Agreement Between Two-Concentration and One-Concentration Area Under the Curve (AUC) Estimates When Using Bayesian Modeling to Dose Vancomycin in Patients With Obesity

Abstract

Background: Vancomycin Bayesian modeling provides 24-hour area under the curve (AUC24) estimations. However, the agreement between two-concentration and one-concentration Bayesian estimates in patients with obesity is unknown. Objective: The purpose of this study was to determine the agreement between two-concentration and one-concentration Bayesian AUC24 estimates in patients with obesity receiving vancomycin. Methods: This retrospective within-subjects cohort study included patients with obesity and two vancomycin concentrations. The first concentration was hidden from dosing software to record the one-concentration AUC24. AUC24 estimates were categorized into 1 of 3 groups: <400, 400 to 600, or >600 mg*h/L. Patients were excluded for vancomycin duration less than 48 hours or renal dysfunction. The primary outcome was AUC24 agreement with two versus one concentration. Secondary outcomes included the AUC24 category, matching of AUC24 categorization, and correlation between two-concentration versus one-concentration AUC24. AUC24 estimate agreement was assessed by Bland Altman plot and bias via linear regression. Statistical analyses were performed using SPSS (version 20.0). Results: A total of 31 patients were included. The mean difference in AUC24 between two versus one concentration was 11.4 mg*h/L (95% limits of agreement = −72 to 95 mg*h/L). Linear regression indicated the presence of proportional bias at higher AUC24 values (β = 0.16; P = 0.015). Matching of AUC24 categorization with two versus one concentration was 87% (27/31 patients). Conclusion and Relevance: This study demonstrated overall agreement between AUC24 estimates when using two versus one vancomycin concentration in patients with obesity, though proportional bias was detected at higher AUC24. Future studies with larger sample sizes are needed to confirm these results.