Aims: To develop a population pharmacokinetic-pharmacodynamic (PK-PD) model describing the concentration-effect relationships for morphine on experimental pain caused by skin heat and muscle pressure.
Methods: Data were analysed from a study of 39 healthy volunteers who received 30 mg oral morphine, or placebo. Blood samples were collected up to 150 min post-dose, while experimental skin heat and muscle pressure pain was induced at: 0, 15, 30, 45, 60, 150 min. Nociceptive input was increased until the subjects reported a pain score of 7 on a 0-10 visual analogue scale, where 5 is the pain detection threshold. The PKs and PD of morphine, M6G and M3G were analysed with non-linear mixed effects modelling (NONMEM v. 7.2, ICON Plc) using a sequential approach. One- and two-compartment models were fitted to morphine, M6G, and M3G plasma concentration-time data. First-order and transit compartment absorption models were tested. The placebo-response of both pain metrics was fitted to slope o, linear, and quadratic effect vs time models. The drug effect was tested as proportional or additive to the placebo-response. Drug effect was fitted to slope 0, linear, and Emax models of effect versus plasma or effect compartment concentration. Morphine, M6G, or M3G were tested as the concentration driving analgesia. The influence of the covariates sex, height, weight, body mass index (BMI), and age was also screened.
Results: The PK of morphine, M6G, and M3G were best described with one-compartment distribution, and the absorption of morphine was best described with a six transit compartment model with a combined additive and proportional residual error model. The placebo-response for skin heat was best described by a linear model with between-occasion variability (BOV) on baseline, and additive residual error model. For muscle pressure, a slope 0 model with BOV on baseline and proportional error model best described the placebo response. Morphine effect on both skin heat and muscle pressure was proportional to the placebo-response and described by a linear model with effect compartment and between subject variability (BSV) on drug effect slope. We were not able to differentiate between morphine, M6G, and M3G as the driving concentration for the effect. The inclusion of covariates did not improve the fit of any model to the data.
Conclusion: The data were characterised by high inter- and intra-individual variability. In contrast to previous studies the placebo-response showed little change with time. A linear concentration-effect relationship of morphine was identified for both pain metrics.
