|| Checking for direct PDF access through Ovid
The incremental prognostic value of myocardial perfusion–gated single photon emission computed tomography (MPGS) compared with exercise test has not yet been properly evaluated.Five thousand six hundred seventy-two consecutive patients with known or suspected coronary disease undergoing exercise MPGS between 1997 and 2007 were included. Three-year predictive models for total death and death from cardiovascular causes or acute myocardial infarction (ie, major cardiovascular events [MCE]) were built using Cox-regression modeling, including only the clinical information. Then the exercise and MPGS information was sequentially added. The added discriminative ability of exercise test information and MPGS was assessed by net reclassification improvement and integrated discrimination improvement. The increase in predictive ability of exercise information for death and MCE was high as assessed by net reclassification improvement (0.199 and 0.263) and integrated discrimination improvement (0.042 and 0.021). The only variable of MPGS associated with total death was ejection fraction (hazard ratio, 0.84; 95% confidence interval, 0.79–0.89; P<0.001). Global stress ischemic score emerged as an additional variable associated with MCE (hazard ratio, 1.07; 95% confidence interval, 1.02–1.12; P=0.007). Adding MPGS information barely improved the prognostic value for total death (net reclassification improvement, 0.017; integrated discrimination improvement, 0.013), but it increased for MCE (net reclassification improvement, 0.122; integrated discrimination improvement, 0.033).Adding MPGS information to exercise information does not improve prediction of total death, although it allows a more accurate prediction of MCE.