P3401Comparison of APACHE II, SAPS II and GRACE 2.0 risk scores to estimate early and late mortality in patients with ST-segment elevation acute myocardial infarction

Abstract

Background Many specific prognostic risk scores have been validated for myocardial infarction (MI), such as the Global Registry of Acute Coronary Events (GRACE) 2.0. Other general risk scores are used in intensive care units (ICUs), such as the Acute Physiologic and Chronic Health Evaluation (APACHE) II and the Simplified Acute Physiology Score (SAPS) II. Purpose Compare the ability of GRACE 2.0, APACHE II and SAPS II risk scores to estimate in-hospital, 3- and 12-month mortality after primary angioplasty (PA) for ST-segment elevation MI (STEMI). Methods Retrospective cohort of 427 consecutive STEMI patients (64 years [55–75]; 78% men) admitted to a general ICU between November-2013 and February-2017. We used Area under the Receiver Operating Characteristic (ROC) curve (AUC) analysis to asses performance of risk scores, and the Hosmer-Lemeshow (HL) goodness of fit test and the Standardized Mortality Ratio (SMR) to assess calibration. Results All risk scores were associated with in-hospital, 3- and 12-month mortality (P<0.001). SAPS II had the highest sensitivity for short-term mortality and the highest AUC for in-hospital, 3- and 12-month mortality (Figure). SAPS II had the highest calibration and the less underestimation of mortality in all follow-up periods analysed (Table). Performance of prognostic risk scores Youden index (J) Sensitivity (%) Specificity (%) AUC (95% CI) H-L (P value) SMR In-Hospital mortality (5.4%) SAPS II 32 87.0 89.4 0.938 (0.887–0.988) 1.826 (0.969) 1.64 GRACE 2.0 150 86.9 83.7 0.922 (0.865–0.979) 8.111 (0.423) 1.77 APACHE II 17 82.6 93.8 0.896 (0.841–0.986) 11.941 (0.154) 1.92 3-month mortality (5.6%) SAPS II 32 79.2 89.1 0.913 (0.854–0.973) 3.635 (0.821) 1.85 GRACE 2.0 150 79.1 85.4 0.902 (0.842–0.962) 5.149 (0.742) 2.0 APACHE II 17 79.1 93.8 0.882 (0.792–0.971) 9.244 (0.322) 2.0 12-month mortality (7.0%) SAPS II 32 73.3 89.7 0.880 (0.809–0.951) 1.994 (0.960) 1.58 GRACE 2.0 150 76.7 84.1 0.878 (0.816–0.941) 4.073 (0.850) 2.31 APACHE II 17 70.0 92.4 0.824 (0.722–0.927) 7.464 (0.487) 2.14 AUC: Area under the curve; H-L: Hosmer-Lemeshow; SMR: Standardized Mortality Ratio. Area under the ROC curves (AUC) Conclusions SAPS II represents the best model to estimate mortality in this cohort of STEMI patients, with an appropriate calibration and less underestimation of mortality. Underestimation of mortality showed by all risk scores suggests the need of creating new risk prediction models that improve identification of high risk STEMI patients.