J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6
ª 2 0 1 6 B Y T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y F O UN DA T I O N I S S N 1 9 3 6 - 8 7 9 8 / $ 3 6 . 0 0
P U B L I S H E D B Y E L S E V I E R h t t p : / / d x . d o i . o r g / 1 0 . 1 0 1 6 / j . j c i n . 2 0 1 5 . 1 2 . 0 2 5
Combination of the Thermodilution-Derived Index of MicrocirculatoryResistance and Coronary Flow ReserveIs Highly Predictive of MicrovascularObstruction on Cardiac MagneticResonance Imaging After ST-SegmentElevation Myocardial Infarction
Sung Gyun Ahn, MD, PHD,a,b Olivia Y. Hung, MD, PHD,b Jun-Won Lee, MD,a Ji Hyun Lee, MD,a Young Jin Youn, MD,aMin-Soo Ahn, MD,a Jang-Young Kim, MD, PHD,a Byung-Su Yoo, MD, PHD,a Seung-Hwan Lee, MD, PHD,a
Junghan Yoon, MD, PHD,a Woocheol Kwon, MD, PHD,c Habib Samady, MDa
ABSTRACT
Fro
Se
Sch
Se
da
Vo
rel
Ma
OBJECTIVES The aim of this study was to investigate the predictive accuracy of invasive coronary microvascular
indexes for identifying microvascular obstruction (MVO) on cardiac magnetic resonance imaging (CMR) in patients treated
with primary percutaneous coronary intervention (PCI) for ST-segment elevation myocardial infarction (STEMI).
BACKGROUND We hypothesized that a combination of the index of microcirculatory resistance (IMR) and the
thermodilution-derived coronary flow reserve (CFRthermo) will enhance the predictive accuracy of detecting MVO
compared with either index alone.
METHODS The IMR and CFRthermo were measured using a single pressure sensor/thermistor-tipped guidewire in
40 STEMI patients immediately after PCI and related to MVO assessed by CMR day 7. The primary endpoint was the
predictive accuracy of the IMR for detecting MVO.
RESULTS Patients with an IMR >36 (upper tertile) had a higher rate of MVO compared with those with an IMR #36
(93% vs. 39%; p ¼ 0.001). MVO occurred in all patients with an IMR >36 and a CFRthermo #1.7 and in no patients with an
IMR #36 and a CFRthermo >1.7. The IMR remained an independent predictor of MVO (odds ratio: 1.212, 95% confidence
interval [CI]: 1.004 to 1.464; p ¼ 0.045) after adjustment for age, creatine kinase-myocardial band, myocardial blush
grade, thrombus burden, and CFRthermo. Both the IMR (area under the curve, 0.868, 95% CI: 0.719 to 0.956; p ¼ 0.001)
and the CFRthermo (area under the curve, 0.706, 95% CI: 0.536 to 0.842; p ¼ 0.03) were predictive of MVO. Combined
IMR and CFRthermo increased the area under the curve for MVO to 0.941.
CONCLUSIONS In patients who underwent primary PCI for STEMI, an increased IMR has an independent predictive
value for MVO detection, and combined high IMR and low CFRthermo are highly predictive of MVO. These indexes could be
used to further risk-stratify patients and guide regional and systemic therapies. (J Am Coll Cardiol Intv 2016;9:793–801)
© 2016 by the American College of Cardiology Foundation.
m the aDivision of Cardiology, Department of Internal Medicine, Yonsei University Wonju College of Medicine, Wonju
verance Christian Hospital, Wonju, Republic of Korea; bDivision of Cardiology, Department of Medicine, Emory University
ool of Medicine, Atlanta, Georgia; and the cDepartment of Radiology, Yonsei University Wonju College of Medicine, Wonju
verance Christian Hospital, Wonju, Republic of Korea. Dr. Ahn has received support from the CardioVascular Research Foun-
tion, Korea. Dr. Hung has received an NRSA training grant (5T32HL007745). Dr. Samady has received research funding from
lcano Corporation, St. Jude Medical, Medtronic, Inc., and Abbott Vascular. All other authors have reported that they have no
ationships relevant to the contents of this paper to disclose. Drs. Samady and Kwon contributed equally to this work.
nuscript received October 22, 2015; revised manuscript received December 17, 2015, accepted December 17, 2015.
ABBR EV I A T I ON S
AND ACRONYMS
AUC = area under the curve
CFIp = pressure-derived
collateral flow index
CFRthermo = thermodilution-
derived coronary flow reserve
CI = confidence interval
CMR = cardiac magnetic
resonance imaging
Cw = coronary wedge pressure
FFR = fractional flow reserve
IMR = index of microcirculatory
resistance
MVO = microvascular
obstruction
OR = odds ratio
PCI = percutaneous coronary
intervention
Pa = mean aortic pressure
Pd = mean distal coronary
pressure
STEMI = ST-segment elevation
myocardial infarction
TEI = transmural extent of
infarction
TIMI = Thrombolysis In
Myocardial Infarction
Tmn = mean transit time
Ahn et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6
IMR, CFR, and MVO After Reperfused STEMI A P R I L 2 5 , 2 0 1 6 : 7 9 3 – 8 0 1
794
M icrovascular obstruction (MVO)occurs frequently in patientswith ST-segment elevation myo-
cardial infarction (STEMI) even after timelyrevascularization of the culprit artery (1,2).MVO is associated with the occurrence ofcardiac death, congestive heart failure, andmyocardial reinfarction (1,3). Cardiac mag-netic resonance imaging (CMR) is consideredthe optimal diagnostic imaging modality todetect MVO (4). CMR also provides clinicianswith multifaceted information including car-diac function, chamber sizes, regional wallmotion abnormalities, intramyocardial hem-orrhage, infarct size, transmural extent ofinfarction (TEI), and myocardial salvage(4,5). Among them, MVO is regarded as themost powerful prognostic indicator of poorclinical outcomes and provides incrementalvalue beyond clinical risk and left ventricularfunction assessments (1,3,4).
SEE PAGE 802
Despite these advantages, CMR is notfrequently used clinically to risk-stratify pa-tients after STEMI due to the cost of CMR andconcerns about the safety of performing CMRimmediately after STEMI when the findingsare most helpful. Indeed, early left ventric-
ular remodeling, reperfusion injury, and microvas-cular edema resulting in dynamic microvascularchanges begins during the acute phase of STEMI (6,7).As a result, numerous invasive physiological indexeshave been developed and evaluated in STEMI pa-tients (8–17).
The index of microcirculatory resistance (IMR)and thermodilution-derived coronary flow reserve(CFRthermo) can be measured quickly and simplyusing a single pressure sensor/thermistor-tippedguidewire in the catheterization laboratory immedi-ately after primary percutaneous coronary interven-tion (PCI). An increased IMR, a marker of puremicrovascular dysfunction, after STEMI has beenassociated with the lack of recovery of regionalwall motion abnormalities (13) and mortality (16).Reduced CFRthermo reflecting diminished epicardialand microvascular flow has also been associatedwith adverse cardiac events in STEMI patients afterprimary PCI (18). We hypothesized that the combi-nation of the IMR and CFRthermo has incrementalvalue for predicting MVO compared with either in-dex alone. Accordingly, the aim of this study was toinvestigate the incremental predictive values of the
IMR, CFRthermo, and the combination of both todetect MVO on CMR in patients treated with primaryPCI for STEMI.
METHODS
STUDY POPULATION. Between January 2011 andJune 2013, we prospectively screened 529 patientswith de novo STEMI undergoing primary PCI within12 h after the onset of chest pain. Exclusion criteriaincluded the following: no invasive physiologicalassessment with pressure-temperature guidewires,Killip class 3 or higher, hemodynamic instabilityrequiring hemodynamic support devices, finalThrombolysis In Myocardial Infarction (TIMI) flowgrade of 3 not achieved after PCI, atrial fibrillation,chronic kidney disease (creatinine $2 mg/dl), historyof bleeding diathesis or known coagulopathy, severeinfection state, and known allergy to gadolinium. Ofthe remaining 72 patients in whom the IMR andCFRthermo were measured, patients who either didnot undergo CMR or had poor-quality CMR wereexcluded. The final study cohort comprised of 40STEMI patients who underwent both cor-onary microvascular assessments with a pressure/thermistor-tipped sensor and had adequate-qualityCMR (Figure 1). The study protocol was approved bythe Ethical Review Board of Yonsei University WonjuCollege of Medicine (Wonju, South Korea). Writteninformed consent was obtained from all participants.
PCI PROCEDURES AND ANGIOGRAPHIC ANALYSES.
All patients received 300 mg aspirin and 300 to600 mg clopidogrel immediately after STEMI diag-nosis by electrocardiogram. After an intravenousbolus injection of unfractionated heparin (70 U/kg),intravenous infusion (1,000 U/h) was continued, and,if necessary, additional boluses were administered toachieve an activated clotting time of 300 s. Afterpassing through the lesion with a 0.014-inch guide-wire, thrombus aspiration was performed using a 6-FThrombuster II (Kaneka Medical Products, Osaka,Japan) in selected patients at the physician’s discre-tion. Subsequently, balloon pre-dilation followed bystent deployment was performed. Pre-dilation wasperformed in all patients using an undersized balloonwith below nominal pressure. Application of glyco-protein IIb-IIIa inhibitor, access route (femoral vs.radial), use of intracoronary imaging modalities, andpost-adjuvant ballooning was dictated by the opera-tor’s preference. Quantitative coronary angiographywas performed offline using a primary diagnosticimage review and analysis workstation (Centricity
FIGURE 1 Study Flow Chart
CMR ¼ cardiac magnetic resonance imaging; TIMI ¼ Thrombolysis In Myocardial Infarction.
J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6 Ahn et al.A P R I L 2 5 , 2 0 1 6 : 7 9 3 – 8 0 1 IMR, CFR, and MVO After Reperfused STEMI
795
Cardiology CA1000, GE Healthcare, Milwaukee,Wisconsin) by a technician blinded to the clinicalinformation. Antegrade coronary blood flow in theinfarct-related artery was evaluated using TIMI flowgrade (8). Myocardial blush was graded based on thecoronary angiogram (10), and coronary thrombus wasclassified into 5 grades by the operator (9).
PHYSIOLOGICAL CORONARY AND MICROVASCULAR
ASSESSMENT. After successful primary PCI, a cali-brated coronary pressure guidewire (St. Jude Medical,Minneapolis, Minnesota) was equalized to the guid-ing catheter pressure with the sensor positioned atthe ostium of the guiding catheter and then advancedbeyond the culprit lesion. All pressure tracings wererecorded on the RadiAnalyzer Xpress (St. Jude Med-ical) for offline analysis. Measurements of the meantransit time (Tmn) at baseline and during hyperemiausing a thermodilution technique have previouslybeen described (19). Briefly, 3 injections of saline(3 ml at room temperature) were administered to thecoronary artery, and the baseline Tmn was measured.Pharmacological hyperemia was then induced byintravenous infusion of adenosine (140 mg/kg/min),and 3 more injections of saline were administered tomeasure the hyperemic Tmn. The coronary wedgepressure (Cw) was obtained after 30 s of balloon oc-clusion within the stented segment. CFRthermo wascalculated as the ratio of the baseline Tmn to the hy-peremic Tmn. Fractional flow reserve (FFR) wasdefined as the ratio of mean distal coronary pressure(Pd) to mean aortic pressure (Pa) at maximal hyper-emia. The IMR (mm Hg.s or U [units]) was defined assimultaneously measured Pd multiplied by the hy-peremic Tmn. The corrected IMR was calculated bythe formula: IMR ¼ Pa � hyperemic Tmn [(Pd � Cw)/(Pa � Cw)] (20). The pressure-derived coronarycollateral flow index (CFIp) was calculated as the ratioof Cw to Pa (Figure 2).
CMR PROTOCOL AND ANALYSIS. CMR was per-formed on day 7 (median, day 7; range, days 5 to 11)using a 3-T magnetic resonance imaging system(Achieva Release 2.1, Philips Medical Systems, Eind-hoven, the Netherlands) equipped with a dedicatedcardiac software package, cardiac coil, and vector-cardiogram. After the acquisition of localizing images,we obtained long- and short-axis cine images usingretrospectively gated breath-hold true fast imagingwith a steady-state free precession technique. Theshort-axis cine scans of 10-mm slices were usedto determine the left ventricular mass, volume, andfunction. A bolus of contrast medium (gadoliniumdiethylenetriamine pentaacetic acid) was injected at adose of 0.1 mmol/kg, and images were acquired for
60 heart beats immediately after contrast infusion.Delayed enhancement images were then obtained byacquiring an inversion-recovery segmented gradientecho T1-weighted sequence 10 to 15 min after thebolus. All post-processing and analyses of theleft ventricular mass, volume, function, myocardialinfarct size, and presence of MVO were performedusing the Extended Brilliance Workstation (PhilipsMedical Systems) by a radiologist experienced in CMRand blinded to all clinical and invasive physiologicaldata. Infarct size was assessed manually by planim-etry on each short-axis slice, delineating the hyper-enhanced area, including areas of hypoenhancementsurrounded by the hyperenhanced area, the latterwas considered MVO. Infarct size, as a percentage ofleft ventricular mass, was computed from the sum ofhyperenhanced pixels from each of the 10 short-axisimages divided by the total number of pixelswithin the left ventricular myocardium multipliedby 100% (21). TEI was calculated as the average TEIof all segments with evidence of infarction in a17-segment model.
MAIN OUTCOME MEASURES. The primary endpointwas predictive accuracy of the IMR for detecting MVOon CMR day 7 after STEMI. The incremental predictive
FIGURE 2 Representative Cases of Invasive Microvascular Evaluation and CMR
(A) The index of microvascular resistance (IMR) is 24 U (75 mm Hg, the distal mean coronary pressure, is multiplied by 0.32 s, the mean hyperemic mean transit time
[Tmn]). Thermodilution-derived coronary flow reserve (CFRthermo) is 2.2 (0.70 s, the baseline Tmn, is divided by 0.32 s, the hyperemic Tmn). (B) Pressure-derived collateral
flow index (CFIp) is 0.21 (6 mm Hg, the coronary wedge pressure [Cw], is divided by 76 mm Hg, the mean aortic pressure). (C) Delayed gadolinium-enhanced cardiac
magnetic resonance imaging (CMR) shows 50% transmural hyperenhancement without microvascular obstruction (MVO) in the middle anteroseptal wall and apex.
(D to F) IMR, CFRthermo, Cw, and CFIp are 56 U, 1.1, 37 mm Hg, and 0.37, respectively. CMR indicates 100% hyperenhanced extent of the left ventricular wall with MVO
(arrowheads) in the anterior, anteroseptal, and apical segments.
Ahn et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6
IMR, CFR, and MVO After Reperfused STEMI A P R I L 2 5 , 2 0 1 6 : 7 9 3 – 8 0 1
796
value of the combined IMR and CFRthermo for identi-fying patients with MVO was also compared witheither index alone. The secondary endpoints wereindividual CMR prognosticators such as MVO, infarctsize, and TEI based on the IMR and CFRthermo values.
STATISTICAL ANALYSES. All continuous variablesare presented as mean � SD and analyzed usingStudent’s t test or analysis of variance. Categoricalvariables are presented as frequencies (percentage)and analyzed using the chi-square or Fisher exacttest. Baseline clinical, angiographic, and procedural
parameters and invasive physiological indexeswere compared according to the presence of MVO.The univariate binary logistic regression analyseswere done to investigate the relationship of MVO withclinical and laboratory variables, echocardiographicand angiographic parameters, and invasive physio-logical indexes. Multiple logistic regression analysiswas performed to assess the independent associationof the IMR with MVO after adjustment for variableson the basis of the best results of the univariateregression analyses at the significance level of p < 0.1.Receiver-operating characteristic curve analysis was
TABLE 1 Baseline Clinical and Laboratory Characteristics According to MVO
Total(N ¼ 40)
MVO (�)(n ¼ 17)
MVO (þ)(n ¼ 23) p Value
Age, yrs 56 � 9 60 � 9 53 � 8 0.013
Male 37 (93) 15 (88) 22 (96) 0.379
BMI, kg/m2 24.1 � 3.2 24 � 3.1 24.2 � 3.3 0.845
DM 28 (70) 11 (65) 17 (74) 0.53
Hypertension 31 (78) 12 (71) 19 (83) 0.386
Dyslipidemia 38 (95) 16 (94) 22 (96) 0.826
Current smoker 24 (60) 9 (53) 15 (65) 0.366
Ischemic time, min 276 � 116 232 � 123 304 � 105 0.131
Killip classification 0.826
1 38 (95) 16 (94) 22 (96)
2 2 (5) 1 (6) 1 (4)
LVEF, % 51 � 8 52 � 7 51 � 8 0.704
E/E0 11 � 3 10 � 3 11 � 3 0.168
Peak CK-MB, ng/ml 201.9 � 127 112.2 � 102.2 268.2 � 101 <0.001
hs-CRP, mg/l 0.5 � 0.8 0.6 � 0.8 0.5 � 0.9 0.595
BNP, ng/l 39 � 70.7 45.5 � 71.4 34.2 � 71.3 0.623
Values are mean � SD or n (%).
BMI ¼ body mass index; BNP ¼ B-type natriuretic peptide; CK-MB ¼ creatine kinase-myocardial band;DM ¼ diabetes mellitus; hs-CRP ¼ high-sensitivity C-reactive protein; LVEF ¼ left ventricular ejection fraction;MVO ¼ microvascular obstruction.
TABLE 2 Baseline Angiographic and Procedural Parameters According to MVO
Total(N ¼ 40)
MVO (�)(n ¼ 17)
MVO (þ)(n ¼ 23)
pValue
Target vessel 0.438
Left anterior descending artery 28 (70) 12 (71) 16 (70)
Left circumflex artery 5 (13) 1 (6) 4 (17)
Right coronary artery 7 (17) 4 (23) 3 (13)
No. of obstructive vessels 1 � 0.2 1 1.1 � 0.3 0.426
Baseline TIMI 0/1 33 (83) 11 (65) 22 (96) 0.033
Collateral flow grade 0/1 34 (85) 15 (88) 19 (83) 0.593
Thrombus score 4.3 � 1.3 3.8 � 1.6 4.7 � 0.7 0.026
Aspiration thrombectomy 31 (78) 13 (77) 18 (78) 0.893
Glycoprotein IIb-IIIa inhibitor 26 (65) 9 (53) 17 (74) 0.169
No. of stents per patient 1.1 � 0.3 1.1 � 0.3 1.1 � 0.3 0.775
Total stent length, mm 24.4 � 9.5 25.6 � 10 23.5 � 9.3 0.603
Stent diameter, mm 3.2 � 0.4 3.4 � 0.5 3.1 � 0.3 0.039
Post-dilation 31 (78) 10 (90) 21 (93) >0.999
Multivessel stenting 1 (3) 0 1 (6) >0.999
Final TIMI flow grade 3 40 (10) 17 (100) 23 (100) 1
Final myocardial blush grade 2/3 33 (83) 17 (100) 16 (70) 0.014
Distal embolization 2 (5) 0 2 (9) 0.499
Values are n (%) or mean � SD.
MVO ¼ microvascular obstruction; TIMI ¼ Thrombolysis In Myocardial Infarction.
J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6 Ahn et al.A P R I L 2 5 , 2 0 1 6 : 7 9 3 – 8 0 1 IMR, CFR, and MVO After Reperfused STEMI
797
used to predict MVO according to the IMR andCFRthermo values and the combination of both. Theoccurrence of MVO was compared, grouped by theupper tertile values of both the IMR (IMR, 36) andCFRthermo (CFRthermo, 1.7). The threshold for signifi-cance was defined as p value <0.05. All analyses werecarried out using Statistical Package for the SocialSciences versions 20.0 (SPSS Inc., Chicago, Illinois).
RESULTS
BASELINE DEMOGRAPHIC AND CLINICAL CHARAC-
TERISTICS. Baseline clinical and laboratory charac-teristics between the 2 cohorts were balanced withrespect to sex, history of diabetes mellitus, hyper-tension, and hyperlipidemia or smoking, ischemictime, Killip classification, left ventricular ejectionfraction, E/E0, high-sensitivity chronic reactive pro-tein, and B-type natriuretic peptide (Table 1). How-ever, patients with MVO were younger (53 � 8 vs. 60 �9 years of age; p ¼ 0.013) and had higher peak creatinekinase-myocardial band (268.2 � 101 vs. 112.2 �102.2; p < 0.001) than those without MVO.
ANGIOGRAPHIC AND PROCEDURAL PARAMETERS.
Most patients had single-vessel disease, with the leftanterior descending artery determined to be theinfarct-related artery in 70% of patients (Table 2). Thetarget vessel distribution and collateral flow gradewere similar between the 2 groups. No difference wasobserved in the use of glycoprotein IIb-IIIa inhibitor,application of aspiration thrombectomy, number ofstents, total stent length, mean stent diameter,adjuvant ballooning, or incidence of distal emboliza-tion. However, patients with MVO had a higher per-centage of baseline TIMI flow grade 0/1 (96% vs. 65%;p ¼ 0.033), lower rate of final myocardial blush grade2/3 (70% vs. 100%; p ¼ 0.014), and larger thrombusburden (4.7 � 0.7 vs. 3.8 � 1.6; p ¼ 0.026) than thosewithout MVO.
CORONARY AND MICROVASCULAR INDEXES, AND
CMR FINDINGS. Patients with MVO had slower hy-peremic Tmn (0.57 � 0.20 s vs. 0.28 � 0.13 s; p < 0.001),a higher IMR (42 � 20 vs. 20 � 8, p < 0.001) and lowerCFRthermo values (1.4 � 0.5 vs. 1.9 � 0.9; p ¼ 0.035)compared with those without MVO (Table 3). Nodifference was found in hemodynamic parameters,FFR, Cw, the CFIp, the rate of ST-segment resolution>70%, or left ventricular ejection fraction. Infarct size(29 � 10% vs. 18 � 13%; p ¼ 0.008) and TEI (80 � 24%vs. 53 � 32%; p ¼ 0.003) were larger in patients withMVO. Patients with TEI $75% had lower CFRthermo
compared with those with TEI <75% (1.4 � 0.6 vs.1.9 � 0.8; p ¼ 0.025). In contrast, the IMR was not
significantly different according to TEI (34 � 22 vs.30 � 15; p ¼ 0.490).
Patients with an IMR >36 had higher rate of MVO(93% vs. 39%; p ¼ 0.001). Patients with CFRthermo #1.7had a tendency toward higher MVO occurrence (67%vs. 39%, p ¼ 0.091). When patients were stratified bythe upper tertile values of both the IMR (36) andCFRthermo (1.7), MVO occurred in all patients with an
TABLE 3 Invasive Physiological Indexes and Other Myocardial Perfusion
Parameters According to MVO
Total(N ¼ 40)
MVO (�)(n ¼ 17)
MVO (þ)(n ¼ 23) p Value
Cath lab parameters
Angiographic reflow(TIMI flow grade 3and MBG 2/3)
33 (83) 17 (100) 16 (70) 0.014
Baseline Pa, mm Hg 88 � 12 90 � 12 86 � 11 0.445
Hyperemic Pa, mm Hg 77 � 11 79 � 9 75 � 8 0.244
Hyperemic Pd, mm Hg 73 � 16 77 � 15 70 � 17 0.288
Hyperemic heart rate, beats/min 86 � 10 86 � 11 86 � 8 0.604
Baseline Tmn, s 0.66 � 0.43 0.55 � 0.40 0.75 � 0.45 0.189
Hyperemic Tmn, s 0.45 � 0.23 0.28 � 0.13 0.57 � 0.20 <0.001
FFR 0.93 � 0.05 0.93 � 0.05 0.93 � 0.05 0.892
IMR, mm Hg/s 33 � 19 20 � 8 42 � 20 <0.001
IMRcor, mm Hg/s 17 � 11 10 � 5 22 � 12 0.001
CFRthermo 1.6 � 0.7 1.9 � 0.9 1.4 � 0.5 0.035
Cw, mm Hg 27 � 8 25 � 8 28 � 7 0.168
CFIp 0.30 � 0.09 0.28 � 0.07 0.32 � 0.09 0.127
ST-segment resolution >70% 17 (43) 7 (41) 10 (44) 0.884
Peak CK-MB, ng/ml 201.9 � 127 112.2 � 102.2 268.2 � 101 <0.001
CMR parameters
Ejection fraction, % 51 � 8 53 � 6 50 � 9 0.138
Infarct size, % 24 � 13 18 � 13 29 � 10 0.008
Transmural extent of infarction, % 69 � 33 53 � 32 80 � 24 0.003
Values are n (%) or mean � SD.
CFIp ¼ pressure-derived collateral flow index; CFRthermo ¼ thermodilution-derived coronary flow reserve;CK-MB ¼ creatine kinase-myocardial band; CMR ¼ cardiac magnetic resonance imaging; Cw ¼ coronary wedgepressure; FFR ¼ fractional flow reserve; IMR ¼ index of microcirculatory resistance; IMRcor ¼ corrected IMR;MBG ¼ myocardial blush grade; Pa ¼ mean aortic pressure; Pd ¼ mean distal coronary pressure; Tmn ¼ meantransit time; other abbreviations as in Tables 1 and 2.
FIGURE 3 Incidence of MVO, Grouped by IMR and CFRthermo
When stratifying patients by the upper tertile values of both the IMR (36) and CFRthermo
(1.7), MVO occurred in all patients with an IMR >36 and CFRthermo #1.7, whereas those with
an IMR #36 and CFRthermo >1.7 had no MVO (p < 0.001). Abbreviations as in Figure 2.
Ahn et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6
IMR, CFR, and MVO After Reperfused STEMI A P R I L 2 5 , 2 0 1 6 : 7 9 3 – 8 0 1
798
IMR >36 and CFRthermo #1.7, whereas those with anIMR #36 and CFRthermo >1.7 had no MVO (Figure 3).Basal Tmn was the longest in patients with an IMR>36 and CFRthermo >1.7 (1.4 � 0.22 s) and the shortestin those with an IMR #36 and CFRthermo #1.7 (0.38 �0.16 s) than other groups (p < 0.001). Hyperemic Tmn
was the fastest in patients with an IMR #36 andCFRthermo >1.7 (0.28 � 0.13 s) than other groups(p < 0.001) (Table 4).
PREDICTORS OF THE PRESENCE OF MVO. Univariatelogistic regression analysis demonstrated that age(odds ratio [OR]: 0.906, 95% confidence interval [CI]:0.833 to 0.986; p ¼ 0.022], peak CK-MB level (OR:1.015, 95% CI: 1.007 to 1.024), myocardial blush grade(OR: 0.197, 95% CI: 0.057 to 0.678), thrombus burden(OR: 2.045, 95% CI: 0.996 to 4.199), CFRthermo (OR:0.331, 95% CI: 0.116 to 0.943), and IMR (OR: 1.151, 95%CI: 1.047 to 1.265) correlated with the presence ofMVO on CMR. The IMR (OR: 1.212, 95% CI: 1.004 to1.464; p ¼ 0.045) was independently associated withMVO after adjustment for age, CK-MB, myocardialblush grade, thrombus burden, and CFRthermo. Onreceiver-operating characteristic curve analyses(Figure 4), both IMR (area under the curve [AUC]:0.868, 95% CI: 0.719 to 0.956; p < 0.001) andCFRthermo (AUC: 0.706, 95% CI: 0.540 to 0.871;p ¼ 0.031) were predictive of MVO. The optimal cutoffvalues of the IMR and CFRthermo for predicting MVOwere 27 (sensitivity of 74% and specificity of 88%)and 1.6 (sensitivity of 71% and specificity of 61%),respectively. Taking the combined IMR and CFRthermo
into consideration, AUC improved to 0.941 (p ¼ 0.17for the IMR; p ¼ 0.005 for CFRthermo).
DISCUSSION
This study evaluated the usefulness of the IMR andCFRthermo for predicting MVO on CMR in patientstreated with primary PCI for STEMI. The main find-ings include the following: 1) MVO occurred mostfrequently in patients with a higher IMR and lowerCFRthermo values; 2) the IMR was the only indepen-dent predictor of MVO among univariate correlatesincluding CFRthermo, myocardial blush grade, andpeak CK-MB; and 3) combined IMR and CFRthermo
values (AUC: 0.941) had greater predictive value fordetecting MVO compared with the IMR (AUC: 0.868)or CFRthermo (AUC: 0.706) alone.
Achieving adequate microvascular perfusion isa goal in the treatment of STEMI patients. Evenafter achieving an epicardial TIMI flow grade of 3,poor myocardial perfusion status has been associatedwith extensive myocardial infarction, left ventri-cular remodeling, and increased mortality (10).
TABLE 4 Hemodynamic and Thermodilution-Derived Indexes, Grouped by IMR and
CFRthermo
IMR >36,CFRthermo #1.7
(n ¼ 8)
IMR >36,CFRthermo >1.7
(n ¼ 6)
IMR #36,CFRthermo #1.7
(n ¼ 19)
IMR #36,CFRthermo >1.7
(n ¼ 7)p
Value
Hyperemic Pa, mm Hg 79 � 11 86 � 10 73 � 9 81 � 13 0.072
Hyperemic Pd, mm Hg 71 � 15 86 � 15 67 � 14 82 � 17 0.051
FFR 0.91 � 0.06 0.95 � 0.03 0.92 � 0.05 0.93 � 0.04 0.561
Baseline Tmn, s 0.82 � 0.22 1.4 � 0.22 0.38 � 0.16 0.69 � 0.23 <0.001
Hyperemic Tmn, s 0.75 � 0.13 0.57 � 0.13 0.37 � 0.18 0.28 � 0.13 <0.001
IMR 61 � 15 46 � 11 21 � 7 21 � 6 <0.001
CFRthermo 1.2 � 0.3 2.3 � 0.5 1.2 � 0.3 2.5 � 0.8 <0.001
Cw, mm Hg 31 � 5 32 � 5 26 � 8 22 � 5 0.066
CFIp 0.32 � 0.08 0.32 � 0.04 0.31 � 0.10 0.26 � 0.06 0.482
Values are mean � SD.
Abbreviations as in the Table 3.
FIGURE 4 ROC Analyses to Predict MVO
Both the IMR (p < 0.001) and CFRthermo (p ¼ 0.031) were
predictive of MVO. The optimal cutoff values of IMR and
CFRthermo in predicting MVO were 27 (sensitivity of 74% and
specificity of 88%) and 1.6 (sensitivity of 71% and specificity of
61%), respectively. Taking the combined IMR and CFRthermo into
consideration, AUC improved to 0.941 (p ¼ 0.17 for IMR; p ¼0.005 for CFRthermo). AUC ¼ area under the curve; CI ¼ confi-
dence interval; other abbreviations as in Figure 2.
J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6 Ahn et al.A P R I L 2 5 , 2 0 1 6 : 7 9 3 – 8 0 1 IMR, CFR, and MVO After Reperfused STEMI
799
Angiographic indicators such as TIMI flow grade (8),TIMI frame count (11), and myocardial blush grade(10) have been evaluated for determining the ade-quacy of coronary and microvascular reperfusion;however, these indexes are semiquantitative andhave considerable inter- and intraobserver disagree-ment (10). Guidewire-based invasive indexes such asCFR, the IMR, and hyperemic microvascular resis-tance index have emerged as more reproducible andquantitative measures of coronary microcirculationfunction (12,13,17).
Reduced CFRthermo (18,22), accelerated coronaryflow velocity diastolic deceleration time, and thepresence of systolic flow reversal (12) are poor prog-nostic indicators after acute myocardial infarction.Our findings that lower CFRthermo was related to MVOare consistent with these observations. However,CFRthermo has several limitations for microvascularassessment. First, it provides an assessment of theaggregate of epicardial and microvascular flow andtherefore cannot discriminate between the relativecontribution of epicardial and microvascular resis-tance. Second, because it is a ratio of hyperemic andbasal flow, it has greater variability.
Compared with CFRthermo, the IMR demonstratesless intrinsic variability and better reproducibilitybecause it does not rely as much on hemodynamicstatus (23). Our findings are in keeping with theexisting literature on the IMR and MVO, infarct size,recovery of left ventricular function, and also withmortality after myocardial infarction (15,16,24).Cuculi et al. (24) measured the IMR immediately afterprimary PCI and then performed serial assessments ofthe IMR and CMR on day 1 and at 6 months in 41STEMI patients and found that the IMR was notsignificantly different between the groups at6 months. They argue that coronary microcirculationbegins to recover within 24 h, and recovery pro-gresses further by 6 months. In addition to the IMR,Doppler-derived intracoronary physiology indexessuch as hyperemic microvascular resistance andzero-flow pressure (the calculated pressure at whichcoronary flow would cease) have recently beenshown to be associated with CMR-defined micro-vascular injury in STEMI patients after primary PCI(17,25,26). Our data add to the literature by con-firming the association of the IMR and MVO anddemonstrating the incremental value of assessingboth the IMR and CFRthermo in this setting. Mea-surements of intracoronary microvascular physio-logical indexes at the time of primary PCI mayprovide a unique opportunity for predicting MVOoccurrence and potentially applying therapeuticmeasures to ameliorate it.
A novel finding of the study is that the com-bined IMR and CFRthermo provide a comprehensiveassessment of coronary microvascular function andhas incremental predictive value for identifyingpatients with MVO above either the IMR or CFRthermo
alone. A high IMR and low CFRthermo concordantlydepict microvascular dysfunction, whereas a low
PERSPECTIVES
WHAT IS KNOWN? Even after timely revasculari-
zation of the culprit artery, CMR-defined MVO
develops in almost one-half of STEMI patients, which
is linked to poor outcome. Both the IMR and CFRthermo
can be measured using a single pressure sensor/
thermistor-tipped guidewire at the time of primary
PCI for STEMI. An increased IMR and reduced
CFRthermo correlate with the presence and severity of
MVO on CMR in STEMI patients.
WHAT IS NEW? A high IMR and low CFRthermo
concordantly depict microvascular dysfunction,
whereas a low IMR and high CFRthermo indicate normal
microcirculatory integrity. The combination of the IMR
and CFRthermo has incremental predictive value for
identifying patients with MVO above either the IMR
or CFRthermo alone.
WHAT IS NEXT? These invasive physiological indexes
provide not only prognostic information but may guide
regional and systemic novel therapeutic intervention to
ameliorate MVO and myocardial remodeling.
Ahn et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6
IMR, CFR, and MVO After Reperfused STEMI A P R I L 2 5 , 2 0 1 6 : 7 9 3 – 8 0 1
800
IMR and high CFRthermo indicate normal microcir-culatory integrity. With respect to discordant pairedvalues, patients with both a high IMR and CFRthermo
or both a low IMR and CFRthermo had intermediaterates of MVO (83% and 53%, respectively). Basaltransit time was longest in patients with both a highIMR and CFRthermo, indicating overall slower coronaryblood flow related to impaired microcirculation andmyocardium. In contrast, patients with both a low IMRand CFRthermo had shortest basal transit time, indi-cating that overall faster coronary blood flow is likelylinked to reactive hyperemia after reperfusion. Addi-tionally, the difference between Pd and Pa in patientswith both a low IMR and CFRthermo suggests greaterenergy loss in the flow of blood due to diffuse disease.
A single coronary guidewire with a pressure/temperature sensor tip also provides other physio-logical indexes including FFR, Cw, and the CFIp. FFRto assess residual epicardial stenosis of the culpritartery immediately after primary PCI is not recom-mended because numerous factors including MVO,myocardial edema, and reperfused myocardial masscan influence the FFR value (24). In stable coronaryartery disease, the CFIp is likely to represent theamount of recruitable collateral flow to the myocar-dium distal to the occluded artery (27). However, inthe STEMI setting, the CFIp is unlikely to reflectcollateral flow because increased microvascularresistance due to extensive myocardial vascularobstruction impedes ejection of myocardial bloodinto the venous system. Indeed, a higher CFIp isassociated with worse functional recovery after acutemyocardial infarction (28). These findings are inkeeping with our observations that Cw and the CFIptended to be higher in patients with MVO.
STUDY LIMITATIONS. This study was conducted ona limited patient population at a single center inSouth Korea. Our findings therefore could not beextrapolated to high-risk STEMI or cardiogenic shockpatients who might have even more extensive MVO.However, it would be difficult to consent these un-stable patients in detailed mechanistic investigationssuch as the present study.
CONCLUSIONS
Increased coronary microvascular resistance assessedby invasive physiological indexes (a high IMR and lowCFRthermo) was associated with MVO on CMR afterreperfused STEMI. Physiological evaluation by a sin-gle pressure sensor/thermistor-tipped guidewire mayallow risk stratification and potential selection ofpatients to undergo regional myocardial therapies.
REPRINT REQUESTS AND CORRESPONDENCE: Dr.Habib Samady, Emory University School of Medicine,1364 Clifton Road F622, Atlanta, Georgia 30322.E-mail: [email protected] OR Dr.Woocheol Kwon,Department of Radiology, Yonsei University WonjuCollege of Medicine, 20 Ilsan-ro, Wonju 220-701,Republic of Korea. E-mail: [email protected].
RE F E RENCE S
1. Hamirani YS, Wong A, Kramer CM, Salerno M.Effect of microvascular obstruction and intra-myocardial hemorrhage by CMR on LV remodelingand outcomes after myocardial infarction: a sys-tematic review and meta-analysis. J Am Coll Car-diol Img 2014;7:940–52.
2. Weir RA, Murphy CA, Petrie CJ, et al. Micro-vascular obstruction remains a portent of adverse
remodeling in optimally treated patients withleft ventricular systolic dysfunction after acutemyocardial infarction. Circ Cardiovasc Imaging2010;3:360–7.
3. van Kranenburg M, Magro M, Thiele H, et al.Prognostic value of microvascular obstructionand infarct size, as measured by CMR in STEMIpatients. J Am Coll Cardiol Img 2014;7:930–9.
4. Eitel I, de Waha S, Wohrle J, et al. Compre-hensive prognosis assessment by CMR imagingafter ST-segment elevation myocardial infarction.J Am Coll Cardiol 2014;64:1217–26.
5. Ortiz-Pérez JT, Lee DC, Meyers SN, Davidson CJ,Bonow RO, Wu E. Determinants of myocardialsalvage during acute myocardial infarction: eval-uation with a combined angiographic and CMR
J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 9 , N O . 8 , 2 0 1 6 Ahn et al.A P R I L 2 5 , 2 0 1 6 : 7 9 3 – 8 0 1 IMR, CFR, and MVO After Reperfused STEMI
801
myocardial salvage index. J Am Coll Cardiol Img2010;3:491–500.
6. Fraccarollo D, Galuppo P, Bauersachs J. Noveltherapeutic approaches to post-infarction remod-elling. Cardiovasc Res 2012;94:293–303.
7. Park HW, Yoon CH, Kang SH, et al. Early- andlate-term clinical outcome and their predictors inpatients with ST-segment elevation myocardialinfarction and non-ST-segment elevation myo-cardial infarction. Int J Cardiol 2013;169:254–61.
8. Chesebro JH, Knatterud G, Roberts R, et al.Thrombolysis in Myocardial Infarction (TIMI) Trial,phase I: a comparison between intravenous tissueplasminogen activator and intravenous streptoki-nase. Clinical findings through hospital discharge.Circulation 1987;76:142–54.
9. Gibson CM, de Lemos JA, Murphy SA, et al.Combination therapy with abciximab reduces an-giographically evident thrombus in acutemyocardial infarction: a TIMI 14 substudy. Circu-lation 2001;103:2550–4.
10. Henriques JP, Zijlstra F, van ’t Hof AW, et al.Angiographic assessment of reperfusion in acutemyocardial infarction by myocardial blush grade.Circulation 2003;107:2115–9.
11. Wong GC, Frisch D, Murphy SA, et al. Time forcontrast material to traverse the epicardial arteryand the myocardium in ST-segment elevationacute myocardial infarction versus unstable anginapectoris/non-ST-elevation acute myocardialinfarction. Am J Cardiol 2003;91:1163–7.
12. Yamamuro A, Akasaka T, Tamita K, et al.Coronary flow velocity pattern immediately afterpercutaneous coronary intervention as a predictorof complications and in-hospital survival afteracute myocardial infarction. Circulation 2002;106:3051–6.
13. Fearon WF, Shah M, Ng M, et al. Predictivevalue of the index of microcirculatory resistancein patients with ST-segment elevation myocardialinfarction. J Am Coll Cardiol 2008;51:560–5.
14. Lim HS, Yoon MH, Tahk SJ, et al. Usefulness ofthe index of microcirculatory resistance for inva-sively assessing myocardial viability immediatelyafter primary angioplasty for anterior myocardialinfarction. Eur Heart J 2009;30:2854–60.
15. McGeoch R, Watkins S, Berry C, et al. The indexof microcirculatory resistance measured acutelypredicts the extent and severity of myocardialinfarction in patients with ST-segment elevationmyocardial infarction. J Am Coll Cardiol Intv 2010;3:715–22.
16. Fearon WF, Low AF, Yong AS, et al. Prognosticvalue of the index of microcirculatory resistancemeasured after primary percutaneous coronaryintervention. Circulation 2013;127:2436–41.
17. Kitabata H, Kubo T, Ishibashi K, et al. Prog-nostic value of microvascular resistance indeximmediately after primary percutaneous coronaryintervention on left ventricular remodeling in pa-tients with reperfused anterior acute ST-segmentelevation myocardial infarction. J Am Coll CardiolIntv 2013;6:1046–54.
18. Takahashi T, Hiasa Y, Ohara Y, et al. Usefulnessof coronary flow reserve immediately afterprimary coronary angioplasty for acute myocardialinfarction in predicting long-term adverse cardiacevents. Am J Cardiol 2007;100:806–11.
19. Fearon WF, Balsam LB, Farouque HM, et al.Novel index for invasively assessing the coronarymicrocirculation. Circulation 2003;107:3129–32.
20. Aarnoudse W, Fearon WF, Manoharan G, et al.Epicardial stenosis severity does not affect mini-mal microcirculatory resistance. Circulation 2004;110:2137–42.
21. Thiele H, Kappl MJ, Conradi S, Niebauer J,Hambrecht R, Schuler G. Reproducibility of chronicand acute infarct size measurement by delayedenhancement-magnetic resonance imaging. J AmColl Cardiol 2006;47:1641–5.
22. Garot P, Pascal O, Simon M, et al. Impact ofmicrovascular integrity and local viability on left
ventricular remodelling after reperfused acutemyocardial infarction. Heart 2003;89:393–7.
23. Ng MK, Yeung AC, Fearon WF. Invasiveassessment of the coronary microcirculation:superior reproducibility and less hemodynamicdependence of index of microcirculatory resis-tance compared with coronary flow reserve.Circulation 2006;113:2054–61.
24. Cuculi F, De Maria GL, Meier P, et al. Impact ofmicrovascular obstruction on the assessment ofcoronary flow reserve, index of microcirculatoryresistance, and fractional flow reserve afterST-segment elevation myocardial infarction. J AmColl Cardiol 2014;64:1894–904.
25. Teunissen PF, de Waard GA, Hollander MR,et al. Doppler-derived intracoronary physiologyindices predict the occurrence of microvascularinjury and microvascular perfusion deficits afterangiographically successful primary percutaneouscoronary intervention. Circ Cardiovasc Interv 2015;8:e001786.
26. Patel N, Petraco R, Dall’Armellina E, et al.Zero-Flow Pressure Measured Immediately AfterPrimary Percutaneous Coronary Intervention forST-Segment Elevation Myocardial Infarction Pro-vides the Best Invasive Index for Predicting theExtent of Myocardial Infarction at 6 Months:An OxAMI Study (Oxford Acute MyocardialInfarction). J Am Coll Cardiol Intv 2015;8:1410–21.
27. Meier P, Gloekler S, Zbinden R, et al. Beneficialeffect of recruitable collaterals: a 10-year follow-up study in patients with stable coronary arterydisease undergoing quantitative collateral mea-surements. Circulation 2007;116:975–83.
28. Yamamoto K, Ito H, Iwakura K, et al. Pressure-derived collateral flow index as a parameter ofmicrovascular dysfunction in acute myocardialinfarction. J Am Coll Cardiol 2001;38:1383–9.
KEY WORDS blood flow velocity, cardiacmagnetic resonance imaging, coronarycirculation, myocardial infarction