Choice reaction time in patients with post-operative cognitive dysfunction

Abstract

Background: Post-operative cognitive dysfunction (POCD) is detected by administration of a neuropsychological test battery. Reaction time testing is at present not included as a standard test. Choice reaction time (CRT) data from the first International Study of Post-operative Cognitive Dysfunction study were collected, but the association between POCD and reaction time has not been presented before. We hypothesized that CRT could be used as a screening tool for POCD.

Methods: Patients aged 60 years or older scheduled for major surgery with general anaesthesia were recruited from 13 centres in nine countries. CRT was measured 52 times using the four boxes test. Patients performed the test before surgery (n=1083), at 1 week (n=926) and at 3 months (n=852) post-operatively. CRT for the individual patient was determined as the median time of correct responses. The usefulness of the CRT as a screening tool for POCD was determined by the receiver-operator characteristic (ROC) curve.

Results: Patients with POCD 1 week after surgery had a significantly longer reaction time compared with patients without POCD: 857 (221) vs. 762 (201) ms, respectively (P<0.0001). Also at 3 months, patients with POCD had a significantly longer CRT. ROC curves revealed that a reaction time of 813 ms was the most appropriate cut-off at 1 week and 762 ms at 3 months but the positive predictive value for POCD was low: 34.4% and 14.7%, respectively.

Conclusions: Post-operative cognitive dysfunction is associated with impaired performance in the CRT test but the test is a poor predictor of POCD.