Author + information
- Received September 22, 1998
- Revision received September 3, 1999
- Accepted October 19, 1999
- Published online February 1, 2000.
- Jennifer M.F Kwok, MB, ChB, MRCPa,b,
- Timothy F Christian, MD, FACCa,b,*,
- Todd D Miller, MD, FACCa,b,
- David O Hodge, MSa,b and
- Raymond J Gibbons, MD, FACCa,b
- ↵*Reprint requests and correspondence: Dr. Timothy F. Christian, Mayo Clinic E-16B, 200 First Street SW, Rochester, Minnesota 55902
The aim of this study was to determine which clinical, exercise and thallium variables can aid in the identification of three-vessel or left main coronary artery disease (3VLMD) in patients with one abnormal coronary territory (either a reversible or fixed defect) on exercise thallium testing and to test the prognostic value of these variables.
Although the sensitivity of detection of coronary artery disease by thallium-201 imaging is high, the actual detection of 3VLMD by thallium tomographic images alone is not optimal.
A multivariate model for prediction of 3VLMD was developed from several clinical, exercise and thallium-201 variables in a training population of 264 patients who had one abnormal coronary artery territory on exercise thallium testing and had undergone coronary angiography. Using this model, patients were stratified into risk groups for prediction of 3VLMD. A separate validation cohort of 474 consecutive patients who were treated initially with medical therapy and who had one abnormal coronary territory were divided into identical risk groupings by the variables derived from the training population, and they were followed for a median of 7.0 years to evaluate the prognostic value of this model.
The prevalence of 3VLMD was 26% in the training population despite one abnormal thallium coronary territory. Four clinical and exercise variables—diabetes, hypertension, magnitude of ST segment depression, and exercise rate-pressure product—were found to be independent predictors of 3VLMD. In the training population, the prevalence of 3VLMD in low-, intermediate- and high-risk groups was 15%, 22% and 51%, respectively. When the multivariate model was applied to the validation population, the eight-year overall survival rates in the low-, intermediate- and high-risk groups were 89%, 73% and 75%, respectively (p < 0.001).
A substantial proportion of patients with one abnormal thallium coronary territory have 3VLMD with subsequent divergent outcomes based upon risk stratification by clinical and exercise variables. Consequently, the finding of only a single abnormal coronary territory by thallium-201 perfusion imaging does not necessarily confer a benign prognosis in the absence of consideration of nonimaging variables.
Identifying the presence of a disease by a diagnostic test is not equivalent to predicting the severity of the disease. Although the sensitivity for the detection of coronary artery disease (CAD) in patients with three-vessel or left main disease (3VLMD) by exercise perfusion imaging is exceedingly high, the actual prediction that such anatomy is present in a particular patient can be suboptimal if based on perfusion images alone (1–11). A prior study has shown that only 29% of patients with three-vessel CAD had three abnormal thallium coronary territories, but exercise and clinical variables added significant independent information to perfusion imaging for predicting the presence of 3VLMD (8). Because revascularization in a substantial proportion of patients with 3VLMD will decrease mortality (12–15), identification of these patients can have a beneficial impact on outcome. Consequently, the prediction of 3VLMD has important clinical implications.
The aim of this study was to develop a model based upon clinical and exercise variables that could aid in the detection of 3VLMD in patients who had perfusion scans with only a single abnormal coronary territory. This model was developed from a cohort of such patients who underwent coronary angiography and applied to a second population of such patients to predict subsequent mortality and cardiovascular events. This approach assumes that patients with 3VLMD are more likely to have subsequent cardiac events.
The study population consisted of a training cohort and a validation cohort. The training cohort was comprised of a consecutive series of 264 patients with suspected or known CAD who had exercise (following the Bruce or Naughton protocols) thallium-201 scintigraphy done at the Mayo Clinic between January 1987 and December 1989. All patients had only one abnormal coronary territory by exercise thallium imaging (either a reversible or fixed defect), and had coronary angiography within six months of the exercise thallium test. There were 4,481 patients who underwent exercise thallium testing during this time and 1,602 had one abnormal coronary territory (36%). Exclusion criteria were as follows: 1) left bundle branch block, preexcitation syndrome on the resting electrocardiogram or permanent pacemaker; 2) valvular, congenital or cardiomyopathic disease; 3) previous coronary revascularization by either percutaneous transluminal coronary angioplasty (PTCA) or coronary artery bypass grafting (CABG); 4) any intervening cardiac events (revascularization procedure, myocardial infarction) between the two studies; 5) digoxin use within two weeks of the exercise study; and 6) technically inadequate perfusion study. The final training population consisted of 264 patients.
In an effort to avoid the referral bias inherent in studying a population of patients referred for coronary angiography, a second population was studied for outcomes. The validation sample consisted of a separate group of 474 consecutive patients who had one abnormal coronary territory by exercise thallium-201 tomographic imaging at the Mayo Clinic between January 1988 and August 1989. Referral to cardiac catheterization was not an inclusion or exclusion criteria in this group of patients. Exclusion criteria were identical to the training cohort.
Clinical data concerning cardiac history and physical examination for all patients were obtained prospectively before testing and entered into a laboratory database. Diabetes was defined as a fasting glucose level >140 mg/dl, or chronic use of insulin or oral hypoglycemic agents and analyzed by insulin and noninsulin use. Hypertension was defined as a systemic blood pressure >140/90 mm Hg on repeated measurements, or the chronic use of antihypertensive medication (defined as treated hypertension). Hypercholesterolemia was defined as a total cholesterol level >250 mg/dl or chronic use of a cholesterol-lowering agent.
Treadmill exercise testing
All patients underwent symptom-limited treadmill exercise testing using either the Bruce or Naughton protocol (see Table 1for percentages). For patients performing the Naughton protocol, a conversion factor was applied to equate exercise duration with the Bruce protocol (16). Twelve-lead electrocardiography and heart rate were recorded at rest, every minute during exercise, at the peak of exercise and every 3 min of recovery. Maximum ST depression was defined as the maximum change of ST-segment depression at 80 ms after the J point during or after exercise. Maximum ST depression was categorized into groups: <1.0 mm, 1.0 mm to 1.4 mm, 1.5 mm to 1.9 mm, 2.0 mm to 2.4 mm and >2.4 mm. Systemic blood pressure was recorded by cuff sphygmomanometry before exercise, at the end of each stage and at peak exercise. At peak exercise, 4 mCi of thallium-201 were injected intravenously and patients were encouraged to exercise for another minute (this study preceded the introduction of thallium reinjection into our laboratory).
Single-photon emission computed tomography (SPECT) imaging
Images were acquired while the patient was in the supine position, using a large field-of-view, single-crystal, rotating gamma camera (Elscint 409) with an all-purpose, parallel hole collimator and previously described methodology (8). A single 5-min anterior planar image was started at 5 to 10 min postinjection to assess cardiac size and pulmonary activity. Tomographic imaging was then performed from 45° right anterior oblique to 45° left posterior oblique, using the step-and-shoot method over a 180° arc; 30 images were obtained at 6° intervals for 40 s each. Each projection was corrected for nonuniformity, and filtered-back projection was performed with a Ramp-Hanning filter. Orthogonal images were generated by oblique angle reconstruction producing horizontal long-axis, short-axis and vertical long-axis slices that were each 6 mm thick. Redistribution thallium images were acquired in a similar fashion 4 h postexercise without reinjection of thallium.
Visual analysis of tomographic images
Two experienced observers interpreted the tomographic images qualitatively by consensus. Cardiac enlargement and pulmonary uptake were assessed qualitatively as present or absent from the anterior planar image. Pulmonary uptake was quantified in borderline cases (8). Uptake was considered to be elevated if the pulmonary/myocardial counts were >0.5. Short-axis images at the apex, middle and base of the left ventricle were divided into anterior, septal, inferior and lateral segments. The septal segments of middle and basal slices were further divided into anterior and inferior portions (Fig. 1). Consequently, the left ventricle was divided into 14 short-axis segments as previously described (17). Each segment was assessed with a 5-point scoring system (4 = normal perfusion; 3 = mild, 2 = moderate, 1 = severe hypoperfusion; 0 = absent perfusion). A reversible defect was defined as ≥1 grade improvement in any segment on the delayed image when compared to the stress image. A fixed defect had to be of at least moderate hypoperfusion (grade 2) to be classified as abnormal. Mild fixed defects were considered normal, likely reflecting soft tissue attenuation artifacts.
The 14 short-axis segments were assigned to three coronary territories as demonstrated in Figure 1. Defects at the middle and base of left ventricle were classified according to the vascular territories of the three major coronary arteries: anterior and anteroseptal defects represented disease in the left anterior descending artery, inferior and inferoseptal defects represented disease in the right coronary artery and lateral defects represented disease in the left circumflex artery. Because of the variation of blood supply to the apex, the apex was not used for coronary artery assignment. However, pure apical defects were considered as one abnormal coronary territory but were not assigned to any specific coronary territory. The presence of only one abnormal coronary territory was defined as a reversible or fixed perfusion defect that occurred in one vascular territory. In the presence of overlap between coronary territories, minimal extension of a perfusion defect into the adjacent territory was considered as a perfusion defect of the predominant territory. This ambiguity was confined to septal segments where the arterial supply is variable. If both septal segments were abnormal, the patient was still classified as having one abnormal territory if only one of the two contiguous segments (anterior or inferior) was also abnormal.
All patients in the training population underwent coronary angiography within six months of thallium imaging (median time, 1 day; range, 177 days before to 167 days after). Coronary artery narrowing was visually assessed and reported as percentage luminal diameter stenosis. Significant coronary stenosis was defined as ≥70% narrowing of the internal diameter of the left anterior descending artery the left circumflex artery, the right coronary artery, or their major branches; and ≥50% narrowing of the left main coronary artery (12). Additionally, an angiographic jeopardy score was calculated for each patient using two prognostically validated scoring methods (Califf et al. , and Mark et al. ). These systems have shown increased five-year event rates (18)and greater survival benefit from coronary revascularization surgery (19)for patients with higher scores.
The study was designed so that follow-up was not initiated until at least five years from the thallium stress test. Consequently, patients with follow-up under five years represent those with events qualifying as an end point or a censoring criteria. The validation population was followed by mailed questionnaire, telephone interview or review of the patient’s medical record or physician contact. Events were defined as death, cardiac death, myocardial infarction (MI) and late revascularization (>90 days from study). Death was confirmed by reviewing the hospital chart, a death certificate or a clinician’s report. The cause of death of each patient was coded as cardiac or noncardiac by a reviewer blinded to the baseline information and the exercise thallium results. Myocardial infarction was defined on the basis of chest pain, electrocardiographic (ECG) changes and elevated serum creatine kinase or isoenzyme levels. Revascularization included PTCA or CABG.
The Student ttest or chi-square test for independence was performed to identify variables significantly associated with 3VLMD for the training population, all of whom were referred for coronary angiography (a p value <0.05 was considered significant). Variables considered for this analysis are shown in Table 1. Logistic regression analysis was used to develop a multivariate model for predicting the presence or absence of 3VLMD. All variables from Table 1were considered for stepwise inclusion in the model until variables not entered into the model were not significant at the p < 0.10 level. The regression model was then utilized to estimate the probability of 3VLMD for each patient by classifying them as low (<15%), intermediate (15% to 35%) and high (>35%) probability. These probability groupings were chosen prospectively to match prior studies for the noninvasive prediction of 3VLMD from this laboratory (8,20,21). We then applied the multivariate model developed from the training population to the validation population. Each patient from this group was placed into a risk group in an identical manner as with the training cohort. This approach assumes that patients predicted to be at higher risk for 3VLMD (and higher jeopardy scores) will have a higher event rate than those predicted to be at lower risk for 3VLMD.
Four end points were analyzed, and the Kaplan-Meier method was used to estimate event-free survival. Three survival curves (for high-, intermediate-, and low-risk patients) were generated for each of the four end points: 1) overall survival (no patients censored from analysis); 2) survival free of cardiac death (censoring of patients with noncardiac death or PTCA/CABG at any time after exercise testing); 3) survival free of cardiac death and nonfatal MI (censoring of patients with noncardiac death or PTCA/CABG at any time after exercise testing); 4) survival free of cardiac death, nonfatal MI and late revascularization (censoring of patients with noncardiac death or early PTCA/CABG). The log-rank test was used to compare each set of curves.
Comparison of the training and validation populations by univariate analysis
Clinical characteristics, exercise variables and thallium variables of the two populations are shown in Table 1. The training population was a slightly “sicker” cohort; they had a higher prevalence of typical angina, a worse exercise performance (a lower peak exercise heart rate, peak exercise rate-pressure product and exercise workload) and a higher prevalence of exercise-induced chest pain and exercise-induced ST-segment depression of >1 mm. There were similar proportions of patients in the training and validation groups with cardiac enlargement or elevated pulmonary uptake. The mean total number of abnormal post-stress thallium segments was five in the training group and four in the validation group (p < 0.001). A higher proportion of patients in the training population had a left anterior descending artery territory defect. An isolated apical defect was present in none of the training population patients and 56 (12%) of the validation population.
Training population (development of the model)
Overall, 68 patients (26%) had 3VLMD. Five patients had isolated left main disease, 49 patients had three-vessel but not left main disease, 14 patients had both three-vessel and left main disease. The majority of left main lesions (14/19) were between 50% to 70% luminal obstruction. The prevalence of two-vessel, single-vessel and no disease were 30%, 27% and 17%, respectively. The median Califf jeopardy score was 4 and the median Mark score was 32.
The results of the univariate analysis for clinical, exercise and thallium variables are shown in Table 2. Significant differences existed in multiple clinical and exercise variables between patients with 3VLMD and without. Patients with 3VLMD had a significantly higher prevalence of diabetes, hypertension and history of typical angina. During exercise, patients with 3VLMD had a higher prevalence of ST-segment depression, achieved a lower peak heart rate, a lower peak rate-pressure product and a smaller increase in systolic blood pressure. Cardiac enlargement and pulmonary/myocardial count ratio were not associated with 3VLMD. The presence of an anterior defect and the number of abnormal post-stress segments (a measure of defect extent) did not help to identify patients with 3VLMD.
A multivariate logistic regression analysis was employed to develop a model for the prediction of 3VLMD based upon the clinical, exercise and thallium variables listed in Table 3. The magnitude levels of exercise-induced ST depression, peak rate-pressure product, diabetes and hypertension were the only independent predictors of 3VLMD in patients with one abnormal coronary territory by exercise thallium-201 imaging. No thallium scintigraphic variables, including cardiac enlargement, elevated pulmonary uptake, number of hypoperfused segments, anterior location, or presence of defect reversibility, added independently for the prediction of severe CAD in this cohort. The overall chi-square value of the model was 23.7 and the individual chi-square values are listed in Table 3.
Risk stratification for the training population
A predicted probability for 3VLMD was calculated for each patient by the regression equation derived from the multivariate model. Patients were then placed into risk groups defined as low (<15%), intermediate- (15 to 35%) and high probability (>35%) for 3VLMD. The prevalence of 3VLMD in low-, intermediate-, and high-risk groups was 15%, 22% and 51%, respectively (Table 4). Consequently, >50% of patients defined as high-risk actually had 3VLMD despite one abnormal thallium territory. The angiographic jeopardy scores by the Califf method (18)and Mark method (19)demonstrated a monotonic increase as the risk of severe CAD increased, which was highly significant.
To express the predictive value of this model in a clinically usable format, nomograms based on the variables from the model were generated to estimate the risk of 3VLMD in individual patients (Fig. 2). Based on the absence or presence of diabetes and hypertension, patients can be classified into four categories: those who have neither diabetes nor hypertension, those who have diabetes alone, those who have hypertension alone and those who have both diabetes and hypertension. The probability of 3VLMD for each patient can be estimated by applying the exercise variables—magnitude of maximum ST segment depression and rate-pressure product. For example, a normotensive, nondiabetic patient with one abnormal thallium coronary territory who has a rate-pressure product of 25,000 and <1 mm ST depression is at low risk to have 3VLMD. A patient with both diabetes and hypertension is automatically classified as at least intermediate risk.
Validation population (testing of the model)
The validity of these independent predictors to detect severe disease was tested in the validation population to predict outcome. Follow-up was initiated at least five years from the exercise study. The median follow-up period was seven years, with a range of 19 days (reflecting early events or censoring) to 11.5 years, and was 95% complete. Twenty-three patients were lost to follow-up, and two patients excluded during follow-up because of missing peak exercise data. As a result, 449 patients qualified for survival analysis. Using the multivariate model derived from the training population, each patient in the validation population was classified into low-, intermediate-, or high-risk group for 3VLMD. The outcome of the patients by risk group are shown in Table 5.
Figure 3shows the Kaplan-Meier survival curves of 1) overall survival; 2) survival free of cardiac death; 3) survival free of cardiac death and nonfatal MI; and 4) survival free of cardiac death, nonfatal MI and late revascularization. The survival curves for the validation population demonstrate a significant association between risk stratification by the multivariate model (based upon angiographic results) and subsequent outcome in patients with one abnormal coronary perfusion territory on exercise thallium-201 imaging with the exception of the end point of cardiac death or MI. The eight-year overall survival rates in the low-, intermediate- and high-risk group patients were 89%, 73% and 75%, respectively (p < 0.001). The eight-year survival free of cardiac death, and cardiac death, MI or late revascularization also had significantly divergent outcomes based upon risk stratification from the model derived from the angiographic cohort (Fig. 3A–D).
Identification of 3VLMD by perfusion scintigraphy
Although the sensitivity of exercise thallium myocardial perfusion imaging for detecting CAD in patients with severe disease is excellent, only a minority of patients with three-vessel CAD have perfusion defects in all three coronary territories. Christian et al. (8)reported that only 29% of patients with three-vessel CAD had perfusion defects in all three coronary territories. Iskandrian et al. (11)showed that 39% of patients with three-vessel CAD had perfusion defects in all three coronary territories by exercise thallium imaging if they achieved an adequate exercise level.
For several reasons, we chose to study the scenario of patients with one abnormal territory and severe CAD. The prevalence of severe CAD with this finding is not low (26% in this study), whereas it is uncommon in patients with normal images (<4% of patients referred to angiography ). Patients with multiterritory defects are more likely to be referred to angiography; patients with one abnormal territory may be medically managed on the presumption of single-vessel disease. This strategy may be appropriate for the majority of patients, but a cohort will be missed where revascularization has been shown to alter prognosis (12–15). The present study was designed to detect those patients.
Identification of 3VLMD and prediction of outcome in this study
Because patients with 3VLMD have a higher cardiac event rate, we attempted to validate a multivariate model for predicting 3VLMD (derived from the angiographic cohort) by prospectively applying it to a separate, unbiased cohort of patients with one abnormal coronary territory for the prediction of outcome. The prevalence of 3VLMD in the angiographic cohort ranged from 15% in the low-risk group to 51% in the high-risk group, and the jeopardy scores increased as the predicted risk of severe CAD increased. Patients with a Califf jeopardy score of 6 (the median value in the high-risk group) have an event rate fivefold greater than do patients with a score of 2 (the median value in the low-risk group) (18).
Patients in the validation population were stratified into risk groups by the training model. Significant differences existed between the low-risk group and the intermediate and high-risk groups in terms of 1) overall survival, 2) survival free of cardiac death and 3) survival free of cardiac death, nonfatal MI, or late revascularization. However, it should be emphasized that the overall prognosis in this cohort was good, consistent with the limited extent of scintigraphic ischemia defined by this study. There were few cardiac deaths and MIs (8% and 3% of the validation population, respectively). Consequently, no significant difference was seen among the three risk groups concerning survival free of cardiac death or nonfatal MI. This lack of a significant difference may be related to the subsequent occurrence of nonfatal MI as a result of plaque rupture of coronary lesions of <50% in diameter, which may not be detected by perfusion imaging, as well as the limited number of such events.
Parameters such as perfusion defect extent, which have been strong predictors of outcome in multiple prior studies, were constrained in range by the exclusion of patients with more than one abnormal territory. Consequently, they did not add significantly to the model for the prediction of 3VLMD in the training population as might have been expected. The high degree of random variability of multivariate models when constructed from different populations has been well described (22). Consequently, we did not pursue a “better fit” multivariate model to predict outcome in this group. Interestingly, neither cardiac enlargement nor increased pulmonary thallium-201 uptake was predictive of 3VLMD in the angiographic cohort. Prior studies employing tomographic imaging (including this laboratory) have shown these variables to be a univariate predictor of severe CAD but not an independent predictor once other measures of ischemia are adjusted for (8,11). This may be due to the constraint on the extent of ischemia required by this study.
Previous studies for the detection of 3VLMD by exercise thallium imaging
The evidence that clinical and exercise variables used in conjunction with stress thallium scintigraphy improve accuracy for the prediction of 3VLMD is strong (4–10). Maddahi et al. (4)reported that quantitative thallium stress distribution and washout, or exercise hypotension response, or marked stress electrocardiographic response, detected 86% of patients with 3VLMD by planar thallium study. Christian et al. (8)concluded that magnitude of ST depression with exercise is the most important independent predictor of 3VLMD, followed by the number of abnormal thallium segments, diabetes and the change in systolic blood pressure with exercise. Iskandrian et al. (9)found that the most important independent predictor of 3VLMD was a multivessel thallium abnormality, but exercise heart rate and ST-segment depression were also independent predictors. In patients with normal resting ECGs, the incremental value of thallium imaging for predicting 3VLMD over clinical and exercise variables is minimal (20).
This study supports and extends previous observations that the clinical and exercise variables are useful for the prediction 3VLMD and outcome (23–25). Mark et al. (26)emphasized the importance of combining test variables to strengthen the overall predictive value of exercise treadmill testing. The present study reaffirms the value of such an approach for exercise perfusion imaging when applied to specific conditions.
One abnormal thallium coronary territory with 3VLMD
Several potential mechanisms explain the presence of only one abnormal perfusion coronary territory in the setting of 3VLMD. First, exercise may be stopped owing to a single severe lesion producing symptoms or signs necessitating test termination. While thallium perfusion imaging measures flow heterogeneity, the relative decrease of coronary blood flow in the other diseased vessels may be insufficient at that point to produce a perfusion defect. This is evident by pooled data showing that only 69% of the diseased vessels can be detected in patients with three-vessel disease, as compared to 83% of the diseased vessels in patients with single-vessel disease (27). The sensitivity for detection of disease in an individual coronary artery differs, particularly in the left circumflex territory where the detection of disease is reduced compared to the other two regions (28). This may be partially responsible for producing a single abnormal territory in the setting of moderate left main disease. Relative hypoperfusion of myocardium can be well appreciated by thallium scintigraphy. Nevertheless, in patients with three-vessel disease, the two less ischemic coronary territories may have a similar degree of impaired flow reserve, resulting in apparently only one abnormal thallium coronary territory despite three-vessel disease.
The number of abnormal post-stress perfusion segments tended to be higher in patients with 3VLMD, but the difference was not significant. This was likely due to the constraints on the extent of abnormal segments dictated by the study design. Additionally, there was a somewhat higher proportion of patients with anterior defects in the non-3VLMD group, which often involves the apex (an additional four abnormal segments), as opposed to nonanterior defects.
This study was performed at a tertiary referral center. Consequently, the prevalence of severe coronary artery disease was likely to be higher than that in community-based hospitals. To identify the ability of a noninvasive test to predict 3VLMD, the model was developed on a test group of patients who all underwent coronary angiography. This posttest referral bias was unavoidable. It was for this reason that we applied the results of the angiographic population to an independent population, which was less selected but also “less sick.” Although there is some inherent inaccuracy in this approach in comparison to applying the model to an unselected population of patients randomized to undergo coronary angiography, it is preferable to no validation procedure. The limitations of using coronary angiography (which reflects anatomical severity of CAD) as the gold standard for comparison with perfusion scintigraphy (which reflects functional coronary perfusion) are well recognized (29).
Transient ischemic dilation, which is an independent determinant of prognosis and is associated with 3VLMD, was not routinely assessed in our patients. Washout rate quantitation, which may have helped to identify more severe disease, was not employed owing to the use of tomographic acquisition techniques. We did not consider mild fixed defects as constituting an abnormal territory as the vast majority of these represent attenuation artifacts. However, there is evidence that the event rate in patients with mild scintigraphic abnormalities (without gated imaging) is increased compared to those with clearly normal scans (30). Multiple variables were considered for the generation of the multivariate model for the prediction of 3VLMD, which might result in some “overfitting” of the model, but application to an independent patient cohort minimizes the impact of this phenomenon.
Patients with one-vessel coronary artery disease generally have an excellent prognosis, with an annual mortality rate of approximately 1% (19). A substantial minority of patients with only one abnormal thallium territory have 3VLMD, and they have divergent prognoses. Patients with one abnormal coronary territory should not be presumed to have single-vessel disease and, therefore, a benign clinical outcome. For low-risk patients, the prevalence of 3VLMD was low and mortality was comparable to that of patients with angiographically demonstrated single-vessel disease (13). Therefore, the clinical decision to refer a patient from the low-risk group to coronary angiography should be based primarily on the patient’s symptoms.
In contrast, intermediate- and high-risk patients have a considerable prevalence of 3VLMD, and a higher mortality. Coronary angiography should be considered in such patients. The use of gating of the tomographic images for regional wall motion, which is becoming a common practice, may provide superior identification of patients with 3VLMD by resolving the etiology of mild fixed defects. This possibility needs to be explored.
☆ This study was supported by the Mayo Foundation.
- coronary artery bypass grafting
- coronary artery disease
- myocardial infarction
- percutaneous transluminal coronary angioplasty
- single-photon emission computed tomography
- three-vessel or left main disease
- Received September 22, 1998.
- Revision received September 3, 1999.
- Accepted October 19, 1999.
- American College of Cardiology
- DePasquale E.E,
- Nody A.C,
- Depuey E.G,
- et al.
- Maddahi J,
- Abdulla A,
- Garcia E.V,
- Swan H.J.C,
- Berman D.S
- Nygaard T.W,
- Gibson R.S,
- Ryan J.M,
- Gascho J.A,
- Watson D.D,
- Beller G.A
- Dash H,
- Massie B.M,
- Botvinick E.H,
- Brundage B.H
- Iskandrian A.S,
- Heo J,
- Lemlek J,
- Ogilby J.D
- Chae S.C,
- Heo J,
- Iskandrian A.S,
- Wasserleben V,
- Cave V
- Iskandrian A.S,
- Heo J,
- Kong B,
- Lyons E
- Alderman E.L,
- Bourassa M.G,
- Cohen L.S,
- et al.
- Participants in the Study Group,
- Takaro T,
- Hultgren H.N,
- Lipton M.J,
- Detre K.M
- Pollock M.L,
- Wilmore J.H
- Califf R.M,
- Phillips H.R,
- Hindman M.C,
- et al.
- Mark D.B,
- Nelson C.L,
- Califf R.M,
- et al.
- Hachamovitch R,
- Berman D.S,
- Kiat H,
- et al.