Patients with LV dysfunction are known to have a high risk for future coronary events. This study aimed to determine the incremental prognostic value of perfusion defects in patients with scintigraphic evidence of LV dysfunction. Patients showing either transient ischemic dilatation or Tl-201 lung:heart ratio of >0.5 on exercise or dipyridamole SPECT Tl- 201 myocardial perfusion scan were included in the study. Perfusion defects were scored semiquantitatively using a 17 -segment, 5-point scale (0 = normal, 4 = absent uptake). The extent and severity of defects were quantified using number of abnormal segments (NAS) and the summed stress score (SSS). Defect reversibility was quantified using the summed difference score (SDS) between stress and rest defects. Patients were followed up for the development of coronary events over a period of 12 - 30 months. There were 6 cardiac deaths and 7 myocardial infarctions in the 65 patients included in the study (20% overall event rate). ROC curve analysis of the scores revealed the following suggested cut-off values for predicting cardiac event: NAS ~ 7 (85% sensitivity, 56% specificity), SSS ~ 19 (77% sensitivity, 67% specificity) and SDS ~ 7 (31 % sensitivity, 60% specificity). ,Event rate was significantly higher above the cut-off value in SSS «19 = 9%, >19 = 32%,p = 0.04). With NAS, the event rate was higher above the cut-off value, with the difference approaching significance (NAS 7 = 26%, P = 0.06). There was no statistically significant difference in the event rate with high or low SDS (SDS 7 = 16%, P = 0.32). The odds ratios for NAS and SSS (5.8,4.9, respectively) were higher compared with SDS (0.56). The extent and severity of stress perfusion defects provided incremental prognostic information in patients with LV dysfunction. Our data suggest that the degree of defect reversibility showed no prognostic value in this subset of patients.