Differences in the in vitro and in vivo concentration-effect relationship of selective and non-selective COX inhibitors : role of translational pharmacology in pain research

Background and purpose. In this manuscript we assess the concentration-effect relationships of three COX inhibitors varying in selectivity (i.e., rofecoxib, diclofenac and ketolorac) by modelling their inhibitory effect on PGE2 and TXB2 in vitro and in vivo in rats and compare their predictive value for the treatment effect in humans. Experimental approach. PGE2 inhibition was measured by whole blood LPS-stimulation whilst TXB2 inhibition was assessed by blood clotting. For the evaluation of drug effects in vitro, pre-defined concentrations were added to blood samples. For the evaluation of the effects in vivo, rats were given an intraperitoneal dose of each compound. Subsequently, serial samples were collected for analysis of concentrations and drug effect on PGE2 and TXB2. In vivo human data from previous publications was used for comparison. PKPD analysis was performed using nonlinear mixed effects modelling. Key results. Inhibition of PGE2 and TXB2 was characterised by a sigmoid Emax model. The IC80 values for PGE2 and TXB2 inhibition were used as parameter of interest for the prediction of the analgesic effect in vivo. All three COX inhibitors showed significant differences in vitro and in vivo (p>0.05) as well as between species. In vitro-in vivo potencies showed a correlation in rats and in humans. Conclusions and implications. The assessment the effect of COX inhibitors in vitro enables evaluation of in vitro-in vivo correlations. In vitro data also provided better estimates of the selectivity and potency of different compounds within and between species. These findings strongly suggest that the use of in vitro human data, instead of rodent models of pain as basis for determining the effective exposure for analgesia in patients.