How can postoperative delirium be predicted in advance? A secondary analysis comparing three methods of early assessment in elderly patients.

BACKGROUND: Postoperative impairment of the cerebral function can appear immediately after general anesthesia and may be predictive for a postoperative delirium. We compared three tools assessing patients on recovery room admission in order to detect early signs of postoperative brain dysfunction: the Postanesthetic Recovery Score (PARS), the Richmond Agitation-Sedation Scale (RASS) and the Nursing Delirium Screening Scale (Nu-DESC).

METHODS: Inclusion criteria of this secondary analysis of the randomized SuDoCo trial were: age ≥60 years, schedule for elective non-cardiac surgery with an anticipated duration of ≥60 minutes, general anesthesia, ability to communicate in German language. A total of 996 patients were analyzed. Investigated scores were assessed 10 minutes after recovery room admission and analyzed in terms of association with postoperative delirium, mortality and length of stay in the recovery room. Multivariate analysis: linear or logistic regression.

RESULTS: Abnormal scores in patients: RASS (values ≤-2 and ≥1) 36.8%, Nu-DESC (≥2 points) 54.2%, PARS (≤7 points) 20.3%. Abnormal RASS values were associated with postoperative delirium (RASS values ≥1: odds ratio (OR) 3.1, 95% CI: 1.7-5.6, P<0.001; RASS values ≤-2: OR=2.1, 95% CI: 1.3-3.3, P=0.001). Abnormal Nu-DESC values were also associated with postoperative delirium (OR=2.4, 95% CI: 1.5-3.9, P<0.001). Abnormal PARS values were associated with a longer recovery room stay (in minutes, OR=16.6, 95% CI: 1.7-31.4, P=0.029).

CONCLUSIONS: The RASS and Nu-DESC were independent predictors for a delirium within seven postoperative days. Very early assessment of the cerebral function may help to advance detection, prevention and treatment of postoperative delirium in elderly patients.