GRACE risk score: sex-based validity of in-hospital mortality prediction in Canadian patients with acute coronary syndrome

Background: Although there are sex differences in management and outcome of acute coronary syndromes (ACS), sex is not a component of Global Registry of Acute Coronary Events (GRACE) risk score (RS) for in-hospital mortality prediction. We sought to determine the prognostic utility of GRACE RS in men and women, and whether its predictive accuracy would be augmented through sex-based modification of its components. Methods: Canadian men and women enrolled in GRACE and Canadian Registry of Acute Coronary Events were stratified as ST-segment elevation myocardial infarction (STEMI) or non-ST-segment elevation ACS (NSTE-ACS). GRACE RS was calculated as per original model. Discrimination and calibration were evaluated using the c-statistic and Hosmer-Lemeshow goodness-of-fit test, respectively. Multivariable logistic regression was undertaken to assess potential interactions of sex with GRACE RS components. Results: For the overall cohort (n = 14,422), unadjusted in-hospital mortality rate was higher in women than men (4.5% vs. 3.0%, p < 0.001). Overall, GRACE RS c-statistic and goodness-of-fit test p-value were 0.85 (95%CI 0.83–0.87) and 0.11, respectively. While the RS had excellent discrimination for all subgroups (c-statistics > 0.80), discrimination was lower for women compared to men with STEMI [0.80 (0.75–0.84) vs. 0.86 (0.82–0.89), respectively, p < 0.05]. The goodness-of-fit test showed good calibration for women (p = 0.86), but suboptimal for men (p = 0.031). No significant interaction was evident between sex and RS components (all p > 0.25). Conclusions: The GRACE RS is a valid predictor of in-hospital mortality for both men and women with ACS. The lack of interaction between sex and RS components suggests that sex-based modification is not required.