Improvement of local anesthetics agents' simulation using Monte Carlo simulation considering correlation among parameters

抄録

Animal experiments using guinea pigs are performed to elucidate the effects of local anesthetic agents. Simulators are used in various fields as an animal alternative. Herein, we aimed to develop a simulator for local anesthetic agents. In a previous study, we developed a statistical model to simulate the effects of local anesthetic agents. We estimated the parameters of the distribution (mean [μ] and logarithm of standard deviation [logσ]) for each drug based on the results of animal experiments. We reported that Monte Carlo simulation yielded results consistent with those from the animal experiments. However, since this simulation did not account for parameter correlations, we observed a large variation in drug parameter values within individual subjects. This led to the order of drug duration being different from the original order in many individuals. In the present study, we investigated correlations among these parameter values, performed simulations using parameter values that followed a multivariate normal distribution, and examined the correlation of duration among drugs to address these shortcomings. Correlation coefficients between μ and logσ (r_μ-logσ) were -0.4 to 0.01. Correlation coefficients for μ (r_μ), and logσ (r_logσ) were 0.4 to 0.6, and 0.3 to 0.6, respectively. In Monte Carlo simulation, when r_μ-logσ was small, the standard deviation of duration within one drug was large. Moreover, when r_μ and r_logσ were large, the correlation of duration between two drugs was large. When correlation among parameters was not set, the correlation coefficient for duration was small (-0.12 to 0.26), but when these were set to parameter values obtained from animal experiments, the correlation coefficient for duration became large (0.22 to 0.47). These findings suggest that by considering the correlation among parameters, it is possible to create a simulator for local anesthetic agents that obtains results closer to those of animal experiments.