Article Text
Abstract
Background In placebo (PBO)-controlled trials of anti-TNF therapy in spondyloarthritis (SpA), background NSAIDs are maintained at stable levels to facilitate demonstration of the anti-TNF treatment effect. In daily practice, most patients (pts) receive NSAIDs on demand and reduce/stop their NSAID intake as soon as the anti-TNF agent is effective.
Objectives To compare the capacity of conventional clinical outcome measures and changes in NSAID therapy alone vs in combination to discriminate between treatment effects at 8 weeks.
Methods In a prospective, randomized, PBO-controlled trial (SPARSE), the NSAID-sparing effect of etanercept (ETN) was evaluated in pts with active axial SpA (axSpA). NSAID intake was evaluated using the ASAS-NSAID equivalent score.1 Post hoc logistic regression analyses were performed at week 8 for single conventional clinical measures: BASDAI50, BASDAI ≤3, ASAS20 and ASAS40, ASDAS remission, and ASDAS LDAS; single NSAID intake measures: decrease ≥50%, score <10, and score=0; and each clinical measure in combination (AND/OR) with each NSAID measure. No adjustments were made for multiple comparisons.
Results In 90 randomized pts, mean (±SD) NSAID score (ETN vs PBO, 98.2±39.0 vs 93.0±23.4), ASDAS-CRP (3.4±0.9 vs 3.2±0.8), and BASDAI (6.0±1.6 vs 5.9±1.5) were similar between groups at baseline (BL). Observed treatment effects for clinical and NSAID score measures alone and combined ranged from an odds ratio (OR) of 1.6 (95% CI 0.5-5.4; P=0.42) for ASDAS remission alone to 6.4 (95% CI 2.2-19.2; P=0.0008) for [ASDAS LDAS OR NSAID score decrease ≥50%], with most measures reaching statistical significance (α=0.05), in favor of ETN (32 of 45). Most combined outcome variables using “OR” were more discriminant than single outcome measures (see ASAS20 in table).
Conclusions This study suggests that 1) potential changes in NSAID intake during treatment do not prevent demonstration of a clinically relevant anti-TNF treatment effect; and 2) combined (ie, clinical+NSAID-sparing) outcome variables were at least as discriminant as single (ie, clinical) outcome variables. Such results may impact future clinical trial design.
References
-
Dougados M, et al. Ann Rheum Dis 2011;70(2):249-51.
Acknowledgements This study was sponsored by Pfizer Inc. Medical writing support was provided by Donna McGuire of Engage Scientific and was funded by Pfizer Inc.
Disclosure of Interest M. Dougados Grant/research support: AbbVie, Celgene, Eli Lilly, Novartis, Pfizer, Roche, and Sanofi-Aventis., Consultant for: AbbVie, Celgene, Eli Lilly, Novartis, Pfizer, Roche, and Sanofi-Aventis., E. Wood Consultant for: Full-time employee of Quanticate International, contracted by Pfizer Inc. to provide statistical input to the study and manuscript., L. Gossec Consultant for: AbbVie, Janssen, Pfizer, Roche and UCB., D. van der Heijde Grant/research support: AbbVie, Amgen, AstraZeneca, Bristol-Myers Squibb, Centocor, Chugai, Daiichi, Eli Lilly, GlaxoSmithKline, Janssen Biologics, Merck, Novartis, Novo Nordisk, Otsuka, Pfizer, Roche, Sanofi-Aventis, Schering-Plough, UCB, and Vertex., Consultant for: AbbVie, Amgen, AstraZeneca, Bristol-Myers Squibb, Centocor, Chugai, Daiichi, Eli Lilly, GlaxoSmithKline, Janssen Biologics, Merck, Novartis, Novo Nordisk, Otsuka, Pfizer, Roche, Sanofi-Aventis, Schering-Plough, UCB, and Vertex., Employee of: Director of Imaging Rheumatology bv., A. Dubanchet Shareholder of: Pfizer., Employee of: Pfizer., I. Logeart Shareholder of: Pfizer., Employee of: Pfizer.
DOI 10.1136/annrheumdis-2014-eular.1469