Automated responsiveness monitor to titrate propofol sedation.

Background: In previous studies, we showed that failure to respond to automated responsiveness monitor (ARM) precedes potentially serious sedation-related adversities associated with loss of responsiveness, and that the ARM was not susceptible to false-positive responses. It remains unknown, however, whether loss and return of response to the ARM occur at similar sedation levels. We hypothesized that loss and return of response to the ARM occur at similar sedation levels in individual subjects, independent of the propofol effect titration scheme.

Methods: Twenty-one healthy volunteers aged 20-45 yr underwent propofol sedation using an effect-site target-controlled infusion system and two different dosing protocol schemes. In all, we increased propofol effect-site concentration (Ce) until loss of response to the ARM occurred. Subsequently, the propofol Ce was decreased either by a fixed percentage (20%, 30%, 40%, 50%, 60%, and 70%; fixed percentage protocol, n = 10) or by a linear deramping (0.1, 0.2, and 0.3 microg x mL(-1) x min(-1); deramping protocol, n = 11) until the ARM response returned. Consequently, the propofol Ce was maintained at the new target for a 6-min interval (Ce plateau) during which arterial samples for propofol determination were obtained, and a clinical assessment of sedation (Observer's Assessment of Alertness/Sedation [OAA/S] score) performed. Each participant in the two protocols experienced each percentage or deramping rate of Ce decrease in random order. The assumption of steady state was tested by plotting the limits of agreement between the starting and ending plasma concentration (Cp) at each Ce plateau. The probability of response to the ARM as a function of propofol Ce, Bispectral Index (BIS) of the electroencephalogram, and OAA/S score was estimated, whereas the effect of the protocol type on these estimates was evaluated using the nested model approach (NONMEM). The combined effect of propofol Ce and BIS on the probability for ARM response was also evaluated using a fractional probability model (P(BIS/Ce)).

Results: The measured propofol Cp at the beginning and the end of the Ce plateau was almost identical. The Ce50 of propofol for responding to the ARM was 1.73 (95% confidence interval: 1.55-2.10) microg/mL, whereas the corresponding BIS50 was 75 (71.3-77). The OAA/S50 probability for ARM response was 12.5/20 (12-13.4). A fractional probability (P(BIS/Ce)) model for the combined effect of BIS and Ce fitted the data best, with an estimated contribution for BIS of 63%. Loss and return of ARM response occurred at similar sedation levels in individual subjects.

Conclusions: Reproducible ARM dynamics in individual subjects compares favorably with clinical and electroencephalogram sedation end points and suggests that the ARM could be used as an independent instrumental guide of drug effect during propofol-only sedation.