Abstract
INTRODUCTION: This study compared the calculated vancomycin area under the curve (AUC(0-24)) using trapezoidal and non-trapezoidal first-order pharmacokinetic equations. METHODS: This retrospective observational study included adult patients with documented MRSA bacteremia who received ≥ 48 h of intravenous vancomycin and had two consecutive serum levels after the first dose. AUC(0-24) was calculated using trapezoidal and non-trapezoidal equations. Correlation and agreement between methods were assessed using Pearson's correlation coefficient (r) and Bland-Altman plots. Significant predictors (p < 0.05) from simple linear regression were included in a multiple linear regression model to evaluate their impact on AUC(0-24) for both methods. RESULTS: Fifty-two patients were included. The median age was 63 years (interquartile range [IQR]: 50-73), and the median vancomycin clearance was 4 l/h (IQR: 2-6). Median vancomycin AUC(0-24) was 399 mg∙h/l (IQR: 257-674) for the trapezoidal method and 572 mg∙h/l (IQR: 466-807) for the non-trapezoidal method. There was a strong correlation between the methods (r = 0.87 [95% CI, 0.79-1]; P < 0.01), but Bland-Altman analysis showed poor agreement, with a bias of - 198 mg∙h/l and 95% limits of agreement from - 482 to 86 mg∙h/l. In multiple linear regression, total daily dose and vancomycin clearance were independent predictors of AUC(0-24) for both methods, with a stronger impact on non-trapezoidal AUC(0-24) (adjusted R(2) = 0.70) than trapezoidal AUC(0-24) (adjusted R(2) = 0.59). CONCLUSIONS: Trapezoidal and non-trapezoidal equations are not interchangeable for estimating vancomycin AUC(0-24). The trapezoidal method consistently results in lower AUC(0-24) estimates than the non-trapezoidal method.