A model for predicting mortality in acute ST-segment elevation myocardial infarction treated with primary percutaneous coronary intervention: results from the Assessment of Pexelizumab in Acute Myocardial Infarction Trial.

BACKGROUND: Accurate models to predict mortality are needed for risk stratification in patients with ST-segment elevation myocardial infarction (STEMI) undergoing primary percutaneous coronary intervention (PCI). METHODS AND RESULTS: We examined 5745 patients with STEMI undergoing primary PCI in the Assessment of Pexelizumab in Acute Myocardial Infarction Trial within 6 hours of symptom onset. A Cox proportional hazards model incorporating regression splines to accommodate nonlinearity in the log hazard ratio (HR) scale was used to determine baseline independent predictors of 90-day mortality. At 90 days, 271 (4.7%) of 5745 patients died. Independent correlates of 90-day mortality were (in descending order of statistical significance) age (HR, 2.03/10-y increments; 95% CI, 1.80 to 2.29), systolic blood pressure (HR, 0.86/10-mm Hg increments; 95% CI, 0.82 to 0.90), Killip class (class 3 or 4 versus 1 or 2) (HR, 4.24; 95% CI, 2.97 to 6.08), heart rate (>70 beats per minute) (HR, 1.45/10-beat increments; 95% CI, 1.31 to 1.59), creatinine (HR, 1.23/10-μmol/L increments >90 μmol/L; 95% CI, 1.13 to 1.34), sum of ST-segment deviations (HR, 1.25/10-mm increments; 95% CI, 1.11 to 1.40), and anterior STEMI location (HR, 1.47; 95% CI, 1.12 to 1.93) (c-index, 0.82). Internal validation with bootstrapping confirmed minimal overoptimism (c-index, 0.81). CONCLUSIONS: Our study provides a practical method to assess intermediate-term prognosis of patients with STEMI undergoing primary PCI, using baseline clinical and ECG variables. This model identifies key factors affecting prognosis and enables quantitative risk stratification that may be helpful in guiding clinical care and for risk adjustment for observational analyses.

Duke Scholars

Author Rajendra H. Mehta Medicine, Cardiology

Author Paul Wayne Armstrong Medicine, Cardiology

Author Kerry L. Lee Biostatistics & Bioinformatics, Division of Biostatistics

Author Christopher Bull Granger Medicine, Cardiology