Long-term survival and resource use in critically ill cardiac surgery patients: a population-based study

Purpose Most cardiac surgery patients recover well; a substantial minority become critically ill after surgery. The epidemiology of critical illness after cardiac surgery is poorly described. We measured the association of prolonged critical illness with long-term survival and resource use after cardiac surgery. Methods This was a historical population-based cohort study in Ontario, Canada (2002-2013), of adult cardiac surgery patients. Validated methods were used to measure postoperative intensive care unit (ICU) length of stay (LOS). We categorized patients into short (0-2 day), moderate (3-9 day), and long (10+ day) ICU LOS groups. The adjusted associations of ICU LOS with one-year survival (primary outcome) and costs, hospital readmissions, and institutional discharge were measured using multilevel, multivariable regression. Pre-specified sensitivity analyses were performed. Results We included 111,740 patients having their first cardiac surgery during the study period who survived >= ten postoperative days. Most patients had a short ICU LOS (75.9%); 20.9% and 3.3% had moderate or long ICU LOS, respectively. The short-stay one-year mortality rate was 2.1%. Longer ICU LOS was independently associated with decreased one-year survival (moderate LOS: hazard ratio [HR], 1.79; 95% confidence interval [CI], 1.6 to 1.94; long LOS: HR, 8.66; 95% CI, 7.93 to 9.44). Sensitivity analyses supported the findings of the primary analysis. Secondary outcomes were independently associated with longer ICU LOS. Long ICU LOS patients occupied 30% of all ICU bed days, and 55% died or were discharged to an institution. Conclusion Prolonged ICU LOS after cardiac surgery is associated with decreased 1-year survival and increased healthcare resource use.