Determinants of risk and its temporal variation in patients with normal stress myocardial perfusion scans: What is the warranty period of a normal scan?

OBJECTIVES: The aim of this study was to determine the predictors of risk and the temporal characteristics of risk associated with normal myocardial perfusion single photon emission computed tomography (MPS). BACKGROUND: No empiric data exist regarding predictors of risk after normal MPS and their temporal characteristics. METHODS: Follow-up (mean: 665 ± 200 days, 96% complete) of 7,376 consecutive patients with normal exercise or adenosine MPS identified 78 hard events (HE) (45 cardiac deaths, 33 non-fatal myocardial infarction; 1.1% cumulative HE rate, 0.6%/year). Cox proportional hazards analysis was used to identify predictors of HE. Parametric survival analysis was used to model predicted time to HE. RESULTS: The HE rates were greater in patients with versus without previous coronary artery disease (CAD). The Cox proportional hazards model identified pharmacologic stress, known CAD, diabetes mellitus (DM), male gender, and increasing age, with interactions between stress type and previous CAD (lower risk in patients without previous CAD undergoing exercise stress vs. all others) and between DM and gender (higher risk in DM females vs. all others) as the model most predictive of HE. The highest risk subgroups had a maximal event rate of 1.4% to 1.8%/year. Parametric survival models revealed that in patients without previous CAD the level of risk was uniform with time, but in patients with known CAD, risk increased with time (e.g., risk in the first year was less than in the second year, hence, a dynamic temporal component of risk was present). CONCLUSIONS: Multiple clinical factors add incremental prognostic value in patients with normal MPS, affecting their risk and its temporal pattern, and may alter the appropriate timing of repeat testing, hence establishing the existence of a "warranty" period for normal MPS studies.