Author + information
- Received May 13, 1998
- Revision received July 13, 1998
- Accepted August 6, 1998
- Published online December 1, 1998.
- Lauro Cortigiani, MDa,*,
- Claudio Dodi, MDa,
- Emilio A Paolini, MDa,
- Daniele Bernardi, MDa,
- Gabriele Bruno, MDa and
- Eugenio Nannini, MDa
- ↵*Address for correspondence: Dr. Lauro Cortigiani, U.O. Malattie Cardiovascolari, Ospedale Campo di Marte, 55032 Lucca-Italy
Objectives. In this study we sought to investigate the prognostic value of pharmacological stress echocardiography in women referred for chest pain, having unknown coronary artery disease.
Background. The noninvasive identification of a high-risk subgroup among women with chest pain and unknown coronary artery disease is an unresolved task to date.
Methods. A total of 456 women (mean [±SD] age 63 ± 10 years) underwent pharmacological stress echocardiography with either dipyridamole (n = 305) or dobutamine (n = 151) for evaluation of chest pain and were followed-up for 32 ± 19 months. None of them had a previous diagnosis of coronary artery disease.
Results. No major complication occurred during stress testing. Five tests (1.1%) were prematurely interrupted because of the appearance of side effects. Echocardiographic positivity was identified in 51 patients. During the follow-up, 23 cardiac events occurred: 3 deaths, 10 infarctions and 10 cases of unstable angina; an additional 21 patients underwent coronary revascularization. At Cox analysis, the echocardiographic evidence of ischemia was found as the only independent predictor of hard cardiac events (death, infarction) (odds ratio [OR] = 27.5; 95% confidence interval [CI] = (6.5 to 115.5; p = 0.0000). When spontaneous cardiac events (death, infarction and unstable angina) were considered as endpoints, the positive echocardiographic result (OR = 23.9; 95% CI = 8.6 to 66.8; p = 0.0000) and family history of coronary artery disease (OR = 3.7; 95% CI = 1.5 to 9.1; p = 0.0037) were independently correlated with prognosis. By using an interactive stepwise procedure, the prognostic value of stress echocardiography was found to be incremental to that provided by clinical variables, both considering hard and spontaneous cardiac events as endpoints. The 3-year survival rate for the negative and the positive population was respectively, 99.5% and 69.5% (p = 0.0000) considering hard cardiac events, 99.2% and 50.6% (p = 0.0000) considering spontaneous cardiac events.
Conclusions. Pharmacological stress echocardiography is safe, highly feasible and effective in risk stratification of women with chest pain and unknown coronary artery disease, also when hard endpoints are considered. Its use can have relevant implications in daily clinical practice for selection of patients needing further investigations.
The noninvasive identification of a high-risk subgroup among women referred for chest pain having unknown coronary artery disease (CAD) is a particularly difficult challenge in daily clinical practice. In fact, although angina is about two times more frequently the initial manifestation of CAD in women than in men (65 vs. 35%) (1), in women it is associated with high prevalence of normal epicardial coronary arteries (2–5)and, therefore, with a generally benign prognosis (6–8). Moreover, though exercise testing is the most widely used method for the noninvasive prognostic assessment of CAD (9), it is unreliable for evaluation of women since it carries a high incidence of false positive responses (10–13), adversely affecting the positive predictive value.
Echocardiography combined with physical or pharmacological stress is an effective and not gender-related tool for diagnostic purposes in women, having shown to provide significant higher diagnostic accuracy when compared with exercise testing (14–17). Despite this, little is known (18)about the value of stress echocardiography in risk stratification of women with chest pain syndrome and unknown CAD. In particular, the prognostic value of stress echocardiography in women has been demonstrated for exercise (18), but not for pharmacological stress echocardiography.
Based on these findings, we sought to determine the prognostic value of pharmacological stress echocardiography in a series of 456 women referred for chest pain, in the absence of a previous diagnosis of CAD.
From November 1990 to October 1996, a total of 456 women (mean [±SD] age 63 ± 10 years) underwent pharmacological stress echocardiography with either dipyridamole (n = 305) or dobutamine (n = 151) for evaluation of chest pain, that was typical in 162 (36%) patients and atypical in the remaining 294 (64%), at two different Institutions. All patients had interpretable echocardiographic images both in resting condition and during stress testing. None of them had known CAD, as defined by history of myocardial infarction, unstable angina, coronary revascularization and/or angiographically assessed stenosis (>50%) of any of the three coronary arteries or their major branches (of the total cohort, 18 patients had previously [>6 months from stress testing] undergone coronary angiography). Furthermore, no patient had significant valvular disease, dilated or hypertrophic cardiomyopathy.
In 445 (97.6%) patients, stress echocardiography was performed after an adequate wash-out from antianginal drugs, discontinuing beta-adrenergic blocking agents for at least 48 h and long-acting nitrates and calcium channel blockers for at least 24 h before the test. In the remaining 11 (2.4%) patients, the test was performed under therapy with beta-blockers (n = 6), calcium channel blockers (n = 5) and/or nitrates (n = 6). All patients evaluated by dipyridamole were also not taking phylline-containing drugs or beverages for at least 24 h.
In our echo laboratories, the choice of one test over the other is governed by several possible factors such as clinical issue, personal experience of individual echocardiographer and known contraindications to the use of the available drug (such as preexistence of complex ventricular arrhythmias in the case of dobutamine or severe obstructive pulmonary disease in the case of dipyridamole).
Dipyridamole was given intravenously according to the high dosage protocol (up to 0.84 mg/kg over 10 minutes) (19). Dobutamine was administered using an infusion pump in steps of three minutes each with an initial dose of 5 μg/kg/min up to 40 μg/kg/min as a maximum (20). Starting in April, 1993, the two protocols were modified with the coadministration of atropine, if no signs of ischemia were detected and the heart rate was inferior than 85% of the age- and sex-predicted maximum at the end of the drug’s infusion (21,22).
Two-dimensional echocardiogram and 12-lead electrocardiogram were continuously monitored during drug administration and until the heart rate has returned to baseline value ±10% during the recovery phase. During the procedure, cuff blood pressure and the electrocardiogram were recorded every minute. Criteria for test interruption were: onset of obvious new wall motion abnormalities, severe chest pain, horizontal or downsloping ST-segment depression ≥2 mm, ST-segment elevation ≥1.5 mm, systolic blood pressure >220 mm Hg, diastolic blood pressure >120 mm Hg, reduction in systolic blood pressure ≥30 mm Hg, supraventricular or ventricular tachyarrythmias, untolerable symptoms. Intravenous aminophylline (up to 240 mg) and metoprolol (up to 5 mg) were promptly available to reverse the effects of dipyridamole and dobutamine, respectively.
Echocardiographic images were obtained continuously from the standard apical and parasternal views using commercially available instruments (Sonos 2000, Hewlett Packard, Palo Alta, California; Sonotron 800 or System-Five, Vingmed, Santa Clara, California; Ultramark-7, ATL; SSH 140, Toshiba, New York, New York). From 1990 to 1995, images were continuously recorded using S-VHS videotape recorders (Panasonic MD 830; Panasonic 7330; Panasonic 6200, Osaka, Japan) for off-line visual analysis; starting in 1996, images were also recorded using a quad-screen cine-loop system. Images were evaluated independently by two expert observers. In case of disagreement, a third observer, whose judgment was binding, evaluated the images.
Regional wall motion was semiquantitatively assessed using a 16 segment model of the left ventricle according to the recommendations of the American Society of Echocardiography (23). A four point score was assigned to each segment as follows: 1 = normal; 2 = hypokinesia; 3 = akinesia; and 4 = dyskinesia. A wall motion score index (WMSI) was obtained dividing the sum of individual segment scores by the number of considered segments. For each patient both the WMSI at baseline and at the peak of drug infusion were obtained.
The result of a test was considered positive for ischemia when any new or worsening of preexisting regional wall motion abnormalities were detected. Moreover, electrocardiographic (ECG) changes and chest pain were not considered per se as markers of ischemia in absence of induced new wall motion abnormalities.
Electrocardiographic changes were considered significant for ischemia if a horizontal or downsloping ST segment depression or elevation ≥0.1 mV from baseline at 80 msec after the J point was assessed in at least two contiguous leads. In the case of right bundle branch block, the ST-segment shift was considered significant when it also occurred in leads V5and/or V6(24).
The mean length of the follow-up was 32 ± 19 months. Follow-up data were obtained in 80% of cases from telephone interviews with the patients and in the remaining 20% of cases, from review of the patient’s hospital chart, contact with the patient’s physician, or periodic visits to our outpatient clinic.
The clinical events recorded during the follow-up were: cardiac and noncardiac deaths, nonfatal myocardial infarction, unstable angina and coronary revascularization procedures (surgery or angioplasty).
The cause of death was elucidated from hospital or physician records; death was attributed to a cardiac origin in the case of documentation of significant arrhythmias and/or cardiac arrest, congestive heart failure or myocardial infarction. Moreover, any death occurring suddenly out of the hospital was ascribed to a cardiac cause. The diagnosis of acute myocardial infarction was made on the basis of symptoms, ECG changes and cardiac enzyme level increase. Unstable angina was defined as angina at rest or change in pattern of preexisting angina requiring hospitalization.
Values were expressed as ± mean SD for continuous variables and as frequency and percentage for categorical variables. Continuous variables were compared using the Student’s unpaired ttest, while differences of categorical variables were assessed by the chi-square test. The Kaplan–Meier method was used for estimation of infarction-free survival and of spontaneously occurring event-free survival. For survival analysis, only one event was considered in each patient. When patients underwent coronary revascularization, they were censored at the time of the procedure. Likewise, when patients died for noncardiac cause, they were censored at the time of death. The differences in survival curves were analyzed using the log-rank test. The capability of certain variables to predict subsequent outcome was assessed by the Cox proportional hazard model using univariate and stepwise multivariate procedure (SPSS for windows, 1995, Chicago, Illinois). Furthermore, in an attempt to investigate the prognostic value of stress echocardiography incremental to clinical data, an interactive stepwise procedure was performed, where variables were included into the model in the same order as in the clinical practice. Therefore, clinical data were first analyzed and the global chi-square was calculated. Subsequently, a second step was created after addition of stress echocardiography results to the independent predictors at first step. The incremental prognostic value of the added variables was assessed by comparison of the global chi-square at each step. At each step, a p value ≤0.1 was taken as the required level of significance for entering a variable into the model.
The differences in risk were expressed as odds ratio (OR) with the corresponding 95% confidence interval (CI). The variables included in the analysis were: age (< or ≥65 years), typical chest pain, coronary risk factors (family history of CAD, hypertension, hypercholesterolemia, cigarette smoking and diabetes) and stress testing results (positive/negative echocardiographic result, WMSI at peak of drug infusion, ECG changes during test and angina during test). A value of p < 0.05 was considered statistically significant.
Feasibility and side effects
No major complication occurred. Of 456 tests, 5 were prematurely interrupted because of the appearance of limiting side effects consisting in symptomatic hypotension (n = 2) and severe chest pain in the absence of new wall motion abnormalities (n = 1) during dipyridamole, atrial fibrillation (n = 1) and nonsustained ventricular tachycardia (n = 1) during dobutamine stress testing. Therefore, the overall feasibility was 99% for dipyridamole and 98.7% for dobutamine stress echocardiography. All side effects reversed by antidote administration (aminophylline for dipyridamole and metoprolol for dobutamine). Results for these five patients, who underwent submaximal stress testing, were included in the analysis as negative ones.
Stress echocardiography results
Echocardiographic evidence of ischemia was identified in 51 (11%) patients. Of them, 34 had been investigated by dipyridamole and 17 by dobutamine stress echocardiography. In the positive population, the WMSI increased from 1.03 ± 0.12 in resting condition to 1.31 ± 0.29 at peak of drug infusion.
Of the remaining 405 patients with negative echocardiographic result, 68 (17%) developed an isolate ST-segment depression during the test; of them, 31 (46%) had a history of hypertension. Table 1illustrates the clinical characteristics, as well as the baseline and stress echocardiography findings, for the positive and the negative population.
Thirteen patients (2.8%) were lost to follow-up; all of them had stress testing negativity. In the remaining 443 patients, 23 cardiac events occurred: 3 cardiac deaths, 10 myocardial infarctions, and 10 cases of unstable angina. Moreover, five patients died of noncardiac causes: three of cancer and two of stroke. They all had negative echocardiographic result. Finally, 21 patients underwent coronary revascularization with either surgery (n = 10) or angioplasty (n = 11), of whom 14 within three months (mean 1.2 ± 0.7 months) and 7 after three months (mean 12.0 ± 14.7 months) from stress testing. Table 2illustrates the incidence of cardiac events and revascularization procedures in the positive and in the negative population. Of note was that no cardiac death occurred among patients with negativity of test.
Considering the positive population, there were 10 events (2 cardiac deaths, 4 infarctions, and 4 cases of unstable angina) among 37 patients evaluated by dipyridamole, and 6 events (1 cardiac death, 3 infarctions, and 2 cases of unstable angina) among 17 patients evaluated by dobutamine stress testing.
Finally, no cardiac event was experienced by 68 patients who developed ECG but not echocardiographic positive response during test.
Seven variables were identified as univariate predictors of hard cardiac events (death, myocardial infarction). In descending order they were: positive echocardiographic result (p = 0.0000), peak-stress WMSI (p = 0.0000), angina during test (p = 0.0001), ECG changes during test (p = 0.0053), typical chest pain (p = 0.0150), hypercholesterolemia (p = 0.0208) and family history of CAD (p = 0.0251) (Table 3).
A nonsignificant higher positive predictive value was found for dobutamine in comparison with dipyridamole stress echocardiography (23.5% vs. 17.6%, p = 0.62).
With a Cox analysis, the positive echocardiographic result (OR = 27.5; 95% CI = 6.5 to 115.5; p = 0.0000) was found as the only independent predictor of hard cardiac events. The infarction-free survival rate was 99.5% for patients with negative and 69.5% for patients with positive echocardiographic result (p = 0.0000) (Fig. 1).
Among clinical variables analyzed at first step of the interactive stepwise procedure, typical chest pain (OR = 3.6; 95% CI = 1.1 to 12.1; p = 0.0346) and hypercholesterolemia (OR = 3.1; 95% CI = 1.0 to 9.5; p = 0.0424) were independent predictors of hard cardiac events; this model showed a global chi-square of 15.8 (p = 0.0012). In the second step, with the addition of stress testing findings, the positive echocardiographic result (OR = 29.9; 95% CI = 7.0 to 127.0; p = 0.0000) was the only prognostic predictor; the global chi-square of this combined clinical and stress testing model was 81.0 (p = 0.0000).
When spontaneous cardiac events (death, myocardial infarction and unstable angina) were considered as endpoints, six variables showed univariate prognostic importance: positive echocardiographic result (p = 0.0000), angina during test (p = 0.0000), peak-stress WMSI (p = 0.0000), family history of CAD (p = 0.0002), ECG changes during test (p = 0.0002) and typical chest pain (p = 0.0003) (Table 4).
No significant difference was evidenced for the positive predictive value provided by dobutamine and dipyridamole stress testing (35.3% vs. 29.4%, p = 0.67).
At multivariate analysis, the positive echocardiographic result (OR = 23.9; 95% CI = 8.6 to 66.8; p = 0.0000) and family history of CAD (OR = 3.7; 95% CI = 1.5 to 9.1; p = 0.0037) were independently correlated with prognosis. The event-free survival rate was 99.2% for patients with negative and 50.6% for patients with positive echocardiographic result (p = 0.0000) (Fig. 2).
Of note, in the subset of patients with positive echocardiographic result, the concomitant evidence of ischemia at ECG was predictive of worse prognosis (p = 0.06) (Fig. 3); on the contrary, the presence or absence of ischemic ECG changes did not portend different clinical outcome (p = 0.24) among patients having negative echocardiographic result of test. At first model of the interactive stepwise procedure, including clinical covariates only, typical chest pain (OR = 4.4; 95% CI = 1.7 to 11.4; p = 0.0022) and family history of CAD (OR = 3.4; 95% CI = 1.5 to 8.2; p = 0.0050) were predictors of spontaneous cardiac events; the global chi-square was 30.4 (p = 0.0000). After the addition of stress testing findings, the positive echocardiographic result (OR = 23.7; 95% CI = 8.6 to 64.9; p = 0.0000) and family history of CAD (OR = 3.7; 95% CI = 1.5 to 8.9; p = 0.0034) had independent prognostic importance; the global chi-square at this second step increased to 146.7 (p = 0.0000).
This study shows that, in women referred for chest pain and without prior diagnosis of CAD, the development of new wall motion abnormalities during pharmacological stress has a strong independent prognostic value, incremental to that provided by clinical data, when both hard and spontaneous cardiac events are considered as endpoints. On the contrary, the negative echocardiographic response to stress identifies a low-risk population having cardiac event-rate of <1% over 3 year follow-up. It is notable that patients developing ST-segment depression in the absence of new wall motion abnormalities have an excellent prognosis (in our study none of them experienced a cardiac event). Importantly, the low event rate evidenced in our patients showing negative echocardiographic result is comparable to that observed in patients with chest pain and normal coronary angiography (25,26).
Comparison with previous studies
For a group of patients having low prevalence of organic CAD and, therefore, a very favorable prognosis, such as women complaining of chest pain, it is difficult for any test to have capability to discriminate among low-and high-risk subsets. Moreover, the results of one test can be affected by the gender of the study population, such as in the case of exercise electrocardiography, providing a high incidence of false-positive responses in women (10–13)and, consequently, the inability to accurately stratify patients’ risk. The echocardiographic sign of ischemia during stress echocardiography is not gender-related and it is well-recognized that this technique is accurate for the noninvasive diagnosis of CAD in women (14–17). Despite this, little is known (18)about the value of stress echocardiography in risk stratification of women with chest pain syndrome. Heupler et al. (18)have recently investigated the prognostic implications of exercise echocardiography in a large series of women with chest pain, having known (18%) or unknown (82%) CAD. When the group of 416 patients without previous CAD was analyzed using sequential Cox models, a positive echocardiographic test result was the most powerful predictor of future cardiac events (cardiac-related death, myocardial infarction and late revascularization procedure), providing incremental prognostic value to clinical and exercise data.
Our results, obtained using a pharmacological stress, expand data by Heupler et al., since they demonstrate an independent and incremental prognostic power to clinical variables by echocardiographic evidence of ischemia also when hard cardiac events are considered as endpoints.
Stress test technique
Although dipyridamole and dobutamine act by different mechanisms, it is now clear that their accuracy is virtually identical (27)as is their prognostic power (28,29). In our opinion, the choice of one test over the other can be made interchangeably according to the presence or absence of specific contraindications of either drug. This is our standard practice in daily routine, and we find it useful and feasible. To use only one test, means for the cardiologist to self-limit his diagnostic tools; to become familiar with both means to get the most from this extremely flexible and versatile technique (30).
The results of this study have major implications in clinical practice, since the use of pharmacological stress echocardiography can be of help to discriminate patients needing further invasive investigations from those needing conservative treatment in the context of a low-risk population, such as that represented by women with chest pain and unknown CAD. However, stress echocardiography is not to be considered an alternative, but rather a complementary method to exercise testing, that remains the first choice of the screening phase, being physiological, simpler, very low-cost and providing a high negative predictive value, similar to that shown by stress echocardiography (31). This last can be recommended in selected conditions, such as in women with equivocal or ischemic ECG response during exercise and in women with the inability to exercise or whose electrocardiogram is uninterpretable.
Limitations of the study
Since it is an observational study, the decision to perform coronary revascularization was taken on the basis of the integration of stress testing results with the individual clinical condition. As a consequence, revascularization procedures were performed in 33% of patients with positive and in only 1% of patients with negative echocardiographic test results. It is presumable that this dropout process can have lowered the positive predictive value of the test, although the echocardiographic evidence of ischemia was found as a very strong and independent predictor of outcome.
The stress protocol was not performed with coadministration of atropine in all patients since atropine was introduced in the clinical practice (21,22)when the study was started. This fact can have reduced the rate of positive echocardiographic results, but its impact on prognosis seems not to have been relevant, based on the excellent prognosis of patients with echocardiographic negativity.
In the majority of the cases, the wall motion analysis was performed from video tape and not in the cine-loop fashion. Surely digital acquisition represents a more practical form of stress testing storage, producing the on-line image evaluation more easily and quickly. However, it has not been found to afford significant diagnostic advantages when compared with traditional video tape analysis (32).
Pharmacological stress echocardiography is safe, highly feasible and effective in risk stratification of women with chest pain syndrome and unknown CAD, also when hard endpoints are considered. Its use can have relevant implications in daily clinical practice for selection of patients needing further investigations.
- confidence interval
- coronary artery disease
- odds ratio
- wall motion score index
- Received May 13, 1998.
- Revision received July 13, 1998.
- Accepted August 6, 1998.
- American College of Cardiology
- Chaitman B.R,
- Bourassa M.G,
- Davis K,
- et al.
- Orencia A,
- Bailey K,
- Yawn B.P,
- Kottke T.E
- Detry J.M,
- Kapita B.M,
- Cosyns J,
- Sottiaux B,
- Brasseur L.A,
- Rousseau M.F
- Barolsky S.M,
- Gilbert C.A,
- Faruqui A,
- Nutter D.O,
- Schlant R.C
- Sawada S.G,
- Ryan T,
- Fineberg N.S,
- Armstrong W.F,
- McHenry P.L,
- Feigenbaum H
- Williams M.J,
- Marwick T.H,
- O’Gorman D,
- Foale R.A
- Marwick T.H,
- Anderson T,
- Williams M.J,
- et al.
- Heupler S,
- Metha R,
- Lobo A,
- Leung D,
- Marwick T.H
- Picano E,
- Lattanzi F,
- Masini M,
- Distante A,
- L’Abbate A
- Picano E,
- Pingitore A,
- Conti U,
- et al.
- Chaitman B.R
- Kemp H.G,
- Kronmal R.A,
- Vlietstra R.E,
- Frye R.L
- Pingitore A,
- Picano E,
- Quarta Colosso M,
- et al.
- Picano E
- Cortigiani L,
- Dodi C,
- Paolini E,
- Iori E,
- Bruno G,
- Nannini E
- Castini D,
- Gentile F,
- Ornaghi M,
- et al.