Predictive performance of Bayesian and nonlinear least-squares regression programs for lidocaine

Area of Science:

• Pharmacokinetics and Pharmacodynamics
• Computational Biology
• Clinical Pharmacology

Background:

• Accurate prediction of drug serum concentrations is crucial for effective patient treatment.
• Lidocaine is commonly used for managing cardiac arrhythmias.
• Evaluating computational models for drug dosing optimizes therapeutic outcomes.

Purpose of the Study:

• To compare the predictive performance of a two-compartment Bayesian model against nonlinear least-squares regression for lidocaine dosing.
• To assess the accuracy of these models in forecasting serum lidocaine concentrations in acute and chronic arrhythmia patients.

Main Methods:

• Two computer programs, a two-compartment Bayesian model and nonlinear least-squares regression, were utilized.
• Lidocaine was administered to two patient groups: 15 acute arrhythmia and 14 chronic arrhythmia patients.
• Serum lidocaine concentrations were measured at various time points and used to forecast future levels.

Main Results:

• The Bayesian model predicted a significant difference in lidocaine concentrations between acute and chronic arrhythmia groups at 12 hours.
• The Bayesian method demonstrated significantly higher precision compared to nonlinear least-squares regression at 8, 12, and 48 hours for the acute group.
• Mean error and mean-squared error were used to assess predictive performance.

Conclusions:

• The two-compartment Bayesian model shows superior accuracy in predicting future lidocaine serum concentrations compared to nonlinear least-squares regression.
• These findings suggest the Bayesian approach may offer a more reliable tool for optimizing lidocaine therapy in arrhythmia management.
• Further research is warranted to validate these predictive capabilities across diverse patient populations.