1326. Vancomycin Area Under the Concentration-Time Curve (AUC) Estimation Using a Bayesian Approach Versus First-Order Pharmacokinetic Equations: A Pilot Study. (31st December 2020)

• Record Type: Journal Article
• Authors: Pearson, Jeffrey; Alsowaida, Yazed S; Pharm, B S; Kubiak, David W; Kovacevic, Mary P; Dionne, Brandon
• Is Part Of: Open forum infectious diseases. Volume 7:Number 1(2020) Supplement
• Page Start: S673
• Page End: S673
• Publication Date: 2020-12-31
• Subjects: Communicable diseases -- Periodicals; Medical microbiology -- Periodicals; Infection -- Periodicals; 616.9
• DOI: 10.1093/ofid/ofaa439.1508
• Languages: English
• ISSNs: 2328-8957

Abstract

Abstract: Background: Current guidelines endorse area under the concentration-time curve (AUC)-based monitoring over trough-only monitoring for systemic vancomycin. Vancomycin AUC can be estimated using either Bayesian modeling software or first-order pharmacokinetic (PK) calculations. The objective of this pilot study was to evaluate and compare the efficiency and feasibility of these two approaches for calculating the estimated vancomycin AUC. Methods: A single-center crossover study was conducted in four medical/surgical units at Brigham and Women's Hospital over a 3-month time period. All adult patients who received vancomycin were included. Patients were excluded if they were receiving vancomycin for surgical prophylaxis, were on hemodialysis, if vancomycin was being dosed by level, or if vancomycin levels were never drawn. The primary endpoint was the amount of time study team members spent calculating the estimated AUC and determining regimen adjustments with Bayesian modeling compared to first-order PK calculations. Secondary endpoints included the number of vancomycin levels drawn and the percent of those drawn that were usable for AUC calculations. Results: One hundred twenty-four patients received vancomycin during the study, of whom 47 met inclusion criteria. The most likely reasons for exclusion were receiving vancomycin for surgical prophylaxis (n=40) or never having vancomycin levels drawn (n=32). The median time taken to assess levels in the Bayesian arm was 9.3 minutes [interquartile range (IQR) 7.8-12.4] versus 6.8 minutes (IQR 4.8-8.0) in the 2-level PK arm (p=0.004). However, if Bayesian software is integrated into the electronic health record (EHR), the median time to assess levels was 3.8 minutes (IQR 2.3-6.8, p=0.019). In the Bayesian arm, 30 of 34 vancomycin levels (88.2%) were usable for AUC calculations, compared to 28 of 58 (48.3%) in the 2-level PK arm. Conclusion: With EHR integration, the use of Bayesian software to calculate the AUC was more efficient than first-order PK calculations. Additionally, vancomycin levels were more likely to be usable in the Bayesian arm, thereby avoiding delays in estimating the vancomycin AUC. Disclosures: All Authors : No reported disclosures