DOP071 Tofacitinib plasma concentration monitoring is not needed for optimisation of induction therapy in moderate-to-severe ulcerative colitis: results of pooled exposure-response analyses of Phase 3 induction studies
A. Mukherjee1, S. Tsuchiwata2, C. Deng3, C. Vong4, R. Xie3, S.W. Martin4, D. Yu5, D. Woodworth5, W. Niezychowski5, C. Su*5
1Pfizer Inc, Groton, United States, 2Pfizer Japan Inc, Tokyo, Japan, 3Pfizer Inc, Shanghai, China, 4Pfizer Inc, Cambridge, United States, 5Pfizer Inc, Collegeville, Pennsylvania, United States
Tofacitinib is an oral, small molecule Janus kinase inhibitor that is being investigated for ulcerative colitis (UC). Two recently completed, identically-designed, placebo-controlled, Phase 3 induction studies evaluating tofacitinib 10 mg twice daily (BID) in patients with moderately to severely active UC met their primary efficacy endpoints at Week 8 and did not demonstrate any unexpected safety events with respect to tofacitinib studies in other populations.
We characterised the exposure-response (ER) relationship between tofacitinib plasma concentration and efficacy using pooled data from Phase 3 induction studies.
Plasma samples for population pharmacokinetic (PK) analysis were collected from all patients. PK analysis was performed to derive individual tofacitinib exposure metrics, average concentration (Cavg) and trough concentration (Ctrough). The efficacy endpoints evaluated were remission and mucosal healing at Week 8 (primary and key secondary endpoint, respectively, in each study; centrally read) based on Mayo scores. The ER analysis implemented Emax models in a logistic regression analysis in SAS v9.2. Baseline disease (Mayo score, C-reactive protein, albumin), prior use of tumour necrosis factor inhibitors and immunosuppressants, baseline corticosteroid and aminosalicylate use, and age, sex, and body mass index were evaluated as covariates.
The PK analysis included concentrations from 866 patients on tofacitinib, and the ER analysis included 217 placebo patients with no exposure. Estimated PK parameters for a typical study subject were oral clearance (CL/F) = 25 L/hr and oral volume of distribution (V/F) = 105 L. Inter-patient variability (coefficient of variation) in CL/F was 19.9%. Cavg was a significantly better predictor of efficacy than Ctrough for both efficacy endpoints. Predicted difference from placebo in remission for patients in the lowest to highest quartiles of Cavg at the 10 mg BID dose was 13.7%, 14.7%, 14.4%, and 17.6%, respectively, and 15.1% for all patients at this dose. Observed and model-predicted probabilities of response and difference from placebo in response rates are shown in Figure 1.
Figure 1. Probability of remission (Panel A) and mucosal healing (Panel B) as a function of average tofacitinib concentration (Cavg) based on an Emax model in a logistic regression analysis. Observed proportions were summarised by placebo and quartiles of Cavg within subjects receiving tofacitinib 10 mg BID. Observed and model predicted difference from placebo in proportion of subjects achieving remission (Panel C) and mucosal healing (Panel D) were also summarised by quartiles of Cavg at 10 mg BID and by tofacitinib dose.
The highly similar remission rate observed across the 4 quartiles in plasma exposure at the 10 mg BID dose indicates that monitoring of tofacitinib plasma concentrations is not necessary for treatment optimisation during induction therapy with tofacitinib.