Abstract
OBJECTIVES
The pathophysiology of delirium after cardiac surgery is complex. The present study aims to determine perioperative risk factors and construct a scoring system for postoperative delirium based on the type of surgery.
METHODS
Three hundred patients undergoing coronary artery bypass grafting (CABG; n = 150) or valve and/or aortic surgery ± CABG (n = 150) were retrospectively evaluated.
RESULTS
The incidence of delirium (32%) was similar in subgroups (28.7% and 33.33%, P = 0.384). Delirium patients were older [71.3 (standard deviation: 8.5) vs 66.6 (standard deviation: 9.5), P < 0.001; 70.0 (standard deviation: 9.6) vs 62.5 (standard deviation: 12.6), P < 0.001] and required more packed red blood cell units [2.1 (standard deviation: 2.1) vs 4.2 (standard deviation: 4.0), P < 0.001; 2.4 (standard deviation: 3.3) vs 5.4 (standard deviation: 5.9), P < 0.001] and fresh frozen plasma units [6.1 (standard deviation: 2.9) vs. 8.0 (standard deviation: 4.2), P < 0.001; 6.3 (standard deviation: 3.4) vs 10.8 (standard deviation: 7.2), P < 0.001] in CABG and valve/aortic subgroups, respectively. Delirium was associated with longer operation time [298.3 (standard deviation: 98.4) vs 250.6 (standard deviation: 67.8) min, P < 0.001], cardiopulmonary bypass (CPB) time [171.5 (standard deviation: 54.9) vs 140.98 (standard deviation: 45.8) min, P < 0.001] and cardiac arrest time [112 (standard deviation: 35.9) vs 91.9 (standard deviation: 28.6), P < 0.001] only in the valve/aortic group (versus non-delirium). Multivariate regression analysis identified an association between delirium and age [odds ratio: 1.056 (95% confidence interval: 1.002-1.113), P = 0.041], CPB time [odds ratio: 1.1014 (95% confidence interval: 1.004-1.025), P = 0.007], fresh frozen plasma transfusion [odds ratio: 1.127 (95% confidence interval: 1.006-1.262), P = 0.039] and atrial fibrillation [odds ratio: 4.801 (95% confidence interval: 1.844-12.502), P < 0.001] after valve/aortic surgery (area under the curve 0.835, P < 0.001) and between delirium and age [odds ratio: 1.089 (95% confidence interval: 1.023-1.160), P = 0.007] and ventilation time [odds ratio: 1.068 (95% confidence interval: 1.026-1.113), P = 0.001] after isolated CABG (area under the curve 0.798, P < 0.001). The cross-validation of the results by k-fold logistic regression revealed for the entire patient cohort an overall average accuracy of the prediction model of 0.764, with a false-positive rate of 0.052 and a false-negative rate of 0.18.
CONCLUSIONS
Age, CPB time, ventilation, transfusion and atrial fibrillation are differently associated with delirium depending on the operative characteristics. Optimization of intraoperative parameters and use of risk calculators may enable early institution of pharmacotherapy and improve overall outcome after cardiac surgery.
Collapse