Variables Associated With Cardiac Surgical Waitlist Mortality From a Population-Based Cohort

Abstract

BackgroundCardiac surgery waitlist recommendations, which were developed based on expert opinion, poorly predict preoperative mortality. Studies reporting risk factors for waitlist mortality have not evaluated the risks including nonadherence to waitlist benchmarks. MethodsIn patients who underwent cardiac surgery or died on the waitlist between 2005 and 2015, we used a Fine and Gray competing risk model to identify independent predictors of waitlist mortality in 12,106 patients scheduled for urgent, semiurgent, or nonurgent surgery. The predictive variables were compared with Canadian Cardiovascular Society (CCS) waitlist recommendations using the Akaike information criterion. ResultsA total of 101 (0.8%) patients died awaiting surgery. The median wait times and frequency waitlist deaths among emergent, urgent, semi-urgent, and nonurgent surgery were 0.6, 7.4, 69.0, 55.5 days (P < 0.001) and 6.3%, 0.8%, 0.3%, 0.6% (P < 0.001), respectively. Adherence to CCS waitlist recommendations was higher in patients who died on the waitlist (51.6% vs 70.8%, P = 0.001) and was not predictive of waitlist mortality (hazard ratio 1.48, 95% confidence interval 0.62-0.56). Independent predictors of waitlist mortality were age, aortic surgery, ejection fraction < 35%, urgent surgery, prior myocardial infarction, haemodynamic instability during cardiac catheterization, hypertension, and dyslipidemia. These variables were superior to current CCS guidelines (Akaike information criterion 1251 vs 1317, likelihood ratio test P < 0.001). ConclusionsCCS waitlist recommendations were poorly predictive of waitlist mortality and the majority of waitlist deaths occur within recommended benchmarks. We identified variables associated with waitlist mortality with improved clinical performance. Our findings suggest a need to re-evaluate cardiac surgical triage criteria using evidence-based data.