Establishing a Population Pharmacokinetic Model for Vancomycin Dosing Optimisation in Critically Ill Nigerian Patients Using Bayesian Adaptive Methods

Background: Vancomycin is a cornerstone agent for methicillin-resistant Staphylococcus aureus infections in ICU settings. Achieving target AUC/MIC exposures is critical for efficacy and nephrotoxicity avoidance. Nigerian critically ill patients may exhibit pharmacokinetic profiles distinct from Western populations due to differences in body composition, renal function distribution, and comorbidity burden. A population pharmacokinetic (popPK) model derived from local patients is needed to guide dosing optimisation. Methods: A prospective opportunistic sampling study was conducted in the ICUs of teaching hospitals in Lagos, Kano, and Enugu. Trough and random vancomycin concentrations were collected from 120 critically ill adult patients. Non-linear mixed effects modelling (NONMEM software) was used to develop a two-compartment popPK model. Bayesian adaptive dosing recommendations were simulated to achieve target AUC24/MIC of 400 to 600 mg.h/L. Covariates including creatinine clearance, body weight, and albumin were evaluated. Results: A two-compartment model adequately described vancomycin pharmacokinetics. Creatinine clearance and albumin were significant covariates on clearance. Nigerian ICU patients showed higher apparent volume of distribution compared to published Caucasian models. Bayesian-guided dosing improved target attainment from 38 to 71 percent in simulation. Conclusion: This first Nigerian popPK model for vancomycin ICU dosing identifies local pharmacokinetic characteristics necessitating adapted dosing strategies. Bayesian adaptive dosing offers superior target attainment over standard regimens. Keywords: vancomycin, population pharmacokinetics, ICU, Bayesian dosing, Nigeria