Silent Myocardial Ischemia in Men With Systemic Hypertension and MMhout Clinical Evidence of

Coronary Artery Disease David Siegel, MD, MPH, Melvin D. Cheitlin, MD, Dana G. Seeley, MS,

Dennis M. Black, PhD, and Stephen B. Hulley, MD, MPH

Ths prevatence, characteristics and circadian pat- tern of dtent myocardial ischemia, and its associa- tion with ventrfcubr arrhythmias was studied in hypertensive men aged 35 to 70 years (mean 61) without clbdcal cardiac disease. Participants were withdrawn from dkuetic treatment and received 1 month of oral etectrotyte repletion wfth 40 mmol of potasshrm chloride, and 400 mg of magnesium oxide daily. Twenty-four-hour Hefter monitoring was then performed. Episodes of spent myocardial ischemia occurred in SO of 186 men (27%) and lasted from 2 to 289 minutes (mean 30 and medi- an 18). Statistical anafyds comparing the interval from midnight to 6 A.M. wfth each of the other three 6-hour time intervals revealed that parttci- pants were lass Hkefy to have dfent myocardial is&em& in this period (p <O.Ol for each compari- son) than at other tfmes of the day. There was lit- tle dHference in the proportion of men with a fre= quent or complex ventricular arrhythmia during the entire day or wiWn 1 hour of the spent myo- cardial lschemic episode (or during a comparable time period) comparing those with to those wfth- out sftent myocardial tschemia. These findings in- dicate that sfkmt myocardial ischemia occurs in approximately 25% of an older populatton of hy- pertenstve men without history of symptomatic cardiac disease. The circadtan pattern of frequen- cy of spent ischemic events in men free of clinical card&c disease is simibr to that reported for pa- tients with cardiac disease and cdncides with that reported for sudden death. There was no signffi- cant association between spent myocardial isch- emia and ventricular arrhythmias.

(Am1Cardtol1992;70:86-SO)

From the Division of Clinical Epidemiology, Department of Epidemiol- ogy and Biostatistics, and the Department of Medicine, University of Callfomia, San Francisco, California. This research was supported by Grant HL36821 and by National Heart, Lung, and Blood Institute Preventive Cardiology Academic Award HLO208 1 to Dr. Siegel from the National Institutes of Health, Bethesda, Maryland. Manuscript received December 3 1,199 1; revised manuscript received and accepted March 2,1992.

Address for reprints: David Siegel, MD, MPH, Prevention Sciences Group, 74 New Montgomery, San Francisco, California 94105.

T ransient myocardial ischemia in the absence of an- gina or other cardiac symptoms has been labeled silent myocardial ischemia and is usually ob-

served in patients with demonstrated coronary artery disease (CAD). I4 Myocardial ischemia, whether silent or symptomatic, is associated with changes in myocardi- al perfusion, wall motion abnormalities and hemody- namic evidence of myocardial impairment.2,5 It is also associated with an increased risk of death for several categories of CAD patients including those after myo- cardial infarction and with unstable angina.6-8 Less is known about the prevalence of silent myocardial isch- emia in patients with risk factors for CAD but without clinical cardiac disease, even though these persons are also at increased risk for cardiac death. Furthermore, the role of silent myocardial ischemia in the develop ment of lethal ventricular arrhythmias is not known, al- though studies have shown that the presence of ventric- ular arrhythmias on resting electrocardiogram or 24- hour Holter monitoring is associated with a significant increase in mortality, even in patients with no history of angina or myocardial infarction.9Jo If silent myocardial ischemia is associated with ventricular arrhythmias in asymptomatic but high-risk patients, intervention stud- ies designed to evaluate the benefit of reversing silent myocardial ischemia with the aim of decreasing ventrio- ular arrhythmias and thus presumably the risk of cardi- ac death would be indicated. We evaluated silent myo cardial ischemia on 24-hour ambulatory Holter moni- toring in hypertensive men who had been withdrawn from diuretic therapy, and repleted with oral potassium and magnesium for 1 month. Men with a history of an- gina or myocardial infarction were eliminated from analysis. This study design enabled us to evaluate the prevalence and pattern of silent myocardial ischemia, and its association with ventricular arrhythmias, in an asymptomatic population of hypertensive men free of electrolyte abnormalities and clinical cardiac disease.

METHODS Study sempk and design overview: Details of the

study design and participant selection process were pre- viously presennxll’ Briefly, the Hypertension Arrhyth- mia Reduction Trial is a clinical trial designed to assess the frequency and severity of ventricular arrhythmias associated with the use of various diuretic combinations. Hypertensive men aged 35 to 70 years with resting elec- trocardiographic abnormalities were selected as study subjects, because the Multiple Risk Factor Intervention Trial suggested that they may be especially at risk for

86 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 70 JULY 1, 1992

sudden death with diuretic use.12 The portion of the study described in this report was performed after diu- retic withdrawal and electrolyte repletion (see later). The study group was recruited both from clinical popu- lations and by a direct mail campaign. Subjects were included in the study if they were receiving diuretics for >6 months and had a diastolic blood pressure <95 mm Hg, or if they were not receiving diuretic therapy, but had history of hypertension and were either receiving nondiuretic antihypertensive drugs or had a diastolic blood pressure 290 but <105 mm Hg.

We excluded men who were receiving medications that may influence the development of ventricular ar- rhythmias, such as antiarrhythmic drugs, fi blockers, theophyllme and digitalis preparations, phenothiazines and tricyclic antidepressants. We also excluded men with history of myocardial infarction, congestive heart failure, angina pectoris, renal insufficiency (creatinine >2 mg/dl) or other serious illness (including psychiatric disability), or inability or unwillingness to give informed consent. Men with electrocardiographic findings (left ventricular hypertrophy and left bundle branch block) that would make interpretation of silent myocardial ischemia on Holter recordings difficult were also ex- cluded.

Subjects were withdrawn from diuretic treatment and received 1 month of oral electrolyte repletion with both 40 mm01 of potassium chloride and 400 mg of magnesium oxide (containing 241 mg of elemental magnesium) daily. Serum potassium and magnesium were then measured, and a continuous 24-hour Holter monitoring test was performed

continuaur Hot& monitaringr Continuous 24-hour Holter monitoring was performed using the Cardio techorder III dual lead system with a frequency re- sponse of 0.5 to 100 Hz, thereby meeting the American Heart Association’s specifications for heart rate and ST-segment changes. I3 The electrocardiogram was re- corded continuously on tape using a cassette system. The 2 lead systems used were V5 and Vi. Patient elec- trodes were selected to minimize patient discomfort, noise, skin-electrode impedance, polarization or other malfunction.

Tapes were sent to Cardio Data Systems (Haddon- field, New Jersey) for analysis. Whole tapes were print- ed at high speed on recording paper using the Cardio Data Corporation Mark IV Holter Analyzer. Specific areas of interest were printed at real time and evaluated by an analyst. Abnormal ST-segment changes were de- fmed as I1 mm of ST-segment depression occurring 80 ms after the J point, lasting for Ll minute and sepa- rated from other episodes by I1 minute. The onset and end of each episode was measured at the point of first deviation from the baseline.

For arrhythmia analysis, specific areas of interest (identified by irregularity in the QRS pattern) were printed at real time and evaluated by an analyst. Ven- tricular arrhythmias were classified (in a manner simi- lar to that of Lown et a1)14 according to the presence of the following arrhythmia types: ventricular premature complexes, multiform extrasystoles, ventricular cou- plets, ventricular tachycardia, and R-wave on T-wave

ventricular premature complexes, We created a summa- ry category of the presence of a frequent (130 ventricu- lar premature complexes/hour) or complex (presence in 24 hours of any of the following: multiform extrasys- toles, ventricular couplets or tachycardia, or R-wave on T-wave ventricular premature complexes) ventricular arrhythmia for 11 of 24 hours. The precision of these measurements is indicated by their within-individual correlation of ventricular arrhythmias on continuous 24- hour Holter monitoring before and after diuretic treat- ment; kappa coefficients for the whole cohort were 0.74 (95% confidence interval 0.61, 0.88) for the presence of 230 ventricular premature complexes/hour, and 0.47 (95% confidence interval 0.33,0.60) for the presence of a frequent or complex ventricular arrhythmia.

Statistical analysis: Silent myocardial ischemia, ventricular arrhythmias and mean heart rate were re- corded per hour for a 24-hour period. To determine if silent ischemic episodes were associated with the occur- rence of ventricular arrhythmias, we compared the fre- quency of ventricular arrhythmias that occurred in a 3- hour time span surrounding the silent ischemic event with that of a randomly assigned, frequency-matched (by diurnal distribution of silent ischemia) 3-hour time period in control subjects without silent ischemia. Pro- portions of participants with a complex or frequent ven- tricular arrhythmia in the 1 hour before the silent isch- emit event, in the 1 hour after the event and in a 3-hour time span (including the hour of the event) were com- pared between those with and without silent myocardial ischemia using chi-square tests of homogeneity.

The proportion of participants with silent myocardi- al ischemia in 11 hour of each 6-hour time block (6 A.M. to noon, noon to 6 P.M., 6 P.M. to midnight, and mid- night to 6 A.M.) was determined. The analysis compar- ing the risk of having silent myocardial ischemia at a given time of day to that at midnight to 6 A.M. was per- formed using a method that takes into account the with- in-person correlation among outcomes (silent ischemic episodes over time) for each participant. This approach (based on generalized linear models) regards the corre- lation among multiple observations in the same subject as a “nuisance” parameter. l5 We used a logit function to estimate regression coefficients and standard errors, and an exchangeable correlation matrix that assumes correlations, which represent the weighted average cor- relation among observed outcomes, are similar over timeel Associations are presented as odds ratios with 95% confidence intervals. Statistical analysis was per- formed using Statistical Analysis System software on an IBM 4341 mainframe computer.

RESULTS Study sampk The demographic and clinical char-

acteristics of the study participants are listed in Table I. The study sample included 186 men, of whom 50 (27%) had episodes of silent myocardial ischemia. There was no difference in age between men with and without si- lent ischemia (mean 61 years in both groups). Mean serum potassium was 4.3 mEq/liter, and mean serum magnesium was 2 mEq/liter in participants both with and without silent ischemia, indicating successful repls

SILENT MYOCARDIAL ISCHEMIA IN ASYMF’TOMATIC MEN 87

TABLE I Characteristics of Study Sample Comparing TABLE II Distribution of Ventricular Arrhythmias in Participants With and Without Silent Myocardial lschemia Participants With and Without Silent Myocardial lschemia

Silent Myocardial lschemia Silent Myocardial lschemia

Characteristic Present Absent

Number 50 136

Age (yr)* 61 (9) 61 (8) Race (%I

Asian 6 6 Black 6 15 Hispanic 4 5 Indian 0 1 White a4 72 Other 0 2

Weight (Ibs)* 181 (27) 192.3 (33)t Body mass index* 27 (3) 28 (4)t Systolic BP (mm Hg)* 144 (15) 145 (18) Diastolic BP (mm Hg)* 85 (10) 88 (10) Body surface area (m2)* 1.9 (0.2) 1.9 (0.2) Serum potassium* 4.3 (0.3) 4.3 (0.4) Serum magnesium* 2.0 (0.2) 2.0 (0.2) Alcohol (current, %I 72 66 Smoke (%)

Current 18 18 Past 60 54 Never 22 27

Antihypertensive other than diuretic (%) 30 34

wean (SO). tp co.05. BP - blood pressure.

Present Absent p Value

Number 50 136 Any complex or frequent 46 48 0.82

ventricular arrhythmia (%I* Complex or frequent ventricular

arrhythmia (%) Within * 1 hour of event 31 24 0.29 Within 1 hour after event 21 17 0.54 Within 1 hour before event 21 15 0.39

‘Complex or frequent ventricular arrhythmia defined as presence of any of the following: * 30 ventricular premature complexes/hour, or any of the following grades: multiform extrasystole, ventricular couplet or ventricular tachycardia.

25% of events lasted from 2 to 10 minutes, 50% were 118 minutes, and 75% were 138 minutes.

tion of these electrolytes. In terms of other characteris- tics, there was little difference between men with or without silent myocardial &hernia, except that those without silent myocardial ischemia were heavier and had a greater body mass index (p <0.05).

Figure 1 presents the hour-by-hour frequency of si- lent myocardial ischemic episodes. There was a consis- tent pattern of less silent myocardial ischemic episodes between midnight and 6 A.M. than during other times of the day. Statistical analysis comparing each of the other three 6-hour time intervals with that from midnight to 6 A.M. revealed that participants were less likely to have silent myocardial ischemia in this period (p <O.Ol) than at other times of day.

Echocardiograms were obtained (data not shown). There was no difference in the proportion of men with silent myocardial ischemia, between those who met echocardiographic criteria for left ventricular hypertro- phy (left ventricular mass index >134 g/m2) and those who did not.

Pattem of siknf myocardial ischnk episodes: There were 14 episodes of ischemia in 50 participants (range 1 to 4 per participant); most participants (64%) who had silent myocardial ischemia had only 1 episode in 24 hours (data not &own). Silent myocardial isch- emit episodes lasted from 2 to 289 minutes (mean 30);

Siknt myourdhl isdtemia and venbiculsr anlmyth- miss: There was little difference in the proportion of participants with a frequent or complex ventricular ar- rhythmia on 24-hour Holter monitoring between those with (46%) and without (48%) silent myocardial isch- emia (Table II). The proportion of men with a frequent or complex ventricular arrhythmia within 1 hour of the silent myocardial ischemic event was somewhat in- creased compared with that of a similar time period in participants without silent myocardial ischemia, but the differences were not statistically significant.

DISCUSSION Most prior studies of silent myocardial ischemia in-

cluded large proportions of participants with various cardiac disorders.1-4*6-8J7 In our study, men with history of angina or myocardial infarction, as well as those with electrocardiographic left ventricular hypertrophy, were eliminated from analysis, and all participants were re pleted with potassium and magnesium. This enabled us to study men without clinical cardiac disease, and to avoid the potential effects of deficiencies in potassium and magnesium on silent myocardial ischemia and ven- tricular arrhythmias.

m 12 5 5 10 .o

E 8 5 ur E 8

w4

:2

Midnight Noon

1

In our study, 27% of men had episodes of silent myo- cardial ischemia. This is a much higher proportion of participants with silent myocardial ischemia than that noted by Selwyn et all8 in a study of healthy normal subjects, where only 2 of 100 participants had horizon- tal ST-segment depression on 24-hour Holter monitor- ing. Their population included younger men and women (age range 20 to 50 years) without hypertension. Simi- lar findings to those of Selwyn were noted in another

88 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 70 JULY 1, 1992

study of healthy younger men and women.19 In another study, 11 of 17 asymptomatic men who had ischemic- type ST-segment depression (12 mm) during tread- mill exercise testing were found to have silent myocardi- al ischemia.20 Participants were older (mean age 62 years), and all but 1 had single or multiple cardiac risk factors, in addition to being men. All 11 participants with silent myocardial ischemia had significant CAD (150% stenosis) on angiography. These findings sug- gest that cardiac risk factors are associated with silent myocardial ischemia and that the presence of silent myocardial ischemia identifies a group of subjects at risk for CAD. Considerable day-to-day variability has been noted in the frequency of silent ischemic events in patients with CAD as recorded by Holter monitoring.21 Thus, our finding of 27% represents the minimum pro- portion of patients in our study with silent myocardial ischemia.

Myocardial infarction and sudden death do not oc- cur randomly throughout the day, but rather in a circa- dian pattern with an increased frequency in the mom- ing.22*23 Several physiologic processes such as heart rate, coronary blood flow, systemic pressure, platelet aggre- gability, plasma cortisol, tissue-type plasminogen activi- ty and plasma epinephrine have a similar pattem.23 We previously reported that ventricular arrhythmias be come more prevalent when subjects go from sleep to awakening.24 We now note that there is a similar circa- dian variation in silent myocardial ischemia. This pat- tern of silent myocardial ischemia was previously re- ported in men with clinical cardiac disease.3*2s,26 Our results extend these findings to men with hypertension but without symptomatic cardiac disease; they are of interest because many episodes of sudden death occur as the initial manifestation of cardiac disea~e.~~

Most studies have been unable to document a strong association between silent myocardial ischemia and ventricular arrhythmias.28 However, 1 study of 15 pa- tients who survived out-of-hospital ventricular fibrilla- tion found that 12 had exercise-induced silent myocar- dial ischemia in the absence of angina.29 Another study found that ventricular arrhythmias were associated with 18% of silent myocardial ischemic episodes in patients with unstable angina. I7 However, in this study there was no comparison with the frequency of ventricular ar- rhythmias during a similar time period in comparable patients without silent myocardial ischemia. In another recent report, ventricular arrhythmias were associated with ischemic episodes in 10 of 97 participants.26 Again, there was no comparison with the frequency of ventricu- lar arrhythmias during a similar time period in compa- rable patients without silent myocardial ischemia. We were unable to demonstrate a statistically significant difference in the proportion of participants with vent& ular arrhythmias between those with and without silent ischemia. However, our study lacks the power to detect small differences in ventricular arrhythmias between the 2 groups. If, as our study suggests, there is a 20% increase between the proportion of participants with a frequent or complex ventricular arrhythmia within 1 hour, either before or after the silent myocardial isch-

emit episode, compared with that of a similar time peri- od in participants without silent myocardial ischemia, a sample size of approximately 1,900 would be needed to detect such a small increased risk (based on a 2-sided significance level of 0.05, and a power of 80%, and as- suming that 25% of participants had silent ischemic epi- Mb).

Silent myocardial ischemia has been associated with a poor prognosis in patients with unstable angina and after infarction6-* Some investigators have suggested that patients with silent myocardial ischemia should be treated in the same manner as those with classic angi- na.30 Whether a similar prognosis is predicted by silent myocardiil ischemia in patients without clinical cardiac disease but with cardiac risk factors and whether these patients would benefit from medical or surgical inter- vention needs further study.

REFERERCES 1. Cohn PF. Silent myocardial ischemla in patients with defective angina1 warn- ing system. Am J Cardiol 1980;45:697-702. 2. Cecchi AC, Dorellini EV, Morchi F, Pucci P, Santoro GM, Fazzini PF. Silent mvocardial ischemia during ambulatorv electrocardioaranhic monitorine in na- t&s with effort angina. iAm Co11 dardiol 1983;3:53i-939. - ’ 3. Rmo MB, Bany J, Campbell S, Nabel E, Cook EF, Goldman L, Selwyn AP. Circadian variation-of tran.&t myocardii ischemia in patients with c&nary arterv dii. Chdarion 1987:75:395-400. 4. P&ley WW. Prevalence aid clinical significance of myocardiil ischemia. Circularion 1989;8O(suppl IV):IV-68-IV-73. 1. Deantield JE, Shea M, Ribiero P, de Landshcere CM, Wilson RA, Horlock P, Selwyn AP. Transient ST-segment depression as a marker of myocardial ischemia during daily life. Am J Cardioi 1984;54:1195-1200. 6. Gottlieb SO, Weisfeld ML, Ouyang P, Mellits ED, Gerstenblith G. Silent ischemia as a marker for early unfavorable outcomes in ptients with unstable angina. N Engl J Med 1986;314:1214-1219. 7. Got&b SO, Gottlieb SH, Achuff SC, Baumgardner R, Mellits ED, Weisfeldt ML, Gerstenblith G. Silent ischemia on Holter monitoring predicts mortality in high-risk p&infarction ptients. JAMA 1988;259:1030-1035. 0. Deedwania PC, Carbajal EV. Silent ischemia during daily life is an indepen- dent predictor of mortality in stable angina. Circulalion 1990,81:748-756. 9. Cullen K. Weame KL, Stenhouse NS, Cumpston GN. Q waves and ventricular extrasyxtoles in resting electrocardiograms. A 16 year follow-up in Busselton study. Br Heart J 1983;50:465-468. 10. Abdalla ISH, Princas RJ, Neaton JD, Jacobs DR Jr, Crow RS. Relation between ventricular premature complex~ and sudden cardiac death in apparently healthy men. Am J Cardiol 1987;60:1036-1042. 11. Siegel D, Hulley SB, Black DM, Cheitlm MD, Sebastian A, Seeley DG, Hearst N, Fine R. Diuretics, serum and intracellular electrolyte levels, and ven- tricular arrhythmias in hypertensive men. JAMA 1992;267:1083-1089. 12. Multiple Risk Factor Intervention Trial Research Group. Baseline rest elec- trocardicgraphic abnormalities, antihypertensive treatment, and mortality in the MuItiple Risk Factor Intervention Trial. Am J Cardiol 1985;55:1-15. 13. Sheffield LT, Bcxson A, Bragg-Rem&e1 D, Gillette PC, Hermes RE, Hinkle L, Kennedy H, Mirvis DM, Oliver C. Recommendations for standards of instru- mentation and practice in the use of ambulatory electrocardiography. Circulation 1985;11:626A-636A. 14. Lawn B, Calvert AF, Armington R, Ryan M. Monitoring for serious arrhyth- mias and high risk of sudden death. Circularion 1975;5l(suppl III):III-189-III- 198. IS. Zcger SL, Liing KY. Longitudinal data analysis for discrete and continuous outcomes. Biombics 1986;42:121-130. 16. Snedecor GW, Cc&an WG. Statistical Methods. 6th ed. Ames. Iowa: Iowa State University Press, 1967:185- 188. 17. Nademanee K. Intarachot V. Sinah PN. Jose&son MA. Sinah BN. Charac- , ~- teristlca and clinical signifmce if s&t myocard’utl ischemia in unstable angjna. Am J Cardid 1986:58:26B-338. 18. Selwyn AP, Shea M, Deanfield JE, Wilson R, Horlock P, O’Brien HA. Character of transient ischemia in angina pectoris. Am J Cardiol 1986;58:21 B- 25B. 19. Deanfleld JE, Ribiero P, Oakley K, Krikler S, Selwyn AP. Analysis of ST- segment change in normal subjects: implications for ambulatory monitoring in angina pectoris. Am J Cardiol 1984,54:1321-1325.

SILENT MYOCARDIAL ISCHEMIA IN ASYMPTOMATIC MEN 89

20. Coy KM, Imperi GA, Lambert CR, Pepine CJ. Silent myocardial ischemia during daily activities in asymptomatic men with positive exercise test responses. Am J Curdiol 198759~45-49. 21. Tzivoni D, Gavish A, Benhorin .I, Banai S, Kcren A, Stem S. Day-to-day variability of myocardiil ischemic episcdes in coronary artery dii. Am J Cardiol 1987;60:1003-1005. 22. Muller JE, Ludmer PL, Willich SN, Tofler GH, Aylmer G, Klangos I, Stone PH. Circadian variation in the frequency of sudden cardiac death. Circulation 1987;75:131-138. 23. Mulls JE, Toffler GH, Stone PH. Ciidiin variation and triggers of onset of acute cardiovascular d&case. Circulation 1989;79:733-743. 24. Siegel D, Black DM, Seeley DG, Hulley SB. Circadiin variation in ventricu- lar arrhythmias in hypertensive men. Am J Cordiol 1992;69:344-347. 25. Nademanee K, Intarachot V, Josephson MA, Singh BN. Cidian variation

in occurrence of transient overt and silent myocardial ischemia in chronic stable angina and comparison with Prinzmetal angina in men. Am J Cardiol1987;601 494498. 26. Hausmann D, Niitta P, Trappe H-J, Daniel WG, Wenzlaff P, Lichtlen PR. Incidence of ventricular arrhythmias during transient myccardial ischemia in patients with stable coronary artery dii. J Am Co11 Cordiol 1990;16:49-54. 27. Kaanel W. Prevalence and cliical aspects of unrecognized myocardial in- farction and sudden unexpected death. Circulation 1987;75(suppl 11):11-4-11-S. 21). Amsterdam EA. Silent myocardial ischemia, arrhythmias and sudden death: are they related? Am J Cordial 1987;59:919-920. 23. Sharma B, Asmger R, Francis GS, Hodges M, Wyeth RP. Painless ischemia in out of hospital ventricular fibrillation. Am J Cordial 1987;59:740-745. 30. Pepine CJ, Hill JA. Management of the total ischemic burden in angina pectoris. Am J Cardiol 1987;5937C-12C.

90 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 70 JULY 1, 1992