Complex Assessment of the Incidence and Risk Factors of Delirium in a Large Cohort of Cardiac Surgery Patients: A Single-Center 6-Year Experience

Abstract

BACKGROUND: Previous reports provided inconsistent data on the occurrence of postoperative delirium and emphasized its considerable impact on outcome. This study sought to evaluate the incidence and predictors of delirium, together with its relation to cerebral ischemia in a large cohort of cardiac surgery patients in a tertiary high-volume center. METHODS AND RESULTS: Consecutive patients (n = 8792) were prospectively enrolled from 2003 to 2008. Exclusion criteria were history of psychiatric disorders, use of psychoactive drugs, alcohol abuse, and data incompleteness. Finally, 5781 patients were analyzed in terms of 100 perioperative patient-specific and treatment variables. The incidence of postoperative delirium (DSM IV criteria) was 4.1% and it coexisted with cerebral ischemia in 1.1% of patients. In bivariate analysis, 49 variables were significantly linked to postoperative delirium. Multivariate analysis confirmed that delirium was independently associated with postoperative stroke (logistic odds ratio (logOR) = 2.862, P = 0.004), any blood transfusions (logOR = 4.178, P < 0.0001), age > 65 years (logOR = 2.417, P = 0.002), carotid artery stenosis (logOR = 2.15, P = 0.01), urgent/emergent surgery (logOR = 1.982, P = 0.02), fasting glucose level, intraoperative oxygen partial pressure fluctuations, and hematocrit. Area under ROC curve for the model was 0.8933. CONCLUSIONS: Early identification of nonpsychiatric perioperative determinants of delirium facilitates its diagnosis and might help develop preventive strategies to improve long-term outcome after cardiac surgery procedures.