Ketercapaian Target PK/PD Meropenem Berbasis Model-informed Precision Dosing dan Hubungannya dengan Kegagalan Terapi pada Pasien Sepsis di Intensive Care Unit

NOVIA DANI AISYAH, Prof. Dr. apt. Akhmad Kharis Nugroho, M.Si; Prof. Dr. apt. Tri Murti Andayani, Sp.FRS

2026 | Tesis | S2 Mag.Farmasi Klinik

Sepsis di intensive care unit (ICU) merupakan penyebab utama kematian di Indonesia dengan mortalitas mencapai 69%. Meropenem, antibiotik pilihan akhir, memerlukan pencapaian target farmakodinamik 100% fT>MIC pada pasien kritis, namun variabilitas farmakokinetik (PK) pada sepsis menyebabkan ketidakpastian dosis. Penelitian ini bertujuan mengoperasionalisasikan model- informed precision dosing (MIPD) berbasis prediksi a priori (berbasis kovariat) menggunakan model farmakokinetik populasi eksternal Asia Tenggara, tanpa pengukuran konsentrasi aktual, untuk menggambarkan ketercapaian target PK/PD serta hubungannya dengan kegagalan terapi 72 jam dan mortalitas pada pasien sepsis dewasa di ICU RS Akademik UGM.
Studi farmakometrik kohort retrospektif ini melibatkan 69 pasien periode Januari–Desember 2024. Parameter PK individual dihitung menggunakan model dua kompartemen Boonpeng dkk. (2022) dengan kovariat klirens kreatinin Cockcroft-Gault, albumin serum, dan status syok. Data albumin yang hilang diimputasi menggunakan MICE. Nilai minimum inhibitory concentration (MIC) ditetapkan berdasarkan distribusi MIC50 sub-populasi ATLAS Asia 2020–2024. Simulasi profil konsentrasi-waktu dilakukan menggunakan package deSolve di R. Analisis statistik mencakup uji bivariat dengan validasi bootstrap (B=2.000), regresi logistik multivariat Firth, serta tiga lapis analisis sensitivitas.
Sebanyak 54 pasien (78,3%) diprediksi mencapai target 100% fT>MIC; seluruh 15 pasien inadekuat (21,7%) memiliki MIC = 32 mg/L yang didominasi patogen carbapenem-resistant Acinetobacter baumannii. Ketercapaian target PK/PD tidak berasosiasi signifikan dengan kegagalan terapi 72 jam (RR 1,16; 95% CI 0,58–2,29; p = 0,665) maupun dengan mortalitas rumah sakit (RR 0,74; 95% CI 0,35–1,56; p = 0,535). Hasil analisis multivariat dan sensitivitas konsisten. Pendekatan ini berhasil mengidentifikasi variasi klirens individual lebih dari enam kali lipat (2,21–13,58 L/jam) dan menstratifikasi 21,7% pasien berisiko paparan subterapeutik, seluruhnya hanya berdasarkan data kovariat klinis rutin tanpa memerlukan pemantauan kadar obat (TDM). Prediksi a priori berbasis kovariat menggunakan model populasi eksternal berpotensi mendukung penerapan MIPD di ICU dengan sumber daya terbatas. Validasi prospektif menggunakan TDM dan pengukuran MIC aktual tetap diperlukan sebelum implementasi klinis. 

Sepsis in the intensive care unit (ICU) represents a leading cause of mortality in Indonesia, with a reported death rate of up to 69%. Meropenem, the antibiotic of last resort, requires attainment of the pharmacodynamic target of 100% fT>MIC in critically ill patients; however, pharmacokinetic (PK) variability in sepsis introduces considerable dosing uncertainty. This study aimed to operationalise covariate-based a priori model-informed precision dosing (MIPD) using an external Southeast Asian population pharmacokinetic model, without actual drug concentration measurements, to characterise PK/PD target attainment and explore its association with 72-hour therapeutic failure and mortality in adult sepsis patients admitted to the ICU of UGM Academic Hospital.
This pharmacometric retrospective cohort study enrolled 69 patients from January to December 2024. Individual PK parameters were estimated using the two-compartment model developed by Boonpeng et al. (2022), incorporating covariates of Cockcroft-Gault creatinine clearance, serum albumin, and shock status. Missing albumin data were imputed using multivariate imputation by chained equations (MICE). Minimum inhibitory concentration (MIC) values were assigned based on the MIC?? distribution of the ATLAS Asia sub-population (2020 2024). Concentration-time profiles were simulated using the deSolve package in R. Statistical analyses included bivariate testing with bootstrap validation (B=2,000), Firth penalised multivariate logistic regression, and three-layer sensitivity analyses.
A total of 54 patients (78.3%) were predicted to attain the 100% fT>MIC target. All 15 patients with inadequate attainment (21.7%) had MIC values of 32 mg/L, predominantly attributable to carbapenem-resistant Acinetobacter baumannii. PK/PD target attainment was not significantly associated with 72-hour therapeutic failure (RR 1.16; 95% CI 0.58–2.29; p = 0.665) or in-hospital mortality (RR 0.74; 95% CI 0.35–1.56; p = 0.535). Multivariate and sensitivity analyses yielded consistent results. This approach successfully identified more than a six fold variation in individual clearance (2.21–13.58 L/h) and stratified 21.7% of patients at risk of subtherapeutic exposure, entirely based on routinely available clinical covariate data without the need for therapeutic drug monitoring (TDM). Covariate-based a priori prediction using an external population model holds promise for supporting MIPD implementation in resource-limited ICU settings. Nevertheless, prospective validation incorporating TDM and actual MIC measurements remains essential prior to clinical implementation. 

Kata Kunci : meropenem, sepsis, model-informed precision dosing, pharmacometrics, target attainment