L.G.W. Franken (1), A.D. Masman (1), B.C.M de Winter (1), B.C.P. Koch (1), F.P.M. Baar (2), D. Tibboel (1), T. van Gelder (1) and R.A.A. Mathot (3)
(1) Erasmus Medical Centre, Rotterdam, Netherlands, (2) Laurens Cadenza, Rotterdam, Netherlands, (3) Academic Medical Centre, Amsterdam, Netherlands
Objectives: To develop a population pharmacokinetic (PK) model for morphine and its two major metabolites: morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G) in terminally ill patients, to determine covariates that can be used for dose individualisation.
Methods: A population pharmacokinetic analysis was conducted with 199 randomly collected samples obtained from 47 terminally ill patients, using non-linear mixed effects modelling NONMEM 7.2.0 [1]. In the covariate analysis blood chemistry levels (e.g. albumin and creatinine), patient characteristics (e.g. age, diagnosis) and time to death were analysed by forward inclusion followed by, backward elimination. For the evaluation of the final model a normalised prediction distribution error (NPDE) analysis was performed.
Results: The data were best described by a two-compartment model for morphine, a one-compartment model for both M3G and M6G and proportional residual error models for all the compounds. Between-subject variability (BSV) could be estimated for the bioavailability of oral morphine, morphine clearance, metabolite clearance and the volumes of distribution of the metabolites. Typical clearance values were 49.4 l/h (morphine, BSV 53%), 1.56l/h (M3G, BSV 31%) and 1.94 l/h (M6G, BSV 31%). Serum creatinine was negatively correlated with metabolite clearance and serum albumin was positively correlated, together explaining 78% of the between-subject variability in metabolite clearance. Furthermore we showed that morphine clearance decreased from 49.4 L/h to 31.4 L/h in the last three weeks before death. Evaluating the final model by NPDE analysis showed accurate predictive ability (global adjusted P value > 0.05).
Conclusion: The population pharmacokinetics of morphine, M3G and M6G in terminally ill patients were accurately quantified. Serum creatinine and albumin levels together were a better predictor of metabolite clearance than creatinine alone. This may be caused by the presence of cachexia and loss of muscle mass in terminally ill patients, in whom serum creatinine may overestimate renal function. Lower albumin levels may be an indicator of cachexia, and for overestimation of GFR based on creatinine levels. Our results show that morphine clearance declines as death approached and that M3G and M6G can accumulate in patients with decreased renal function. Dose adjustment might therefore be required in these patients, however the clinical effect of this requires further study.
References:
[1] Beal SL, Sheiner LB, Boeckmann AJ, Bauer RJ (Eds.) NONMEM Users Guides. 1989-2011. Icon Development Solutions, Ellicott City, Maryland, USA.
Reference: PAGE 24 () Abstr 3334 [www.page-meeting.org/?abstract=3334]
Poster: Drug/Disease modeling - Other topics