Nuo Xu1, Yufei Shi1, Xin Liu1, Bijue Liu1, Gangfeng Yan2, Xiaoqiang Xiang1, Yixue Wang2, Xiao Zhu1
1Fudan University, 2Children's Hospital of Fudan University
Introduction Caspofungin is currently the first-line treatment for invasive fungal infections (IFIs) in the pediatric intensive care unit (PICU) population [1,2]. However, as these children often present with hypoalbuminemia, organ dysfunction, and needing ECMO support, the pharmacokinetic (PK) of caspofungin may become unpredictable [3]. Furthermore, most PK studies focus on total caspofungin levels, but its high plasma protein binding (97%) may substantially reduce free drug concentrations, impacting efficacy and safety [4]. The current lack of PK data derived from critically ill children raises concerns about the adequacy of current dosing guidelines. Objective ?To develop a population pharmacokinetic (PopPK) model describing both total and unbound caspofungin concentrations in Chinese critically pediatric patients ?To evaluate and optimize the application of current standard treatment dosing regimen Method This was a prospective, observational clinical study of intravenous caspofungin administered to patients in PICU for the treatment or prophylaxis of fungal infections (Clinical Trial, NCT04961593). The standard dosing strategy based on body surface area (BSA) using the Mosteller formula was adopted: loading dose of 70 mg/m² on the first day, followed by 50 mg/m² daily for maintenance. Two sampling strategies were adopted: A rich sampling was first used (day 6 sampling on baseline, 1, 2,4, 8 and 16 hours), along with a sparse sampling (post optimal design; pre-dose, round infusion (±0.5 hours), 6-7 hours post-dose, and a a pre-dose sample on day 7) The PopPK model was developed using the NONMEM (version 7.5.0) with theFOCE-I algorithm. Given the limited data, a linear model and a specific protein-binding model would be tested to describe PKs of total and unbound caspofungin simultaneously [5]. The final model was graphically evaluated by GoF, VPC plots and bootstrap. We assessed the PopPK model’s ability to predict unbound caspofungin concentrations in critically ill pediatric patients by comparing model-derived and fixed 3% unbound fraction estimates to observed values using mean prediction error (MPE) and normalized root mean squared error (NRMSE). Monte Carlo simulations were conducted to evaluate and optimize the standard dosing regimen using PK/PD target attainment (AUC0–24h/MIC values exceeding 450, 865, and 1185 corresponded to C. glabrata, C. albicans, and C. parapsilosis, respectively). Results A total of 29 children were enrolled in the study, with 138 concentrations for both total (0.155-58.3 mg/L) and unbound (0.00305-2.11 mg/L) caspofungin. The median (range) age and weight are 5.33 years (0.33-16) and 16 kg (4.9-74). Four patients underwent ECMO during caspofungin treatment. GoF and VPC demonstrated that the two-compartment, specific protein-binding model with allometric scaling, incorporating BSA (standardized at 0.79 m², CL exponent of 0.66 and V of 1) effectively described both total and unbound caspofungin concentrations, with parameter estimates showing good precision (RSE <50%). The model-derived approach demonstrated higher accuracy in predicting unbound caspofungin concentrations (MPE -48.85%) than the method assuming a constant unbound fraction of 3% (MPE -186.03%). The PopPK model identified BSA as a suitable factor for dosing adjustments, as allometric scaling incorporating BSA provided a superior description of caspofungin pharmacokinetics compared to weight (?OFV =-12.184) or BMI (?OFV = -9.546). The maximum binding capacity of caspofungin to albumin increased with rising albumin levels. While ECMO significantly increased the central distribution volume (by 4.91-fold), it had no notable impact on exposure. Bayesian posterior estimates further supported this conclusion (ECMO [n=4, median: 2.25 h*mg/L, (0.16-3.04)] vs. non-ECMO [n=25, median: 3.75 h*mg/L, (0.44-8.56)], p=0.1458). Simulations suggested that at the lowest albumin level of 15 g/L, lower doses were predicted for the unbound compared to the total target, which indicated that using total caspofungin concentration as the target for dose prediction may lead to overdosing in patients with hypoalbuminemia. The current BSA-based dosing regimen could ensure adequate drug exposure in critically ill pediatric patients with high variability. Dosing simulations indicated that a loading dose (LD) of 50 mg/m² and a maintenance dose (MD) of 60 mg/m² could achieve at least 90% probability of target attainment (PTA) at aminimum inhibitory concentration (MIC) of 0.03 mg/L for C. albicans.For C. glabrata, the same regimen of 50 mg/m² LD and 60 mg/m² MD was sufficient to achieve at least 90% PTA at an MIC of 0.06 mg/L. However, for C. parapsilosis, no dosing regimens were able to achieve the target of at least 90% PTA across all tested MIC values. Conclusions Our study generated new evidence to inform dosing in critically ill pediatric patients and, for the first time, reaffirmed the adequacy of current dosing guideline for pediatric patients receiving ECMO treatment.This work demonstrated the utility of pharmacometrics to help integrate data, optimize dosing algorithms and avoid unnecessary high-dose individualization of medication.
[1] ISHIWADA N, et al. Nationwide survey of neonatal invasive fungal infection in Japan. Med Mycol, 2018, 56(6): 679-86. [2] Prevalence and attributable health burden of chronic respiratory diseases, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Respir Med, 2020, 8(6): 585-96. [3] RICE T W, et al. Therapeutic intervention and targets for sepsis. Annu Rev Med, 2005, 56: 225-48. [4] KURLAND S, et al. Human plasma protein levels alter the in vitro antifungal activity of caspofungin: An explanation to the effect in critically ill? Mycoses, 2022, 65(1): 79-87. [5] Sime FB, et al. Population pharmacokinetics of total and unbound concentrations of intravenous posaconazole in adult critically ill patients. Crit Care. 2019;23(1):205.
Reference: PAGE 33 (2025) Abstr 11755 [www.page-meeting.org/?abstract=11755]
Poster: Clinical Applications