In patients with normal results on stress single-photon emission computed tomographic (SPECT) studies, coronary artery disease risk factors (RFs) and the mode of testing can influence the trajectory of long-term outcomes. Nevertheless, the combined prognostic impact of these commonly assessed factors has heretofore not been considered. In this study, all-cause mortality rates were assessed in 5,762 patients with normal results on stress SPECT studies. Patients were divided according to mode of stress testing, exercise or pharmacologic, and by number of coronary artery disease RFs. Patients were followed for a mean of 8 ± 4.2 years for all-cause mortality. There were 1,051 deaths (18%), with an annualized mortality rate of 2.2% per year. The RF-adjusted event rate was significantly higher for pharmacologic versus exercise SPECT studies (3.6% per year vs 1.2% per year, p <0.0001) and for patients with increasing numbers of coronary artery disease RFs (p <0.0001). Kaplan-Meier survival analysis revealed wide heterogeneity in all-cause mortality rates when RF burden and performance of exercise versus pharmacologic testing were considered, ranging from only 0.8% per year in exercise patients with no RFs to 4.2% per year in pharmacologic patients with ≥2 RFs. Mortality rates in exercise patients with ≥2 RFs were comparable to those in pharmacologic patients with no RFs. In conclusion, long-term outcomes after cardiac stress testing are synergistically and strongly influenced by RF burden and inability to exercise. Given these findings, prospective study is indicated to determine whether enhanced risk categorization that combines the consideration of these 2 factors improves patient counseling and physician risk management among patients manifesting normal results on stress SPECT studies.
Stress-rest single-photon emission computed tomographic (SPECT) myocardial perfusion imaging is commonly used to risk-stratify patients with suspected or known coronary artery disease (CAD). A central tenet of this application is the understanding that normal stress SPECT results indicate a low likelihood of sustaining adverse clinical events. Accordingly, stress-rest SPECT myocardial perfusion imaging is commonly used for many acute-term decisions in cardiac practice, such as determining the need for cardiac catheterization in patients with chest pain symptoms, the appropriateness for referral for myocardial revascularization procedures, and preoperative risk assessment. Increasingly, however, it is recognized that long-term outcomes in patients with normal stress-rest SPECT results are influenced by factors such as CAD risk factor burden and plaque burden as assessed by coronary artery calcium scanning. Heretofore, however, no study has further considered how long-term outcomes after stress SPECT imaging might be influenced by risk factor burden and patients’ ability to exercise, which are routinely assessed at the time of stress testing. Accordingly, in this study, we assessed long-term mortality in patients who underwent exercise versus pharmacologic stress testing, with the 2 groups stratified according to risk factor burden.
Methods
We retrospectively identified 15,047 unique patients who underwent stress-rest SPECT myocardial perfusion imaging at St. Luke’s and Roosevelt Hospitals from January 1995 to October 2010. All studies were performed on the basis of physician referral, because of clinical suspicion of CAD, and/or for workup for risk stratification. Before stress testing, a brief clinical history was obtained from each patient by a supervising physician. The collected data included age, gender, height, weight, ethnicity, medication regimen, the presence of CAD risk factors (hypertension, smoking, diabetes mellitus, hyperlipidemia, and family history of premature CAD), the presence and type of chest pain symptoms, and history of cardiac disease (i.e., status post myocardial infarction or myocardial revascularization procedures). Exclusion criteria for this study were a history of cardiac disease and equivocal, probably abnormal, or abnormal SPECT results. Among patients who underwent multiple SPECT studies, only the first SPECT study was considered in our analyses. After applying these exclusion criteria, our final study population consisted of 5,762 patients (mean age 60 ± 13 years, 37% men) who had no known histories of heart disease and normal results on stress SPECT myocardial perfusion imaging. This study was approved by the St. Luke’s-Roosevelt Hospital institutional review board.
All patients underwent same-day rest-stress or 2-day stress-rest myocardial perfusion SPECT imaging with technetium-99m sestamibi. Patients were asked to withhold β blockers and calcium channel blockers for 24 hours before testing. For rest studies, patients were injected with 9 to 10 mCi of technetium-99m sestamibi, and SPECT images were acquired 30 to 45minutes after rest injection. All SPECT studies were acquired on commercially available cameras using a 180° arc, high-resolution parallel-hole collimators, and 64 stops with 20 s/stop for a total imaging time of 25 minutes. All images were acquired using a 64 × 64 image matrix. Patients underwent symptom-limited treadmill exercise testing using the Bruce protocol with continuous 12-lead electrocardiographic monitoring. At peak heart rate, 25 to 30 mCi of technetium-99m sestamibi (the dose was weight based) was injected. For those patients who were unable to exercise, pharmacologic stress testing was performed using dipyridamole or regadenoson. Dipyridamole at a rate of 0.142 mg/kg/min was infused over 4 minutes, with sestamibi injection at 7 minutes. Regadenoson 0.4 mg was infused over 10 seconds according to standard protocol. Patient then underwent poststress SPECT image acquisition, which included poststress gated images to assess left ventricular function. Gated SPECT images were acquired at 8 frames/cardiac cycle.
All stress and SPECT data were analyzed by expert readers for perfusion and wall motion abnormalities. Semiquantitative visual interpretation was performed using segmentation of short-axis slices into apical, mid, and basal slices and vertical long-axis myocardial tomograms. All studies were assessed by an expert reader and the results classified as normal, probably normal, equivocal, probably abnormal, or definitely abnormal. Only patients determined to have normal or probably normal perfusion and normal left ventricular function were included in this study.
All-cause mortality was determined for all 5,762 patients using the Social Security Death Index. The mean follow-up period was 8 ± 4.2 years (interquartile range 4.5 to 11.1). All analyses were carried out using SPSS version 16.0 (IBM, Armonk, New York). Continuous variables are expressed as mean ± SD. Categorical variables are expressed as numbers and percentages. Univariate analyses of continuous variables were performed using 2-tailed Student’s t tests and chi-square tests for categorical variables. One-way analysis of variance with a post hoc Bonferroni test was used to compare the means of continuous variables among multiple groups. Adjusted multivariate analysis was performed using a Cox proportional-hazards model including age, gender, body mass index, chest pain symptoms, ethnicity, stress mode, and CAD risk factors. Actuarial survival curves for the groups were calculated using the Kaplan-Meier estimate, and the difference between curves was calculated using log-rank statistics. Cox proportional-hazards analyses was applied to estimate adjusted odds ratios and 95% confidence intervals (CIs) for potential predictors of survival. A p value <0.05 was considered significant.
Results
The baseline clinical characteristics of all patients according to the mode of stress testing are listed in Table 1 . The pharmacologic group was older, had fewer men, had more patients with hypertension and diabetes, and had higher prescribed use of antihypertensive medications. Among the 5,762 patients, there were 1,051 deaths (18%) and an overall event rate of 2.2% per year during the follow-up period. All-cause mortality and annualized event rates were substantially higher in the pharmacologic patients than the exercise patients ( Tables 1 and 2 ). After adjusting for age, gender, CAD risk factors, ethnicity, chest pain symptoms, and body mass index, pharmacologic patients had >2 times the mortality risk of exercise patients (hazard ratio [HR] 2.3, 95% CI 1.9 to 2.6, p <0.0001). When dividing patients into 5-year age ranges, a higher event rate was observed for pharmacologic patients in all age groups, in men and in women (p <0.0001; Figure 1 ). After adjusting of all important clinical variables, within the exercise cohort, patients who achieved higher heart rate responses had lower all-cause mortality (HR 0.99, 95% CI 0.98 to 0.99, p <0.001) compared to those who did not. Similarly, within the pharmacologic cohort, patients who had higher heart rate responses had lower mortality (HR 0.99, 95% CI 0.98 to 0.99, p = 0.005).
| Variable | Exercise Group (n = 2,994) | Pharmacologic Group (n = 2,768) | p Value |
|---|---|---|---|
| Age (yrs) | 55 ± 12 | 65 ± 13 | <0.0001 |
| Men | 1,204 (41%) | 897 (33%) | <0.0001 |
| Ethnicity | <0.0001 | ||
| Caucasian | 423 (15%) | 304 (11%) | |
| Hispanic | 1,340 (46%) | 1,158 (43%) | |
| African American | 1,066 (37%) | 1,215 (45%) | |
| Symptoms | <0.0001 | ||
| Asymptomatic | 985 (33%) | 995 (36%) | |
| Nonanginal chest pain | 886 (29%) | 661 (24%) | |
| Atypical chest pain | 513 (17%) | 429 (15%) | |
| Typical chest pain | 322 (11%) | 297 (11%) | |
| Dyspnea only | 288 (10%) | 386 (14%) | |
| Hypertension ∗ | 1,931 (65%) | 2,272 (82%) | <0.0001 |
| Diabetes mellitus | 644 (22%) | 988 (36%) | <0.0001 |
| Hyperlipidemia † | 1,164 (39%) | 1,041 (38%) | 0.289 |
| Smokers | 721 (24%) | 725 (26%) | 0.065 |
| Family history | 463 (16%) | 314 (11%) | <0.0001 |
| Body mass index (kg/m 2 ) | 29.7 ± 8.4 | 31 ± 11.1 | <0.0001 |
| Body mass index (kg/m 2 ) | <0.0001 | ||
| <25 | 641 (22%) | 662 (25%) | |
| 25–29.9 | 1,055 (37%) | 772 (29%) | |
| ≥30 | 1,127 (39%) | 1,206 (45%) | |
| All-cause mortality | 262 (9%) | 789 (29%) | <0.0001 |
| Chest pain history | 1,729 (59%) | 1,376 (50%) | <0.0001 |
| Rest heart rate (beats/min) | 73 ± 14 | 72 ± 15 | 0.054 |
| Heart rate response | 106 ± 26 | 13 ± 28 | <0.0001 |
| Rest systolic blood pressure (mm Hg) | 137 ± 20 | 147 ± 26 | <0.0001 |
| Rest diastolic blood pressure (mm Hg) | 84 ± 12 | 86 ± 13 | <0.0001 |
| Calcium channel antagonist | 436 (15%) | 543 (20%) | <0.0001 |
| Vasodilators | 422 (15%) | 550 (20%) | <0.0001 |
| β blockers | 458 (15%) | 802 (29%) | <0.0001 |
| Diuretics | 357 (12%) | 489 (18%) | <0.0001 |
∗ History of hypertension and/or use of antihypertensive medications.
| Patient Group | Event Rate (% per year) | HR | 95% CI | p Value |
|---|---|---|---|---|
| Exercise patients | <0.0001 | |||
| 0 risk factor | 0.8 | 1.00 (reference) | ||
| 1 risk factor | 1.0 | 1.34 | 0.92–1.96 | 0.13 |
| ≥2 risk factors | 1.8 | 2.7 | 1.82–3.9 | <0.0001 |
| Pharmacologic patients | <0.0001 | |||
| 0 risk factor | 2.6 | 1.00 (reference) | ||
| 1 risk factor | 3.5 | 1.3 | 0.96–1.67 | 0.09 |
| ≥2 risk factors | 4.2 | 2.0 | 1.5–2.6 | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
