Background Biological DMARDs are recommended by EULAR for patients with an inadequate response (IR) to MTX or anti-TNF. Health Technology Assessments have classified the evidence provided by indirect or mixed treatment comparisons(ITC/MTC) as too uncertain for expressing a preference across agents. The studies included in the ITC/MTCs have been considered too different in terms of included patients and study designs. Furthermore, patient characteristics and designs evolved, hindering comparability between studies over time. Consequently, it will be crucial for ITC/MTCto adequately address uncertainty resulting from heterogeneity and bias across trials.
Objectives To perform a critical appraisal of the quality of available ITC/MTCs and evaluate whether these would allow decision makers or prescribers to discriminate between drugs.
Methods A systematic literature search (2002–2011) was performed in the Cochrane library, Medline and Embase. The quality of the identified ITC/MTCs was assessed following PRISMA statement.1
Results The search identified 13 ITC/MTCs. ITCs (n=5) and Bayesian MTCs (n=7) primarily evaluated relative differences in efficacy up to 1 year on ACR response rate. Comparability across ITC/MTCs is limited, as these consider different types of RA patients: all types (n=3), those failing DMARD or anti-TNF (n=8) or those who did not fail on biologics (n=1). Three MTCs considering all RA patients and 1 in patients who did not fail on biologics reported anakinra as the least effective agent. However, as these MTCs used a meta-regression design (n=3) or were at high risk for confounding (n=1) they are exploratory by nature and caution any definite conclusions. The ITCs (n=5)/MTCs (n=3) that focused on DMARD-IR or anti-TNF IR patients did not report consistent results regarding relative differences in short-term efficacy across agents. Five MTCs/ITCs did not perform any risk assessment for bias. Individual studies included in the MTCs/ITCs were conducted at different time periods or in different regions, with different designs (e.g. early escape/rescue therapy) that can lead to heterogeneity across studies. Although the risk of heterogeneity is expected to be high, six studies either ignored or addressed this issue inadequately. One ITC did not report any differences among agents after anti-TNF failure. Conversely, the recent safety Cochrane MTC met all quality criteria and reported a significantly lower risk of serious adverse events for abatacept and anakinra vs. most biologics and a lower risk for abatacept for serious infections relative to infliximab, certolizumab and tocilizumab.2
Conclusions The review suggests that it is unlikely that decision makers and prescribers can discriminate between agents on short-term efficacy as none of the efficacy-focused ITC/MTCs satisfactorily addressed prior concerns regarding heterogeneity and bias. However, the Cochrane review on safety did suggest a favorable safety profile for abatacept over biologics with regards to risk for serious infections.
Moher D. et al. Phys Ther 2009; 89: 873–80.
Singh JA, et al. CMAJ 2009; 181: 787–96.
Disclosure of Interest M. Schiff Consultant for: Bristol-Myers Squibb, W. Stam: None Declared, I. Gilloteau Employee of: Bristol-Myers Squibb, C. Poncet: None Declared, M. Hochberg Consultant for: Abbott Laboratories, Amgen Inc., BMS, Genentech and Biogen IDEC Inc, Roche, UCB, Inc., AstraZeneca, Bayer, Bioberica, Combinatorx, Eli Lilly and Company, Endo Pharmaceuticals Inc., GlaxoSmithKline, Merck Pharmaceuticals, Merck Serono International, NicOx, Novartis Pharmaceuticals Corporation, POZEN Inc., Pfizer Inc, sanofi-aventis