Correspondence analysis between remifentanil concentrations and monitored physiological measures in children.

Abstract

Background & Objective: There is scarce research done towards the development of target-controlled infusion models (TCI) for the opioid analgesic drug known as remifentanil in anaesthesia. TCI models reduce the incidence of adverse events and contribute towards personalized care of patients, but their development requires ample research around the relationships amongst the variables involved. A correspondence analysis between remifentanil plasma concentration and monitored physiological measures during induction to anaesthesia in a population of children aged 3 to 16 years old, could be beneficial towards deepening our understand-ing of the effects of remifentanil on physiological responses during anaesthesia. Methods: The dataset used has information of monitored measures for state (SE) and response (RE) spectral entropy of the electroencephalography (EEG) signal originating from brain activity, heart rate (HR), blood pressure (BP), and surgical stress index (SSI), of 60 patients aged 3-16 years during surgical procedures. Annotations in the University of Michigan Sedation Scale (UMSS) were included in the dataset, as well as annotations of when the remifentanil concentration was measured (R-points). A nonlinear measure of association known as prediction probability (Pk) was used to evaluate the correspondence between the remifentanil concentration and the averaged values of the monitored measures, and UMSS. Pk values have a range from 0 to 1, 1 for correspondence in same direction, 0.5 for no correspondence, and 0 for correspondence in opposite direction. Lastly, to evaluate the Pk results, a Wilcoxon Signed Rank test was used to determine statistical significance of the Pk results at the alpha of 5%, with the null hypothesis being that Pk values would be equal to 0.5. Bonferroni-Holm correction was applied to reduce the incidence of type I errors for such a large number of tests. Results: The Pk values from remifentanil concentration and UMSS scores passed the hypothesis test in all groups, indicating correspondence in the same direction. Evaluations for correspondence with RE and SE entropy failed to pass the hypothesis test across all age groups, the same case was seen with BP measures. HR and SSI passed the hypothesis test in the old-est age groups. Conclusions: The hypnotic component of remifentanil is evidenced in the correspondence with UMSS scores, however, the lack of covariation with entropy measures could be indicative that the effect of remifentanil over patient awareness must be studied with a different indicator. Age groups of 3-6 and 7-11 years could be affected by developmental factors, that made the behaviour for SSI and HR in relation with remifentanil different to the one expected for adults. There isn’t a measure associated only to analgesia that has a significant correspondence with remifentanil, future works could aim to develop a type of index measure that fulfils this role.