Objectives: Postoperative Delirium (POD) frequently occurs after coronary artery bypass graft (CABG). Patients screened with an appropriate POD predictive rule (PR) may benefit from targeted preventive treatment. The accuracy of current available PRs, however, remained modest with the C statistic below 0.75, partially because none of which include intra- and post-operative (IPOP) as well as preoperative (preop) risk factors at the same time. Since POD is a postoperative (postop) phenomenon, IPOP factors might play important roles. This study aimed to develop and validate a PR based on perioperative (periop) factors to predict POD after CABG. Methods: This prospective study included two cohorts (derivation [n = 132] and validation [n = 53]) of patients who underwent first isolated CABG. POD was assessed using the Confusion Assessment Method for Intensive Care Unit twice daily until 5 days after operation. Logistic regression and a bootstrap resampling procedure were used for model development and internal validation. Rule performance was measured with the C statistic. Points were assigned to each variable and a simplified rule was derivated to identify subjects at low, moderate, and high risk of POD based on their total risk score. The proportion of subjects with POD in each risk group was calculated. Results: The PR included 5 periop factors: age (≥68y), female, intraoperative intra-aortic balloon pump use, postop atrial fibrillation and serum IL-6 ≥ 478 pg/ml at 18th hour. The relative risk of POD in moderate (2~5 points) and high risk group (6~8 points) were 6.7 and 10.9 folds higher compared with low risk group (0~1 point), respectively (Fig A). The model C statistic was 0.861 (Fig B) in the derivation cohort, 0.829 in internal validation, and 0.754 in validation cohort. Conclusion: We developed a POD PR for patients undergoing CABG, which includes IPOP risk factors as well as preop risk factors. The new developed rule demonstrated more powerful in predicting POD after CABG.
Read full abstract