The association between postoperative cognitive dysfunction and cerebral oximetry during cardiac surgery: a secondary analysis of a rendomised trial

Background: Postoperative cognitive dysfunction (POCD) occurs commonly after cardiac surgery. Near-infrared spectroscopy (NIRS) has been used to monitor regional cerebral oxygen saturation (rScO(2)) in order to minimise the occurrence of POCD by applying dedicated interventions when rScO(2) decreases. However, the association between rScO(2) intraoperatively and POCD has not been clarified. Methods: This is a secondary analysis of a randomised trial with physician-blinded NIRS monitoring and cognitive testing at discharge from hospital and at 3 months after surgery. The association between intraoperative rScO(2) values and POCD at discharge from hospital and at 3 months after surgery was investigated. The prespecified candidate predictive variable of interest was cumulative time during surgery with rScO(2) >= 10% below its preoperative value. Results: One hundred and fifty-three patients had complete NIRS data and neurocognitive assessments at discharge, and 44 of these patients (29%) had POCD. At 3 months, 148 patients had complete data, and 12 (8%) of these patients had POCD. The median time with rScO(2) >10% below preoperative values did not differ for patients with and without POCD at discharge (difference=0.0 min; Hodges-Lehmann 95% confidence interval, -3.11-1.47, P=0.88). Other rScO(2) time thresholds that were assessed were also not significantly different between those with and without POCD at discharge. This applied both to absolute rScO(2) values and relative changes from preoperative values. Similar results were found in relation to POCD at 3 months. Conclusions: No significant association was found between intraoperative rScO(2) values and POCD. These findings bring into question the rationale for attempting to avoid decreases in rScO(2) if the goal is to prevent POCD.