L.G.W. Franken, PharmD (1), A.D. Masman, MD (2,3), B.C.M de Winter, PharmD, PhD (1), B.C.P. Koch, PharmD, PhD (1), F.P.M. Baar, MD (2), D. Tibboel, MD, PhD (3,4), T. van Gelder, MD, PhD (1) and R.A.A. Mathôt, MD, PhD (5)
1 Department of Hospital Pharmacy, Erasmus Medical Centre, Rotterdam, the Netherlands 2 Palliative Care Centre, Laurens Cadenza, Rotterdam, the Netherlands 3 Intensive Care, Department of Paediatric Surgery, Erasmus MC-Sophia Children’s Hospital, Rotterdam, the Netherlands 4 Pain Expertise Centre, Erasmus MC-Sophia Children’s Hospital, Rotterdam, the Netherlands 5 Hospital Pharmacy – Clinical Pharmacology, Academic Medical Centre, Amsterdam, the Netherlands
Objectives: Midazolam is a commonly used sedative drug in terminally ill adult patients and is titrated to achieve the desired level of sedation. As terminally ill patients are a very heterogeneous population with severe co-morbidity (e.g. hepatic and renal impairment) patients could potentially benefit from an individualised dose that is determined beforehand. To find clinical relevant parameters for dose individualisation we performed a pharmacokinetic study on midazolam, 1OH-midazolam (1OH-M) and 1OH-midazolam-glucuronide (1OH-MG) in adult terminally ill patients.
Methods: 192 samples from 47 patients who had received midazolam orally and subcutaneously were available. Population pharmacokinetic parameters were estimated using non-linear mixed effects modelling (NONMEM 7.2). The covariates analysed were patient characteristics, co-medication and blood chemistry levels. The predictive performance of the model was evaluated with a normalised prediction distribution errors (NPDE) analysis.
Results: The data were best described by a one-compartment model for midazolam, 1OH-M and 1OH-MG. Between-subject variability (BSV) was shown for the bioavailability of midazolam, clearance of midazolam, 1OH-M and 1OH-MG and for the volume of distribution of midazolam. The population mean estimates for midazolam, 1OH-M and 1OH-MG clearance were 9.1 L/h (BSV 46%), 48 L/h (BSV 58%) and 5.7L/h (BSV 49%) respectively. Low albumin levels corresponded with low midazolam clearance and explained 19 % of the BSV in midazolam clearance. 1OH-MG clearance was correlated with the estimated glomular filtration rate (eGFR) explaining 40% of the BSV in 1OH-MG clearance. The NPDE analysis showed good model predictability with the distribution of the NPDEs not significantly deviating from a normal distribution (P-value 0.789).
Conclusions: The population pharmacokinetics of midazolam and its two major metabolites were accurately quantified. A decreased eGFR resulted in lower clearance of 1OH-MG and could therefore result in increased sedation. Low albumin levels were associated with decreased midazolam clearance, probably as a result of inflammatory response. CRP as a covariate had a similar effect as albumin yet less significant. eGFR and albumin might be useful clinical parameters to develop an individualized dosing regimen. Additional studies on the pharmacodynamic effects of midazolam in this population are recommended.
Reference: PAGE 25 (2016) Abstr 5836 [www.page-meeting.org/?abstract=5836]
Poster: Drug/Disease modeling - Other topics