The number needed to treat for second-generation biologics when treating established rheumatoid arthritis: a systematic quantitative review of randomized controlled trials

• Published in: Scandinavian journal of rheumatology 2011-01, Vol.40 (1), p.1-7
• Main Authors: Kristensen, LE, Jakobsen, AK, Bartels, EM, Geborek, P, Bliddal, H, Saxne, T, Danneskiold-Samsøe, B, Christensen, R
• Format: Article
• Language: English
• Subjects: Adult; Antibodies, Monoclonal - therapeutic use; Antirheumatic Agents - therapeutic use; Arthritis, Rheumatoid - drug therapy; Biological and medical sciences; Biological Products - therapeutic use; Clinical Medicine; Databases, Bibliographic; Diseases of the osteoarticular system; Double-Blind Method; Female; Humans; Inflammatory joint diseases; Klinisk medicin; Male; Medical and Health Sciences; Medical sciences; Medicin och hälsovetenskap; Middle Aged; Randomized Controlled Trials as Topic; Research Design; Reumatologi och inflammation; Rheumatology and Autoimmunity; Sample Size
• ISSN: 0300-9742; 1502-7732; 1502-7732
• DOI: 10.3109/03009742.2010.491834

Objective: To evaluate the number needed to treat (NNT) and the number needed to harm (NNH) of the second-generation biologics abatacept, certolizumab, golimumab, rituximab, and tocilizumab in patients with established rheumatoid arthritis (RA) taking concomitant methotrexate (MTX). Methods: A systematic literature search of MEDLINE, EMBASE, Web of Science, and the Cochrane Register of Controlled Trials was conducted up to 1 November 2009. We selected any published randomized, double-blind, MTX-controlled study including RA patients with a mean disease duration of at least 5 years before entering a pivotal trial on second-generation biological therapy. Studies eligible for inclusion involved patients, who had previously shown inadequate response to conventional disease-modifying anti-rheumatic drug (DMARD) therapy. Pre-specified binary outcomes were extracted with a preference for 1-year data (6-month data were used if no data were available for 1 year). Two reviewers independently extracted the data necessary to estimate the absolute measures in a non-responder intention-to-treat (ITT) analysis. Results: Five randomized controlled trials, one for each of the drugs, were selected and data extracted according to published data at endpoint for American College of Rheumatology 50% (ACR50)-responding patients, and withdrawals due to adverse events. NNT ranged from four to six treated patients to achieve one ACR50 response, while withdrawals due to adverse events were few and non-significant compared to the placebo group, except for rituximab administered as 1000 mg. Conclusion: Comparable efficacy was shown by the five biological agents studied, with few adverse events. However, for rituximab, tocilizumab, and golimumab, only 6-month data were available, hampering the external validity with regard to long-term efficacy and tolerability. A low dose (500 mg) of rituximab may be as effective as the recommended dose of 1000 mg.