Tigecycline population pharmacokinetics in critically ill patients with decompensated cirrhosis and severe infections
Carla Bastida (1), Marcial Cariqueo (1), María Hernández-Tejero (2), Fátima Aziz (2), Virginia Fortuna (3), Mercè Brunet (4), Javier Fernández (2,5), Dolors Soy (1).
(1) Pharmacy Department, Division of Medicines, Hospital Clinic of Barcelona, University of Barcelona, Barcelona, Spain; (2) Liver ICU, Hospital Clinic of Barcelona, University of Barcelona, IDIBAPS, CIBERehd, Barcelona, Spain; (3) Pharmacology and Toxicology Section, Biochemistry and Molecular Genetics Department, Biomedical Diagnostic Centre, Hospital Clinic of Barcelona, Barcelona, Spain; (4) Pharmacology and Toxicology Section, Biochemistry and Molecular Genetics Department, Biomedical Diagnostic Centre, Hospital Clinic of Barcelona, University of Barcelona, IDIBAPS, CIBERehd, Barcelona, Spain; (5) European Foundation for the Study of Chronic Liver Failure (EF-Clif), EASL-CLIF Consortium and Grifols Chair, Barcelona, Spain
Introduction: Tigecycline is considered an appropriate agent for the treatment of complicated infections.[1] Its efficacy has been related to the ratio AUC0–24 to the MIC of the pathogen.[2-4] Physiopathological changes that occur in cirrhotic patients could alter tigecyclin’s PK and compromise target attainment.[5]
Objectives:
- Describe tigecycline's PK in patients with decompensated cirrhosis and severe bacterial infections,
- Identify sources of PK variability,
- Assess the performance of different dosing regimens to optimize the PK/PD target based on clinical variables.
Methods: Prospective observational study conducted at the Liver Intensive Care Unit of a tertiary hospital. Patients’ demographics, biochemistry and clinical data were collected. Blood samples were obtained at pre-dose and at 1, 2, 5 and 8-12h after drug administration.
Non-linear effects modelling was performed using NONMEM®v.7.4[6]. Internal validation of the PK model was assessed with visual predictive check (VPC)[7] and non-parametric bootstrap resampling techniques[8].
Monte Carlo simulations using the final population PK (popPK) model were performed. Three PK/PD targets were assessed based on literature[2–4]: (i) an AUC0–24/MIC≥4.5 for hospital-acquired pneumonia; (ii) ≥6.96 for intra-abdominal infections; and (iii) ≥17.9 for skin infections.[3,4] Steady-state AUC0–24 was calculated as the ratio between 24h-tigecycline dose and total individual clearance (CL). Probability target attainment (PTA) values were calculated corresponding to the percentage of patients that achieve the PK/PD targets according to different clinical susceptibility breakpoints.[9] Treatment was considered successful if PTA≥90%.
Results: A total of 20 critically-ill patients were included in the study. Sixteen (80%) were men with a median age of 59 years [range 51–67].
The popPK model was developed using data from 99 tigecycline serum concentration samples. Data were best described by a two-compartment linear model with IIV incorporated into the total CL and central volume of distribution (V1). Residual variability was characterized by a proportional error model of 21.67%.
Concerning the covariate analysis, MELD score (Model for End-stage Liver Disease) significantly influenced tigecycline CL, reducing the IIV from 59.5% to 46.4%. Higher MELD values, associated with increasing severity of hepatic dysfunction, relate to a lower tigecycline CL. Total serum proteins significantly influenced V1, leading to a decrease in its IIV (from 60.2% to 51.5%), but without a clinically relevant change in PTA and was not included in the simulations.
Considering a target AUC0–24/MIC≥4.5, >90% of the patients would be successfully treated for bacteria with a MIC up to 0.5 mg/L, with a loading dose(D*) of 100mg followed by 25mg q12h and D*150mg followed by 100mg q12h for MELD scores>16 and ≤15, respectively. For an MIC=1 mg/L and a MELD score>8, >90% of the patients would achieve the PK/PD target with a D*150mg followed by 100mg q12h, and 70% of those with a MELD≤7. Only patients with a MELD score>16 could be satisfactory treated with high-dose tigecycline (100mg q12h) for MIC≥2 mg/L.
Considering a target AUC0–24/MIC≥6.96, a standard tigecycline dose could be used for the treatment of bacteria with a MIC=0.5 mg/L if MELD score>8. Regarding an MIC=1 mg/L, the same dosing regimen could be applied in patients with a MELD score>26, but a higher maintenance dose (100mg q12h) would be required if MELD score is between 8-25. Only 32% of the patients with MELD≤7 would reach the target with the SPC recommended dose.
A target of AUC0–24/MIC≥17.9 could be attained in those patients with a MELD score>36 and >16 with a tigecycline dose of 50mg/12h and 100mg q12h, respectively, considering a MIC=0.5 mg/L. Less than 80% and 20% of the patients with lower MELD scores (<15 and <7, respectively) would be adequately treated with a dose of 150mg+100mg q12h.
Conclusions:
We present the first popPK study of tigecycline in critically-ill patients with cirrhosis and severe infections. We identified MELD score as a covariate that significantly influences tigecycline CL, and total serum proteins as a modifier of tigecycline V1. Different dosing regimens are presented to reach three targets considering the effect of the above-mentioned variables, which can easily be measured in routine clinical practice.
References:
- Xie J, Roberts JA, Alobaid AS et al. Population pharmacokinetics of tigecycline in critically ill patients with severe infections. Antimicrob Agents Chemother 2017; 61: e00345-17.
- Passarell J, Meagher A, Liolios K et al. Exposure-response analyses of tigecycline efficacy in patients with complicated intra-abdominal infections. Antimicrob Agents Chemother 2008; 52: 204–10.
- Meagher AK, Passarell JA, Cirincione BB et al. Exposure-response analyses of tigecycline efficacy in patients with complicated skin and skin-structure infections. Antimicrob Agents Chemother 2007; 51: 1939–45.
- Bhavnani S, Rubino C, Hammel J et al. Pharmacological and patient-specific response determinants in patients with hospital-acquired pneumonia treated with tigecycline. Antimicrob Agents Chemother 2012; 56: 1065–72.
- Takikawa H. Hepatobiliary transport of bile acids and organic anions. J Hepatobiliary Pancreat Surg 2002; 9: 443–7.
- Beal SL et al. 1989-2011. NONMEM Users Guides. Icon Development Solutions, Ellicott City, Maryland, USA.
- Bergstrand M, Hooker AC, Wallin JE et al. Prediction-corrected visual predictive checks for diagnosing nonlinear mixed-effects models. AAPS J 2011; 13: 143–51.
- Cheng A, Yeager M. Bootstrap resampling for voxel-wise variance analysis of three-dimensional density maps derived by image analysis of two-dimensional crystals. J Struct Biol 2007; 158: 19–32.
- EUCAST. Clinical Breakpoints. http://www.eucast.org.