Search

We have developed an automatic anaesthesia system for closed-loop administration of anaesthesia drugs. The control variables used were bispectral index (BIS) and Analgoscore for hypnosis and antinociception, respectively.One hundred and eighty-six patients were randomly enrolled in two groups. Propofol, remifentanil, and rocuronium were administered using closed-loop feedback control (closed-loop, n = 93) or manually (control group, n = 93). The clinical performance of hypnosis control was determined by calculating the offset from a BIS of 45: 'excellent', 'good', 'poor', and 'inadequate' control was defined as BIS values within 10%, from 11% to 20%, from 21% to 30%, or30% offset from the target. The clinical performance of analgesia was defined as the offset from Analgoscore values. Data presented as mean (standard deviation) (95% confidence interval).Excellent or good control of hypnosis was achieved significantly longer in the closed-loop group [47.0 (9.8%) (45.0/49.0), 34.4 (4.7%) (33.5/35.4)] than in the control group [37.3 (14.3%) (34.4/40.2) and 32.3 (7.6%) (30.7/33.7)], respectively (P0.0001 and 0.0085). Poor and inadequate control of hypnosis was significantly shorter in the closed-loop group [10.8 (5.0%) (9.8/11.8) and 7.7 (6.2%) (6.4/9.0)] than in the control group [14.7 (6.8%) (13.3/16.0) and 15.8 (14.7%) (12.8/18.8)], respectively (P0.0001). Excellent control of analgesia was achieved significantly longer in the closed-loop group [78.7 (16.2%) (75.4/82.0)] than in the control group [73.7 (17.8%) (70.1/77.3)] (P=0.0456).The closed-loop system was better at maintaining BIS and Analgoscore than manual administration.