I-19 John Maringwa

Model-based meta-analysis of summary clinical outcome data in idiopathic pulmonary fibrosis (IPF)

J. Maringwa (1), E. Cox (1), L. Harnisch (2) and X. Gao (3)

(1) Quantitative Solutions BV, Breda, The Netherlands (2) Global Pharmacometrics, Pfizer Inc, Sandwich, UK; (3) Clinical Pharmacology, Pfizer Inc, Groton, USA.

Objectives: (1) To perform a model-based meta-analysis of comparative effectiveness of various drugs investigated in treating IPF patients, and to assess the effect of trial- and patient disease variables on drug treatment effect. (2) To simulate expected treatment- and trial performance in IPF.

Methods: Publicly available longitudinal summary level data of clinical efficacy outcomes (forced vital capacity, FVC % predicted) from 7 drugs in 11 placebo-controlled IPF trials (2,834 patients) were analyzed using an asymptotic repeated measures regression model to describe the time course of treatment effect (drug, dose) on FVC. A non-parametric placebo model was used to account for heterogeneity in the time course of placebo effect, both between- and within trials.

Results: Significant, time-dependent treatment effects on FVC were identified for pirfenidone and the investigational compound BIBF1120 and for the latter so in a linear dose dependent manner. No statistically significant treatment effect was observed for the other drugs (etanercept, interferon-gamma, sildenafil, bosentan, and azathioprine). Patient disease variables such as baseline FVC values did not appear to impact treatment effect in a significant manner. The maximum estimated treatment effect is 3.29- and 5.14% for pirfenidone and 300mg/day BIBF1120, respectively, while the estimated time to achieve 50% of this maximum effect for both drugs is ~15 weeks. For BIBF1120 doses of 200mg/day and higher, a larger treatment effect on FVC compared to pirfenidone is expected. Trial simulations suggest that a trial with mean baseline FVC of 70%, duration of 52 weeks, and enrolling 40 patients per group will have a power of ~80% to detect a treatment effect of 1.53%.

Conclusions: Pirfenidone and BIBF1120 slow down the decline of lung function (FVC) in IPF patients in a significant manner. At doses of 200mg/day and higher, the effect of BIBF1120 on FVC decline is expected to be larger than for pirfenidone. Model-based simulations allowed exploring several trial design aspects.

Reference: PAGE 21 (2012) Abstr 2422 [www.page-meeting.org/?abstract=2422]

Poster: Study Design

PDF poster / presentation (click to open)