Landewé, Robert2; Østergaard, Mikkel1; Keystone, Edward C2; Florentinus, Stefan2; Liu, Shufang2; van der Heijde, Désirée2
1 Videncenter for Reumatologi og Rygsygdomme, HovedOrtoCentret Rigshospitalet, Rigshospitalet, The Capital Region of Denmark2 unknown
OBJECTIVE: A longitudinal integration approach evaluated all radiographic scores assessed over 10 years, rather than only completer data, from 2 studies of adalimumab (ADA) for rheumatoid arthritis (RA). METHODS: The DE019 (methotrexate [MTX]-inadequate responders, longstanding RA) and PREMIER (MTX-naive, early RA) studies, respectively, had 1- or 2-year double-blind periods followed by 9- or 8-year open-label extensions (OLEs). Patients received ADA ± MTX in both OLEs. Radiographic progression was assessed using change from baseline in modified total Sharp score (ΔmTSS). A mixed-effects model was used post hoc to evaluate repeated measurements of different data campaigns and to estimate ΔmTSS through up to 10 years of treatment based on original randomization groups (placebo [PBO] + MTX or standard dose ADA + MTX). RESULTS: Data from patients with baseline and ≥1 postbaseline radiograph were included (n = 327 for DE019; n = 452 for PREMIER). Integrated and 10-year completer ΔmTSS progression curves differed slightly. In DE019, for patients originally assigned PBO + MTX, accrued ΔmTSS at year 10 was 6.6 units (integrated model) and 6.2 units (completers). For patients originally assigned ADA + MTX, accrued ΔmTSS was 0.9 units by integrated analysis and 0.7 units in completers. In PREMIER, for patients originally assigned PBO + MTX, accrued ΔmTSS at year 10 was 11.2 units (integrated analysis) and 11.0 units (completers). For patients originally assigned ADA + MTX, accrued ΔmTSS was 5.1 units (integrated analysis) and 4.0 units (completers). A higher radiographic progression rate was observed in patients who received delayed versus immediate ADA + MTX treatment. CONCLUSIONS: Longitudinal integrated analysis provided robust estimates of radiographic progression that only slightly differed from completers-only scores and confirmed the effects.
Arthritis Care and Research, 2015, Vol 67, Issue 2, p. 180-6