: James A. Reiffel, MD, 161 Fort WashingtonAvenue, New York, New York 10032. E-mail: [email protected].
atic episodes with physician notification, and asymptom-atic atrial fibrillation.
At the time of each transtelephonic transmission,the patient was asked about the symptoms that werepresent during the recording and the symptoms thatwere present at the time of the transmission, whenadditional live transmissions could be made. Figure 2displays a typical report. The criteria for immediatenotification of a physician about the recorded event, incontrast to routine transmission of the results, whichwere prespecified after patient enrollment, included(1) narrow QRS tachycardia of �150 beats/min of�30 seconds; (2) bradycardia of �30 seconds in anage-adjusted manner (�70 beats/min for 1 to 2 yearsof age, �50 beats/min for 3 to 10 years of age, and�40 beats/min for �10 years of age); (3) pauses of�3 seconds; (4) wide QRS tachycardia of �120 beats/min of �4 beats; (5) atrial flutter and/or fibrillationthat was of new (first) onset and �30 seconds or had
a ventricular response of �150 beats/min for �30seconds; (6) any second- or third-degree atrioventric-ular block; (7) all reported syncopes; (8) any pace-maker failure or malfunction; and (9) any documentedor reported implantable cardioverter-defibrillator dis-charge. In this study, programmable features for theAT-MLR included tachycardia trigger criteria, brady-cardia trigger criteria, beep for event recorded (on),atrial fibrillation event (on), pre-event memory period(60 seconds), postevent memory (30 seconds), andother default criteria.
RESULTSThere were no differences in age across the 3
groups. There were no differences in symptoms thatresulted in monitoring across the 3 groups (Table 1).Sixty-five percent to 80% (p � NS) had palpitations,dizziness, and/or syncope as the clinical event thatbeing evaluated.
FIGURE 1. The form used at Life-watch lists criteria for promptphysician notification (used withpermission from Lifewatch, ACard Guard Company).
1056 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
The AT-MLR approach provided a higher total yieldof diagnostic events (Table 1), including the total numberof patients who had events that required physician noti-fication, the total number of events that resulted in phy-sician notification, and percent recordings with a diag-nostic yield. The AT-MLR approach also provided ahigher capture of specific diagnostic events (Table 1),including the number of recordings that documented asymptomatic event that resulted in physician notification,the number of recordings that documented an asymp-tomatic event that resulted in physician notification, thenumber of documented asymptomatic atrial fibrillation
episodes, and the total number of atrial fibrillation epi-sodes that required physician notification. Thirty-onepercent of patients who had asymptomatic atrial fibrilla-tion were �65 years old. The AT-MLR approach alsoprovided a higher yield of other diagnostic events and anearlier diagnosis (Table 1), including documentation of awide QRS tachycardia, documentation of a narrow QRStachycardia, documentation of a significant bradycardiaor excessive pause (as previously described), and docu-mentation of second- or third-degree atrioventricularblock. The number of days to initial immediate physiciannotification was 1 to 5 days in 2 patients who underwent
TABLE 1 Indications for Monitoring and Results Obtained by 24-hour Holter Monitoring (HM), Standard 30-day Memory LoopRecording (MLR), and Autotriggered Memory Loop Recording (AT-MLR)
Data Collected HM MLR AT-MLR
Total no. of patients 600 600 600Total no. of transmissions 600 4,862 11,124Total no. of events 600 11,562 21,390Total no. of patients who had PN 37 108 216Total no. of transmissions with PN 37 212 524Patient demographics
Symptomatic episodes with PN 37 (per diary) 204 268Asymptomatic episodes with PN Inconclusive data due to patients
who had no diary8 256
Asymptomatic atrial fibrillation NA—some patients had nodiary
1 52
NA � not available; PN � physician notification.
ARRHYTHMIAS AND CONDUCTION DISTURBANCES/AUTOTRIGGERED MEMORY LOOP RECORDERS ARE SUPERIOR 1057
HM, 69 patients who underwent MLR, and 131 patientswho underwent AT-MLR. With AT-MLR, the averagenumbers of days to the first immediate physician notifi-cation because of an indicated event were 7.5 days forasymptomatic episodes and 8.1 days for a symptomaticepisode.
DISCUSSIONTo better determine the relative efficacy of the 2
monitoring techniques most commonly used, HM andMLR, plus that of the recently developed AT-MLR,which, if superior, would likely replace the standardMLR in clinical practice, the present study was per-formed. Our results were very clear (Table 1). Forvirtually all dysrhythmias detected, the AT-MLRmethod captured a larger number of underlying symp-tom-provoking arrhythmias than did the alternativeapproaches. Moreover, the AT-MLR approach wasmore effective at capturing clinically significant butasymptomatic arrhythmias, such as episodes of atrialfibrillation and transient bradyarrhythmias, than wasHM or MLR. In this respect, the AT-MLR data par-
allels the data currently availableonly from continuously monitoringimplanted devices, such as frompacemaker memory interrogation orfrom the diagnostic implanted RE-VEAL monitor (Medtronic, Inc.,Minneapolis, Minnesota).13–16 In ad-dition, the AT-MLR seems to be anideal alternative in most circumstancesto commercially available outpatienttelemetric monitoring systems, such asthat of CardioNet (San Diego, Califor-nia).17
Documentation of asymptomaticarrhythmias that are clinically signif-icant is currently of great inter-est.10–12 For example, asymptomaticatrial fibrillation that remains despitethe absence of further symptomaticevents after antiarrhythmic drug ini-tiation or a procedural intervention,such as ablation or Maze surgery, isrecognized as important to detect be-cause of its implications regardinganticoagulation therapy. Similarly, inpatients who have infarction withouta sustained ventricular tachyarrhyth-mia and a decreased left ventricularejection fraction that remains �30%,implantation of a defibrillator is cur-rently not supportable unless the pa-tient has nonsustained ventriculartachycardia and is inducible at elec-trophysiologic testing.18–20 The abil-ity of AT-MLR to detect such ectopyis superior to HM and MLR. Like-wise, in a patient who has dizzinessor syncope but no recorded symp-tomatic event, the detection of a dys-rhythmia, such as Mobitz II block,
would result in a clinical intervention (e.g., pacemakerimplantation) without the need for further monitoring.
In addition, the AT-MLR approach in our studyprovided a dysrhythmic diagnosis of a severity thatjustified rapid physician notification more quicklythan did HM or standard MLR. This early notificationof a serious dysrhythmia should translate into im-proved clinical care. Our results clearly suggest anadvantage to AT-MLR over standard MLR and agreater likelihood than HM to provide diagnosticallyuseful information.
There are several limitations with respect to ourdata. The selection of the monitoring technique to usein each patient was not randomized; rather, it was bydecision of the referring physician. Biases must existin such a process. For example, 1 might assume thatthe 24-hour HM technique was used for patients whohad frequent symptoms, such as daily, whereas anMLR system might be preferentially used in patientswhose symptoms are less frequent. Despite this pos-sible bias, the MLR systems used in this study weremore, not less, effective in capturing events, although
FIGURE 2. A representative report provided to the enrolling physician after transmis-sion of a recorded event (used with permission of Lifewatch, A Card Guard Com-pany).
1058 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
they may expectedly have occurred less frequently inpatients who had been referred for MLR or AT-MLRrather than HM. Presumably, memory loop unitswould capture, by patient trigger, all the symptomaticperiods captured by HM as correlated with the pa-tient’s diary and all the significant asymptomatic ar-rhythmias noted during the HM recording if the AT-MLR were the unit used. Another potential limitationis the unknown bias that may exist with regard to theabsence of information concerning underlying struc-tural heart disease and drug consumption in the pa-tients included in our study, which may not necessar-ily have been the same in our 3 patient groups.Nonetheless, because the study concerns detection ofsymptomatic and asymptomatic dysrhythmias, not oftheir etiologic or contributing mechanisms, the ab-sence of such information should not negate the sig-nificance of our findings, although, admittedly, drugadministration could alter the findings with regard tothe incidence of symptomatic versus asymptomaticarrhythmias and total arrhythmic events. Accordingly,our data clearly show that the diagnostic yields ofMLR and AT-MLR are greater than that of HM anddemonstrate the benefit of prolonged monitoring; thediagnostic yield of AT-MLR is greater than that ofstandard MLR and demonstrates the benefits of pro-grammable AT-MLR; AT-MLR is a superior methodfor detecting symptomatic and asymptomatic tachy-cardias, significant bradycardias, and episodes of atrialfibrillation (with or, especially, without symptoms)than HM or standard MLR; and the use of AT-MLRshould improve clinical care over that available withthe use of HM or standard MLR techniques.
1. Reiffel JA, Schulhof E, Joseph B, Severance E, Wyndus P, McNamara A. Theoptimum duration of transtelephonic ECG monitoring when used for transientsymptomatic event detection. J Electrocardiogr 1991;24:165–168.
2. Bigger JT Jr, Reiffel JA, Coromilas J. Ambulatory electrocardiography. In: PlatiaE, ed. Non-Pharmacologic Management of Cardiac Arrhythmias. Philadelphia: JBLippincott, 1986:36–61.3. Kowey PR, Kocovic DZ. Cardiology patient pages. Ambulatory electrocar-diographic recording. Circulation 2003;108:31–33.4. Heilbron EL. Advances in modern electrocardiographic equipment for long-term ambulatory monitoring. Card Electrophysiol Rev 2002;6:185–189.5. Belgian Society of Cardiology, Working Group on Electrocardiology, ExerciseTesting and Cardiac Rehabilitation. Clinical value of vectorcardiography, Holtermonitoring and quantitative electrocardiology. Acta Cardiol 2000;55:157–162.6. Zimetbaum PJ, Josephson ME. The evolving role of ambulatory arrhythmiamonitoring in general clinical practice. Ann Intern Med 1999;130:848–856.7. Zimetbaum P, Josephson ME. Evaluation of patients with palpitations. N EnglJ Med 1998;338:1369–1373.8. Bell C, Kapral M. Use of ambulatory electrocardiography for the detection ofparoxysmal atrial fibrillation in patients with stroke. Canadian Task Force onPreventive Health Care. Can J Neurol Sci 2000;27:25–31.9. Gula LJ, Krahn AD, Masse D, Skanes A, Yee R, Klelin GJ. External looprecorders: determinants of diagnostic yield in patients with syncope. Am Heart J2004;147:644–648.10. Page RL, Wilkinson WE, Clair WK, McCarthy EA, Pritchett EL. Asymp-tomatic arrhythmias in patients with symptomatic paroxysmal atrial fibrillationand paroxysmal supraventricular tachycardia. Circulation 1994;89:224–227.11. Page RL, Reiffel JA, Tavel ME. Asymptomatic atrial fibrillation: problems ofmanagement. Chest 2001;119:628–631.12. Kaufman ES, Waldo AL. The impact of asymptomatic atrial fibrillation. J AmColl Cardiol 2004;43:53–54.13. Pollak WM, Simmons JD, Interian A Jr, Castellanos A, Myerburg RJ,Mitraini RD. Pacemaker diagnostics: a critical appraisal of current technology.Pacing Clin Electrophysiol 2003;26:76–98.14. Nowak B. Taking advantage of sophisticated pacemaker diagnostics. Am JCardiol 1999;83:172–179.15. Krahn AD, Klein GJ, Skanes AC, Yee R. Insertable loop recorder use fordetection of intermittent arrhythmias. Pacing Clin Electrophysiol 2004;27:657–664.16. Benditt DG, Ermis C, Pham S, Hiltner L, Vrudney A, Lurie KG, SakaguchiS. Implantable diagnostic monitoring devices for evaluation of syncope, andtachy- and brady-arrhythmias. J Interv Card Electrophysiol 2003;9:137–144.17. Kowey PR, Joshi A, Prystowsky EN, Benditt DG, Cannom DS, Pratt CM,McNamara A, Sangrigoli RM. First experience with a mobile cardiac outpatienttelemetry (MCOT) system for the diagnosis and management of cardiac arrhyth-mia. PACE 2003;26:1090.18. Moss AJ. MADIT-I and MADIT-II. J Cardiovasc Electrophysiol 2003;14:96–98.19. Nisam S, Henry S, Wilber DJ. Implementation of MADIT and MUSTT inclinical practice: results of an international survey. Ann Noninvasive Electrocar-diol 2002;7:399–405.20. Nisam S, Henry S, Wilbur DJ. Implementation of MADIT and MUSTT inclinical practice: results of an international survey. Ann Noninvasive Electrocar-diol 2002;7:399–405.
ARRHYTHMIAS AND CONDUCTION DISTURBANCES/AUTOTRIGGERED MEMORY LOOP RECORDERS ARE SUPERIOR 1059
Long-Term Effects of Stress Reduction onMortality in Persons >55 Years of Age
With Systemic HypertensionRobert H. Schneider, MD, Charles N. Alexander, PhD,* Frank Staggers, MD,
Maxwell Rainforth, PhD, John W. Salerno, PhD, Arthur Hartz, MD, Stephen Arndt, PhD,Vernon A. Barnes, PhD, and Sanford I. Nidich, EdD
Psychosocial stress contributes to high blood pressure andsubsequent cardiovascular morbidity and mortality. Previ-ous controlled studies have associated decreasing stresswith the Transcendental Meditation (TM) program withlower blood pressure. The objective of the present studywas to evaluate, over the long term, all-cause and cause-specific mortality in older subjects who had high bloodpressure and who participated in randomized controlledtrials that included the TM program and other behavioralstress-decreasing interventions. Patient data were pooledfrom 2 published randomized controlled trials that com-pared TM, other behavioral interventions, and usual ther-apy for high blood pressure. There were 202 subjects,including 77 whites (mean age 81 years) and 125 African-American (mean age 66 years) men and women. In thesestudies, average baseline blood pressure was in the pre-hypertensive or stage I hypertension range. Follow-up ofvital status and cause of death over a maximum of 18.8
years was determined from the National Death Index.Survival analysis was used to compare intervention groupson mortality rates after adjusting for study location. Meanfollow-up was 7.6 � 3.5 years. Compared with combinedcontrols, the TM group showed a 23% decrease in theprimary outcome of all-cause mortality after maximumfollow-up (relative risk 0.77, p � 0.039). Secondary anal-yses showed a 30% decrease in the rate of cardiovascularmortality (relative risk 0.70, p � 0.045) and a 49% de-crease in the rate of mortality due to cancer (relative risk0.49, p � 0.16) in the TM group compared with combinedcontrols. These results suggest that a specific stress-de-creasing approach used in the prevention and control ofhigh blood pressure, such as the TM program, may con-tribute to decreased mortality from all causes and cardio-vascular disease in older subjects who have systemichypertension. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1060–1064)
Quantitative reviews of previous studies of reduc-tion approaches to stress have either shown neg-
ative results or heterogeneity of effects on decreasingblood pressure depending on the experimental designand selection of specific technique.1,2 However, recentmeta-analytic reviews have indicated that decreasingstress with the Transcendental Meditation (TM) pro-gram may be particularly effective in decreasing highblood pressure, psychosocial stress, morbidity, andmortality.3–5 Two previous randomized controlled tri-als of behavioral interventions associated the TM pro-gram with significantly decreased blood pressure after 3months in independent samples of older white and Af-rican-American subjects.6–8 In addition, preliminary
data suggested trends toward decreases in mortality inthe TM group compared with active and usual-carecontrols in these trials.8–10 Therefore, the objective ofthe present study was to evaluate long-term all-causeand cause-specific mortality in older subjects who hadhigh blood pressure and who participated in random-ized controlled trials that included the TM programand other behavioral stress-reducing interventions.
METHODSClinical trials: Computerized and manual searches
of the literature were used to locate published reportsthat met the following criteria: (1) randomized con-trolled trial design, (2) blood pressure as a primaryoutcome, and (3) studies that compared the effects ofthe TM program with those of other behavioral controlinterventions. Two trials met the search criteria andprovided data for the present pooled analysis.6–8
The methods of the trial by Alexander et al8 havebeen described in detail. In brief, the 77 white partic-ipants (18% men and 82% women; mean age 81 � 9.3years) were residents of homes for the elderly in theBoston area. Baseline systolic blood pressure aver-aged 138 � 15.9 mm Hg. Outcome measurements inthe original study were cognitive functioning, mentalhealth, and systolic blood pressure after 3 months andmortality rate after 3 years. Subjects were randomizedto 4 treatment groups: a TM group (n � 21), a mind-fulness training group (n � 23), a mental relaxation
From the Institute for Natural Medicine and Prevention, MaharishiUniversity of Management, Fairfield, Iowa; the West Oakland HealthCenter, Oakland, California; the Departments of Family Medicine andPsychiatry, University of Iowa College of Medicine, Iowa City, Iowa;and the Georgia Prevention Institute, Medical College of Georgia,Augusta, Georgia. This study was supported in part by a SpecializedCenter of Research grant from the National Institutes of Health, Be-thesda, Maryland (Grant 1P50AT00082 from the National Center forComplementary and Alternative Medicine). Manuscript received Au-gust 30, 2004; revised manuscript received and accepted December28, 2004.
Address for reprints: Robert H. Schneider, MD, Institute for NaturalMedicine and Prevention, Maharishi University of Management,2100 Mansion Drive, Maharishi Vedic City, Iowa 52556. E-mail:[email protected].
group (n � 22), and a usual-care control group (n � 11).In this study, the 3 active interventions (TM, mindful-ness training, and mental relaxation) were matched toeach other for teaching format, instructional time,home practice requirements (20 minutes 2 times aday), and expectancy of beneficial health outcomes.None of the interventions required other changes inlifestyle or personal beliefs.
The distinctive features of the 3 active treatmentswere as follows, The TM program was described asthe principal mind-body technique of Maharishi Con-sciousness-Based Health Care, a comprehensive tra-ditional system of natural medicine derived from theancient Vedic tradition.11,12 The TM program wasreported to include a simple, yet precise techniquewhereby the ordinary thinking process becomes qui-escent and a unique psychophysiologic state of “rest-ful alertness” appears to be gained.13 Details of theinstructional protocol have been published.14 Mindful-ness training, developed by Alexander et al,8 involvedlearning a guided attention technique to engage par-ticipants in novel and creative mental activities.8 Themental relaxation technique was designed to simulategeneric features of the TM program, hypothesized bysome researchers to be responsible for its relaxationeffects.15 The mental relaxation technique requiredsubjects to repeat silently to themselves familiar andbrief mental stimuli (e.g., a phrase or verse) during therelaxation period. Home practice compliance ratesafter 3 months for each of the 3 active treatmentgroups were 80% for TM subjects who practicedregularly compared with 48% for the mindfulnesstraining group and 47% for the mental relaxationgroup. Systolic blood pressure was significantly lowerin the TM group than in the control group at 3 months.
Methods of the second trial have been published indetail.6,7 Briefly, this was a randomized controlledtrial of community-dwelling older African-Americans(43% men and 57% women; mean age 67 � 7.6 years)who had stage I and II hypertension (stages I and II;baseline blood pressure averaged 147/92 � 13.2/8.7mm Hg) and lived in Oakland, California. The 125subjects were randomized to a TM group (n � 40), aprogressive muscle relaxation group (PMR; n � 42),or a health education control group (n � 43). The 2active interventions (TM and PMR) were matched forexpectancy of benefits, teaching format, instructionaltime, and home practice (20 minutes 2 times a day).The TM program was taught according to the samestandardized procedures as in the Alexander et al8trial. The PMR technique followed previously stan-dardized and validated procedures of Bernstein andBorkovic,16 based on a classic muscle relaxation pro-gram by Jacobson.17 This technique involved system-atically relaxing the major muscle groups of the bodywith the goal of physical and mental relaxation. Thehealth education control group received lifestyle mod-ification information consistent with national guide-lines.18 The education control intervention was designedto reflect usual-care community practice. Original out-comes in this study included systolic and diastolicblood pressures and other cardiovascular risk factors
for cardiovascular disease (diet, exercise, weight, andsubstance use). The duration of the intervention was 3months. Home practice regularity rates for TM andPMR after 3 months were 97% and 81%, respectively.There were no significant changes in diet, exercise, orblood pressure medications. Systolic and diastolicblood pressures were significantly lower in the TMgroup than in the PMR and education controlgroups.6,7
Determination of mortality: Data from 202 subjectsfrom the 2 clinical trials were pooled.19 Vital statuswas obtained from the National Death Index (NationalCenter for Health Statistics, Research Triangle Park,North Carolina), which provided an estimated accu-racy of 93% to 98% in identifying vital status ofpatients.20 The database of subjects’ records wassearched against the National Death Index databasethrough December 31, 1998, the most recent yearavailable as of autumn 2000. There were 101 mortal-ities. Causes of deaths were obtained by using theNational Death Index Plus service, which providedcoded causes of death according to the InternationalClassification of Diseases, Ninth Revision. Cause-specific mortality was based on the underlying causeof death. Codes 401 to 438 were defined as cardio-vascular diseases, and codes 140 to 239.9 were de-fined as cancer.
Statistical analysis: Survival analyses were per-formed on the pooled individual patient data from the2 trials. This approach used meta-analysis of individ-ual patient data.19 Survival time was measured startingfrom the date of randomization in the original studies.The primary outcome was all-cause mortality aftercompletion of maximum follow-up. Secondary out-comes were mortality due to cardiovascular diseaseand cancer. Maximum follow-up periods were 18.8years in trial 1 and 9.0 years in trial 2. In addition,secondary analyses were conducted of survival during3 years (for comparison with the short-term publishedfindings from trial 1) and during 10 years.
Statistical comparisons were made across the fol-lowing treatment categories: the TM program (used inthe 2 trials), “other active behavioral treatments”(mindfulness training and mental relaxation in trial 1and PMR in trial 2), and the “usual-care” conditions(usual care in trial 1 and education control in trial 2).The usual-care groups constituted minimally con-trolled conditions for comparison with the 2 activetreatment categories. Comparisons of the TM programwith the other active behavioral therapies allowedcontrol of nonspecific therapeutic effects, such as mo-tivational factors and generalized effects of relaxationtherapies (e.g., effects of sitting with eyes closed andnovel deployment of attention). This was feasible be-cause all active therapies were taught and practiced instructurally similar formats.
Survival analyses were performed with the SAS sta-tistical package (SAS Institute, Cary, North Carolina).Data were checked for validity of the proportionalhazards assumption.21 This was done using a time-dependent covariate, Z � ln(t), where Z is the indi-cator variable for TM, and t is the time since random-
SYSTEMIC HYPERTENSION/MEDITATION AND MORTALITY 1061
ization, in Cox’s regression analysis of survival times.The time-dependent covariate was highly statisticallysignificant in this analysis (p �0.0001), indicating thatthere was a significant violation of the proportionalhazards assumption. Therefore, survival times of treat-ment groups were statistically compared using thegeneralized Wilcoxon’s test that has been recom-mended for analysis of nonproportional hazards.21
Stratification on study location was used to adjust fordemographic differences between studies. The effectof treatment on survival was quantified as relative riskthat was estimated from Cox’s regression analysis ofsurvival times. Two-sided p values �0.05 were con-sidered statistically significant. Survival analyses wereconducted with “intention-to-treat” and “on-trial”principles. Subjects who were randomly allocated andcompleted post-testing at 3 months were included in
the on-trial analyses. Missing data were handled in theintention-to-treat analysis by the method of multipleimputation.22
RESULTSFor all subjects in the pooled study, mean age �
SD was 72.0 � 10.6 years, 67% were women, andsystolic blood pressure was 144 � 15.6 mm Hg. Meanfollow-up times were 8.45 � 4.83 years in study 1,7.13 � 2.15 years in study 2, and 7.63 � 3.48 yearsfor studies 1 and 2. Table 1 presents the number ofall-cause and disease-specific mortality events in eachstudy for the on-trial analysis. Table 2 presents thenumber of mortality events for the intention-to-treatanalysis. Table 3 presents the relative risks and sig-nificance levels for the primary and secondary out-comes of cardiovascular disease and cancer mortality
TABLE 1 Numbers of Deaths by Treatment Group in Each Clinical Trial*
Clinical Trial Treatment Group No. of Subjects
No. of deaths
All Causes Cardiovascular Diseases Cancers
Boston (Alexander et al8) TM 20 20 11 1Mindfulness 20 20 10 4Mental relaxation 22 21 12 4Usual care 11 10 4 1
*For subjects included in intention-to-treat analyses.
TABLE 3 Effects of Interventions on Mortality in the Pooled Clinical Trials
Cause of Death Statistical Comparison
Relative Risk
Significance (2 trials)Boston8 Oakland6 2 Trials
All cause TM vs usual care 1.13 0.60 0.89 0.21TM vs other active therapy 0.79 0.58 0.73 0.03Other active therapy vs usual care 1.33 1.16 1.26 0.43TM vs combined controls 0.85 0.61 0.77 0.04
Cardiovascular disease TM vs usual care 1.65 0.16 0.82 0.10TM vs other active therapy 0.80 0.23 0.68 0.06Other active therapy vs usual care 1.79 0.69 1.23 0.74TM vs combined controls 0.92 0.19 0.70 0.05
Cancer TM vs usual care 0.54 0.69 0.65 0.47TM vs other active therapy 0.20 0.69 0.44 0.09Other active therapy vs usual care 2.49 1.25 1.56 0.52TM vs combined controls 0.24 0.73 0.51 0.15
1062 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
for each study and for the 2 studies combined. Aspresented in Table 3, the relative risk of all-causemortality for the TM group compared with combinedcontrol groups was 0.77 (p � 0.039; Figure 1). Formortality due to cardiovascular disease, the relativerisk was 0.70 (p � 0.045); for mortality due to cancer,the relative risk was 0.51 (p � 0.153). The intention-to-treat results showed a similar pattern. That is, forthe 2 trials pooled, the relative risks were 0.75 (p �0.062) for the TM group compared with combinedcontrol groups for all-cause mortality, 0.62 (p �0.025) for cardiovascular mortality, and 0.53 (p �0.190) for cancer mortality.
To further evaluate the effects of stress-decreasingintervention on intermediate time points of survival,analyses were conducted for 3 and 10 years of follow-up. The results of the TM versus combined controlscomparison for all-cause mortality showed relativerisks of 0.11 (p � 0.013) at 3 years and 0.60 (p �0.025) at 10 years of follow-up. Intention-to-treatanalyses showed similar results, i.e., relative riskswere 0.21 (p � 0.046) at 3 years and 0.62 (p � 0.044)at 10 years.
DISCUSSIONThe results of this retrospective long-term fol-
low-up of 2 randomized controlled trials of decreasingstress and high blood pressure in older subjects sug-gest that a selected behavioral approach, the TM pro-gram, is associated with significantly lower mortalityrates compared with other behavioral interventionsand usual care. The decreases in risk were 23% forall-cause mortality and 30% for cardiovascular mor-tality, and there was a trend for cancer mortality overthe 7.6-year mean, 18.8-year maximum follow-up pe-riod. The results of intention-to-treat analyses weresimilar to those of on-trial analyses. There were sim-ilar risk decreases for the 2 independent study sam-ples: elderly white subjects who had prehypertensionand who resided in Boston, Massachusetts, and olderAfrican-Americans who had stage I and II hyperten-sion and lived in Oakland, California.
This is the first pooled analysis of randomizedcontrolled trials on long-term effects of nonpharma-cologic interventions for hypertension on mortality.
Although the sample of 202 African-American andwhite men and women in the present study was rela-tively modest, these preliminary results suggest thatan effective stress-reducing intervention may decreasemortality rates for all causes and cardiovascular dis-ease.
A meta-analysis of pooled clinical trials of phar-macologic antihypertensive therapy indicated an av-erage decrease of 12 to 13 mm Hg in systolic bloodpressure over 4 years.23 This was associated with a13% decrease in all-cause mortality and a 25% de-crease in cardiovascular mortality.24 The short-termfollow-up of blood pressure changes in the 2 trials ofdecreasing stress included in the present analysis as-sociated the practice of TM with a decrease of 10 to11 mm Hg in systolic blood pressure. Diastolic bloodpressure was measured only in study 2, and there thedecrease averaged 6 mm Hg. Whether these or otherdecreases in blood pressure were maintained duringthe 7.6 years of average follow-up of survival is un-known, because in the present retrospective analysisonly mortality data from national databases were col-lected. However, the 23% decrease in the primaryoutcome of all-cause mortality and 30% decrease incardiovascular mortality may be considered somewhatsimilar to decreased risks reported with antihyperten-sive therapy in middle-aged and older adults who havehypertension.23,24 Larger clinical trials that are pro-spectively designed to assess blood pressure, treat-ment compliance, mechanistic variables, and mortalityover the long-term are required to more definitivelyevaluate these effects.
Another relevant consideration when comparingthe effectiveness and utility of interventions for hy-pertension is the significant rate of adverse effectsassociated with antihypertensive drugs.25,26 Previousstudies on physiologic outcomes and quality-of-lifeeffects of the TM program reported a range of im-provements in physical and psychological health inadults generally and in older subjects specifical-ly.12,27,28 These and the present effects on mortalityare consistent with theoretical mechanisms of actionof the TM program, which suggest a diverse yetintegrated effect on mind–body health.11,29
In conclusion, the present findings suggest thatnonpharmacologic decrease of stress with the TMprogram, in addition to usual care, may be associatedwith long-term decreases in mortality in older popu-lations that have high blood pressure.
Acknowledgment: The investigators are grateful toWilliam Sheppard, PhD, and Marcelino Aguilar-Cer-vantes, MS, for technical assistance and editorial consul-tation.
1. Eisenberg DM, Delbanco TL, Berkey CS, Kaptchuk TJ, Kupelnick B, Kuhl J,Chalmers T. Cognitive behavioral techniques for hypertension: are they effec-tive? Ann Intern Med 1993;118:964–972.2. Spence J, Barnett P, Linden W, Ramsden V, Taenzer P. Lifestyle modificationsto prevent and control hypertension. 7 Recommendations on stress management.Can Med Assoc J 1999;160:46–50.3. Orme-Johnson D, Walton K. All approaches to preventing or reversing effectsof stress are not the same. Am J Health Promot 1998;12:297–299.
FIGURE 1. Kaplan-Meier survival curves for all-cause mortalityfor subjects who completed post-testing in the original studies(relative risk 0.77, p � 0.04).
SYSTEMIC HYPERTENSION/MEDITATION AND MORTALITY 1063
4. Lin MC, Nahin R, Gershwin ME, Longhurst JC, Wu KK. State of comple-mentary and alternative medicine in cardiovascular, lung, and blood research:executive summary of a workshop. Circulation 2001;103:2038–2041.5. Alexander CN, Robinson P, Orme-Johnson DW, Schneider RH, Walton KG.Effects of Transcendental Meditation compared to other methods of relaxationand meditation in reducing risk factors, morbidity and mortality. Homeostasis1994;35:243–264.6. Schneider RH, Staggers F, Alexander C, Sheppard W, Rainforth M, KondwaniK, Smith S, King CG. A randomized controlled trial of stress reduction forhypertension in older African Americans. Hypertension 1995;26:820–827.7. Alexander CN, Schneider R, Staggers F, Sheppard W, Clayborne M, RainforthM, Salerno J, Kondwani K, Smith S, Walton K, Egan B. A trial of stress reductionfor hypertension in older African Americans (part II): sex and risk factor sub-group analysis. Hypertension 1996;28:228–237.8. Alexander CN, Langer EJ, Newman RI, Chandler HM, Davies JL. Transcen-dental Meditation, mindfulness, and longevity: an experimental study with theelderly. J Pers Social Psychol 1989;57:950–964.9. Barnes VA, Schneider RH, Alexander CN, Rainforth M, Staggers F, SalernoJ. Impact of transcendental meditation on mortality in older African Americans—eight year follow-up. J Social Behav Pers 2005;17:201–216.10. Alexander CN, Barnes VA, Schneider RH, Langer EJ, Newman RI, ChandlerHM, Davies JL, Rainforth M. A randomized controlled trial of stress reduction oncardiovascular and all-cause mortality in the elderly: results of 8 and 15 yearfollow-ups. Circulation 1996;93:P19.11. Nader T. Human Physiology: Expression of Veda and the Vedic Literature.Vlodrop, Holland: Maharishi Vedic University, 2000.12. Schneider RH, Alexander CN, Salerno J, Robinson D, Fields J, Nidich S.Disease prevention and health promotion in the aging with a traditional system ofnatural medicine: Maharishi Vedic Medicine. J Aging Health 2002;14:57–78.13. Jevning R, Wallace RK, Biedebach M. The physiology of meditation: areview. A wakeful hypometabolic integrated response. Neurosci Biobehav Rev1992;16:415–424.14. Roth R. Maharishi Mahesh Yogi’s Transcendental Meditation. Washington,DC: Primus, 2002:90–102.15. Smith JC. Therapeutic effects of transcendental meditation with controls forexpectation of relief and daily sitting. J Consult Clin Psychol 1976;44:630–637.16. Bernstein DA, Borkovic TD. Progressive Relaxation Training: A Manual forthe Helping Profession. Chicago: Illinois University Press, 1973.
17. Jacobson E. Progressive Relaxation. Chicago: University of Chicago Press,1992:28–100.18. Chobanian A, Bakris G, Black H, National High Blood Pressure EducationProgram Coordinating Committee. The seventh report of the Joint NationalCommittee on prevention, detection, evaluation, and treatment of high bloodpressure. JAMA 2003;289:2560–2572.19. Stewart LA, Clarke M. Practical methodology of meta-analysis using updatedindividual patient data. Stat Med 1995;14:2057–2079.20. Edlavitch SA, Baxter J. Comparability of mortality follow-up before and afterthe National Death Index. Am J Epidemiol 1988;127:1164–1178.21. Collett D. Modelling Survival Data in Medical Research. London: Chapman& Hall, 1994.22. Little RJA, Rubin DB. Statistical Analysis With Missing Data. New York:Wiley, 1987.23. Psaty BM, Smith NL, Siscovick DS, Koepsell TD, Weiss NS, Heckbert SR,Lemaitre RN, Wagner EH, Furberg CD. Health outcomes associated with anti-hypertensive therapies used as first-line agents. A systematic review and meta-analysis. JAMA 1997;277:739–745.24. Staessen J, Gasowski J, Wang J, Thijs L, Den Hond E, Boissel JP, Coope J,Ekborm T, Gueyffier F, Liv L, et al. Risks of untreated and treated isolatedsystolic hypertension in the elderly: meta-analysis of outcome trials. Lancet2000;355:865–872.25. Hansson L, Zanchetti A, Carruthers S. Effects of intensive blood pressurelowering and low-dose aspirin in patients with hypertension: principle results ofthe Hypertension Optimal Treatment (HOT) randomized trial. HOT study group.Lancet 1998;351:1755–1762.26. Cutler J. Calcium channel blockers for hypertension—uncertainty continues.N Engl J Med 1998;338:679–680.27. Walton KG, Schneider RH, Nidich SI, Salerno JW, Nordstrom CK, Bairey-Merz CN. Psychosocial stress and cardiovascular disease 2: effectiveness of thetranscendental meditation technique in treatment and prevention. Behav Med2002;28:106–123.28. Walton K, Schneider R, Nidich S. Review of controlled research on thetranscendental meditation program and cardiovascular disease—risk factors, mor-bidity and mortality. Cardiol Rev 2004;12:262–266.29. Nader T, Rothenberg S, Averbach R, Charles B, Fields JZ, Schneider RH.Improvements in chronic diseases with a comprehensive natural medicine ap-proach: a review and case series. Behav Med 2000;26:34–46.
1064 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Matrix Metalloproteinases and TheirInhibitors in Malignant and Autoreactive
Pericardial EffusionSteffen Lamparter, MD, Michael Schoppet, MD, Michael Christ, MD,
Sabine Pankuweit, PhD, and Bernhard Maisch, MD
Matrix metalloproteinases (MMPs) are proteolytic en-zymes essentially involved in tissue remodeling and tu-mor invasion, and their activity is counterbalanced byendogenous antagonists, the tissue inhibitors of matrixproteinases (TIMPs). Recent reports have suggested apotential role of MMPs in the evolution of pericardialeffusion (PE). In this study, we determined the levels ofMMP-2 and MMP-9 and their inhibitors TIMP-1 andTIMP-2 in 19 patients who had malignant PE that wasconfirmed by histology or cytology and 30 patients whohad nonmalignant, autoreactive PE compared with peri-cardial fluid of 19 patients who had preserved left ventric-ular function and who underwent aortocoronary bypasssurgery for control. Samples were assayed by zymogra-phy, immunoblotting, and quantitative enzyme-linkedimmunosorbent assay. We found significantly higherMMP-2 levels in malignant PE than in pericardial fluid(2,906 � 348 vs 1,493 � 114 ng/ml, p � 0.0005) orautoreactive PE (2,079 � 269 ng/ml, p � 0.01). No
significant differences in MMP-9 levels were found be-tween malignant PE and autoreactive PE (83 � 28.6 vs106 � 30.4 ng/ml, p � 0.22), whereas MMP-9 wasbelow the detection limit in pericardial fluid. No differ-ences in TIMP-1 levels were found across the differentstudy groups, whereas compared with pericardial fluid,TIMP-2 levels were significantly lower in autoreactive PE(113 � 18.9 vs 187 � 12.2 ng/ml, p � 0.002). Inaddition, there was a trend to lower TIMP-2 levels inmalignant PE (137 � 27.1 ng/ml, p � 0.07). Thepresent findings indicate that proteolytic enzymes andtheir inhibitors are involved in the pathogenesis of PE,with an expression pattern that depends on etiology.The involvement of MMP-2 in the pathogenesis of ma-lignant PE may indicate a potential role of MMP inhibi-tors in the control of malignant PE. �2005 by ExcerptaMedica Inc.
(Am J Cardiol 2005;95:1065–1069)
Matrix metalloproteinases (MMPs) constitute afamily of functionally related, zinc-containing
proteases that are capable of degrading components ofthe extracellular matrix. Activation of some species ofMMPs has been demonstrated in cardiovascular re-modeling1 and is believed to be involved in the evo-lution and progression of heart failure.2,3 IncreasedMMP activity has been associated with ventricularrupture after myocardial infarction, dilated cardiomy-opathy, hypertension, and abdominal aortic aneu-rysm.4–6 Of note, activation of MMPs is involved inneovascularization at sites of tissue repair by disrupt-ing basement membranes, thereby facilitating the mi-gration of endothelial cells,7 and in the pathogenesisof inflammatory processes.
Pericardial effusion (PE) is a feature in differentpathologic conditions, but the etiology of PE is diffi-cult to elucidate on the basis of clinical assessment.Because PE may progress to fatal cardiac compres-sion, evaluation of PE and accurate diagnosis of theunderlying disease are of paramount interest. PE maybe caused by viral or bacterial infections and fre-
quently occurs in malignancy.8 Similarly, malignancymay cause pleural effusion in which MMP activity hasbeen demonstrated to be increased.9 Of interest, inhi-bition of MMPs by intrapleural instillation of specificMMP inhibitors appears to retard tumor invasion andneovascularization in this condition.10 However, thepathophysiologic role of MMPs in malignant PE andautoreactive PE is currently unknown. Thus, we soughtto determine whether MMPs are differentially expressedin malignant PE versus autoreactive PE.
METHODSPatients: We studied samples from 49 consecutive
patients who had PE and who underwent pericardio-centesis and pericardioscopically guided pericardialbiopsy for therapeutic and/or diagnostic reasons afterinformed consent and approval of the institutionalreview board. Pericardial fluid from 19 patients thatwas obtained during open heart surgery for coronaryartery disease served as the control. In addition, thepresence of cardiotropic viruses (coxsackievirus BRNA, parvovirus B19 DNA, cytomegalovirus DNA,and adenovirus type 2 DNA) was examined in allsamples of PE and pericardial biopsies. For extractionof DNA/RNA from PE and pericardial biopsies, theQIAamp Blood Mini Kit and the QIAamp Tissue Kit(Qiagen, Hilden, Germany) were used. Conditions forpolymerase chain reaction and primers have been de-scribed previously.11 The diagnosis of malignant PEwas based on the presence of malignant cells in the PE
From the Diakonie Krankenhaus Wehrda, Department of Internal Med-icine; and the Department of Internal Medicine and Cardiology,Philipps-University, Marburg, Germany. Manuscript received August13, 2004; revised manuscript received and accepted December 21,2004.
Address for reprints: Steffen Lamparter, MD, Diakonie KrankenhausWehrda, Department of Internal Medicine, Hebronberg 5, D-35041Marburg, Germany. E-mail: [email protected].
or the presence of an invasive tumor in pericardio-scopically guided pericardial biopsy specimens. Thediagnosis of autoreactive PE was established with thefollowing criteria: (1) increased numbers of lympho-cytes and mononuclear cells and the presence of an-tibodies against heart muscle tissue (antisarcolemmalantibodies) in PE; (2) exclusion of active viral infec-tion in PE and pericardial/epicardial biopsies (no virusdetection, no immunoglobulin M titer against cardio-tropic viruses in PE, and negative polymerase chainreaction for major cardiotropic viruses); (3) exclusionof infection by Mycobacterium tuberculosis, Borreliaburgdorferi, Chlamydia pneumoniae, or other bacte-rial agents as assessed by negative polymerase chainreaction and/or cultures; (4) exclusion of malignancyby cytology of PE and in pericardial/epicardial biop-sies by histopathologic analysis; and (5) exclusion of
patients who developed PE after ra-diation and chemotherapy or patientswho had metabolic disorders anduremia. Patients who had autoreac-tive PE and acute or chronic recur-rent pericarditis were classified as hav-ing autoreactive PE.12,13 All pericardialfluid samples were obtained from pa-tients who had coronary artery dis-ease with preserved left ventricularfunction and underwent aortocoro-nary bypass surgery.
Echocardiographic studies: Compre-hensive echocardiographic examina-tions were performed in all patients,and left ventricular end-diastolic diam-eter and ejection fraction were mea-
sured according to guidelines of the American Societyof Echocardiography.14 PE was identified according tothe classification reported by Horowitz et al.15
Sampling of PE or pericardial fluid: All PE and peri-cardial fluid samples were immediately transferred intochilled sterile tubes that contained a proteinase inhibitorcocktail (Complete, Roche, Penzberg, Germany) andsubsequently stored at �80°C.
Total protein content in PE and pericardial fluid:Amounts of protein in pericardial fluid and PE weredetermined by Bradford’s method using bovine serumalbumin as a standard.16 Pericardial samples were notconcentrated before analysis to avoid loss of enzy-matic activity.
Gelatin zymography: MMP activity was determinedby zymographic analysis under denaturing and nonre-ducing conditions with an equal amount of 10 �g of
TABLE 1 Demographic Characteristics of Patients With or Without Pericardial Effusion
Values are mean � SEM, or numbers of patients (percentages).*Analysis of variance/chi-square test.†p �0.05 versus PE; ‡p �0.05 versus aPE.aPE � autoreactive PE; mPE � malignant PE; PF � pericardial fluid.
TABLE 2 Characteristics of Pericardial Fluid and Pericardial Effusions inStudy Patients
Values are mean � SEM.*Analysis of variance/chi-square test.†p �0.05; ‡p � 0.07 versus PF; §p �0.05 versus aPE.Abbreviations as in Table 1.
1066 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
protein of each sample as previously described.17
Bands were visualized by negative staining with Coo-massie blue. Prestained standard proteins were usedfor identification of the molecular size of lytic bands.
Western blot analysis: Sodium dodecyl sulfatepolyacrylamide gel electrophoresis was performed us-ing the buffer system described by Laemmli.18 Pro-teins were transferred to nitrocellulose according tothe method of Towbin et al,19 and equal loading wasconfirmed by Ponceau S staining. Immunoblotting
was performed by sequential incuba-tion with anti–MMP-2 or anti–MMP-9(Oncogene, Cambridge, Massachu-setts) as primary antibodies for 2 hoursin phosphated buffered saline andTween 0.05% (v/v) at 22°C, followedby staining with a peroxidase-conju-gated secondary antibody. Bands werevisualized with enhanced chemilumi-nescence (Amersham, Buckingham-shire, United Kingdom).
Measurements of MMPs and tissueinhibitors of matrix proteinases by en-zyme-linked immunosorbent assay:MMP and tissue inhibitor of matrixproteinase (TIMP) protein levels inPE and pericardial fluid were mea-sured by a quantitative enzyme im-munoassay (Amersham Biotech,Freiburg, Germany). The MMP en-zyme-linked immunosorbent assaydetects free pro–MMP-2 and pro–MMP-2 complexed with TIMP-2and free pro–MMP-9 and pro–MMP-9 complexed with TIMP-1,whereas active MMP-2 and MMP-9are not detected. The TIMP assaysdetect total TIMP (free TIMP andTIMP complexed with MMPs). In-tra- and interassay variations were5% and 9%, respectively, for theMMP enzyme-linked immunosor-bent assays and 9% and 13%, respec-tively, for the TIMP enzyme-linkedimmunosorbent assays. The lower lim-its of detection were 0.37, 0.6, 1.25,and 3.0 ng/ml for MMP-2, MMP-9,TIMP-1, and TIMP-2, respectively.
Statistical analysis: Data aremeans � SEM, and results werecompared with 1-way analysis ofvariance. Log transformation of datawas performed when data were notnormally distributed. Fisher’s exact ttest and Tukey’s post hoc analysisfollowed for comparison of groups.Chi-square analysis was used for cat-egorical data. A p value �0.05 wasregarded as statistically significant,and Bonferroni’s correction was per-formed for multiple comparisons.
RESULTSWe identified malignant PE in 19 patients (18
patients had carcinoma and 1 patient had high-gradelymphoma) and autoreactive PE in 30 patients (Table 1).Cardiotropic viruses (coxsackievirus B RNA, parvo-virus B19 DNA, cytomegalovirus DNA, and adeno-virus type 2 DNA) were undetectable in pericardialbiopsies and in all malignant PE and autoreactive PEsamples. Compared with autoreactive PE, we identi-fied significantly more leukocytes in malignant PE
FIGURE 1. (A) Samples of pericardial fluid (PF), malignant PE (mPE), and nonmalig-nant autoreactive PE (aPE) were studied by gelatin zymography; 2 representative pa-tients per study group are shown. Analysis demonstrated lytic bands at 72 and 92kD, corresponding to MMP-2 and MMP-9 (representative zymogram), respectively. (B)Compared with PF, immunoreactivity against MMP-2 and MMP-9 is increased in aPEand aPE (representative immunoblot). (C) MMP-2 levels as measured by an enzyme-linked immunosorbent assay are significantly increased in aPE versus PF (*p <0.01)and aPE (#p <0.05). Values are means � SEM.
MISCELLANEOUS/MMP AND TIMP IN PE 1067
(Table 2). Lymphocytically rich PEs were identified in20 of 30 patients who had autoreactive PE. Withgelatin zymography, lytic bands were detected at 72and 92 kD, which corresponded to MMP-2 and MMP-9,respectively. Compared with pericardial fluid, lytic ac-tivity was increased in malignant PE and autoreactivePE and was associated with increased immunoreactiv-ity for MMP-2 and MMP-9 (Figure 1). Using a quan-titative enzyme-linked immunosorbent assay, MMP-2activity was detectable in all PE and pericardial fluidsamples. MMP-2 levels were significantly higher inmalignant than in autoreactive PE (2,906 � 348 ng/ml, 95% confidence interval 2,346 to 3,772 vs 2,079 �269 ng/ml, 95% confidence interval 1,531 to 2,559, p �0.02), or pericardial fluid (1,493 � 114 ng/ml, 95%confidence interval 1,268 to 1,718, p � 0.01; Figure 1).We found no significant differences in MMP-9 levelsbetween malignant PE and autoreactive PE, whichwas most likely due to a high variation of MMP-9expression in the 2 groups: 7 of 19 patients (37%) whohad malignant PE (range 16.5 to 220.3 ng/ml) and 15of 30 patients (50%) who had autoreactive PE (range16.26 to 397.3 ng/ml) had detectable MMP-9 levels.MMP-9 was below the detection limit in all pericar-dial fluid samples (Table 2). No significant associa-tions were found for white blood cell counts andMMP-2 or MMP-9 levels in PE samples (MMP-2: y� 2,115 � 14.1 � x, r2 � 0.025; MMP-9: y � 36.6� 0.3 � x, r2 � 0.005). Moreover, no significantdifferences in TIMP-1 and TIMP-2 levels betweenmalignant PE and autoreactive PE were detected.However, compared with pericardial fluid, TIMP-2levels were significantly lower in autoreactive PE, andthere was a trend to lower TIMP-2 levels in malignantPE versus pericardial fluid (Table 2). We found noassociation of MMP-2/MMP-9 and TIMP-1/TIMP-2levels in PE or pericardial fluid with intrapericardialcytokine levels of interferon-�, interleukin-6, interleu-kin-8, or transforming growth factor-�1 (data notshown). Tumor necrosis factor-�, a critical modulatorof MMP activity, was above the detection limit of theassay in PE samples from only 2 patients (1 hadautoreactive PE and 1 had malignant PE).
DISCUSSIONThus far, the role of MMPs in the pathogenesis of
PE is not fully understood. In this study, we deter-mined MMP and TIMP levels in pericardial fluid ofcontrol patients, patients who had autoreactive PE,and those who had malignant PE. Compared withpericardial fluid and autoreactive PE, MMP-2 levelswere high in malignant PE. No significant differencesin MMP-9 levels were found between autoreactive PEand malignant PE, whereas MMP-9 was below thedetection limit in pericardial fluid. Because MMP-2and MMP-9 are antagonized by endogenous TIMPs,we further determined TIMP-1 and TIMP-2 levels. Nosignificant differences in TIMP-1 levels were ob-served across groups, whereas TIMP-2 levels weresignificantly lower in autoreactive PE than in pericar-dial fluid, and there was a trend to lower TIMP-2levels in malignant PE than in pericardial fluid. Hy-
drostatic and/or colloid osmotic forces and pathologicmembrane permeability are involved in the pathogen-esis of transudative or exudative PE. Cytokines, en-zymes, and mediators are formed in heart residentcells and secreted in a paracrine or autocrine mannerthat leads to an increase of their concentration in thepericardial cavity. It has been suggested that increasedlevels of these mediators reflect the pathogenetic pro-cesses that initiate PE, and several cytokines such asbasic fibroblast growth factor, endothelin-1, and atrialnatriuretic peptide have been identified in PE and arelikely to be involved in the pathogenesis of autoreac-tive PE and malignant PE.20,21 Because MMP expres-sion and activation are tightly regulated by cytokines,MMPs may be increased in inflammatory or malignantPE. MMP-2 levels were high in malignant PE com-pared with pericardial fluid of controls, whereas nodifferences were evident in autoreactive PE. Our dataare in accordance with the findings of Eickelberg etal,9 who reported an increase of MMPs in malignantpleural effusion, which supports the hypothesis thatMMPs have an important role in malignant PE. In arecent study, high MMP-2 levels were reported inpericardial fluid of patients who had dilated cardio-myopathy and impaired left ventricular function,22
and high serum levels of MMPs have been found inpatients who have congestive heart failure.23 An in-crease of MMP levels in congestive heart failure mayresult from the stimulation of cytokines, which areincreased in the myocardium of these patients.24 How-ever, only patients who had preserved left ventricularfunction were included in our study. We found noassociation between cytokine and MMP levels in PEand therefore speculate that MMP expression may bedue to MMP secretion by malignant cells and the largevariation of MMP levels may rather reflect the tumorload in PE.25 PE was evident for variable time inter-vals, which may have caused differences in MMPlevels. In the present study, we did not address thecellular source of MMPs and TIMPs that may includecardiac myocytes.26 In malignant PE, tumor cells are apotential cellular source of increased MMP activity,which may be associated with tumor progression andmetastatic disease.27 Tumor spreading has been asso-ciated with angiogenesis, a process tightly controlledby MMP expression.28 Although we included onlypatients who had histologically proved intrapericardialneoplasia, a systemic rather than an intrapericardialorigin of MMPs has to be considered. In autoreactivePE, direct visualization of the pericardial surfaceshows an inflammatory response of the pericardi-um,12,29 suggesting that inflammatory tissues may alsobe a source of MMPs and TIMPs.27 Because we didnot include patients who had active viral pericarditis,MMP expression by virus-infected cells seems un-likely.
There are some limitations of our study. (1) Due tothe small number of patients included in this observa-tional study, the conclusions of our study may not betransferable to the general population and thereforeneed further evaluation. (2) Patients who had coronaryartery disease were used as a control population. It
1068 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
may be hypothesized that MMP expression in pericar-dial fluid of patients who have coronary artery diseaseis upregulated because increased collagenolytic activ-ity has been demonstrated in coronary atheroscleroticplaques and in the circulation, which may subse-quently accumulate in pericardial fluid.30 However,MMP-2 levels were lowest in pericardial fluid, indi-cating that, even if upregulation of MMPs is a featurein atherosclerosis, gelatinase levels in PE remain sig-nificantly higher than those in pericardial fluid. (3) Apotential limitation may be that the repertoire of cy-tokines analyzed may not be sufficient to excludeinflammation as significant player in altering the bal-ance of MMPs to TIMPs in autoreactive PE. (4) Dueto the large variation of data, which was not expectedduring the sample calculation before the study, ourstudy may have been underpowered to detect impor-tant differences in MMP/TIMP abundance.
1. Woessner JFJ. The family of matrix metalloproteinases. Ann NY Acad Sci1994;732:11–21.2. Armstrong PW, Moe GW, Howard RJ, Grima EA, Cruz TF. Structuralremodeling in heart failure: gelatinase induction. Can J Cardiol 1994;10:214–220.3. Spinale FG, Coker ML, Thomas CV, Walker JD, Mukherjee R, Hebbar L.Time-dependent changes in matrix metalloproteinase activity and expressionduring the progression of congestive heart failure: relation to ventricular andmyocyte function. Circ Res 1998;82:482–495.4. Campbell SE, Diaz-Arias AA, Weber KT. Myocardial rupture postinfarction.In: Weber KT, ed. Wound Healing in Cardiovascular Disease. Armonk, NY:Futura Publishing, 1995:49–60.5. Weber KT. Cardiac interstitium in health and disease: the fibrillar collagennetwork. J Am Coll Cardiol 1989;13:1637–1652.6. Armstrong PJ, Johanning JM, Calton WC, Delatore JR, Franklin DP, Han DC,Carey DJ, Elmore JR. Differential gene expression in human abdominal aorta:aneurysmal versus occlusive disease. J Vasc Surg 2002;35:346–355.7. Mignatti P, Rifkin DB. Plasminogen activators and matrix metalloproteinasesin angiogenesis. Enzyme Protein 1996;49:117–137.8. Maisch B, Ristic AD, Pankuweit S, Neubauer A, Moll R. Neoplastic pericar-dial effusion. Efficacy and safety of intrapericardial treatment with cisplatin. EurHeart J 2002;23:1625–1631.9. Eickelberg O, Sommerfeld CO, Wyser C, Tamm M, Reichenberger F, Bardin PG,Soler M, Roth M, Perruchoud AP. MMP and TIMP expression pattern in pleuraleffusions of different origins. Am J Respir Crit Care Med 1997;156:1987–1992.10. Macaulay VM, O’Byrne KJ, Saunders MP, Braybrooke JP, Long L, GleesonF, Mason CS, Harris AL, Brown P, Talbot DC. Phase I study of intrapleuralbatimastat (BB-94), a matrix metalloproteinase inhibitor, in the treatment ofmalignant pleural effusions. Clin Cancer Res 1999;5:513–520.11. Maisch B, Ristic A, Pankuweit S. Intrapericardial treatment of autoreactivepericardial effusion with triamcinolone. The way to avoid side effects of systemiccorticosteroid therapy. Eur Heart J 2002;23:1503–1508.
12. Maisch B. Pericardial diseases, with a focus on etiology, pathogenesis,pathophysiology, new diagnostic imaging methods, and treatment. Curr OpinCardiol 1994;9:379–388.13. Maisch B, Ristic AD. The classification of pericardial disease in the age ofmodern medicine. Curr Cardiol Rep 2002;4:13–21.14. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, FeigenbaumH, Gutgesell H, Reichek N, Sahn D, Schnittger I. Recommendations for quanti-tation of the left ventricle by two-dimensional echocardiography. AmericanSociety of Echocardiography Committee on Standards, Subcommittee on Quan-titation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr 1989;2:358–367.15. Horowitz MS, Schultz CS, Stinson EB, Harrison DC, Popp RL. Sensitivityand specificity of echocardiographic diagnosis of pericardial effusion. Circulation1974;50:239–247.16. Bradford MM. A rapid and sensitive method for the quantitation of micro-gram quantities of protein utilizing the principle of protein-dye binding. AnalBiochem 1976;72:248–254.17. Tyagi SC, Matsubara L, Weber KT. Direct extraction and estimation ofcollagenase(s) activity by zymography in microquantities of rat myocardium anduterus. Clin Biochem 1993;26:191–198.18. Laemmli UK. Cleavage of structural proteins during the assembly of the headof bacteriophage T4. Nature 1970;227:680–685.19. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins frompolyacrylamide gels to nitrocellulose sheets: procedure and some applications.Proc Natl Acad Sci USA 1979;76:4350–4354.20. Fujita M, Ikemoto M, Kishishita M, Otani H, Nohara R, Tanaka T, TamakiS, Yamazato A, Sasayama S. Elevated basic fibroblast growth factor in pericar-dial fluid of patients with unstable angina. Circulation 1996;94:610–613.21. Horkay F, Szokodi I, Selmeci L, Merkely B, Kekesi V, Vecsey T,Vuolteenaho O, Ruskoaho H, Juhasz-Nagy A, Toth M. Presence of immunore-active endothelin-1 and atrial natriuretic peptide in human pericardial fluid. LifeSci 1998;62:267–274.22. Miyamoto S, Nagaya N, Ikemoto M, Tambara K, Kihara Y, Onozawa Y,Hasegawa K, Yamazato A, Fujita M. Elevation of matrix metalloproteinase-2level in pericardial fluid is closely associated with left ventricular remodeling.Am J Cardiol 2002;89:102–105.23. Wilson EM, Gunasinghe HR, Coker ML, Sprunger P, Lee-Jackson D,Bozkurt B, Deswal A, Mann DL, Spinale FG. Plasma matrix metalloproteinaseand inhibitor profiles in patients with heart failure. J Card Fail 2002;8:390–398.24. Diwan A, Tran T, Misra A, Mann DL. Inflammatory mediators and the failingheart: a translational approach. Curr Mol Med 2003;3:161–182.25. Vihinen P, Kahari VM. Matrix metalloproteinases in cancer: prognosticmarkers and therapeutic targets. Int J Cancer 2002;99:157–166.26. Coker ML, Doscher MA, Thomas CV, Galis ZS, Spinale FG. Matrix metal-loproteinase synthesis and expression in isolated LV myocyte preparations. Am JPhysiol 1999;277:H777–H787.27. Visse R, Nagase H. Matrix metalloproteinases and tissue inhibitors ofmetalloproteinases: structure, function, and biochemistry. Circ Res 2003;92:827–839.28. Luttun A, Dewerchin M, Collen D, Carmeliet P. The role of proteinases inangiogenesis, heart development, restenosis, atherosclerosis, myocardial isch-emia, and stroke: insights from genetic studies. Curr Atheroscler Rep 2000;2:407–416.29. Di Girolamo N, Tedla N, Lloyd A, Wakefield D. Expression of matrixmetalloproteinases by human plasma cells and B lymphocytes. Eur J Immunol1998;28:1773–1784.30. Pasterkamp G, Schoneveld AH, Hijnen DJ, de Kleijn DP, Teepen H, van derWal AC, Borst C. Atherosclerotic arterial remodeling and the localization ofmacrophages and matrix metalloproteases 1, 2, and 9 in the human coronaryartery. Atherosclerosis 2000;150:245–253.
MISCELLANEOUS/MMP AND TIMP IN PE 1069
Prevalence of T-Wave Inversion BeyondV1 in Young Normal Individuals and
Usefulness for the Diagnosis ofArrhythmogenic Right Ventricular
Cardiomyopathy/DysplasiaFrank I. Marcus, MD
T-wave inversion in precordial leads V1 to V3 ispresent in <3% of apparently healthy subjects whoare 19 to 45 years of age but is present in 87% ofpatients who have arrhythmogenic right ventricularcardiomyopathy/dysplasia. T-wave inversion in leadV2 or V3 in a young or middle-aged patients who have
no apparent heart disease but do have ventriculararrhythmias of left bundle branch morphology shouldraise the suspicion of arrhythmogenic right ventricularcardiomyopathy/dysplasia. �2005 by Excerpta Med-ica Inc.
(Am J Cardiol 2005;95:1070–1071)
T-wave inversion in precordial leads beyond leadV1 is often dismissed as nonspecific or as persis-
tence of the juvenile T-wave pattern. The observationthat inverted T waves beyond lead V1 are common inchildren was reported by Lepeschkin1 in 1938. Hestudied normal chest electrocardiograms of 50 chil-dren from 2 weeks to 15 years of age and stated, “TheT wave is inverted on the right chest anteriorly andupright on the left. On transition, a diphasic T isfound. This transition is more to the left anteriorly inchildren than in grownups, and this deviation from themidline is greater the younger the child.” After the ageof 12 years, the T wave is usually inverted in lead V1and is usually upright in leads V2 to V6. If the T wavesremain inverted beyond lead V1 in a young patientwho has no overt heart disease, it is generally consid-ered to be of no clinical significance. However, afeature of an unusual disease, arrhythmogenic rightventricular cardiomyopathy/dysplasia (ARVC/D), isthe presence of T-wave inversion beyond lead V1 thatcan extend to lead V5 or V6. The sensitivity of thisfinding in patients who otherwise meet the task forcecriteria2 for having this disease is 87%.3 In a youngpatient who presents with ventricular arrhythmias, in-cluding premature ventricular beats, of left bundlebranch block morphology, the presence of T-waveinversion beyond lead V1 should raise the suspicion ofARVC/D. It is important to diagnose this disease,because it can cause arrhythmic death and may begenetically transmitted.
The concern arises as to the specificity of thejuvenile T-wave pattern for ARVC/D because it can
occur in patients who have normal electrocardio-grams. In other words, what is the prevalence of thejuvenile T-wave pattern in patients who have normalelectrocardiograms? It has been believed that the ju-venile T-wave pattern may be more common amongblacks and highly trained athletes. The literature wasreviewed in an attempt to answer these questions.Studies were excluded if electrocardiograms weretaken before the era of the precordial V-lead config-uration. Studies were also excluded if ST-T–wave ab-normalities were not specifically defined. The search waslimited to studies that identified T-wave inversionsin specific precordial leads. Biphasic T waves in precor-dial leads are not considered to be a juvenile pattern.
Suarez and Suarez4 studied the T-wave configura-tion in precordial leads at 3 age levels: 5 to 11, 12 to18, and 19 to 45 years. The results are presented inTable 1. At �12 years of age, the T waves wereinverted in lead V2 in 45% to 65% of children, 30% to40% had T-wave inversion in lead V3, and 5% to 20%had T-wave inversion beyond lead V3. In females andmales 12 to 18 years old, there was a 10% to 20%incidence of T-wave inversion in lead V2 and a 5%incidence of T-wave inversion in lead V3. No patienthad a T-wave inversion beyond lead V3. In 19- to45-year-old women, there was a 13% prevalence ofT-wave inversion in lead V2 and a 3% prevalence inlead V3. Of the 50 men in this age group, there was noT-wave inversion beyond lead V1. Hiss et al5 analyzedT-wave changes in 67,375 presumably healthy menwho had no symptoms and were on flying status withthe United States Air Force. All men were 20 to 45years of age. Isolated T-wave changes in the rightprecordial leads were observed in 2.7% of these men.5
Gottschalk and Craige6 addressed the problem ofwhether there is a difference in ST-T–wave morphol-ogy in healthy blacks compared with whites. Subjects’average age was 20 to 21 years. There were 150patients in each group of white women, white men, blackwomen, and black men. Among white men and black
From the Sarver Heart Center, University of Arizona, Tucson, Arizona.This work was supported in part by Grant UO1-HL65594 from theNational Heart, Lung, and Blood Institute, National Institutes of Health,Bethesda, Maryland. Manuscript received November 1, 2004; re-vised manuscript received and accepted December 27, 2004.
Address for reprints: Frank I. Marcus, MD, Sarver Heart Center,University of Arizona, Department of Medicine, Section of Cardiology,P.O. Box 245037, Tucson, Arizona 85724-5037. E-mail: [email protected].
men, none had inverted T waves in leads V2 and V3.Two white women (1.3%) had T waves inverted inlead V2 and 4% of black women had T waves invertedin leads V2 and V3. Another study examined T-wavechanges in young black men (n � 277) who were anaverage age of 26 years (range 20 to 40) and 43 blackwomen who were an average of 24 years (range 20 to38).7 Of the 320 subjects, only 3 (0.9%) had invertedT waves over the precordium as far to the right as leadV3. Two of the 3 subjects were women. Thus, theredoes not seem to be an increase in the frequency ofT-wave inversion in the black population.
It has been a clinical impression that ST-T–waveabnormalities are more common in trained athletesthan in a sedentary population. The prevalence of thetypical persistent juvenile precordial pattern of T-wave inversion is difficult to assess from the literature,because many of these electrocardiograms showed aninitial ST elevation with a rounded upslope that is notcharacteristic of the juvenile pattern. With this caveatin mind, T-wave inversion in leads V1 to V3 (or V4)that was associated with incomplete right bundlebranch block was found in 16 of 1,005 trained athletes(1.6%) who were �15 years of age.8 Sharma et al9evaluated electrocardiographic changes in 1,000 jun-ior (�18 years of age) elite athletes and 300 nonath-letic controls matched for gender, age, and body sur-face area. The frequency of T-wave inversions inleads V2 and V3 was 4% in athletes and nonathleteswho were �16 years old, but no subject whose agewas �16 years had similar T-wave changes.
The T-wave morphology recorded in normal adultson the surface of the chest does not correspond to thatfound on the epicardial surface just underneath thechest electrodes.10 The observation that the deep in-spiration frequently causes an upright change in thepolarity of the negative T wave suggests that thesewaves are due to the relation of the heart to the chestwall, possibly due to a change in the interposition ofair between the heart and the chest wall.
It has been a consistent observation that the in-
verted precordial T waves becomeupright with exercise in the normalpatient.11,12 A systematic study thatcompared changes in T-wave mor-phology with exercise has not beendone in ARVC/D; however, in somepatients who have ARVC/D, T-waveinversion in leads V1 to V3 or V4persists with exercise (personal ob-servation).
Thus, the juvenile pattern of aninverted T wave in leads V1 to V3appears to be present in 1% to 3% ofa healthy population that is 19 to 45
years of age. There does not seem to be a greaterprevalence of persistent juvenile pattern in blacks or intrained athletes. The finding of T-wave inversion be-yond lead V1 in a patient in this age group who hasventricular arrhythmias of left bundle branch blockmorphology raises the likelihood that this diagnosismay be present. This is particularly applicable for menof this age, because ARVC/D is expressed more fre-quently in men and there is a trend for a lower prev-alence of this electrocardiographic pattern in men.
1. Lepeschkin E. The normal electrocardiogram in childhood. Arch Kreislaufforsch1938;3:321–339.2. McKenna WJ, Thiene G, Nava A, Fontaliran F, Blomstrom-Lundquist C,Fontaine G, Camerini F. Diagnosis of arrhythmogenic right ventricular dysplasia/cardiomyopathy: task force of the Working Group Myocardial and PericardialDisease of the European Society of Cardiology and of the Scientific Council onCardiomyopathies of the International Society and Federation of Cardiology. BrHeart J 1994;71:215–218.3. Nasir K, Bomma C, Tandri H, Roquin A, Dalel D, Prakasa K, Tichnell C,James C, Jspevak P, Marcus F, Calkins H. Electrocardiographic features ofarrhythmogenic right ventricular dysplasia/cardiomyopathy according to diseaseseverity: a need to broaden diagnostic criteria. Circulation 2004;110:1527–1534.4. Suarez RM, Suarez RM Jr. The T wave of the precordial electrocardiogram atdifferent age levels. Am Heart J 1946;32:480–493.5. Hiss RG, Averill KH, Lamb LE. Electrocardiographic findings in 67,375asymptomatic subjects VIII non-specific T wave changes. Am J Cardiol 1960;6:178–189.6. Gottschalk CW, Craige E. A comparison of the precordial S-T and T waves inthe electrocardiograms of 600 healthy young negro and white adults. South Med J1956;49:453–457.7. Thomas J, Harris E, Lassiter G. Observations on the T wave and S-T segmentchanges in the precordial electrocardiogram of 320 young Negro adults. Am JCardiol 1960;5:468–472.8. Pelliccia A, Maron BJ, Culasso F, DiPaolo M, Spataro A, Biffi A, Caselli G,Piovano P. Clinical significance of abnormal electrocardiographic patterns intrained athletes. Circulation 2000;102:278–284.9. Sharma S, Whyte G, Elliot P, Padula M, Kaushal R, Mahon N, McKenna WJ.Electrocardiographic changes in 1000 highly trained junior elite athletes. Br JSports Med 1999;33:319–324.10. Blackman NS, Kuskin L. Inverted T waves in the precordial electrocardio-gram of normal adolescents. Am Heart J 1964;67:304–312.11. Zeppilli P, Pirrami MM, Sassara M, Fenici R. T-wave abnormalities intop-ranking athletes: effects of isoproterenol, atropine, and physical exercise. AmHeart J 1980;100:213–221.12. Serra-Grima R, Estorch M, Carrio I, Subirana M, Berna L, Prat T. Markedventricular repolarization abnormalities in highly trained athletes’ electrocar-diograms: clinical and prognostic implications. J Am Coll Cardiol 2000;36:1310 –1336.
TABLE 1 Percent T-wave Negativity in Leads V1 to V6 in 161 Healthy PuertoRicans at Various Ages*
Comparison of All-Cause Mortality in Women WithKnown or Suspected Coronary Artery Disease Referredfor Dobutamine Stress Echocardiography With Normal
Versus Abnormal Test Results
Elena Biagini, MD, Abdou Elhendy, MD, Arend F.L. Schinkel, MD, Vittoria Rizzello, MD,Ron T. van Domburg, PhD, Boudewijn J. Krenning, MD, Olaf Schouten, MD,
Fabiola B. Sozzi, MD, Angelo Branzi, MD, Guido Rocchi, MD, Maarten L. Simoons, MD,Jeroen J. Bax, MD, and Don Poldermans, MD
The presence of myocardial ischemia during dobut-amine stress echocardiography is independently as-sociated with an increased risk of all-cause mortalityin women after adjustment for clinical data. This as-sociation is observed in patients who have provedcoronary artery disease and in patients who have nohistory of coronary artery disease. �2005 by ExcerptaMedica Inc.
(Am J Cardiol 2005;95:1072–1075)
This study evaluated predictors of all-cause mortal-ity in women who had known or suspected coro-
nary artery disease (CAD) and were referred for do-butamine stress echocardiography (DSE).
• • •The study population consisted of 1,172 consecu-
tive women who had limited exercise capacity andwere referred for DSE between January 1990 andJanuary 2003 at the Thoraxcenter (Rotterdam, TheNetherlands). Follow-up was successful in 1,168 pa-tients (99%). Sixty-three patients (5%) underwentearly coronary revascularization �60 days after DSE.These patients were excluded from analysis becausereferral to myocardial revascularization �60 days af-ter stress testing tends to be based on results of thetest, whereas referral to revascularization at �60 daysafter testing tends to be based on worsening of thepatient’s clinical status.1 The final population con-sisted of 1,105 patients. The protocol was approved bythe hospital ethics committee. All patients gave in-formed consent before the test.
A structured interview and clinical history weretaken and cardiac risk factors were assessed beforeDSE. Cardiac risk factors included a history of currentor stable angina pectoris, previous myocardial infarc-tion, congestive heart failure, hypertension, diabetesmellitus, and hypercholesterolemia. Congestive heartfailure was defined according to Framingham criteria.2
Hypertension was defined as a blood pressure �140/90mm Hg or use of antihypertensive medication. Diabetesmellitus was defined as a fasting glucose level �7.8mmol/L or the need for insulin or oral hypoglycemicagents. Hypercholesterolemia was defined as a totalcholesterol level �6.4 mmol/L or use of a lipid-lower-ing medication. History of CAD was defined as previousmyocardial infarction, angiographically documented cor-onary artery stenosis, and/or myocardial revasculari-zation.
Dobutamine/atropine stress testing was performedaccording to a standard protocol as previously report-ed.3 After obtaining a baseline echocardiogram, do-butamine was administered intravenously, starting at adose of 10 �g/kg/min for 3 minutes (5 �g/kg/min inpatients who had left ventricular dysfunction at rest).Incremental dobutamine doses of 10 �g/kg/min weregiven at 3-minute intervals to a maximum dose of 40�g/kg/min. If the test end point was not reached at adobutamine dose of 40 �g/kg/min, atropine (�2 mg)was given intravenously. Blood pressure, heart rate,and electrocardiography were constantly monitored.Test end points were achievement of a target heart rate(85% of maximum age- and gender-predicted heartrates), maximal doses of dobutamine and atropine,extensive new wall motion abnormalities, �2 mVdownsloping ST-segment depression measured 80 msafter the J point compared with baseline, hypertension(blood pressure �240/120 mm Hg), a decrease insystolic blood pressure of �40 mm Hg compared withat rest, significant arrhythmias, or any intolerable ad-verse effect considered to be the result of dobutamineor atropine. An intravenous � blocker (metoprolol 1 to5 mg) was available to reverse the adverse effects ofdobutamine/atropine.
Two-dimensional echocardiographic images wereacquired at rest, during dobutamine stress, and duringrecovery using the standard views. Echocardiogramswere recorded in a quad-screen format. Two experi-enced observers who were unaware of the clinical datascored the echocardiograms using a standard 16-seg-ment model. In case of disagreement, a consensusdecision was achieved by a third observer. Regionalwall motion and systolic wall thickening were scoredon a 5-point scale. Ischemia was defined as new orworsened wall motion abnormalities during stress thatwas indicated by a wall motion score increase of �1
From the Department of Cardiology, Thoraxcenter, Erasmus MC,Rotterdam, The Netherlands; the Department of Cardiology, LeidenUniversity Medical Center, Leiden, The Netherlands; and the Instituteof Cardiology, S. Orsola Hospital, Bologna, Italy. Dr. Poldermans’saddress is: Department of Cardiology, Thoraxcenter Room Ba 300,Erasmus MC, Dr. Molewaterplein 40, 3015 GD Rotterdam, TheNetherlands. E-mail: [email protected]. Manuscript re-ceived October 25, 2004; revised manuscript received and ac-cepted December 16, 2004.
grade in �1 segment. Ischemia was not consideredpresent when akinetic segments at rest became dyski-netic during stress. For each patient, a wall motionscore index was calculated by dividing the sum ofsegment scores by the total number of interpretedsegments.
Follow-up was obtained by review of the hospitalrecords and by contacting general practitioners. In
addition, vital status was verified through civil regis-tries. If no further information from the general prac-titioner could be obtained, patients were approachedafter retrieval of vital status. The primary follow-upevent was death from any cause. Myocardial revascu-larization and nonfatal myocardial infarction werealso noted.
Continuous data were expressed as mean � SD.Student‘s t test was used to analyze continuous data.Differences between proportions were compared withchi-square test. Univariate and multivariate Cox’s pro-portional hazard regression models (BMDP StatisticalSoftware Inc., Los Angeles, California) were used toidentify independent predictors of death.4 Variableswere selected in a stepwise forward selection manner,with entry and retention set at a significance level of0.05. The risk of a variable was expressed as a hazard
ratio with a corresponding 95% con-fidence interval. The probability ofsurvival was calculated with theKaplan-Meier method, and survivalcurves were compared with the log-rank test. A p value �0.05 was con-sidered statistically significant.
Mean age was 61 � 12 years.Clinical and hemodynamic data arepresented in Tables 1 and 2, respec-tively. There were 346 patients(31%) who had known CAD and 759patients (69%) who had no history ofCAD. Heart rate and systolic bloodpressure increased significantly fromrest to peak stress (p �0.001). Ar-rhythmias during the test were non-sustained ventricular tachycardia in25 patients (2%) and atrial fibrilla-tion in 11 patients (1%). Hypoten-sion occurred in 9 patients (1%). Thetest was terminated for achievementof the target heart rate in 1,004 pa-tients (91%), maximal dobutamine/atropine dose in 29 (3%), chills ornausea in 31 (3%), arrhythmias in 18(1%), abnormal blood pressure in 12(1%), and ST-segment changes in 11patients (1%).
Dobutamine stress echocardio-grams were normal in 483 patients(44%). Fixed wall motion abnormal-ities were detected in 206 patients(19%) and ischemia (new or worsen-ing wall motion abnormalities) wasdetected in 416 patients (38%).Among patients who had ischemia,
364 (33%) also had wall motion abnormalities at rest.During a mean follow-up of 5.1 � 3.1 years (range
2 months to 9.5 years), there were 216 deaths (19%).Thirty-two patients (3%) developed nonfatal myocar-dial infarction. Late myocardial revascularization wasperformed in 178 patients (16%).
Univariable and multivariable predictors of deathin the entire population are presented in Table 3.
Independent clinical predictors were age, history ofheart failure, smoking, and hypertension. Multivariateanalysis demonstrated that DSE provided incrementalprognostic information for the prediction of eventsover clinical and stress test data. Independent echo-cardiographic predictors were wall motion score indexat rest and the presence of ischemia during DSE.Among the 346 patients who had previous CAD,independent predictors of death were age, history ofheart failure, smoking, hypertension, and the presenceof ischemia (Table 4). In patients who had no historyof CAD, independent predictors of death were age,history of heart failure, smoking, diabetes mellitus,and the presence of ischemia (Table 4). Kaplan-Meiersurvival curves based on the results of DSE are pre-sented in Figure 1. The annual death rates were 2.7%in women who had a normal dobutamine stress echo-cardiogram, 3.4% in patients who had fixed wall mo-
tion abnormalities, and 4.4% in pa-tients who had ischemia. The annualevent rate (cardiac death or nonfatalmyocardial infarction) for womenwho had a normal dobutamine stressechocardiogram was 1.3% (Figure2).
• • •This study assessed the predictors
of all-cause mortality in 1,105women who had known or suspectedCAD and underwent DSE. Duringfollow-up, 216 women (19%) died.Annual mortality rates were 2.7% inwomen who had a normal dobut-amine stress echocardiogram and4.4% in women who had new wallmotion abnormalities (p �0.001).The annual event rate (cardiac deathor nonfatal myocardial infarction)for women who had a normal dobut-amine stress echocardiogram was1.3%. Predictors of mortality wereage, history of heart failure, smok-ing, hypertension, wall motionscore index at rest, and presence ofischemia during DSE. Multivariateanalysis demonstrated that DSEprovided incremental prognosticinformation for the prediction ofevents over clinical and stress testdata.
Information on the utility of DSEto predict all-cause mortality inwomen is scarce. Cortigiani et al5evaluated 456 women (mean age 63� 10 years) who underwent pharma-cologic stress echocardiography withdipyridamole (305 patients) or do-butamine (151 patients) for evalua-tion of chest pain. During a meanfollow-up of 32 � 19 months, 3deaths (0.6%) and 10 myocardial in-farctions (2%) occurred. The pres-
ence of ischemia during the stress test was the onlyindependent predictor of cardiac death and myocardialinfarction. The death rate was low in that study andanalysis was based on a composite end point. Previousstudies have demonstrated an incremental prognosticvalue of DSE for prediction of cardiac death and othercomposite end points of cardiac events.6–8 However, afew studies used all-cause mortality as the primaryend point. Marwick et al6 studied 3,156 patients (meanage 63 � 12 years; 1,355 women) who underwentDSE. After a mean follow-up of 3.8 � 1.9 years, deathoccurred in 716 patients (23%). Ischemia was an in-dependent predictor of total mortality rate. However,the study did not analyze the incremental prognosticvalue of DSE in women separately. Studies of prog-nostic utility of stress myocardial perfusion imaging inwomen have focused mainly on composite end pointsor cardiac death.9–11 A recent study12 demonstrated an
FIGURE 1. Kaplan-Meier survival curves (all-cause mortality) in women according todobutamine stress echocardiographic results.
TABLE 4 Univariate and Multivariate Predictors of Mortality in Women With andWithout a History of Coronary ArteryDisease (CAD)
Univariate Multivariate
HR 95% CI HR 95% CI
History of CADAge (yrs) 1.01 0.99–1.03 1.1 1.03–1.17Previous myocardial infarction 1.9 1.1–3.3Previous heart failure 2.1 1.3–3.2 3.2 1.4–7.5Smoking 1.8 1.1–3.5 2.3 1.1–5.5Systemic hypertension 1.6 1.0–3.1 2.3 1.1–5.1Wall motion score index at rest 1.7 1.1–2.8Myocardial ischemia 2.2 1.1–3.9 2.1 1.1–4.5
No history of CADAge (yrs) 1.04 1.02–1.07 1.1 1.05–1.16Previous heart failure 2.2 1.3–3.6 2.2 1.1–5.2Smoking 1.9 1.2–4.1 2.4 1.2–5.0Diabetes mellitus 2.1 1.1–4.9 2.4 1.2–5.0Myocardial ischemia 2.0 1.1–4.0 2.2 1.2–4.3
Abbreviations as in Table 3.
1074 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
incremental value of stress technetium-99m tetrofos-min imaging in predicting all-cause mortality in 503women (mean age 58 � 19 years) who had known orsuspected CAD and were followed for 3.5 years.
In the present study, a history of heart failure,smoking, hypertension, and diabetes were indepen-dent clinical predictors of mortality. Previous epide-miologic studies have shown that risk factors, such ashypertension, current smoking, and diabetes consideredindividually or in association are positively associatedwith all-cause mortality in women.13–15 Moreover, ahistory of diabetes, particularly if associated with knownCAD and heart failure, has been shown to identify aparticularly high-risk group for cardiac death and all-cause mortality in women.15 The effect of left ventric-ular function and ischemia was not evaluated in thesestudies.
1. Hachamovitch R, Berman DS, Kiat H, Cohen I, Cabico JA, Friedman J,Diamond GA. Exercise myocardial perfusion SPECT in patients without knowncoronary artery disease: incremental prognostic value and use in risk stratifica-tion. Circulation 1996;93:905–914.2. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history ofcongestive heart failure: the Framingham study. N Engl J Med 1971;285:1441–1446.
3. Poldermans D, Fioretti PM, Boersma E, Bax JJ,Thomson IR, Roelandt JR, Simoons ML. Long-termprognostic value of dobutamine-atropine stress echo-cardiography in 1737 patients with known or suspectedcoronary artery disease: a single-center experience. Cir-culation 1999;99:757–762.4. Cox DR. Regression models and life-tables. J R StatSoc B 1972;34:187–202.5. Cortigiani L, Dodi C, Paolini EA, Bernardi D, BrunoG, Nannini E. Prognostic value of pharmacologicalstress echocardiography in women with chest pain andunknown coronary artery disease. J Am Coll Cardiol1998;32:1975–1981.6. Marwick TH, Case C, Sawada S, Rimmerman C,Brenneman P, Kovacs R, Short L, Lauer M. Predictionof mortality using dobutamine echocardiography. J AmColl Cardiol 2001;37:754–760.7. Pingitore A, Picano E, Varga A, Gigli G, Cortigiani L,Previtali M, Minardi G, Colosso MQ, Lowenstein J,Mathias W Jr, Landi P. Prognostic value of pharmacolog-ical stress echocardiography in patients with known orsuspected coronary artery disease: a prospective, large-scale, multicenter, head-to-head comparison between di-pyridamole and dobutamine test. Echo-Persantine Inter-national Cooperative (EPIC) and Echo-Dobutamine Inter-national Cooperative (EDIC) Study Groups. J Am CollCardiol 1999;34:1769–1777.8. Sicari R, Picano E, Landi P, Pingitore A, Bigi R,Coletta C, Heyman J, Casazza F, Previtali M, Mathias WJr, et al. Prognostic value of dobutamine-atropine stressechocardiography early after acute myocardial infarction.
Echo Dobutamine International Cooperative (EDIC) Study. J Am Coll Cardiol1997;29:254–260.9. Hachamovitch R, Berman DS, Kiat H, Bairey CN, Cohen I, Cabico A,Friedman J, Germano G, Van Train KF, Diamond GA. Effective risk stratificationusing exercise myocardial perfusion SPECT in women: gender-related differ-ences in prognostic nuclear testing. J Am Coll Cardiol 1996;28:34–44.10. Amanullah AM, Berman DS, Erel J, Kiat H, Cohen I, Germano G,Friedman JD, Hachamovitch R. Incremental prognostic value of adenosinemyocardial perfusion single-photon emission computed tomography inwomen with suspected coronary artery disease. Am J Cardiol 1998;82:725–730.11. Pancholy SB, Fattah AA, Kamal AM, Ghods M, Heo J, Iskandrian AS.Independent and incremental prognostic value of exercise thallium single-photon emission computed tomographic imaging in women. J Nucl Cardiol1995;2:110 –116.12. Elhendy A, Schinkel AF, van Domburg RT, Bax JJ, Valkema R, PoldermansD. Prediction of all-cause mortality in women with known or suspected coronaryartery disease by stress technetium-99m tetrofosmin myocardial perfusion imag-ing. Am J Cardiol 2004;93:450–452.13. Houterman S, Verschuren WM, Kromhout D. Smoking, blood pressureand serum cholesterol-effects on 20-year mortality. Epidemiology 2003;14:24 –29.14. Stamler J, Stamler R, Neaton JD, Wentworth D, Daviglus ML, Garside D,Dyer AR, Liu K, Greenland P. Low risk-factor profile and long-term cardio-vascular and noncardiovascular mortality and life expectancy: findings for 5large cohorts of young adult and middle-aged men and women. JAMA1999;282:2012–2018.15. Hu FB, Stampfer MJ, Solomon CG, Liu S, Willett WC, Speizer FE, NathanDM, Manson JE. The impact of diabetes mellitus on mortality from all causes andcoronary heart disease in women: 20 years of follow-up. Arch Intern Med2001;161:1717–1723.
FIGURE 2. Kaplan-Meier survival curves (cardiac death or nonfatal myocardial infarc-tion) in women according to dobutamine stress echocardiographic results.
BRIEF REPORTS 1075
Ability for Visualization, Reasons for NonassessableImage, and Diagnostic Accuracy of 16-Slice
Multidetector Row Helical ComputedTomography for the Assessment of
Takayuki Suzuki, MD, Kouichi Fujikawa, MD, and Hiroshi Yamaguchi, RT
It was demonstrated that 16-slice multidetector rowhelical computed tomography may permit more ac-curate assessment of entire coronary arteries, with alow rate of arteries unable to be evaluated. The de-tection of coronary artery stenoses with this modalitywas applicable to assessable segments. �2005 byExcerpta Medica Inc.
(Am J Cardiol 2005;95:1076–1079)
Recently, a new generation of multidetector rowhelical computed tomography (MDCT) scanners,
equipped with more and thinner detector rows andincreased rotation speed, has been introduced.1 In thisstudy, we investigated the ability for the visualizationof entire coronary arteries by 16-slice MDCT and thereasons for nonassessable images. In addition, weexamined the accuracy of 16-slice MDCT in detectingcoronary artery stenoses in comparison with conven-tional coronary angiography (CAG).
• • •Sixteen-slice MDCT was performed in 70 consec-
utive patients (49 men; aged 69 � 11 years) withsuspected coronary disease. The patients’ characteris-tics are listed in Table 1. Exclusion criteria included(1) irregular heart rate (chronic atrial fibrillation, fre-quent premature beats, and so forth), (2) renal dys-function (serum creatinine �1.5 mg/dl), (3) allergy tocontrast media, (4) severe deafness (breath hold im-possible), and (5) unstable hemodynamics. The studywas approved by the hospital’s ethics committee, andinformed consent was obtained from all patients.
Helical computed tomography angiography using16-slice MDCT (LightSpeed Ultra 16, GE Healthcare,Waukesha, Wisconsin; gantry rotation time 0.5 seconds,16 � 0.625 mm detector collimation) was performedusing the retrospective electrocardiographic gatedmethod. Of 70 patients, 65 (heart rate before theMDCT scan �60 beats/min, 93%) received metopro-
lol 40 mg (Seroken, AstraZeneca PLC, London,United Kingdom) orally 60 minutes before MDCTscanning, and all received nitroglycerin 0.3 mg(Myocolspray, Toa Eiyo Ltd., Tokyo, Japan) sublin-gually 5 minutes before MDCT scanning. First, a plainlocalization scan that yielded an anteroposterior viewof the chest was performed to decide the position(z-axis coverage) of the imaging volume for coronaryartery (and bypass graft) imaging. Next, after imagingat the level of the left ventricle and positioning theregion of interest in the cavity of the left ventricle, abolus of contrast medium 100 ml (iohexol 300 or 370mgI/ml, Schering AG, Berlin, Germany) was injectedinto the right cubital vein at 2.5 to 4.0 ml/s, and abolus of saline 60 ml was subsequently injected intothe right cubital vein at 3.0 ml/s (using Dual Shot,Nemoto Kyorindou Co. Ltd., Tokyo, Japan). As soon
From the Departments of Cardiology and Diagnostic Imaging,Hiroshima General Hospital, Hatsukaichi, Japan; and Department ofMolecular and Internal Medicine, Graduate School of BiomedicalSciences, Hiroshima University, Hiroshima, Japan. Dr. Kitagawa’saddress is: Department of Molecular and Internal Medicine, GraduateSchool of Biomedical Sciences, Hiroshima University, 1-2-3 KasumiMinami-Ku, Hiroshima 734-8551, Japan. E-mail: [email protected]. Manuscript received October 6, 2004; revised manu-script received and accepted December 28, 2004.
TABLE 1 Patients Characteristics (n � 70)
Characteristic Value
Age (yrs) 69 � 11 (range 44–84)Men/women 49/21Hypertension 33 (47%)Hyperlipidemia 21 (30%)Diabetes mellitus 29 (41%)Smoker 21 (30%)Body mass index �25 kg/m2 24 (34%)Previous percutaneous coronary
intervention (after stenting)10 (14%)9 (13%)
Previous coronary artery bypassgrafting
26 (37%)
Data are presented as mean � SD or number (%) of patients.
TABLE 2 Phase With the Fewest Motion Artifacts in the Leftand Right Coronary Arteries (n � 70)
as the monitored signal density of the region of inter-est reached a predefined threshold of 100 HU, the scanwas started (Smart Prep, GE Healthcare). Image ac-quisition was performed during an inspiratory breathhold without the inhalation of oxygen. The volumedata sets were acquired in spiral mode, with the si-multaneous acquisition of 16 parallel slices (slicethickness 0.625 mm, gantry rotation time 0.5 to 0.6seconds, helical pitch 0.275:1 to 0.325:1, table feed2.75 to 3.25 mm/rotation, tube current 350 to 400 mAs,tube voltage 120 to 140 kV). The patient’s electrocar-diogram was digitized and continuously monitoredduring scanning. The raw data of the scans weretransferred to a computer workstation (AdvantageWorkstation, version 4.05, GE Healthcare), and theimage reconstruction was performed using image anal-ysis software (CardIQ I or II, GE Healthcare).
According to the heart rate during scanning, a halfreconstruction algorithm (heart rate �65 beats/min, 1sector, temporal window 250 ms) or a multisectorreconstruction algorithm (heart rate �65 beats/min, 2
to 4 sectors, temporal window �125 ms) was selected.Initially, volume rendering images of the 20 phaseswere created at an interval of 5% from the beginning(0%) to the end (95%) of the cardiac cycle (R- toR-wave interval), with a slice thickness of 1.25 mm, andthe phase with the fewest motion artifacts was chosen.
Next, the data sets of the phase with the fewestmotion artifacts were reconstructed by volume render-ing and curved multiplanar reconstruction methods,with a slice thickness of 0.625 mm. In each patient,the phase with the fewest motion artifacts was used forfurther evaluation for the left and right coronary ar-teries. On the basis of the axial images, volume ren-dering images, and curved multiplanar reconstruction im-ages, the image quality of all coronary arteries wasjudged by consensus of 3 observers (a radiologic tech-nologist and 2 cardiologists) as sufficient or insufficientfor morphologic evaluation. In all patients, each of the 14coronary segments (American Heart Association classi-fication) visualized with the axial images, volume ren-dering images, and curved multiplanar reconstructionimages was categorized into an assessable group (goodand fair) or a nonassessable group, and the latter wasdivided into subgroups according to the reason for non-assessable images: (1) small branch, (2) banding artifact,(3) calcification, and (4) nondetectable positions contain-ing postocclusion segments, stented segments, and non-detectable segments because of mixed factors (smallbranch, motion artifact, and small signal-to-noise ratio).Furthermore, in 20 of 70 patients in whom CAG wasperformed �4 weeks after MDCT, stenoses in the 14coronary segments visualized with MDCT images andcategorized into the assessable group were assessed bythe preceding 3 observers. Stenoses �50% (50%, 75%,90%, 99%, and 100%, American Heart Association clas-sification) were considered to be significant. When theassessment varied among the observers, the agreement of2 observers was used. CAG was evaluated by a blindedreviewer using quantitative CAG (Integris HM3000,Philips Medical Systems, Best, The Netherlands) and theusual gold standard for stenosis detection. The severityof stenoses in the 14 coronary segments was graded
FIGURE 1. Coronary artery visualization by 16-slice MDCT ineach segment of the left coronary artery (American Heart Associ-ation classification).
FIGURE 2. Coronary artery visualization by 16-slice MDCT ineach segment of the right coronary artery (American Heart Asso-ciation classification), using the same categories as in Figure 1.PD � posterior descending artery, AV � atrioventricular nodeartery.
BRIEF REPORTS 1077
using the same categories for the MDCT analysis. Theseresults were compared with those of MDCT.
Mean heart rate before scanning was 58.4 � 10.2beats/min (range 45 to 95). Mean heart rate after
scanning was 64.1 � 12.2 beats/min(range 42 to 122) during breathhold.The mean scan time (the mean dura-tion of breath hold) was 23.8 � 5.8seconds (range 20 to 40). The z-axiscoverage and helical pitch (tablefeed) determined the scan time andvaried among the patients. In all pa-tients, there were no obvious compli-cations of the MDCT scanning. In 30of 70 patients, the half reconstructionalgorithm was used; in the remaining40 patients, the multisector recon-struction algorithm was used. Thephase with the fewest motion arti-facts is listed in Table 2.
The result of each coronary segment visualizationis shown in Figure 1 (the left coronary artery) andFigure 2 (the right coronary artery). The assessabilityof the proximal portion of coronary arteries (segments5, 7, 11, 1, and 2), excluding segment 6, was good(�85%), and that of segment 6 was less (79%) thanthat of the other proximal segments because of calci-fication. With respect to the distal portion of coronaryarteries, the assessability of segments 8 and 3 wasvery good (95% and 95%), that of segment 4 (poste-rior descending artery and atrioventricular node ar-tery) was relatively good (83% and 83%), and that ofsegments 9 and 13 was moderate (79% and 77%). Theassessability of segments 12 and 14 was low (68% and66%). Nonassessable images due to calcification weremore frequent in segments 5 to 7 (the proximal portionof the left coronary artery) than others; therefore, theassessability of segment 8 (the distal portion of the leftcoronary artery) with less calcification was better thanthat of segments 6 and 7. The assessability of the rightcoronary artery was good, but nonassessable imagesdue to banding artifacts were more frequent in seg-ment 2 than in others. A summary of coronary arteryvisualization is listed in Table 3.
In CAG, 17 of the 20 patients had �1 significantcoronary stenosis (total number of stenoses 34). Incomparison with CAG, MDCT correctly classified in14 of 17 patients (82.4%) as having �1 coronarystenosis. After excluding 46 of 280 (20 patients � 14segments) coronary segments classified as nonassess-
FIGURE 3. MDCT (volume rendering [A], curved multiplanar re-construction [B]) and CAG (C) images in a 71-year old womanwith stable angina pectoris. Significant stenosis (90%) of segment6 was detected in MDCT and CAG images (arrow). In MDCT im-ages, calcification in segment 6 was also detected (arrow head).LAD � left anterior descending coronary artery, LCX � left cir-cumflex artery.
TABLE 3 Summary of Coronary Artery Visualization by 16-slice MDCT
Total of 980 Coronary Segments (70 Patients � 14 Segments)
Data are the number of segments.*Segments with complete agreement between the MDCT and CAG findings.
1078 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
able by MDCT (16.4%), 24 of 28 significant stenoticsegments were detected, and the absence of stenosiswas correctly identified in 202 of 206 segments (sen-sitivity 85.7%, specificity 98.1%, accuracy 96.6%,positive predictive value 85.7%, negative predictivevalues 98.1%; Table 4). Figure 3 shows the case withagreement between the MDCT and CAG findings.
• • •Our study demonstrated that 16-slice MDCT per-
mitted the assessment of entire coronary arteries, witha low rate of coronary segments that could not be eval-uated (16.2%) in comparison with 4-slice MDCT.2–4
Our study also suggests that the clinical utility ofMDCT may currently be limited for different reasonsin each coronary segment. The reasons for nonassess-able images were small branch (12.6%, 20 of 159segments), banding artifact (9.4%, 15 of 159 seg-ments), calcification (23.3%, 37 of 159 segments), andnondetectable positions (54.7%, 87 of 159 segments),that is, 3 with postocclusion (1.9%), 8 with stentartifact (5.0%), and 76 with nondetectable segmentsdue to mixed factors (small branch, motion artifact,and small signal-to-noise ratio; 47.8%, 76 of 159segments).
It is believed that the motion artifacts due torespiration and heart beat that were the main reasonfor nonassessability of cardiac computed tomogra-phy were decreased greatly with the use of 16-sliceMDCT, but banding artifact due to myocardial mo-tion was more likely to limit the proper assessmentof segment 2. The motion artifacts were likely to beless in end-diastole, which was the at rest momentof the cardiac cycle, but in some patients, the bestreconstructed images were created with the datasets of the systole and other stages. Therefore, it isbelieved that the search of the phase with the fewestmotion artifacts in each patient was necessary forthe best reconstruction. In the previous studies ofcardiac computed tomography,2– 8 the assessabilityof the proximal portion of coronary arteries was
better than that of the distal portion. In our study,the motion artifacts were less influential on theassessability of the proximal portion of coronaryarteries, excluding segment 2, whereas calcificationin the left main coronary artery and the proximalportion of the left anterior descending artery wasmost likely to be severe, and it decreased the as-sessability of segments 5 to 7. It was consideredthat the reason for less assessability of segments 12and 14 (the periphery of the left circumflex artery)than of segment 4 (posterior descending artery andatrioventricular node artery) (the periphery of theright coronary artery) was the effect of the activemyocardial motion and the overlapping of the cor-onary sinus.
1. Ropers D, Baum U, Pohle K, Anders K, Ulzheimer S, Ohnesorge B, SchlundtC, Bautz W, Daniel WG, Achenbach S. Detection of coronary artery stenoseswith thin-slice multi-detecter row spiral computed tomography and multiplanarreconstruction. Circulation 2003;107:664–666.2. Nieman K, Oudkerk M, Rensing BJ, van Ooijen P, Munne A, van Geuns RJ,de Feyter PJ. Coronary angiography with multislice computed tomography.Lancet 2001;357:599–603.3. Achenbach S, Giesler T, Ropers D, Ulzheimer S, Derlien H, Schlte C, WenkelE, Moshage W, Bautz W, Daniel WG, et al. Detection of coronary artery stenosisby contrast-enhanced, retrospectively electrocardiographically-gated, multislicespiral computed tomography. Circulation 2001;103:2535–2538.4. Nieman K, Rensing BJ, van Geuns RJM, Munne A, Ligthart JMR, PattynamaPMT, Krestin GP, Serruys PW, de Feyter PJ. Usefulness of multislice computedtomography for detecting obstructive coronary artery disease. Am J Cardiol2002;89:913–918.5. Kachelriess M, Kalender WA. Electrocardiogram-correlated image reconstruc-tion from subsecond spiral computed tomography scans of the heart. Med Phys1998;25:2417–2431.6. Ohnesorge B, Flohr T, Becker C, Kopp AF, Schoepf UJ, Baum U, Knez A,Klingenbeck-Regn K, Reiser MF. Cardiac imaging by means of electrocardio-graphically gated multisection spiral CT: initial experience. Radiology 2000;217:564–571.7. Knez A, Becker CR, Leber A, Ohnesorge B, Becker A, White C, Haberl R,Reiser MF, Steinbeck G. Usefulness of multislice spiral computed tomographyangiography for determination of coronary artery stenosis. Am J Cardiol 2001;88:1191–1194.8. Otsuka M, Hirohashi S, Uemura S, Watanabe M, Ishigami K, Maekura T, ItoA, Haze K, Saito Y, Kichikawa K. Assessment of coronary artery by four-detector multislice computed tomography: diagnostic accuracy and limitations forcoronary artery lesions. J Cardiol 2003;41:225–234.
BRIEF REPORTS 1079
Effectiveness of High-Intensity Interval Training forthe Rehabilitation of Patients With Coronary
Artery Disease
Darren E.R. Warburton, PhD, Donald C. McKenzie, MD, PhD, Mark J. Haykowsky, PhD,Arlana Taylor, PT, Paula Shoemaker, MSc, Andrew P. Ignaszewski, MD,
and Sammy Y. Chan, MD
We found that interval training provides an effectivemeans to improve the cardiovascular fitness andhealth status of highly functional patients with coro-nary artery disease. We also revealed that intervaltraining improves anaerobic tolerance to a greaterextent than the traditional exercise training modelwithout increasing the risk to the patient. This re-search supports the implementation of interval train-ing for highly functional patients with coronary arterydisease. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1080–1084)
Cardiac (exercise) rehabilitation is a valuable non-pharmacologic intervention to improve cardiore-
spiratory fitness and overall health status in patientswith coronary artery disease (CAD). Furthermore,continuous aerobic exercise training has been effec-tive in reducing all-cause and cardiac mortality rates.1.Despite these benefits, this form of training may besuboptimal for patients with CAD with higher base-line levels of fitness. We have shown that high-inten-sity interval training is an effective modality to im-prove maximal aerobic power in sedentary men.2Preliminary research has also supported the healthbenefits of interval training and/or circuit weight train-ing in patients with cardiac disease (including chronicheart failure and CAD).3–5 However, the interval train-ing research has generally used low work to rest ratiosand/or lower work intensities. Accordingly, the pri-mary purpose of this study was to examine the effectsof high-intensity interval training versus the tradi-tional aerobic program in highly functional (�9METs) patients with CAD. We hypothesized that in-terval training would be safely tolerated by patientsand lead to a similar improvement in peak aerobicpower (VO2peak) as the traditional model. Due to theimportance of anaerobic capacity for the performance
of many activities of daily living,6,7 we were alsointerested in evaluating the effects of interval andcontinuous aerobic training on anaerobic capacity.Given the anaerobic requirements of interval training,we hypothesized that this form of training wouldresult in greater tolerance to an anaerobic challengethan the traditional program.
• • •Fourteen men with CAD (Table 1) who had under-
went (�6 months previously) bypass surgery or an-gioplasty were stratified (age, body mass, andVO2peak) and randomly assigned to traditional (n �7) or interval training (n � 7). Ethical approval andwritten informed consent were obtained. All partici-pants were recruited from a pool of patients who hada negative stress test (i.e., no significant ST depression(�1 mm) during a symptom-limited exercise test) and aVO2peak �9 METs. As such, these participants repre-sented stable, highly functional patients with CAD.
All participants were assessed on 2 separate days atbaseline and after 16 weeks of aerobic training. Ontesting day 1, the patients performed a symptom-limited, incremental to maximal exercise treadmill test(Bruce protocol) to assess VO2peak. Patients weremonitored continuously via 12-lead electrocardiogra-phy, and blood pressure was taken at rest, everyminute during exercise, and for �3 minutes during therecovery. The participants’ rating of perceived exer-tion was also evaluated throughout the test.8 Expiredgas and ventilatory parameters were acquired using aportable metabolic cart (K4, Cosmed, Italy). Anaero-bic threshold and ventilatory efficiency were deter-mined via standard procedures.9 The major criteria forterminating the test were volitional fatigue and a pla-teau in VO2.
On testing day 2, the participants performed ahigh-intensity time to exhaustion test on a treadmill at90% of heart rate reserve. This test provides an ob-jective marker of endurance capacity, with the laterstages of the test being particularly dependent uponanaerobic capacity. Each participant had previouslyengaged in 1 familiarization time to exhaustion test.Heart rate and the electrocardiogram were monitoredcontinuously via 3-lead telemetry. Blood pressure wasalso taken before, throughout (every minute), and for�3 minutes after exercise. The speed and grade of thetreadmill were kept identical between the baseline andpost-training tests.
The traditional cardiac rehabilitation model con-sisted of a standardized 10-minute warm-up, 30 min-
From the School of Human Kinetics and Faculty of Medicine, Universityof British Columbia, Vancouver, British Columbia; Healthy Heart Pro-gram, St. Paul’s Hospital, Vancouver, British Columbia; and Faculty ofRehabilitation Medicine, University of Alberta, Edmonton, Alberta,Canada. This study was supported by the Natural Sciences andEngineering Council of Canada, the Canada Foundation for Innova-tion, the British Columbia Knowledge Development Fund, and theMichael Smith Foundation for Health Research, Canada. Dr. Warbur-ton’s address is: University of British Columbia, Unit II Osborne Centre,Room 205, 6108 Thunderbird Blvd., Vancouver, British Columbia,Canada V6T 1Z3. E-mail: [email protected]. Manu-script received July 22, 2004; revised manuscript received and ac-cepted December 16, 2004.
utes of continuous aerobic exercise at 65% of heartrate/VO2 reserve, standardized resistance training, anda 10-minute cool-down period. Interval training con-sisted of the identical warm-up, resistance training,and cool-down procedures as performed by the tradi-tional training group. However, the interval traininggroup exercised using 2-minute, high-intensity workphases (90% of heart rate/VO2 reserve [range 85% to95%]) followed by 2-minute recovery bouts (40% ofheart rate/VO2 reserve [range 35% to 45%]). Bothgroups were required to train for 30 minutes/day, 2days/week for 16 weeks. Training involved 3 bouts of10 minutes of exercise on 3 different types of exerciseequipment, including a treadmill, a stairclimber, andcombined arm and leg cycle ergometer. These differ-ent exercises were utilized to allow for total bodytraining as per standard cardiac rehabilitation proce-dures. Each group was also instructed to engage in 3additional training days per week consisting of con-tinuous exercise at 65% of heart rate/VO2 reserve(range 60% to 70%) (compliance rates 98.5 � 2.0 vs.98.8 � 2.0%, respectively, for interval and traditionaltraining).The electrophysiologic response to exercisewas evaluated by 3-lead telemetry and a portable heartrate monitor (Polar Vantage XL, Kemple, Finland).Individual workloads were adjusted daily according toa heart rate range to reflect changes in fitness.2 Theaverage training volume was similar between groups.
All dependent measures were reported as mean �SD. All variables were analyzed using repeated mea-sures analysis of variance (with Tukey’s post hoccomparisons) with the � level set a priori at p �0.05.
At baseline, the groups were wellmatched according to body mass,VO2peak, and age (Table 1 and Fig-ures 1 to 3). Sixteen weeks of cardiacrehabilitation (both traditional andinterval training) did not result insignificant changes in resting mea-sures of heart rate, systolic bloodpressure, diastolic blood pressure,pulse pressure, and rate–pressureproduct, and maximal exercise mea-sures of heart rate, systolic bloodpressure, and rate–pressure product(Table 2). There were significant im-provements in resting and maximalexercise oxygen pulse, VO2peak,Bruce treadmill time, and time to ex-haustion (Table 2 and Figures 1 to 3)after both training programs. Therewas also a concomitant reduction inthe rating of perceived exertion andpulse pressure during maximal exer-cise in both training groups (Table2). The improvement in time to ex-haustion was significantly greater inthe interval versus the traditional train-ing group (Figure 3). Ventilatory effi-ciency (calculated from the whole testor before the anaerobic threshold) wasnot significantly changed as a result of
either training program (Table 2). Anaerobic thresholdwas significantly increased in both groups but to agreater extent in the interval training group. Ventilationat stage 2 of exercise (submaximal) was decreased,and peak ventilation was increased to a similar extentin both groups (p �0.05). There were no adverseeffects as the result of participating in either trainingprogram.
• • •The key finding of this investigation was that high-
intensity interval training results in similar improve-ments in aerobic fitness and a greater tolerance to ananaerobic challenge in comparison to traditional con-tinuous aerobic exercise training. Also, it appears thathigh-intensity interval training can be conducted withminimal risk to highly functional patients with CAD.
Interval training is widely utilized in healthy pop-ulations to improve aerobic performance. In athletes,the greatest improvements in VO2peak are observedwhen high-intensity (i.e., 90% to 100% heart ratereserve) exercise is incorporated into the training re-gime. Despite the widespread usage of interval train-ing in healthy populations, few studies have evaluatedthe effect of interval training on the health status ofpatients with cardiovascular disease. In recent years,interval training has been advocated for the rehabili-tation of patients with severe chronic heart failure.10
Preliminary research also indicated that interval train-ing may be more effective in improving exercise ca-pacity than continuous aerobic training.4 This researchhas generally dealt with less functional patients withcardiovascular disease. However, a recent study in
Data presented as mean � SD or numbers.*Significant change with exercise intervention (p �0.05).†Some patients had multiple diagnoses, including myocardial infarction and coronary artery bypass
surgery.
BRIEF REPORTS 1081
patients with CAD (with similar fitness levels to ourparticipants) revealed that interval training may resultin a 10% greater improvement in VO2peak than lowerintensity training,11 which is contrary to our findings.This discrepancy is commonly found in the literaturewith healthy subjects and is likely to be the result ofseveral factors related to differences in experimentaldesigns and/or procedures.2 In comparison to ourstudy, Rogmo et al11 utilized interval training boutsthat were of longer duration and lower intensity. Theyalso included rest phases (during interval training) thatwere similar to the average training intensity for thecontinuous training group. Therefore, although theauthors attempted to maintain a consistent volume ofwork, the average training intensity and duration weredifferent between conditions. Also, in the previousstudy, the control patients exercised at a markedlylower training intensity (i.e., 50% to 60% of peakVO2) than our participants. These differences may
explain the divergent findings between studies. Wetook every effort to create training environments thatwere comparable in terms of intensity, total work, andprogression between groups. As such, we feel confi-dent that we were able to comprehensively evaluatethe differential effects of traditional versus intervaltraining on VO2peak in highly functional patients withCAD. It is also important to note, that we believed thatit was imperative to maintain the average trainingintensity of 65% of VO2 reserve between the tradi-tional and interval training groups. This was due toseveral factors including: (1) previously showing thatboth intensity and duration have independent effectson cardiovascular function,12 and (2) at our institutiontraditional cardiac rehabilitation for highly functionalpatients with CAD involves training at 60% to 70% ofVO2 reserve. Furthermore, research with enduranceathletes has consistently shown that the average train-ing intensity is approximately 65% of VO2 reserve(even if 2 or 3 interval training sessions per week areincluded).13 Because the aim of our study was toevaluate the effects of a training program that is com-monly used in the healthy athletic realm it was essen-tial that we keep the average training intensity at 65%of VO2 reserve. Our findings are consistent with otherresearch that has applied stringent controls betweeninterval and continuous training.2
Another important and novel finding of the presentinvestigation was that neither training program signif-icantly improved ventilatory efficiency,9, which is incontrast to what occurs after training in persons withchronic heart failure.14 The ventilatory efficiency val-ues of our participants were consistent with that ob-served in healthy age-matched men,15, and well belowthe level (i.e., �34) that has been associated with anincreased risk for premature mortality.16 Thus, itwould appear, based on our participants’ ventilatoryefficiency values, that they have a more favorableprognosis than less functional patients (such as pa-tients with heart failure). This gives further support forthe usefulness of ventilatory efficiency for the evalu-
FIGURE 1. Effects of interval and traditional cardiac rehabilitationon peak aerobic power (mean � SD). *Significant training effect(p <0.05).
FIGURE 2. Effects of interval and traditional cardiac rehabilitationon Bruce treadmill time to exhaustion (mean � SD). *Significanttraining effect (p <0.05).
FIGURE 3. Effects of interval and traditional cardiac rehabilitationon treadmill time to exhaustion (mean � SD). *Significant trainingeffect (p <0.05). **Significantly greater change after interval train-ing (p <0.05).
1082 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
TABLE 2 Individual Cardiorespiratory Measures During Exercise Before and After 16 Weeks of Rehabilitation
Values reported for heart rate, pulse pressure, rate–pressure product (RPP), rating of perceived exertion (RPE), oxygen pulse, and ventilation (peak VE) were measured during peak exercise.*Significant change with exercise intervention (p �0.05).†Significantly greater change after interval training (p �0.05).Stage 2 VE � ventilation at the end of Bruce stage 2; VE-VCO2 � slope of the relationship between ventilation and carbon dioxide during exercise; VO2 at AT � oxygen consumption at the anaerobic threshold (AT).
BRIEFREPO
RTS1083
ation of the risk of premature mortality in persons withcardiovascular disease.
Unique to this investigation, we revealed that in-terval training resulted in a greater improvement intime to exhaustion during a high-intensity exercise testand anaerobic threshold during incremental exercise.This is consistent with findings in healthy individu-als.17 Thus, although interval training may not neces-sarily improve VO2peak to a greater extent than thetraditional rehabilitation program, it does appear tolead to adaptations that allow for a greater tolerance toa strenuous exercise challenge. These adaptationswould be of particular benefit for the performance ofmany activities of daily living.6,7 Further support forthe importance of high-intensity training for func-tional status is provided by research that revealed thatthis form of training improves cardiac function (in-cluding submaximal stroke volume, improved myo-cardial contractility, and an increased ejection fractionat peak exercise) and exercise tolerance.18–20 Theseadaptations have also been shown to be greater afterhigh-intensity training in comparison to low-intensitytraining.19 Furthermore, high-intensity training hasbeen shown to reduce the incidence of angina and todecrease ST-segment depression at a given rate–pres-sure product during exercise in previously symptom-atic patients.18 Thus, it is apparent that high-intensitytraining can improve submaximal and maximal car-diovascular function while reducing ischemia at thesame myocardial oxygen consumption. Accordingly,our research has important implications for overallhealth status, as our participants may be able to per-form activities of daily living with less effort and fora longer period of time. Thus, it appears that intervaltraining may provide additional health benefits tohighly functional patients with CAD.
1. Jolliffe JA, Rees K, Taylor RS, Thompson D, Oldridge N, Ebrahim S.Exercise-based rehabilitation for coronary heart disease. Cochrane Database SystRev 2000;4:CD001800.
2. Warburton DE, Haykowsky MJ, Quinney HA, Blackmore D, Teo KK, TaylorDA, McGavock J, Humen DP. Blood volume expansion and cardiorespiratoryfunction: effects of training modality. Med Sci Sports Exerc 2004;36:991–1000.3. Meyer K, Samek L, Schwaibold M, Westbrook S, Hajric R, Lehmann M,Essfeld D, Roskamm H. Physical responses to different modes of intervalexercise in patients with chronic heart failure—application to exercise training.Eur Heart J 1996;17:1040–1047.4. Meyer K, Lehmann M, Sunder G, Keul J, Weidemann H. Interval versuscontinuous exercise training after coronary bypass surgery: a comparison oftraining-induced acute reactions with respect to the effectiveness of the exercisemethods. Clin Cardiol 1990;13:851–861.5. Haennel RG, Quinney HA, Kappagoda CT. Effects of hydraulic circuit trainingfollowing coronary artery bypass surgery. Med Sci Sports Exerc 1991;23:158–165.6. Warburton DE, Gledhill N, Quinney A. Musculoskeletal fitness and health.Can J Appl Physiol 2001;26:217–237.7. Warburton DE, Gledhill N, Quinney A. The effects of changes in musculo-skeletal fitness on health. Can J Appl Physiol 2001;26:161–216.8. Borg G. Psychophysical bases of perceived exertion. Med Sci Sports Exerc1982;14:377–387.9. Clark AL, Skypala I, Coats AJ. Ventilatory efficiency is unchanged afterphysical training in healthy persons despite an increase exercise tolerance. J Car-diovasc Risk 1994;1:347–351.10. Meyer K. Exercise training in heart failure: recommendations based oncurrent research. Med Sci Sports Exerc 2001;33:525–531.11. Rognmo O, Hetland E, Helgerud J, Hoff J, Slordahl SA. High intensityaerobic interval exercise is superior to moderate intensity exercise for increasingaerobic capacity in patients with coronary artery disease. Eur J Cardiovasc PrevRehabil 2004;11:216–222.12. Warburton DE, Gledhill N, Quinney HA. Blood volume, aerobic power, andendurance performance: potential ergogenic effect of volume loading. ClinJ Sport Med 2000;10:59–66.13. Robinson DM, Robinson SM, Hume PA, Hopkins WG. Training intensity ofelite male distance runners. Med Sci Sports Exerc 1991;23:1078–1082.14. Davey P, Meyer T, Coats A, Adamopoulos S, Casadei B, Conway J, SleightP. Ventilation in chronic heart failure: effects of physical training. Br Heart J1992;68:473–477.15. Sun XG, Hansen JE, Garatachea N, Storer TW, Wasserman K. Ventilatoryefficiency during exercise in healthy subjects. Am J Respir Crit Care Med2002;166:1443–1448.16. Gitt AK, Wasserman K, Kilkowski C, Kleemann T, Kilkowski A, Bangert M,Schneider S, Schwarz A, Senges J. Exercise anaerobic threshold and ventilatoryefficiency identify heart failure patients for high risk of early death. Circulation2002;106:3079–3084.17. Poole DC, Gaesser GA. Response of ventilatory and lactate thresholds tocontinuous and interval training. J Appl Physiol 1985;58:1115–1121.18. Ehsani AA, Biello DR, Schultz J, Sobel BE, Holloszy JO. Improvement ofleft ventricular contractile function by exercise training in patients with coronaryartery disease. Circulation 1986;74:350–358.19. Oberman A, Fletcher GF, Lee J, Nanda N, Fletcher BJ, Jensen B, CaldwellES. Efficacy of high-intensity exercise training on left ventricular ejection frac-tion in men with coronary artery disease (the Training Level Comparison Study).Am J Cardiol 1995;76:643–647.20. Hagberg JM, Ehsani AA, Holloszy JO. Effect of 12 months of intenseexercise training on stroke volume in patients with coronary artery disease.Circulation 1983;67:1194–1199.
1084 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Effectiveness of Inhibition of Cholesteryl Ester TransferProtein by JTT-705 in Combination With Pravastatin in
Type II Dyslipidemia
Jan Albert Kuivenhoven, PhD, Greetje J. de Grooth, MD, PhD, Hitoshi Kawamura, PhD,Anke H. Klerkx, PhD, Francois Wilhelm, MD, PhD, Mieke D. Trip, MD, PhD, and
John J. P. Kastelein, MD, PhD
The inhibition of cholesteryl ester transfer protein(CETP) has recently been shown to effectively increasehigh-density lipoprotein (HDL) cholesterol. This studyexamined the use of the CETP inhibitor JTT-705 com-bined with pravastatin. In a randomized, double-blind, placebo-controlled trial, 155 patients with typeII dyslipidemia using pravastatin 40 mg were treatedwith placebo or JTT-705 300 or 600 mg. Four weeksof treatment with JTT-705 600 mg led to a 30%decrease in CETP activity (p <0.001), a 28% increasein HDL cholesterol (p <0.001), and a 5% decrease inlow-density lipoprotein cholesterol (p <0.03). Combina-tion therapy using JTT-705 and pravastatin effectivelyincreases HDL cholesterol levels and is safe and welltolerated up to 4 weeks of administration. �2005 byExcerpta Medica Inc.
(Am J Cardiol 2005;95:1085–1088)
Increasing high-density lipoprotein (HDL) choles-terol levels by the inhibition of cholesteryl ester
transfer protein (CETP) to reduce atherosclerotic car-diovascular disease is gaining momentum after recentpublications on 2 small molecule inhibitors of CETP:JTT-7051 and torcetrapib.2,3 Enthusiastic reviews4 un-derline the great expectations of these novel drugs.The number of clinical trials investigating pharmaceu-tical CETP inhibition, however, is very limited, andthere is an urgent need for large randomized trials tojudge safety and efficacy and to assess the long-termeffects of CETP inhibition on atherosclerosis. Giventoday’s clinical practice of managing dyslipidemia,another issue that warrants investigation is the safetyand efficacy of combination therapy of CETP inhibi-tion with evidence-based low-density lipoprotein(LDL) cholesterol reduction. This was addressed inthe present study, which investigated the combinationof JTT-705 and pravastatin in mildly dyslipidemicpatients.
• • •
The study cohort comprised 120 men and 35women with type II dyslipidemia with LDL choles-terol �160 mg/dl, HDL cholesterol �60 mg/dl, andtriglycerides �400 mg/dl. Exclusion criteria wereidentical to those described for JTT-705 mono-therapy.1 The trial was a 16-week, multicenter, ran-domized, double-blinded, placebo-controlled studyevaluating the efficacy and safety of the combined useof JTT-705 and pravastatin. A run-in period of 8weeks of treatment with pravastatin 40 mg/day wasfollowed by a 4-week period of combination treatmentwith placebo or JTT-705 300 or 600 mg and a 4-weekfollow-up period with pravastatin only. Participantswere randomized to 1 of the 3 treatment regimens.Blood samples were drawn after overnight fasting.The counting of returned tablets and empty packageswas used to monitor compliance. The trial was ap-proved by the ethics committees of all participatingcenters, and all participants gave informed consent.
Total cholesterol and triglycerides were measuredby established methods (reagents from BoehringerMannheim GmbH, Mannheim, Germany, and Techni-con). HDL cholesterol was determined with a heparinmanganese chloride precipitation reagent (Sigma, St.Louis, Missouri),5 and LDL cholesterol was calculatedby the Friedewald formula.6 HDL cholesterol subfrac-tions were determined after serial ultracentrifugation,and apolipoproteins were measured using establishedmethods (reagents from Dade Behring Holding GmbH,Liederbach, Germany). CETP activity was measured asdescribed elsewhere.7 The intra- and interassay coeffi-cients of variation were 1.43% and 1.26%, respec-tively. Plasma CETP concentration was measured by asandwich enzyme-linked immunosorbent assay using2 monoclonal antibodies specific to human CETP,JHC1 and JHC2, as described.8 The intra- and inter-assay coefficients of variation were 3.56% and 5.42%,respectively. Safety monitoring included physical ex-amination (vital signs and weight), electrocardiogram,routine hematology, biochemistry (including aspartateaminotransferase, alanine aminotransferase, �2-mi-croglobulin, and creatinine), and urinalysis.
Analysis of all parameters was assessed by fittingan analysis of variance model with percentage changeas the dependent variable and treatment and center asthe fixed effects. Analyses included all patients whoentered the run-in period and received �1 dose of thestudy medication. Statistical analyses were performedwith SAS software (SAS Institute Inc., Cary, NorthCarolina).
From the Department of Vascular Medicine, Academic Medical Cen-ter, Amsterdam, The Netherlands; and AKROS Pharma Inc., Princeton,New Jersey. This study was supported by a grant from AKROS PharmaInc., Princeton, New Jersey. Dr. Klerkx was supported by a grant fromthe Dutch Heart Foundation, Amsterdam, The Netherlands (NHS2000.073 and NHS 2003B191). Dr. Kastelein’s address is: Depart-ment of Vascular Medicine, Academic Medical Center, Meibergdreef9, Room F4-159.2, 1105 AZ Amsterdam, The Netherlands. E-mail:[email protected]. Manuscript received August 13, 2004;revised manuscript received and accepted December 20, 2004.
Table 1 indicates that most of the randomizedpatients were men. The 3 groups did not differ regard-ing baseline demographic characteristics, lipids, li-poproteins, apolipoproteins, CETP activity, or CETPconcentration.
Table 2 lists the absolute and percentage changes
from baseline for CETP activity,CETP concentration, lipids, and apo-lipoproteins after 4 weeks of treat-ment with placebo or JTT-705 (300and 600 mg) in combination withpravastatin 40 mg/day (p values de-note comparisons with baseline val-ues). After 4 weeks of treatment,CETP activity was inhibited by 30%in the JTT-705 600-mg dose group(p �0.001) and resulted in a 28%increase in HDL cholesterol levels.This increase in HDL cholesterolwas due to significant increases inHDL2 and HDL3 and was accompa-nied by significant increases in apo-lipoprotein-AI (p �0.001). These ef-fects were furthermore accompaniedby a 103% increase in CETP concen-tration (p �0.001). In the large-dosegroup, we also recorded a 5% de-crease of LDL cholesterol levels (p�0.03 at 4 weeks). Total cholesterol,triglycerides, apolipoprotein-B, andapolipoprotein-E levels were, how-ever, not significantly changed afterJTT-705 administration. Comparedwith the placebo group, the athero-genicity index, represented by LDLcholesterol/HDL cholesterol ratios,was reduced by 11% and 23% in theJTT-705 300- and 600-mg dosegroups, respectively. JTT-705 hadsimilar effects in men and women(data not included).
One 59-year-old man died duringthe pravastatin run-in phase of acuteheart failure. JTT-705 300 and 600mg was well-tolerated and exhibiteda clean safety profile. Furthermore,the discontinuation rate for adverseevents was limited to 1 patient in theplacebo group and 1 patient in theJTT-705 300 mg dose group. Oneserious adverse event in the placebogroup was reported in the 4-weekperiod of combination therapy: a 62-year-old man discontinued treatmentbecause of cholelithiasis. Anotherpatient discontinued JTT-705 300mg because of moderate gastroenter-itis. Slightly more gastrointestinal,respiratory, and musculoskeletal ad-verse events were noted in the JTT-705 600-mg dose group than the pla-cebo group (see Table 3).
Safety tests revealed 6 significant laboratory ab-normalities: 1 increased eosinophilic count in the pla-cebo group, 2 increased chloride values and 1 revers-ible increase of blood potassium value in the smalldose group, and 1 reversible increase of white bloodcell count and 1 low lymphocyte count in the large
TABLE 1 Demographic Characteristics, Baseline Lipid and Lipoprotein Parametersof Randomized Type II Dyslipidemic Patients Using Pravastatin 40 mg
Values are means � SDsNone of the parameters were statistically different between the groups.*For control, we used plasma drawn from 49 normolipidemic healthy volunteers.
TABLE 2 Changes in CETP Activity and Concentration, Lipids, and LipoproteinParameters From Baseline in Type II Dyslipidemic Patients Receiving Pravastatin40 mg After 4 Weeks of Treatment With Placebo or JTT-705 300 or 600 mg
Values are given as mean � SD. Numbers in parentheses refer to percent changes from baseline.Statistically significant compared with baseline values:*p �0.001; †p � 0.01; ‡p � 0.03; §p � 0.05.
1086 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
dose group. Overall, no dose-related changes of com-plete blood cell count, liver function, creatinine phos-phokinase, renal function, blood glucose, and bloodchemistry were noted. During and after the study, nodose-related changes of vital signs, blood pressure,and body weight were observed.
• • •In a previous analysis, JTT-705 monotherapy was
shown to induce a dose-dependent decrease of CETPactivity with a concomitant increase in HDL choles-terol in mild dyslipidemic patients.1 In clinical prac-tice, HDL cholesterol–increasing medication willlikely be used in combination with evidence-basedstatin therapy.9 We therefore assessed such combina-tion therapy and show that the increase in HDL cho-lesterol with the daily use of JTT-705 combined withpravastatin daily equals that seen with JTT-705 mono-therapy.1 In the large dose group, JTT-705 further-more induced a significant albeit moderate 5% LDLcholesterol reduction on top of the LDL cholesterol–reducing effect of pravastatin. Importantly, the JTT-705–pravastatin combination therapy did not raise no-table safety concerns or exhibit serious adverseeffects. Mild gastrointestinal effects, as seen withmonotherapy, were again observed in this study, butthese did not appear to be dose dependent and werealso present in placebo-treated patients.
During the preparation of this report, torcetrapib, asecond CETP inhibitor, was shown to also effectivelyincrease HDL cholesterol levels.2,3 Overall, the effectsof this compound on most lipid parameters at dosagesranging from 10 to 240 mg/day were more profoundthan those observed for JTT-705 at dosages up to 900mg/day. At present, however, it is entirely unknownhow much CETP inhibition is needed to achieve theinhibition of atherosclerosis, an issue that needs fur-ther investigation.
When comparing the results of the 2 CETP inhib-itors in more detail, some intriguing findings emerge:JTT-705 900 mg/day was previously shown to in-
crease HDL cholesterol by 34% anda 37% reduction of CETP activitylevels in subjects with mild dyslipi-demia (HDL cholesterol 1.17 mmol/L). In patients with lower baselineHDL cholesterol levels (1.01 mmol/L), torcetrapib 120 mg/day, however,induced a 46% increase in HDL cho-lesterol with only 28% CETP inhibi-tion. These data may suggest thatCETP inhibition is more effective inpatients with lower HDL cholesterollevels. Regarding the effect of CETPinhibition on plasma triglyceride lev-els, we previously reported that JTT-705 900 mg/day monotherapy pro-duced a nonsignificant 10% reductionin triglyceride levels, whereas combi-nation therapy with pravastatin in thepresent investigation resulted in asmall but significant 8% reduction (p� 0.05). Although torcetrapib 120 mg
twice daily induced a 26% reduction (p � 0.05) intriglycerides, no effect was observed when this drugwas administered at the same dose but once daily.Thus, CETP inhibition by the 2 investigational drugsappears to affect plasma triglycerides only to a limitedextent. This is interesting given that under normalphysiologic conditions, triglycerides are generally in-versely associated with HDL cholesterol levels. Itshould be noted, however, that the data are limited andthat larger trials are needed to establish the effect ofCETP inhibition on plasma triglycerides. With respectto apolipoprotein levels, the effects of JTT-705 andtorcetrapib on apolipoprotein levels are quite similarexcept for apolipoprotein-E. Torcetrapib, whethercombined with atorvastatin or not, induced a markedincrease of apolipoprotein-E levels (up to 66%). Incontrast, JTT-705 did not affect apolipoprotein-E con-centration when used either as monotherapy or incombination with pravastatin. This suggests that the 2drugs, although very similar in their modes of action,may have different effects on the residence time ofapolipoprotein-E in the circulation or, alternatively,may affect apolipoprotein-E gene expression. Fromthe available data, it is not clear whether it is the HDLcholesterol fraction that is apolipoprotein-E enrichedand whether this will affect the anticipated antiathero-genicity of CETP inhibition. Finally, JTT-705 andtorcetrapib induce a marked dose-dependent increasein CETP concentration. In contrast, CETP inhibitionby antisense oligonucleotides in rabbits reduced CETPconcentration levels.10 This interesting aspect is cur-rently being investigated, but it is possible that theCETP:CETP inhibitor complex acquires the half-lifeof the HDL cholesterol particle, thereby increasing theapparent CETP mass in the circulation.
Focusing on the main objective of HDL cholesterol-increasing therapy, that is, reducing cardiovascularevents, the present results allow only careful specula-tion. Only 300 patients who received CETP inhibitorshave been described. Combination treatment with sta-
TABLE 3 Most Frequent Adverse Events (�5% in �1 Treatment Group) and SafetyData After 4 Weeks of Placebo, or JTT-705 300 or 600 mg/day inCombination With Pravastatin 40 mg in Type II Dyslipidemic Patients
tins was restricted to 101 patients with JTT-705 andonly 9 patients with torcetrapib. The effects on lipidprofiles are promising, but these small numbers un-derline that our knowledge of pharmacologic CETPinhibition in humans is limited. Although the effectson surrogate end points for atherosclerosis are antici-pated soon, many questions still need to be answered.For example, how will CETP inhibition affect lipidmetabolism in patients with a high-triglyceride, low–HDL cholesterol phenotype? Together, the currentdata provide a basis for optimism, but only the long-term use of CETP inhibitors, either as monotherapy orcombined with other lipid-lowering medications, willtell whether this novel drug will reduce the burden ofcardiovascular disease. Fortunately, the clean safetyprofile is encouraging.
1. de Grooth GJ, Kuivenhoven JA, Stalenhoef AF, de Graaf J, Zwinderman AH,Posma JL, van Tol A, Kastelein JJ. Efficacy and safety of a novel cholesteryl estertransfer protein inhibitor, JTT-705, in humans: a randomized phase II dose-response study. Circulation 2002;105:2159–2165.
2. Brousseau ME, Schaefer EJ, Wolfe ML, Bloedon LT, Digenio AG, Clark RW,Mancuso JP, Rader DJ. Effects of an inhibitor of cholesteryl ester transfer proteinon HDL cholesterol. N Engl J Med 2004;350:1505–1515.3. Clark RW, Sutfin TA, Ruggeri RB, Willauer AT, Sugarman ED, Magnus-Aryitey G, Cosgrove PG, Sand TM, Wester RT, Williams JA, et al. Raisinghigh-density lipoprotein in humans through inhibition of cholesteryl ester transferprotein: an initial multidose study of torcetrapib. Arterioscler Thromb Vasc Biol2004;24:1–9.4. Barter PJ, Brewer HB Jr, Chapman MJ, Hennekens CH, Rader DJ, Tall AR.Cholesteryl ester transfer protein: a novel target for raising HDL and inhibitingatherosclerosis. Arterioscler Thromb Vasc Biol 2003;23:160–167.5. Warnick GR, Albers JJ. A comprehensive evaluation of the heparin-manganeseprecipitation procedure for estimating high density lipoprotein cholesterol. JLipid Res 1978;19:65–76.6. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration oflow-density lipoprotein cholesterol in plasma, without use of the preparativeultracentrifuge. Clin Chem 1972;18:499–502.7. Kato H, Nakanishi T, Arai H, Nishida HI, Nishida T. Purification, microhet-erogeneity, and stability of human lipid transfer protein. J Biol Chem 1989;264:4082–4087.8. Sawada S, Sugano M, Makino N, Okamoto H, Tsuchida K. Secretion of prebeta HDL increases with the suppression of cholesteryl ester transfer protein inHepG2 cells. Atherosclerosis 1999;146:291–298.9. Illingworth DR. Achievement of low-density lipoprotein cholesterol goals:new strategies to address new guidelines. Cardiol Clin 2003;21:363–375.10. Sugano M, Makino N, Sawada S, Otsuka S, Watanabe M, Okamoto H,Kamada M, Mizushima A. Effect of antisense oligonucleotides against cho-lesteryl ester transfer protein on the development of atherosclerosis in cholester-ol-fed rabbits. J Biol Chem 1998;273:5033–5036.
Effect of Chronic Total Coronary Occlusion onTreatment Strategy
Ryan D. Christofferson, MD, Kenneth G. Lehmann, MD, Gary V. Martin, MD,Nathan Every, MD, James H. Caldwell, MD, and Samir R. Kapadia, MD
In a registry analysis of 8,004 consecutive patientspresenting for diagnostic catheterization at a singleinstitution from 1990 to 2000, chronic total occlusion(CTO) was found in 52% of patients with significant(>70% diameter stenosis) coronary artery disease.Peripheral vascular disease was the strongest clinicalpredictor of the presence of a CTO. In a multivariateanalysis, CTO was the strongest predictor against theselection of percutaneous coronary intervention (PCI)as a treatment strategy, indicating that efforts to im-prove the success rate of PCI in CTO may have asignificant impact on management of coronarydisease. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1088–1091)
The frequency of chronic total occlusion (CTO) ofcoronary arteries in an unselected population is not
well known, and previous studies from major referralcenters are subject to selection bias. In addition, theinfluence of CTO on the management of coronarydisease, particularly the selection of percutaneous cor-
onary intervention (PCI) versus bypass surgery, is alsounknown. The purpose of this study was to evaluatethe frequency of CTO in an unselected population andto investigate the clinical predictors of CTO and theinfluence of CTO on treatment recommendations.
• • •This study was conducted with the approval of the
institutional review board of the University of Washington.We reviewed diagnostic cardiac catheterization datafrom patients investigated for suspected or knowncoronary artery disease at the Veterans AdministrationHospital in Seattle, Washington, from 1990 to 2000.The Puget Sound Health Care System has 291 acutehospital beds and approximately 80,000 enrolled pa-tients. Nine staff cardiologists practiced in this centerduring the study period. According to institutionalnorms, data were prospectively collected at the time ofcardiac catheterization, including age, gender, medicalhistory, coronary angiographic results, the ejectionfraction, and treatment recommendations. These datawere retrospectively reviewed to assess for the fre-quency of CTO. Patients with previous bypass surgeryor recent (�3 months) myocardial infarctions wereexcluded to identify patients with CTO. To analyzerisk factors for CTO, we compared patients with sig-nificant coronary disease but no CTO with patientswith CTO.
Significant coronary disease was defined as �1lesion of �70% luminal diameter stenosis. Total cor-
From the Division of Cardiovascular Diseases, Cleveland Clinic Founda-tion, Cleveland, Ohio; and the Division of Cardiology, VA Puget SoundHealth Care System, University of Washington, Seattle, Washington.Dr. Kapadia’s address is: Cleveland Clinic Foundation, 9500 EuclidAvenue, F25, Cleveland, Ohio 44195. E-mail: [email protected] received August 31, 2004; revised manuscript receivedand accepted December 29, 2004.
onary artery occlusion was defined as 100% luminaldiameter stenosis without a discernable lumen and theabsence of anterograde flow. CTO was defined as totalcoronary artery occlusion of �3 months in duration.Multivessel disease was defined as �70% stenosis in�2 vessels.
Data are reported as mean � SD.Continuous variables were comparedusing Student’s t test, and categoricalvariables were compared using thechi-square test. Multiple logistic re-gression was used to perform a mul-tivariate analysis of risk factors forCTO and influence on the choice oftherapy. A p value of �0.05 wasconsidered statistically significant.Data were analyzed with the Statis-tica software package, version 6.0(StatSoft, Inc., Tulsa, Oklahoma).
We identified 8,004 consecutivepatients from 1990 to 2000 who un-derwent diagnostic cardiac catheter-ization at a single institution. We ex-cluded 1,423 patients with previousbypass surgery or recent myocardial
infarctions. Of the remaining 6,581 patients, 3,087patients (47%) had significant coronary artery disease.Of patients with significant coronary artery disease, atotally occluded coronary artery was present in 1,612(52%). In all, 2,130 vessels were totally occluded,with 375 patients (12%) having �1 CTO.
Compared with patients without CTO, those withCTO had more frequent hypertension, smoking, andperipheral vascular disease (Table 1). The ejectionfraction was significantly less in patients with CTO.Multivessel disease was present more frequently inpatients with CTO. A multivariate analysis found thatperipheral vascular disease was the strongest predictorof CTO (odds ratio [OR] 1.67, 95% confidence inter-val [CI] 1.35 to 2.08, p �0.001).
The territory supplied by the left anterior descend-ing artery was involved in 444 patients (28%), the leftcircumflex artery in 560 patients (35%), and the rightcoronary artery in 1,027 patients (64%) (Table 2).
Compared with patients without CTO, those withCTO were recommended for PCI less frequently (11%vs 36%, p �0.0001) and bypass surgery (40% vs 28%,p �0.0001) or medical therapy (49% vs 35%, p�0.0001) more frequently (Figure 1).
A multivariate analysis of predictors of recommen-dation for bypass surgery or PCI was performed witha priori selection of the following variables for inclu-sion: age, the presence of diabetes, CTO, and mul-tivessel disease (Figures 2 and 3). This analysis foundthat CTO was not an independent predictor of recom-mendation for bypass surgery (OR 1.1, 95% CI 0.94 to1.31, p �0.21). The strongest predictor of recommen-dation for bypass surgery was multivessel disease (OR6.2, 95% CI 5.1 to 7.4, p �0.0001), followed byincreased age (OR 1.014 per year increase, 95% CI1.006 to 1.022, p �0.001). A second a priori multi-variate analysis of predictors of PCI with the samevariables found that increased age, the presence ofdiabetes, CTO, and multivessel disease were all inde-pendent risk factors against PCI, with the strongestpredictor being CTO (OR 0.26, 95% CI 0.22 to 0.31,p �0.0001).
This study is the largest series to describe theincidence of CTO in an unselected, community-basedpopulation. We found that a large percentage of pa-tients in our population with significant coronary dis-ease had 1 CTO (52%) or �1 CTO (12%). Thiscontrasts with previous studies, which have docu-mented a lower rate of CTO in such patients (9% to40%).1–4 Of these studies, the only community-basedstudy, performed by Kahn,1 found that in the study’scohort, 35% of patients with significant coronary dis-ease (�50% stenosis) had CTOs.1 Our greater preva-lence of CTO may be the result of a more stringentdefinition of significant coronary artery disease(�70% diameter stenosis), leading to selection biasfor more extensive coronary disease. It may also re-flect a greater coronary disease quantity of plaque inthis veteran population.
Our study is the first to report that hypertension,smoking, and peripheral vascular disease confer agreater risk for CTO in patients with significant cor-onary disease. Previous studies have either found nounivariate risk factors,5 younger age,6 or previousmyocardial infarction2,6 to be univariate predictors ofCTO. These differences may be a reflection of thegreater prevalence of hypertension, smoking, and pe-ripheral vascular disease in our group than the previ-ously mentioned cohorts. Peripheral vascular disease
may be the strongest risk factor for CTO, becauseclaudication could limit exercise tolerance, preventinganginal symptoms from developing and allowing astenosis to reach a high grade without clinical recog-nition.
The presence or absence of CTO appears to play apivotal role in the recommendation for therapy.2,7 Inour cohort, patients with CTO were recommended forbypass surgery or medical therapy more frequentlythan PCI. It is intriguing to note that in a multivariableanalysis, we found that CTO was the strongest pre-dictor against PCI but did not predict a recommenda-tion for bypass surgery. Instead, multivessel diseasewas the strongest predictor of bypass surgery. In de-ciding to recommend bypass surgery, it appears thatother factors, such as multivessel disease and age,were more important than CTO in this cohort.
This study has significant limitations. Most pa-tients were men, so our findings may not apply towomen. During the study, significant advances weremade in the medical and catheter-based treatment ofcoronary disease, which may have influenced thetreatment choices and introduced bias. Finally, thisstudy reflects the treatment recommendations givenonly at a single institution and may be a product of thelocal institutional bias.
The trend in cardiovascular medicine mirrors asimilar trend in all areas of medicine, whereby pa-tients and physicians are seeking less invasive alter-natives to traditional surgical treatments. The aging ofthe population and the increasing prevalence of obe-sity, end-stage renal disease, and other surgical riskfactors may also lead to an increased number of pa-tients who are less suitable for cardiac surgery. Newcatheter-based developments in CTO management, suchas intravascular ultrasound-guided wiring,8 hydrophiliccoated guidewires,9 tapered-tip guidewires,10 optical re-flectometry,11 and coronary artery stenting12,13 havethe potential to increase the initial success rate of PCI.The high rate of restenosis and reocclusion after PCIof CTO may be significantly reduced by drug-elutingstents.14 These advances are likely to make the per-cutaneous treatment of CTO more feasible, durable,and safe in the future.
1. Kahn JK. Angiographic suitability for catheter revascularization of totalcoronary occlusions in patients from a community hospital setting. Am Heart J1993;126:561–564.2. Delacretaz E, Meier B. Therapeutic strategy with total coronary artery occlu-sions. Am J Cardiol 1997;79:185–187.3. Srinivas VS, Brooks MM, Detre KM, King SB III, Jacobs AK, Johnston J,Williams DO. Contemporary percutaneous coronary intervention versus balloonangioplasty for multivessel coronary artery disease: a comparison of the NationalHeart, Lung and Blood Institute Dynamic Registry and the Bypass AngioplastyRevascularization Investigation (BARI) study. Circulation 2002;106:1627–1633.4. El Gaylani N, McAdam BF, White U, Gearty GF, Walsh MJ, Crean PA.Immediate and follow-up results of coronary angioplasty—lessons for the future.Ir Med J 1996;89:60–61.5. Suero JA, Marso SP, Jones PG, Laster SB, Huber KC, Giorgi LV, JohnsonWL, Rutherford BD. Procedural outcomes and long-term survival among patientsundergoing percutaneous coronary intervention of a chronic total occlusion innative coronary arteries: a 20-year experience. J Am Coll Cardiol 2001;38:409–414.
FIGURE 2. OR plot for the multivariable analysis of predictors ofcoronary artery bypass grafting (CABG).
FIGURE 3. OR plot for the multivariable analysis of predictors ofPCI.
1090 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
6. Violaris AG, Melkert R, Serruys PW. Long-term luminal renarrowing aftersuccessful elective coronary angioplasty of total occlusions. A quantitative an-giographic analysis. Circulation 1995;91:2140–2150.7. Hamm CW, Reimers J, Ischinger T, Rupprecht HJ, Berger J, Bleifeld W. Arandomized study of coronary angioplasty compared with bypass surgery inpatients with symptomatic multivessel coronary disease. German AngioplastyBypass Surgery Investigation (GABI). N Engl J Med 1994;331:1037–1043.8. Matsubara T, Murata A, Kanyama H, Ogino A. IVUS-guided wiring tech-nique: promising approach for the chronic total occlusion. Catheter CardiovascInterv 2004;61:381–386.9. Kahler J, Koster R, Brockhoff C, Reimers J, Baldus S, Terres W, Meinertz T,Hamm CW. Initial experience with a hydrophilic-coated guidewire for recanalizationof chronic coronary occlusions. Catheter Cardiovasc Interv 2000;49:45–50.10. Saito S, Tanaka S, Hiroe Y, Miyashita Y, Takahashi S, Satake S, Tanaka K.Angioplasty for chronic total occlusion by using tapered-tip guidewires. CatheterCardiovasc Interv 2003;59:305–311.
11. Morales PA, Heuser RR. Chronic total occlusions: experience with fiber-opticguidance technology—optical coherence reflectometry. J Interv Cardiol 2001;14:611–616.12. Sirnes PA, Golf S, Myreng Y, Molstad P, Emanuelsson H, Albertsson P,Brekke M, Mangschau A, Endresen K, Kjekshus J. Stenting in Chronic CoronaryOcclusion (SICCO): a randomized, controlled trial of adding stent implantationafter successful angioplasty. J Am Coll Cardiol 1996;28:1444–1451.13. Buller CE, Dzavik V, Carere RG, Mancini GB, Barbeau G, Lazzam C,Anderson TJ, Knudtson ML, Marquis JF, Suzuki T, et al. Primary stenting versusballoon angioplasty in occluded coronary arteries: the Total Occlusion Study ofCanada (TOSCA). Circulation 1999;100:236–242.14. Hoye A, Tanabe K, Lemos PA, Aoki J, Saia F, Arampatzis C, Degertekin M,Hofma SH, Sianos G, McFadden E. Significant reduction in restenosis after theuse of sirolimus-eluting stents in the treatment of chronic total occlusions. J AmColl Cardiol 2004;43:1954–1958.
Outcomes Following Elective Percutaneous CoronaryIntervention Without On-Site Surgical Backup in a
Community Hospital
Alexander Paraschos, MD, PhD, Dwayne Callwood, MD,Marilyn B. Wightman, MSN, MBA, James E. Tcheng, MD, Harry R. Phillips, MD,
Gary L. Stiles, MD, John M. Daniel, BA, and Michael H. Sketch, Jr., MD
Despite guidelines to the contrary, limited numbers ofelective percutaneous coronary intervention (PCI) pro-cedures without on-site surgical backup are beingperformed, particularly in Europe and Canada. In theUnited States, many hospitals are considering estab-lishing on-site surgical programs, in part to facilitatePCI. At a hospital with only off-site surgical backup,562 elective PCI procedures were performed on 489consecutive patients. Of these, 551 (98.0%) were suc-cessfully completed without major in-hospital compli-cations; 5 patients (1.0%) had in-hospital complica-tions, and 4 (0.8%) were urgently transferred. It isconcluded that elective PCI with off-site surgical backupis feasible and safe for selected patients under specificconditions. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1091–1093)
Even before the use of stents became widespread,emergency percutaneous coronary intervention
(PCI) during acute myocardial infarction without on-site surgical backup was found to be feasible andsafe.1–5 In Europe and Canada, this paradigm has beenextended in many institutions to elective PCI. Al-though not endorsed by guidance committees,6 similarprograms now also exist in the United States.7,8 Themain criticism of these programs has not been thatthey are unsafe but that they are unnecessary: there arenearly 800 open-heart surgical programs throughout
the United States.9 About 500 additional centers havecatheterization laboratories but no open-heart pro-grams10; although many perform only diagnostic cath-eterizations, those that do perform PCI without on-sitesurgical facilities are considered by some not to havelegitimate interventional programs. To add to the de-bate on whether it is wise to expend health careresources to establish on-site cardiac surgery pro-grams, we sought to determine whether elective PCIwith off-site surgical backup might be feasible andsafe in selected patients, given a well-developed trans-fer plan.
• • •From February 1998 to October 2002, we per-
formed 562 PCI procedures on 489 patients. All pa-tients provided standard clinical informed consent,which included information that only off-site surgicalbackup was being provided; as an alternative, patientswere offered PCI at Duke University Medical Center.To identify patients at lower risk for cardiovascularcomplications, patients selected for PCI met exclusioncriteria agreed on by Alamance Regional MedicalCenter and Duke University Medical Center; specifi-cally, no patient’s age was �75 years, and no patienthad class III or IV heart failure, left ventricular func-tion �30%, acute myocardial infarction, cardiogenicshock, PCI immediately after thrombolytic therapy,refractory unstable angina, left main or 3-vessel dis-ease, PCI of �2 major vessels, collaterals originatingfrom the vessel targeted for intervention, complexlesion morphology (ostial location, bifurcation, heavycalcification, intracoronary thrombus, or total occlu-sion), or vein grafts. These characteristics have beenshown to be associated with a moderate to high riskfor in-hospital adverse events.6,11,12 Data on patientsexcluded were not systematically collected during the
From Alamance Regional Medical Center, Burlington, North Carolina;the Duke Clinical Research Institute; and Duke University MedicalCenter, Durham, North Carolina. Dr. Sketch’s address is: Duke Uni-versity Medical Center, PO Box 3157, Durham, North Carolina27710. E-mail: [email protected]. Manuscript received Sep-tember 3, 2004; revised manuscript received and accepted Decem-ber 27, 2004.
study except from January to November 2003, whenan average of 16.1 patients per month were excludedby these criteria. On average, 11.1 patients per monthunderwent PCI at Alamance during the study.
To ensure quality and compliance with the speci-fied exclusion criteria, the first 20 patients were pre-viewed prospectively with interventionalists fromDuke. In addition, the first 50 patients were reviewedretrospectively to assess procedural success and com-plications. All procedures were performed by 2 fel-lowship-trained, board-eligible interventional cardiol-ogists; since their fellowship training, the 2 operatorshad maintained caseloads of �75 PCI procedures peryear. Procedural success was defined as a residualstenosis �20% without in-hospital major complica-tions (emergency coronary bypass surgery, transmuralmyocardial infarction, or death). Myocardial infarc-tion after PCI was defined as the elevation of creatinekinase-MB levels to �3 times the upper limit ofnormal. If patients arrived with elevated enzymes,myocardial infarction was defined as a further 50%increase from baseline after PCI. All major complica-tions were reviewed by a performance improvementcommittee comprising representatives from the 2 in-stitutions.
PCI was performed through the femoral arteryusing standard techniques. All patients received aspi-rin and intracoronary nitroglycerin. In coronary stentcases, patients received either ticlopidine or clopi-dogrel. Heparin was administered to achieve an acti-vated clotting time of 300 to 350 seconds or 200 to250 seconds when glycoprotein IIb/IIIa inhibitorswere given. Before October 2001, abciximab was theglycoprotein IIb/IIIa inhibitor of choice; subsequently,virtually all patients received eptifibatide. All patientswere monitored overnight in the coronary care unit.
Off-site surgical backup was provided by DukeUniversity Medical Center, located 34 miles (mostlyInterstate highway) from Alamance Regional MedicalCenter. A Duke mobile intensive care ambulance wasalways present on-site during a PCI procedure, and incases of transfer to Duke, patients were accompaniedby the interventional cardiologist. An operating roomat Duke was also required to be on standby. Beforeinitiation of this program, we rehearsed the transfer ofmock patients to Duke to ensure that transfer could beaccomplished smoothly and quickly. After the pro-gram began, mock transfers continued to be practicedevery 6 months.
Demographic and angiographic characteristics aswell as in-hospital events were captured in Alamance’sPCI database for quality. Cases in the quality databasewere verified with appropriate charge codes inAlamance’s cost accounting system. Angiographic char-acteristics were also tabulated in the Duke InformationSystem for Cardiac Care database. In-hospital eventswere captured by case managers, staff members, andmedical records coders. Vascular access site complica-tions, including retroperitoneal bleeds, psuedoaneu-rysms, and arteriovenous fistulae, were verified by com-paring event data in the quality database withInternational Classification of Diseases, Ninth Revision,Clinical Modification diagnosis codes.
Descriptive statistics were performed on demo-graphic characteristics, vessel disease, and in-hospitalevents. Data analysis was performed using statisticalsoftware from SAS Institute. Inc. (Cary, North Carolina).
Table 1 lists the clinical and procedural character-istics of the 489 patients who underwent elective PCIduring the study. Most of the patients were men withsingle-vessel disease. With rare exceptions, patientshad ejection fractions �35%. Most patients were in-patients admitted to the emergency department withacute ischemic syndromes. As Table 1 indicates, 87%of patients received intracoronary stents, and 93%were treated with glycoprotein IIb/IIIa receptor antag-onists.
Of the 562 procedures, 551 (98.0%) were success-ful (Table 2). There were 11 unsuccessful procedures(2.0%). In 6 patients (1.1%), we were unable to crossthe lesion with a guidewire. Three of these patientswere referred to Duke, where they eventually under-went successful PCI. The other 3 patients were treatedmedically without sequelae.
Major in-hospital complications were observed in5 patients (0.9%); overall, there were 22 major andminor complications, for an event rate of 3.9% (Table3). Four patients (0.7%) were urgently transferred
Procedural (n � 562)Stent use 87%Glycoprotein IIb/IIIa inhibitor use 93%
TABLE 2 Procedural Success* (562 Procedures)
Procedural Outcome n %
Successful 551 98.0Unsuccessful 11 2.0
Unable to cross 6 1.1Major complications 5 0.9
*Defined as �20% residual stenosis without a major in-hospital complica-tion (coronary artery bypass graft, transmural myocardial infarction, or death).
TABLE 3 Adverse Cardiac Events
Adverse Event n% (95%
confidence interval)
Transmural myocardial infarction 0 0.0 (0.0–0.0)Emergency revascularization 4 0.7 (0.01–1.41)In-hospital death 1 0.2 (0.17–0.52)Enzyme elevation to 3 times the
1092 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
without incident and underwent successful bypass sur-gery. One patient who was urgently transferred under-went further attempts at PCI at Duke University Med-ical Center but eventually required a bypass graft thefollowing day. The mean door-to-door time (fromdeparture from the Alamance catheterization suite toarrival in the Duke catheterization suite in preparationfor surgery) for the transferred patients was 40 min-utes; the mean time from departure from Alamance tothe operating room at Duke was 83 minutes.
One in-hospital death occurred, although not as aresult of acute vessel closure. The patient had a suc-cessful PCI with an excellent angiographic result.However, the patient developed acute renal failurebelieved to be multifactorial in origin secondary to acombination of newly initiated angiotensin-convertingenzyme inhibitor therapy, transient hypotension, andcontrast nephropathy. The patient died at another in-stitution while awaiting emergency dialysis.
Two deaths (0.4%) occurred out of the hospitalseveral days after PCI. We assume that these patientslikely experienced in-stent thrombosis, although theexact nature of their deaths is unknown. Over theentire study, only 30 patients (5.3%, with 100% fol-low-up) were readmitted during the first 6 months forclinical restenosis within the same vessel (althoughnot necessarily the same lesion).
• • •Using selective criteria coupled with attention to
logistics, we found PCI with off-site surgical backupto be safe and very successful. We observed a 98.0%PCI procedural success rate, comparable with ratesfound in the Stent Restenosis Study and the BelgianNetherlands Stent Study (96.1% and 92.7%, respec-tively),13,14 and the total event rate for major andminor in-hospital complications was only 3.9%. Therate of major in-hospital complications was only 0.9%.We did not observe a single case of vessel closure aftersuccessful PCI over 4.5 years of follow-up.
The evolution of stent technology and adjunctivepharmacology has contributed to the improved suc-cess of PCI,15 emboldening sites to perform PCI with-out surgical backup. Concurrently, some argue thatPCI without on-site surgical backup is not warrantedbecause of the abundance of open-heart programs inthe United States. The impetus to establish a propor-tion of these programs may have been to legitimize aninterventional program rather than to serve actualpatient demand.16,17 Studies have associated worsepatient outcomes with small-volume open-heart pro-grams versus institutions that perform greater volumesof surgical procedures.18–20 In the report of Ting andcolleagues,8 a 99.5% success rate (195 of 196 consec-utive patients) was achieved at an institution at whichsurgical backup, even off-site, was not considered.The success of the program appeared to be largely dueto rigorous selection criteria and the availability of
immediate consultation using telemedicine. As an al-ternative, on the basis of our experience, we concludethat PCI can be safely performed with off-site surgicalbackup and may obviate the need for a small-volumeopen-heart program at every site that performs PCIwhile facilitating the performance of PCI (in con-trolled circumstances) in the community.
1. Grines LL, Browne KF, Marco J, Rothbaum D, Stone GW, O’Keefe J, OverlieP, Donohue B, Chelliah N, Timmis GC, for the PAMI Study Group. A compar-ison of immediate angioplasty with thrombolytic therapy for acute myocardialinfarction. N Engl J Med 1993;328:673–679.2. Zijlstra F, Jan de Beer M, Hoorntje JCA, Reiffers S, Reiber JH, SuryapranataH. A comparison of immediate coronary angioplasty with intravenous streptoki-nase in acute myocardial infarction. N Engl J Med 1993;328:680–684.3. Weaver WD, Litwin PE, Martin JS. Use of direct angioplasty for treatment ofpatients with acute myocardial infarction in hospitals with and without on-sitecardiac surgery. Circulation 1993;88:2067–2075.4. Wharton TP, McNamara NS, Fedele FA, Jacobs MI, Gladstone AR, Funk EJ.Primary angioplasty for the treatment of acute myocardial infarction: experienceat two community hospitals without cardiac surgery. J Am Coll Cardiol 1999;33:1257–1265.5. Aversano T, Aversano CT, Passamani E, Knatterud GL, Terrin ML, WilliamsDO, Forman SA. Thrombolytic therapy vs. primary percutaneous coronary in-tervention for myocardial infarction in patients presenting to hospitals withouton-site cardiac surgery. JAMA 2002;287:1943–1951.6. American College of Cardiology/American Heart Association Task Force onPractice Guidelines. ACC/AHA guidelines for percutaneous coronary interven-tion (revision of the 1993 PTCA guidelines). J Am Coll Cardiol 2001;37:2239i–2239li.7. Vogel JHK. Changing trends for surgical standby in patients undergoingpercutaneous transluminal coronary angioplasty. Am J Cardiol 1992;69(suppl):25F–32F.8. Ting HH, Garratt KN, Singh M, Kjelsberg MA, Timimi FK, Cragun KT,Houlihan RJ, Boutchee KL, Crocker CH, Cusma JT, et al. Low-risk percutaneouscoronary interventions without on-site cardiac surgery: two years’ observationalexperience and follow-up. Am Heart J 2003;145:278–284.9. Hospital Statistics. 1988 Ed. Chicago: American Hospital Association, 1988.10. Directory of Cardiac Catheterization Laboratories in the United States. 4thEd. Raleigh, North Carolina: Society for Cardiac Angiography and Interventions,1997.11. Tan K, Sulke N, Taub N, Sowtno E. Clinical and lesion morphologicdeterminants of coronary angioplasty success and complications: current experi-ence. J Am Coll Cardiol 1995;25:855.12. Van Domburg RT, Saia F, Lemos PA. Coronary artery bypass surgery andpercutaneous transluminal coronary angioplasty in patients with multivesseldisease. Minerva Cardioangiol 2003;51:599–608.13. Fischman DL, Leon MB, Baim DS, Schatz RA, Savage MP, Penn I, Detre K,Veltri L, Ricci D, Nobuyoshi M. A randomized comparison of coronary stentplacement and balloon angioplasty in the treatment of coronary artery disease.N Engl J Med 1994;331:496–501.14. Serruys PW, de Jaegere P, Kiemeneij F, Macaya C, Rutsch W, HeyndrickxG, Emanuelsson H, Marco J, Legrand V, Materne P. A comparison of balloon-expandable-stent implantation with balloon angioplasty in patients with coronaryartery disease. N Engl J Med 1994;331:489–495.15. Altmann DB, Racz M, Battleman DS, Bergman G, Spokojny A, Hannan EL.Reduction in angioplasty complications after the introduction of coronary stents:results from a consecutive series of 2242 patients. Am Heart J 1996;132:503–507.16. Bonche LI. Should surgical support within the same institution be required forpercutaneous transluminal coronary angioplasty? Ann Thorac Surg 1989;48:159–160.17. Ullyot DJ. Surgical standby for coronary angioplasty. Ann Thorac Surg1990;50:3–4.18. Hannan EL, Kilburn HJ, Bernard H, O’Donnell JF, Lukacik G. Coronaryartery bypass surgery: the relationship between in-hospital mortality rate andsurgical volume after controlling for clinical risk factors. Med Care 1991;29:1094–1107.19. Clark RE. Outcome as a function of annual coronary artery bypass graftvolume: the Ad Hoc Committee on Cardiac Surgery Credentialing of the Societyof Thoracic Surgeons. Ann Thorac Surg 1996;61:21–26.20. Birmeyer JD, Siewers AE, Finlayson EVA, Stukel TA, Lucas FL, Batista I,Welch HG, Wennberg DE. Hospital volume and surgical mortality in the UnitedStates. N Engl J Med 2002;246:1128–1137.
BRIEF REPORTS 1093
Follow-Up of Coronary Artery Bypass Graft Patency byMultislice Computed Tomography
Emilio Chiurlia, MD, Mila Menozzi, MD, Carlo Ratti, MD, Renato Romagnoli, MD,and Maria Grazia Modena, MD
Fifty-two consecutive asymptomatic patients with atotal of 166 bypass grafts were investigated by 16-slice computed tomography (CT) 15 � 5 days beforeinvasive coronary angiography. Overall, 165 grafts(99.4%) were assessable with multislice CT (MSCT).Coronary angiography showed that 111 grafts (67%)were patent and 54 (33%) were occluded. Of thepatent grafts, 22 had high-grade stenoses. MSCT cor-rectly classified 111 grafts as patent and 54 as oc-cluded. Of the patent grafts, 16-slice CT correctlydetected 21 bypass stenoses (95%). These resultsyielded 100% sensitivity and specificity of 16-slice CTfor detecting bypass grafts occlusion and 96% sensi-tivity and 100% specificity for detecting high-gradestenoses in patent grafts. �2005 by Excerpta Med-ica Inc.
(Am J Cardiol 2005;95:1094–1097)
Coronary angiography is the gold standard to eval-uate the status of grafts, but it is an invasive
procedure that includes x-ray exposure, requires hos-pitalization, and includes a (small) risk for complica-tions.1,2 These disadvantages make use of coronaryangiography less attractive as a diagnostic method forthe evaluation of early and late postoperative graftspatency. Electron beam computed tomography (CT)and magnetic resonance imaging have been investi-gated for noninvasive coronary artery and bypass graftimaging.3–5 However, neither of these approaches hasresulted in an accurate diagnostic procedure that iswidely accepted. Multislice CT (MSCT) with 4 rowsof detectors has shown promising results, despite alarge percentage of nonassessable bypass grafts due toinadequate spatial and temporal resolution and to ascan time of approximately 40 seconds.6 Our aim wasto compare the newest generation of cardiac MSCTscanners that permit the simultaneous acquisition ofup to 16 submillimeter slices per rotation, with con-ventional angiography for the evaluation of graft oc-clusion and the presence of significant coronary ste-noses.
• • •From January to May 2004, 52 consecutive asymp-
tomatic patients (45 men and 7 women; mean age 63years, range 35 to 84) with a total of 166 bypass grafts
(117 venous and 49 arterial grafts) were investigatedby 16-slice CT 15 � 5 days before invasive coronaryangiography. The mean time between bypass opera-tion and MSCT investigation was 95 months (range 36to 168). Only patients in sinus rhythm and stableclinical condition, without implanted pacemakersand contraindications to the administration of iodin-ated contrast agent, were included in the study. Allpatients gave their written informed consent, andthe ethics committee of our university approved thestudy protocol.
Each patient was placed within the gantry of amultidetector row CT scanner (LightSpeed Ultra 16,GE Healthcare, Waukesha, Wisconsin) in a supineposition. Leads were attached for simultaneous elec-trocardiographic and image recording, which is nec-essary for inter-related image reconstruction. Our im-aging protocol consisted of the following steps. First,a noncontrast coronal view of the chest was obtainedto determine the position of the heart, define the scanvolume for further imaging, and recognize eventualcoronary calcification. Then, the patient contrast agenttransit time was measured from the injection into aperipheral vein of a 20-ml bolus of contrast agent(Iomeron 350, Bracco Ltd., Milan, Italy) to the in-crease in density within the ascending aorta. At last,the volume data set for coronary bypass imaging wasacquired in spiral mode with the simultaneous acqui-sition of 16 parallel slices using the algorithm of 180°multislice cardiac interpolation, during the intrave-nous injection of 120 ml of contrast agent at a rate of3.5 ml/s. Scan parameters were as follows: detectorcollimation 16 � 0.625, tube current 440 mA, andtube voltage 120 kVp. Depending on patient heart
From the Departments of Cardiology and Radiology, University ofModena and Reggio Emilia, Policlinico Hospital, Modena, Italy. Dr.Chiurlia’s address is: Via del Pozzo 71, 41100 Modena, Italy.E-mail: [email protected]. Manuscript received October 13,2004; revised manuscript received and accepted December 20,2004.
FIGURE 1. Volume rendering images before (A) and after (B) theautomatic removal of large cardiac chambers. Three patentgrafts, a left IMA (LIMA) graft and 2 saphenous vein grafts(SVGs), are visible simultaneously.
frequency, the scanner adapted table speed and as aconsequence the spiral pitch. Patients with heart rates�65 beats/min received metoprolol 5 to 10 mg intra-venously immediately before MSCT examination. Allimages were acquired during inspiratory breath hold(20 to 25 seconds) and reconstructed with a slicethickness of 0.6 mm and an interslice gap of 0.6 mmusing retrospective electrocardiographic gating at dif-ferent R-to R-wave intervals (from 35% to 95% inincrements of 10%). Finally, images were transferredto an offline dedicated Advantage Workstation (ver-sion 4.1) equipped with CardIQ software (GE Health-care) for image editing. For each graft and coronarytree segment, we selected the R- to R-wave interval
that allowed the best image quality.MSCT data sets were displayed in a 3-di-mensional volume rendering fashion andwith multiplanar reformation. MSCT im-ages were evaluated by 2 independent ra-diologists blinded to the conventional cor-onary angiographic findings.
Cardiac catheterization and contrast-enhanced x-ray coronary angiographywere done according to standard tech-niques. Multiple views of the coronaryarteries were obtained and stored on aCD-ROM.
The MSCT images were evaluatedby 2 independent radiologists experi-enced in cardiac CT and blinded to theangiographic results. Bypass grafts wereconsidered occluded if they were notidentified in cross-sectional images and3-dimensional images. Significant ste-noses of bypass grafts were defined asluminal diameter stenoses �50% on thebasis of cross-sectional images and mul-
tiplanar 3-dimensional images. Analysis of all angio-grams was performed by 2 independent experiencedcardiologists, who scored arterial and venous grafts byflow, in accordance with the system used in theThrombolysis In Myocardial Infarction Trial.7 Grades0 or 1 were taken as occluded grafts and all others aspatent grafts. A lesion was considered to be a signif-icant stenosis if the mean diameter reduction in 2orthogonal planes was �50%.
The sensitivity, specificity, and accuracy for thedetection of bypass graft patency and the presence ofsignificant stenosis was evaluated using conventionalcontrast angiography as a reference. The concordancebetween observers for the detection of coronary ste-nosis by MSCT was calculated using Cohen’s � value.
All patients completed MSCT examinations with-out complications. The average investigation time was15 minutes, although reconstruction took a further 20minutes. The average heart rate during the scan was 58� 6 beats/min. The total scan time was 19 � 2seconds. Overall, 165 of the possible 166 grafts wereassessable with MSCT. One graft was excluded fromanalysis because of numerous metal clips along itscourse. Of the 165 assessable bypass grafts, 117(71%) were saphenous vein grafts; 46 (28%) wereinternal mammary artery (IMA) grafts, including 37left IMA grafts and 9 right IMA grafts; and 2 were radialartery grafts. Of the 117 saphenous vein grafts, 47 wereto the left anterior descending artery, 25 to the obtusemarginal artery, 29 to the right coronary artery, 11 to thediagonal branch, 7 to the left circumflex coronary artery,and 7 to the posterior interventricular branch. Of theIMA grafts, there were 5 sequential bypasses to thediagonal branch and left anterior descending coronaryartery; the remaining were single grafts to the leftdescending anterior artery in 30 patients, to the diag-onal branch in 9 patients, and to the right coronaryartery in 2 patients. The 2 radial artery grafts were tothe obtuse marginal artery. Coronary angiography of
FIGURE 2. (A) A volume rendering image shows the occlusion (arrow) of saphe-nous vein grafts. (B) In the postcontrast transverse image, the occlusion of the sa-phenous vein grafts (arrow) is clearly seen.
FIGURE 3. A volume rendering image shows 2 significant steno-ses (arrows) in the middle segment of a saphenous vein graft tothe left anterior descending coronary artery; farther up, stentplacement (asterisk) is shown.
BRIEF REPORTS 1095
the bypass grafts showed that 111 bypass grafts (67%)were patent and that 54 (33%) were occluded (42saphenous vein grafts, 10 IMA grafts, and 2 radialartery grafts). Of the patent grafts, 89 were free ofsignificant stenoses, and 22 (19 saphenous vein grafts,2 left IMA grafts, and 1 right IMA graft) had high-grade stenoses. In 15 patients, the stenosis was local-ized in the graft itself, and in 7 patients, the stenosiswas in the region of the distal anastomosis. MSCTcorrectly classified 111 grafts as patent (Figure 1) and54 as occluded (Figure 2). Of the patent grafts, MSCTcorrectly detected 21 of the 22 bypass stenoses (Fig-ures 3 and 4); the only false-negative stenosis waslocalized at distal anastomosis. These results yieldedsensitivity and specificity of 100% of MSCT for de-tecting bypass graft occlusion and sensitivity of 96%and specificity of 100% for detecting high-grade ste-noses in patent grafts. The interobserver agreementbetween the 2 readers was excellent (� � 0.82, 95%confidence interval 0.76 to 0.87).
• • •The newest generation of MSCT scanners permits
the simultaneous acquisition of up to 16 submillimeterslices per rotation, and these scanners have greaterspatial and temporal resolution compared with previ-ous scanners.8 So far, only 1 study has used this newtechnology compared with conventional angiographyfor bypass evaluation, yielding images of adequatequality in all examinations.9 Compared with the datareported by Ropers et al,10 in this study, the number ofbypasses that could be evaluated increased from 62%to 74%, combined with improved diagnostic accuracyfor the detection of the stenosis. Our study confirmsthe excellent image quality of the new scanners. Only1 of the 166 grafts was not eligible for analysis be-cause of numerous metal clips along its course.
Despite these encouraging results, several limita-tions of our study need to be addressed. First, to obtaingood image quality, all patients with heart rates �65
beats/min received � blockers intravenously immedi-ately before the MSCT examination. This strategy isnot applicable to patients with chronic obstructivepulmonary disease. However, this is a minor limita-tion that concerns particularly native coronary arter-ies; bypass grafts, because of their size and relativeimmobility, are less affected by heart rate. In ourstudy, few grafts had extensive calcific deposits ormetal artifacts. The lower frequency of clip artifacts inour study may have also been influenced by differentsurgical techniques. Finally, we need to underline themajor drawback of retrospectively electrocardiograph-ically gated MSCT data acquisition: its large radiationexposure, which has been reported at between 6.7 and13.0 mSv.11 However, by reducing the tube outputduring heart phases that are not likely to be targeted bythe electrocardiographically gated reconstruction, adose reduction of 48% is possible.12
Even with these technical limitations, our studyshows that the newest generation of MSCT scannersallows the very accurate assessment of graft patencyand provides relevant information concerning thepresence of significant obstructive disease in bypassgrafts.
1. Ryan TJ. The coronary angiogram and its seminal contributions to cardiovas-cular medicine over five decades. Circulation 2002;10:752–756.2. Kennedy JW. Complications associated with cardiac catheterization and an-giography. Cathet Cardiovasc Diagn 1982;8:5–11.3. Achenbach S, Moshage W, Ropers D, Nossen J, Bachmann K. Noninvasive,three-dimensional visualization of coronary artery bypass grafts by electron beamtomography. Am J Cardiol 1997;79:856–861.4. Rubinstein RI, Askenase AD, Thickman D, Feldman MS, Agarwal JB, HelfantRH. Magnetic resonance imaging to evaluate patency of aortocoronary bypassgrafts. Circulation 1987;76:786–791.5. Langerak SE, Vliegen HW, de Roos A, Zwinderman AH, Jukema JW, KunzP, Lamb HJ, van Der Wall EE. Detection of vein graft disease using high-resolution magnetic resonance angiography. Circulation 2002;105:328–333.6. Engelmann MG, von Smekal A, Knez A, Kurzinger E, Huehns TY, Hofling B,Reiser M. Accuracy of spiral computed tomography for identifying arterial andvenous coronary graft patency. Am J Cardiol 1997;80:569–574.
FIGURE 4. Curved multiplanar reconstruction (A) and corresponding conventional coronary angiography (B) confirm the presence ofmultiple stenoses (arrows) in the middle segment of the venous graft; farther down, stent placement (asterisk) is shown.
1096 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
7. Chesebro JH, Knatterud G, Roberts R, Borer J, Cohen LS, Dalen J, Dodge HT,Francis CK, Hillis D, Ludbrook P. Thrombolysis In Myocardial Infarction (TIMI)Trial, phase I: a comparison between intravenous tissue plasminogen activatorand intravenous streptokinase. Clinical findings through hospital discharge. Cir-culation 1987;76:142–154.8. Pannu HK, Flohr TG, Corl FM, Fishman EK. Current concepts in multi-detector row CT evaluation of the coronary arteries: principles, techniques.Radiographics 2003;23:S111–S125.9. Kuettner A, Trabold T, Schroeder S, Feyer A, Beck T, Brueckner A, Heus-chmid M, Burgstahler C, Kopp AF, Claussen CD. Noninvasive detection ofcoronary lesions using 16-detector multislice spiral computed tomography tech-nology: initial clinical results. J Am Coll Cardiol 2004;44:1230–1237.
10. Ropers D, Ulzheimer S, Wenkel E, Baum U, Giesler T, Derlien H, MoshageW, Bautz WA, Daniel WG, Kalender WA, et al. Investigation of aortocoronaryartery bypass grafts by multislice spiral computed tomography with electrocar-diographic-gated image reconstruction. Am J Cardiol 2001;88:792–795.11. Hunold P, Vogt FM, Schmermund A, Debatin JF, Kerkhoff G, Budde T,Erbel R, Ewen K, Barkhausen J. Radiation exposure during cardiac CT: effectivedoses at multi-detector row CT and electron-beam CT. Radiology 2003;22:145–152.12. Jakobs TF, Becker CR, Ohnesorge B, Flohr T, Suess C, Schoepf UJ, ReiserMF. Multislice helical CT of the heart with retrospective ECG gating: reductionof radiation exposure by ECG-controlled tube current modulation. Eur Radiol2002;12:1081–1086.
Do Statins Increase the Risk ofIdiopathic Polyneuropathy?
Jeffrey L. Anderson, MD, Joseph B. Muhlestein, MD, Tami L. Bair, BS, Steven Morris,Aaron N. Weaver, MD, Donald L. Lappé, MD, Dale G. Renlund, MD,
Robert R. Pearson, BS, Kurt R. Jensen, MS, and Benjamin D. Horne, MStat, MPH
A recent European case-control study suggested thatstatins increase the risk for polyneuropathy, a rare butserious neurologic condition. This risk was assessed in272 patients with idiopathic polyneuropathy and 1,360matched controls in the Intermountain Health Care elec-tronic database. It was found that statin use beforediagnosis was not significantly greater in patients thancontrols (odds ratio 1.30, 95% confidence interval 0.3 to2.1, p � 0.27), nor were doses different between pa-tients and controls. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1097–1099)
Hydroxymethylglutaryl CoA reductase inhibitors (sta-tins) have become the dominant lipid-lowering
pharmacotherapy.1,2 Accumulating clinical trial datasupport the benefit of statins for chronic secondary andhigh-risk primary prevention.3–5 Statins recently havebeen shown to benefit patients when begun early afteracute coronary syndromes.6,7 Simvastatin is now avail-able in the United Kingdom from pharmacies on a non-prescription basis. Recent National Cholesterol Educa-tion Program Adult Treatment Panel III guidelines havebeen revised to allow the targeting of lower low-densitylipoprotein cholesterol levels (i.e., �70 mg/dl) in high-risk patients.8,9 Thus, statin use and doses are expected tocontinue to increase through prescription and futureover-the-counter use.
• • •With widespread statin use, there is growing con-
cern about rare but severe adverse reactions (e.g.,myopathy, rhabdomyolysis) as well as less severe butmore common reactions (e.g., low-grade myopathy
with normal creatine kinase levels).2,10,11 A recentDanish case-control study12 linked statin use withidiopathic polyneuropathy (IPN), which has not beenpreviously recognized as a potential hazard of statintherapy. This study followed smaller, earlier reports ofan association of neuropathy with statin use.13–18 Itwas suggested that statins might induce IPN by alter-ing nerve function by excessive depletion of choles-terol, a ubiquitous component of nerve cell mem-branes.12 Statins also co-inhibit the synthesis ofubiquinone, a key mitochondrial respiratory chain en-zyme, potentially disturbing neuronal energy produc-tion and use, inducing neuropathy.15
Polyneuropathy is not listed as a potential adversereaction to statin therapy in current product labeling,nor was it recognized as a potential side effect by arecent expert clinical advisory on statin safety.2 Thus,the finding of an association of statins with IPN, iftrue, might importantly affect the assessment of statinsafety as well as an understanding of the pathophysi-ologic mechanisms of IPN. Therefore, we tested thehypothesis that statin use predisposes to IPN in theIntermountain Health Care (IHC) database.
After receiving institutional and corporate consentfor the study, a search of the IHC electronic databasewas performed to find patients with an initial diagno-sis of IPN within the past 4 years. The IHC databasecontains records of all admissions to any of the 23IHC hospitals, located in Utah and its neighboringstates. It also contains records of outpatient visits tohospital-affiliated IHC clinics and physicians, and, inrecent years, has provided a link to prescription infor-mation from IHC pharmacies, which are preferentiallyused by IHC health insurance plan enrollees (1.2 millioncovered lives). The following codes, registered fromApril 2000 to April 2003, were sought: progressiveidiopathic neuropathy (International Classification ofDiseases, 9th Revision, codes 356.4 and 356.8), drug-induced polyneuropathy (code 357.6), or unspecifiedIPN (code 356.9). To eliminate cases of possible sec-ondary polyneuropathy, patients were excluded if they
From the Cardiovascular Department, LDS Hospital, IntermountainHealth Care; and the University of Utah School of Medicine, Salt LakeCity, Utah. This study was partially supported by the Deseret Founda-tion, Salt Lake City, Utah. Dr. Anderson’s address is: LDS Hospital,Cardiology Division, 8th Avenue and C Street, Salt Lake City, Utah84143. E-mail: [email protected]. Manuscript received Oc-tober 4, 2004; revised manuscript received and accepted December27, 2004.
had ever had diagnoses of diabetes, renal insuffi-ciency, alcohol abuse, cancer, hypothyroidism, ac-quired immunodeficiency syndrome, Lyme disease, orheavy metal intoxication. Electronically recorded pre-scription use and doses for atorvastatin, cerivastatin,fluvastatin, lovastatin, pravastatin, and simvastatinwere sought. One dose equivalent was defined asatorvastatin 10 mg/day, cerivastatin 0.2 mg/day, flu-vastatin 40 mg/day, lovastatin 40 mg/day, pravastatin40 mg/day, or simvastatin 20 mg/day.
Of 915,066 patients with electronic pharmacyclaims data, 272 met IPN criteria. Each of these pa-tients was matched to 5 controls (n � 1,360) selectedby the computer as the nearest IHC electronic recordentry date of the same gender and age (within 2 years).Statin use (timing, dose, type) was then determined forpatients and controls from the electronic records andcompared by chi-square test and analysis of variance.
The ages of patients and matched controls aver-aged 47 � 13 years; 57% were women. Statins wereprescribed for 7.2% overall (Table 1). The mean du-ration of statin use in these patients and controls at thetime of inquiry was 10.1 months. Statins were pre-scribed for 8.8% of patients (n � 24) and 6.9% ofcontrols (n � 94; odds ratio 1.30, 95% confidenceinterval [CI] 0.82 to 2.08, p � 0.27). Furthermore, theduration of statin use did not differ between treatedpatients and controls (18 � 10 vs 14 � 10 months, p �0.10), and average statin dose equivalents were similar(1.63 � 1.08 vs 1.83 � 1.37, respectively, p � 0.72).
• • •The potential for long-term exposure to statins to
substantially increase the risk for polyneuropathy wasraised by a 2002 report by Gaist et al.12 Their studyfrom Funen County, Denmark, was based on 166patients (35 definite, 54 probable, and 77 possible) and25 control subjects per case without polyneuropathyrandomly chosen and matched for age, gender, andindex date. Their results suggested an odds ratio forIPN with statin use of 3.7 (95% CI 1.8 to 7.6). Giventheir background IPN incidence of 1.7 per 10,000subject-years in unexposed subjects, they estimated anexcess of 4.5 per 10,000 patient-years, or 1 excesscase for every 2,200 patient-years of statin use. Thisrate is comparable with the excess incidence of overtmyopathy (with or without rhabdomyolysis) noted inlarge, placebo-controlled, long-term studies of statin
therapy.19 Moreover, for patients treated for �2 yearsand with definite IPN, the odds ratio was 26 (95% CI8 to 45).
This Danish case-control study followed smaller,earlier reports of an association of neuropathy withstatin use, including several case reports and 1 caseseries.13–18 A previous, smaller cohort study of sub-jects treated by general practitioners in the UnitedKingdom by Gaist et al17 estimated a relative risk forpolyneuropathy for current statin users of 2.5, a non-significant trend (95% CI 0.3 to 14).
In contrast, we were unable to confirm a significantassociation between IPN and statin use. Our studyhypothesis was prospectively tested, and our databasecontained a larger number of patients selected from anindependent database. We note that our 95% CIs (0.8to 2.1) did not allow us to exclude a modest, up totwofold, increase in IPN risk, but we were able toexclude the very large (fourfold to 14-fold) increasesreported on the basis of the Danish cohort. Our find-ings are supported by the 20,536-patient Heart Pro-tection Study, which did not report an excess of IPNduring an average of 5 years of therapy with simva-statin (40 mg/day).19 However, the Heart ProtectionStudy did not specifically assess IPN incidence in the2 treatment groups, so that small difference trendscould have been missed. Our study thus providescomplementary safety information by its specific fo-cus on IPN in a large, general population of medicallyinsured subjects.
Our study has certain limitations inherent to obser-vational studies, including potential ascertainment andother biases. The study power, although good, did notallow the exclusion of a modest increase (up to two-fold) in IPN risk. Also, given our average exposuretime of about 1 year, the possibility of increased IPNrisk during longer term (�2 to 5 years) exposurecannot be excluded. Medical records for IPN studypatients were not reviewed by a neurologist to ascer-tain and verify the basis for an IPN diagnosis; how-ever, patients in the IHC system have ready access tosubspecialty consultation, and an IPN diagnosis isgenerally not made by a general practitioner alone.
Given the substantial cardiovascular protective ef-fect of statins in higher risk populations, concernsabout IPN should not limit their use, according tocurrent recommendations.1 Our findings in this regardfor IPN are reassuring. However, as more aggressivelipid-lowering therapy is recommended8 and treat-ment is extended to larger and lower risk cohorts, rarebut serious adverse reactions can be expected to occurmore frequently, and these should be recognized andappropriate interventions undertaken, including thediscontinuation of offending agents.
In this large medical database of patients and con-trols, a link between IPN and statin use could not beconfirmed. Although we cannot exclude a modestincrease in the risk of up to twofold, the dramaticproposed increase, so as to affect 1 case per 2,200patient-years of statin use, is unlikely.12 These obser-vations are reassuring in view of increasingly wide-spread statin use, but additional observations are war-
TABLE 1 Statin Use in Patients Versus Controls
VariablePatients
(n � 272)Controls
(n � 1,360) p Value*
Age (yrs) 47 � 13 47 � 13 —Women 57% 57% —Statin use 8.8% 6.9% 0.27Duration of statin
use (mo)18 � 10 14 � 10 0.10
Dose equivalents† 1.63 � 1.08 1.83 � 1.37 0.72
*Chi-square test or analysis of variance. Means are presented as � SD.†One dose equivalence was defined as atorvastatin 10 mg/day, cerivasta-
tin 0.2 mg/day, fluvastatin 40 mg/day, lovastatin 40 mg/day, pravastatin 40mg/day, or simvastatin 20 mg/day.
1098 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
ranted to more clearly define or exclude a modestassociation of IPN with statin therapy.
1. NCEP Expert Panel. Executive Summary of the Third Report of the NationalCholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation,and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III).JAMA 2001;285:2486–2497.2. Pasternak RC, Smith SCJ, Bairey-Merz CN, Grundy SM, Cleeman JI, LenfantC. ACC/AHA/NHLBI clinical advisory on the use and safety of statins. Circu-lation 2002;106:1024–1028.3. Scandinavian Simvastatin Survival Group. Randomized trial of cholesterollowering in 4444 patients with coronary heart disease: the Scandinavian Simva-statin Survival Study. Lancet 1994;344:1383–1389.4. Shepherd J, Cobbe SM, Ford I, Isles CG, Lorimer AR, MacFarlane PW,McKillop JH, Packard CJ. Prevention of coronary heart disease with pravastatinin men with hypercholesterolemia. West of Scotland Coronary Prevention Group.N Engl J Med 1995;333:1301–1307.5. LaRosa JC, He J, Vupputuri S. Effect of statins on risk of coronary disease: ameta-analysis of randomized controlled trials. JAMA 1999;282:2340–2346.6. Schwartz GG, Olsson AG, Ezekowitz MD, Ganz P, Oliver MF, Waters D,Zeiher A, Chaitman BR, Leslie S, Stern T. Effects of atorvastatin on earlyrecurrent ischemic events in acute coronary syndromes: the MIRACL study: arandomized controlled trial. JAMA 2001;285:1711–1718.7. Cannon CP, Braunwald E, McCabe CH, Rader DJ, Rouleau JL, Belder R, JoyalSV, Hill KA, Pfeffer MA, Skene AM. Intensive versus moderate lipid loweringwith statins after acute coronary syndromes. N Engl J Med 2004;350:1495–1504.
8. Grundy SM, Cleeman JI, Bairey-Merz CN, Brewer HB Jr, Clark LT,Hunninghake DB, Pasternak RC, Smith SC Jr, Stone NJ. Implications of recentclinical trials for the National Cholesterol Education Program Adult TreatmentPanel III Guidelines. Circulation 2004;110:227–239.9. O’Keefe JH, Cordain L, Harris WH, Moe RM, Vogel R. Optimal low-densitylipoprotein is 50 to 70 mg/dL. J Am Coll Cardiol 2004;43:2142–2146.10. Phillips PS, Haas RH, Bannykh S, Hathaway S, Gray NL, Kimura BJ,Vladutiu GD, England JD. Statin-associated myopathy with normal creatinekinase levels. Ann Intern Med 2002;137:581–585.11. Donaghy M. Assessing the risk of drug-induced neurologic disorders. Statinsand neuropathy. Neurology 2002;58:1321–1322.12. Gaist D, Jeppesen U, Andersen M, Garcia Rodriguez LA, Hallas J, SindrupSH. Statins and risk of polyneuropathy. Neurology 2002;58:1333–1337.13. Jacobs MB. HMG-CoA reductase inhibitor therapy and peripheral neuropa-thy. Ann Intern Med 1994;120:970.14. Ahmad S. Lovastatin and peripheral neuropathy. Am Heart J 1995;130:1321.15. Phan T, McLeod JG, Pollard JD, Peiris O, Rohan A, Halpern JP. Peripheralneuropathy associated with simvastatin. J Neurol Neurosurg Psychiatry 1995;58:625–628.16. Jeppesen U, Gaist D, Smith T, Sindrup SH. Statins and peripheral neuropathy.Eur J Clin Pharmacol 2001;54:835–838.17. Gaist D, Garcia-Rodriguez LA, Huerta C, Hallas J, Sindrup SH. Are users oflipid-lowering drugs at increased risk of peripheral neuropathy? Eur J ClinPharmacol 2001;56:931–933.18. Lovastatin Study Groups I Through IV. Lovastatin 5-year safety and efficacystudy. Arch Intern Med 1993;153:1079–1087.19. Heart Protection Study Collaborative Group. MRC/BHF Heart ProtectionStudy of cholesterol lowering with simvastatin in 20,536 high-risk individuals: arandomised placebo-controlled trial. Lancet 2002;360:7–22.
Relation of Echocardiographic Wall Motion ScoreIndex and Response to Dobutamine Stress to
Defibrillation Threshold at the Time of Implantation ofa Cardiac Defibrillator
Ali M. Kizilbash, MD, Paul A. Grayburn, MD, Prasad Anand, MD,Robert C. Kowal, MD, PhD, Richard L. Page, MD, Kim Smith, RN,
and Mohamed H. Hamdan, MD
Predictors of defibrillation threshold (DFT) at implan-tation remain poorly defined. It was hypothesizedthat a greater wall motion score index (WMSI) at reston echocardiography and an ischemic or biphasic re-sponse on dobutamine stress echocardiography (DSE)would predict a greater DFT. Consecutive patientswho underwent implantable cardiac-defibrillator im-plantation underwent echocardiography and DSE.DFT at implantation was assessed using the step-down method. Thirteen men aged 68 � 10 years withleft ventricular ejection fractions of 28 � 10% partic-ipated in the study. DFTs ranged from 5 to 25 J (mean14 � 8). A WMSI at rest of >2.5 had a 71% positivepredictive value and an 83% negative predictive
value for a DFT >15 J. An ischemic or biphasic re-sponse during DSE did not predict a greater DFT atimplantation. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1099–1101)
Predictors of defibrillation threshold (DFT) at im-plantation remain poorly defined. DFT and safety
margin influence the choice of implantable cardiac-defibrillators (ICDs) used in patients referred for de-vice implantation. In addition, DFT affects the type ofantiarrhythmic agents offered to patients with atrialfibrillation or recurrent ventricular arrhythmias. Lim-ited data from animal studies suggest that ischemiaand larger heart sizes are associated with greater DFTs.1The purpose of this study was to prospectively assessthe utility of simple echocardiographic predictors ofDFT. We hypothesized that (1) a greater wall motionscore index (WMSI) at rest on echocardiography and(2) an ischemic or biphasic response on dobutaminestress echocardiography (DSE) would predict a greaterDFT.
• • •The study was conducted at the Dallas Veterans
Affairs Hospital after obtaining approval from the in-stitutional review board committee. All patients gave
From the University of Texas Southwestern Medical Center; the VeteransAffairs Medical Center; and Baylor University Medical Center, Dallas,Texas; the University of Washington School of Medicine, Seattle,Washington; and the University of Utah School of Medicine, Salt LakeCity, Utah. This study was partially supported by a grant from GuidantCorporation, Indianapolis, Indiana. Dr. Hamdan’s address is: Univer-sity of Utah, Department of Medicine, Division of Cardiology, 30North 1900 East, Room 4A100, Salt Lake City, Utah 84132-2401.E-mail: [email protected]. Manuscript received Sep-tember 2, 2004; revised manuscript received and accepted January5, 2005.
written informed consent. Consecutive patients withindications for ICDs were enrolled to avoid selectionbias. Before ICD implantation, each patient underwenttransthoracic echocardiography at rest and DSE ac-cording to the American Society of Echocardio-graphy’s2 published guidelines. WMSI at rest wasdefined as the sum of wall motion scores divided bythe number of segments scored. Ischemic responsewas defined as the development of a new or worseningwall motion abnormality in 1 of the myocardial seg-ments. New or worsening wall motion of 2 adjacentsegments was used to define ischemic response if thesegments were (1) adjacent to a scar or (2) present inthe basal inferior or posterior wall. A biphasic re-sponse was defined as an improvement of thickeningwith small-dose dobutamine in �1 segment that was
hypokinetic or akinetic at baseline,followed by deterioration at peakdose of dobutamine infusion. Isch-emic or biphasic response duringDSE was used to define myocardiumat risk.
ICDs were implanted using stan-dard transvenous techniques. DFTsat implantation were assessed using astep-down method starting at 20 Jand decrementing by 5 J. DFT wasreported as the smallest energy shockthat resulted in successful defibrilla-tion.
Thirteen men aged 68 � 10 yearswith left ventricular ejection frac-tions of 28 � 10% participated in thestudy. Eleven patients had coronaryartery disease, and 2 had dilated car-diomyopathy. Indications for ICD im-plantation included primary preven-tion (n � 6), syncope (n � 3),ventricular tachycardia (n � 3), andsudden cardiac death (n � 1). Two of13 patients were taking amiodarone.All patients had their ICDs im-planted in the left pectoral region. A2-coil lead was placed in the rightventricular apex in all patients. Anadditional atrial lead was placed in 5patients for dual-chamber pacing. Asummary of patients’ characteristicsis provided in Table 1.
DFTs ranged from 5 to 25 J(mean 14 � 8). A WMSI at rest of�2.5 had a 71% positive predictivevalue and an 83% negative predic-tive value for a DFT �15 J. Thecorrelation between WMSI at restand DFT was weak and did not reachstatistical significance. A graph de-picting the patients’ WMSIs andDFTs is provided in Figure 1. Elevenof 13 patients completed DSE. Fivepatients had ischemic or biphasic re-sponses. The presence of an ischemic
or biphasic response did not predict DFT at implan-tation.
• • •In this study, we explored whether commonly
available echocardiographic indexes predict DFT atimplantation. Our results support the use of echocar-diographically based WMSI at rest to predict a large(�15 J) or small (�15 J) DFT on the basis of athreshold value of 2.5. We found no relation betweenmyocardial ischemia on DSE and DFT.
WMSI at rest represents the extent and severity ofmyocardial dysfunction,3 is easy to derive, and hassignificant bearing on prognosis in patients with cor-onary artery disease. In addition, it is an independentprognostic factor and performs better than Killip class
FIGURE 1. Patients’ WMSIs and DFTs.
TABLE 1 Clinical Characteristics
Patient Age (yrs) EF (%) WMSI at Rest Ischemic/Biphasic DFT (J)
EF � left ventricular ejection fraction by echocardiography; ischemic/biphasic � myocardial responseon DSE.
1100 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
in predicting future cardiac events.4 The results of thisstudy support the use of echocardiographically basedWMSI at rest to predict DFT. A WMSI at rest of �2.5had a 71% positive predictive value and an 83%negative predictive value for a DFT �15 J. In con-trast, the presence of an ischemic or a biphasic re-sponse on DSE did not predict DFT at implantation.The apparent disparity from animal data is probablydue to the differences in the study design. In animalstudies, ischemic provocation, such as with coronaryocclusion, is often done before ventricular fibrillationinduction. In the present study, an ischemic or bipha-sic response represents myocardium with poor bloodflow during stress, not necessarily at rest. Thus, itappears that the timing of ischemia plays a major rolein determining DFT. The role of DSE in predicting
DFT after spontaneous ventricular fibrillation remainsto be determined.
1. Walcott GP, Killingsworth CR, Smith WM, Ideker RE. Biphasic waveformexternal defibrillation thresholds for spontaneous ventricular fibrillation second-ary to acute ischemia. J Am Coll Cardiol 2002;39:359–365.2. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, Feigenbaum H,Gutgesell H, Reichek N, Sahn D, Schnittger I, et al. Recommendations forquantitation of the left ventricle by two-dimensional echocardiography. AmericanSociety of Echocardiography Committee on Standards, Subcommittee on Quan-titation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr 1989;2:358–367.3. Heger JJ, Weyman AE, Wann LS, Rogers EW, Dillon JC, Feigenbaum H.Cross-sectional echocardiographic analysis of the extent of left ventricular asyn-ergy in acute myocardial infarction. Circulation 1980;61:1113–1118.4. Nishimura RA, Reeder GS, Miller FA Jr, Ilstrup DM, Shub C, Seward JB,Tajik AJ. Prognostic value of predischarge 2-dimensional echocardiogram afteracute myocardial infarction. Am J Cardiol 1984;53:429–432.
Effect of Syncope-Related Traumatic Injuries on theDiagnostic Evaluation and Syncope Recurrence of
Patients With Syncope and Apparently Normal Hearts
Andreas Schuchert, MD, Renke Maas, MD, Kai Mortensen, MD,Muhammet Ali Aydin, MD, Christina Kretzschmar, MD, and Thomas Meinertz, MD
The aims of this study were to assess the frequency oftraumatic injuries in patients with syncope and ap-parently normal hearts and their influence on thediagnostic evaluation and recurrence of syncope. Pa-tients presenting with syncope before head-up tilttesting frequently had a history of syncope-relatedinjuries. Syncope-related injuries seemed to be ran-dom: they was not related to patients’ histories, in-cluding the number of previous syncopal attacks, andthey had no predictive value for the outcome of head-up tilt testing or for the recurrence of syncope during1-year follow-up. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1101–1103)
Syncope is a symptom defined as a transient, self-limited loss of consciousness, usually leading to
falling.1 Because the onset of syncope is relativelyrapid in most patients, it may result in traumaticinjuries. Syncope-related injuries are severe compli-cations.2–4 A specific group is patients with syncope-related injuries in whom cardiac or neurologic causeshave been ruled out.5 The occurrence of syncope-related traumatic injuries seems to contradict the gen-eral assumption that syncope is a benign disorder inthis patient group.4 The present study assessed theprevalence of injuries in patients with first and recur-
rent syncope and apparently normal hearts and theirinfluence on the diagnostic evaluation and recurrenceof syncope after head-up tilt testing.
• • •This prospective study included 152 consecutive
patients with syncope at the time of their evaluationwith the head-up tilt test. Patients with syncope dueto any cardiac, neurologic, or psychiatric cause wereexcluded. Cardiac causes were assumed or excludedon the basis of history, physical examination, andelectrocardiogram at rest.6 When indicated in selectedcases, additional diagnostic procedures were per-formed.7 Additional exclusion criteria were carotidsinus hypersensitivity, depressed left ventricular func-tion, and conduction abnormalities on 12-lead electro-cardiography.
The patients underwent head-up tilt testing after 10minutes of lying supine.1 The drug-free stage lasted 45minutes, with a tilt angle of 60°. The subsequentpharmacologic provocation consisted of nitroglycerin400 �g given sublingually, with an additional 10minutes of head-up tilting. The test results were de-fined as positive as soon as syncope occurred.1 Afterthe evaluation, patients were given specific recom-mendations to prevent syncope recurrence. In individ-ual clinical decisions, 59 patients received � blockers,10 midodrine, and 5 a selective serotonin reuptakeinhibitor. After 12-month follow-up, the patients wereasked in telephone interviews whether they had hadsyncopal recurrences.
The patients were classified as those with firstsyncope, with 2 episodes of syncope in the previous 6months, and with �2 episodes of syncope in the
From Medical Clinic III, University-Hospital Hamburg-Eppendorf,Hamburg, Germany. Dr. Schuchert’s address is: Medical Clinic III,University-Hospital Hamburg-Eppendorf, Martinistr. 52, D-20246Hamburg, Germany. E-mail: [email protected]. Manu-script received September 20, 2004; revised manuscript receivedand accepted January 3, 2005.
previous 6 months. For each group, patients with andwithout syncope-related injuries were compared witheach other. An injury was defined as any traumarelated to the transient loss of consciousness. Thepatients were compared with respect to their histories,diagnostic evaluation, and outcomes after head-up tilttesting. All patients completed the 1-year follow-up.Statistical analyses were performed with the 2-tailedStudent’s t test or Fisher’s exact test. A p value �0.05was considered statistically significant.
The prevalence of traumatic injury was not signif-
icantly different distributed amongthe 3 patient groups, with 39% inpatients with first and 2 syncope ep-isodes within the previous 6 monthsand 51% in patients with �2 syn-cope episodes in the previous 6months. The 3 groups had similardemographic data (Table 1).
The mean number of diagnostictests before head-up tilt testing was notsignificantly different in patients withand without injury (Table 2). No sig-nificant differences were observed inthe occurrence of positive head-up tilttest results among and within thegroups (Figure 1). The recurrence ofpresyncope and syncope showed nosignificant difference between patientswith and without syncope-related inju-ries (Figure 2).
• • •The present study restricted analysis to patients
with apparently normal hearts. The frequency of syn-cope-related injuries was similarly distributed in thepatient groups with �1 previous episode of syncope.Our findings indicate that patients with syncope-re-lated injuries had not received specific diagnosticworkups compared with the remaining patients. Thisseemed to be unnecessary, because syncope-relatedtrauma did not predict an increased risk for syncoperecurrence. After the extended evaluation withhead-up tilt testing, the 1-year recurrence rate was low
FIGURE 1. Percentage of head-up tilt tests with positive results.
TABLE 1 Clinical Data of the Three Patient Groups
Variable
1 Syncopal Episode2 Syncopal EpisodesWithin Past 6 Months
1102 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
in the studied population compared with some previ-ous investigations.8–12 This applied to patients withfirst and recurrent syncopal spells.
There are 2 possible explanations. It is well acknowl-edged that many patients have fewer recurrences afterhead-up tilt testing than expected in patients withoutthis diagnostic test. Another explanation is that mostpatients have only 1 or very few syncope episodesduring their lifetimes.13 This could apply to the stud-ied patients, because the present study had consecu-tive inclusion. Frequent recurrences were observed inselected patients groups as indicated by the inclusioncriteria in these studies.8–12
All patients received head-up tilt tests for the eval-uation of their syncope. The diagnostic yield of head-up tilt testing has become questionable in recent yearsbecause of its low sensitivity and reproducibility.14
Nevertheless, most published randomized therapystudies included patients with positive head-up tilttest results. Any therapy has hardly been evaluatedin patients with negative head-up tilt test results. Anew concept for syncope prevention is to encouragepatients to perform isometric counter-pressure ma-neuvers.15,16 Such recommendations must be care-
fully considered in patients withsyncope-related injuries because ofthe rapid onset of their syncope.
1. Task Force on Syncope, European Society of Cardi-ology. Guidelines. Eur Heart J 2001;22:1256–1306.2. Chiu T, Pang P, Ying SY, Burd A. Syncope and burns.Burns 2004;30:438–442.3. Martikainen K, Seppa K, Viita P, Rajala S, LaippalaP, Keranen T. Transient loss of consciousness with andwithout injuries: where to treat these patients? EurJ Gen Pract 2003;9:91–95.4. Owings JT, Wisner DH, Battistella FD, Perlstein J,Walby WF, Tharratt RS. Isolated transient loss of con-sciousness is an indicator of significant injury. ArchSurg 1998;133:941–946.5. Goldschlager N, Epstein AE, Grubb BP, OlshanskyB, Prystowsky E, Roberts W, Scheinman M. Etiologicconsiderations in the patient with syncope an apparentlynormal heart. Arch Intern Med 2003;163:151–162.6. Farwell DJ, Sulke AN. Does the use of a syncopediagnostic protocol improve the investigation and man-agement of syncope? Heart 2004;90:52–58.7. Sarasin FP, Junod AF, Carballo D, Slama S, UngerPF, Louis-Simonet M. Role of echocardiography in the
evaluation of syncope: a prospective study. Heart 2002;88:363–367.8. Sutton R, Brignole M, Menozzi C, Raviele A, Alboni P, Giani P, Moya A.Dual-chamber pacing in the treatment of neurally mediated tilt-positive cardioin-hibitory syncope: pacemaker versus no therapy: a multicenter randomized study.The Vasovagal Syncope International Study (VASIS) Investigators. Circulation2000;102:294–299.9. Ammirati F, Colivicchi F, Santini M, Syncope Diagnosis and Treatment StudyInvestigators. Permanent cardiac pacing versus medical treatment for the preven-tion of recurrent vasovagal syncope: a multicenter, randomized, controlled trial.Circulation 2001;104:52–57.10. Moya A, Brignole M, Menozzi C, Garcia-Civera R, Tognarini S, Mont L,Botto G, Giada F, Cornacchia D, on behalf of the International Study on Syncopeof Uncertain Etiology (ISSUE) Investigators. Mechanism of syncope in patientswith isolated syncope and in patients with tilt-positive syncope. Circulation2001;104:1261–1267.11. Connolly SJ, Sheldon R, Thorpe KE, Roberts RS, Ellenbogen KA, Wilkoff BL,Morillo C, Gent M, VPS II Investigators. Pacemaker therapy for prevention ofsyncope in patients with recurrent severe vasovagal syncope: Second VasovagalPacemaker Study (VPS II): a randomized trial. JAMA 2003;289:2224–2229.12. Krahn AD, Klein GJ, Yee R, Skanes AC. Randomized assessment of syncopetrial: conventional diagnostic testing versus a prolonged monitoring strategy.Circulation 2001;104:46–51.13. Soteriades ES, Evans JC, Larson MG, Chen MH, Chen L, Benjamin EJ, LevyD. Incidence and prognosis of syncope. N Engl J Med 2002;347:878–885.14. Sagrista-Sauleda J, Romero B, Permanyer-Miralda G, Moya A, Soler-Soler J.Reproducibility of sequential head-up tilt testing in patients with recent syncope,normal ECG and no structural heart disease. Eur Heart J 2002;23:1706–1713.15. Brignole M, Croci F, Menozzi C, Solano A, Donateo P, Oddone D, PuggioniE, Lolli G. Isometric arm counter-pressure maneuvers to abort impending vaso-vagal syncope. J Am Coll Cardiol 2002;40:2053–2059.16. Krediet CTP, van Dijk N, Linzer M, van Lieshout JJ, Wieling W. Manage-ment of vasovagal syncope controlling or aborting faints by leg crossing andmuscle tensing. Circulation 2002;106:1684–1689.
FIGURE 2. Recurrent presyncope and syncope during 1-year follow-up.
BRIEF REPORTS 1103
Monitoring of Physical Activity and Heart RateVariability in Patients With Chronic Heart Failure
Using Cardiac Resynchronization Devices
Frieder Braunschweig, MD, PhD, Peter T. Mortensen, MD, Daniel Gras, MD,Wolfgang Reiser, MD, Thomas Lawo, MD, Hassan Mansour, MD, Peter Sogaard, MD,
Berthold Stegemann, PhD, Hans-Jürgen Bruns, PhD, and Cecilia Linde, MD, PhD,on behalf of the InSync III Study Investigators
Cardiac resynchronization therapy (CRT) devices notonly deliver effective treatment but may also serve asvaluable diagnostic tools in heart failure manage-ment. In the present study, the minutes of daily phys-ical activity and heart rate variability, measured bysensors incorporated into such a device, reflected theeffects of CRT and were related to New York HeartAssociation functional class. �2005 by Excerpta Med-ica Inc.
(Am J Cardiol 2005;95:1104–1107)
Apart from delivering effective therapy,1,2 cardiacresynchronization therapy (CRT) devices also of-
fer a unique possibility to collect critical diagnosticinformation on various physiologic variables over time.Such monitoring may provide valuable data on treat-ment efficacy and disease state and thus improve themanagement of patients with heart failure. Heart ratevariability (HRV), an established tool for the nonin-vasive assessment of the autonomic nervous tone, isrelated to the progression of heart failure.3–7 Further-more, an index of mean daily physical activity(MDPA) may reflect patients’ tolerance of exercisesimilarly to other exercise tests commonly used for theassessment of heart failure.8–11 The aim of the presentstudy was to evaluate if trend data on these physio-logic variables can be obtained from a CRT device.We hypothesized that such monitoring might provideuseful information on CRT efficacy and heart failureseverity. In contrast to previous trials on CRT focusingon patients in New York Heart Association (NYHA)classes III and IV, our study also included patients inclass II. Therefore, the response to CRT in relation toNYHA functional class was part of our analysis.
• • •Patients of legal age with symptomatic heart failure
in NYHA classes II to IV, QRS duration �130 ms (forpacemaker-dependent patients, paced QRS duration�180 ms), left ventricular ejection fractions �35%,left ventricular end-diastolic diameters �55 mm, andnormal sinus rhythm were included in the prospective,nonrandomized InSync III Study. Patients had to bereceiving optimal medical treatment before inclusion,and medication was kept stable until 3 months offollow-up were completed. Exclusion criteria wereany indication for an implantable cardioverter-defi-brillator, chronic atrial tachyarrhythmias, and the pres-ence of a mechanical right-sided cardiac valve. Thelocal ethics committees approved the study protocol,and all patients gave their signed informed consent toparticipation.
All patients received an InSync III atrio-biventricularpacemaker (model 8042, Medtronic, Inc., Minneapolis,Minnesota) together with a transvenous left ventricu-lar lead (Attain, Medtronic, Inc.) and conventionalright atrial and right ventricular leads. Before deviceimplantation, patients were assessed by NYHA func-tional class, the 6-minute walking test,12 and standardDoppler echocardiography. The walking test andassessment of NYHA class were repeated after 3months. Information on the incidence of high-rateatrial episodes or mode switch during atrial fibrillationwas stored in the device’s memory.
HRV and MDPA were measured from the day afterdevice implantation at midnight and stored as dailymeans in the device’s memory. HRV was analyzed asthe time-domain parameter SDANN, defined as the
From the Department of Cardiology, Karolinska University Hospital,Stockholm, Sweden; the Department of Cardiology, Skejby Hospital,Aarhus University, Aarhus, Denmark; the Department of Cardiology,Polyclinique Saint Henri, Nantes, France; Cardio Clinic, Nürnberg,Germany; the Department of Cardiology, Ruhr Hospital, BochumUniversity, Bochum, Germany; the Department of Cardiology, Cler-mont-Ferrand Hospital, Clermont-Ferrand, France; and the BakkenResearch Center, Medtronic, Maastricht, The Netherlands. This studywas supported by grants from the Swedish Heart and Lung Founda-tion, Stockholm, Sweden, and Medtronic, Inc., Minneapolis,Minnesota. Dr. Braunschweig’s address is: Department of Cardiology,Karolinska University Hospital, S-17176 Stockholm, Sweden. E-mail:[email protected]. Manuscript received November 3,2004; revised manuscript received and accepted December 28,2004.
SD of 5-minute medians of all normal atrial-to-atrialintervals. For this analysis, all data recorded duringatrial high-rate episodes or mode switch were ex-cluded from analysis. MDPA was measured as thetime per day (in minutes) with physical activitygreater than a threshold of 70 steps/min walking usingan accelerometer sensor.
Device-stored trend data on MDPA and HRV, col-lected before hospital discharge and at 4 and 12 weeksof follow-up, were saved to disk and entered into acentral database together with the clinical data.Weekly means of the MDPA and HRV were calcu-lated. Data obtained during the first week were used asa baseline reference. Because these measurementsmay be affected by postoperative hospitalization, fol-low-up data were also compared with the second weekafter implantation.
Statistics were calculated using standard commer-cial software (SPSS version 12.0, SPSS, Inc., Chi-cago, Illinois). Results are given as mean � SD orSEM. Repeated-measures analysis of variance wasused to compare trend effects between groups of dif-ferent NYHA classes. For the assessment of differ-ences in the 6-minute walking test, the nonparametricMann-Whitney U test was applied. For all methods, ap value of 0.05 was considered statistically significant.
Fifty-six patients with complete trend data for the first
12 weeks after pacemaker implanta-tion were included (Table 1). Nine pa-tients had previously been implantedwith pacemakers for conventional in-dications. Notably, only 43% were re-ceiving � blockers at baseline. How-ever, according to the study protocol,medication was kept stable throughoutthe study, with the exception of minorchanges in diuretics.
At baseline and after 3 months, 0,11, 37, and 8 and 18, 29, 8, and 1patient were in NYHA classes I, II,III, and IV, respectively (p �0.001).Accordingly, 43 of 56 patients (77%)had improved by �1 NYHA class.This improvement was observed inall groups of NYHA functional classas determined at baseline (Figure 1).
Walking distance increased from 361 � 104 m atbaseline to 429 � 135 m after 3 months (p �0.001).Figure 2 presents changes in walking distance accordingto NYHA class at baseline. Patients in NYHA classes IIand III improved significantly from 432 � 115 to 557 �97 m (p �0.05) and from 350 � 78 to 402 � 122 m (p�0.01), respectively. In patients in NYHA class IV, thewalking distance increased from 194 � 78 m at base-line to 327 � 107 m (p � NS) after 12 weeks.
MDPA was 108 � 81 minutes at baseline and168 � 115, 211 � 130, and 225 � 140 minutes duringweeks 2, 4, and 12, respectively (p �0.001 for all com-parisons with baseline). When compared with week 2,there was still a significant increase in MDPA atweeks 4 and 12 (p �0.001; Table 2). The MDPA trendover 3 months was significantly greater in patientswith better functional status at baseline (Figure 3).During the first 4 weeks of treatment, MDPA in-creased significantly in all patient groups, as definedby NYHA classification. Thereafter, MDPA furtherincreased gradually in patients in NYHA class II,whereas it remained stable in those in NYHA classesIII and IV. After 12 weeks of CRT, patients in NYHAclass IV had 49 and 153 minutes less daily activitycompared with those in NYHA classes III and II,respectively.
Two patients were excluded from HRV analysisbecause of frequent atrial high-rate episodes. SDANNwas 64 � 23 ms at baseline and 72 � 24, 72 � 26, and72 � 26 ms during weeks 2, 4, and 12, respectively(p �0.001 for all comparisons with baseline). Withinthe first 12 weeks of CRT, SDANN trends were sig-nificantly different among baseline NYHA classes(Figure 3). Although SDANN remained large in pa-tients in NYHA class II, it increased gradually overtime in those in class III, reaching almost the level ofpatients in NYHA class II after 12 weeks. Patients inNYHA class IV, however, showed small SDANNvalues at baseline, which remained unchanged over 12weeks of CRT.
• • •In the present study, MDPA and HRV trend data
analysis reflected the response to CRT itself. Beyond
FIGURE 1. Improvement in NYHA functional class after 3 months of CRT for sub-groups of NYHA class as determined at baseline (p <0.05, p <0.001, and p <0.05for changes in NYHA classes II, III, and IV, respectively).
FIGURE 2. Improvement in 6-minute walking distance after 3months (3M) of CRT for subgroups of NYHA class as determinedat baseline. *p <0.05; **p <0.01.
BRIEF REPORTS 1105
that, monitoring features incorporated in CRT devicescan add valuable prognostic information13 and maycontribute to improved patient management by pro-viding useful information on changes in disease se-
verity over time. Furthermore, thesefeatures have the potential to guidepharmacologic and other therapeuticinterventions in heart failure. Suchbenefit has recently been proposedfor implantable heart failure moni-tors with hemodynamic sensors.14,15
Thus, together with the extensive in-formation on arrhythmia events fromthe device’s memory, long-termphysiologic trend data may add im-portant diagnostic value to the ther-apeutic benefit by cardiac resynchro-nization. Concerning the rapidlyincreasing number of pacemakersand defibrillators implanted in pa-tients with heart failure, these novelmonitoring features offer a potentialfor wide clinical use, and their po-tential impact on heart failure man-agement should be further evaluatedin controlled studies.
Currently, guideline recommen-dations for CRT include only pa-tients in NYHA functional classes IIIand IV.16,17 In our study, patients inNYHA class II had a similar im-provement in functional class andwalking distance as shown for pa-tients in classes III and IV. Further-more, the improvement of MDPAwas greatest in patients in NYHAclass II. This finding suggests thateven patients with mild heart failuremay benefit from CRT. This hypoth-esis must be tested in a prospective,controlled trial.
Because of the protocol of themain trial, trend data could be com-pared neither with a period withoutactive CRT nor with a parallel con-trol group. Thus, postoperative dis-tress and immobilization probablyinfluenced trend data in the first daysafter implantation. However, theimplantation procedure is today con-
sidered “minor surgery,” with low indexes of pain anddiscomfort, immediate postoperative mobilization,and short hospital stay. Therefore, in agreement withother reports on short-term electromechanical and he-
FIGURE 3. Trend data analysis performed on a weekly basis on (A) MDPA (mean �SEM) and (B) SDANN (mean � SEM). Results are grouped by preimplantation assess-ment of disease severity (NYHA classification). *p <0.05; ***p <0.001 (repeated-measures analysis of variance) for differences between trends.
TABLE 2 Device-based Trend Data by NYHA Class at Baseline
*p �0.05 compared with the first week; †p �0.05 compared with the second week.
1106 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
modynamic effects of CRT,18,19 it is conceivable thatthe marked increase in MDPA and HRV could reflecta true treatment effect. The greater variability of trenddata for NYHA classes II and IV may be due to theirsmaller sample sizes compared with the NYHA classIII group.
Acknowledgment: We thank Medtronic Bakken Re-search Center and the Medtronic Heart Failure Man-agement Department for excellent data management.
APPENDIXThe InSync III Study Investigators: F. Hintringer (Innsbruck, Austria),
H. J. Nesser (Linz, Austria), B. Coutu (Notre-Dame, Quebec, Canada), A. D.Krahn (London, Ontario, Canada), R. A. Leather (Victoria, British Columbia,Canada), A. Tang (Ottawa, Ontario, Canada), R. Yee (London, Ontario, Canada),B. Hansky and J. Vogt (Bad Oeynhausen, Germany), B. Lemke (Bochum, Ger-many), W. Reiser (Nürnberg, Germany), C. Wolpert (Bonn, Germany), P. T.Mortensen (Aarhus, Denmark), D. Gascon (Sevilla, Spain), J. Leal (BellavistaSevilla, Spain), C. Barnay (Aix-en-Provence, France), J. C. Daubert (Rennes,France), D. Gras (Paris, France), H. Mansour (Clermont-Ferrand, France), A.Psapia (Marseille, France), P. E. Vardas (Crete, Greece), F. Gaita (Asti, Italy), M.Lunati (Milan, Italy), G. Perego (Milan, Italy), C. L. Verweij (Leiden, TheNetherlands), O. J. Ohm (Bergen, Norway), C. Morais (Amadora, Portugal),C. M. Linde (Stockholm, Sweden), T. Mabin (Somerset West, South Africa), andI. W. P. Obel (Johannesburg, South Africa).
1. Abraham WT, Fisher WG, Smith AL, Delurgio DB, Leon AR, Loh E, KocovicDZ, Packer M, Clavell AL, Hayes DL, et al. Cardiac resynchronization in chronicheart failure. N Engl J Med 2002;346:1845–1853.2. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De Marco T, CarsonP, DiCarlo L, DeMets D, White BG, et al. Cardiac-resynchronization therapywith or without an implantable defibrillator in advanced chronic heart failure.N Engl J Med 2004;350:2140–2150.3. Nolan J, Batin PD, Andrews R, Lindsay SJ, Brooksby P, Mullen M, Baig W,Flapan AD, Cowley A, Prescott RJ, et al. Prospective study of heart ratevariability and mortality in chronic heart failure: results of the United Kingdomheart failure evaluation and assessment of risk trial (UK-heart). Circulation1998;98:1510–1516.4. Bonaduce D, Petretta M, Marciano F, Vicario ML, Apicella C, Rao MA,Nicolai E, Volpe M. Independent and incremental prognostic value of heart ratevariability in patients with chronic heart failure. Am Heart J 1999;138:273–284.5. Alonso C, Ritter P, Leclercq C, Mabo P, Bailleul C, Daubert JC. Effects ofcardiac resynchronization therapy on heart rate variability in patients with chronic
systolic heart failure and intraventricular conduction delay. Am J Cardiol2003;91:1144–1147.6. Makikallio TH, Huikuri HV, Hintze U, Videbaek J, Mitrani RD, CastellanosA, Myerburg RJ, Moller M, DIAMOND Study Group. Fractal analysis and time-and frequency-domain measures of heart rate variability as predictors of mortalityin patients with heart failure. Am J Cardiol 2001;87:178–182.7. Kruger C, Lahm T, Zugck C, Kell R, Schellberg D, Schweizer MW, Kubler W,Haass M. Heart rate variability enhances the prognostic value of established param-eters in patients with congestive heart failure. Z Kardiol 2002;91:1003–1012.8. Francis GS, Rector TS. Maximal exercise tolerance as a therapeutic end pointin heart failure—are we relying on the right measure? Am J Cardiol 1994;73:304–306.9. Madsen BK, Hansen JF, Stokholm KH, Brons J, Husum D, Mortensen LS.Chronic congestive heart failure. Description and survival of 190 consecutivepatients with a diagnosis of chronic congestive heart failure based on clinicalsigns and symptoms. Eur Heart J 1994;15:303–310.10. Gottlieb SS, Fisher ML, Freudenberger R, Robinson S, Zietowski G, Alves L,Krichten C, Vaitkevicus P, McCarter R. Effects of exercise training on peakperformance and quality of life in congestive heart failure patients. J Card Fail1999;5:188–194.11. Walsh JT, Charlesworth A, Andrews R, Hawkins M, Cowley AJ. Relation ofdaily activity levels in patients with chronic heart failure to long-term prognosis.Am J Cardiol 1997;79:1364–1369.12. Lipkin DP, Scriven AJ, Crake T, Poole-Wilson PA. Six minute walking testfor assessing exercise capacity in chronic heart failure. BMJ 1986;292:653–655.13. Adamson PB, Smith AL, Abraham WT, Kleckner KJ, Stadler RW, Shih A,Rhodes MM. Continuous autonomic assessment in patients with symptomaticheart failure: prognostic value of heart rate variability measured by an implantedcardiac resynchronization device. Circulation 2004;110:2389–2394.14. Adamson PB, Magalski A, Braunschweig F, Bohm M, Reynolds D, SteinhausD, Luby A, Linde C, Ryden L, Cremers B, et al. Ongoing right ventricularhemodynamics in heart failure: clinical value of measurements derived from animplantable monitoring system. J Am Coll Cardiol 2003;41:565–571.15. Braunschweig F, Linde C, Eriksson MJ, Hofman-Bang C, Ryden L. Contin-uous haemodynamic monitoring during withdrawal of diuretics in patients withcongestive heart failure. Eur Heart J 2002;23:59–69.16. Gregoratos G, Abrams J, Epstein AE, Freedman RA, Hayes DL, Hlatky MA,Kerber RE, Naccarelli GV, Schoenfeld MH, Silka MJ, et al. ACC/AHA/NASPE2002 guideline update for implantation of cardiac pacemakers and antiarrhythmiadevices. J Am Coll Cardiol 2002;40:1703–1719.17. Remme WJ, Swedberg K. ESC guidelines for the diagnosis and treatment ofchronic heart failure. Eur Heart J 2001;22:1527–1560.18. Yu CM, Chau E, Sanderson JE, Fan K, Tang MO, Fung WH, Lin H, KongSL, Lam YM, Hill MR, et al. Tissue Doppler echocardiographic evidence ofreverse remodeling and improved synchronicity by simultaneously delayingregional contraction after biventricular pacing therapy in heart failure. Circula-tion 2002;105:438–445.19. Kass DA, Chen CH, Curry C, Talbot M, Berger R, Fetics B, Nevo E.Improved left ventricular mechanics from acute VDD pacing in patients withdilated cardiomyopathy and ventricular conduction delay. Circulation 1999;99:1567–1577.
BRIEF REPORTS 1107
Echocardiographic Examination of Atrioventricular andInterventricular Delay Optimization in Cardiac
Resynchronization Therapy
Maria Cristina Porciani, MD, Cristina Dondina, PhD, Roberto Macioce, MD,Gabriele Demarchi, MD, Paolo Pieragnoli, MD, Nicola Musilli, MD, Andrea Colella, MD,
Giuseppe Ricciardi, MD, Antonio Michelucci, MD, and Luigi Padeletti, MD
In 21 patients implanted with a biventricular pacingdevice with programmable interventricular delay(VVd), the myocardial performance index (MPI) wasevaluated during spontaneous sinus rhythm, simulta-neous biventricular pacing, and sequential biventricu-lar pacing at different VVds and atrioventricular de-lays (AVds). The AVd-VVd combination associatedwith the minimum MPI defined patient-tailored biven-tricular pacing. Simultaneous biventricular pacingsignificantly improved MPI compared with spontane-ous sinus rhythm. An additional improvement wasobtained by tailored biventricular pacing. The opti-mal AVds were significantly shorter during right ven-tricular preactivation than during left ventricularpreactivation. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1108–1110)
The present study was designed to evaluate the acuteeffect of individualized interventricular (VVd) and
atrioventricular (AVd) delays on left ventricular (LV)performance as measured by the myocardial perfor-mance index (MPI),1 which reflects systolic and dia-stolic function.2 This is a simple Doppler-derived in-dex based on the evaluation of different time intervals,and as such it is easy to assess in any echocardio-graphic laboratory, providing reliable results.
• • •Twenty-one patients who underwent implantation
between January 2004 and March 2004 were includedin the present study. All patients were in sinus rhythmand had left bundle branch block with a QRS duration�130 ms. Patients’ demographic and clinical vari-ables are listed in Table 1.
Three transvenous pacing leads were implanted: 1at the right atrial appendage (model 4480, GuidantCo., Indianapolis, Indiana), 1 at the right ventricularapex (model 0148, Guidant Co.), and the other in theposterolateral vein through the coronary sinus (models4512 and 4513, Guidant Co.). The leads were con-nected to a biventricular pacing device with defibril-lation backup (Contak Renewal II, model H155,Guidant Co.).
The Contak Renewal II model H155 heart failuredevice is designed to provide cardiac resynchroniza-
tion therapy by delivering electrical stimulation to theright and left ventricles, and to adjust the delay be-tween delivery of the LV and right ventricular (RV)pacing pulse. No major perioperative complicationsoccurred.
Within 24 hours after implantation, all patientsunderwent echo/Doppler evaluation using a commer-cially available imaging system (GE-Vingmed Ultra-sound, Horten, Norway).
The MPI was calculated as the sum of isovolumiccontraction and relaxation times divided by ejectiontime. Ejection time was evaluated as the interval be-tween the beginning and the end of aortic flow. Thesum of the isovolumic contraction and relaxationtimes was derived from the interval between the endof mitral inflow and the onset of the next mitral inflowsignal minus ejection time (Figure 1). The MPI hasalready been used to evaluate the effects of cardiacresynchronization therapy3 and to identify optimalAVd in patients with traditional pacemakers.4 Weused this index, which is comprehensive of systolicand diastolic function, as a single parameter for bothoptimizing AVd and VVd, to simplify the protocol ofoptimization and to encourage its execution in clinicalpractice.
In all patients, echo/Doppler evaluation was per-formed at baseline (spontaneous sinus rhythm), atsimultaneous biventricular pacing (VVd � 0 ms), andat different VVds with either RV preactivation (�20and �40 ms) or LV preactivation (�20 and �40 ms),and with different AVds (from 80 to 200 ms by stepsof 20 ms) programmed in a randomized sequence.
Thus, 35 AVd-VVd combinations were obtained.For each combination, a recording period was pre-ceded by a stabilization period of 3 minutes. In pre-vious invasive studies, a stabilization period of 5 or 20seconds was considered sufficient to achieve hemody-namic equilibrium.5–7 To minimize the variability ofthe measurements, all echo/Doppler evaluations wereperformed and analyzed by the same physician. Opti-mum AVds and VVds were identified by the mini-mum MPI and defined individually tailored biven-tricular pacing in each patient. The matrix in Table 2shows an example of our echocardiographic protocol.
Mean � SD values were computed to describecontinuous variables. Absolute frequencies were cal-culated for categorical variables. A p value �0.05 wasretained for statistical significance. The Bonferronicorrection for multiple tests was used for post hoccomparisons.
From the Division of Cardiology, Careggi Hospital, Florence; andGuidant Italia, Milan, Italy. Dr. Porciani’s address is: Division ofCardiology, Careggi Hospital, Viale Morgagni 85, 50134 Florence,Italy. E-mail: [email protected]. Manuscript received August 12,2004; revised manuscript received and accepted January 5, 2005.
Both simultaneous and tailored biventricular pac-ing were associated with a significant decrease in theMPI compared with those at baseline (from 1.09 �0.41 to 0.78 � 0.19 [p � 0.0004 for simultaneousbiventricular pacing] and to 0.60 � 0.14 [p � 0.0001for tailored biventricular pacing]); a significant differ-ence was also found between the simultaneous andtailored biventricular pacing MPIs (p � 0.04; Figure 2).In the tailored biventricular pacing group, the optimalpacing configuration was LV preactivation in 10 pa-tients (48%), RV preactivation in 10 patients (48%),and simultaneous biventricular pacing in 1 patient (4%).
Optimal LV preactivation was 20 ms in 7 patientsand 40 ms in 3 patients. Optimal RV preactivation was20 ms in 8 patients and 40 ms in 2 patients. The meanMPI in RV preactivated patients was 0.66 � 0.12(range 0.45 to 0.87) and was significantly lower(p � 0.00008) than the corresponding value during
simultaneous biventricular pacing (0.85 � 0.12, range0.53 to 1.01). The mean percent reduction in the MPIduring RV preactivation was 22.3%. The mean MPI inLV preactivated patients was 0.54 � 0.15 (range 0.28to 0.79) and was significantly lower (p �0.005) thanthe corresponding value during simultaneous biven-tricular pacing (0.74 � 0.23, range 0.37 to 1.18). Themean percent reduction in the MPI during LV preac-tivation was 37%.
A significant difference was observed in the MPIbetween optimal RV and LV preactivated patients(p � 0.02). The MPI obtained during simultaneousand tailored biventricular pacing was similar in only 1patient; the AVd-VVd combination providing thehighest cardiac output was programmed.
In considering the underlying disease, optimal RVpreactivation was observed in 7 patients with ischemiccardiomyopathy and in 3 patients with dilated cardio-myopathy, whereas optimal LV preactivation was ob-served in 5 patients with ischemic cardiomyopathy, in4 patients with dilated cardiomyopathy, and in 1 pa-tient with valve disease (Figure 3).
Optimal AVd was 97.14 � 27 ms (range 80 to 160)during simultaneous biventricular pacing and 114 �25.2 ms (range 80 to 160) during tailored biventricularpacing . The optimal AVd was 126 � 25 ms (range 80to 160) in the 10 patients in whom the left ventriclewas preactivated, and 102 � 19.9 ms (range 80 to140) in the 10 patients in whom the right ventricle waspreactivated (p � 0.02).
• • •The major findings of our study were (1) an im-
mediate beneficial effect of simultaneous biventricularpacing on LV function, as observed in previous he-modynamic and echocardiographic investigations,5,7–9
and (2) a significant additional benefit to the effects ofcardiac resynchronization therapy on LV performanceinduced by tailored biventricular pacing.
Despite a similar QRS morphology, patients withheart failure and left bundle branch block may presentwith a different location of ventricular asynchrony,which appears to be related to the underlying etiology.In a recent study,10 it has been demonstrated that inpatients affected with heart failure, the time needed toachieve global depolarization of the left ventricle dur-ing LV pacing is prolonged due to the presence ofslow conduction areas. Thus, the depolarization initi-ated by LV stimulus is delayed. This delay differsfrom 1 patient to another, and different LV preactiva-tion values are needed to obtain an optimal coalitionof the LV and RV wave front activation. VVd optimi-zation is crucial to restore intraventricular synchrony.
In a previous study, Sogaard et al8 demonstratedthat in patients with idiopathic dilated cardiomyopa-thy, the myocardium with delayed contraction tendedto be located in the lateral and posterior LV wall.However, in patients with ischemic cardiomyopathy,these delayed areas were more frequent in the septum.This explains why LV preactivation produced the bestsystolic performance in patients with idiopathic di-lated cardiomyopathy, whereas the latter was obtainedwith RV preactivation in patients with ischemic car-
FIGURE 1. Evaluation of MPI by mitral and aortic pulsed echo/Doppler. The sum of the isovolumic contraction (x) and relax-ation (y) times is derived from the interval between the end ofmitral inflow and the onset of the next mitral inflow signal (a)minus ejection time (b). Ejection time is calculated as the intervalbetween the beginning and the end of aortic flow.
TABLE 1 Demographics and Clinical and EchocardiographicVariables (n � 21)
Demographics and clinical variablesAge (yrs) 69.7 � 12.7Men/women 14/7IHD/non-IHD 13/8QRS duration (ms) 160 � 25PR interval (ms) 200 � 20NYHA class III/IV 78%/22%ACE inhibitors 16 (82%)Diuretics 17 (80%)� blockers 16 (78%)Spironolactone 3 (15%)Digoxin 8 (36%)Quality of life (Minnesota Questionnaire) 28.2 � 18.86-min walk test (m) 310 � 169.7
Values are expressed as mean � SD or percentage.ACE � angiotensin-converting enzyme; IHD � ischemic heart disease;
NYHA � New York Heart Association.
BRIEF REPORTS 1109
diomyopathy. These findings were not confirmed byPerego et al,5 who reported that LV preactivation wasoptimal in an entire group of 12 patients (5 of whomwere affected by ischemic cardiomyopathy). Using aninvasive optimization protocol, the investigators dem-onstrated that AVd optimization and tailored biven-tricular pacing increased the effects of biventricularpacing on LV systolic and diastolic mechanical effi-
ciency. In a recent study, Van Gelder et al6 reportedthat LV preactivation was the best pacing configura-tion in 44 of 53 patients, regardless of the underlyingcardiac disease.
Our data and all these data are related to optimalLV stimulation (i.e., lateral or posterolateral); dataconcerning different sites of pacing are lacking.
Another finding of our study was that optimal AVdwas significantly shorter during RV preactivation thanLV preactivation. The delay in LV contraction inducedby RV preactivation prolongs the interval between LVand left atrial contraction; this explains why shorterAVds are required to optimize LV performance. Due tothis interference between VVd and optimal AVd, testingdifferent AVds for each VVd is required.
Despite the time-consuming aspects of this proto-col, its execution should be recommended to maxi-mize the beneficial effects of cardiac resynchroniza-tion therapy on LV performance.
1. Tei C. New noninvasive index for combined systolic and diastolic function.J Cardiol 1995;26:135–136.2. Tei C, Nishimura RA, Seward J, Jamil Tajik A. Noninvasive Doppler-derivedmyocardial performance index: correlation with simultaneous measurements ofcardiac catheterization measurements. J Am Soc Echocardiogr 1997;10:169–178.3. Yu C-M, Fung W-H, Lin H, Zhang Q, Sanderson JE, Lau C-P. Predictors ofleft ventricular reverse remodeling after cardiac resynchronization therapy forheart failure secondary to idiopathic dilated or ischemic cardiomyopathy. Am JCardiol 2002;91:684–688.4. Porciani MC, Corbucci G, Fantini F, Musilli N, Sabini A, Michelucci A,Coolella A, Pieragnoli P, Demarchi G, Padeletti L. A perspective on atrioven-tricular delay optimization in patients with a dual chamber pacemaker. PacingClin Electrophysiol 2004;27:333–338.5. Perego GB, Chianca R, Facchini M, Frattola A, Balla E, Zucchi S, Cavaglià S,Vicini I, Negretto M, Osculati G. Simultaneous vs. sequential biventricularpacing in dilated cardiomyopathy: an acute hemodynamic study. Eur J Heart Fail2003;5:305–313.6. Van Gelder B, Bracke F, Meijer A, Lakerveld L, Pijls N. Effect of optimizingthe VV interval on left ventricular contractility in cardiac resynchronizationtherapy. Am J Cardiol 2004;93:1500–1503.7. Auricchio A, Stellbrink C, Block M, Sack S, Vogt J, Bakker P, Klein H, for thePacing Therapies for Congestive Heart Failure Study Group, and Kramer A, DingJ, Salo R, Tockman B, Pochet T, Spinelli J, for the Guidant Congestive HeartFailure Research Group. Effects on pacing chamber and atrioventricular delay onacute systolic function of paced patients with congestive heart failure. Circulation1999;99:2993–3001.8. Sogaard P, Egeblad H, Pedersen A, Kim WY, Kristensen B, Hansen P,Mortensen P. Sequential versus simultaneous biventricular resynchronization forsevere heart failure. Evaluation by tissue Doppler imaging. Circulation 2002;106:2078–2084.9. Kass DA, Chen CH, Curry C, Talbot M, Berger R, Fetics B, Nevo E. Improvedleft ventricular mechanism from acute VDD pacing in patients with dilated cardio-myopathy and ventricular conduction delay. Circulation 1999;99:1567–1573.10. Lambiase PD, Rinaldi A, Hauck J, Mobb M, Elliott D, Mohammad S, Gill JS,Bucknall CA. Non-contact left ventricular endocardial mapping in cardiac resyn-chronization therapy. Heart 2004;90:44–51.
TABLE 2 Matrix of 65-year-old Woman With Dilated Cardiomyopathy, Defining MPI Values at Different Combinations ofAtrioventricular Delays (AVds) and Interventricular Delays (VVds) Tested During Echo Cardiographic Evaluation*
AVd (ms) VVd � �40 ms VVd � �20 ms VVd � 0 ms VVd � �20 ms VVd � �40 ms
*VVd was considered negative when LV stimulation preceded RV stimulation.†Optimal MPI was obtained at a VVd of 20 ms and an AVd of 80 ms.
FIGURE 2. MPI values at baseline (spontaneous sinus rhythm), dur-ing simultaneous biventricular pacing, and tailored biventricularpacing in 21 patients. **p � 0.0004, simultaneous biventricularpacing compared with baseline; *p � 0.0001, tailored biventricularpacing compared with baseline; §p � 0.04, tailored compared withsimultaneous biventricular pacing.
FIGURE 3. Distribution of the patient group for optimal preactivation.DCM � dilated cardiomyopathy; IHD � ischemic heart disease.
1110 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Effect of Cardiac Resynchronization Therapy onInducibility of Ventricular Tachyarrhythmias in Cardiac
Arrest Survivors With Either Ischemic or IdiopathicDilated Cardiomyopathy
Philippine Kiès, MD, Jeroen J. Bax, MD, PhD, Sander G. Molhoek, MD,Gabe B. Bleeker, MD, Katja Zeppenfeld, MD, PhD, Marianne Bootsma, MD, PhD,Lieselot van Erven, MD, PhD, Paul Steendijk, PhD, Ernst E. van der Wall, MD, PhD,
and Martin J. Schalij, MD, PhD
We evaluated whether long-term cardiac resynchro-nization therapy affects the inducibility of ventriculartachyarrhythmias in relation to reverse remodeling incardiac arrest survivors with either ischemic or idio-pathic dilated cardiomyopathy. Clinical, electrophysi-ologic, and echocardiographic data of 18 patientswere obtained before and after 6 months of cardiacresynchronization. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1111–1114)
Agrowing number of randomized clinical trials hasestablished the beneficial effects of cardiac re-
synchronization therapy (CRT) on symptoms, exer-cise capacity, left ventricular (LV) systolic perfor-mance, and LV reverse remodeling.1–3 In addition,several reports have suggested that CRT may alsoreduce the frequency of ventricular ectopy and appro-priate implantable cardioverter-defibrillator (ICD)therapy.4–7 Preliminary data have also shown a reduc-tion in the prevalence of ventricular tachyarrhythmicepisodes in patients treated with CRT.8 The reductionin ventricular arrhythmias was related to reverse re-modeling after CRT, suggesting that reduced walltension may lead to a reduction in ventricular arrhyth-mias. To further substantiate this hypothesis, we eval-uated the inducibility of ventricular tachyarrhythmiasduring an electrophysiologic study after 6 months ofCRT in cardiac arrest survivors treated with a com-bined CRT-ICD device and its relation to reverseremodeling as a result of CRT.
vors with advanced heart failure who were implantedwith a CRT-ICD device were evaluated. Implantationof an ICD in these patients was based on standardtherapy guidelines.9 Eligibility for CRT was based onthe following criteria: (1) advanced heart failure withNew York Heart Association functional class III orIV, (2) LV ejection fraction �35%, and (3) wide QRS
complex (�120 ms) with a left bundle branch blockpattern on the electrocardiogram. Patients with isch-emic or nonischemic cardiomyopathy were evaluated.Patients with atrial fibrillation were not excluded. Sixmonths after a baseline electrophysiologic study andthe implantation of a CRT-ICD device, the inducibil-ity of ventricular tachycardia/ventricular fibrillation(VT/VF) was evaluated as part of standard clinicalpractice.
First, a coronary sinus venogram was obtainedusing the balloon catheter. Next, the LV pacing lead(Easytrack 4512 to 80, Guidant, Minneapolis, Minne-sota, or Attain-SD model 4189, Medtronic Inc., St.Paul, Minnesota) was inserted transvenously via thesubclavian route with the help of an 8Fr guidingcatheter, and positioned in a posterolateral vein.10 Theright atrial and 2 ventricular leads were positionedconventionally. All leads were connected to a dual-chamber biventricular ICD (Contak CD or ContakRenewal, Guidant, or InSync CD, Medtronic Inc.).
Three quadripolar catheters were inserted trans-venously and positioned in the high right atrium, at theHis bundle, and in the right ventricular (RV) apex.Pacing was performed at twice the pacing threshold.To induce VT/VF, programmed electrical stimulationwas performed from the RV apex and subsequentlyfrom the RV outflow tract according to a standardizedprotocol, applying up to 3 extrastimuli after 8-beatdrive trains at 3 different basic cycle lengths (600,500, and 400 ms). The extrastimuli were decrementedwith 10 ms until refractoriness or the shortest couplinginterval (200 ms) was reached. Inducibility of VT/VFwas defined as the induction of sustained VT (�30seconds) and/or VF.
At 6 months, a second electrophysiologic studywas performed according to the same protocol butthrough the ICD (via the RV lead). Consequently,stimulation through the RV outflow tract could not beperformed. To compare the results of both electro-physiologic studies, we only considered the results ofstimulation from the RV apex of the first electrophysi-ologic study. Accordingly, 1 patient who had onlybeen inducible with stimulation in the RV outflowtract during the first study was excluded.
Patients were imaged in the left lateral decubitusposition using a commercially available system(Vingmed system FiVe/Vivid 7, General Electric-
From the Department of Cardiology, Leiden University Medical Center,Leiden, The Netherlands. Drs. Molhoek and Bleeker were supported byGrants 2001D015 and 2002B109, respectively, from the Dutch HeartFoundation, The Hague, The Netherlands. Dr. Bax’s address is: Depart-ment of Cardiology, Leiden University Medical Center, Albinusdreef 2,2333 ZA Leiden, The Netherlands. E-mail: [email protected]. Manuscript received November 8, 2004; revised manuscriptreceived and accepted December 30, 2004.
Vingmed, Milwaukee, Wisconsin). Images were ob-tained using a 3.5-MHz transducer at a depth of 16 cmin the parasternal and apical views (standard long-axisand 2- and 4-chamber images). Standard 2-dimen-sional and color Doppler data, triggered to the QRScomplex, were saved in cineloop format. LV ejectionfraction and LV volumes were derived from the con-ventional apical 2- and 4-chamber images using thebiplane Simpson’s rule.11 The severity of mitral re-gurgitation was graded semiquantitatively from colorflow Doppler in the conventional parasternal long-axisand apical 4-chamber images. Mitral regurgitationwas characterized as mild � 1� (jet area/left atrialarea �10%), moderate � 2� (jet area/left atrial area10% to 20%), moderately severe � 3� (jet area/leftatrial area 20% to 45%), and severe � 4� (jet area/leftatrial area �45%).12 All data (baseline, follow-up)were obtained by 2 independent observers blinded toall other data.
Patients were evaluated clinically at baseline andafter 6 months of CRT. Heart failure symptoms wereclassified according to the New York Heart Associa-tion score. Quality-of-life score was assessed usingthe Minnesota Living with Heart Failure question-naire.13 Exercise capacity was evaluated using the6-minute hall walk test.14 A surface electrocardiogramwas obtained (12 leads at a study speed of 50 mm/s) toestablish QRS morphology and duration before andafter 6 months of CRT.
Data are expressed as mean � SD. Comparison ofdata was performed using the Student‘s t test forpaired and unpaired data when appropriate, whereasthe Wilcoxon signed rank test was performed in caseof markedly skewed distribution of the data. McNemar’stest for paired dichotomous data was performed tocompare inducibility before and after 6 months ofCRT. For all tests, a p value �0.05 was consideredsignificant. Patients who were part of this study werenot included in other published studies.
We analyzed 18 patients (15 men and 3 women)after the exclusion of 1 patient. The mean age of thestudy group at the time of implantation of the CRT-ICD device was 62 � 11 years (range 38 to 81).Etiology of underlying heart failure was ischemiccardiomyopathy in 10 patients and idiopathic dilatedcardiomyopathy in 8 patients. All patients had beenreferred with aborted sudden death, 6 deaths due toVF and 12 due to a fast sustained VT (average cyclelength of VT, 244 � 22 ms) with hemodynamiccollapse. All patients received optimized (if tolerated)medical therapy, including diuretics and oral antico-agulants in all, angiotensin-converting enzyme inhib-itors in 83%, � blockers in 66%, and spironolactone in50% of patients. In addition, amiodarone was admin-istered to 13 patients (72%) and sotalol to 4 patients(22%). Medication remained unchanged during thestudy.
At baseline, all patients were in New York HeartAssociation class III (n � 18), with an average QRSduration of 185 � 29 ms. The electrocardiogramshowed a left bundle branch block pattern in all 18patients. After 6 months of CRT, the New York Heart
Association class decreased from 3.0 � 0 to 2.1 � 0.8(p �0.01). The quality-of-life score decreased from41 � 12 to 24 � 13 (p �0.01). In addition, exercisecapacity improved, reflected by a significant improve-ment in the 6-minute hall walk test from 321 � 73 to444 � 136 m (p �0.01) after 6 months of CRT(Figure 1).
Results of the first electrophysiologic study aredisplayed in Table 1. In 15 of 18 patients (83%), asustained VT was inducible. Three patients had noinducible VT/VF. After an average of 7.1 � 0.8months of CRT, the second electrophysiologic studywas performed (Table 1). As opposed to the firststudy, a sustained VT was inducible in only 6 patients(33%, p �0.01). Thus, in 9 of 15 patients (60%),ventricular arrhythmias were no longer inducible atthe second study. Of note, no patient experiencedincreased inducibility; the 3 patients in whom induc-ibility could not be achieved remained noninducibleduring the second electrophysiologic study. Hemody-namics (heart rate and blood pressure) were compa-rable during both electrophysiologic studies. This wasalso the case during both echocardiograhic exami-nations.
LV ejection fraction at baseline was 19 � 5%. Allpatients had severely dilated left ventricles as evi-denced by LV end-diastolic and end-systolic volumes
FIGURE 1. Clinical evaluation before (left bars) and after (rightbars) 6 months of CRT . (A) New York Heart Association (NYHA)classification, (B) quality-of-life (QOL) score, and (C) the 6-minutehall walk test (6-MWT).
1112 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
of 274 � 61 and 224 � 62 ml, respectively. After 6months of CRT, reverse remodeling was observed.End-diastolic volume decreased to 254 � 69 ml (p�0.05) and end-systolic volume decreased to 198 �66 ml (p �0.01). However, close examination of thedata demonstrated that this decrease in volumes issolely attributable to the patient group that had loss ofinducibility during the second electrophysiologicstudy (Table 2). In addition, before CRT, 4 patientshad severe (grade 3�), 10 patients had moderate(grade 2�), and 4 patients had mild (grade 1�) mitralregurgitation. After 6 months of CRT, in 9 patients(50%), the degree of mitral regurgitation had im-proved by �1 grade.
• • •The main finding of our study was that the induc-
ibility of VT/VF was decreased in patients with sig-nificant LV reverse remodeling after 6 months oftreatment with a CRT-ICD device.
We observed a striking difference between patientswith loss of inducibility and patients who remainedinducible in terms of LV remodeling. Patients withloss of inducibility during CRT had a significant re-
duction in LV end-diastolic volume,from 293 � 67 to 258 � 69 ml (p�0.01), whereas LV end-diastolicvolumes in patients with continuinginducibility did not change (from258 � 30 to 252 � 41 ml, p � NS).Similar results were observed forend-systolic volumes. In patientswith loss of inducibility, LV end-systolic volumes were decreasedfrom 244 � 69 to 198 � 69 ml (p�0.01). Patients who remained in-ducible had almost similar end-sys-
tolic volumes before and after 6 months of CRT (208� 27 and 199 � 35 ml, p � NS).
St. John Sutton et al15 recently demonstrated thatongoing LV remodeling resulted in a higher incidenceof ventricular arrhythmias due to increased LV wallstress and its consequences on LV structure and com-position. As a result, anisotropic reentry, an importantmechanism underlying ventricular arrhythmias, maybe facilitated because this originates from areas ofslowed impulse propagation and unidirectional con-duction block.16 This process may be amplified byelevated plasma levels of adrenergic neurohormonesin patients with end-stage heart failure.17,18 Con-versely, reverse remodeling, as observed after pro-longed CRT, may reduce wall stress and invert thisprocess.19 In addition, CRT may lead to decreasedplasma-norepinephrine levels.20 Reduction in wallstress may therefore be explanatory for the decreasedprevalence of ventricular arrhythmias, such as werecently demonstrated in 18 patients after 12 monthsof CRT.8 The loss of inducibility of ventricular tachy-arrhythmias observed in this study may also be ex-plained by reduced wall stress.
TABLE 1 Results of Electrophysiologic Studies Both Before (study 1) and After (study 2) Six Months of Resynchronization Therapy
1. Abraham WT, Fisher WG, Smith AL, Delurgio DB, Leon AR, Loh E, KocovicDZ, Packer M, Clavell AL, Hayes DL, et al. Cardiac resynchronization in chronicheart failure. N Engl J Med 2002;346:1845–1853.2. St John Sutton MG, Plappert T, Abraham WT, Smith AL, Delurgio DB, LeonAR, Loh E, Kocovic DZ, Fisher WG, Ellestad M, et al. Effect of cardiacresynchronization therapy on left ventricular size and function in chronic heartfailure. Circulation 2003;107:1985–1990.3. Yu CM, Chau E, Sanderson JE, Fan K, Tang MO, Fung WH, Lin H, Kong SL,Lam YM, Hill MR, Lau CP. Tissue Doppler echocardiographic evidence ofreverse remodeling and improved synchronicity by simultaneously delayingregional contraction after biventricular pacing therapy in heart failure. Circula-tion 2002;105:438–445.4. Higgins SL, Yong P, Sheck D, McDaniel M, Bollinger F, Vadecha M, DesaiS, Meyer DB. Biventricular pacing diminishes the need for implantable cardio-verter defibrillator therapy. Ventak CHF Investigators. J Am Coll Cardiol 2000;36:824–827.5. Martinelli FM, Pedrosa AA, Costa R, Nishioka SA, Siqueira SF, Tamaki WT,Sosa E. Biventricular pacing improves clinical behavior and reduces prevalenceof ventricular arrhythmia in patients with heart failure. Arq Bras Cardiol 2002;78:110–113.6. Walker S, Levy TM, Rex S, Brant S, Allen J, Ilsley CJ, Paul VE. Usefulnessof suppression of ventricular arrhythmia by biventricular pacing in severe con-gestive cardiac failure. Am J Cardiol 2000;86:231–233.7. Zagrodzky JD, Ramaswamy K, Page RL, Joglar JA, Sheehan CJ, Smith ML,Hamdan MH. Biventricular pacing decreases the inducibility of ventricular tachy-cardia in patients with ischemic cardiomyopathy. Am J Cardiol 2001;87:1208–1210.8. Kies P, Bax JJ, Molhoek SG, Bleeker GB, Zeppenfeld K, Bootsma M, St JohnSutton M, van Erven L, van der Wall EE, Schalij MJ. Effect of left ventricularremodeling after cardiac resynchronization therapy on frequency of ventriculararrhythmias. Am J Cardiol 2004;94:130–132.9. Gregoratos G, Abrams J, Epstein AE, Freedman RA, Hayes DL, Hlatky MA,Kerber RE, Naccarelli GV, Schoenveld MH, Silka MJ, et al. ACC/AHA/NASPE2002 guideline update for implantation of cardiac pacemakers and antiarrhythmiadevices: summary article. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/
NASPE Committee to Update the 1998 Pacemaker Guidelines). J CardiovascElectrophysiol 2002;13:1183–1199.10. Butter C, Auricchio A, Stellbrink C, Fleck E, Ding J, Yu Y, Huvelle E,Spinelli J. Effect of resynchronization therapy stimulation site on the systolicfunction of heart failure patients. Circulation 2001;104:3026–3029.11. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, FeigenbaumH, Gutgesell H, Reichek N, Sahn D, Schnittger I. Recommendations for quanti-tation of the left ventricle by two-dimensional echocardiography. AmericanSociety of Echocardiography Committee on Standards, Subcommittee on Quan-titation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr 1989;2:358–367.12. Heinle SK, Grayburn PA. Doppler echocardiographic assessment of mitralregurgitation. Coron Artery Dis 2000;11:11–18.13. Rector TS, Kubo SH, Cohn JN. Validity of the Minnesota Living with HeartFailure questionnaire as a measure of therapeutic response to enalapril or placebo.Am J Cardiol 1993;71:1106–1107.14. Lipkin DP, Scriven AJ, Crake T, Poole-Wilson PA. Six minute walking testfor assessing exercise capacity in chronic heart failure. BMJ 1986;292:653–655.15. St John Sutton SM, Lee D, Rouleau JL, Goldman S, Plappert T, BraunwaldE, Pfeffer MA. Left ventricular remodeling and ventricular arrhythmias aftermyocardial infarction. Circulation 2003;107:2577–2582.16. de Bakker JM, van Capelle FJ, Janse MJ, Wilde AA, Coronel R, Becker AE,Dingemans KP, van Hemel NM, Hauer RN. Reentry as a cause of ventriculartachycardia in patients with chronic ischemic heart disease: electrophysiologicand anatomic correlation. Circulation 1988;77:589–606.17. Kjaer A, Hesse B. Heart failure and neuroendocrine activation: diagnostic,prognostic and therapeutic perspectives. Clin Physiol 2001;21:661–672.18. Aronson D, Burger AJ. Neurohumoral activation and ventricular arrhythmiasin patients with decompensated congestive heart failure: role of endothelin.Pacing Clin Electrophysiol 2003;26:703–710.19. Nguyen T, El Salibi E, Rouleau JL. Postinfarction survival and inducibility ofventricular arrhythmias in the spontaneously hypertensive rat: effects of ramipriland hydralazine. Circulation 1998;98:2074–2080.20. Hamdan MH, Zagrodzky JD, Joglar JA, Sheehan CJ, Ramaswamy K, ErdnerJF, Page RL, Smith ML. Biventricular pacing decreases sympathetic activitycompared with right ventricular pacing in patients with depressed ejection frac-tion. Circulation 2000;102:1027–1032.
Long-Term Effects of Carvedilol or Metoprolol onLeft Ventricular Function in Ischemic and
Nonischemic Cardiomyopathy
Philip Green, MS, Michael Anshelevich, MD, Ashok Talreja, MD, Joyce L. Burcham, PhD,Srinivas M. Ravi, MD, Jamshid Shirani, MD, and Thierry H. Le Jemtel, MD
Data regarding the effects of � blockers on left ven-tricular (LV) function after 12 months are scarce inischemic and nonischemic cardiomyopathy. Echocar-diograms of 72 patients with ischemic and nonisch-emic cardiomyopathy, who were free of clinicalevents susceptible to alter LV function while receivingcarvedilol or metoprolol for at least 24 months, wereprospectively reanalyzed. Twelve months after�-blocker initiation, LV ejection fraction (EF) increasedby >5% in 75% of patients, whereas EF failed toincrease by 5% or decreased in the remaining 25%.Over the subsequent 32 months, LVEF increased fur-ther in patients who had experienced an initial EFincrease by >5%, whereas EF tended to further de-
crease in patients who had experienced an initial EFincrease of <5% or a decrease. Thus, the benefits ofcarvedilol or metoprolol on LV function are long last-ing in patients with ischemic or nonischemic cardio-myopathy who are free of events susceptible to alterLV function while receiving � blockade. �2005 byExcerpta Medica Inc.
(Am J Cardiol 2005;95:1114–1116)
The present study was undertaken to prospectivelyreanalyze serial 2-dimensional echocardiograms
collected from patients with ischemic and nonisch-emic cardiomyopathy who received �-blocker therapyfor �24 months and did not experience events thatcould have interfered with the effect of � blockade onleft ventricular (LV) function.
• • •Among the 648 patients with congestive heart fail-
ure followed at the Montefiore Medical Center heartfailure service, 72 were eligible for the study based onthe following criteria: (1) they had received �-blockertherapy for �24 months; (2) they had undergone high-
From the Department of Medicine, Albert Einstein College of Medi-cine, Bronx; and Our Lady of Mercy Medical Center, Bronx, NewYork. Dr. Le Jemtel’s address is: Albert Einstein College of Medicine,Room G-46, Forchheimer Building, 1300 Morris Park Avenue, Bronx,New York 10461. E-mail: [email protected]. Manuscript re-ceived September 14, 2004; revised manuscript received and ac-cepted January 4, 2005.
quality, yearly 2-dimensional and Doppler echocardio-graphic examination after the initiation of �-blocker ther-apy; and (3) they had been free of cardiac events thatcould have altered LV function, such as myocardialinfarction, coronary revascularization, valvular repairor replacement, and cardiac resynchronization ther-apy, and did not demonstrate poor adherence to amedical regimen and diet. Echocardiographic exami-nations of eligible patients were prospectively reana-lyzed by an echocardiologist who was blinded topatient identification and time of the study. All pa-tients had a LV ejection fraction (EF) �40% beforeinitiation of �-blocker therapy. Fifty-six patients(78%) were receiving angiotensin-converting enzymeinhibitors, and the remaining 16 (22%) were receivingangiotensin receptor blockers. Sixty patients (83%)were receiving loop diuretics and 46 (64%) werereceiving digoxin at a maximal daily dose of 0.125
mg. Fifty patients (69%) were re-ceiving warfarin for the preventionof systemic embolism in the contextof atrial fibrillation (10 patients,14%) or, when in sinus rhythm, amarkedly dilated LV cavity (40 pa-tients, 55%). At the start of�-blocker therapy, 2 patients werereceiving aldosterone receptor block-ers, which increased to 37 patients(51%) at the time of the last echocar-diographic examination. Aldosteronereceptor blockade was initiated afterthe first year of follow-up. The studywas approved by the institutionalcommittee on clinical investigation.
All echocardiographic examina-tions were performed by 2 techni-cians from the Montefiore MedicalCenter cardiac noninvasive labora-tory. Transthoracic echocardiogra-phy was performed using a Sonos5500 system (Philips Medical Sys-tems, Andover, Massachusetts) withan S3 transducer. Harmonic 2-di-mensional and M-mode images wereobtained using standard views. Im-ages were recorded on 0.5-in VHSvideotapes and analyzed offline.End-diastolic and end-systolic LVdimensions and left atrial size weremeasured using 2-dimensional–guided M-mode recordings accord-ing to standard guidelines.1 LV vol-umes and EF were obtained from theapical 4-chamber view by manuallytracing the endocardial border inend-diastole and end-systole and us-ing the disk summation method.2Three consecutive representative cy-cles were measured and the averagevalues were used for analysis. Echo-cardiographic examinations were ob-tained immediately before initiation
of �-blocker therapy, at 1 year (�1 month) and yearly(�2 months) thereafter.
The response to �-blocker therapy was measured bychanges in LVEF and LV end-diastolic dimension.Based on LVEF changes at 12 months, relative to LVEFat the first visit, patients were grouped into respondersand nonresponders. Patients who experienced a LVEFincrease of �5% at 12 months were defined as respond-ers, and those who experienced a LVEF increase of�5% or a decrease at 12 months were defined as non-responders. Using best-fit linear regression analysis, theslopes (rates of change) of LVEF and LV end-diastolicdimension were used to compare the long-term responseto �-blocker therapy between groups and between sub-groups based on various medications.
The study group included 40 men and 32 women.Their mean age was 60 years (range 27 to 86). Thirty-eight patients received carvedilol at a mean daily dose of
FIGURE 1. Absolute change in LVEF (percent) during long-term �-adrenergic block-ade. Solid line, changes in responders; broken line, changes in nonresponders.
FIGURE 2. Change in LV end-diastolic dimension (millimeters) during long-term �-adren-ergic blockade. Solid line, changes in responders; broken line, changes in nonresponders.
BRIEF REPORTS 1115
37.5 mg, 21 patients received metoprolol succinate at amean daily dose of 150.0 mg, and the remaining 13patients received metoprolol tartrate at a mean daily doseof 155.5 mg. Forty patients had ischemic cardiomyopa-thy and 32 had nonischemic cardiomyopathy.
LVEF increased by �5% during the first 12 monthsof therapy in 54 patients (75%). Mean LVEF in these 54patients increased from 28 � 7% to 41 � 9% (p�0.001), whereas mean LV end-diastolic dimension de-creased from 62 � 8 to 56 � 8 mm (p � 0.001). Overthe subsequent 32 months of therapy, LVEF furtherincreased to 46 � 12% (p � 0.02), whereas LV end-diastolic dimensions decreased to 55 � 8 mm (p � 0.34)(Figures 1 and 2). Among responders at 44 months, LVsize and function were not affected by the initiation ofaldosterone receptor blockade. LVEF was similar in pa-tients receiving aldosterone blockade and in those whodid not (47% and 45%, respectively). Similarly, LVend-diastolic dimension was similar in patients receivingaldosterone blockade and in those who did not (56 and54 mm, respectively).
LVEF failed to increase by 5% during the first 12months of therapy in 18 patients (25%). Mean LVEFin these 18 patients decreased from 32 � 7% to 29 �6% (p � 0.21), whereas LV end-diastolic dimensionincreased from 63.5 � 8.8 to 64.6 � 8.6 mm (p �0.71). During the subsequent 32 months of therapy,LVEF further decreased to 28 � 8% (p � 0.71), andLV end-diastolic dimension increased to 65.3 � 8.2mm (p � 0.81) (Figures 1 and 2).
The percentage of responders was similar in pa-tients receiving carvedilol and metoprolol (74% and77%, respectively). The percentage of responders wasalso similar in patients with ischemic and nonischemiccardiomyopathy (72% and 78%, respectively). Theduration of congestive heart failure tended to beshorter in responders than in nonresponders (26 and33 months, respectively; p � 0.62).
• • •The present data indicate that, in the absence of
concurrent cardiovascular events, the beneficial effectsof carvedilol or metoprolol on LV function and dimen-sion clearly persist after the first 12 months of therapy inpatients with ischemic and nonischemic cardiomyopathy.
When compared with placebo, administration ofcarvedilol or metoprolol for 12 months substantiallyimproves LVEF and reduces LV end-diastolic dimen-sion in patients with ischemic and nonischemic car-diomyopathy.3–5 Few nonplacebo-controlled studieshave assessed the effects of � blockade on LV func-tion and dimension for �12 months. After adminis-tration of metoprolol for 18 months, LV function,mass, and geometry were still improved in patientswith ischemic and nonischemic cardiomyopathy com-pared with baseline values.5 However, these LV pa-rameters were not compared at 12 and 18 months.5LVEF and LV end-diastolic dimension continued toimprove from 12 to 18 months after starting �-blockertherapy in 33 patients with ischemic and idiopathicdilated cardiomyopathy who were switched frommetoprolol to carvedilol or vice versa.6 After admin-istration of carvedilol for 20 months to patients with
ischemic and nonischemic cardiomyopathy, 1/3 ofpatients had significantly lower LVEF and greater LVend-diastolic dimension at 20 months than those notedat 12 months.7 In contrast, the effects of metoprolol onLVEF and LV end-diastolic dimension were fullymaintained 40 months after initiation of therapy inpatients with idiopathic dilated cardiomyopathy.8
Patients with nonischemic cardiomyopathy, and es-pecially patients with idiopathic dilated cardiomyopathy,experience greater improvement in LVEF and a reduc-tion in LV end-diastolic dimension than patients withischemic cardiomyopathy during long-term �-blockertherapy.9 Considering previous experience, the long-last-ing effects of carvedilol or metoprolol on LV functionand dimension were somewhat expected in patients withnonischemic cardiomyopathy. The present data are thefirst to document the long-lasting effects of carvedilol ormetoprolol on LV function and dimension in patientswith ischemic cardiomyopathy who are free of clinicalcardiovascular events that could alter LV function. Ourfindings contrast with the previously reported attenuationof the effects of � blockade on LV function and dimen-sion in patients with ischemic cardiomyopathy.7 Themost plausible explanation for the steady effects of �blockade on LV function is the absence of clinical eventsthat could alter LV function and dimension in our patients.
The major limitation of the present study is its obser-vational nature and the absence of a control group ofpatients. However, the striking benefits of � blockers inpatients with chronic heart failure due to LV systolicdysfunction precludes withdrawal of this therapy.10,11 Inthat context, nonplacebo-controlled observational studiesmay, despite their limitations, provide useful informationto better our understanding of �-blocker therapy in pa-tients with chronic heart failure.
1. Sahn DJ, DeMaria A, Kisslo J, Weyman A. Recommendations regardingquantitation in M-mode echocardiography: results of survey of echocardiographicmeasurements. Circulation 1978;58:1072–1083.2. Schiller NB, Shah PM, Crawford M. Recommendations for quantitation of theleft ventricle by two-dimensional echocardiography. J Am Soc Echocardiogr1989;2:58–68.3. Doughty RN, Whalley GA, Gamble G, MacMahon S, Sharpe N. Left ventric-ular remodeling with carvedilol in patients with congestive heart failure due toischemic heart disease. J Am Coll Cardiol 1997;29:1060–1066.4. Waagstein F, Hjalmarson A, Swedberg K, Bristow MR, Gilbert EM, CameriniF, Fowler MB, Silver MA, Johnson MR, Goss FG. Beneficial effects of meto-prolol in idiopathic dilated cardiomyopathy. Lancet 1993;342:1441–1446.5. Hall SA, Cigarroa CG, Marcoux L, Risser RC, Grayburn PA, Eichhorn EJ.Time course improvement in left ventricular function, mass and geometry inpatients with congestive heart failure treated with beta-adrenergic blockade. J AmColl Cardiol 1995;25:1154–1161.6. Maack C, Elter T, Nickenig G, LaRosee K, Crivaro M, Stäblein A, Wuttke H,Böhm M. Prospective crossover comparison of carvedilol and metoprolol inpatients with chronic heart failure. J Am Coll Cardiol 2001;38:939–946.7. Levine TB, Levine AB, Bolenbaugh J, Green PR. Reversal of heart failureremodeling: is it maintained? Clin Cardiol 2003;26:419–423.8. DeMaria R, Di Lenarda A, Gavazzi A. Metoprolol-induced functional benefitin dilated cadiomyopathy is sustained over four years and favorably influencesoutcome. Ital Heart J 2001;2:130–138.9. Metra M, Nodari S, Parrinello G, Giubbini R, Manca C, Cas LD. Markedimprovement in left ventricular ejection fraction during long term �-blockade inpatients with chronic heart failure: clinical correlates and prognostic significance.Am Heart J 2003;145:292–299.10. Packer M, Bristow MR, Cohn JN, Colucci WS, Fowler MB, Gilbert EM,Shusterman NH, for the U.S. Carvedilol Heart Failure Study Group. The effect ofcarvedilol on morbidity and mortality in patients with chronic heart failure.N Engl J Med 1996;334:1349–1355.11. MERIT-HF Study Group. Effect of metoprolol CR/XL in chronic heartfailure: Metoprolol CR/XL randomized Interventional Trial in Congestive HeartFailure. (MERIT-HF). Lancet 1999;353:2001–2007.
1116 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Prognostic Impact of Diabetes Mellitus in PatientsWith Acute Decompensated Heart Failure
Andrew J. Burger, MD, Lana Tsao, MD, and Doron Aronson, MD
The presence of diabetes mellitus (DM) adversely af-fects the natural history of heart failure (HF), but itsprognostic significance is unknown in acute decom-pensated HF. Of the 498 patients enrolled with de-compensated HF requiring intravenous vasoactivetherapy, 236 (47.4%) had a previous diagnosis ofDM. After 6 months, 113 patients (22.7%) had died.A Cox proportional-hazards model showed a signif-icant association between DM and worse survivalafter hospital discharge. DM is common among pa-tients admitted with decompensated HF, and diabe-tes-related biologic differences contribute to the pro-gression of HF. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1117–1119)
D iabetes mellitus (DM) has been associated withan increased risk for the development of heart
failure (HF).1,2 Not only is DM associated with anincreased risk for ischemic heart disease and subse-quent development of HF,3 but DM, in conjunctionwith hypertension and the microangiopathic processof DM, can lead to HF. Population-based studies haverevealed the adverse effect of DM on prognosis ofpatients with HF.4 However, it is not known whetherthe effect of DM on survival extends to patients withsevere decompensated HF. The aim of this study wasto determine whether the presence of DM in patientswith severe decompensated HF adversely affectsprognosis.
• • •The study group was derived from the Vasodilation
in the Management of Acute Congestive HF (VMAC)trial. A detailed description of the study has beenpublished elsewhere.5 Briefly, this trial was a prospec-tive, multicenter, placebo- and active-controlled trialin which 498 patients with acute decompensated HFwere randomized to vasoactive therapy. Randomiza-tion was stratified based on the investigator’s clinicaldecision to use a right-sided cardiac catheter to man-age the decompensated HF. Eligible patients weremen or women with decompensated HF with symp-toms of dyspnea at rest or immediately upon minimalactivity. The following major exclusion criteria were:(1) systolic blood pressure consistently �90 mm Hg; (2)cardiogenic shock, volume depletion, or any other clin-ical condition that would contraindicate the administra-
tion of an intravenous vasodilatory agent; (3) most recentpulmonary capillary wedge pressure �20 mm Hg within24 hours before randomization; or (4) need for me-chanical ventilation.
Information about DM status was systematicallyobtained as part of the recruitment and follow-upprocedure of the VMAC study. Patients were classi-fied as having DM on the basis of history, regardlessof duration of disease or need for antidiabetic agents.The diagnosis could have been based on a previousphysician telling patients that they had DM or patientstaking or having previously taken oral hypoglycemicagents or insulin or receiving diet therapy.
Data are expressed as mean � SD. Baseline char-acteristics of the groups were compared using the t testfor continuous variables and by the chi-square statisticfor noncontinuous variables. Baseline hemodynamicmeasurements in patients with and without DM werecompared using the Mann-Whitney rank-sum test.The multivariate Cox proportional-hazards regressionmodel was the primary statistical analysis used todetermine the independent relation of DM and otherbaseline characteristics to survival. The followingbaseline clinical characteristics were considered in themodeling procedure: age, gender, weight, baselinecreatinine, primary etiology of HF (stratified as isch-emic or nonischemic), presence of atrial fibrillation,automatic implantable cardioverter-defibrillator im-plantation, left ventricular ejection fraction, baselineheart rate and systolic blood pressure, use of medica-tions (digoxin, angiotensin converting-enzyme inhib-itors, angiotensin II receptor blockers, � blockers, andamiodarone), and vasoactive therapy assignments. Allpotential confounders were forced into the model.Differences were considered statistically significant atthe 2-sided p �0.05 level. All statistical analyses wereperformed using SPSS statistical software, version11.5 (SPSS Inc., Chicago, Illinois).
At study entry, all patients had dyspnea at rest, and84% were in chronic New York Heart Association classIII or IV HF before decompensation. Of 498 patients,236 (47.3%) had a previous diagnosis of DM. Invasivehemodynamic measurements were available in 108 pa-tients with DM and in 126 patients without DM. Nosignificant differences were found in baseline pulmonarycapillary wedge pressure (28.1 vs 27.7 mm Hg;p � 0.53), cardiac index (2.1 vs 2.1 L/min/m2; p � 0.40),systemic vascular resistance (1,469 vs 1,428 dynes/s/cm�5;p � 0.42), and right atrial pressure (15.0 vs 14.9mm Hg; p � 0.85) among patients with and withoutDM.
Analysis of the clinical characteristics of patientswith and without DM also revealed several differ-ences (Table 1). Patients with DM were older and had
From the Beth Israel Deaconess Medical Center, Boston, Massachusetts;and Rambam Medical Center, Haifa, Israel. Dr. Burger’s address is:Beth Israel Deaconess Medical Center, Noninvasive Cardiology Lab-oratory, Baker 3, 1 Deaconess Road, Boston, Massachusetts 02215.E-mail: [email protected]. Manuscript received November1, 2004; revised manuscript received and accepted December 20,2004.
greater body weight. They were less likely to haveischemic heart disease as the primary cause of theirHF. Patients with DM had higher ejection fractionsand were less likely to be treated with �-adrenergicblockers.
During a mean follow-up of 6 months, 113 patients(22.7%) died. Results of a Cox proportional-hazardsmodel for 6-month mortality are listed in Table 2. DMwas found to be a significant independent predictor ofworse survival after consideration of numerous base-line clinical characteristics (RR 1.78, 95% confidenceinterval [CI] 1.19 to 2.65, p � 0.005). Figure 1 showsthe adjusted survival curves for patients with andwithout DM.
• • •We sought to determine whether DM in acute
decompensated HF affects long-term survival. Ourstudy demonstrates that patients with acute decompen-sated HF have a worse prognosis in the presence ofDM than patients without DM. These findings suggest
that DM-related biologic differencesare present in the progression of HF,even with advanced disease.6
The high prevalence of DM amongpatients with advanced HF enrolledin the VMAC study is striking andunderscores the need for furtherknowledge of the interaction be-tween DM and HF. This finding isprobably not explained by the higherrisk for HF conferred by DM.2 Apossible contributing factor to thevery high prevalence of DM in thepresent study is the association be-tween HF and increased insulin re-sistance.7 Recent studies have shownthat advanced HF is associated witha significantly higher risk of devel-oping DM.8,9
The prognosis of diabetic patientswith HF has only been studied in asmall number of trials and in thesetting of stable HF. The DanishInvestigations of Arrhythmia andMortality on Dofetilide trial10 was arandomized, double-blinded, placebo-controlled study (n � 5,491) evaluat-ing the efficacy of the antiarrhythmicdofetilide in patients with advancedHF. Compared with non-DM pa-tients, patients with DM had a longerduration of their HF, worse systolicdysfunction, greater frequency of hy-pertension and ischemic heart dis-ease, and were taking more medica-tions. DM emerged as a potent,independent risk factor for death inpatients hospitalized for HF (RR 1.5,95% CI 1.3 to 1.76, p �0.0001). Thisrisk was more striking in womenthan in men (RR 1.7, 95% CI 1.4 to1.9, p � 0.0001).
The �-Blocker Evaluation of Survival Trial11 was amulticenter, randomized, double-blinded, placebo-controlled study (n � 2,708) evaluating the efficacy ofthe � blocker bucindolol in patients with New YorkHeart Association class III and IV HF and left ven-tricular ejection fraction �35%. This investigationdemonstrated that DM portended a worse prognosisfor patients with advanced chronic HF (i.e., all-causeand cardiovascular death, pump failure death, andhospitalization for HF), but this finding was limited topatients with ischemic cardiomyopathy.
In the Studies of Left Ventricular Dysfunction tri-als (n � 6,797) and registry (n � 6,273), DM wasshown to be an independent risk factor for de-creased survival, or hospitalization, or both.12 In theFramingham Heart study of new-onset HF (n � 652)with a mean follow-up of 3.9 years, DM was animportant prognostic, survival parameter.2 In a largeScottish cohort study13 (n � 66,547), DM was pre-dictive of long-term mortality. In a retrospective
*The following baseline clinical characteristics were considered in the Cox modeling procedure: age,gender, weight, baseline creatinine, primary etiology of HF (stratified as ischemic or nonischemic),presence of atrial fibrillation, automatic implantable cardioverter-defibrillator implantation, left ventric-ular ejection fraction, baseline heart rate, and systolic blood pressure, use of medications, and vaso-active therapy assignments.
1118 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
French study14 (n � 1,246), DM was an independentpredictor of cardiovascular mortality in ischemic pa-tients (RR 1.54, 95% CI 1.13 to 2.09, p � 0.006) butnot in nonischemic patients (RR 0.65, 95% CI 0.39 to1.07, p � 0.09). In contrast, DM had no prognosticimplications for long-term survival in several otherHF trials.15,16
The mechanism by which DM confers a worseprognosis is unknown. Prognostic implications of DMmay reflect, in part, the higher prevalence of hyper-tension, atherosclerotic heart disease, renal disease,and the combined cardiorenal syndrome.17 The exis-tence of a specific diabetic cardiomyopathy may beanother important potential factor.
1. Iribarren C, Karter AJ, Go AS, Ferrara A, Liu JY, Sidney S, Selby JV. Glycemiccontrol and heart failure among adult patients with diabetes. Circulation 2001;103:2668–2673.2. Ho KK, Pinsky JL, Kannel WB, Levy D. The epidemiology of heart failure: theFramingham Study. J Am Coll Cardiol 1993;22:6A–13A.3. Dries DL, Sweitzer NK, Drazner MH, Stevenson LW, Gersh BJ. Prognosticimpact of diabetes mellitus in patients with heart failure according to the etiologyof left ventricular systolic dysfunction. J Am Coll Cardiol 2001;38:421–428.4. Bauters C, Lamblin N, Mc Fadden EP, Van Belle E, Millaire A, De Groote P.Influence of diabetes mellitus on heart failure risk and outcome. CardiovascDiabetol 2003;2:1.5. VMAC Investigators. Intravenous nesiritide vs nitroglycerin for treatment ofdecompensated congestive heart failure: a randomized controlled trial. JAMA2002;287:1531–1540.6. Solang L, Malmberg K, Ryden L. Diabetes mellitus and congestive heartfailure. Further knowledge needed. Eur Heart J 1999;20:789–795.7. Grundy SM. Higher incidence of new-onset diabetes in patients with heartfailure. Am J Med 2003;114:331–332.8. Tenenbaum A, Motro M, Fisman EZ, Leor J, Freimark D, Boyko V,Mandelzweig L, Adler Y, Sherer Y, Behar S. Functional class in patients withheart failure is associated with the development of diabetes. Am J Med2003;114:271–275.9. Witteles RM, Tang WH, Jamali AH, Chu JW, Reaven GM, Fowler MB.Insulin resistance in idiopathic dilated cardiomyopathy: a possible etiologic link.J Am Coll Cardiol 2004;44:78–81.10. Gustafsson I, Brendorp B, Seibaek M, Burchardt H, Hildebrandt P, Kober L,Torp-Pedersen C. Influence of diabetes and diabetes-gender interaction on therisk of death in patients hospitalized with congestive heart failure. J Am CollCardiol 2004;43:771–777.11. Domanski M, Krause-Steinrauf H, Deedwania P, Follmann D, Ghali JK,Gilbert E, Haffner S, Katz R, Lindenfeld J, Lowes BD, et al. The effect ofdiabetes on outcomes of patients with advanced heart failure in the BEST trial.J Am Coll Cardiol 2003;42:914–922.12. Shindler DM, Kostis JB, Yusuf S, Quinones MA, Pitt B, Stewart D, PinkettT, Ghali JK, Wilson AC. Diabetes mellitus, a predictor of morbidity and mortalityin the Studies of Left Ventricular Dysfunction (SOLVD) trials and registry. Am JCardiol 1996;77:1017–1020.13. MacIntyre K, Capewell S, Stewart S, Chalmers JW, Boyd J, Finlayson A,Redpath A, Pell JP, McMurray JJ. Evidence of improving prognosis in heartfailure: trends in case fatality in 66 547 patients hospitalized between 1986 and1995. Circulation 2000;102:1126–1131.14. De Groote P, Lamblin N, Mouquet F, Plichon D, McFadden E, Van Belle E,Bauters C. Impact of diabetes mellitus on long-term survival in patients withcongestive heart failure. Eur Heart J 2004;25:656–662.15. Simon T, Mary-Krause M, Funck-Brentano C, Jaillon P. Sex differences inthe prognosis of congestive heart failure: results from the Cardiac InsufficiencyBisoprolol Study (CIBIS II). Circulation 2001;103:375–380.16. Adams KF Jr, Sueta CA, Gheorghiade M, O’Connor CM, Schwartz TA, KochGG, Uretsky B, Swedberg K, McKenna W, Soler-Soler J, Califf RM. Genderdifferences in survival in advanced heart failure. Insights from the FIRST study.Circulation 1999;99:1816–1821.17. Aronson D, Mittleman MA, Burger AJ. Elevated blood urea nitrogen level asa predictor of mortality in patients admitted for decompensated heart failure. Am JMed 2004;116:466–473.
FIGURE 1. Adjusted survival curves for patients with and withoutDM. Cox proportional-hazards method revealed a significantsurvival advantage for patients without DM compared with pa-tients with DM (p � 0.005) after adjusting for a variety of base-line clinical characteristics.
BRIEF REPORTS 1119
Effect of a Seated Exercise Program to ImprovePhysical Function and Health Status in Frail Patients
>70 Years of Age With Heart Failure
Miles D. Witham, BM, BCh, Joan M. Gray, MCSP, Ishbel S. Argo, SRN,Derek W. Johnston, PhD, Allan D. Struthers, MD, and Marion E.T. McMurdo, MD*
Eighty-two patients aged >70 years with heart fail-ure were randomized to a gentle, seated exerciseprogram or to usual care. Six-minute walk distanceand quality of life did not change between groups,but daily activity as measured by accelerometry in-creased in the exercise group relative to the controlgroup. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1120–1124)
We hypothesized that a seated exercise programdesigned specifically for older frail patients
with heart failure (HF) would be well attended andwould lead to improvements in exercise capacity, ev-eryday activity, and health status.1–6 To test this hy-pothesis, we undertook a randomized, single-blind,controlled trial comparing a multifaceted, seated ex-ercise program (designed specifically for older, frailpatients with HF) with a usual-care program.
• • •We recruited patients from the local specialist HF
clinic and from the local Medicine for the Elderlyclinics. Patients aged �70 years with a clinical diag-nosis of chronic heart failure according to EuropeanSociety of Cardiology guidelines, New York HeartAssociation class II or III symptoms, and evidence ofleft ventricular systolic dysfunction on echocardiog-raphy, contrast ventriculography, or radionuclide ven-triculography were eligible to participate. Exclusioncriteria were patients with uncontrolled atrial fibrilla-tion, significant aortic stenosis, sustained ventriculartachycardia, recent myocardial infarction, inability towalk without human assistance, abbreviated mentaltest score �6 of 10, or those currently undergoingphysiotherapy or rehabilitation. Written informed con-sent was obtained from all participants; the studyprotocol was approved by Tayside Committee onMedical Research Ethics.
We randomized participants after performing base-line assessments. A researcher not otherwise con-nected with the operation of the study prepared cardscontained in numbered, sealed envelopes from com-puter-generated random number tables. An experi-enced research nurse who was blinded to treatmentallocation performed all assessments. Participants un-derwent assessments at baseline (before randomiza-tion) and at 3 and 6 months.
The primary outcome measure was the 6-minutewalk distance.7,8 Participants undertook a 6-minutewalk along a 25-m corridor with standardized encour-agement. Daily activity over a 7-day period was mea-sured using the Stayhealthy RT3 triaxial accelerometer(Stayhealthy Inc, Monrovia, California).9 The devicewas mounted anteriorly on the participant’s waist-band, and recorded summed acceleration counts at1-minute intervals. We asked participants to wear thedevice from when they first dressed in the morning towhen they retired at night. Data from the first and lastdays were discarded to reduce the influence of incom-plete days and transport artifacts. We administered theGuyatt chronic heart failure questionnaire7,10 (a dis-ease-specific health-related quality-of-life measure),the Hospital Anxiety and Depression score,11 thePhiladelphia Geriatric Morale Scale,12 and the modi-fied Functional Limitations Profile—a United Kingdomversion of the Sickness Impact Profile13 in interviewformat during a home visit.
An experienced physiotherapist delivered the ex-ercise intervention, which was divided into supervisedand home phases. In the supervised phase (0 to 3months), participants attended exercise classes as out-patients in groups of 3 to 4, twice a week during thefirst 3 months. Between 17 and 20 sessions were
From the Section of Ageing and Health and Department of ClinicalPharmacology, University of Dundee, Ninewells Hospital, Dundee;and School of Psychology, University of Aberdeen, Aberdeen, UnitedKingdom. This study was supported by Grant 2006/918 from TheHealth Foundation (formerly PPP Health Foundation), London, UnitedKingdom. Dr. Witham’s address is: Section of Ageing and Health,Ninewells Hospital, Dundee DD1 9SY, United Kingdom. E-mail:[email protected]. Manuscript received September 22,2004; revised manuscript received and accepted January 4, 2005.
*Dr. McMurdo is a director of DD Developments, a University ofDundee company promoting exercise classes for subjects aged �60years. Any profits are pledged to support research into the health ofolder people. We confirm that the conduct of the study, analysis, andpublication of the results are independent of the study funders.
offered during the 3-month period. Each session be-gan with a warm-up and ended with a cool-downsequence of movements. Further segments of exercisewere added in between these sequences, consisting ofupper limb exercise, lower limb exercise, slow wholebody aerobic movements, and quicker whole bodyaerobic movements. A new segment was added ateach session until the whole 6-part program was per-formed. The 6-part program took approximately 20minutes to complete and was set to music. At thispoint, wrist and ankle weights were introduced in asimilar sequential fashion until the 6 sessions wereperformed using 500-g wrist weights and 1.1-kg ankleweights. Weights and participation were adjusted toeach participants ability and progress. We encouragedparticipants to use the Borg rating of perceived exer-tion scale,14 aiming for a Borg level of between 11 and13. We advised participants who rated their perceivedexertion as �13 to rest or reduce the level of activityduring the exercise session. After performing the ex-ercises, participants undertook a series of breathingexercises and a 10-minute relaxation session to finishthe session. We encouraged spouses or other familymembers to attend any or all of the sessions. Partici-pants kept a diary detailing their main daily activities
over this 3-month period; the phys-iotherapist reviewed the diary withthe participant weekly and set newtargets for daily walking activity.
In the second phase (3 to 6months), we asked participants tocontinue performing the exercises athome 2 to 3 times per week with theaid of a video or audio cassette withdemonstrations, instructions, and mu-sic. There was no face-to-face con-tact with the physiotherapist duringthis period. Participants continued tokeep a diary of their daily activities,which we used as a basis for a weeklytelephone liaison. During these tele-phone calls, the physiotherapist gaveencouragement and agreed on newtargets for daily walking activity.
Participants in the control groupreceived usual care. We gave stan-dardized written information aboutthe diagnosis and management ofheart failure to participants in bothgroups. We told participants in thecontrol group that exercise was notharmful for their condition and wedid not ask the control group to re-strict their activities in any way.
We calculated that 33 patients pergroup were required to provide 90%power to detect a 30-m difference atthe 0.05 significance level, assuminga baseline 6-minute walk distance of230 m and a SD of 50 m.7,15 A 30-mchange has previously been shown tobe the minimum clinically important
change in the 6-minute walk. Allowing for dropouts,the final target number for recruitment to the trial wasthus 84 patients. Data were analyzed using SPSSstatistical software, version 11.5 (SPSS., Chicago, Il-linois). Baseline variables were compared using Stu-dent’s t test for continuous variables and chi-squaretesting for discrete variables. Percent changes betweenbaseline and 3 months and also between baseline and6 months were calculated and compared using theStudent’s t test for normally distributed variables, andthe Mann-Whitney U statistic test for skewed vari-ables. A p value �0.05 was considered statisticallysignificant.
Participants were recruited between January 2002and October 2003, and the 6-month follow-up wascompleted in April 2004. Figure 1 gives details ofparticipant flow and follow-up; Table 1 lists baselinedetails of patients randomized into the study. Majorreasons for declining participation were poor health,lack of interest in exercise, and frequent current ac-tivity and exercise. We offered a total of 758 person-sessions of exercise to the 41 participants in the ex-ercise group, a mean of 18.5 sessions (range 12 to 20)per participant. Participants attended a total of 626person-sessions, a mean of 15.3 sessions (range 0 to
TABLE 1 Baseline Characteristics of Patients Randomized
CharacteristicExercise Group
(n � 41)Control Group
(n � 41)
Mean age � SD 80 � 6 81 � 4Men 26 (63%) 19 (46%)New York Heart Association class II/III 25 vs 16 21 vs 20Ischemic etiology 31 (76%) 23 (56%)Left ventricular systolic dysfunction
*p �0.05 exercise versus control group.ACE � angiotensin-converting enzyme; BMI � body mass index.
BRIEF REPORTS 1121
20) attended per participant. The attendance rate wasthus 15.3 of 18.5 (82.7%). Thirty-one of 41 particiants(75.6%) attended �80% of their allocated sessions.We noted no adverse events during the delivery of theexercise sessions. Twenty-seven of 41 participants(66%) in the exercise group progressed to using theheaviest weights offered by the end of the supervisedphase of the exercise program. Outcome measures aregiven in Tables 2 to 4. The mean/median percentchange in each variable between 0 and 3 months, andalso between 0 and 6 months, is given, together withthe raw scores. Details of deaths and hospitalizationsare given in Table 5.
• • •This is the largest trial of exercise in patients aged
�70 years with HF to date, with a mean participantage of 80 years. It is also one of the few trials ofexercise to enroll patients with multiple co-morbidi-ties and to assess a comprehensive set of outcomes ofpractical relevance to the care of older people withHF. In contrast to most HF trials, our exercise pro-gram did not require the use of specialist equipmentlike bicycles or treadmills16 and was specifically de-signed for use by older HF patients with significantco-morbid disease. The supervised exercise phase waswell attended and tolerated by the participants, re-flected by the high participation rate. Furthermore, theprogram appeared to be safe, with no adverse eventsnoted during delivery of the exercise sessions and noexcess of deaths or hospitalizations in the exercise armof the study.
Our most important finding was the significantincrease in everyday physical activity measured byaccelerometry in the exercise group at 6 months. Thisoccurred despite the absence of any face-to-face con-tact with either the therapist or trial team between 3and 6 months. This is the first report of an objectiveincrease in everyday physical activity levels 3 monthsafter the end of a formal exercise program. The fewstudies of exercise training for HF that have measuredeveryday activity failed to find an improvement de-spite improvements in exercise capacity.17–19 The ben-eficial effects of exercise on exercise capacity persistonly for as long as the exercise is maintained,19 andstudies to date have been dogged by rapid reversion topreexercise function when the sedentary state is re-sumed. Unlike physical activity questionnaires relyingon self-report, our observations were based on objec-tive accelerometry data, using a device validated foruse in elderly people.9 The findings are intriguing,especially given the lack of improvement in the6-minute walk test. The increase in everyday activitywas most likely to have resulted from behavioralchanges brought about by participation in the pro-gram, telephone follow-up, keeping a diary, and thesetting of walking targets, although this is impossibleto prove in the absence of a suitable control group.The home-based seated exercises were peformed dur-ing the week-long accelerometry assessments, but de-tailed analysis of the accelerometry data suggests thatonly a small fraction of the increase in accelerometrycounts seen in the exercise group was attributable to
TABL
E2
Out
com
eM
easu
res
(six
-min
ute
wal
kdi
stanc
e)
Out
com
e
0m
o3
mo
6m
o
Exer
cise
Con
trol
Exer
cise
Con
trol
pVa
lue
Exer
cise
Con
trol
pVa
lue
Six-
min
wal
k(m
)(m
ean
�SD
)26
1�
117
240
�93
254
�10
722
8�
102
—26
2�
110
246
�11
1—
Mea
n%
chan
gefro
mba
selin
e(9
5%C
I)—
—2.
1(�
4.2
to8.
4)�
3.8
(�11
.4to
3.7)
0.23
4.4
(�3.
1to
11.8
)3.
2(�
6.0
to12
.4)
0.84
Exer
cise
vers
usco
ntro
lcom
paris
ons
byin
depe
nden
tsam
ple
Stud
ent’s
ttes
t.C
I�co
nfide
nce
inte
rval
.
1122 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
TABLE 3 Outcome Measures (daily activity and function)
Outcome
0 mo 3 mo 6 mo
Exercise Control Exercise Control p Value Exercise Control p Value
Accelerometry (mean � SD) 92,565 � 51,227 95,163 � 62,190 10,6021 � 98,315 93,373 � 62,862 10,915 � 70,385 82,463 � 53,832Median % change from baseline (IQR) — — 18.7 (�28.5 to 51.8) 7.0 (�29.1 to 36.8) 0.51* 2.3 (�11.1 to 46.6) �14.0 (�37.7 to 25.4) 0.036*FLP physical (mean � SD) 299 � 85 310 � 65 309 � 64 326 � 39 303 � 79 322 � 37Median % change from baseline (IQR) — — 0 (�4.7 to 4.6) 4.6 (�0.8 to 8.9) 0.085* 0 (�2.3 to 4.4) 2.9 (�2.1 to 10.2) 0.21*FLP psychosocial (mean � SD) 344 � 85 360 � 67 338 � 87 363 � 70 339 � 77 364 � 63Mean % change from baseline (95% CI) — — �0.8 (�9.7 to 8.2) 3.6 (�6.6 to 13.8) 0.52 1.0 (�7.2 to 9.2) 3.3 (�2.9 to 9.4) 0.65FLP total (mean � SD) 1,050 � 198 1,088 � 139 1,053 � 181 1,106 � 116 1,054 � 180 1,108 � 105Mean % change from baseline (95% CI) — — �0.2 (�3.6 to 3.3) 2.7 (�1.6 to 7.0) 0.29 1.2 (�2.3 to 4.8) 4.4 (0.1 to 8.7) 0.26
*Exercise versus control comparisons by Mann-Whitney U test. All other comparisons by independent sample Student’s t test.FLP � Functional Limitations Profile; IQR � interquartile range; other abbreviation as in Table 2.
TABLE 4 Outcome Measures (health-related quality of life and psychological status)
Outcome
0 mo 3 mo 6 mo
Exercise Control Exercise Control p Value Exercise Control p Value
Guyatt score (mean � SD) 67 � 13 70 � 12 68 � 11 69 � 12 65 � 10 69 � 13Mean % change from baseline (95% CI) — — 6.5 (�1.1 to 14.1) �0.3 (�4.1 to 3.6) 0.11 1.7 (�5.2 to 8.6) �1.9 (�9.6 to 5.7) 0.48PGMS (mean � SD) 12.0 � 3.3 11.4 � 2.3 11.5 � 3.3 11.5 � 2.7 12.0 � 3.4 11.6 � 2.6Mean % change from baseline (95% CI) — — 1.1 (�7.8 to 9.9) 0.7 (�6.0 to 7.4) 0.95 3.8 (�3.5 to 11.1) �0.2 (�6.3 to 5.9) 0.40HADS depression (mean � SD) 4.1 � 2.5 5.5 � 2.8 4.7 � 2.8 5.0 � 3.0 4.5 � 2.9 5.1 � 3.0Mean % change from baseline (95% CI) — — 31.6 (1.0 to 62.2) �4.6 (�22.0 to 12.7) 0.04 17.7 (�3.5 to 38.9) �3.4 (�18.3 to 11.5) 0.10HADS anxiety (mean � SD) 3.5 � 4.1 2.5 � 2.3 3.2 � 3.8 2.0 � 1.8 3.4 � 4.3 2.5 � 2.9Mean % change from baseline (95% CI) — — 8.5 (�34.8 to 51.8) �26.5 (�44.5 to �8.6) 0.13 �3.1 (�30.0 to 23.7) 1.5 (�16.1 to 19.2) 0.77
Exercise versus control comparisons by independent sample Student’s t test.HADS � Hospital Anxiety and Depression Score; PGMS � Philadelphia Geriatric Morale Scale; other abbreviation as in Table 2.
BRIEFREPO
RTS1123
wearing the accelerometers during the seated exercisesessions. Importantly, our accelerometry data weremirrored by preservation of functional capacity in theexercise group as measured by the Functional Limi-tations Profile, compared with a decrease in functionalcapacity in the control group, although this did notreach statistical significance. The effect of suchchanges in everyday physical activity on long-termprognosis from HF, even when the 6-minute walk testremains unchanged, merits further study; daily activ-ity has been shown to be a powerful independentpredictor of prognosis in 1 study.20 If our data onpreservation of functional capacity can be confirmedover a longer time period, the possibility arises thatsuch an exercise program may be able to prevent theonset of institutionalization and dependence; the meanbaseline Functional Limitations Profile scores confirmthat this patient group was severely functionally im-paired and thus at significant risk of institutionaliza-tion had their conditions deteriorated.
1. Cowie MR, Wood DA, Coats AJ, Thompson SG, Poole-Wilson PA, Suresh V,Sutton GC. Incidence and aetiology of heart failure; a population-based study.Eur Heart J 1999;20:421–428.2. Lien CT, Gillespie ND, Struthers AD, McMurdo ME. Heart failure in frailelderly patients: diagnostic difficulties, co-morbidities, polypharmacy and treat-ment dilemmas. Eur J Heart Fail 2002;4:91–98.
3. Masoudi FA, Havranek EP, Wolfe P, Gross CP, Rathore SS, Steiner JF, OrdinDL, Krumholz HM. Most hospitalized older persons do not meet the enrollmentcriteria for clinical trials in heart failure. Am Heart J 2003;146:250–257.4. Lloyd-Williams F, Mair FS, Leitner M. Exercise training and heart failure: asystematic review of current evidence. Br J Gen Pract 2002;52:47–55.5. Piepoli MF, Davos C, Francis DP, Coats AJ. Exercise training meta-analysisof trials in patients with chronic heart failure (ExTraMATCH). BMJ 2004;328:189–192.6. Witham MD, Struthers AD, McMurdo ME. Exercise training as a therapy forchronic heart failure: can older people benefit? J Am Geriatr Soc 2003;51:699–709.7. O’Keeffe ST, Lye M, Donnellan C, Carmichael DN. Reproducibility andresponsiveness of quality of life assessment and six minute walk test in elderlyheart failure patients. Heart 1998;80:377–382.8. Guyatt GH, Sullivan MJ, Thompson PJ, Fallen EL, Pugsley SO, Taylor DW,Berman LB. The 6-minute walk: a new measure of exercise capacity in patientswith chronic heart failure. Can Med Assoc J 1985;132:919–923.9. Kochersberger G, McConnell E, Kuchibhatla MN, Pieper C. The reliability,validity, and stability of a measure of physical activity in the elderly. Arch PhysMed Rehabil 1996;77:793–795.10. Guyatt GH, Nogradi S, Halcrow S, Singer J, Sullivan MJ, Fallen EL.Development and testing of a new measure of health status for clinical trials inheart failure. J Gen Intern Med 1989;4:101–107.11. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. ActaPsychiatr Scand 1983;67:361–370.12. Lawton MP. The Philadelphia Geriatric Center Morale Scale: a revision. JGerontol 1975;30:85–89.13. Pollard B, Johnston M. Problems with the Sickness Impact Profile: a theo-retically based analysis and a proposal for a new method of implementation andscoring. Soc Sci Med 2001;52:921–934.14. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc1982;14:377–381.15. Beynon JH, Pathy MS. An open, parallel group comparison of quinapril andcaptopril, when added to diuretic therapy, in the treatment of elderly patients withheart failure. Curr Med Res Opin 1997;13:583–592.16. McKelvie RS, Teo KK, Roberts R, McCartney N, Humen D, Montague T,Hendrican K, Yusuf S. Effects of exercise training in patients with heart failure:the Exercise Rehabilitation Trial (EXERT). Am Heart J 2002;144:23–30.17. Gottlieb SS, Fisher ML, Freudenberger R, Robinson S, Zietowski G, Alves L,Krichten C, Vaitkevicus P, McCarter R. Effects of exercise training on peakperformance and quality of life in congestive heart failure patients. J Card Fail1999;5:188–194.18. Berg-Emons R, Balk A, Bussmann H, Stam H. Does aerobic training lead toa more active lifestyle and improved quality of life in patients with chronic heartfailure? Eur J Heart Fail 2004;6:95–100.19. Willenheimer R, Rydberg E, Cline C, Broms K, Hillberger B, Oberg L,Erhardt L. Effects on quality of life, symptoms and daily activity 6 months aftertermination of an exercise training programme in heart failure patients. IntJ Cardiol 2001;77:25–31.20. Walsh JT, Charlesworth A, Andrews R, Hawkins M, Cowley AJ. Relation ofdaily activity levels in patients with chronic heart failure to long-term prognosis.Am J Cardiol 1997;79:1364–1369.
1124 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Usefulness of Enalapril Versus Propranolol or Atenololfor Prevention of Aortic Dilation in Patients With the
Marfan Syndrome
Anji T. Yetman, MD, Renee A. Bornemeier, MD, and Brian W. McCrindle, MD, MPH
Despite variable clinical results, � blockers have be-come the primary therapy for prevention of aorticdilation in patients with the Marfan syndrome. Thisstudy examines the use of the angiotensin-convertingenzyme inhibitor enalapril for treatment of these pa-tients. We sought to examine the effects of enalaprilversus �-blocker therapy in patients with the Marfansyndrome and noted improved aortic distensibility(3.0 � 0.3 vs 1.9 � 0.4 cm2 dynes�1; p <0.02) anda reduced aortic stiffness index (8.0 � 2.9 vs 18.4 �3.8; p <0.05) in patients receiving enalapril com-pared with those receiving � blockers. These favor-able hemodynamic changes were associated with asmaller increase in aortic root diameter (0.1 � 1.0 vs5.8 � 5.2 mm) and fewer clinical end points duringfollow-up. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1125–1127)
Treatment with angiotensin-converting enzyme (ACE)inhibitors has been shown to increase aortic dis-
tensibility in patients without Marfan syndrome,1 aneffect that appears to be related to improved endothe-lial performance.2 Although there are no clinical trialsof ACE inhibitors in patients with the Marfan syn-drome, in vitro studies suggest promising results. An-giotensin II-regulated vascular smooth muscle cell(VSMC) apoptosis plays a substantial role in the his-tologic abnormality of aortic medial degeneration re-sponsible for aortic dilation and dissection in patientswith the Marfan syndrome.3 ACE inhibitors have beenshown to inhibit VSMC apoptosis in cultured aorticmedia from patients with the Marfan syndrome.3 It hasbeen proposed that ACE inhibitors may, through in-hibition of VSMC apoptosis, reduce the risk of aorticdissection in these patients.3 We sought to assess theeffect of enalapril and � blockers on aortic elasticproperties and aortic growth rate in a group of patientswith the Marfan syndrome.
• • •We prospectively enrolled 58 patients with the
Marfan syndrome diagnosed in accordance with the
revised Ghent criteria.4 All patients had evidence ofcardiovascular involvement at the time of enrollment.Patients were excluded from the study if they hadundergone previous cardiovascular surgery, werepregnant, or had greater than mild aortic or mitralinsufficiency. The study was approved by the institu-tional review board and was performed in accordancewith institutional guidelines.
New patients were nonrandomly assigned to open-label treatment with either a � blocker (propranolol inchildren weighing �12.5 kg or atenolol in thoseweighing �12.5 kg) or enalapril. Random assignmentof medication was not performed because there wereno data to support the clinical use of ACE inhibitors atthe initiation of this study, and � blockers were con-sidered standard care. All patients were given theoption of either medication. Beta-blocker therapy con-sisted of either atenolol at a starting dose of 1 mg/kgbody weight per day, with the dose titrated to a max-imal dose of 2 mg/kg body weight per day or propran-olol at a dose of 1 mg/kg twice daily. Therapy withenalapril involved a starting dose of 2.5 mg twicedaily, titrated upward to a maximal dose of 10 mgtwice daily. The final dose was the highest dose tol-erated by the patient without important adverse ef-fects. No effort was made to adjust therapy to thepatients’ heart rate or blood pressure. Subjects wereassessed on a semiannual basis from the time of en-rollment until surgery, pregnancy, or completion of 3years of treatment. Indications for surgical interven-tion on a dilated aortic root included an absolute aorticroot diameter of 50 mm in a patient weighing �50 kgor an aortic root diameter �150% than the uppernormal value for age using previously establishednormograms.5
At the time of enrollment in the study, subjects
From the Division of Cardiology, Department of Pediatrics, The Chil-dren’s Hospital, University of Colorado Health Sciences Center, Denver,Colorado; Arkansas Children’s Hospital, University of Arkansas Med-ical Sciences, Little Rock, Arkansas; and The Hospital for Sick Chil-dren, University of Toronto, Toronto, Ontario, Canada. This report wassupported by a grant from the National Marfan Foundation, PortWashington, New York. Dr. Yetman’s address is: Department ofPediatric Cardiology, B-100, The Children’s Hospital, 1056 East19th Avenue, Denver, Colorado 80220. E-mail: [email protected]. Manuscript received September 29, 2004; re-vised manuscript received and accepted January 4, 2005.
FIGURE 1. Patients segregated by which cardiovascular medica-tion was being taken before and during the course of the study.BB � � blocker (propranolol or atenolol).
underwent echocardiographic assessment off of allcardiac medications. Thereafter, biannual assessmentswere performed while the subject was taking the as-signed medication. Echocardiographic assessment in-cluded measurement of systolic and diastolic dimen-sions of the aorta at the level of the sinuses of Valsalvaimaged from a parasternal long-axis view6 and leftventricular end-diastolic dimension and percent short-ening fraction using M-mode analysis imaged fromthe parasternal short-axis view. The subject’s bloodpressure was obtained in the supine position at thetime of assessing aortic root dimensions using a cuffsphygmomanometer. Three readings were obtainedand the average value used. The following calcula-tions were then made in accordance with previouslywell established guidelines7: aortic stiffness index:natural logarithm (systolic/diastolic blood pressure)/(systolic-diastolic aortic diameter)/diastolic aortic di-ameter; and aortic distensibility: 2 (systolic-diastolicaortic diameter)/(diastolic aortic diameter) · (pulsepressure).
Because normal aortic dimensions are related topatient body size, aortic root dimensions were notrecorded as absolute values. Rather, to adjust forchanges related to body size and somatic growth, theaortic root dimensions were expressed as a percentage ofpredicted dimension based on body surface area usingpublished nomograms.5 Echocardiographic analysiswas performed by a cardiologist blinded to medica-tions the patient was receiving.
All data are expressed as frequencies, medianswith ranges, and mean � SD as appropriate. Thechange in each subject’s echocardiographic measure-ments was related to the appropriate time interval overwhich the change occurred and calculated as a rate ofchange per year. The serial rates of change for eachvariable were then related to other clinical variables,particularly the type of medication used during theinterval, in repeated-measures, mixed linear regres-sion analysis. All analyses were performed using SASstatistical software, version 8 (SAS Institute, Inc.Cary, North Carolina) using default settings.
Fifty-seven patients (mean age 12.8 � 7.8 years
and body surface area 1.4 � 0.5 m2 at enrollment)were followed prospectively for a mean interval of 3.0 �0.2 years. During prospective data collection, 32 pa-tients received enalapril, 24 atenolol, and 2 receivedpropranolol (Figure 1). Six new patients initially opt-ing for �-blocker therapy discontinued treatment dur-ing the study because of adverse effects (depression in2, fatigue in 2, and short-term memory loss in 2) andwere subsequently enrolled in the enalapril group. Nopatient withdrew from the enalapril group and noserious side effects were reported. Two patients re-ceiving � blockers before initiation of the study choseto change to enalapril because of adverse effects(mood swings in 1 and poor school performance inanother). Four patients initially receiving no medicationbecause of adverse effects on previous �-blocker therapyopted for treatment with enalapril. The median dose ofenalapril was 5 mg twice daily, the median atenolol dosewas 25 mg twice daily, and the median propranolol dosewas 1 mg/kg/dose 3 times/day.
Despite lack of randomization, there were no im-portant differences at enrollment between treatmentgroups (Table 1). The rate of change per year forpercent predicted aortic root dimension (n � 314observations) was significantly related to the medica-tion used during the interval (Table 1). After adjustingfor medications, the only other variable that was sig-nificantly related to the rate of change was subject age(after logarithmic transformation, p � 0.008) and per-cent predicted aortic root dimension (after logarithmictransformation, p � 0.008), suggesting that, indepen-dent of medication, the rate of change increases withadvancing age and with greater percent predicted aor-tic root dimension. The rate of change was not signif-icantly related to height, weight, gender, or bloodpressure at each assessment.
Repeated measures of stiffness index and aorticdistensibility (n � 379 observations) were signifi-cantly related to the medication that the subject wastaking at the time of the measurement (Table 1). Agreater stiffness index and lower aortic distensibilitywere also independently associated with higher systolicblood pressure (p �0.02 and p �0.001, respectively).
TABLE 1 Demographic and Echocardiographic Measurements
VariableEnalapril(n � 32)
Propranolol/Atenolol(n � 25) p Value
Mean � SD age (yrs) at start of therapy 14.6 � 7.7 12.0 � 7.6 0.2Mean initial aortic diameter (mm) 33 � 8 31 � 6 0.7Mean initial % predicted aortic size 118 � 18 118 � 14 0.7Mean change in aortic diameter (mm) during the study 0.1 � 1.0 5.8 � 5.2 �0.001Mean heart rate on therapy (beats/min) 78 � 20 66 � 13 0.05Mean systolic BP on therapy (mm Hg) 100 � 20 98 � 20 0.8Rate of change in % predicted aortic size over time (%/yr) �2.5 � 1.0 3.7 � 1.4 0.005Aortic stiffness index 8.0 � 2.9 18.4 � 3.8 �0.05Aortic distensibility (cm2/dynes�1) 3.0 � 0.3 1.9 � 0.4 �0.02Male (48%) (51%) NSAdverse outcomes
Death 0 1Aortic root replacement 2 7
BP � blood pressure.
1126 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
Neither aortic stiffness index nor aortic distensibilitywere associated with height, weight, gender, diastolicblood pressure, aortic root dimension, left ventricularsize, or shortening fraction.
There was 1 death, occurring in a subject withdocumented ventricular arrhythmias while taking�-blocker therapy. The subject died after a ventriculararrhythmia and had no evidence of aortic dissection onautopsy. Nine patients underwent aortic root replace-ment for progressive dilation (2 while on enalapril and7 while on � blockers). The 2 patients on ACE inhib-itor therapy were approaching surgical criteria at thetime of enrollment into the study, with aortic rootdiameters of 49 mm (148% of predicted diameter) and30 mm (147% of predicted diameter), respectively.
• • •We have shown that patients treated with enalapril
have a reduction in the aortic stiffness index, accom-panied by improved aortic distensibility. These alter-ations in aortic elastic tone are associated with aslower rate of aortic growth. The improvement inaortic root dilation in patients treated with enalaprilmay result in fewer adverse outcomes. There are sev-eral potential mechanisms by which ACE inhibitorsmay alter aortic form or function. First, patients withthe Marfan syndrome have been documented to haveabnormal endothelial function contributing to in-creased aortic stiffness.8,9 However, they have a pre-served response to the vasodilator bradykinin.9 ACEinhibitors, through their inhibition of bradykininbreakdown, may lead to a direct alteration in aorticelastic tone. The reduction in aortic stiffness may thenlead to less pulse-wave reflectance and less pulsatilestress on the aortic root. Second, early initiation ofmedical therapy with ACE inhibitors may preventaortic degradation through inhibition of angiotensin-mediated VSMC apoptosis. It has previously beendocumented that children diagnosed with the Marfansyndrome at a young age and started on medicaltherapy early in the course of their disease have lessaortic root dilation over time, despite more advanceddisease involvement at the time of diagnosis.10 Fi-nally, patients with Marfan syndrome associated withcardiovascular complications may have higher levels
of plasma homocysteine.11 Hyperhomocysteinemiaincreases vascular stiffness through an increase inangiotensin II responsiveness.12,13 This effect may beblocked by ACE inhibitor therapy.
Irrespective of its mechanism of action, ACE in-hibitors appear to provide previously undocumentedtherapeutic benefit in patients with the Marfan syn-drome. A multicenter, randomized control trial ofenalapril and � blockers may help better define whichof these therapies is of greater benefit for long-termmanagement of these patients.
1. Topouchian J, Brisac AM, Pannier B, Vicaut E, Safar M, Asmar R. Assessmentof the acute effects of converting enzyme inhibition in essential hypertension: adouble-blind, comparative and crossover study. J Hum Hypertens 1998;12:181–187.2. Drakos SG, Papamichael CM, Alexopoulos GP, Anastasiou-Nana MI,Stathopoulos JV, Nanas JN. Effects of high doses versus standard doses ofenalapril on endothelial cell function in patients with chronic congestive heartfailure secondary to idiopathic dilated or ischemic cardiomyopathy. Am JCardiol 2003;91:885– 888.3. Nagashima H, Sakomura Y, Aoka Y, Uto K, Kameyama K, Ogawa M, Aomi S,Koyanagi H, Ishizuka N, Naruse M, Kawana M, Kawanuki H. Angiotensin II type 2receptor mediates vascular smooth muscle cell apoptosis in cystic medial degenera-tion associated with Marfan’s syndrome. Circulation 2001;104:282–287.4. DePaepe A, Devereux RB, Dietz HC, Hennekam RCM, Pyeritz RE. Reviseddiagnostic criteria of the Marfan syndrome. Am J Med Genetics 1996;62:417–426.5. Roman MJ, Devereux RB, Kramer-Fox R, O’Loughlin J. Two-dimensionalechocardiographic aortic root dimensions in normal children and adults. Am JCardiol 1989;64:507–512.6. Yetman AT, Bornemeier RA, McCrindle BW. Long-term outcome in patientswith Marfan syndrome: is aortic dissection the only cause of sudden death? J AmColl Cardiol 2003;41:329–332.7. Hirata K, Triposkiadis F, Sparks E, Bowen J, Wooley CF, Boudoulas H. TheMarfan syndrome: abnormal aortic elastic properties. J Am Coll Cardiol 1991;18:57–62.8. Wilson DG, Bellamy MF, Ramsey MW, Goodfellow J, Brownlee M, Davies S,Wilson JF, Lewis MJ, Stuart G. Endothelial function in Marfan syndrome—selectiveimpairment of flow-mediated vasodilation. Circulation 1999;99:909–915.9. Nakamura M, Itoh S, Makita S, O’hira A, Arakawa N, Hiramori K. Peripheralresistance vessel dysfunction in Marfan syndrome. Am Heart J 2000;139:661–666.10. Yetman AT, Huang P, Bornemeier RA. Comparison of outcome of Marfansyndrome in patients diagnosed at age �6 years versus those diagnosed �6 yearsof age. Am J Cardiol 2003;91:102–103.11. Giusti B, Porciani MC, Brunelli T, Evangelisti L, Fedi S, Gensini G, AbbateR, Sani G, Yacoub M, Pepe G. Phenotypic variability of cardiovascular mani-festations in Marfan syndrome. Possible role of hyperhomocysteinemia andC677T MTHFR gene polymorphism. Eur Heart J 2003;24:2038–2045.12. Neves MF, Endemann D, Amiri F, Virdis A, Pu Q, Rozen R, Schiffrin EL.Small artery mechanics in hyperhomocysteinemic mice: effects of angiotensin II.J Hypertens 2004;22:959–966.13. Bonaventura D, Tirapelli CR, Haddad R. Chronic methionine load-inducedhyperhomocysteinemia enhances rat carotid responsiveness for angiotensin II.Pharmacology 2004;70:91–99.
BRIEF REPORTS 1127
Mechanism of Spaceflight-Induced Changes in LeftVentricular Mass
Richard L. Summers, MD, David S. Martin, BS, Janice V. Meck, PhD,and Thomas G. Coleman, PhD
Decrements in left ventricular (LV) mass observed af-ter microgravity exposure have been previously pos-tulated to be a central component of spaceflight-induced cardiovascular deconditioning. In this study,echocardiographic measurements of LV mass in as-tronauts demonstrated a comparative 9.1% reductionin postflight LV mass that returned to preflight valuesby the third day of recovery. A ground-based study innormal subjects determined that these pre- to post-flight LV mass changes could be reproduced by sim-ple dehydration. Reductions in LV mass observed im-mediately after spaceflight may be secondary tosimple physiologic fluid exchanges. �2005 by Ex-cerpta Medica Inc.
(Am J Cardiol 2005;95:1128–1130)
Cardiovascular adaptations during spaceflight arean important concern for returning astronauts.1,2
In this study, we measured the sequential changes inleft ventricular (LV) mass by echocardiography beforeflight, on landing day, and 3 days after landing inastronauts who flew on missions lasting from 9 to 16days. In addition, because astronauts are hypovolemicon landing day, we collected LV mass measurementsin ground-based subjects during normovolemic andhypovolemic states and compared the data with theastronaut pre-/postflight data. Our results suggest thatthe reductions in LV mass after spaceflight may be anartifact due to the loss of fluid volume.
• • •Thirty-eight astronauts (11 men and 27 women;
aged 42 � 6 years) from shuttle missions lasting from9 to 16 days were evaluated 10 days before flight, 2 to3 hours after landing, and 3 days after flight by either2-dimensional echocardiography (Philips HDI 5000,Bothell, Washington) or M-mode echocardiography(Genesis II, Biosound, Indianapolis, Indiana). On eachtest day, subjects abstained from taking any medica-tions, caffeine, or alcohol for the previous 12 hours
and had not eaten a heavy meal for 4 hours norexercised maximally within 24 hours. All were ingood health, having recently completed Air Forceclass III physicals before flight. On landing day, allsubjects had consumed the standard fluid load of oralsalt tablets and water to equal 15 ml/kg of isotonicsaline before re-entry and had consumed water adlibitum between landing and the time of echocardio-graphic measurements. Two-dimensional LV masswas calculated from echocardiographic 2-dimensionalimages using an area–length method (1.05 · (((5/6) · epicardial area · (length � average myocardialthickness)) � ((5/6) · endocardial area · length))). Av-erage wall thickness was calculated as the square rootof the LV area in short axis at end-diastole divided by3.14 minus the square root of the LV area in short axisat peak systole. M-mode measurements were: LVmass � (1.04 · [diastolic ventricular septal thickness� diastolic LV posterior wall thickness � diastolicLV internal diameter]3 � diastolic LV internal di-ameter3) � 13.6 g. The thickness of the ventricularseptum and the LV posterior wall were also mea-sured on each view, along with LV diameter atend-diastole. Images from both the 2-dimensionaland M-mode studies were stored on VHS videotapefor off-line analysis (Prosolv, Indianapolis, Indiana, orMicrosonics, Indianapolis, Indiana). Three images ofeach 2-dimensional study and 4 images of each M-mode study were digitized and each image was ana-lyzed by 2 different observers.
Preliminary analysis of the astronaut data sug-gested that the well-known occurrence of dehydrationon landing day may have influenced LV mass mea-surements.3 Therefore, an additional ground-basedstudy was initiated to determine if the pre- to post-flight LV mass changes could be reproduced by sim-ple dehydration in normal subjects who had not flownin space. Thirteen healthy, normotensive, nonsmok-ing men and women (8 men and 5 women; aged39 � 7 years) were studied twice: once while nor-movolemic and once while hypovolemic. The sameechocardiographic techniques were used. All sub-jects had also passed an Air Force class III physical.On each test day, subjects had abstained from tak-ing any medications, caffeine, or alcohol for theprevious 12 hours, had not eaten a heavy mealwithin 4 hours, and had not exercised maximallywithin 24 hours.
Subjects in the normovolemic state were given (inaddition to their normal diet) two 1.0-g salt tabletsevery morning for 7 days to ensure a high normalsodium intake. Fluid intake was as needed. On the
From the Department of Emergency Medicine, University of MississippiMedical Center, Jackson, Mississippi; Wyle Laboratories, NationalAeronautics and Space Administration Johnson Space Center, Houston,Texas; Human Adaptation and Countermeasures Office, Space andLife Sciences Directorate, National Aeronautics and Space Adminis-tration Johnson Space Center, Houston, Texas; and Department ofPhysiology and Biophysics, University of Mississippi Medical Center,Jackson, Mississippi. Dr. Summers’ address is: Department of Emer-gency Medicine, University of Mississippi Medical Center, 2500North State Street, Jackson, Mississippi 39216. E-mail: [email protected]. Manuscript received August 18, 2004; revised manuscriptreceived and accepted December 30, 2004.
eighth day, blood volume and plasma volume weremeasured each day using a previously described car-bon dioxide rebreathing technique.4,5 Echocardio-graphic measurements of LV mass were also obtainedat this time.
The hypovolemic state was induced in these samesubjects with the following protocol: for 24 hoursbefore the study, subjects were asked to restrict saltin their diet. The following day, a urine samplewas collected, and if the sodium/creatinine ratio was�100 mEq/g, an intravenous catheter was placed,through which furosemide (0.5 mg/kg) was adminis-tered at a rate of 4 mg/min. Blood pressure and urineoutput were monitored. Subjects were allowed toleave the testing area once urine output had stabilized(after approximately 1 hour). At that time, subjectswere given a prepared low salt (10 mEq/24 hours),caffeine-free diet (Hermann Hospital and UniversityClinical Research Center, Houston Texas), diet log,and distilled water. Subjects were instructed to eatonly the prepared diet for the following 36 hours.Fluid intake was as needed, but consisted of onlydistilled water or bottled water that contained 0 mg ofsodium content. At the end of the 36 hours, plasmavolume and echocardiographic measurements wererepeated.
One-way analysis of variance was used for theflight studies and paired t tests were used to evaluatethe ground-based study. Statistical significance was
determined by a p value �0.05. The Johnson SpaceCenter institutional review board approved proto-cols and all subjects gave their written informedconsent.
Table 1 lists astronaut data. There was a 9.1%decrease in measured LV mass from before flight tolanding day, but this loss was completely recovered 3days after flight. LV end-diastolic volume followedthe same trend. Wall thickness did not have a signif-icant decrease on landing day. There were no dif-ferences in mass, volume, or wall thickness betweenbefore flight and after 3 days flight. There was alsono correlation between flight duration and eitherLV mass or LV end-diastolic volume (9 to 16 days;r � 0.086).
In the ground-based subjects, the low salt regimenreduced plasma volume significantly (3.05 � 0.15 to2.61 � 0.14 liters, p �0.001). Measured LV mass, LVend-diastolic volume, and average wall thickness dur-ing hypovolemia were all significantly lower than theywere during normovolemia. Most notably, there was asignificant correlation between plasma volume andLV mass (r � 0.71; Figure 1).
• • •The findings presented here suggest that the loss of
LV mass measured in astronauts after spaceflight maynot be primarily due to cardiac atrophy. The mostpersuasive argument for this assertion is that the mea-sured LV mass was completely restored within 3 days,and it is unlikely that cardiac muscle can remodel tothat extent in such a short time period.6,7 This led us tosuspect that the measured decrease in LV mass may berelated to the known dehydration on landing day.3,8
The ground-based study reinforces this conclusionby reproducing the astronaut landing day findingsthrough dehydration of normal subjects, demonstrat-ing a significant correlation between measured LVmass and plasma volumes. Corroboration for thismechanistic interpretation of our data comes from theevidence provided by Hunold et al9 using magneticresonance imaging to measure cardiac volumes. Theyfound significant acute changes in LV mass as fluidwas removed during the course of dialysis. They alsoconcluded that the mean decrease in LV mass of 3.7%could be explained by a loss of cardiac interstitialvolume with the treatment.
FIGURE 1. Significant correlation (r � 0.71) was found betweenplasma volume (liters) and LV mass (grams) in ground-basedsubjects with simple dehydration.
TABLE 1 Sequential Changes in Left Ventricular (LV) mass, Volume, and Wall Thickness from Preflight to Landing day to theRecovery Period (3 days postflight)*
Preflight Landing Day RecoveryPreflight vs Landing
*Differences between preflight and recovery measurements were not statistically significant.
BRIEF REPORTS 1129
adrenergic receptor responses before flight and central autonomic dysregulationpostflight. Am J Physiol Heart Circ Physiol 2004;286:H1486–H1495.2. Perhonen MA, Franco F, Lane LD, Buckey JC, Blomqvist CG, Zerwekh JE,Peshock RM, Weatherall PT, Levine BD. Cardiac atrophy after bed rest andspaceflight. J Appl Physiol 2001;91:645–653.3. Fritsch-Yelle JM, Whitson PA, Bondar RL, Brown TE. Subnormal norepi-nephrine release relates to presyncope in astronauts after spaceflight. J ApplPhysiol 1996;81:2134–2141.4. Dahms TE, Horvath SM. Rapid, accurate technique for determination ofcarbon monoxide in blood. Clin Chem 1974;20:533–537.5. Thomsen JK, Fogh-Andersen N, Bulow K, Devantier A. Blood and plasmavolumes determined by carbon monoxide gas, 99mTc-labelled erythrocytes,125I-albumin and T 1824 technique. Scand J Clin Lab Invest 1991;51:185–190.
6. Dunn FG, Ventura HO, Messerli FH, Kobrin I, Frohlich ED. Time course ofregression of left ventricular hypertrophy in hypertensive patients treated withatenolol. Circulation 1987;76:254–258.7. Mirsky I, Laks MM. Time course of changes in the mechanical properties ofthe canine right and left ventricles during hypertrophy caused by pressureoverload. Circ Res 1980;46:530–542.8. Waters WW, Ziegler MG, Meck J. Postflight orthostatic hypotension occursmostly in women and is predicted by low vascular resistance. J Appl Physiol2002;92:586–594.9. Hunold P, Vogt FM, Heemann UW, Zimmermann U, Barkhausen J. Myocar-dial mass and volume measurement of hypertrophic left ventricles by MRI—study in dialysis patients examined before and after dialysis. J Cardiovasc MagnReson 2003;5:553–561.
Usefulness of Contrast Echocardiography for Diagnosisof Left Ventricular Noncompaction
Lotte E. de Groot-de Laat, MD, Boudewijn J. Krenning, MD,Folkert J. ten Cate, MD, and Jos R. T. C. Roelandt, MD
Many imaging modalities have been evaluated for thediagnosis of left ventricular noncompaction (LVNC).Echocardiography with color Doppler flow imaging iscurrently the diagnostic modality of choice. In this study,the usefulness of contrast echocardiography to diag-nose LVNC was examined, and the diagnostic criteriawere evaluated. �2005 by Excerpta Medica Inc.
(Am J Cardiol 2005;95:1130–1134)
Many imaging modalities have been evaluated forthe diagnosis of left ventricular noncompaction
(LVNC), including echocardiography,1 angiography,2computed tomography,3,4 and magnetic resonance im-aging.4,5 However, echocardiography complementedwith color Doppler flow imaging is currently the di-agnostic modality of choice.1 We present a series ofadult patients with suspected LVNC on conventional2-dimensional echocardiography and evaluate the incre-mental diagnostic value of contrast echocardiography.
• • •From October 2002 to December 2003, 3,807 pa-
tients were referred for transthoracic echocardio-graphic examination from an outpatient clinic. In 18patients (mean age 40 � 11 years; 9 men), a diagnosisof LVNC was suspected on standard 2-dimensionalechocardiography after cardiac examination, includ-ing history, physical examination, and electrocardio-gram (Table 1). Five patients were in heart failure, 5patients were already known to have unclassified car-diomyopathy, and 7 patients were first-degree rela-tives of patients with LVNC. One patient had a historyof ventricular tachycardia, and 3 patients had congen-ital heart defects (1 hypertrophic obstructive cardio-myopathy; 1 right and left ventricular outflow obstruc-
tion, and 1 atrial septal defect, type secundum, withpulmonary valve stenosis).
Complete 2-dimensional echocardiographic and colorDoppler examinations (Philips Sonos 5500, PhilipsMedical Systems, Andover, Massachusetts) were per-formed in the 18 patients, followed by the intravenousinjection of SonoVue 0.5 to 1.0 ml (Bracco S.p.A.,Milan, Italy)6 (Figure 1). Standard settings were usedfor left ventricular opacification at a mechanical indexof 0.3. Images were collected from 3 apical views andthe parasternal short-axis view. A left ventricularthrombus was detected in 1 patient. The standard16-segment model was used for analysis and localiza-tion of the noncompacted segments. The presence orabsence of noncompaction in patient segments wasevaluated using a 4-point score: (1) segment inade-quately visualized, (2) normal, (3) dubious for non-compaction, and (4) noncompaction present. All im-ages were independently analyzed by 2 experiencedechocardiographers. In case of disagreement, a majoritydecision was achieved by a third reviewer. The leftventricular ejection fraction was measured on contrastimages, using the biplane area–length method.7
Previously proposed diagnostic echocardiographiccriteria for LVNC were used: (1) a 2-layer wall struc-ture with a compacted thin epicardial band and thickernoncompacted endocardial layer of trabecular meshworkwith deep endomyocardial spaces (ratio of noncom-pacted to compacted thickness �2), (2) a predominantlocalization of the trabeculation in the mid-lateral, apical,and mid-inferior segments, and (3) evidence of colorDoppler flow in deep perfused intertrabecular recess-es.1 To calculate the noncompacted/compacted thick-ness ratio, the thickness of the layers was measured atthe site of maximal thickness in the apical 3- or4-chamber view at end-diastole on contrast-enhancedimages.
Categorical data are presented as counts and per-centages, whereas continuous data are summarized asmean � SD or median value with range.
No patient showed signs of side effects after the
From the Department of Cardiology, Thoraxcenter, Erasmus MC,Rotterdam, The Netherlands. Dr. Roelandt’s address is: Erasmus MC,Thoraxcenter, Room H536, Dr. Molewaterplein 40, 3015 GDRotterdam, The Netherlands. E-mail: [email protected] received August 10, 2004; revised manuscript receivedand accepted December 9, 2004.
Patient Age (yrs)/Sex History Clinical Presentation NYHA Class Associated Cardiac Defect Echo Abnormalities EF (%) X/Y
1 21 F HC Congenital heart defect I 0 � (RBBB) 85 1.72 33 F HC CHF III 0 � (LVH) 83 2.03 38 M Unclassified HC Screening I SSS, PAF 0 59 3.44 44 F Unclassified HC, myectomy Control II 0 � (LVH) — —5 45 F Unclassified HC, Graves’ disease Control I Interatrial septum aneurysm � (Left axis deviation) 67 2.06 28 F 0 CHF II 0 0 57 2.27 31 F Crohn’s disease Screening I 0 0 54 2.28 33 M 0 Screening I 0 0 44 2.59 33 M 0 Viral infection I 0 � (VES, VT) — —
10 34 M Crohn’s disease Screening I 0 0 60 2.511 35 M Radiotherapy (leukemia) Acute CHF IV 0 � (LBBB) 17 6.012 35 F Pulmonary homograft, ASD
closure, PM (SSS)Congenital heart defect II Pulmonary valve stenosis,
ASD� (PM) 46 3.6
13 36 F 0 Screening I 0 0 58 3.014 45 F 0 CHF, AP, palpitations III 0 0 63 1.815 48 M Insulin-dependent diabetes
mellitusScreening I 0 0 32 2.5
16 57 M VT, cardiomyopathy Control I 0 � (LBBB) 25 2.517 58 M Cognitive disorder Screening I 0 0 49 3.218 62 M Pneumonia 3-vessel disease Acute CHF IV Atrial flutter � (VES) 31 2.3
AP � angina pectoris; ASD � atrial septal defect; CHF � congestive heart failure; Echo � echocardiographic; EF � ejection fraction; HC � hypertrophic cardiomyopathy; LBBB � left bundle branch block; LVH � left ventricular hypertrophy;NYHA � New York Heart Association; PAF � paroxysmal atrial fibrillation; PM � pacemaker; RBBB � right bundle branch block; SSS � sick sinus syndrome; VES � ventricular extrasystole; VT � ventricular tachycardia; X � measure offree-wall thickness to peak of trabeculation; Y � distance between epicardial surface and trough of recess; 0 � absent.
BRIEFREPO
RTS1131
injection of the ultrasound contrast agent. Eighteenpatients had suspected LVNC on their standard 2-di-mensional echocardiograms, and the diagnoses weremade definitive with contrast echocardiography in 16patients. However, in 2 patients, the prominent myo-cardial trabeculations were in the normal range. In 7patients (44%), morphologic abnormalities in the rightventricle were found that appeared quite similar tothose of the noncompacted left ventricle segments, butno specific criteria for the right ventricle are available.In 9 patients (56%), the characteristic noncompactionmorphology was isolated in the left ventricle.
Figure 2 shows the number of segments adequatelyvisualized for analysis: with standard echocardiogra-phy, 269 of 288 segments (93.4%) and with contrastenhancement, 287 of 288 segments (99.6%). Of the163 segments interpreted as normal on the standard
echocardiograms, 23 myocardial seg-ments (14%) were noncompacted on thecontrast-enhanced images. Thirty-sixsegments were dubious for noncompac-tion on standard echocardiographic ex-amination, and 20 were classified asnoncompacted and 10 as normal aftercontrast enhancement. All segments di-agnosed as noncompacted on standardechocardiography were also diagnosedas noncompacted after contrast enhance-ment.
The median myocardial noncom-pacted-to-compacted left ventricular ratiowas 2.5 (range 1.7 to 6.0), measured withcontrast echocardiography (Figure 3). Thelocalization of noncompacted myocardialsegments is shown in Figure 4. Thesewere predominantly localized in the apexand the mid-posterior wall (11 of 16patients and 9 of 16 patients, respec-tively) on standard echocardiography.With contrast enhancement, the involve-ment appeared in 15 of 16 patients and
14 of 16 patients, respectively. The mid-ventricularsegments of the lateral and inferior wall appeared tobe affected on standard echocardiography in 5 of 16patients. In contrast-enhanced images in 10 of 16patients and 7 of 16 patients, respectively, these seg-ments were interpreted as noncompacted. The in-volvement of the interventricular septum and otherbasal segments is less frequent.
The measurement of mitral inflow velocities showednormal diastolic function in 11 patients (69%), pseudo-normalization in 1 patient, a restrictive pattern in 1 pa-tient, and impaired relaxation in 3 patients. The meanejection fraction was 52% � 19%, and 7 patients (44%)had ejection fractions �50%.
• • •Ventricular angiography, magnetic resonance imag-
ing, and computed tomography have been used for thediagnosis of LVNC and are described in case reportswithout proposing specific diagnostic criteria.2–5 Cur-rently, the imaging modality of choice for LVNC isechocardiography, and diagnostic criteria were proposedby Jenni et al.1 The patients included in our studyfulfilled most of these criteria. They all had suspicionof LVNC on conventional 2-dimensional echocardio-graphic examination by hypertrophic segments andprominent trabeculations. Although in the classic ap-pearance, the diagnosis of LVNC is not difficult, inmore subtle noncompaction, the differential diagnosisamong hypertrophic cardiomyopathy, prominent nor-mal myocardial trabeculations, dilated cardiomyopa-thy, and left ventricular apical thrombus may be dif-ficult. In this situation, magnetic resonance imaging,computed tomography, or ventricular angiography canbe helpful in making a definitive diagnosis. In a pre-vious case report, we described the use of contrast-enhanced echocardiography in a patient with noncom-paction cardiomyopathy.8 In this study, we tested
FIGURE 1. Apical 4-chamber view with noncompacted myocardium of the lateralwall (A) and apical 3-chamber view with perfused intertrabecular recesses oncolor Doppler imaging (B).
FIGURE 2. Presence of LVNC for the total number of segmentsevaluated in all patients before (2-dimensional) and during con-trast (2-dimensional contrast) echocardiography.
1132 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
contrast echocardiography to confirm the definitive di-agnosis of noncompaction cardiomyopathy in 18 pa-tients.
The use of contrast allows clear endocardial borderdelineation and better evaluation of almost all seg-ments not visible on standard echocardiography(Figure 2). Noncompacted myocardial segments dur-ing contrast echocardiography are more readily iden-tified. One of the main diagnostic criteria of LVNC isa predominant localization of the trabeculation in mid-lateral, mid-inferior, and apical segments. In our pa-tients, the apex and mid-ventricular segments of theposterior, inferior, and lateral wall were most affected,more with contrast than with standard echocardiogra-phy, followed by the mid-anterior and basal posteriorwall (Figure 4), whereas involvement of the septumand other basal segments was uncommon. Anotherproposed echocardiographic diagnostic criterion is acommunication between multiple deep intertrabecularrecesses and the ventricular cavity, demonstrated by
color Doppler imaging. Because bloodflow velocity is low in the spaces be-tween the trabeculations, the communi-cation between the recesses and theventricular cavity can be missed withcolor Doppler flow imaging. This wouldbe possible with echo contrast opacifyingthe deep perfused recesses (Figure 3). Toquantify LVNC, several investigatorssuggested determining the ratio of non-compacted to compacted layers of themyocardium by measuring the depth ofthe intertrabecular recesses in relation tothe thickness of the myocardium.1,9 Forthe best visual differentiation, they as-sessed the thickness of the noncom-pacted and the compacted layers in theshort-axis view during end-systole.Some of our patients had only promi-nent trabeculations in the apex, and itwas best visualized in the long-axisview. We used the contrast-enhancedlong-axis 3- or 4-chamber view to as-sess the end-diastolic myocardial wallthickness and to determine the noncom-
pacted-to-compacted ratio. We found only clear casesof LVNC to have hypertrabeculation on end-systolicimages. In the more subtle noncompacted segmentsand in patients with normal left ventricular ejectionfractions, the intertrabecular recesses were not visibleat end-systole because of a good contraction state ofthe left ventricular wall. The ratio of noncompacted tocompacted layer ranged from 1.7 to 6.0 in our patients(definition �2). Nevertheless, the 2 patients with ra-tios of 1.7 and 1.8 were included because of theirtrabeculation and specific echocardiographic images.The depths of the recesses are difficult to assess echo-cardiographically, better with contrast enhancement, butto determine the ratio is dependent on which location andvolume status measurements are performed.
LVNC may present with systolic and diastolic dys-function, cardiac arrhythmias (ventricular tachycar-dia), and embolic events (systemic and pulmonaryembolism). Suspected unusual hypertrophy or hyper-trabeculation on 2-dimensional echocardiography inpatients with these symptoms is a reason to do acontrast study. The screening of family members with2-dimensional and contrast echocardiography leads toearlier diagnoses in asymptomatic patients.10–12
1. Jenni R, Oechslin E, Schneider J, Attenhofer Jost C, Kaufmann PA. Echocar-diographic and pathoanatomical characteristics of isolated left ventricular non-compaction: a step towards classification as a distinct cardiomyopathy. Heart2001;86:666–671.2. Baumhakel M, Janzen I, Kindermann M. Images in cardiovascular medicine.Cardiac imaging in isolated noncompaction of ventricular myocardium. Circu-lation 2002;106:e16–e17.3. Conces DJ Jr, Ryan T, Tarver RD. Noncompaction of ventricular myocardium:CT appearance. Am J Roentgenol 1991;156:717–718.4. Hamamichi Y, Ichida F, Hashimoto I, Uese KHK, Miyawaki T, Tuskano S,Ono Y, Echigo S, Kamiya T. Isolated noncompaction of the ventricular myocar-dium: ultrafast computed tomography and magnetic resonance imaging. IntJ Cardiovasc Imaging 2001;17:305–314.
FIGURE 3. Apical long-axis view with suspected abnormal apex on standard echocar-diography (A) and apical noncompaction of the myocardium on contrast echocardiog-raphy (B), with measurement of noncompacted inner (X) and compacted outer layer(Y).
FIGURE 4. Distribution of noncompacted segments of the left ven-tricle in conventional and contrast echocardiography.
BRIEF REPORTS 1133
5. Daimon Y, Watanabe S, Takeda S. Two-layered appearance of noncompactionof the ventricular myocardium on magnetic resonance imaging. Circ J 2002;66:619–621.6. Bokor D. Diagnostic efficacy of SonoVue. Am J Cardiol 2000;86(suppl):19G–24G.7. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, Feigenbaum H,Gutgesell H, Reichek N, Sahn D, Schnittger I, et al. Recommendations for quanti-tation of the left ventricle by two-dimensional echocardiography. American Societyof Echocardiography Committee on Standards, Subcommittee on Quantitation ofTwo-Dimensional Echocardiograms. J Am Soc Echocardiogr 1989;2:358–367.8. De Laat LE, Galema TW, Krenning BJ, Roelandt JR. Diagnosis of non-compaction cardiomyopathy with contrast echocardiography. Int J Cardiol 2004;94:127–128.
9. Chin TK, Perloff JK, Williams RG, Jue K, Mohrmann R. Isolated noncom-paction of left ventricular myocardium. A study of eight cases. Circulation1990;82:507–513.10. Dusek J, Ostadal B, Duskova M. Postnatal persistence of spongy myocardiumwith embryonic blood supply. Arch Pathol 1975;99:312–317.11. Jenni R, Rojas J, Oechslin E. Isolated noncompaction of the myocardium.N Engl J Med 1999;340:966–967.12. Richardson P, McKenna W, Bristow M, Maisch B, Mautner B, O’Connell J,Olsen E, Thiene G, Goodwin J, Gyarfas I, et al. Report of the 1995 World HealthOrganization/International Society and Federation of Cardiology Task Force onthe definition and classification of cardiomyopathies. Circulation 1996;93:841–842.
1134 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 MAY 1, 2005
