High Density Near-Infrared Cerebral Monitoring in Cardiac Surgery: Prediction of Perioperative Cognitive Decline.

Abstract

Approximately 500,000 patients undergo cardiac surgery every year. Of those, 50-80% have a cognitive decline at time of discharge, which negatively impacts quality of life. Near infrared spectroscopy (NIRS) can measure cerebral oxygenation, allowing physicians to monitor patients intra-operatively. However results from current NIRS monitors poorly predict cognitive decline after decreases in cerebral saturation. We hypothesized that a larger and denser NIRS array will provide more accurate monitoring of patient cerebral oxygen saturation. We performed validation studies on two Bonnet Macaque monkeys to confirm our array’s temporal and spatial accuracy. Next, we recruited a small group of patients to monitor during cardiac surgery to characterize cerebral vascular heterogeneity. Finally, using an array of 104 data channels, we monitored an additional 17 patients intra-operatively, and performed pre- and post-operative cognitive testing to identify those with cognitive decline. The animal studies showed that our large array was sensitive to large and focal changes in the cerebral vasculature of the monkey. Data from the clinical pilot revealed large heterogeneity in the vascular response to severe, sudden drops in blood pressure. Additionally, data show that methods used to remove contamination introduced from tissues above the cortex do not perform as expected. Data from the 17 patients monitored with the large array and received neurocognitive testing demonstrated that biomarkers other than decreases in saturation are better predictors of cognitive decline. In particular, a surrogate maker for cerebral autoregulation was statistically larger (p<0.01) in patients with cognitive decline, with 100% sensitivity. Markers of cerebral autoregulation should be further investigated in larger studies as method of predicting cognitive decline.