1090. Does calculation method matter for targeting vancomycin AUC?

Abstract Background Recent vancomycin (VAN) guidelines recommend targeting an area under the curve (AUC) concentration of 400-600 for treatment of methicillin resistant Staphylococcus aureus infections. Multiple strategies for calculating AUC exist, including first order pharmacokinetic (foPK) equations and Bayesian models. Most clinical applications of foPK assume unchanged patient status and project ideal administration times to estimate exposure. Bayesian modeling provides the best estimate of true drug exposure and can incorporate changing patient covariates and exact doses. We compared two commonly used foPK methods to Bayesian estimates of VAN AUC. Graphs depict calculated AUCs using the three different methods: 1) Population PK estimated (foPOPPK) 2) Two-level first dose estimated (foFDPK) 3) Bayesian estimated. Methods First order equations were performed using population PK estimates (foPOPPK) to estimate steady state (SS) AUC and initial doses. Two concentrations after first dose were used to estimate SS AUC (foFDPK). A 2-compartment Bayesian model allometrically scaled for weight and adjusted for creatinine clearance was used to determine 24-48 hour AUCs. Differences between AUCs were compared using a mixed-effects analysis, and correlation of foPK equations to Bayesian estimates was described using Spearman’s correlation. Patient results from each method were classified as below (< 400), within (400-600), or above ( >600) targets. Results 65 adult patients were included. The median and IQR for calculated AUCs using foPOPPK, foFDPK, and Bayesian methods were 495.6 (IQR: 76.6), 498.2 (IQR: 107.4), and 472.1 (IQR: 177.9), respectively with p >0.65 for both foPK methods vs. the Bayesian method. AUCs predicted by foPK equations were poorly correlated with Bayesian AUCs (Spearman’s rho= -0.08, p=0.55), while AUCs from foFDPK better correlated with Bayesian AUCs (Spearman’s rho= 0.48, p=0.00). AUCs were within, above, and below target for 54%, 20%, and 26% for the Bayesian model; 95%, 5% and 0% for foPOPPK; and 74%, 12%, and 14% for foFDPK. foPK AUC estimates concurred with Bayesian estimates only 52% of the time. Conclusion AUCs calculated by the three methods did not differ on average, but dosing recommendations for foPK at the patient level varied substantially compared to the Bayesian method. This difference is because Bayesian estimation incorporates actual patient exposures while foPK equations rely on idealized dose timing to predict AUCs. Disclosures Kimberly C. Claeys, PharmD, GenMark (Speaker’s Bureau) Marc H. Scheetz, PharmD, MSc, Nevakar (Grant/Research Support)SuperTrans Medical (Consultant)US Patent #10688195B2 (Other Financial or Material Support, Patent holder)

QSP/PBPK model forecast of ADG20 300 mg IM in adults.
Predicted median serum ADG20 concentration is shown with the dotted line representing 100× in vitro IC90 of 0.011 mg/L or 1.1 mg/L; the solid black line represents the simulated median; the shaded area represents the 90% prediction interval. The predicted median half-life of ADG20 300 mg IM exceeded 74 days. PBPK model inputs include Ln-normal Kd,FcRn of 9.55 nM (10% IIV); IM bioavailability of 100%; 15% IIV on muscle lymph RC; and Centers for Disease Control and Prevention weight distribution of 45-150 kg. FcRn, neonatal Fc receptor; IIV, inter-individual variability; Kd, dissociation constant; Ln, log-normal; RC, reflection coefficient. Multiple strategies for calculating AUC exist, including first order pharmacokinetic (foPK) equations and Bayesian models. Most clinical applications of foPK assume unchanged patient status and project ideal administration times to estimate exposure. Bayesian modeling provides the best estimate of true drug exposure and can incorporate changing patient covariates and exact doses. We compared two commonly used foPK methods to Bayesian estimates of VAN AUC.
Methods. First order equations were performed using population PK estimates (foPOPPK) to estimate steady state (SS) AUC and initial doses. Two concentrations after first dose were used to estimate SS AUC (foFDPK). A 2-compartment Bayesian model allometrically scaled for weight and adjusted for creatinine clearance was used to determine 24-48 hour AUCs. Differences between AUCs were compared using a mixed-effects analysis, and correlation of foPK equations to Bayesian estimates was described using Spearman's correlation. Patient results from each method were classified as below (< 400), within (400-600), or above ( >600) targets.
Conclusion. AUCs calculated by the three methods did not differ on average, but dosing recommendations for foPK at the patient level varied substantially compared to the Bayesian method. This difference is because Bayesian estimation incorporates actual patient exposures while foPK equations rely on idealized dose timing to predict AUCs. Disclosures

Validation of an Allometrically Scaled Body Weight Equation to Predict Vancomycin Clearance and Guide 24-Hour Vancomycin AUC Dosing in Obese Patients
Brent Footer, PharmD, BCPS 1 ; Arthur Nguyen, PharmD 1 ; Meagan Greckel, PharmD 1 ; Colton Taylor, PharmD 2 ; Alyssa Christensen, PharmD, BCIDP 2 ; Gregory Tallman, PharmD, BCIDP, BCPS 3 ; 1 Providence Portland Medical Center, Portland, Oregon; 2 Providence Saint Vincent Medical Center, Portland, Oregon; 3 School of Pharmacy, Pacific University, Portland, Oregon Session: P-62. PK/PD Studies Background. Accurately determining empiric vancomycin (VAN) doses in obese patients represents a clinical challenge. A recent population pharmacokinetic (PK) study provided an equation to estimate vancomycin clearance (CL) based on age, sex, serum creatinine (Scr), and allometrically scaled body weight. The purpose of this study was to validate this equation in a population of obese adults treated with vancomycin at eight community-based hospitals and use the CL estimate to guide empiric VAN dosing.
Methods. The study period was November 1, 2020 and March 30, 2021. Patients were included if they were ≥ 18-year-old with a body mass index (BMI) ≥ 30 kg/m 2 , had an empiric dose targeting an AUC24 determined using the above referenced equation, and had a calculated AUC24. Only the first vancomycin course and AUC calculation for each patient were included. Patients with a creatinine clearance < 30ml/min and pregnant women were excluded. AUC24 and other PK parameters were calculated using two levels and noncompartmental analysis. Observed versus predicted CL and AUC24 were plotted to determine correlation.
Results. Sixty patients were included, of which 60% were male and 33% had a confirmed methicillin-resistant Staphylococcus aureus infection. The mean age, BMI, and baseline Scr were 61.8 years, 37.8 kg/m 2 , and 0.99 mg/dL, respectively. Fifty-three