Frank E. Harrell

Prognostic value of a coronary artery jeopardy score

The prognostic value of a coronary artery jeopardy score was evaluated in 462 consecutive nonsurgically treated patients with significant coronary artery disease, but without significant left main coronary stenosis. The jeopardy score is a simple method for estimating the amount of myocardium at risk on the basis of the particular location of coronary artery stenoses. In patients with a previous myocardial infarction, higher jeopardy scores were associated with a lower left ventricular ejection fraction. When the jeopardy score and the number of diseased vessels were considered individually, each descriptor effectively stratified prognosis. Five year survival was 97% in patients with a jeopardy score of 2 and 95, 85, 78, 75 and 56%, respectively, for patients with a jeopardy score of 4, 6, 8, 10 and 12. In multivariable analysis when only jeopardy score and number of diseased vessels were considered, the jeopardy score contained all of the prognostic information. Thus, the number of diseased vessels added no prognostic information to the jeopardy score. The left ventricular ejection fraction was more closely related to prognosis than was the jeopardy score. When other anatomic factors were examined, the degree of stenosis of each vessel, particularly the left anterior descending coronary artery, was found to add prognostic information to the jeopardy score. Thus, the jeopardy score is a simple method for describing the coronary anatomy. It provides more prognostic information than the number of diseased coronary arteries, but it can be improved by including the degree of stenosis of each vessel and giving additional weight to disease of the left anterior descending coronary artery.(ABSTRACT TRUNCATED AT 250 WORDS)

Prognostic indicators from radionuclide angiography in medically treated patients with coronary artery disease

The purpose of this investigation was to determine which variables obtained when performing radionuclide angiography predict subsequent survival or total events (cardiovascular death or nonfatal myocardial infarction) in stable patients with symptomatic coronary artery disease (CAD). Univariable and multivariable analyses of 6 variables, including ejection fraction (EF) at rest and exercise, change in EF with exercise, development of ischemic chest pain or electrocardiographic changes, left ventricular (LV) wall motion abnormalities and exercise time were examined in 386 patients followed up to 4.5 years. Univariate analyses revealed that the exercise EF was the variable most closely associated with future events (p less than 0.01), followed by EF at rest, wall motion abnormalities and exercise time. Multivariable analyses revealed that once the exercise EF was known, no other radionuclide variables contributed independent information about the likelihood of future events. Multivariable analyses also revealed that the exercise EF describes much of the prognostic information of coronary anatomy. Our findings suggest that the radionuclide angiogram is useful in predicting future events in patients with stable CAD, although examination in conjunction with other clinical descriptors will be necessary to further quantify this contribution.

Comparison of predictions based on observational data with the results of randomized controlled clinical trials of coronary artery bypass surgery

Clinical decisions are most secure when based on findings from several large randomized clinical trials, but relevant randomized trial data are often unavailable. Analyses using clinical data bases might provide useful information if statistical methods can adequately correct for the lack of randomization. To test this approach, the findings of the three major randomized trials of coronary bypass surgery were compared with predictions of multivariable statistical models derived from observations in the Duke Cardiovascular Disease Databank. Clinical characteristics of patients at Duke University Medical Center who met eligibility requirements for each major randomized trial were used in the models to predict 5 year survival rates expected for medical and surgical therapy in each randomized trial. Model predictions agreed well with randomized trial results and were within the 95% confidence limits of the observed survival rates in 24 (92%) of 26 clinical subgroups. The overall correlation between predicted and observed survival rates was good (Spearman coefficient 0.73, p less than 0.0001). These results suggest that carefully performed analyses of observational data can complement the results of randomized trials.

Outcome in suspected acute myocardial infarction with normal or minimally abnormal admission electrocardiographic findings.

Seven hundred seventy-five consecutive patients with symptoms suggestive of acute myocardial infarction (AMI) who were admitted to the cardiac care unit from the emergency room were studied; 107 had normal electrocardiographic findings and 73 had only minimal nonspecific changes. AMI subsequently evolved in 11 patients (10%) with normal electrocardiographic findings and in 6 (8%) with minimal changes, compared with 245 (41%) with frankly abnormal emergency room findings. Only 1 (1%; 95% confidence limits 0.02 to 5%) and 4 (6%; 95% confidence limits 2 to 15%) of those with normal and nonspecific initial electrocardiographic findings, respectively, had a complication for which they potentially benefited from intensive care unit intervention, although many patients received prophylactic therapy. Thus, the initial emergency room electrocardiogram can effectively separate patients into high- and low-risk groups for AMI and serious complications. Admission to a monitored intermediate care ward may be an acceptable practice in patients with chest pain and a normal or minimally changed initial electrocardiogram.

Prognostic implications of ventricular arrhythmias during 24 hour ambulatory monitoring in patients undergoing cardiac catheterization for coronary artery disease

The prognostic importance of ventricular arrhythmias detected during 24 hour ambulatory monitoring was evaluated in 395 patients with and 260 patients without significant coronary artery disease. Ventricular arrhythmias were found to be strongly related to abnormal left ventricular function. A modification of the Lown grading system (ventricular arrhythmia score) was the most useful scheme for classifying ventricular arrhythmias according to prognostic importance. When only noninvasive characteristics were considered, the score contributed independent prognostic information, and the complexity of ventricular arrhythmias as measured by this score was inversely related to survival. However, when invasive measurements were included, the ventricular arrhythmia score did not contribute independent prognostic information. Furthermore, ejection fraction was more useful than the ventricular arrhythmia score in identifying patients at high risk of sudden death.

Prognostic value of ventricular arrhythmias associated with treadmill exercise testing in patients studied with cardiac catheterization for suspected ischemic heart disease

The prognostic information provided by ventricular arrhythmias associated with treadmill exercise testing was evaluated in 1,293 consecutive nonsurgically treated patients undergoing an exercise test within 6 weeks of cardiac catheterization. The 236 patients with simple ventricular arrhythmias (at least one premature ventricular complex, but without paired complexes or ventricular tachycardia) had a higher prevalence of significant coronary artery disease (57 versus 44%), three vessel disease (31 versus 17%) and abnormal left ventricular function (43 versus 24%) than did patients without ventricular arrhythmias. Patients with paired complexes or ventricular tachycardia had an even higher prevalence of significant coronary artery disease (75%), three vessel disease (39%) and abnormal left ventricular function (54%). In the 620 patients with significant coronary artery disease, patients with paired complexes or ventricular tachycardia had a lower 3 year survival rate (75%) than did patients with simple ventricular arrhythmias (83%) and patients with no ventricular arrhythmias (90%). Ventricular arrhythmias were found to add independent prognostic information to the noninvasive evaluation, including history, physical examination, chest roentgenogram, electrocardiogram and other exercise test variables (p = 0.03). Ventricular arrhythmias made no independent contribution once the cardiac catheterization data were known. In patients without significant coronary artery disease, no relation between ventricular arrhythmias and survival was found.

Type A behavior and survival: A follow-up study of 1,467 patients with coronary artery disease

Patients with documented coronary artery disease, admitted to Duke Medical Center between 1974 and 1980, were assessed for type A behavior pattern and were followed until 1984. The relation of type A behavior to survival was tested using data from coronary angiography to control for disease severity. Cox model regression analyses demonstrated an interaction (p less than 0.01) between type A behavior and an index of disease severity in the prediction of cardiovascular death. Among those with relatively poor left ventricular function, type A patients had better survival than type B. This difference was not present among patients with better prognoses. Type A behavior did not predict the subsequent incidence of nonfatal myocardial infarctions. Differential risk modification and differential selection into postinfarction status are possible explanations for the findings. These results need not conflict with the proposition that type A behavior plays a role in the pathogenesis of coronary artery disease.

The restricted cubic spline hazard model

A cubic spline hazard model where the tails are linearly constrained (Stone and Koo, 1985) has considerable flexibility in describing data which has been generated from distributions having a variety of hazard function shapes. This model is as efficient as the Kaplan-Meier (1958) estimator for estimating survival probabilities.

Prognostic effect of bundle branch block related to coronary artery bypass grafting

The incidence and prognostic effect of the development of new perioperative ventricular conduction abnormalities were examined in all patients undergoing coronary artery bypass surgery at Duke University Medical Center between 1976 and 1981. Of the 913 patients included, transient (resolved before discharge) ventricular conduction abnormalities developed in 156 (17%) and persistent (until discharge) changes developed in 126 (14%). Complete right bundle branch block (BBB) was the most frequent type of new ventricular conduction abnormality, followed by left anterior hemiblock and incomplete right BBB (found in 60%, 26%, and 9%, respectively, of all patients with transient changes and 29%, 33% and 26% of all patients with persistent changes). Development of new ventricular conduction abnormalities was most strongly related to date of operation (p less than 0.0001, univariate chi 2 = 122), increasing from 2% transient and 7% persistent in 1976 to 36% transient and 22% persistent in 1981. The incidence was also higher in older patients. Preoperative ejection fraction and number of diseased vessels were related to development of perioperative ventricular conduction abnormalities but were not independently related after adjustment for other baseline characteristics. Contrary to findings in other studies, development of new perioperative ventricular conduction abnormalities, including isolated new left BBB, did not worsen the survival rate in patients followed up to 3 years after surgery.

Clinical correlates and prognostic significance of type A behavior and silent myocardial ischemia on the treadmill

Type A patients with coronary artery disease (CAD) tend to ignore or underreport symptoms, especially during challenging tasks such as the treadmill exercise test. To determine whether type A CAD patients might be more likely than type B patients to have silent ischemia during exercise and consequently a worse prognosis, 403 patients with stable CAD who had significant coronary disease on angiography, a positive Bruce protocol treadmill test and a structured interview to assess type A behavior were studied. Median follow-up time was 6 years. Type A patients were more likely to experience silent ischemia during exercise than were type B patients (35 vs 25%, p = 0.05). Patients with silent ischemia during exercise had a history of fewer anginal episodes/week, and type A patients with silent ischemia were less likely to have had a history of typical angina. However, using the Cox model, there were no significant differences in survival between type A patients and B patients with silent ischemia (4-year survival 86 vs 79%, p = 0.44) and no significant differences in survival between type A patients with silent ischemia and type A patients with symptomatic ischemia (6-year survival 86 vs 80%, p = 0.59). Similar results were obtained for infarction-free survival. Type A patients are more likely than type B patients to have silent ischemia during exercise, but long-term survival is not affected.