Background We performed a systematic review, a standard meta-analysis and network meta-analysis (NMA), which updates the 2009 Cochrane Overview, 'Biologics for rheumatoid arthritis (RA)'. This review is focused on biologic monotherapy in people with RA in whom treatment with traditional disease-modifying anti-rheumatic drugs (DMARDs) including methotrexate (MTX) had failed (MTX/other DMARD-experienced). Objectives To assess the benefits and harms of biologic monotherapy (includes anti-tumor necrosis factor (TNF) (adalimumab, certolizumab pegol, etanercept, golimumab, infliximab) or non-TNF (abatacept, anakinra, rituximab, tocilizumab)) or tofacitinib monotherapy (oral small molecule) versus comparator (placebo or MTX/other DMARDs) in adults with RA who were MTX/other DMARD-experienced. Methods We searched for randomized controlled trials (RCTs) in the Cochrane Central Register of Controlled Trials (CENTRAL; The Cochrane Library 2015, Issue 6, June), MEDLINE (viaOVID 1946 to June 2015), and Embase (via OVID1947 to June 2015). Article selection, data extraction and risk of bias and GRADE assessments were done in duplicate. We calculated direct estimates with 95% confidence intervals (CI) using standard meta-analysis. We used a Bayesian mixed treatment comparisons (MTC) approach for NMA estimates with 95% credible intervals (CrI). We converted odds ratios (OR) to risk ratios (RR) for ease of understanding. We calculated absolute measures as risk difference (RD) and number needed to treat for an additional beneficial outcome (NNTB). Main results This update includes 40 new RCTs for a total of 46 RCTs, of which 41 studies with 14,049 participants provided data. The comparator was placebo in 16 RCTs (4,532 patients), MTX or other DMARD in 13 RCTs (5,602 patients), and another biologic in 12 RCTs (3,915 patients). Monotherapy versus placebo Based on moderate-quality direct evidence, biologic monotherapy (without concurrent MTX/other DMARDs) was associated with a clinically meaningful and statistically significant improvement in American College of Rheumatology score (ACR50) and physical function, as measured by the Health Assessment Questionnaire (HAQ) versus placebo. RR was 4.68 for ACR50 (95% CI, 2.93 to 7.48); absolute benefit RD 23% (95% CI, 18% to 29%); and NNTB = 5 (95% CI, 3 to 8). The mean difference (MD) was -0.32 for HAQ (95% CI, -0.42 to -0.23; a negative sign represents greater HAQ improvement); absolute benefit of -10.7% (95% CI, -14% to 7.7%); and NNTB = 4 (95% CI, 3 to 5). Direct and NMA estimates for TNF biologic, non-TNF biologic or tofacitinib monotherapy showed similar results for ACR50, downgraded to moderate-quality evidence. Direct and NMA estimates for TNF biologic, anakinra or tofacitinib monotherapy showed a similar results for HAQ versus placebo with mostly moderate quality evidence. Based onmoderate-quality direct evidence, biologic monotherapy was associated with a clinically meaningful and statistically significant greater proportion of disease remission versus placebo with RR 1.12 (95% CI 1.03 to 1.22); absolute benefit 10% (95% CI, 3% to 17%; NNTB = 10 (95% CI, 8 to 21)). Based on low-quality direct evidence, results for biologic monotherapy for withdrawals due to adverse events and serious adverse events were inconclusive, with wide confidence intervals encompassing the null effect and evidence of an important increase. The direct estimate for TNF monotherapy for withdrawals due to adverse events showed a clinically meaningful and statistically significant result with RR 2.02 (95% CI, 1.08 to 3.78), absolute benefit RD 3% (95% CI, 1% to 4%), based on moderate-quality evidence. The NMA estimates for TNF biologic, non-TNF biologic, anakinra, or tofacitinib monotherapy for withdrawals due to adverse events and for serious adverse events were all inconclusive and downgraded to low-quality evidence. Monotherapy versus active comparator (MTX/other DMARDs) Based on direct evidence of moderate quality, biologic monotherapy (without concurrent MTX/other DMARDs) was associated with a clinically meaningful and statistically significant improvement in ACR50 and HAQ scores versus MTX/other DMARDs with a RR of 1.54 (95% CI, 1.14 to 2.08); absolute benefit 13% (95% CI, 2% to 23%), NNTB = 7 (95% CI, 4 to 26) and a mean difference in HAQ of -0.27 (95% CI, -0.40 to -0.14); absolute benefit of -9% (95% CI, -13.3% to -4.7%), NNTB = 2 (95% CI, 2 to 4). Direct and NMA estimates for TNF monotherapy and NMA estimate for non-TNF biologic monotherapy for ACR50 showed similar results, based on moderate-quality evidence. Direct and NMA estimates for non-TNF biologic monotherapy, but not TNF monotherapy, showed similar HAQ improvements, based on mostly moderate-quality evidence. There were no statistically significant or clinically meaningful differences for direct estimates of biologic monotherapy versus active comparator for RA disease remission. NMA estimates showed a statistically significant and clinically meaningful difference versus active comparator for TNF monotherapy (absolute improvement 7% (95% CI, 2% to 14%)) and non-TNF monotherapy (absolute improvement 19% (95% CrI, 7% to 36%)), both downgraded to moderate quality. Based on moderate-quality direct evidence from a single study, radiographic progression (scale 0 to 448) was statistically significantly reduced in those on biologic monotherapy versus active comparator, MD-4.34 (95% CI, -7.56 to -1.12), though the absolute reduction was small, -0.97% (95% CI, -1.69% to -0.25%). We are not sure of the clinical relevance of this reduction. Direct and NMA evidence (downgraded to low quality), showed inconclusive results for withdrawals due to adverse events, serious adverse events and cancer, with wide confidence intervals encompassing the null effect and evidence of an important increase. Authors' conclusions Based mostly on RCTs of six to 12-month duration in people with RA who had previously experienced and failed treatment with MTX/other DMARDs, biologic monotherapy improved ACR50, function and RA remission rates compared to placebo or MTX/other DMARDs. Radiographic progression was reduced versus active comparator, although the clinical significance was unclear. Results were inconclusive for whether biologic monotherapy was associated with an increased risk of withdrawals due to adverse events, serious adverse events or cancer, versus placebo (no data on cancer) or MTX/other DMARDs.
