- Email: [email protected]
Prognostic Value of Stress Echocardiography in the Evaluation of Atypical Chest Pain Patients Without Known Coronary Artery Disease
Prognostic Value of Stress Echocardiography in the Evaluation of Atypical Chest Pain Patients Without Known Coronary Artery Disease Percy J. Colon III, MD, Sameh K. Mobarek, MD, Richard V. Milani, MD, Carl J. Lavie, MD, Mark M. Cassidy, MD, Joseph P. Murgo, MD, and Jorge Cheirif,
MD
Patients with atypical chest pain frequently lack significant coronary artery disease (CAD) and are, therefore, at low risk for future adverse cardiovascular events. We hypothesized that in this group of patients, stress echocardiography could identify those at risk for cardiac events. We retrospectively reviewed (mean follow-up 23.0 6 7.2 months) the prognostic value of stress echocardiography for major (cardiac death, myocardial infarction, congestive heart failure, and unstable angina) and total (major events plus coronary revascularization) cardiac events in 661 patients with atypical chest pain, normal global left ventricular (LV) systolic function, and no history of CAD. A positive stress echocardiogram was defined as the development of new or worsening wall motion abnormalities with exercise stress (80%) or dobutamine (20%). A total of 41 cardiac and 16 major events were noted. The event-free survival for total car-
diac events was 97% for a normal stress echocardiogram and 93% for a normal stress electrocardiogram (ECG) at 30 months. A positive stress ECG predicted an event-free rate of 86% compared with 74% for stressinduced wall motion abnormalities and 42% if stressinduced LV dysfunction accompanied the wall motion abnormalities. A strategy recommending invasive studies based on positive stress echocardiogram results increased the per-patient cost, but led to greater savings per cardiac event predicted and provided incremental prognostic value for future cardiac events beyond clinical and stress electrocardiographic data. Thus, stress echocardiography in low-risk patients for CAD appears to be more cost effective than a stress ECG. Q1998 by Excerpta Medica, Inc. (Am J Cardiol 1998;81:545–551)
hoosing the optimal cost-effective strategy for evaluating patients with atypical chest pain reC mains controversial. These patients are at low risk for
either exercise (81%) or dobutamine (19%) stress echocardiography. After retrospective review of the medical records, 1,310 patients were excluded due to a history of CAD, typical angina, depressed left ventricular (LV) function (ejection fraction at rest ,50%), valvular heart disease, or cardiac transplantation. Of the remaining 688 patients, 27 (4%) were excluded due to inadequate clinical information or follow-up. Thus, the final cohort consisted of 661 patients at low risk for cardiac events. Stress protocol: Based on the patient’s physical abilities, cardiac function was assessed using either an exercise treadmill protocol or pharmacologic stress with intravenous dobutamine. Exercise treadmill testing was symptom limited using the Bruce, modified Bruce, or Ramp-based protocols. Pharmacologic stress was provoked by intravenous dobutamine, starting at 5 mg/kg/min and progressing to a maximum dose of 50 mg/kg/min at 3-minute intervals until the target heart rate (85% of maximum predicted heart rate for age) was achieved. If needed, $1 mg of atropine was given intravenously in divided intervals to attain the target heart rate. Clinical symptoms and blood pressure were monitored in the standard fashion throughout the study. Stress testing was terminated prematurely if the patient developed significant chest pain, serious ventricular arrhythmia, severe hypertension (systolic .220 mm Hg; diastolic .120 mm Hg), a progressive decrease in systolic blood pressure (,90 mm Hg), or $2 mm of ST-segment depression.
developing future cardiovascular events.1 The decision to recommend stress testing depends upon the physician’s initial clinical assessment, which is strongly influenced by the patient’s presentation2 as well as by the coronary risk factor profile. Stress echocardiography, a noninvasive diagnostic tool used to detect coronary artery disease (CAD), has grown in popularity, in part due to its superior accuracy over stress electrocardiography3–5 and its lower cost over radionuclide imaging. Although the prognostic value of stress echocardiography in high-risk patients for CAD has previously been established,6 –11 the potential value of stress echocardiography to assess prognosis in patients at low risk for CAD has not been thoroughly studied. The purpose of this study was to determine the prognostic value of stress echocardiography in a large cohort of low-risk patients referred to our laboratory for functional cardiac assessment.
METHODS
Study population: During 1993, 1,998 patients were referred to our laboratory for evaluation of CAD using From the Section of Cardiology, Department of Internal Medicine, Ochsner Medical Institutions, New Orleans, Louisiana. Manuscript received August 6, 1997; revised manuscript and accepted November 25, 1997. Address for reprints: Jorge Cheirif, MD, North Texas Heart Center, 8440 Walnut Hill Lane, Dallas, Texas 75231. ©1998 by Excerpta Medica, Inc. All rights reserved.
0002-9149/98/$19.00 PII S0002-9149(97)00987-9
545
Stress electrocardiography: Throughout the study, 12-lead electrocardiograms (ECGs) were recorded; these were considered to be diagnostic for ischemia (positive) if there was $0.1 mV of horizontal or downsloping ST-segment depression or $1.5 mV of upsloping ST-segment depression 80 ms after the J point in the lateral precordial leads. A stress ECG was categorized as negative (nonischemic) if the above criteria were not met or nondiagnostic if there was left bundle branch block or a paced rhythm. Stress echocardiography: Baseline echocardiographic images were acquired in the 4 standard views recommended by the American Society of Echocardiography12 in a quad-screen cine loop format (TomTec Imaging Systems, Boulder, Colorado). Those patients undergoing a treadmill protocol were imaged again immediately after exercise (within 60 to 90 seconds) in the same 4 views. Pre- and poststress images were compared to assess ejection fraction and regional wall motion using a 16-segment model. Patients receiving pharmacologic stress testing were imaged at each dose of dobutamine and atropine. Representative cardiac cycle cine loops were digitized at baseline, low dose (5 mg/kg/min), peak dose, and in recovery. Continuous imaging during the pharmacologic protocol allowed premature termination of the test if $2 worsening wall motion abnormalities or worsening LV function developed before achievement of the target heart rate. Although electrocardiographic and echocardiographic imaging were performed simultaneously, they were interpretated independently for this study. LV ejection fraction was visually estimated and reported using intervals of 5% to 10% (with normal considered $50%). A positive stress echocardiogram was defined as the development of new or worsening wall motion abnormalities in any wall segment relative to baseline. All other studies were negative for ischemia, including those with baseline wall motion abnormalities that remained unchanged during stress testing. Clinical follow-up: Follow-up data on each patient were obtained retrospectively from extensive review of the clinical records and telephone interviews, which were blinded to the results of stress testing. Major adverse cardiac events were defined as hospitalization for unstable angina, heart failure, acute myocardial infarction, or cardiac death. Total adverse cardiac events were defined as major cardiac events plus coronary artery revascularization by either percutaneous angioplasty or coronary artery bypass surgery. Statistical analysis: Results are reported as mean 6 1 SD. Univariate analysis was performed using an unpaired t test for continuous variables and the chisquare test for nominal variables. Predictors of future cardiac events by multivariate analysis were determined using a Cox proportional hazards regression model. The probability of a cardiac event-free period after the index stress test was estimated using the Kaplan-Meier method and compared using the logrank test. A p value ,0.01 was considered statistically significant. Incremental benefits of clinical variables, stress 546 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 81
electrocardiography, and stress echocardiography were studied using a Cox proportional hazards regression model in a stepwise fashion. First, all clinical variables found to be predictive of event-free survival by univariate analysis were added to the model and retained if the model maintained a significance level of 0.05. Those variables left in the model were used to represent the clinical data and were not removed during the next 2 steps. Next, stress variables discovered to be significant by univariate analysis were added and retained if significant. Finally, echocardiographic variables were added. The chi-square likelihood ratio for major and total cardiac events was determined for each step by summation of all variables retained in the model. Cost implications: Cost analysis was examined, comparing the referral for invasive studies based on positive stress electrocardiography and stress echocardiography using the reimbursable amounts paid by Medicare during 1993 in Louisiana (stress ECG $50, dobutamine $50, stress echocardiogram $225, coronary angiogram $1,401). Assuming that a positive stress test warranted referral for coronary angiography, the potential cost of invasive testing was added to the cost of stress testing for each group. From these totals, the average cost per patient tested and per cardiac event were determined, as well as the number of cardiac events not predicted.
RESULTS
Patient population: During a mean follow-up period of 23.0 6 7.2 months, 41 of 661 patients experienced an index cardiac event (1 sudden cardiac death, 7 myocardial infarctions, 5 congestive heart failures, 3 unstable angina, and 25 percutaneous angioplasty or coronary artery bypass surgery). For the overall group, the mean age was 58 6 13 years, 48% were men, and the cohort averaged 2.6 6 1.4 risk factors for CAD (Table I). Those who experienced any (total) cardiac event had a significantly greater number of cardiac risk factors and presence of baseline wall motion abnormalities, whereas major events were more frequently seen in those unable to perform exercise stress and were unrelated to the presence of baseline wall motion abnormalities. Stress study characteristics: Cardiovascular stress was provoked with an exercise protocol in 498 patients (80%) and with dobutamine in 163 (20%). Stress studies were more frequently diagnosed as positive in those patients who underwent an exercise protocol compared with dobutamine when evaluated by stress electrocardiographic criteria (24% vs 10%, chi-square 12.7, p 5 0.002) and stress echocardiographic criteria (18% vs 8%, chi-square 9.5, p 5 0.002). Peak heart rate for all patients was 149 6 24 beats/min, which represented 86 6 25% of the agepredicted maximum heart rate. Overall, target heart rate was achieved in 471 patients (71%)—78% of those using the exercise protocol and 50% of those using the dobutamine protocol (p ,0.0001). The mean estimated maximum heart rate– blood pressure product for the exercise group was 10,100 6 7,800. MARCH 1, 1998
TABLE I Comparison of Clinical and Stress Variables in Relation to Major and Total Cardiac Events
Age (yrs) Men Cardiac risk factors Description of chest discomfort Heavy, pressure, tight Sharp, stabbing Dull, aching Other Presence of baseline wall motion abnormalities Exercise protocol Reached target heart rate Follow-up duration (months)
No Cardiac Event (n 5 645)
Major Cardiac Events (n 5 16)
58 6 13 309 (48%) 2.5 6 1.4
63 6 11 8 (50%) 3.7 6 1.4
166 56 45 379 81
8 0 1 7 2
(26%) (9%) (7%) (58%) (13%)
492 (76%) 463 (72%) 23.1 6 7.1
(50%)
p Value*
No Cardiac Event (n 5 620)
Total Cardiac Events (n 5 41)
p Value*
NS NS 0.002
58 6 13 291 (47%) 2.5 6 1.4
63 6 10 26 (63%) 4.0 6 1.3
NS NS ,0.0001
157 55 39 369 70
17 1 6 17 13
NS NS NS NS NS
(6%) (44%) (13%)
6 (38%) 8 (50%) 19.1 6 9.6
0.001 NS NS
(26%) (9%) (6%) (59%) (11%)
472 (76%) 440 (72%) 23.1 6 7.2
(41%) (3%) (15%) (41%) (32%)
26 (63%) 28 (70%) 23.0 6 7.4
NS NS 0.05 NS ,0.005 NS NS NS
Data are expressed as mean 6 1 SD or number of patients (%). *Comparison with no cardiac events.
TABLE II Univariate Predictors of Major and Total Cardiac Events Major Cardiac Events Predictor Presence of stress-induced wall motion abnormality Stress-induced reduction in LV function† Diabetes mellitus History of hypertension Use of calcium channel blocker .1 mm ST-segment depression .1 segmental wall motion abnormality at baseline Active or recent tobacco use Use of digoxin Description of chest discomfort‡ Use of diuretic Advanced age (.60 yrs) Gender Use of b blockers
Chi-square*
RR
Total Cardiac Events p Value
Chi-square*
RR
p Value
14.8
5.4
,0.0001
83.1
10.5
,0.0001
1.3 7.1 12.1 12.4 1.7 0.0
0 3.9 14.8 5.7 1.8 1.0
NS 0.008 0.0005 0.0004 NS NS
55.3 26.2 15.7 14.6 14.0 13.3
10.8 4.5 4.1 3.1 2.8 3.4
,0.0001 ,0.0001 ,0.0001 ,0.0001 0.001 0.0003
7.2 5.0 6.9 2.3 0.7 0.1 0.5
3.8 5.2 3.6 2.5 1.5 1.1 1.6
0.007 NS NS NS NS NS NS
12.3 11.1 6.2 7.5 4.7 4.3 0.1
2.8 4.7 2.2 2.5 2.0 1.9 1.1
0.0005 0.001 0.01 0.01 NS NS NS
*Degrees of freedom 5 1. † Fall in visually assessed ejection fraction $10% from baseline. ‡ Classified as heavy, pressure, tightness, or other.
Clinical variables associated with an inability to achieve target heart rate for all patients included hypertension (present in 69% of patients who did not reach target heart rate vs 41% who did, chi-square 40.0, p ,0.0001), use of calcium channel blockers (39% vs 18%, chi-square 28.6, p ,0.0001), b blockers (26% vs 10%, chi-square 24.0, p ,0.0001), and diuretics (22% vs 11%, chi-square 10.9, p ,0.001), as well as the presence of diabetes mellitus (24% vs 12%, chi-square 14.0, p 5 0.0002) or peripheral vascular disease (12% vs 5%, chi-square 7.9, p 5 0.005). Stress testing was classified as positive in 129 patients (20%) by electrocardiographic criteria and in 103 patients (16%) by echocardiographic visualization of new or worsening wall motion abnormalities in $1 wall segments. Concordance between stress electrocardiography and stress echocardiography was found
in 84% of the patients. Only 12 with a positive stress echocardiogram exhibited concomitant stress-induced LV dysfunction. Additionally, 15 (2%) had an ECG that was nondiagnostic for ischemia due to left bundle branch block or ventricular paced rhythm. Clinical outcome of total cardiac events: Univariate analysis identified 11 variables that were predictive of any cardiac event (Table II). The strongest independent variables were derived from echocardiographic imaging (stress-induced regional wall motion abnormalities or reduced LV function, both of which carried a relative risk .10 for developing any cardiac event). Multivariate analysis of all 11 variables identified only 4 that were predictive of total cardiac events (Table III). When separated by gender, only stress-induced wall motion abnormalities and diabetes mellitus were multivariate predictors in both men and women,
CORONARY ARTERY DISEASE/PROGNOSIS WITH STRESS ECHO
547
TABLE III Multivariate Predictors of Major and Total Cardiac Events Major Cardiac Events Predictors Presence of stress-induced wall motion abnormality Use of calcium channel blocker Diabetes mellitus Stress-induced reduction in LV function‡ Description of chest discomfort† $1 mm ST-segment depression
Total Cardiac Events
Chi-square*
p Value
Chi-square*
11.3
0.0008
44.9
,0.0001
10.8 — —
0.001 NS NS
13.7 18.7 12.7
0.0002 ,0.0001 0.0009
— —
NS NS
— —
p Value
NS NS
*Degrees of freedom 5 1, except for stress-induced LV function 5 2. † Classified as heavy, pressure, tightness, or other. ‡ Fall in visually assessed ejection fraction $10% from baseline.
electrocardiography and stress echocardiography contributed to clinical variables in the form of a summation of the chi-square values produced by each variable. Separation of the data based on the patient’s ability to achieve target heart rate did not affect the prognostic accuracy of either stress electrocardiography or stress echocardiography. Likewise, choice of the stress protocol utilized did not alter the results. Total cardiac event-free survival for patients grouped by the development of stress-induced new or worsening wall motion abnormalities and/or worsening LV function was excellent (96%) in the group with no stress-induced wall motion abnormalities and an improved LV ejection fraction (Figure 3); however, if the ejection fraction decreased and new or worsening wall motion abnormalities occurred during stress, the event-free rate was a dismal 42%. An intermediate event-free rate of 73% was found with new or worsening wall motion abnormalities without worsening of LV function. Clinical outcome of major cardiac events: Limiting the analysis to in-
clude only major cardiac events identified 6 variables with univariate FIGURE 1. Kaplan-Meier estimates of the total cardiac event-free survival based on the analysis (Table II). Stress-induced combination of stress electrocardiographic and stress echocardiographic results. regional wall motion abnormalities proved to be an important independent variable, but a reduction in LV whereas stress-induced LV dysfunction (chi-square function was not, because these patients were referred 17.0; p ,0.0001) and calcium channel blocker utili- for cardiac catheterization and intervention, if indization (chi-square 13.8; p 5 0.0002) were predictive cated; thus preventing potential cardiac events. The remaining 5 variables were derived from the clinical in men only. Figure 1 illustrates the Kaplan-Meier cumulative history. Multivariate analysis identified 2 variables event-free survival plots for total cardiac events that were predictive of major cardiac events: the degrouped according to the stress electrocardiographic velopment of stress-induced wall motion abnormaliand stress echocardiographic criteria. Those patients ties and the use of calcium channel blockers (Table with a positive stress echocardiogram, despite a neg- III). Figure 4 illustrates the Kaplan-Meier cumulative ative stress ECG, had a poor prognosis (66% eventfree survival, chi-square 99.0, p ,0.0001). Patients event-free survival plots for major cardiac events with both a positive stress ECG and stress echocar- grouped according to stress electrocardiographic and diogram carried a similar event-free survival (70%). stress echocardiographic criteria. Those patients with Alternatively, regardless of whether a positive or neg- a positive stress echocardiograms and negative stress ative stress ECG was obtained, a negative stress echo- ECGs had the poorest prognosis (82% event-free surcardiogram predicted an excellent prognosis (97% and vival, chi-square 18.9, p 5 0.0003). Those with both 96%, respectively). Comparison of prognostic values a positive stress ECG and stress echocardiogram carfor stress electrocardiography and stress echocardiog- ried a slightly better prognosis of 94% (which was raphy in isolation yielded varying degrees of useful- unexpected), whereas a negative stress echocardioness. Although stress echocardiography proved to be a gram with either a positive or negative stress ECG powerful predictor of future adverse events, stress predicted an excellent prognosis (97% and 98%, reelectrocardiography was less impressive (Table IV). spectively). Again, comparison of prognostic values In fact, stress electrocardiographic results were not for stress echocardiography in isolation proved to be a significant enough to be left in the multivariate model. significant, but less impressive, predictor of future Figure 2 illustrates the incremental value that stress adverse events, whereas stress electrocardiographic 548 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 81
MARCH 1, 1998
TABLE IV Prognostic Value for Major and Total Cardiac Events When Stress Electrocardiography (ECG) and Stress Echocardiography (echo) Are Analyzed Individually for 2.5-Year Event-free Survival Major Cardiac Events n Stress ECG Positive Negative
5 11
Stress echo Positive Negative
8 8
Event-Free Survival
Total Cardiac Events n
Event-Free Survival
96% 97% Chi-square 1.6, NS
17 24
86% 93% Chi-square 14.8, p 5 0.0001
92% 99% Chi-square 15.4, p ,0.0001
27 14
74% 97% Chi-square 98.1, p ,0.0001
FIGURE 2. Comparison of cumulative chi-square for stepwise analysis of clinical, stress electrocardiographic, and stress echocardiography variables for the prediction of total and major cardiac events.
results failed to provide any significant discriminatory ability (Table IV). This is further highlighted in Figure 2, which demonstrates the lack of incremental value from stress electrocardiography using a stepwise multivariate model. However, a modest benefit was observed with the addition of stress echocardiography to the model. Cost implications: Clinical management decisions based solely on the results of stress electrocardiography would have resulted in 129 patient referrals for coronary angiography. Additionally, 15 patients would have required concurrent cardiac imaging due to a nondiagnostic ECG, resulting from left bundle branch block or a paced rhythm. The mean cost of stress electrocardiography testing and coronary angiography for the cohort would have averaged $340 per patient. Because a stress ECG correctly predicted 17 total cardiac events and 5 major cardiac events, the cost to correctly predict each patient with a cardiac event would have been $13,209 for total events and $44,910 for major events. In addition, a negative stress ECG would miss 24 total events (59%) and 11 major events (69%). Using stress echocardiography, 103 patients would have been referred for coronary angiography for a positive study, resulting in a cost of $456 per patient.
Stress echocardiography correctly predicted 27 total cardiac events and 8 major cardiac events costing $11,155 and $37,647 per cardiac event predicted, respectively. Furthermore, negative stress echocardiography would miss 10 fewer total events (24%) and 3 fewer major events (19%) compared with stress electrocardiography.
DISCUSSION
Multiple studies have demonstrated the high sensitivity and specificity of stress echocardiography to detect CAD3–5; however, these studies have examined heterogeneous groups of patients, often with significant numbers of patients with known CAD. Others have shown the prognostic importance of stress echocardiography to predict future cardiac events in high-risk patients.6 –11 In clinical practice, it is common to extrapolate these findings to patients with a low-risk profile for CAD; however, data to support this action are lacking. The focus of this investigation was to determine the predictive value of stress echocardiography in patients with ,50% pretest likelihood for CAD. It is this group of patients who are commonly examined by primary care physicians and cardiologists for ‘‘atypical’’ chest pain complaints or as asymptomatic patients undergoing CAD risk stratification. Our results demonstrate that despite the overall low incidence of cardiac events in this cohort, stress echocardiography correctly identified a subset of patients at increased risk for experiencing future cardiac events. Furthermore, the benefit of stress echocardiography was incrementally beneficial over clinical and stress electrocardiographic results. The lack of stressinduced wall motion abnormalities or a history of diabetes mellitus were powerful negative predictors of any future (total) adverse event in both men and women. In agreement with previous studies that examine patients with higher rates of predetermined CAD,9,13–15 ischemic ST-segment changes (positive stress ECG) were prognostically important for total cardiac events with univariate analysis. However, in our low-risk group of patients, stress electrocardiography was not a significant independent predictor of major cardiac events. Furthermore, stress electrocardiography was not predictive of future total or major cardiac events using multivariate analysis. As previously shown in asymptomatic patients, the prognostic value of ischemic stress electrocardiographic changes predicts a cardiac event (angina, myocardial infarction, or sudden death) rate of 13% to 20% during long-term follow-up.16,17 Improvement is seen in symptomatic patients with no history of myocardial infarction18; however, the value of stress electrocardiography appears less important when combined with radionuclide imaging. Brown et al19 discovered that in patients without known CAD, stress electrocardiogra-
CORONARY ARTERY DISEASE/PROGNOSIS WITH STRESS ECHO
549
anatomy, baseline ejection fraction, and age were accounted for. Exercise thallium scintigraphy has also demonstrated powerful independent prognostic value for future cardiac events by multivariate analysis.19,24,25 The magnitude of incremental improvement varied with benefits confined to intermediate- to high-risk subgroups.26 The total cardiac event-free rate for a negative stress echocardiogram (no stress-induced wall motion abnormalities) in our study was 99% and 96% at 12 and 30 months, respectively; these findings are similar to the 96% event-free rate found by FIGURE 3. Kaplan-Meier estimates of the total cardiac event-free survival based on Sawada et al,27 who used a high-risk the results of stress echocardiographic wall motion analysis and change in overall LV cohort of patients with a mean folejection fraction (EF). WMA 5 wall motion abnormality. low-up of 28 months. We observed that an exercise-induced decrease in LV function was associated with poor prognosis. As previously shown by Limacher et al,28 this likely relates to a larger ischemic burden accounting for measurable LV dysfunction. In our patients with abnormal studies, stressinduced LV dysfunction was accompanied by new or worsening wall motion abnormalities in an average of 7.1 segments compared with an average of 3.2 segments in those without worsening LV function (p 5 0.0005). Interestingly, the inability to achieve target heart rate did not adFIGURE 4. Kaplan-Meier estimates of the major cardiac event-free survival based on versely affect the prognostic value of the combination of stress electrocardiographic and stress echocardiographic results. stress electrocardiography or stress echocardiography to predict cardiac events. Patients failing to achieve phy had no prognostic value by multivariate analysis. target heart rate compared with those reaching target This finding is supported by Schalet et al20 who found heart rate were 1.5 to 2 times more likely to have a that despite positive stress electrocardiographic testing history of hypertension, diabetes mellitus, or using b ($2 mm ST depression), a normal single-photon blockers, calcium channel blockers, or diuretics. emission computed tomography thallium-201 study Revascularization procedures were the most compredicted the need for revascularization in only 5% of mon indexing event in our study, accounting for 62% patients, whereas in high-risk patients, Krivokapich et of the cardiac events, thereby increasing the prognosal21 found the prognostic value of both stress electro- tic power of our data; however, the inclusion of recardiography and stress echocardiography to be pre- vascularization as a clinical end point clearly limits dictive of future cardiac events by multivariate anal- our findings. To conduct a study in this era and to withhold revascularization would be unethical, and ysis. The addition of radionuclide imaging to stress test- exclusion of these patients would create significant ing unsurprisingly transcends the prognostic value of bias. Cardiac mortality, occurring in only 1 patient, clinical and electrocardiographic variables. Pryor et was expectantly low given our low-risk population. al22 were the first to demonstrate that the exercise Although the classification of mortality of cardiac and ejection fraction assessed by radionuclide angiogra- noncardiac events is sometimes subjective, all-cause phy was the most important predictor of cardiovascu- mortality was ,1%/year in this cohort and evenly lar death or myocardial infarction. However, in a large distributed between those with a positive and negative heterogeneous population of patients, Taliercio et al23 stress echocardiogram. A second limitation of this study involves the comfailed to demonstrate the independent prognostic value of radionuclide angiography once coronary bining of patients who were stressed with an exercise 550 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 81
MARCH 1, 1998
protocol with those undergoing pharmacologic stress. Recent work has shown dobutamine stress electrocardiography to have very poor diagnostic accuracy29; however, this was not the case in our study. The exclusion of those patients pharmacologically stressed with intravenous dobutamine resulted in a lowering of the chi-squared values for stress electrocardiography for both total (14.0 to 5.8) and major events (1.7 to 0.4) using univariate analysis. Echocardiography has shown great progress over the past decade with improvements in imaging quality and digital acquisition techniques. Furthermore, the development of contrast echocardiography will ultimately make echocardiography a practical option in most patients. Our findings, as outlined above, are comparable to radionuclide imaging. However, stress echocardiography has the advantage of portability, direct visualization of cardiac structures, lower cost, and takes less of the patient’s time to perform the study. In the upcoming era of managed care, the importance of cost effectiveness will likely elevate the role of stress echocardiography in the evaluation of low-risk patients. Acknowledgment: We gratefully acknowledge the technical assistance of Carlos Moreno, Lauren Camardelle for her assistance in the preparation of the manuscript, and the staff of the Ochsner Medical Institutions’ echocardiography laboratory for the efforts, which were necessary to complete this study.
1. Pryor DB, Shaw L, McCants CB, Lee KL, Mark DB, Harrell FE, Muhlbaier LH, Califf RM. Value of the history and physical in identifying patients at increased risk of coronary artery disease. Ann Int Med 1993;118:81–90. 2. Birdwell B, Herbers J, Kroenke K. Evaluating chest pain. The patient’s presentation style alters the physicians diagnostic approach. Arch Int Med 1993; 153:1991–1995. 3. Crouse L, Harbrecht J, Vacek J, Rosamond T, Kramer P. Exercise echocardiography as a screening test for coronary artery disease and correlation with coronary arteriography. Am J Cardiol 1991;67:1213–1218. 4. Sawada SG, Segar DS, Ryan T, Brown SE, Dohan AM, Williams R, Fineberg NS, Armstrong WF, Feigenbaum H. Echocardiographic detection of coronary artery disease during dobutamine infusion. Circulation 1991;83:1605–1614. 5. Marwick T, Nemec J, Pashkow F, Stewart W, Salcedo E. Accuracy and limitations of exercise echocardiography in a routine clinical setting. J Am Coll Cardiol 1992;19:74 – 81. 6. Afridi I, Quinones M, Zoghbi W, Cheirif J. Dobutamine stress echocardiography: sensitivity, specificity, and predictive value for future cardiac events. Am Heart J 1994;127:1510 –1515. 7. Eriksson S, Erhardt L, Lindvall K, Melcher A, Rehnqvist N. Long-term prognostic importance of exercise echocardiography after an episode of unstable angina. Cardiology 1995;86:426 – 431. 8. Kamaran M, Teague S, Finkelhor R, Dawson N, Bahler R. Prognostic value of dobutamine stress echocardiography in patients referred because of suspected coronary artery disease. Am J Cardiol 1995;76:887– 891. 9. Mazeika P, Nadazdin A, Oakley C. Prognostic value of dobutamine echocardiography in patients with high pretest likelihood of coronary artery disease. Am J Cardiol 1993;71:33–39.
10. Quintana M, Lindvall K, Ryden L, Brolund F. Prognostic value of predis-
charge exercise stress echocardiography after acute myocardial infarction. Am J Cardiol 1995;76:1115–1121. 11. Williams M, Odabashian J, Lauer M, Thomas J, Marwick T. Prognostic value of dobutamine echocardiography in patients with left ventricular dysfunction. J Am Coll Cardiol 1996;27:132–139. 12. American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. J Am Soc Echocardiogr 1989;2:358 –367. 13. Weiner DA, Ryan TJ, McCabe CH, Chaitman BR, Sheffield LT, Ferguson JC, Fisher LD, Tristani F. Prognostic importance of a clinical profile and exercise test in medically treated patients with coronary artery disease. J Am Coll Cardiol 1984;3:772–779. 14. Mark DB, Shaw L, Harrell FE, Hlatky MA, Lee KL, Bengtson JR, McCants CB, Califf RM, Pryor DB. Prognostic value of a treadmill exercise score in outpatients with suspected coronary artery disease. N Engl J Med 1991;325:849 – 853. 15. Moriel M, Rozanski A, Klein J, Berman D, Bairey N. The limited efficacy of exercise radionuclide ventriculography in assessing prognosis of women with coronary artery disease. Am J Cardiol 1995;76:1030 –1035. 16. Froelicher V, Thomas M, Pillow C, Lancaster M. Epidemiologic study of asymptomatic men screened by maximal treadmill testing for latent coronary artery disease. Am J Cardiol 1974;34:770 –776. 17. Giagnoni E, Secchi MB, Wu SC, Morabito A, Oltrona L, Mancarella S, Volpin N, Fossa L, Bettazzi L, Arangio G, Sachero A, Folli G. Prognostic value of exercise EKG testing in asymptomatic normotensive subjects. A prospective matched study. N Engl J Med 1983;309:1085–1089. 18. Ellestad M, Wan M. Predictive implications of stress testing. Follow-up of 2700 subjects after maximum treadmill stress testing. Circulation 1975;51:363– 369. 19. Brown KA, Boucher CA, Okada RD, Guiney TE, Newell JB, Strauss HW, Pohost GM. Prognostic value of exercise thallium-201 imaging in patients presenting for evaluation of chest pain. J Am Coll Cardiol 1983;1:994 –1001. 20. Schalet B, Kegel J, Heo J, Segal B, Iskandrian A. Prognostic implications of normal exercise SPECT thallium images in patients with strongly positive exercise electrocardiograms. Am J Cardiol 1993;72:1201–1203. 21. Krivokapich J, Child J, Gerber R, Lem V, Moser D. Prognostic usefulness of positive or negative exercise stress echocardiography for predicting coronary events in ensuing twelve months. Am J Cardiol 1993;71:646 – 651. 22. Pryor DB, Harrell FE, Lee KL, Rosati RA, Coleman RE, Cobb FR, Califf RM, Jones RH. Prognostic indicators from radionuclide angiography in medically-treated patients with coronary artery disease. Am J Cardiol 1984;53:18 –22. 23. Taliercio CP, Clements IP, Zinsmeister AR, Gibbons RJ. Prognostic value and limitations of exercise radionuclide angiography. Mayo Clin Proc 1988;63: 573–582. 24. Kaul S, Lilly DR, Gascho JA, Watson DD, Gibson RS, Oliner CA, Ryan JM, Beller GA. Prognostic utility of the exercise thallium-201 test in ambulatory patients with chest pain: comparison with cardiac catheterization. Circulation 1988;77:745–758. 25. Hachamovitch R, Berman DS, Kiat H, Cohen I, Cabico JA, Friedman J, Diamond GA. Exercise myocardial perfusion SPECT in patients without known coronary artery disease. Incremental prognostic value and use in risk stratification. Circulation 1996;93:905–914. 26. Ladenheim M, Kotler T, Pollock B, Berman D, Diamond G. Incremental prognostic power of clinical history, exercise electrocardiography and myocardial perfusion scintigraphy in suspected coronary artery disease. Am J Cardiol 1987; 59:270 –277. 27. Sawada S, Ryan T, Conley M, Corya B, Feigenbaum H, Armstrong W. Prognostic value of a normal exercise echocardiogram. Am Heart J 1990;120: 49 –55. 28. Limacher M, Quinones M, Poliner L, Nelson J, Winters W, Waggoner A. Detection of coronary artery disease with exercise two-dimensional echocardiography. Description of a clinically applicable method and comparison with radionuclide ventriculography. Circulation 1983;67:1211–1218. 29. Mairesse GH, Marwick TH, Vanoverschelde JJ, Baudhuin T, Wijns W, Melin JA, Detry JR. How accurate is dobutamine stress electrocardiography for detection of coronary artery disease? Comparison with two-dimensional echocardiography and Technetium-99m methoxyl isobutyl isonitrile (Mibi) perfusion scintigraphy. J Am Coll Cardiol 1994;24:920 –927.
CORONARY ARTERY DISEASE/PROGNOSIS WITH STRESS ECHO
551
MD
Patients with atypical chest pain frequently lack significant coronary artery disease (CAD) and are, therefore, at low risk for future adverse cardiovascular events. We hypothesized that in this group of patients, stress echocardiography could identify those at risk for cardiac events. We retrospectively reviewed (mean follow-up 23.0 6 7.2 months) the prognostic value of stress echocardiography for major (cardiac death, myocardial infarction, congestive heart failure, and unstable angina) and total (major events plus coronary revascularization) cardiac events in 661 patients with atypical chest pain, normal global left ventricular (LV) systolic function, and no history of CAD. A positive stress echocardiogram was defined as the development of new or worsening wall motion abnormalities with exercise stress (80%) or dobutamine (20%). A total of 41 cardiac and 16 major events were noted. The event-free survival for total car-
diac events was 97% for a normal stress echocardiogram and 93% for a normal stress electrocardiogram (ECG) at 30 months. A positive stress ECG predicted an event-free rate of 86% compared with 74% for stressinduced wall motion abnormalities and 42% if stressinduced LV dysfunction accompanied the wall motion abnormalities. A strategy recommending invasive studies based on positive stress echocardiogram results increased the per-patient cost, but led to greater savings per cardiac event predicted and provided incremental prognostic value for future cardiac events beyond clinical and stress electrocardiographic data. Thus, stress echocardiography in low-risk patients for CAD appears to be more cost effective than a stress ECG. Q1998 by Excerpta Medica, Inc. (Am J Cardiol 1998;81:545–551)
hoosing the optimal cost-effective strategy for evaluating patients with atypical chest pain reC mains controversial. These patients are at low risk for
either exercise (81%) or dobutamine (19%) stress echocardiography. After retrospective review of the medical records, 1,310 patients were excluded due to a history of CAD, typical angina, depressed left ventricular (LV) function (ejection fraction at rest ,50%), valvular heart disease, or cardiac transplantation. Of the remaining 688 patients, 27 (4%) were excluded due to inadequate clinical information or follow-up. Thus, the final cohort consisted of 661 patients at low risk for cardiac events. Stress protocol: Based on the patient’s physical abilities, cardiac function was assessed using either an exercise treadmill protocol or pharmacologic stress with intravenous dobutamine. Exercise treadmill testing was symptom limited using the Bruce, modified Bruce, or Ramp-based protocols. Pharmacologic stress was provoked by intravenous dobutamine, starting at 5 mg/kg/min and progressing to a maximum dose of 50 mg/kg/min at 3-minute intervals until the target heart rate (85% of maximum predicted heart rate for age) was achieved. If needed, $1 mg of atropine was given intravenously in divided intervals to attain the target heart rate. Clinical symptoms and blood pressure were monitored in the standard fashion throughout the study. Stress testing was terminated prematurely if the patient developed significant chest pain, serious ventricular arrhythmia, severe hypertension (systolic .220 mm Hg; diastolic .120 mm Hg), a progressive decrease in systolic blood pressure (,90 mm Hg), or $2 mm of ST-segment depression.
developing future cardiovascular events.1 The decision to recommend stress testing depends upon the physician’s initial clinical assessment, which is strongly influenced by the patient’s presentation2 as well as by the coronary risk factor profile. Stress echocardiography, a noninvasive diagnostic tool used to detect coronary artery disease (CAD), has grown in popularity, in part due to its superior accuracy over stress electrocardiography3–5 and its lower cost over radionuclide imaging. Although the prognostic value of stress echocardiography in high-risk patients for CAD has previously been established,6 –11 the potential value of stress echocardiography to assess prognosis in patients at low risk for CAD has not been thoroughly studied. The purpose of this study was to determine the prognostic value of stress echocardiography in a large cohort of low-risk patients referred to our laboratory for functional cardiac assessment.
METHODS
Study population: During 1993, 1,998 patients were referred to our laboratory for evaluation of CAD using From the Section of Cardiology, Department of Internal Medicine, Ochsner Medical Institutions, New Orleans, Louisiana. Manuscript received August 6, 1997; revised manuscript and accepted November 25, 1997. Address for reprints: Jorge Cheirif, MD, North Texas Heart Center, 8440 Walnut Hill Lane, Dallas, Texas 75231. ©1998 by Excerpta Medica, Inc. All rights reserved.
0002-9149/98/$19.00 PII S0002-9149(97)00987-9
545
Stress electrocardiography: Throughout the study, 12-lead electrocardiograms (ECGs) were recorded; these were considered to be diagnostic for ischemia (positive) if there was $0.1 mV of horizontal or downsloping ST-segment depression or $1.5 mV of upsloping ST-segment depression 80 ms after the J point in the lateral precordial leads. A stress ECG was categorized as negative (nonischemic) if the above criteria were not met or nondiagnostic if there was left bundle branch block or a paced rhythm. Stress echocardiography: Baseline echocardiographic images were acquired in the 4 standard views recommended by the American Society of Echocardiography12 in a quad-screen cine loop format (TomTec Imaging Systems, Boulder, Colorado). Those patients undergoing a treadmill protocol were imaged again immediately after exercise (within 60 to 90 seconds) in the same 4 views. Pre- and poststress images were compared to assess ejection fraction and regional wall motion using a 16-segment model. Patients receiving pharmacologic stress testing were imaged at each dose of dobutamine and atropine. Representative cardiac cycle cine loops were digitized at baseline, low dose (5 mg/kg/min), peak dose, and in recovery. Continuous imaging during the pharmacologic protocol allowed premature termination of the test if $2 worsening wall motion abnormalities or worsening LV function developed before achievement of the target heart rate. Although electrocardiographic and echocardiographic imaging were performed simultaneously, they were interpretated independently for this study. LV ejection fraction was visually estimated and reported using intervals of 5% to 10% (with normal considered $50%). A positive stress echocardiogram was defined as the development of new or worsening wall motion abnormalities in any wall segment relative to baseline. All other studies were negative for ischemia, including those with baseline wall motion abnormalities that remained unchanged during stress testing. Clinical follow-up: Follow-up data on each patient were obtained retrospectively from extensive review of the clinical records and telephone interviews, which were blinded to the results of stress testing. Major adverse cardiac events were defined as hospitalization for unstable angina, heart failure, acute myocardial infarction, or cardiac death. Total adverse cardiac events were defined as major cardiac events plus coronary artery revascularization by either percutaneous angioplasty or coronary artery bypass surgery. Statistical analysis: Results are reported as mean 6 1 SD. Univariate analysis was performed using an unpaired t test for continuous variables and the chisquare test for nominal variables. Predictors of future cardiac events by multivariate analysis were determined using a Cox proportional hazards regression model. The probability of a cardiac event-free period after the index stress test was estimated using the Kaplan-Meier method and compared using the logrank test. A p value ,0.01 was considered statistically significant. Incremental benefits of clinical variables, stress 546 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 81
electrocardiography, and stress echocardiography were studied using a Cox proportional hazards regression model in a stepwise fashion. First, all clinical variables found to be predictive of event-free survival by univariate analysis were added to the model and retained if the model maintained a significance level of 0.05. Those variables left in the model were used to represent the clinical data and were not removed during the next 2 steps. Next, stress variables discovered to be significant by univariate analysis were added and retained if significant. Finally, echocardiographic variables were added. The chi-square likelihood ratio for major and total cardiac events was determined for each step by summation of all variables retained in the model. Cost implications: Cost analysis was examined, comparing the referral for invasive studies based on positive stress electrocardiography and stress echocardiography using the reimbursable amounts paid by Medicare during 1993 in Louisiana (stress ECG $50, dobutamine $50, stress echocardiogram $225, coronary angiogram $1,401). Assuming that a positive stress test warranted referral for coronary angiography, the potential cost of invasive testing was added to the cost of stress testing for each group. From these totals, the average cost per patient tested and per cardiac event were determined, as well as the number of cardiac events not predicted.
RESULTS
Patient population: During a mean follow-up period of 23.0 6 7.2 months, 41 of 661 patients experienced an index cardiac event (1 sudden cardiac death, 7 myocardial infarctions, 5 congestive heart failures, 3 unstable angina, and 25 percutaneous angioplasty or coronary artery bypass surgery). For the overall group, the mean age was 58 6 13 years, 48% were men, and the cohort averaged 2.6 6 1.4 risk factors for CAD (Table I). Those who experienced any (total) cardiac event had a significantly greater number of cardiac risk factors and presence of baseline wall motion abnormalities, whereas major events were more frequently seen in those unable to perform exercise stress and were unrelated to the presence of baseline wall motion abnormalities. Stress study characteristics: Cardiovascular stress was provoked with an exercise protocol in 498 patients (80%) and with dobutamine in 163 (20%). Stress studies were more frequently diagnosed as positive in those patients who underwent an exercise protocol compared with dobutamine when evaluated by stress electrocardiographic criteria (24% vs 10%, chi-square 12.7, p 5 0.002) and stress echocardiographic criteria (18% vs 8%, chi-square 9.5, p 5 0.002). Peak heart rate for all patients was 149 6 24 beats/min, which represented 86 6 25% of the agepredicted maximum heart rate. Overall, target heart rate was achieved in 471 patients (71%)—78% of those using the exercise protocol and 50% of those using the dobutamine protocol (p ,0.0001). The mean estimated maximum heart rate– blood pressure product for the exercise group was 10,100 6 7,800. MARCH 1, 1998
TABLE I Comparison of Clinical and Stress Variables in Relation to Major and Total Cardiac Events
Age (yrs) Men Cardiac risk factors Description of chest discomfort Heavy, pressure, tight Sharp, stabbing Dull, aching Other Presence of baseline wall motion abnormalities Exercise protocol Reached target heart rate Follow-up duration (months)
No Cardiac Event (n 5 645)
Major Cardiac Events (n 5 16)
58 6 13 309 (48%) 2.5 6 1.4
63 6 11 8 (50%) 3.7 6 1.4
166 56 45 379 81
8 0 1 7 2
(26%) (9%) (7%) (58%) (13%)
492 (76%) 463 (72%) 23.1 6 7.1
(50%)
p Value*
No Cardiac Event (n 5 620)
Total Cardiac Events (n 5 41)
p Value*
NS NS 0.002
58 6 13 291 (47%) 2.5 6 1.4
63 6 10 26 (63%) 4.0 6 1.3
NS NS ,0.0001
157 55 39 369 70
17 1 6 17 13
NS NS NS NS NS
(6%) (44%) (13%)
6 (38%) 8 (50%) 19.1 6 9.6
0.001 NS NS
(26%) (9%) (6%) (59%) (11%)
472 (76%) 440 (72%) 23.1 6 7.2
(41%) (3%) (15%) (41%) (32%)
26 (63%) 28 (70%) 23.0 6 7.4
NS NS 0.05 NS ,0.005 NS NS NS
Data are expressed as mean 6 1 SD or number of patients (%). *Comparison with no cardiac events.
TABLE II Univariate Predictors of Major and Total Cardiac Events Major Cardiac Events Predictor Presence of stress-induced wall motion abnormality Stress-induced reduction in LV function† Diabetes mellitus History of hypertension Use of calcium channel blocker .1 mm ST-segment depression .1 segmental wall motion abnormality at baseline Active or recent tobacco use Use of digoxin Description of chest discomfort‡ Use of diuretic Advanced age (.60 yrs) Gender Use of b blockers
Chi-square*
RR
Total Cardiac Events p Value
Chi-square*
RR
p Value
14.8
5.4
,0.0001
83.1
10.5
,0.0001
1.3 7.1 12.1 12.4 1.7 0.0
0 3.9 14.8 5.7 1.8 1.0
NS 0.008 0.0005 0.0004 NS NS
55.3 26.2 15.7 14.6 14.0 13.3
10.8 4.5 4.1 3.1 2.8 3.4
,0.0001 ,0.0001 ,0.0001 ,0.0001 0.001 0.0003
7.2 5.0 6.9 2.3 0.7 0.1 0.5
3.8 5.2 3.6 2.5 1.5 1.1 1.6
0.007 NS NS NS NS NS NS
12.3 11.1 6.2 7.5 4.7 4.3 0.1
2.8 4.7 2.2 2.5 2.0 1.9 1.1
0.0005 0.001 0.01 0.01 NS NS NS
*Degrees of freedom 5 1. † Fall in visually assessed ejection fraction $10% from baseline. ‡ Classified as heavy, pressure, tightness, or other.
Clinical variables associated with an inability to achieve target heart rate for all patients included hypertension (present in 69% of patients who did not reach target heart rate vs 41% who did, chi-square 40.0, p ,0.0001), use of calcium channel blockers (39% vs 18%, chi-square 28.6, p ,0.0001), b blockers (26% vs 10%, chi-square 24.0, p ,0.0001), and diuretics (22% vs 11%, chi-square 10.9, p ,0.001), as well as the presence of diabetes mellitus (24% vs 12%, chi-square 14.0, p 5 0.0002) or peripheral vascular disease (12% vs 5%, chi-square 7.9, p 5 0.005). Stress testing was classified as positive in 129 patients (20%) by electrocardiographic criteria and in 103 patients (16%) by echocardiographic visualization of new or worsening wall motion abnormalities in $1 wall segments. Concordance between stress electrocardiography and stress echocardiography was found
in 84% of the patients. Only 12 with a positive stress echocardiogram exhibited concomitant stress-induced LV dysfunction. Additionally, 15 (2%) had an ECG that was nondiagnostic for ischemia due to left bundle branch block or ventricular paced rhythm. Clinical outcome of total cardiac events: Univariate analysis identified 11 variables that were predictive of any cardiac event (Table II). The strongest independent variables were derived from echocardiographic imaging (stress-induced regional wall motion abnormalities or reduced LV function, both of which carried a relative risk .10 for developing any cardiac event). Multivariate analysis of all 11 variables identified only 4 that were predictive of total cardiac events (Table III). When separated by gender, only stress-induced wall motion abnormalities and diabetes mellitus were multivariate predictors in both men and women,
CORONARY ARTERY DISEASE/PROGNOSIS WITH STRESS ECHO
547
TABLE III Multivariate Predictors of Major and Total Cardiac Events Major Cardiac Events Predictors Presence of stress-induced wall motion abnormality Use of calcium channel blocker Diabetes mellitus Stress-induced reduction in LV function‡ Description of chest discomfort† $1 mm ST-segment depression
Total Cardiac Events
Chi-square*
p Value
Chi-square*
11.3
0.0008
44.9
,0.0001
10.8 — —
0.001 NS NS
13.7 18.7 12.7
0.0002 ,0.0001 0.0009
— —
NS NS
— —
p Value
NS NS
*Degrees of freedom 5 1, except for stress-induced LV function 5 2. † Classified as heavy, pressure, tightness, or other. ‡ Fall in visually assessed ejection fraction $10% from baseline.
electrocardiography and stress echocardiography contributed to clinical variables in the form of a summation of the chi-square values produced by each variable. Separation of the data based on the patient’s ability to achieve target heart rate did not affect the prognostic accuracy of either stress electrocardiography or stress echocardiography. Likewise, choice of the stress protocol utilized did not alter the results. Total cardiac event-free survival for patients grouped by the development of stress-induced new or worsening wall motion abnormalities and/or worsening LV function was excellent (96%) in the group with no stress-induced wall motion abnormalities and an improved LV ejection fraction (Figure 3); however, if the ejection fraction decreased and new or worsening wall motion abnormalities occurred during stress, the event-free rate was a dismal 42%. An intermediate event-free rate of 73% was found with new or worsening wall motion abnormalities without worsening of LV function. Clinical outcome of major cardiac events: Limiting the analysis to in-
clude only major cardiac events identified 6 variables with univariate FIGURE 1. Kaplan-Meier estimates of the total cardiac event-free survival based on the analysis (Table II). Stress-induced combination of stress electrocardiographic and stress echocardiographic results. regional wall motion abnormalities proved to be an important independent variable, but a reduction in LV whereas stress-induced LV dysfunction (chi-square function was not, because these patients were referred 17.0; p ,0.0001) and calcium channel blocker utili- for cardiac catheterization and intervention, if indization (chi-square 13.8; p 5 0.0002) were predictive cated; thus preventing potential cardiac events. The remaining 5 variables were derived from the clinical in men only. Figure 1 illustrates the Kaplan-Meier cumulative history. Multivariate analysis identified 2 variables event-free survival plots for total cardiac events that were predictive of major cardiac events: the degrouped according to the stress electrocardiographic velopment of stress-induced wall motion abnormaliand stress echocardiographic criteria. Those patients ties and the use of calcium channel blockers (Table with a positive stress echocardiogram, despite a neg- III). Figure 4 illustrates the Kaplan-Meier cumulative ative stress ECG, had a poor prognosis (66% eventfree survival, chi-square 99.0, p ,0.0001). Patients event-free survival plots for major cardiac events with both a positive stress ECG and stress echocar- grouped according to stress electrocardiographic and diogram carried a similar event-free survival (70%). stress echocardiographic criteria. Those patients with Alternatively, regardless of whether a positive or neg- a positive stress echocardiograms and negative stress ative stress ECG was obtained, a negative stress echo- ECGs had the poorest prognosis (82% event-free surcardiogram predicted an excellent prognosis (97% and vival, chi-square 18.9, p 5 0.0003). Those with both 96%, respectively). Comparison of prognostic values a positive stress ECG and stress echocardiogram carfor stress electrocardiography and stress echocardiog- ried a slightly better prognosis of 94% (which was raphy in isolation yielded varying degrees of useful- unexpected), whereas a negative stress echocardioness. Although stress echocardiography proved to be a gram with either a positive or negative stress ECG powerful predictor of future adverse events, stress predicted an excellent prognosis (97% and 98%, reelectrocardiography was less impressive (Table IV). spectively). Again, comparison of prognostic values In fact, stress electrocardiographic results were not for stress echocardiography in isolation proved to be a significant enough to be left in the multivariate model. significant, but less impressive, predictor of future Figure 2 illustrates the incremental value that stress adverse events, whereas stress electrocardiographic 548 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 81
MARCH 1, 1998
TABLE IV Prognostic Value for Major and Total Cardiac Events When Stress Electrocardiography (ECG) and Stress Echocardiography (echo) Are Analyzed Individually for 2.5-Year Event-free Survival Major Cardiac Events n Stress ECG Positive Negative
5 11
Stress echo Positive Negative
8 8
Event-Free Survival
Total Cardiac Events n
Event-Free Survival
96% 97% Chi-square 1.6, NS
17 24
86% 93% Chi-square 14.8, p 5 0.0001
92% 99% Chi-square 15.4, p ,0.0001
27 14
74% 97% Chi-square 98.1, p ,0.0001
FIGURE 2. Comparison of cumulative chi-square for stepwise analysis of clinical, stress electrocardiographic, and stress echocardiography variables for the prediction of total and major cardiac events.
results failed to provide any significant discriminatory ability (Table IV). This is further highlighted in Figure 2, which demonstrates the lack of incremental value from stress electrocardiography using a stepwise multivariate model. However, a modest benefit was observed with the addition of stress echocardiography to the model. Cost implications: Clinical management decisions based solely on the results of stress electrocardiography would have resulted in 129 patient referrals for coronary angiography. Additionally, 15 patients would have required concurrent cardiac imaging due to a nondiagnostic ECG, resulting from left bundle branch block or a paced rhythm. The mean cost of stress electrocardiography testing and coronary angiography for the cohort would have averaged $340 per patient. Because a stress ECG correctly predicted 17 total cardiac events and 5 major cardiac events, the cost to correctly predict each patient with a cardiac event would have been $13,209 for total events and $44,910 for major events. In addition, a negative stress ECG would miss 24 total events (59%) and 11 major events (69%). Using stress echocardiography, 103 patients would have been referred for coronary angiography for a positive study, resulting in a cost of $456 per patient.
Stress echocardiography correctly predicted 27 total cardiac events and 8 major cardiac events costing $11,155 and $37,647 per cardiac event predicted, respectively. Furthermore, negative stress echocardiography would miss 10 fewer total events (24%) and 3 fewer major events (19%) compared with stress electrocardiography.
DISCUSSION
Multiple studies have demonstrated the high sensitivity and specificity of stress echocardiography to detect CAD3–5; however, these studies have examined heterogeneous groups of patients, often with significant numbers of patients with known CAD. Others have shown the prognostic importance of stress echocardiography to predict future cardiac events in high-risk patients.6 –11 In clinical practice, it is common to extrapolate these findings to patients with a low-risk profile for CAD; however, data to support this action are lacking. The focus of this investigation was to determine the predictive value of stress echocardiography in patients with ,50% pretest likelihood for CAD. It is this group of patients who are commonly examined by primary care physicians and cardiologists for ‘‘atypical’’ chest pain complaints or as asymptomatic patients undergoing CAD risk stratification. Our results demonstrate that despite the overall low incidence of cardiac events in this cohort, stress echocardiography correctly identified a subset of patients at increased risk for experiencing future cardiac events. Furthermore, the benefit of stress echocardiography was incrementally beneficial over clinical and stress electrocardiographic results. The lack of stressinduced wall motion abnormalities or a history of diabetes mellitus were powerful negative predictors of any future (total) adverse event in both men and women. In agreement with previous studies that examine patients with higher rates of predetermined CAD,9,13–15 ischemic ST-segment changes (positive stress ECG) were prognostically important for total cardiac events with univariate analysis. However, in our low-risk group of patients, stress electrocardiography was not a significant independent predictor of major cardiac events. Furthermore, stress electrocardiography was not predictive of future total or major cardiac events using multivariate analysis. As previously shown in asymptomatic patients, the prognostic value of ischemic stress electrocardiographic changes predicts a cardiac event (angina, myocardial infarction, or sudden death) rate of 13% to 20% during long-term follow-up.16,17 Improvement is seen in symptomatic patients with no history of myocardial infarction18; however, the value of stress electrocardiography appears less important when combined with radionuclide imaging. Brown et al19 discovered that in patients without known CAD, stress electrocardiogra-
CORONARY ARTERY DISEASE/PROGNOSIS WITH STRESS ECHO
549
anatomy, baseline ejection fraction, and age were accounted for. Exercise thallium scintigraphy has also demonstrated powerful independent prognostic value for future cardiac events by multivariate analysis.19,24,25 The magnitude of incremental improvement varied with benefits confined to intermediate- to high-risk subgroups.26 The total cardiac event-free rate for a negative stress echocardiogram (no stress-induced wall motion abnormalities) in our study was 99% and 96% at 12 and 30 months, respectively; these findings are similar to the 96% event-free rate found by FIGURE 3. Kaplan-Meier estimates of the total cardiac event-free survival based on Sawada et al,27 who used a high-risk the results of stress echocardiographic wall motion analysis and change in overall LV cohort of patients with a mean folejection fraction (EF). WMA 5 wall motion abnormality. low-up of 28 months. We observed that an exercise-induced decrease in LV function was associated with poor prognosis. As previously shown by Limacher et al,28 this likely relates to a larger ischemic burden accounting for measurable LV dysfunction. In our patients with abnormal studies, stressinduced LV dysfunction was accompanied by new or worsening wall motion abnormalities in an average of 7.1 segments compared with an average of 3.2 segments in those without worsening LV function (p 5 0.0005). Interestingly, the inability to achieve target heart rate did not adFIGURE 4. Kaplan-Meier estimates of the major cardiac event-free survival based on versely affect the prognostic value of the combination of stress electrocardiographic and stress echocardiographic results. stress electrocardiography or stress echocardiography to predict cardiac events. Patients failing to achieve phy had no prognostic value by multivariate analysis. target heart rate compared with those reaching target This finding is supported by Schalet et al20 who found heart rate were 1.5 to 2 times more likely to have a that despite positive stress electrocardiographic testing history of hypertension, diabetes mellitus, or using b ($2 mm ST depression), a normal single-photon blockers, calcium channel blockers, or diuretics. emission computed tomography thallium-201 study Revascularization procedures were the most compredicted the need for revascularization in only 5% of mon indexing event in our study, accounting for 62% patients, whereas in high-risk patients, Krivokapich et of the cardiac events, thereby increasing the prognosal21 found the prognostic value of both stress electro- tic power of our data; however, the inclusion of recardiography and stress echocardiography to be pre- vascularization as a clinical end point clearly limits dictive of future cardiac events by multivariate anal- our findings. To conduct a study in this era and to withhold revascularization would be unethical, and ysis. The addition of radionuclide imaging to stress test- exclusion of these patients would create significant ing unsurprisingly transcends the prognostic value of bias. Cardiac mortality, occurring in only 1 patient, clinical and electrocardiographic variables. Pryor et was expectantly low given our low-risk population. al22 were the first to demonstrate that the exercise Although the classification of mortality of cardiac and ejection fraction assessed by radionuclide angiogra- noncardiac events is sometimes subjective, all-cause phy was the most important predictor of cardiovascu- mortality was ,1%/year in this cohort and evenly lar death or myocardial infarction. However, in a large distributed between those with a positive and negative heterogeneous population of patients, Taliercio et al23 stress echocardiogram. A second limitation of this study involves the comfailed to demonstrate the independent prognostic value of radionuclide angiography once coronary bining of patients who were stressed with an exercise 550 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 81
MARCH 1, 1998
protocol with those undergoing pharmacologic stress. Recent work has shown dobutamine stress electrocardiography to have very poor diagnostic accuracy29; however, this was not the case in our study. The exclusion of those patients pharmacologically stressed with intravenous dobutamine resulted in a lowering of the chi-squared values for stress electrocardiography for both total (14.0 to 5.8) and major events (1.7 to 0.4) using univariate analysis. Echocardiography has shown great progress over the past decade with improvements in imaging quality and digital acquisition techniques. Furthermore, the development of contrast echocardiography will ultimately make echocardiography a practical option in most patients. Our findings, as outlined above, are comparable to radionuclide imaging. However, stress echocardiography has the advantage of portability, direct visualization of cardiac structures, lower cost, and takes less of the patient’s time to perform the study. In the upcoming era of managed care, the importance of cost effectiveness will likely elevate the role of stress echocardiography in the evaluation of low-risk patients. Acknowledgment: We gratefully acknowledge the technical assistance of Carlos Moreno, Lauren Camardelle for her assistance in the preparation of the manuscript, and the staff of the Ochsner Medical Institutions’ echocardiography laboratory for the efforts, which were necessary to complete this study.
1. Pryor DB, Shaw L, McCants CB, Lee KL, Mark DB, Harrell FE, Muhlbaier LH, Califf RM. Value of the history and physical in identifying patients at increased risk of coronary artery disease. Ann Int Med 1993;118:81–90. 2. Birdwell B, Herbers J, Kroenke K. Evaluating chest pain. The patient’s presentation style alters the physicians diagnostic approach. Arch Int Med 1993; 153:1991–1995. 3. Crouse L, Harbrecht J, Vacek J, Rosamond T, Kramer P. Exercise echocardiography as a screening test for coronary artery disease and correlation with coronary arteriography. Am J Cardiol 1991;67:1213–1218. 4. Sawada SG, Segar DS, Ryan T, Brown SE, Dohan AM, Williams R, Fineberg NS, Armstrong WF, Feigenbaum H. Echocardiographic detection of coronary artery disease during dobutamine infusion. Circulation 1991;83:1605–1614. 5. Marwick T, Nemec J, Pashkow F, Stewart W, Salcedo E. Accuracy and limitations of exercise echocardiography in a routine clinical setting. J Am Coll Cardiol 1992;19:74 – 81. 6. Afridi I, Quinones M, Zoghbi W, Cheirif J. Dobutamine stress echocardiography: sensitivity, specificity, and predictive value for future cardiac events. Am Heart J 1994;127:1510 –1515. 7. Eriksson S, Erhardt L, Lindvall K, Melcher A, Rehnqvist N. Long-term prognostic importance of exercise echocardiography after an episode of unstable angina. Cardiology 1995;86:426 – 431. 8. Kamaran M, Teague S, Finkelhor R, Dawson N, Bahler R. Prognostic value of dobutamine stress echocardiography in patients referred because of suspected coronary artery disease. Am J Cardiol 1995;76:887– 891. 9. Mazeika P, Nadazdin A, Oakley C. Prognostic value of dobutamine echocardiography in patients with high pretest likelihood of coronary artery disease. Am J Cardiol 1993;71:33–39.
10. Quintana M, Lindvall K, Ryden L, Brolund F. Prognostic value of predis-
charge exercise stress echocardiography after acute myocardial infarction. Am J Cardiol 1995;76:1115–1121. 11. Williams M, Odabashian J, Lauer M, Thomas J, Marwick T. Prognostic value of dobutamine echocardiography in patients with left ventricular dysfunction. J Am Coll Cardiol 1996;27:132–139. 12. American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. J Am Soc Echocardiogr 1989;2:358 –367. 13. Weiner DA, Ryan TJ, McCabe CH, Chaitman BR, Sheffield LT, Ferguson JC, Fisher LD, Tristani F. Prognostic importance of a clinical profile and exercise test in medically treated patients with coronary artery disease. J Am Coll Cardiol 1984;3:772–779. 14. Mark DB, Shaw L, Harrell FE, Hlatky MA, Lee KL, Bengtson JR, McCants CB, Califf RM, Pryor DB. Prognostic value of a treadmill exercise score in outpatients with suspected coronary artery disease. N Engl J Med 1991;325:849 – 853. 15. Moriel M, Rozanski A, Klein J, Berman D, Bairey N. The limited efficacy of exercise radionuclide ventriculography in assessing prognosis of women with coronary artery disease. Am J Cardiol 1995;76:1030 –1035. 16. Froelicher V, Thomas M, Pillow C, Lancaster M. Epidemiologic study of asymptomatic men screened by maximal treadmill testing for latent coronary artery disease. Am J Cardiol 1974;34:770 –776. 17. Giagnoni E, Secchi MB, Wu SC, Morabito A, Oltrona L, Mancarella S, Volpin N, Fossa L, Bettazzi L, Arangio G, Sachero A, Folli G. Prognostic value of exercise EKG testing in asymptomatic normotensive subjects. A prospective matched study. N Engl J Med 1983;309:1085–1089. 18. Ellestad M, Wan M. Predictive implications of stress testing. Follow-up of 2700 subjects after maximum treadmill stress testing. Circulation 1975;51:363– 369. 19. Brown KA, Boucher CA, Okada RD, Guiney TE, Newell JB, Strauss HW, Pohost GM. Prognostic value of exercise thallium-201 imaging in patients presenting for evaluation of chest pain. J Am Coll Cardiol 1983;1:994 –1001. 20. Schalet B, Kegel J, Heo J, Segal B, Iskandrian A. Prognostic implications of normal exercise SPECT thallium images in patients with strongly positive exercise electrocardiograms. Am J Cardiol 1993;72:1201–1203. 21. Krivokapich J, Child J, Gerber R, Lem V, Moser D. Prognostic usefulness of positive or negative exercise stress echocardiography for predicting coronary events in ensuing twelve months. Am J Cardiol 1993;71:646 – 651. 22. Pryor DB, Harrell FE, Lee KL, Rosati RA, Coleman RE, Cobb FR, Califf RM, Jones RH. Prognostic indicators from radionuclide angiography in medically-treated patients with coronary artery disease. Am J Cardiol 1984;53:18 –22. 23. Taliercio CP, Clements IP, Zinsmeister AR, Gibbons RJ. Prognostic value and limitations of exercise radionuclide angiography. Mayo Clin Proc 1988;63: 573–582. 24. Kaul S, Lilly DR, Gascho JA, Watson DD, Gibson RS, Oliner CA, Ryan JM, Beller GA. Prognostic utility of the exercise thallium-201 test in ambulatory patients with chest pain: comparison with cardiac catheterization. Circulation 1988;77:745–758. 25. Hachamovitch R, Berman DS, Kiat H, Cohen I, Cabico JA, Friedman J, Diamond GA. Exercise myocardial perfusion SPECT in patients without known coronary artery disease. Incremental prognostic value and use in risk stratification. Circulation 1996;93:905–914. 26. Ladenheim M, Kotler T, Pollock B, Berman D, Diamond G. Incremental prognostic power of clinical history, exercise electrocardiography and myocardial perfusion scintigraphy in suspected coronary artery disease. Am J Cardiol 1987; 59:270 –277. 27. Sawada S, Ryan T, Conley M, Corya B, Feigenbaum H, Armstrong W. Prognostic value of a normal exercise echocardiogram. Am Heart J 1990;120: 49 –55. 28. Limacher M, Quinones M, Poliner L, Nelson J, Winters W, Waggoner A. Detection of coronary artery disease with exercise two-dimensional echocardiography. Description of a clinically applicable method and comparison with radionuclide ventriculography. Circulation 1983;67:1211–1218. 29. Mairesse GH, Marwick TH, Vanoverschelde JJ, Baudhuin T, Wijns W, Melin JA, Detry JR. How accurate is dobutamine stress electrocardiography for detection of coronary artery disease? Comparison with two-dimensional echocardiography and Technetium-99m methoxyl isobutyl isonitrile (Mibi) perfusion scintigraphy. J Am Coll Cardiol 1994;24:920 –927.
CORONARY ARTERY DISEASE/PROGNOSIS WITH STRESS ECHO
551