Nada Dia (1), Sabrina De Winter (2), Omar Elkayal (1), Beatrijs Mertens (1), Peter Vanbrabant (3), Willy Peetermans (3), Isabel Spriet (1,2), and Erwin Dreesen (1)
(1) Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, Leuven, Belgium, (2) Pharmacy Department, University Hospitals Leuven, Leuven, Belgium, (3) De-partment of Internal Medicine, University Hospitals Leuven, Leuven, Belgium
Objectives:
Sepsis related to Gram-negative infections is a life-threatening organ dysfunction responsible for 11 million deaths per year worldwide [1]. Early and appropriate administration of antibiotics is associated with improved clinical and microbiological responses [2]. Amikacin is prescribed to critically ill adult patients with sepsis or septic shock admitted to the emergency department. The pharmacokinetic-pharmacodynamic (PKPD) index is defined as a ratio of peak concentration (Cpeak) to minimum inhibitory concentration (MIC) of at least eight [3]. Considering an 8 mg/L MIC of the least susceptible bacteria (i.e., Enterobacteriaceae and Pseudomonas aeruginosa), a Cpeak of a minimum 64 mg/L ensures target attainment (TA) [4]. The desired concentration 24 hours after the administration of amikacin is 3 mg/L. Conventional dosing of amikacin (a single 15 mg/kg dose) was associated with a poor probability of PKPD TA [5].
The aim of our population pharmacokinetic (popPK) simulation study was to identify an amikacin dosing strategy with an optimized probability of Cpeak being ≥64 mg/L (efficacy) and an acceptable probability of the C24h being ≤3 mg/L (nephrotoxicity).
Methods:
Monte Carlo simulations were performed using a published two-compartment popPK model [6] to explore four dosing scenarios of a single dose of 30-minute infusion of amikacin:
- 15 mg/kg
- 25 mg/kg capped to 2,000 mg at a total body weight above 80 kg
- 1,500 mg
- 2,000 mg
Covariates were body mass index at inclusion, estimated glomerular filtration rate (eGFR) at 24 hours after inclusion, serum total protein at inclusion, serum sodium at inclusion, and fluid balance for the first 24 hours after inclusion. The (transformed) covariates and the body weight were sampled from a truncated normal distribution, with boundaries for each covariate set to the range as observed in the study dataset [6]. Transformation to normality was applied based on Shapiro’s test. Correlations between covariates were preserved when creating a virtual patient dataset. A total of 6,400 virtual patients were created, undergoing the four dosing regimens. A total of 1,000 simulations per virtual patient was performed in NONMEM (v7.5; Icon Development Solutions, Gaithersburg, MD, USA).
Receiver operating characteristics (ROC) analysis was performed to investigate the clinical relevance of the covariates. The area under the ROC curve (AUROC) with its 95% confidence interval (CI) was calculated across all the covariates and the body weight by applying 1,000 bootstraps. AUROC is an indicator of the overall performance of the diagnostic accuracy. The probability of TA for Cpeak and C24h was evaluated at 1 hour and 24 hours after the start of the infusion, respectively across the range of covariate values per dosing regimen. A probability of TA >90% was defined as clinically acceptable. Additionally, the time to achieve a concentration ≤3 mg/L was evaluated across the covariates ranges per dosing regimen.
Results:
Body weight had numerically the highest diagnostic accuracy in identifying the probability of attaining the desired Cpeak (AUROC 0.76, 95% CI 0.62-0.87). This is likely because the body mass index, which is dependent on the body weight, was a covariate on the volume of distribution, a key driver of the Cpeak.
eGFR had numerically the highest diagnostic accuracy in identifying the probability of attaining the desired C24h (AUROC 0.99, 95% CI 0.96-1.00). eGFR was identified as a covariate on the clearance. Hence it helps predict the exposure 24 hours after the start of the infusion.
The probability of Cpeak TA was >90% across the entire body weight range (36-111 kg) with a flat dose of 2,000 mg. The Cpeak range was 90-128 mg/L.
The probability of Cpeak TA was >90% across the entire eGFR range (6.1-176.1 mL/min/1.73 m2) with dosing regimens of 2,000 mg and 25 mg/kg capped to 2,000 mg. The Cpeak range was 84-110 mg/L, and 69-102 mg/L, respectively.
The probability of C24h TA was never >90% neither across the entire body weight range nor the eGFR range.
Time to achieve a concentration ≤3 mg/L ranged between 21.9 hours and 61.3 hours after the start of the infusion across all dosing regimens in virtual patients with a body weight range of 36 to 111 kg.
Conclusions:
A single flat dose of 2,000 mg amikacin, which can be easily prepared from commercially available vials, may be advised for prospective evaluation in a clinical study. The desired concentration of toxicity is not attained by the majority of patients 24 hours after the start of the infusion.
References:
[1] https://www.who.int/news/item/08-09-2020-who-calls-for-global-action-on-sepsis—cause
-of-1-in-5-deaths-worldwide.
[2] Evans L et al. Intensive Care Med (2021) 47, 1181-1247.
[3] Spriet I et al. J Antimicrob Chemother (2020) 75, 3696-3697.
[4] https://www.eucast.org/clinical_breakpoints/.
[5] De Winter S et al. Int J Antimicrob Agents (2018) 51, 562-570.
[6] De Winter S et al. Eur J Drug Metab Pharmacokinet (2021) 46, 653-663.
Reference: PAGE 30 (2022) Abstr 10042 [www.page-meeting.org/?abstract=10042]
Poster: Clinical Applications