Exercise Thallium Tomography Predicts Future Clinically Manifest Coronary Heart Disease in a High-Risk Asymptomatic Population
Background Exercise testing, even when combined with radionuclide perfusion imaging, does not accurately predict future clinical coronary heart disease (CHD) in low-risk asymptomatic populations. We hypothesized that these tests would perform better in a higher-risk population with a high prevalence of occult CHD. Siblings of persons with premature CHD represent such a group in whom it would be advantageous to identify affected individuals before the occurrence of clinically manifest CHD.
Methods and Results Exercise thallium scintigraphy was performed in 264 asymptomatic individuals less than 60 years of age who had a sibling with documented CHD before age 60. Despite an average age of only 46 years at the time of screening, 19 of the 264 siblings developed clinical CHD (sudden death in 1, myocardial infarction in 10, coronary revascularization in 8) over a mean of 6.2 years (range, 1 to 9 years) of follow-up. Abnormal thallium scans were observed in 29% of men and 9% of women, while abnormal exercise ECGs occurred in 12% and 5%, respectively. Of men ≥45 years of age, 45% had an abnormal exercise ECG, thallium scan, or both. In contrast, only 3% of women <45 years of age had an abnormal test result. Although abnormal exercise ECGs and thallium scans were both predictive of future clinical CHD, the thallium scan was associated with a higher relative risk. After adjustment for age, sex, and exercise ECG results, the relative risk of developing clinical CHD was 4.7 for an abnormal scan. Siblings with a concordant abnormal exercise ECG and thallium scan had a relative risk of 14.5. These siblings were all men >45 years of age at the time of screening and had a strikingly high incidence of clinical CHD (6 of 12, 50%).
Conclusions Exercise thallium scintigraphy appears to be useful in the risk assessment of asymptomatic siblings of patients with premature CHD, particularly in male siblings who are 45 years of age or older.
Premature coronary heart disease (CHD) is known to cluster in families.1 2 3 4 Siblings of persons with CHD events prior to 60 years of age have an excess risk of developing symptomatic CHD that is twice as high in sisters and up to 12 times as high in brothers compared with the general population.5 6 7 Although the mechanisms of familial aggregation of CHD remain to be elucidated, a high prevalence of known coronary risk factors including hypertension, dyslipidemia, and cigarette smoking has been observed in our prior studies in apparently healthy siblings of persons with documented CHD prior to age 60 years.8
In patients with clinically apparent CHD, a positive exercise ECG and/or thallium scintigram is associated with an increased likelihood of subsequent coronary events, but in most asymptomatic populations, noninvasive exercise testing results in a low predictive value due to a low prevalence of CHD.9 10 11 We hypothesized that the predictive value of stress testing and thallium scintigraphy would be increased in an asymptomatic group of individuals with a family history of premature CHD because of the known propensity of this population to develop premature CHD events and because of their high prevalence of predictive risk factors. This study was designed to determine the extent to which detection of reduced myocardial perfusion by thallium scintigraphy and/or exercise-induced ST-segment changes on maximal graded treadmill testing predicted future CHD events in asymptomatic, apparently healthy siblings of persons with clinically manifest CHD prior to 60 years of age.
Sample and Recruitment
The Johns Hopkins Sibling Study is a prospective investigation, begun in 1983, of coronary risk factors in asymptomatic, apparently healthy siblings of persons with documented CHD prior to 60 years of age. The investigation was approved by the Johns Hopkins Institutional Review Committee, and participants gave informed consent. Siblings were identified at the time of a CHD event in an index patient. Index patients were eligible if they had experienced a documented myocardial infarction or anginal symptoms with at least one 50% diameter stenosis in one or more coronary arteries on angiography. Index patients with coronary disease associated with calcific aortic stenosis, collagen vascular disease, cardiac transplantation, or chronic glucocorticosteroid therapy were excluded. Index patients were recruited during specified weeks, during which all eligible patients on the medical and surgical units of the Johns Hopkins Hospital were identified. Recruitment was distributed throughout the week, Monday through Friday, and by season.
Siblings were eligible if they were less than 60 years of age and had no known history of coronary artery disease. Siblings were excluded if they had functional status limitations that precluded exercise testing, if they were receiving chronic or recent glucocorticosteroid therapy, had a collagen vascular disease, or had any comorbidity for which life expectancy was judged to be 5 years or less (for example, cancer). Siblings less than 30 years of age were included only if the index patient had experienced an event prior to 35 years of age. From 1983 until 1987, both male and female siblings were recruited, while from 1987 until the end of recruitment in 1991, only male siblings were included.
During hospitalization, index patients were asked for access to all siblings who did not have known coronary artery disease. Within a week of index patient discharge, eligible siblings were sent a description of the study and a refusal postcard to return if they preferred not to be contacted. Siblings were called within 2 weeks and asked to participate in screening that included a detailed cardiovascular history and physical examination, risk factor assessment, and a maximal graded treadmill test with thallium scintigraphy. A standardized health history questionnaire was administered via telephone to determine eligibility, and the absence of preexisting coronary disease was verified and validated with their primary physician. The health history was repeated by an attending cardiologist on the day of the screening and an anginal assessment and resting ECG were administered to again verify the absence of symptoms or a prior coronary disease event. After the screening, results were sent to the patient and his/her physician.
Maximal Graded Exercise Testing
All siblings underwent a maximal symptom-limited graded treadmill test with a modified Bruce protocol with increments in treadmill speed or inclination or both at 3-minute intervals. MET levels at each exercise stage were as follows: stage 1, 2.1; stage 2, 3.9; stage 3, 5.3; stage 4, 8.3; stage 5, 9.5; stage 6, 11.6; stage 7, 15.1; and stage 8, 16.2. Exercise was continued until the subject had to stop because of fatigue, dyspnea, dizziness, chest pain, or leg weakness or pain. In a few cases, exercise was terminated early by the supervising physician for severe hypertension or ventricular arrhythmias.
Maximal heart rate attained, total time in minutes on the treadmill, peak blood pressure, and MET level achieved were recorded. In men, a positive stress test was defined as horizontal or downsloping ST-segment depression of ≥1 mm over baseline at 0.06 seconds after the J-point in three or more consecutive beats at any time during the exercise test or during the first 3 minutes of recovery after exercise. To reduce the probability of a high rate of false-positive tests in women, an abnormal response was prospectively defined as ≥2.0-mm flat or downsloping ST-segment depression over baseline in leads II, III, or AVF or ≥1.5-mm ST depression in other leads. Two board-certified cardiologists, who were blinded to the risk factor status of the individual, reached a consensus as to whether the exercise test was positive or negative for ischemia.
Stress Thallium Scintigraphy
Thallium-201 scintigraphy was performed in conjunction with the maximal exercise treadmill test in 259 siblings. One minute before the subject anticipated exercise cessation, 3 mCi thallium-201 was injected through a peripheral intravenous line and flushed in rapidly. Five to 10 minutes after cessation of exercise, imaging was begun.
In 243 siblings, tomographic imaging was performed with a Technicare Omega 500 rotating large field of view camera interfaced to a Technicare 560 computer. Sixty 30-second images were acquired through a 180° circular orbit (30° right anterior oblique through 60° left posterior oblique). Raw images were obtained in a 128×128-byte mode with a 1.4 camera magnification and corrected for camera nonuniformity with a 128 million count flood field and for center-of-rotation variations with a line source calibration. Three hours later, delayed imaging was performed without reinjection of thallium.
Images were reconstructed by filtered back projection with the use of a ramp filter to form a set of contiguous transverse slices 3 pixels (6.2 mm) thick. Prior to reconstruction, the projection images underwent prefiltering with a two-dimensional Fourier (Wiener) filter and correction for translational motion. The Wiener filter was optimized for each patient by determining the object (heart) and noise power spectra on a selected projection image, combining these data with the known modulation transfer function of the imaging system, “spinning” the filter to create a two-dimensional circularly symmetrical filter, and applying the filter to all of the other projection images.12 Wiener filters have been shown to recover resolution and reduce image noise and improve quantification of regional myocardial perfusion with thallium-201 tomography.12
Correction for translational heart movement during acquisition caused by patient motion or upward creep after exercise was performed on all studies.13 As previously described, the heart was tracked in serial Wiener-filtered projection images, using the “diverging squares” algorithm; the projection images were then realigned in the x-y plane so that the heart center conformed to the projected position of a fixed point in space and the shifted images were reconstructed as described below. Filtering between slices and attenuation correction were not performed. Visually determined cardiac axes were obtained from mid-myocardial transverse and long-axis sections. The raw data were then reoriented around the true cardiac long axis to form orthogonal short-axis and horizontal and vertical long-axis image sets.
Image interpretation was performed visually by an experienced nuclear cardiologist (L.C.B.) without knowledge of the subject’s identity or exercise test results. Eight paired stress and redistribution images (16 per video screen) were displayed in four projections (transverse, short-axis, horizontal long-axis, and vertical long-axis) on a Sopha NXT system and stored on optical disk. These eight slices usually spanned the entire left ventricle.
Results were coded as positive or negative for reversible ischemia. Borderline reversible defects and mild to moderate fixed defects were recorded as negative. A severe fixed defect was also considered abnormal but did not occur in any of the subjects. A positive thallium tomogram was defined by a segmental perfusion defect on the immediate postexercise images in at least two contiguous tomographic slices and two image orientations, with definite improvement or normalization on the delayed images. Defects that were nonsegmental, especially slitlike apical defects or those limited to the posterior septal wall, were coded as normal. We have previously shown, using receiver-operating characteristic curve analysis, that visual interpretation using these criteria provides a sensitivity for detection of coronary disease of 95% for tomography and 88% for planar imaging at a false-positive rate of 10%.14 In patients without previous infarction, sensitivities for tomographic and planar imaging were 87% and 83%, respectively (see Reference 17).
In 16 siblings, only planar imaging was performed because of unavailability of the tomographic camera on the day of screening. In general, an all-purpose collimator was used to acquire three views (anterior, 45° left anterior oblique, and 75° left anterior oblique), each containing at least 500 000 counts. Paired stress and redistribution images were read visually without background subtraction on the video display monitor. As with the tomographic studies, a positive scintigram was defined by a reversible segmental perfusion defect. Five siblings (all women) declined the thallium study and underwent stress ECG testing only.
On the basis of the stress ECG and thallium scintigraphy results, four mutually exclusive groups were defined for analysis: (1) negative stress ECG and negative thallium scan, (2) positive stress ECG and negative thallium scan, (3) negative stress ECG and positive thallium scan, and (4) concordant positive stress ECG and thallium scan. In addition, the predictive value of an abnormal stress ECG, irrespective of the results of thallium scintigraphy (groups 2 and 4), was examined, as was the value of an abnormal stress ECG or an abnormal thallium scintigram (groups 2, 3, and 4).
Evaluation of CHD Events: Follow-up Questionnaire
A health status and CHD events history questionnaire was administered by telephone in all siblings who were screened. Information on any new CHD or related diagnosis was elicited by a trained interviewer using standardized questions. Information was requested on new diagnostic tests including repeat treadmill testing, thallium scintigraphy, or coronary arteriography. Questions elicited information about CHD interventions including coronary angioplasty, atherectomy, or bypass surgery. All outcomes were assessed in siblings 1 to 9 years after baseline screening. The information was verified by obtaining physician office records, hospital records, death records, and original data on all diagnostic tests performed. All records were reviewed by two cardiologists to verify the presence or absence of a CHD event. CHD events were prospectively defined according to the criteria in Table 1⇓.
Characteristics of the subjects who had a coronary event any time during follow-up and those who did not were compared using 2×2 contingency tables and the χ2 statistic. The relation between exercise test results and the incidence of a first CHD event was examined using Kaplan-Meier survival curves for each of the four test result contingencies: (1) both ECG and thallium normal, (2) only ECG abnormal, (3) only thallium abnormal, and (4) both ECG and thallium abnormal.15 The statistical difference between the curves representing these four different exercise test groups was assessed with a log-rank test.16 A Cox proportional hazards model was also used to study the risk of developing a CHD event for each of the four combinations of exercise test results.15 Multivariate Cox proportional hazards models were developed to adjust for sex and age (≥45 years versus <45 years, divided at the median age value), as well as maximal MET level attained during exercise and the presence of certain coronary risk factors. Estimates of relative risk and corresponding two-sided 95% confidence intervals (CIs) relating exercise test results to CHD events were computed from the Cox models.17 All tests of significance were two-tailed.
The 264 siblings participating in this study were identified from 161 index cases with documented coronary artery disease prior to 60 years of age. Index cases, 86% of whom were men, had their first CHD event at a mean age of 46.2±7.7 years (range, 28 to 59 years). Qualifying CHD events in index cases included myocardial infarction in 92 (57%), coronary bypass surgery in 50 (31%), angioplasty without previous infarct in 15 (9%), and an abnormal coronary angiogram (at least one major coronary artery with a ≥50% diameter stenosis) in 5 (3%). The number of siblings per index case averaged 2.6, with an average of 1.3 already having documented coronary artery disease. Of the 161 index cases, 100 had a single eligible sibling participating in the study; 32 had 2 eligible siblings; 17 had 3 siblings; 11 had 4 siblings; and 1 had 5 siblings (mean of 1.64 siblings per index case).
The characteristics of the sibling population are shown in Table 2⇓. The study population was generally well educated, the majority were men, and almost all were white. There was a high prevalence of traditional coronary risk factors, including elevated low-density lipoprotein (LDL) cholesterol, hypertension, and smoking.
Results of Exercise Testing and Thallium Scintigraphy
Maximal symptom-limited treadmill exercise tests were performed in all siblings. Mean maximal heart rate was 173±16 beats per minute; 96% of siblings achieved >85% of the maximal predicted age-adjusted heart rate. The mean MET level attained was 11.6. No individual experienced typical anginal chest pain during the test. Ischemic-type ST-segment abnormalities were observed during and/or after exercise in 22 of the 181 male siblings (12.2%) and 4 of the 83 female siblings (4.8%). Marked ST-segment depression (≥3 mm) occurred in only 1 sibling; abnormalities first appeared at a heart rate of <120 beats per minute in only 2 male siblings; 1 male sibling had exercise-induced hypotension.
An abnormal exercise thallium perfusion scan, characterized by one or more reversible exercise-induced perfusion defects, was observed in 52 of 181 male siblings (28.7%) and 7 of 78 female siblings (9.0%). Abnormal thallium scans were about twice as frequent as abnormal exercise ECGs. Most perfusion defects were single; multiple defects were observed in only 5 of the 52 men with abnormal scans and none of the 7 women. Forty-nine of the 64 total defects were graded as mild, 14 (22%) as moderate, and only 1 (2%) as severe.
The exercise ECG and thallium results for male and female siblings are shown in Fig 1⇓. The exercise ECG and thallium scan were both abnormal in 12 men but no women. All 12 of these “double-positives” occurred in men who were more than 45 years of age at the time of screening. Abnormal exercise ECGs and/or thallium scans were much more common in male than female siblings and were more common in siblings ≥45 years of age among both sexes. In only 55% of men ≥45 years of age were the exercise test and thallium scan both normal. In comparison, 76% of men <45 years of age, 80% of women ≥45 years of age, and 96% of women <45 years of age had both tests normal. The addition of thallium imaging to the exercise tests resulted in substantially more individuals being identified as abnormal; 23% of men ≥45 years of age, 22% of men <45, 12% of women ≥45, and 4% of women <45 were identified as abnormal based on the thallium results.
Coronary Heart Disease Events
During an average follow-up period of 6.2 years (range, 1 to 9 years), 19 siblings who were asymptomatic and apparently healthy at the time of screening developed clinically manifest CHD (Table 3⇓). One 47-year-old man experienced sudden death 10 months after screening. Although there was no history of angina or myocardial infarction, autopsy showed severe multivessel coronary artery disease and extensive posterior wall scarring.
Ten siblings had an acute myocardial infarction an average of 33.4 months after screening. All presented with chest pain and ST-segment elevation and 6 received thrombolytic therapy. One had ventricular fibrillation and was resuscitated. Seven siblings developed angina pectoris without infarction and underwent a coronary revascularization procedure (5 angioplasty, 2 bypass surgery). One additional sibling with worsening exercise test findings but no chest pain also underwent coronary bypass surgery. All of the siblings undergoing revascularization had severe coronary narrowings at catheterization, and 6 of the 8 had significant multivessel disease. In addition to these initial events, 2 siblings with a first infarct experienced a second one, both requiring bypass surgery; 2 siblings with an infarct subsequently underwent coronary angioplasty, 1 of these also resulted in emergency bypass surgery.
Fifteen of 19 siblings developing clinical CHD were men ≥45 years old at the time of screening. Mean age at screening was 50 in siblings with CHD events compared with 45 in those without events (P<.05). The overall incidence of CHD events was 7.2%; among men ≥45 at screening, the incidence of events was 16.9%. Having an event was significantly associated with being ≥45 years old at screening, being male, having an abnormal exercise ECG, and having an abnormal thallium scan (Table 4⇓).
Prediction of Individuals Developing Clinical Coronary Heart Disease Events
Crude incidence rates for first coronary events in siblings with different exercise ECG and thallium results are shown in Table 5⇓. Clinical CHD developed in 3% of siblings who were normal by both tests compared with 7% of those who were abnormal only by ECG, 13% of those abnormal only by thallium, and 50% of those abnormal by both tests. An abnormal exercise ECG had a predictive value of 27% for the development of future clinical CHD; of those developing CHD, 37% were abnormal by exercise ECG at the time of screening (“sensitivity” of the test). For an abnormal thallium scan, the predictive value for clinical CHD was slightly lower (20%), but the sensitivity was nearly doubled (63%). A combined abnormal ECG and thallium scan had a predictive value of 50% and a sensitivity of 32%.
Survival analysis was conducted to predict the risk of developing clinical CHD for different combinations of exercise test results after accounting for possible differences in censoring and follow-up time among groups (Fig 2⇓). Examination of 5-year cumulative incidences shows a 2.5% incidence in siblings with both tests normal, compared with 8% in those abnormal only by ECG, 12% in those abnormal only by thallium, and 64% in those abnormal by both tests.
The relative risk (RR) of having a first CHD event was significantly increased in siblings who were abnormal by thallium only (RR=4.45, P<.01) and most notably in siblings abnormal by both tests (RR=33.3, P=.0001) compared with siblings with both tests normal. After adjusting for age and sex using a Cox proportional hazards model, relative risk was still significantly increased in siblings with abnormal thallium only (3.6, P<.03) and in siblings with both thallium and ECG abnormal (14.5, P<.0001). The results were similar when first CHD events were restricted to death and myocardial infarction: such events occurred in 2% of siblings normal by both tests, 9% of those abnormal only by ECG, 7% of those abnormal only by thallium, and 17% of those abnormal by both tests. The relative risk of having a myocardial infarction or death was increased in siblings abnormal by ECG (RR=6.8, P=.03), thallium (RR=3.8, P=.10), or both (RR=61.3, P=.01). The statistical significance of these relationships was reduced due to the limited number of these types of events.
Comparing all siblings with abnormal exercise ECGs with those with normal exercise ECGs (in Table 5⇑), the relative risk of having an abnormal exercise ECG, adjusted for age and sex, was 3.9 (95% CI, 1.4 to 10.3, P<.007). Comparing all siblings with abnormal exercise thallium scans to those with normal scans (in Table 5⇑), the relative risk of having an abnormal thallium scan, adjusted for age and sex, was 5.3 (95% CI, 2.0 to 13.9, P=.0006). Among siblings with normal exercise ECGs, those with abnormal thallium scans had a relative risk of 4.5 (95% CI, 1.5 to 14.1, P=.009) compared with those with normal scans. Similarly, among siblings with abnormal exercise ECGs, an abnormal thallium scan resulted in a relative risk of 11.1 (95% CI, 1.3 to 93, P=.03). Finally, comparing siblings with abnormal thallium scans with those with normal ones, relative risk was 4.7 (95% CI, 1.8 to 12.5, P=.002) after adjusting for age, sex, and exercise ECG results.
Compared with siblings with a normal thallium scan, those with an abnormal scan had a lower mean high-density lipoprotein (HDL) cholesterol (45.4 versus 52.6 mg/dL, P=.001) and more frequent hypertension (58% versus 38%, P=.008). There were no significant differences in LDL cholesterol or in the prevalence of current smoking. An abnormal exercise ECG was significantly associated only with a lower HDL cholesterol (44.9 versus 51.7 mg/dL, P=.003). Nevertheless, none of these coronary risk factors were significantly associated with the development of clinical CHD in the Cox proportional hazards model. Controlling explicitly for HDL cholesterol and hypertension had no influence on the results presented in Table 5⇑.
Before adjustment for stress test results, both older age (age ≥45, RR=5.6, P=.003) and male sex (RR=6.2, P=.02) predicted the risk of developing clinically manifest CHD. After adjustment for ECG and thallium results, age was still associated with CHD events (P=.05), while the association of sex with CHD events was still present but was no longer statistically significant (P=.13). Although the primary analysis used age dichotomized at 45 (the median value) to simplify interpretation of the results, the results were virtually identical in all respects when age was used as a continuous variable except that after adjustment for ECG and thallium results, age was no longer statistically significant (P=.13).
Because the overall results were strongly influenced by the high prevalence of CHD events in men (17 of 19), the analyses were all repeated for men alone. The results were the same in men with lower levels of statistical significance, principally because of smaller sample size. Within male siblings, the relative risk was increased in the three groups with any abnormal test results, but only double-positives were statistically significant. Age-adjusted relative risk was 1.4 in male siblings abnormal by ECG only (95% CI, 0.2 to 12.5), 3.1 in those abnormal by thallium only (95% CI, 0.9 to 10.9), and 12.4 in those abnormal by both ECG and thallium (95% CI, 3.6 to 42.8, P=.0001). Male siblings with an abnormal thallium scan had a relative risk of 4.3 (95% CI, 1.5 to 11.9, P=.006) after adjusting for age and exercise ECG results. Outcomes were too few in women to allow the genesis of a stable multivariate analysis.
Additional analyses were performed to examine the possible confounding effect of certain additional variables on the results. The MET level achieved during exercise was significantly lower for those with an abnormal ECG (11.1 versus 12.8, P=.003). However, MET level was not associated with the occurrence of CHD events, and explicit adjustment of the Cox proportional hazards model for MET level did not change the results. Since 16 subjects had planar rather than tomographic thallium studies, all analyses were repeated with these individuals excluded. The results were virtually identical to the primary analysis of all subjects. Finally, the statistical analysis treats each sibling as an independent observation despite multiple siblings coming from some families. To control for the familial clustering of CHD, the analysis was repeated using only the oldest sibling in each family. The main results remained unchanged: the crude incidence of clinical CHD was 4% among the 110 siblings with both normal ECG and thallium scan, 13% among the 30 siblings with normal ECG/abnormal thallium scan, 0% among the 9 siblings with abnormal ECG/normal thallium scan, and 45% among the 11 siblings with both tests abnormal.
Siblings of patients with premature coronary artery disease are known to be at high risk of developing CHD at a relatively young age. This study demonstrates that exercise testing and particularly thallium scintigraphy are useful tools for identifying individuals who will develop future clinically manifest CHD. Despite being apparently healthy, asymptomatic, and relatively young at the time of screening, 7.2% of the 264 siblings developed clinical CHD over a mean of 6.2 years of follow-up. Although abnormal exercise ECGs and thallium scans were both predictive of CHD events, thallium was more predictive; after adjusting for age, sex, and exercise ECG results, the relative risk of having a CHD event in those with an abnormal thallium scan was 4.7, and the relative risk associated with a “double-positive” stress ECG and thallium scan was 14.5.
It is important to note that this is a cohort of apparently healthy individuals. Ninety-six percent surpassed 85% of maximal predicted heart rate during exercise testing, and the mean MET level at maximal exercise averaged 11.6. Very few individuals had a strongly positive stress test or a high-risk thallium scan with multiple or severe perfusion defects. None experienced typical anginal chest pain during exercise or in the postexercise period. Thus, when exercise-induced ischemia occurred in this sibling cohort, it was almost always of a mild degree and clinically silent. Nonetheless, almost half of asymptomatic male siblings 45 years of age or older at the time of screening had an abnormal test result. Abnormal noninvasive tests occurred in almost a quarter of men and a fifth of older women but in only 4% of younger women. The low prevalence of abnormal noninvasive tests in young female siblings despite a strong family history of premature coronary artery disease suggests that stress testing is not warranted in this group.
Previous studies have examined the value of treadmill testing for predicting first CHD events. As summarized by Froelicher,18 eight studies following a total of approximately 5000 clinically healthy, mostly middle-aged men for 3 to 13 years for development of CHD events, including angina pectoris, demonstrated that an abnormal exercise test had a sensitivity of 48%, a positive predictive value of 26%, and a risk ratio of 9 for identifying individuals with any subsequent events. However, the exercise test was found to perform considerably less well for the prediction of “hard” end points (death or myocardial infarction but not angina). Results from the Seattle Heart Watch,19 Multiple Risk Factor Intervention Trial,20 21 and Lipid Research Clinics22 23 studies have demonstrated that for these “hard” end points, an abnormal exercise test had a sensitivity of 27%, a positive predictive value of only 6%, and a risk ratio of 4.18
In the current study, which demonstrated a relative risk of 3.9 for an abnormal exercise ECG and 5.3 for an abnormal thallium scan (after adjusting for age and sex), end points for the onset of clinical CHD included death, myocardial infarction, and revascularization procedures. We did not include the presence of angina per se but instead required that a new anginal syndrome be accompanied by objective ECG changes and/or angiographic evidence of hemodynamically significant coronary artery narrowings and be considered severe enough to warrant angioplasty or bypass surgery. Given these stipulations, we believe that a revascularization procedure in a previously asymptomatic person represents a useful end point that reflects significant worsening of coronary disease and the transition from an occult to a clinically manifest stage of CHD. Although the exercise test at screening could have influenced the later decision for revascularization, this seems unlikely, since no sibling had coronary angiography for clinical reasons after screening, and revascularization did not occur until an average of 45 months (range, 10 to 95 months) later (Table 3⇑).
Fleg and associates24 were the first group to use stress thallium testing to screen for future CHD events in an asymptomatic community-dwelling population. They reported that the prevalence of exercise-induced silent ischemia (defined as a concordant positive stress ECG and thallium scan) rose from 2% in the 5th and 6th decades to 11% in those greater than 65 years of age. In our cohort of male and female siblings less than 60 years of age, 4.6% (12 of 259) had both an abnormal stress ECG and thallium examination. All 12 of these “double-positives,” however, occurred in men more than 45 years of age at the time of screening. Fleg and coworkers found that CHD events (death, myocardial infarction, or angina) occurred in 48% of those with concordant positive findings during a 4.6-year mean follow-up. By comparison, siblings with concordant positive stress ECG and thallium scans had a 64% 5-year cumulative incidence of CHD events in our study.
On the basis of their results, Fleg et al24 concluded that stress thallium testing was probably not a cost-effective strategy for the early diagnosis of CHD in unselected apparently healthy persons, especially if they were less than age 60 years. Froehlicher18 concluded that screening of asymptomatic persons should generally be reserved for those with multiple abnormal risk factors or a family history of premature coronary artery disease. We would generally concur with these views, since our study indicates that in asymptomatic siblings of individuals with documented CHD before age 60 years, many of whom had multiple risk factors, stress thallium scintigraphy does serve to predict individuals who will develop clinically manifest CHD. Our data support exercise thallium testing to screen for silent coronary artery disease in male siblings 45 years of age or older, since the prevalence of abnormal exercise ECGs and/or thallium scans and the incidence of clinical CHD during follow-up were both high (45% and 16.9%, respectively). Screening also may be indicated in younger male siblings and older female siblings, since the prevalence of abnormal exercise tests or thallium scans was also moderately high (24% and 20%, respectively). However, very few of these individuals developed clinical CHD during follow-up, and none had both an abnormal exercise ECG and thallium scan. Nevertheless, the severity of occult coronary disease may have been less pronounced in these individuals, and a follow-up period longer than 6 years might have allowed clinical CHD to become manifest. Evidence that exercise thallium scintigraphy is predictive of future CHD events in these groups will probably require long-term observational studies of 10 years or more.
It should be noted that 6 of the 19 siblings who developed a CHD event had a normal stress thallium test, which may indicate hemodynamically insignificant coronary artery stenoses at the time of screening. Recent data indicate that many acute ischemic events are associated with plaque rupture and the rapid progression of coronary disease in arterial segments that previously had only minimal obstruction.25 26 27 28 An exercise stress test is designed to detect ischemia resulting from an imbalance of oxygen supply and demand during exercise, or flow heterogeneity in the case of stress thallium scintigraphy.9 If there is no flow-limiting stenosis, an abnormal exercise test is unlikely; however, coronary disease that may become clinically manifest in a relatively short period of time nevertheless may be present in this high-risk group.
A recent investigation by Schwartz et al29 in asymptomatic men found that thallium exercise scintigraphy was only 45% sensitive and 78% specific for detecting at least one 50% or greater diameter stenosis in a major epicardial coronary artery. This series, however, did not have long-term follow-up for CHD events. It is possible that asymptomatic persons with evidence of reversible ischemia on provocative exercise thallium testing may develop vasoconstriction at the site of a nonobstructive coronary stenosis. With longer follow-up, these persons with presumed endothelial dysfunction may show a greater propensity to develop clinical CHD events.
A possible rationale for performing exercise thallium testing of asymptomatic siblings might be to identify individuals with extensive occult ischemia, who should undergo coronary angiography to identify significant obstructions and subsequent myocardial revascularization to prevent future myocardial infarction or death. However, very few of our siblings had “high-risk” thallium scans or exercise tests, and even though those individuals with abnormal scans and/or exercise ECGs had a significant increase in the risk of developing clinical CHD, it was still the minority that developed CHD, and even fewer had death or myocardial infarction. In addition, it has not been established that “prophylactic” revascularization is beneficial in preventing death or myocardial infarction in asymptomatic people with normal left ventricular function, even in those with “high-risk anatomy.”
Another and probably more important reason for performing stress thallium scintigraphy in these siblings is that an abnormal test result may influence healthcare providers to be more vigorous in the management of an individual’s modifiable risk factors. We now have good evidence that aggressive risk factor modification can greatly slow the progression of atherosclerosis.30 However, an aggressive approach is often not taken in usual clinical practice, since it may involve prescribing a lipid-lowering medication to an asymptomatic person for many years (perhaps more than a decade) in an effort to prevent or delay the theoretical occurrence of a future CHD event. Evidence of occult ischemia by exercise thallium testing should make the physician more inclined to aggressively prescribe medication to control lipids and should help motivate the patient to undertake more extensive lifestyle modification, including more stringent fat and calorie restriction and smoking cessation. We have, in fact, preliminary evidence from an analogous group of siblings that smoking cessation, cholesterol lowering, and control of blood pressure are significantly better achieved in those with abnormal exercise tests and/or thallium scans than in those with normal tests. In addition, an abnormal test identifies a subset of individuals who should be followed more closely for the onset of symptoms or progression of occult disease.
An abnormal thallium scan was defined as a reversible segmental thallium defect on delayed imaging; reinjection of thallium was not performed, and we did not perform 24-hour delayed imaging31 because these techniques had not yet been described during the time that the siblings were being entered into this study. We may, therefore, have underestimated the prevalence of perfusion defects. However, the majority of defects observed in these siblings were mild, and reinjection is most useful in patients with severe defects after exercise or in those with previous myocardial infarction and not in asymptomatic people with mild stress-induced defects.32 It should also be pointed out that our thallium studies were processed using Wiener filtering of the projection data and correction for motion prior to tomographic reconstruction, methods that we have previously shown improve image quality and reduce artifacts and are now standard in our laboratory.12 13 It is unknown whether the results would have been the same if these procedures had not been used. Finally, almost all of the siblings in our study were white, with only 3% African Americans. It is unknown to what extent the results can be generalized to other racial groups.
We found that the predictive value of exercise stress testing and thallium scintigraphy was greatly increased in an asymptomatic group of individuals less than 60 years of age with a family history of premature coronary disease. Abnormal thallium scans were observed in 29% of men and 9% of women; they occurred about twice as frequently as did abnormal stress ECGs. The development of clinical CHD during a mean 6.2 years of follow-up was significantly associated with being a man ≥45 years of age, having an abnormal stress ECG, and having an abnormal thallium scan. Both the exercise ECG and thallium scans provide predictive information, but the scan data have a much higher sensitivity and remains predictive even after adjusting for the stress ECG results. Individuals with a concordant positive stress test and thallium scan had a strikingly high incidence of subsequent coronary events. Thus, stress thallium scintigraphy may be particularly useful in the risk assessment of men 45 years of age or older with a family history of premature coronary disease.
This study was supported by USPHS Grants NR02241 from the National Center for Nursing Research and HL-49762 from the National Heart, Lung, and Blood Institute and supported in part by the General Clinical Research Center grants MO1-RR00035 and MO1-RR00722 from the National Center for Research Resources and the National Institutes of Health.
Reprint requests to Roger S. Blumenthal, MD, Johns Hopkins Hospital, Ciccarone Preventive Cardiology Center, Carnegie 538–Cardiology, 600 N Wolfe St, Baltimore, MD 21287.
- Received July 20, 1995.
- Revision received October 4, 1995.
- Accepted October 11, 1995.
- Copyright © 1996 by American Heart Association
Rose G. Familial patterns in ischemic heart disease. Br J Prev Soc Med. 1964;18:175-180.
Deutscher S, Ostrander LD, Epstein FH. Familial factors in premature coronary heart disease. Am J Epidemiol. 1970;91:233-243.
Snowden CG, McNamara DM, Garrison RJ, Feinleib M, Kannel WB, Epstein FH. Predicting coronary heart disease in siblings: a multivariate assessment. Am J Epidemiol. 1982;115:217-222.
Slack J, Evans KA. The increased risk of death from ischemic heart disease in first degree relatives of 121 men and 96 women with ischemic heart disease. J Med Genet. 1966;3:239-257.
Roncaglioni MC, Santoro L, D’Avanzo, Negri E, Nobili A, Ledda A, Pietropaolo F, Franzosi MG, La Vecchia C, Feruglio GA, Maseri A. Role of family history in patients with myocardial infarction. Circulation. 1992;85:2065-2072.
Sox HC, Littenberg B, Garber AM. The role of exercise testing in screening for coronary artery disease. Ann Intern Med. 1989;110:456-469.
Links JM, Jeremy RW, Dyer SM, Frank TC, Becker LC. Wiener filtering improves quantification of regional myocardial perfusion with thallium-201 SPECT. J Nucl Med. 1990;31:1230-1236.
Geckle WJ, Frank TL, Links JM, Becker LC. Correction for patient and organ movement in SPECT: application to exercise thallium-201 cardiac imaging. J Nucl Med. 1988;29:441-450.
Fintel DJ, Links JM, Brinker JA, Frank TL, Parker M, Becker LC. Improved diagnostic performance of exercise thallium-201 single photon emission computed tomography over planar imaging in the diagnosis of coronary artery disease: a receiver operating characteristic analysis. J Am Coll Cardiol. 1989;13:600-612.
Peto R, Peto J. Asymptomatically efficient rank invariant test procedures. J R Stat Soc Ser. 1972;135:185-207.
Cox DR. Regression models and life tables. J R Stat Soc Ser. 1972;34:187-200.
Froelicher VF. Special applications: screening apparently healthy individuals. In: Manual of Exercise Testing. St Louis, Mo: Mosby; 1994:160-176.
Gordon DJ, Ekelund L, Karon JM, Probstfield JL, Rubenstein C, Sheffield LG, Weissfeld M. Predictive values of the exercise tolerance test for mortality in North American men: the Lipid Research Clinics Mortality Follow-up Study. Circulation. 1986;74:252-256.
Ekelund LG, Suchindran CM, McMahan RP, Heiss G, Leon AS, Rombitt DW, Rubenstein CL, Probstfield JL, Ruwith JF. Coronary heart disease morbidity and mortality in hypercholesterolemic men predicted from an exercise test: the Lipid Research Clinics Coronary Primary Prevention Trial. J Am Coll Cardiol. 1989;14:557-563.
Fleg JL, Gerstenblith G, Zonderman AB, Becker LC, Weisfeldt ML, Costa PT Jr, Lakatta EG. Prevalence and prognostic significance of exercise-induced silent myocardial ischemia detected by thallium scintigraphy and electrocardiography in asymptomatic volunteers. Circulation. 1990;81:428-436.
Ambrose JA, Winters SL, Arora RR, Eng A, Riccio A, Gorlin R, Fuster V. Angiographic evolution of coronary artery morphology in unstable angina. J Am Coll Cardiol. 1986;7:471-478.
Little WC, Constantinescu M, Applegate RJ, Kutcher MA, Burrows MT, Kahl FR, Santamore WP. Can coronary angiography predict the site of a subsequent myocardial infarction in patients with mild-moderate coronary artery disease? Circulation. 1988;78:1157-1166.
Hackett D, Davies G, Maseri A. Pre-existing coronary stenoses in patients with first myocardial infarction are not necessarily severe. Eur Heart J. 1988;9:1317-1323.
Schwartz RS, Jackson WG, Celio PV, Richardson LA, Hickman JR. Accuracy of exercise 201-Tl myocardial scintigraphy in asymptomatic young men. Circulation. 1993;87:165-172.