Objectives:Clinical observation, as well as randomized controlled trials, indicated an increasing rate of postoperative cognitive dysfunction (POCD) with increasing depth of general anesthesia. However, the findings are subject to bias due to varying degree of analgesia. In this trial, we compared the rate of POCD between patients receiving light versus high anesthesia while holding analgesia comparable using nerve block. Methods:Elderly patients (≧60 years) receiving elective total knee replacement were randomized to receive the surgery under general anesthesia at BIS 40-50 (LOBIS group) or BIS 55-65 (HIBIS group). The femoral nerve and the sciatic nerve were blocked under ultrasonic guidance in all patients before induction. Cognitive performance was assessed with Montreal cognitive assessment (MoCA) at the baseline and 1d, 3d, and 7d after the surgery. POCD was defined by Z score of >1.96 using cross-reference. The extubation time and recovery time were also compared. Results:A total of 66 patients were randomized; 60 (n = 30 per group) completed trial as the protocol specified. POCD occurred in six patients (20%) in the LOBIS group vs. in one patient (3.3%) in the HIBIS group (Figure 3, p = .04). In all seven cases, the diagnosis of POCD was based on MoCA assessment on 1d after the surgery. Assessment in 3d and 7d after surgery did not reveal POCD in any case. Extubation time was longer in the LOBIS group (12.16 ± 2.58 vs. 5.77 ± 3.01 min in the HIBIS group (p < .001)). The time of comeback of directional ability was 13.47 ± 3.14 and 6.17 ± 3.23 min in the LOBIS and HIBIS groups, respectively (p < .001). Conclusions:In elderly patients receiving a total knee replacement, lighter anesthesia could reduce the rate of POCD with complete analgesia during surgery.