This study investigated the clinical performance of a model-based, patient-individualized closed-loop (CL) control system for propofol administration using the bispectral index (BIS) as a controlled variable during the induction and maintenance of anesthesia with propofol and remifentanil and studied the influence of the targeted effect-site concentration of remifentanil (Ce-REMI) on its clinical performance.
In 163 patients, propofol was administered using a CL system (BIS target [BISTARGET] between 40 and 50). Initial Ce-REMI targets between 2 and 7.5 ng/mL were selected as deemed clinically required. Performance parameters during induction were the time required to initially cross the target BIS, the time required to reach the maximal drug effect after induction (T-PEAK,T- BIS) and the corresponding BIS at this moment, and the time required to regain the target BIS at the end of induction. Performance during maintenance was defined as the percentage of case time with target BIS +/- 10 from target and the amount of performance error (PE) between the observed and target BIS values and its derived median PE (MDPE) as a measure of control bias, median absolute PE (MDAPE) as a measure of control inaccuracy, divergence as a measure of the time-related trend of the measured BIS values relative to the target BIS values, and wobble as a measure of intrasubject variability in prediction error. The secondary end point was the hemodynamic stability of the patient during CL control.
The applied CL system induced and maintained anesthesia within clinically accepted ranges. The percentage of case time [mean (standard deviation [SD]) across all study participants] with BIS +/- 10 from the target was 82% (14%). The mean (SD) population MDPE and MDAPE were -6.6% (5.5%) and 11.2% (5.5%), respectively. A negative divergence [-0.001 (0.004)] and acceptable wobble [9.7% (4.0%)] were found. The correlation between the system PE and Ce-REMI was low and only influenced by a Ce-REMI CONCLUSIONS:
The applied CL system for propofol administration has an acceptable performance in the Ce-REMI range of 2.8-7.5 ng/mL during the induction and maintenance of anesthesia. There was no evidence of a strong association between Ce-REM and the CL performance. This study also shows that when the Ce-REMI is