Although stress gated technetium-99m single-photon emission computed tomographic (SPECT) myocardial perfusion imaging (MPI) is useful in differentiating ischemic from nonischemic cardiomyopathy, its prognostic usefulness in this patient population is not well understood. Consecutive unique patients with suspected coronary artery disease who, for clinical indications, underwent technetium-99m rest and stress MPI demonstrating ejection fractions ≤40% by gated SPECT imaging were retrospectively identified. In addition to prescan variables, previously defined cutoffs for gated SPECT parameters using visual and standard 17-segment semiquantitative scoring were applied and related to the occurrence of cardiac death up to 5 years after MPI. Of the 475 patients fulfilling criteria for study inclusion, follow-up was complete in 444 (93%) over 3.7 ± 1.6 years. Of 393 patients without subsequent early (≤60 days) coronary revascularization, cardiac death occurred in 64 (16%). The summed stress score, an MPI measure of the extent and severity of coronary artery disease that also accounts for the ischemic burden, was the gated SPECT parameter most related to cardiac death with Kaplan-Meier 5-year cardiac death–free survival of 85.6% and 67.3% in patients with summed stress scores ≤8 and >8, respectively (p <0.001). In multivariate Cox regression analysis, a summed stress score >8 independently contributed to cardiac death (adjusted hazard ratio 2.20, 95% confidence interval 1.34 to 3.61), and its addition to the model significantly increased the global chi-square value over prescan variables (from 32.46 to 41.67, p = 0.002). In conclusion, stress MPI data from gated technetium-99m SPECT scans are useful for the prediction of cardiac death in patients with moderate to severe left ventricular systolic dysfunction in whom there is suspicion of underlying coronary artery disease.
Differentiating ischemic from nonischemic cardiomyopathy is of considerable importance for optimal patient management. Our group and others have previously demonstrated that stress gated technetium-99m single-photon emission computed tomographic (SPECT) myocardial perfusion imaging (MPI) is useful in differentiating ischemic from nonischemic cardiomyopathy. While the findings from these studies have important diagnostic implications, the prognostic usefulness of this cardiovascular imaging technique in this patient population is not well understood. Accordingly, in the present study, we examined the usefulness of stress gated technetium-99m SPECT MPI for the prediction of cardiac death in patients with moderate to severe left ventricular systolic dysfunction and suspected coronary artery disease (CAD).
Methods
The electronic database within the Nuclear Cardiology Laboratory at Hartford Hospital (Hartford, Connecticut) was queried to identify consecutive unique patients who, for clinical indications, underwent technetium-99m sestamibi rest and stress MPI from January 1, 1996, to December 31, 2004, inclusive, with ejection fractions ≤40% by gated SPECT imaging. Excluded were patients with any of the following at the time of MPI: (1) known significant CAD (≥50% stenosis in any major epicardial vessel by angiography), (2) previous coronary revascularization, or (3) previous myocardial infarction as documented by clinical history or as defined by associated changes (Q-waves or QS complexes in the absence of QRS prolongation ) on rest electrocardiography. Also excluded were patients with implanted cardiac defibrillators or pacemakers at the time of MPI. Patients fulfilling criteria for study inclusion but having undergone coronary revascularization procedures early (≤60 days) after MPI were excluded from the prognostic portion of the analysis. This study was approved by and conducted within guidelines of the institutional review board at Hartford Hospital.
Exercise and pharmacologic stress testing was performed in accordance with standard protocols and techniques. Radiopharmaceutical dosing, gated SPECT acquisition, and image processing were performed using methods previously described and within guidelines established by the American Society of Nuclear Cardiology. Notably, our laboratory currently uses the same gated SPECT protocol for rest and stress technetium-99m MPI except now with attenuation correction.
All images were visually interpreted during daily reading sessions by a consensus of ≥2 experienced readers without clinical information using a standard 17-segment model and semiquantitative scoring. For the assessment of myocardial perfusion on stress and rest imaging, each segment was scored on a scale of 0 to 4 (0 = normal activity, 1 = mild, 2 = moderate, 3 = severe reduction in photon activity, 4 = complete absence of photon activity). For each image, a summed stress score and a summed rest score was calculated by adding the segment scores. A summed difference score was derived for each image by subtracting the summed rest score from the summed stress score. Evaluation of left ventricular systolic function on the stress images was performed by assessment of endocardial border excursion and regional wall thickening with each segment scored on a scale of 0 to 5 (0 = normal, 1 = mild, 2 = moderate, 3 = severe hypokinesia, 4 = akinesia, 5 = dyskinesia). The variance of the average wall motion score for left anterior descending, circumflex and right coronary artery vascular territories was expressed as the regional wall motion variance. The ejection fraction was calculated using an automated quantitative method and confirmed visually. On the basis of previously reported data from our group, identified upper bounds of the 95% confidence intervals for nonischemic cardiomyopathy (8.39, 1.94, and 0.114 for the summed stress score, summed difference score, and regional wall motion variance, respectively) were applied as previously described.
Patient follow-up was achieved through mailed questionnaires and/or scripted telephone interviews as well as by review of hospital admission records and the Social Security Death Index. In patients who were confirmed to have died, experienced personnel unaware of MPI data reviewed information provided by hospital admission records or death certificates to ascertain cause, which was categorized as cardiac (death attributable to lethal arrhythmia, heart failure, myocardial infarction, or sudden death) or noncardiac. Follow-up was considered incomplete if a surviving patient was with <1 year of data (by questionnaire, telephone interview, or hospital admission records) or apparent death in a patient could not be confirmed to ascertain cause. Patients were followed for up to 5 years after MPI.
All numerical values are presented as mean ± SD or as proportions. Discrete or dichotomized variables were compared using chi-square or Fisher’s exact tests and continuous variables using Student’s t tests. Cumulative cardiac death–free survival was calculated using the Kaplan-Meier procedure and compared using the log-rank test. Annualized rates of cardiac death were calculated as the number of occurrences divided by the total exposure years. Multivariate Cox regression was used to identify variables that independently contributed to cardiac death. A forward stepwise (Wald method) selection procedure was applied with prescan followed by gated SPECT imaging and thresholds of p ≤0.05 and p ≥0.10 for variable entry and removal, respectively. A p value <0.05 was considered statistically significant. All statistical analyses were 2 tailed and performed using SPSS version 15.0 (SPSS, Inc., Chicago, Illinois).
Results
Of the 475 patients fulfilling criteria for study inclusion, follow-up was complete in 444 (93%) over 3.7 ± 1.6 years. The prescan characteristics of these 444 patients are listed in Table 1 . Notably, 70% of patients possessed multiple CAD risk factors. Seventy-six percent of patients underwent pharmacologic stress. On gated SPECT imaging, the mean ejection fraction in the patients was 30.8 ± 7.3%. Fifty-five percent of patients demonstrated ≥1 of the following: summed stress score >8 (37%), regional wall motion variance >0.114 (36%), or summed difference score >1 (28%). Substantial reversibility (summed difference score >8) was demonstrated in 5% of patients.
| Variable | Value |
|---|---|
| Age (years) | 63 ± 13 (25–88) |
| Men | 285 (64%) |
| History | |
| Chest pain or dyspnea | 332 (75%) |
| Heart failure | 189 (43%) |
| Angina pectoris | 153 (34%) |
| Systemic hypertension | 344 (77%) |
| Tobacco smoker | 239 (54%) |
| Diabetes mellitus | 162 (36%) |
| Hypercholesterolemia | 137 (31%) |
| Cardiac medications | |
| Angiotensin-converting enzyme inhibitors or angiotensin receptor blockers | 231 (52%) |
| Diuretics | 208 (47%) |
| β blockers | 186 (42%) |
| Aspirin | 166 (37%) |
| Statins | 93 (21%) |
| Digoxin | 76 (17%) |
| Calcium antagonists | 72 (16%) |
| Nitrate | 68 (15%) |
Fifty-one patients (11%) subsequently underwent early coronary revascularization (coronary artery bypass grafting in 32 and percutaneous coronary intervention in 19). Gated SPECT parameters in relation to early coronary revascularization are listed in Table 2 . Of 393 patients without early coronary revascularization, cardiac death occurred in 64 (16%). Cumulative 5-year cardiac death–free survival in these 393 patients was 79.9%, while the rate of cardiac death was 4.3% per year.
| Variable | No Early Revascularization (n = 393) | Early Revascularization (n = 51) | p Value |
|---|---|---|---|
| Summed stress score | 6.9 ± 6.1 | 13.2 ± 7.0 | <0.001 |
| Summed stress score >8 | 127 (32%) | 38 (75%) | <0.001 |
| Summed difference score | 1.3 ± 2.6 | 5.1 ± 5.0 | <0.001 |
| Summed difference score >1 | 90 (23%) | 36 (71%) | <0.001 |
| Regional wall motion variance | 0.237 ± 0.579 | 0.512 ± 0.696 | 0.009 |
| Regional wall motion variance >0.114 | 131 (33%) | 29 (57%) | 0.001 |
| Summed stress score >8 or summed difference score >1 or regional wall motion variance >0.114 | 197 (50%) | 46 (90%) | <0.001 |
| Ejection fraction (%) | 30.6 ± 7.5 | 32.4 ± 6.1 | 0.062 |
Prescan and gated SPECT variables in relation to cardiac death are listed in Table 3 . Because the summed stress score was the gated SPECT parameter most related to cardiac death, its prognostic usefulness was examined. The ejection fraction on gated SPECT imaging was not significantly different between patients with summed stress scores ≤8 and those with scores >8 (30.8 ± 7.4% vs 30.4 ± 7.6%, respectively, p = 0.643). Cumulative 5-year cardiac death–free survival was 85.6% and 67.3% in patients with summed stress scores ≤8 and >8, respectively (p <0.001; Figure 1 ) . The annualized cardiac death rate in patients with summed stress scores ≤8 (2.9%) was significantly lower than in those with summed stress scores >8 (7.3%, unadjusted hazard ratio 2.62, 95% confidence interval 1.58 to 4.33, p <0.001).
| Variable | No Cardiac Death (n = 329) | Cardiac Death (n = 64) | p Value |
|---|---|---|---|
| Age (years) | 62 ± 13 | 66 ± 13 | 0.030 |
| Age ≥70 years | 104 (32%) | 31 (48%) | 0.009 |
| Men | 214 (65%) | 39 (61%) | 0.530 |
| Chest pain or dyspnea | 243 (74%) | 43 (67%) | 0.273 |
| Heart failure | 150 (46%) | 28 (44%) | 0.786 |
| Angina pectoris | 99 (30%) | 22 (34%) | 0.497 |
| Systemic hypertension | 258 (78%) | 42 (66%) | 0.028 |
| Tobacco smoker | 180 (55%) | 29 (45%) | 0.168 |
| Diabetes mellitus | 120 (37%) | 22 (34%) | 0.749 |
| Hypercholesterolemia | 104 (32%) | 10 (16%) | 0.010 |
| Pharmacologic stress | 237 (72%) | 59 (92%) | 0.001 |
| Summed stress score | 6.4 ± 5.7 | 9.2 ± 7.3 | 0.006 |
| Summed stress score >8 | 94 (29%) | 33 (52%) | <0.001 |
| Summed difference score | 1.2 ± 2.6 | 1.7 ± 3.0 | 0.232 |
| Summed difference score >1 | 73 (22%) | 17 (27%) | 0.446 |
| Regional wall motion variance | 0.212 ± 0.505 | 0.366 ± 0.859 | 0.171 |
| Regional wall motion variance >0.114 | 108 (33%) | 23 (36%) | 0.629 |
| Ejection fraction (%) | 30.8 ± 7.3 | 29.6 ± 8.4 | 0.232 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
