Importance of epidemiology and biostatistics in deciding clinical strategies for using diagnostic tests: a simplified approach using examples from coronary artery disease.

The explosion of costly new medical diagnostic technologies demands a common sense approach to help physicians decide appropriate indications and strategies for use of these tests. This simple, nonmathematical review focuses on the assessment of coronary artery disease, but the approach can be generalized to other medical problems. This clinical approach to diagnostic testing strategies is based on seven sequential questions: 1. What is the clinical probability that this patient has a specific disease characteristic based on clinical data? 2. What is the overall objective for management of this patient based on the overall status of the patient? 3. Most importantly, what specific questions need to be answered about the patient's condition before the physician can recommend the most appropriate management (e.g., whether the patient has coronary disease, whether an anatomic lesion is functionally significant, whether a myocardial region is reversibly ischemic or irreversibly infarcted, whether a particular therapy has had good or bad effects or what is the patient's prognosis)? The key point is for the physician to formulate a specific clinical question about the patient before the test. 4. The physician must then ask how well does the test answer the particular clinical question about the patient. Here the physician needs to understand the sensitivity and specificity of the test, especially because they are influenced by various clinical biases. 5. Next, the physician must ask how to interpret the reliability of a positive or negative test result in the individual patient. This requires understanding predictive value and predictive error of a given result and how they are influenced by the clinical data as described by Bayes' theorem. 6. Next, the physician must ask what further tests or therapies will be recommended for the patient. The physician can estimate in advance how different test results would alter management plans and he can then allow this estimate to help determine indications for the test. There is some controversy concerning whether to use Bayes' theorem or multivariate analysis to estimate the final probability of a disease characteristic. 7. Finally, in this era of quality assurance, professional review and cost containment, it behooves each physician to ask whether the data provided by the particular tests were worth the cost, inconvenience and risk for that particular patient.

Gated blood pool scintigraphic monitoring of doxorubicin cardiomyopathy: comparison of camera and computerized probe results in 101 patients.

Serial gated blood pool scintigraphic monitoring of cardiac function with both a nonimaging scintillation probe and a conventional gamma camera-computer imaging system was performed in 101 patients receiving doxorubicin hydrochloride (Adriamycin) chemotherapy. Comparison of probe- and camera-derived ejection fractions (n = 287) correlated significantly (r = 0.70, p less than 0.005) as did the interstudy (n = 183) change in ejection fraction (r = 0.76, p greater than 0.005). Significant discordance in probe- and camera-derived ejection fraction change occurred in 3 (1.6%) of 183 interstudy intervals. Average intrastudy variability of absolute probe-derived ejection fraction was 2.9%. This variability was unrelated to the level of cardiac function. Thirteen patients (13%) developed clinical cardiotoxicity, including four at cumulative Adriamycin levels less than 450 mg/m2. Mean absolute camera ejection fraction decline for these patients was 21% from baseline evaluation, and mean absolute probe ejection fraction decline was 22%. The minimal absolute ejection fraction decline was 11% for patients with clinical congestive heart failure. Eight asymptomatic patients had therapy terminated before the development of clinical cardiotoxicity after a mean decline in absolute camera ejection fraction of 19 +/- 4% (SD) and in probe ejection fraction of 19 +/- 9% into abnormal ranges (a decline in magnitude equivalent to that in patients developing congestive failure). None of these five asymptomatic patients available for clinical follow-up at 6 months after termination of Adriamycin therapy subsequently developed signs of ventricular dysfunction. The majority of patients (83%) studied at 450 mg/m2 cumulative dose levels did not have a 15% or greater decline from baseline into the abnormal range.(ABSTRACT TRUNCATED AT 250 WORDS)

Bayesian comparison of cost-effectiveness of different clinical approaches to diagnose coronary artery disease.

The objective of this study was to compare the cost-effectiveness of four clinical policies (policies I to IV) in the diagnosis of the presence or absence of coronary artery disease. A model based on Bayes' theorem and published clinical data was constructed to make these comparisons. Effectiveness was defined as either the number of patients with coronary disease diagnosed or as the number of quality-adjusted life years extended by therapy after the diagnosis of coronary disease. The following conclusions arise strictly from analysis of the model and may not necessarily be applicable to all situations. As prevalence of coronary disease in the population increased, it caused a linear increase in cost per patient tested, but a hyperbolic decrease in cost per effect, that is, increased cost-effectiveness. Thus, cost-effectiveness of all policies (I to IV) was poor in populations with a prevalence of disease below 10%, for example, asymptomatic people with no risk factors. Analysis of the model also indicates that at prevalences less than 80%, exercise thallium scintigraphy alone as a first test (policy II) is a more cost-effective initial test than is exercise electrocardiography alone as a first test (policy I) or exercise electrocardiography first combined with thallium imaging as a second test (policy IV). Exercise electrocardiography before thallium imaging (policy IV) is more cost-effective than exercise electrocardiography alone (policy I) at prevalences less than 80%. 4) Noninvasive exercise testing before angiography (policies I, II and IV) is more cost-effective than using coronary angiography as the first and only test (policy III) at prevalences less than 80%. 5) Above a threshold value of prevalence of 80% (for example patients with typical angina), proceeding to angiography as the first test (policy III) was more cost-effective than initial noninvasive exercise tests (policies I, II and IV). One advantage of this quantitative model is that it estimates a threshold value of prevalence (80%) at which the rank order of policies changes. The model also allows substitution of different values for any variable as a way of accounting for the uncertainty inherent in the data. In conclusion, it is essential to consider the prevalence of disease when selecting the most cost-effective clinical approach to making a diagnosis.

Atrial kinetics and left ventricular diastolic filling in the healthy elderly.

A delay of left ventricular isovolumic relaxation and decrease in myocardial compliance may result in a decline of measured early filling rates in elderly subjects. Previous studies of diastolic function, however, have not excluded coronary artery disease or addressed the contribution of atrial contraction to diastole. The present study evaluated radionuclide-derived diastolic variables in 13 healthy elderly volunteers aged 75 +/- 6 years without symptoms or risk factors for coronary disease who had normal findings on the stress electrocardiogram, stress gated blood pool imaging and two-dimensional echocardiogram. Results were compared with those of a group of 10 healthy young volunteers aged 26 +/- 5 years. High count, 32 frame, double-buffered gated blood pool acquisitions were obtained at rest in the left anterior oblique view with an RR interval variation less than 5%. Left ventricular time-activity curves were analyzed and flow-volume loops for each group were constructed. In the healthy elderly: peak early diastolic filling rate is decreased, time of peak early filling and time to first third of diastolic filling are delayed, and peak late left ventricular filling rate and percent of atrial filling volume are augmented, suggesting an adaptive response of the atria to diminished left ventricular compliance.

Relation between exercise-induced changes in ejection fraction and systolic loading conditions at rest in aortic regurgitation.

To examine the role of systolic wall stress at rest in determining left ventricular performance during exercise in aortic regurgitation (AR), systolic wall stress (measured by M-mode echocardiography) was related to changes in left ventricular function during maximal exercise (evaluated by radionuclide ventriculography) in 30 patients with chronic aortic regurgitation. Of these 30 patients, 7 had a normal exercise response, defined as an absolute increase in ejection fraction of 5% or greater (Group I) and 23 had abnormal exercise response, defined as no change (less than 5% change) or a decline (less than or equal to 5%) in ejection fraction (Group II). Patients in Group I had a significantly lower radius/wall thickness ratio (2.5 +/- 0.2 versus 3.1 +/- 0.1, p less than 0.01) and lower peak systolic wall stress (123 +/- 11 versus 211 +/- 12 X 10(3) dynes/cm2, p less than 0.01) than patients in Group II. An increase in ejection fraction during exercise was seen in 6 of the 9 patients with normal systolic wall stress at rest (less than 150 X 10(3) dynes/cm2), but in only 1 of 21 patients with elevated systolic wall stress (p less than 0.001). Peak systolic wall stress at rest varied linearly, and inversely with changes in left ventricular ejection fraction during exercise (r = 0.60, p less than 0.001). Groups I and II did not differ in ejection fraction at rest, clinical symptoms or maximal work load achieved.(ABSTRACT TRUNCATED AT 250 WORDS)

Quantification of cardiac conduction abnormalities using segmental vector Fourier analysis of radionuclide gated blood pool scans.

Timing abnormalities of myocardial contractility may occur as inter- or intraventricular asynchrony. Gated blood pool scintigraphy was performed on 21 patients with a normal ejection fraction and the following electrocardiograms: six normal, six with left bundle branch block, four with right bundle branch block and five with right ventricular pacemaker rhythm. A phase and amplitude of the first harmonic of the Fourier transform was obtained for each pixel, and left and right ventricles were trisected. A mean vector phase for each region was obtained by vector summation. Regional and global values were analyzed within each group and compared with normal values. The phase differences between the entire left and right ventricles (mean +/- standard deviation) were: 9 +/- 3 in the normal patients, 38 +/- 8 (p less than 0.01) in patients with left bundle branch block, -6 +/- 7 (p less than 0.05) in patients with right bundle branch block and 15 +/- 9 (difference not significant) in patients with pacemaker rhythm. The phase differences between left ventricular posterolateral and septal regions were -4 +/- 2 in the normal patients, 10 +/- 5 (p less than 0.01) in patients with left bundle branch block, -7 +/- 10 (p less than 0.05) in patients with right bundle branch block and 10 +/- 5 (p less than 0.01) in patients with pacemaker rhythm. Within the right ventricle, phase differences between the apical and septal segments were 14 +/- 9 in the normal patients, 14 +/- 10 (NS) in patients with left bundle branch block, -2 +/- 3 (p less than 0.01) in patients with right bundle branch block and -22 +/- 18 (p less than 0.01) in patients with pacemaker rhythm. Interventricular phase differences were greatest in patients with left bundle branch block and absent or reversed in right bundle branch block.(ABSTRACT TRUNCATED AT 250 WORDS)

Physiological influences on perfusion imaging in transient myocardial ischaemia: importance of early distribution of thallium-201.

We tested the hypothesis that visualisation of defects on thallium-201 (201Tl) myocardial perfusion images (MPI) depends on the duration of the ischaemic state between 201Tl injection and the time of reperfusion of an occluded coronary artery. Praecordial imaging with a gamma camera was performed in 24 anaesthetised, open-chest dogs with transient coronary occlusion. Results indicated that if the duration of the ischaemic state after 201Tl injection was less than 3 min before reperfusion, then the MPI 5 to 15 min after 201Tl injection was falsely negative (201Tl activity in zone (IZ)/normal zone (NZ)greater than 0.85). Dogs which were ischaemic more than 5 min always had MPI defects 5 to 15 min after 201Tl injection (IZ/NZ201Tl ratio less than 0.85). MPI results (201Tl IZ/NZ) 15 min after 201Tl injection were determined by the duration of the ischaemic state after 201Tl injection (r = -0.86) because prolonged ischaemia allowed 201Tl to distribute from blood to myocardium before reperfusion: 201Tl (IZ/NZ( = 0.356 +/- 1.00 (fraction of total 201Tl remaining in blood at the end of the ischaemic state), r = 0.94.

Femoral pseudoaneurysm following nonpenetrating trauma in a patient with aortic insufficiency.

Although not uncommon after penetrating vascular trauma, arterial pseudoaneurysms rarely develop following blunt trauma. A patient is described in whom indirect trauma led to pseudoaneurysm of the profunda femoris artery, and persistent bleeding required surgical intervention. In this case, the coexistence of significant aortic valvular regurgitation suggests that wide pulse pressure may predispose to this arterial complication.

Accuracy and precision of regional multiharmonic Fourier analysis of gated blood-pool images.

In order to estimate the precision and accuracy of parameters derived from segmental multiharmonic Fourier analysis of gated blood-pool images, a Monte Carlo computer noise simulation was tested on five sample regional time-activity curves. The first three Fourier harmonics were retained and the precision and accuracy of parameters of ventricular function were calculated, varying the ejection fraction, segment size, and framing rate. Precision improved with higher ejection fraction, higher counts per frame, or higher framing rate. There was no change in precision as the framing rate changed at fixed total counts. Accuracy changed little with changing framing rate. Thus, for segmental analysis there is no advantage to using a higher framing rate. Regions five or more pixels in size are recommended for reliable results. This study provides useful information for the optimization of acquisition and processing conditions for regional gated blood-pool analysis.

Computer modeling of planar myocardial perfusion imaging: effect of heart rate and ejection fraction on wall thickness and chamber size.

Myocardial perfusion imaging is generally performed as a static acquisition without regard for dynamic changes in the cardiac cycle. The effect of heart rate and ejection fraction on the appearance of left ventricular chamber size and wall thickness as perceived in 201Tl scintigrams has not, to our knowledge, been previously studied. A dynamic computer model of the left ventricle was constructed, capable of varying the heart rate and ejection fraction. Parallel slices through the model were convolved with experimentally derived 201Tl point spread functions at corresponding depths to incorporate the effects of scatter and attenuation. Both gated and static left anterior oblique images were created at three clinically encountered heart rates and ejection fractions, with constant end-diastolic volume and left ventricular mass. Results of the study indicate that perceived and quantified wall thickness increases and chamber size decreases appreciably with increasing ejection fraction and (slightly) with increasing heart rate. Thus, evaluation of wall thickness and chamber size in planar images should take into account variations in heart rate and contractility. This is especially pertinent to estimates of left ventricular hypertrophy and chamber size, attempted from nongated myocardial perfusion images.

Effect of variable left ventricular vertical orientation on planar myocardial perfusion images.

Differences in vertical orientation of the left ventricle within the chest cavity cannot be corrected by gamma camera positioning. The effect of variations in vertical angulation on the appearance of the diagnostically important left anterior oblique (LAO) view has not been previously evaluated. In the current study, a computer simulation of a normal left ventricle was created and "imaged," varying only the degree of vertical rotation. The effect of six vertical positions on the LAO image was assessed visually and with horizontal and circumferential profile analysis. Results indicate a homogenous distribution of counts in the horizontal views. With increasing verticality, there are fewer counts in the valve plane, while the inferoapex initially increases in count density, and then progressively decreases. Quantification revealed count variations of up to 37% in the valve plane and 45% in the inferoapex due entirely to differences in vertical orientation of the left ventricular simulation. A survey of 167 patients who underwent routine stress thallium imaging showed a vertical angulation that varied from 7 degrees to 64 degrees (mean = 37 degrees) as determined from the anterior view. Clinical images were similar in appearance to computer generated images after correction for anterior view foreshortening. The present study suggests that the accuracy of current quantitative thallium methods to detect coronary artery disease might be enhanced by the use of a revised set of normal standards corrected for vertical orientation of the left ventricle.

Serial myocardial perfusion imaging with dipyridamole and rubidium-82 to assess restenosis after angioplasty.

UNLABELLED: The purpose of this study was to determine whether patients at high risk for clinical restenosis, following coronary angioplasty, could be identified by myocardial perfusion imaging performed with dipyridamole- 82Rb PET. METHODS: Forty-five patients (34 men, 11 women; mean age 58.5 yr) who had successful single-vessel angioplasty and were asymptomatic had dipyridamole-82Rb PET at 1 and 3 mo after the procedure. Abnormal flow reserve in the distribution of the angioplasty artery on PET was considered to be a decrease of > or = 1 perfusion grade in response to dipyridamole (assessed qualitatively from tomographic images and polar coordinate maps). Follow-up was performed for 6 mo postangioplasty. Clinical restenosis was defined as recurrent angina similar to that occurring before angioplasty and/or > or = 50% stenosis at the angioplasty site documented angiographically. We analyzed abnormal flow reserve in the distribution of the angioplasty vessel to identify which patients were at high risk for clinical restenosis. RESULTS: Fourteen patients developed clinical restenosis between 1 and 6 mo postangioplasty. Abnormal relative flow reserve in the distribution of the angioplasty vessel was present prior to the development of symptoms in 13 of 14 patients with clinical restenosis and in 8 of 31 patients without clinical restenosis (sensitivity 93%, specificity 74%, p < 0.0001). PET imaging successfully separated postangioplasty patients into groups with high (62%) and low (4%) risk of clinical restenosis. CONCLUSION: Abnormal relative flow reserve in the distribution of the angioplasty vessel on dipyridamole PET identifies asymptomatic postangioplasty patients at risk for clinical restenosis.

Left ventricular pseudoaneurysm causing superior vena caval obstruction and effusive-constrictive pericarditis.

A diabetic woman with a silent myocardial infarction on clinical and electrocardiographic criteria presented with findings on physical examination of superior vena caval obstruction and effusive-constrictive pericarditis. A left ventricular posterior wall pseudoaneurysm and intrapericardial hematoma were found, with extrinsic compression of the right atrium. The diagnosis was first suspected by radionuclide imaging and confirmed by contrast angiography and surgery.

Silent myocardial ischemia during PET.

A resting perfusion defect detected during radioisotope myocardial perfusion imaging is generally thought to represent myocardial infarction. The administration of nitroglycerin during cardiac PET imaging results in improved resting defects, which suggests transiently abnormal resting perfusion with silent myocardial ischemia.

Left ventricular cavity-to-myocardial count ratio: a new parameter for detecting resting left ventricular dysfunction directly from tomographic thallium perfusion scintigraphy.

Patients with reduced left ventricular function or aneurysms have cavities that appear dark on SPECT thallium scintigrams. We hypothesized that a quantitative index, which relates thallium activity in the left ventricular cavity to that in the myocardium (C/M ratio), could provide information on left ventricular function. A group of 80 patients who had both exercise SPECT thallium imaging and cardiac catheterization were studied. The C/M ratio was obtained from the short-axis tomogram on both exercise and rest images. Counts in a 2 x 2 pixel region of interest in the left ventricular cavity were divided by the number of counts in the "hottest" area of the myocardium. Plotting the angiographically determined ejection fraction against the C/M exercise and rest ratios, we observed a linear correlation between ejection fraction and both C/M ratios, r = 0.65 for C/M exercise and r = 0.67 for C/M rest ratio (p < 0.00001). Using data from 12 normal cardiac catheterization patients, we established the lower limit of normal; 50% for ejection fraction and 0.40 for the C/M ratios. A C/M exercise ratio < or = 0.40 identified 26 of 31 patients with an ejection fraction < or = 50%. A C/M exercise ratio > 0.40 identified 39 of 49 patients with an ejection fraction > 50%. These calculations yielded a sensitivity of 83% and specificity of 78% for the C/M exercise ratio. A similar analysis for C/M rest ratio revealed sensitivity of 61% and specificity of 92%. The present study shows that an abnormal C/M ratio correctly distinguishes patients with abnormal from normal ejection fractions with an accuracy of 81%. The C/M ratio is easily obtained, requires minimal processing time and is highly reproducible. These attributes may enable this index to add supplementary information regarding left ventricular function in addition to perfusion from thallium imaging.

Detection of coronary collaterals using dipyridamole PET myocardial perfusion imaging with rubidium-82.

UNLABELLED: This study evaluated the ability of dipyridamole PET myocardial perfusion imaging to detect coronary collaterals. A previous study showed an association between dipyridamole-induced coronary steal on PET imaging and the presence of coronary collaterals on angiography. METHODS: Dipyridamole PET myocardial perfusion imaging using 82Rb was performed in 45 patients who had recent coronary angiography. The stress/rest count ratio (rubidium activity with stress divided by activity at rest)-was used to express the change in regional tracer uptake with dipyridamole and was calculated manually and automatically. The accuracy of the stress/rest count ratio for detecting coronary collaterals was determined. RESULTS: A manual stress/rest count ratio < or = 0.80 identified coronary collaterals with 81% sensitivity, 92% specificity and 90% accuracy (p < 0.0001). An automated ratio < or = 0.80 had 90% sensitivity, 88% specificity and 90% accuracy (p < 0.0001). Vascular beds incorrectly identified by PET as having collaterals had an increased frequency of severe stenoses and abnormal wall motion. CONCLUSION: PET perfusion imaging using the stress/rest count ratio can serve as a unique imaging method to identify coronary collaterals noninvasively.

Defects on SPECT "perfusion" images can occur due to abnormal segmental contraction.

UNLABELLED: Technetium-99m-sestamibi images reflect tracer distribution at the time of injection. This "stay put" indicator allowed us to separate the effects of segmental left ventricular dysfunction per se versus myocardial blood flow on SPECT "perfusion" images in ten dogs. METHODS: An electromagnetic flow probe and hydraulic occluder were placed on the LAD coronary artery. Sonomicrometry was used to measure segmental wall shortening. At peak myocardial blood flow induced by adenosine, 35-45 mCi 99mTc-sestamibi were injected without occlusion. At 1 hr postinjection, during normal contraction, 40-50 msec end-diastolic and end-systolic SPECT images (#1) were acquired to reflect normal myocardial blood flow distribution. Later, during total LAD occlusion, and without reinjection of isotope, another gated scan (#2) was acquired. RESULTS: Coincident with abnormal contraction, large severe systolic defects [(28 +/- 5)% more severe compared to the baseline-scan #1; p < 0.01], and milder diastolic defects [(12 +/- 8)% more severe compared to the baseline-scan #1; p < 0.01] were observed during scan #2. Thus, abnormal contraction alone produced defects on SPECT images. CONCLUSION: Accordingly, defects in myocardial perfusion images must be interpreted as representing the integrated result of the combination of blood flow and segmental contraction heterogeneity.

Response of the right ventricle to exercise in isolated mitral stenosis.

Eight patients in sinus rhythm, with varying degrees of isolated mitral stenosis (mitral valve area 0.6 to 1.3 cm2 and total pulmonary vascular resistance 5.0 to 17.5 U-m2), underwent supine rest and symptom-limited exercise radionuclide ventriculography to determine right ventricular (RV) and left ventricular ejection fraction (EF). Cardiac catheterization with hemodynamic measurements at rest and at peak exercise was performed within 24 hours of radionuclide ventriculography. Four of the 8 patients underwent corrective mitral surgery resulting in normal mean pulmonary artery pressures and total pulmonary vascular resistance at rest. These 4 patients had repeat radionuclide ventriculography at rest and during exercise 1 to 2 months after surgery. Preoperatively, all 8 patients had an abnormal exercise RVEF response (mean change +/- standard deviation [SD], -5.0 +/- 4.5%), coincident with an increase in mean pulmonary artery pressure during exercise (mean change, 15 +/- 5.0 mm Hg). The change in RVEF from rest to exercise, corrected for duration of exercise, correlated with peak exercise mean pulmonary artery pressure (r = -0.71, p = 0.05), as well as total pulmonary vascular resistance at rest (r = -0.82, p = 0.02). Postoperatively, all 4 patients who underwent surgical correction showed a normal RVEF response during exercise (mean change +/- SD, +6.8 +/- 4.0%). Thus, in patients with acquired mitral stenosis and no coronary artery disease (1) loading conditions and not contractility are prime determinants of RV exercise response, and (2) an exercise-induced decrease in RVEF may be a sensitive marker for increased total pulmonary vascular resistance and pulmonary hypertension.

Relation of late potentials to ejection fraction and wall motion abnormalities in acute myocardial infarction.

A prospective study was performed to determine the relation between quantitative signal-averaged parameters and ejection fraction (EF) and wall motion abnormalities determined by radionuclide ventriculography in patients with acute myocardial infarction (AMI). In 50 patients with AMI, signal-averaging of the surface QRS complex (200 beats; filter frequencies of 40 to 250 Hz and 80 to 250 Hz) was performed and radionuclide ventriculograms were recorded 8 +/- 5 days after AMI. Twenty-five of these patients (50%) had anterior wall AMI, 20 (40%) had inferior wall AMI and 5 (10%) had non-Q-wave AMI. The duration of the low-amplitude signals of less than 40 microV, the signal-averaged QRS complex and the root-mean-square voltage of the terminal 40 ms were determined. In addition to EF determinations, wall motion abnormalities were assessed for the presence or absence of dyskinetic, akinetic and hypokinetic segments. A wall motion score was constructed by separating the left and right ventricles into 21 segments in the anterior, left anterior oblique and lateral views. On the basis of the presence or absence of late potentials, the patients were separated into 2 groups: group I comprised 15 patients (30%) with late potentials and group II 35 patients (70%) without late potentials. The low-amplitude signals (49 +/- 12 vs 24 +/- 8 ms) and the signal-averaged QRS complex (122 +/- 20 vs 96 +/- 15 ms) were significantly longer and the root-mean-square voltage (13.8 +/- 4.9 vs 54.3 +/- 27.4 microV) significantly lower in group I than in group II.(ABSTRACT TRUNCATED AT 250 WORDS)

Can noninvasive exercise test criteria identify patients with left main or 3-vessel coronary disease after a first myocardial infarction?

This study attempts to determine whether exercise treadmill testing with clinical, electrocardiographic, and thallium-201 myocardial perfusion imaging data can identify which patients have left main or 3-vessel (anatomically high-risk) coronary artery disease (CAD) after their first transmural myocardial infarct (MI). Twelve exercise test criteria for high-risk disease were compared in 40 patients referred for cardiac catheterization; 34 had a history of chest pain and 17 had angiographically defined high-risk CAD. A thallium image defect outside the vascular distribution of the MI was the most reliable criterion to distinguish patients with high-risk CAD (p = 0.00052 for Fisher's exact test of discrimination). Thallium imaging was somewhat more sensitive (92 versus 65%, p = 0.108) when patients with negative thallium imaging criteria who failed to achieve 85% of the age-predicted maximal heart rate were excluded. Failure to achieve 85% of predicted heart rate was by itself a useful criterion for detecting high-risk CAD (p = 0.017), especially in patients not taking propranolol (p = 0.004). Development of positive S-T segment depression at less than 70% predicted heart rate also discriminated left main or 3-vessel disease from less extensive CAD (p = 0.016). Other criteria failed to discriminate significantly between high-risk and less extensive CAD in patients after their first MI (p greater than 0.05). S-T segment depression (p = 0.199) or chest pain (p = 0.577) during exercise testing were particularly unreliable. Further, none of the criteria for high-risk CAD were influenced by irreversible left ventricular dysfunction. It is concluded that patients with thallium imaging defects outside the region of the infarct, decreasing blood pressure during exercise, failure to achieve 85% of predicted heart rate, or S-T depression at less than 70% of predicted heart rate have a high probability of having left main or 3-vessel disease. Patients without these criteria have a very low probability of having high-risk CAD and probably do not need coronary angiography for the purpose of excluding these high-risk coronary lesions after a first MI.