Exploring the relationship between analgesic event rate and pain intensity in kidney stone surgery: A Repeated Time to Event Pilot Study
RV Juul(1), KV Pedersen(2, 4), LL Christrup(1), AE Olesen(1, 3), AM Drewes(3), PJS Osther(4), TM Lund(1)
1) Department of Drug Design and Pharmacology, University of Copenhagen, Denmark. 2) Department of Clinical Genetics, Vejle Hospital, Denmark. 3) Mech-Sense, Aalborg University Hospital, Denmark, 4) Urological Research Center, Lillebaelt Hospital, University of Southern Denmark, Fredericia, Denmark
Objectives: Opioid consumption has often been reported as an indirect measure of pain in postoperative pain trials. The rate of consecutive analgesic events can be described by repeated time-to-event (RTTE) modelling in order to analyse the dynamical changes and concentration-effect-relationships with analgesic consumption. However a relationship with pain intensity has not yet been established. The aim of this pilot study was to discuss how best to investigate the relationship between RTTE hazard of analgesic events and pain intensity in postoperative pain.
Methods: Data was available from 44 patients undergoing kidney stone surgery (percutaneous nephrolithotomy), who were randomized to morphine or oxycodone administered upon request (1). Pain intensity was recorded on Numerical Rating Scale (NRS) every 15 min until 4 hours after admission to recovery ward. RTTE modelling of analgesic events was performed in NONMEM 7.2 and PsN (2). Gompertz and exponential distribution models were evaluated. Post-hoc linear mixed effect modelling was performed between estimated RTTE hazard and observed NRS using the lme4 package in R (3).
Results: A Gompertz distribution model adequately described data, with a baseline event rate of 0.64h-1 (RSE 25%) and a decline in event rate with a half-life of 1.2h-1 (RSE 22%). No significant differences were found between morphine and oxycodone. Post-hoc linear mixed effects modelling of the estimated RTTE hazard and NRS is demonstrated, but do not optimally describe the categorical nature of NRS.
Conclusions: An RTTE model well described both morphine and oxycodone consumption data. RTTE modelling is a promising tool to investigate correlations between opioid consumption and pain intensity in time, but appropriate methods needs to be applied to study this relationship.
References:
1) Pedersen KV, et al. Urolithiasis, 41.5 (2013): 423-430.
2) Karlsson KE. , et al. AAPS J, 13.1 (2011): 83-91.
3) Bates D. MM & Bolker B. lme4: (2012). http://CRAN.R-project.org/package=lme4