Abstract 727: Long-term Validation of a Prognostic Adenosine Stress Score for SPECT Myocardial Perfusion Imaging
Background: Clinicians face the challenge of incorporating multiple types of data when utilizing the prognostic implications of pharmacologic stress SPECT myocardial perfusion imaging (MPI) to guide management decisions. We previously derived and validated a prognostic adenosine stress score (PASS) for prediction of cardiac death that incorporated data available after adenosine MPI and sought to further validate this score in a large cohort with long-term follow-up (FU).
Methods: We identified 4295 patients (pts) [40% prior coronary artery disease (CAD), 51% male, age 71±11] who underwent adenosine stress MPI between 1991 and 1997 and had FU [7.2±3.7 years (yr), lost to FU 2.8%]. The study endpoint was all-cause death (ACD), based on the Social Security Death Index. PASS was: (age [years]) + (% myocardium ischemic) + (%myocardium fixed) +10 (if dyspnea was a presenting symptom) + (resting heart rate × 0.5) − (peak heart rate × 0.5) + 30 (if rest ECG is abnormal). Cox proportional hazards modeling (CPH) was utilized to assess the performance of PASS.
Results: FU identified 2061 ACD (48%, 6.7%/yr). ACD rates as a function of quartiles of PASS were 34.5% (4.3%/yr), 43.9% (5.8%/yr), 52.1% (7.5%/yr), and 62% (10.4%/yr). Risk of ACD for pts with normal or abnormal MPI was successfully stratified by quartiles of PASS (both p<0.001). Univariable CPH modeling of PASS (chi-square 208, p<0.001, inter-quartile HR: 1.52) was used to estimate predicted survival in pts with PASS values of 50, 75, 100, and 150 at long-term FU (Figure⇓).
Conclusion: PASS, derived previously for relatively short-term cardiac death, can successfully risk stratify patients with respect to long-term risk of ACD.