Loading…
1564. Target Attainment of Empiric Vancomycin Therapy to Achieve Safe and Effective Exposure When it Matters Most: How Much of the Drug Do We Really Need in the First 48 Hours?
Abstract Background Empiric dosing of vancomycin (VAN) to reach targets is a daunting task due to the large variability observed in the pharmacokinetics of this agent. With the change to AUC driven dosing on the horizon, the goal of this study was to establish empiric dosing requirements of vancomyc...
Saved in:
| Published in: | Open forum infectious diseases 2019-10, Vol.6 (Supplement_2), p.S570-S571 |
|---|---|
| Main Authors: | , , , , , , , |
| Format: | Article |
| Language: | English |
| Subjects: | |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract
Background
Empiric dosing of vancomycin (VAN) to reach targets is a daunting task due to the large variability observed in the pharmacokinetics of this agent. With the change to AUC driven dosing on the horizon, the goal of this study was to establish empiric dosing requirements of vancomycin that effectively achieve the desired AUC of 400 to 600 mg hours/L targets in the first 48 hours of therapy.
Methods
VAN TDM data were used in this analysis. A two-compartment model was fitted with Bayesian routines to establish the AUCs. Then, the AUC achieved was used to identify the desired total daily dose (TDD, capped at 4,500 mg) needed to attain an AUC in the target range for days 1 and 2, per patient. Next, multivariable regression was undertaken to predict this desired dose with frequently calculated weight and renal function indices. Last, model validation in a test data set based on calculated ME, RMSE and their 95% CI took place, and then the best model was used to simulate TDDs at 24 h intervals. To evaluate the agreement between the 2 pharmacists selecting the final TDDs by screening the simulated regimens, a weighted Kappa was calculated.
Results
1450 concentrations from 268 patients (60.9% male) with mean (IQR) age of 62.8 (52.7, 75) years, weight of 79.1 (63.2, 90.9) kg and CrCl of 76.7 (36.8, 110.6) mL/minute were analyzed. Fit of the model to data was excellent with R2 = 0.98 (Figure 1). AUC attainment with actual dose vs. the AUCs based on the desired dose was poor (Figure 2). Final regression model [ME (95% CI) 31.58 (−160.38, 217.34) mgs; RMSE (95% CI) 761.98 (628.19, 895.93) mgs] identified adjusted body weight (ABW) (P = 0.02), CrCl (P < 0.001), age (P = 0.05) and sex (P = 0.01) on day 1, vs. CrCl (P < 0.001), age (P = 0.02) and sex (P = 0.002) on day 2 as predictors of TDD. Kappas showed near perfect agreement for day 1 (0.992, P < 0.01) and day 2 (0.883, P < 0.01) between the raters resulting in the selection of the final dosing regimens (Figures 3 and 4).
Conclusion
Our model accurately identified the 4 major variables most likely to explain VAN variability in predicting AUC in the first 48 hours. These detailed dosing recommendations—strengthened by rigorous external validation and near perfect between rater agreements—allow for the design of safe and effective AUC driven empiric dosing regimens.
Disclosures
All authors: No reported disclosures. |
|---|---|
| ISSN: | 2328-8957 2328-8957 |
| DOI: | 10.1093/ofid/ofz360.1428 |