Postoperative Complications and Long-Term Survival
Patients in the high-risk cohort had a higher incidence of prolonged mechanical ventilation (12.58% vs 8.24%, P=0.013), longer median intensive care unit times (47.0 hours [IQR 26.0 to 81.6] vs 43.0 hours [IQR 24.6 to 69.0], P=0.004), and longer hospital length of stay (10 days [IQR 8-13] vs 9 days [IQR 8-12], P=0.013). Rates of operative mortality, reoperation, and need for transfusion were similar between cohorts. Additionally, rates of renal failure, pneumonia, stroke, and sepsis were also similar (Table 2 ).
Thirty-day mortality was similar between high- and low-risk cohorts. Mortality at one (8.94% vs 6.92%, P=0.193) and five years (17.72% vs 14.66%, P=0.149) was also similar. Kaplan Meier survival between cohorts is shown in Figure 1 .
Cox Proportional Hazards modeling was conducted to identify risk-adjusted predictors for 5-year mortality. In this model, preoperative troponin peak above 1.95 ng/mL was not associated with increased hazards for mortality (HR 1.28, 95% CI 0.94 to 1.72, P=0.113). Similar findings were obtained when preoperative troponin level was modeled as a continuous variable (HR 1.00, 95% CI 0.99 to 1.01, P=0.674) (Supplemental Table 1). Risk-adjusted factors associated with increased hazards for mortality included increasing age, history of diabetes mellitus, chronic obstructive pulmonary disease, immunosuppression, cerebrovascular disease, bridge with intravenous inotropes, and increasing preoperative serum creatinine. Factors associated with decreased hazards for mortality included increasing body surface area and increasing preoperative left ventricular ejection fraction (Table 3 ).