Confidential: For Review Only · Date Submitted by the Author: 25-Mar-2015 Complete List of Authors: Hazlewood, Glen; University of Calgary, Medicine ... Arthritis, Rheumatoid (MeSH),

Confidential: For Review O

nly

Methotrexate monotherapy and methotrexate combination

therapy with traditional and biologic disease modifying anti-rheumatic drugs for rheumatoid arthritis: abridged

Cochrane systematic review and network meta-analysis

Journal: BMJ

Manuscript ID BMJ.2015.026116

Article Type: Research

BMJ Journal: BMJ

Date Submitted by the Author: 25-Mar-2015

Complete List of Authors: Hazlewood, Glen; University of Calgary, Medicine Barnabe, Cheryl; University of Calgary, Medicine Tomlinson, George; University of Toronto, Medicine Marshall, Deborah; University of Calgary, Community Health Sciences Devoe, Dan; University of Calgary, Community Health Sciences Bombardier, Claire; University of Toronto, Medicine

Keywords:

Arthritis, Rheumatoid (MeSH), Methotrexate (MeSH), Antirheumatic agents

(MeSH), Randomized Controlled Trial (MeSH), network meta-analysis, biologic therapy

https://mc.manuscriptcentral.com/bmj

BMJ


nly

1

Methotrexate monotherapy and methotrexate combination therapy with traditional

and biologic disease modifying anti-rheumatic drugs for rheumatoid arthritis:

abridged Cochrane systematic review and network meta-analysis

Glen S Hazlewood, Cheryl Barnabe, George Tomlinson, Deborah Marshall, Dan Devoe,

Claire Bombardier

Glen S Hazlewood, clinical assistant professor1,2,3

, Cheryl Barnabe, assistant

professor1,2,4

, George Tomlinson, scientist5, Deborah Marshall, associate professor

2,4,

Dan Devoe, research assistant4, Claire Bombardier, professor

5,6

Affiliations: 1Department of Medicine, University of Calgary, Calgary, Canada, T2N4Z6

2McCaig Institute for Bone and Joint Health, University of Calgary, Calgary, Canada,

T2N4Z6 3Institute of Health, Policy, Management and Evaluation, University of Toronto, Toronto,

Canada, M5T3M6 4Department of Community Health Sciences, University of Calgary, Calgary, Canada,

T2N4Z6 5Department of Medicine, University Health Network and Mount Sinai Hospital,

University of Toronto, Toronto, Canada, M5G2C4 6Institute for Work and Health, University of Toronto, Toronto, Canada, M5G2E9

Correspondence to:

Glen S Hazlewood

3330 Hospital Dr NW, Calgary, AB, T2N1N1

(tel): 403-220-5903 (fax): 403-210-7367

email: [email protected]

Keywords: Arthritis, Rheumatoid (MeSH); Methotrexate (MeSH); Antirheumatic agents

(MeSH); Randomized Controlled Trial (MeSH); network meta-analysis

Word count: 3984

Page 1 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

2

ABSTRACT

Objective: To compare methotrexate based disease-modifying anti-rheumatic drug

(DMARD) treatments for rheumatoid arthritis in patients naïve to or after an inadequate

response (IR) to methotrexate.

Design: Systematic review and Bayesian random-effects network meta-analysis of trials

assessing methotrexate used alone or in combination with other conventional synthetic

DMARDs (csDMARDs), biologics (bDMARDs) or tofacitinib in adult patients with

rheumatoid arthritis. The major outcomes were ACR50 response (major clinical

improvement), radiographic progression and withdrawals due to adverse events (WDAE).

The risk of bias of each study was evaluated using the Cochrane risk of bias tool,

separately for each outcome. Trials judged at high overall risk of bias were excluded.

Data sources: Trials were identified from MEDLINE, EMBASE and CENTRAL

databases from inception to August 13, 2014, abstracts from 2 major rheumatology

meetings from 2009-2014, 2 trial registers, and hand-searches of Cochrane reviews.

Eligibility criteria for selecting studies: Randomised or quasi-randomised trials that

compared methotrexate to any other DMARD or DMARD combination and contributed

to the network of evidence between the treatments of interest.

Results: 150 trials with over 34,000 patients were included. In methotrexate -naïve

patients, ‘triple therapy’ (the combination of methotrexate + sulphasalazine +

hydroxychloroquine), and methotrexate + several bDMARDs (abatacept, adalimumab,

etanercept, infliximab, rituximab, tocilizumab) were similar to each other and

significantly superior to oral methotrexate for ACR50 response. Only methotrexate +

adalimumab and methotrexate + etanercept were significantly superior to oral MTX for

inhibiting radiographic progression. Triple therapy had significantly fewer WDAE

compared to methotrexate + infliximab. In methotrexate-IR patients, several treatments

were significantly superior to oral methotrexate for ACR50 response, including

csDMARD combinations (triple therapy, methotrexate + hydroxychloroquine,

methotrexate + leflunomide, methotrexate + intramuscular gold), methotrexate + most

bDMARDs, and methotrexate + tofacitinib. No treatment was significantly superior to

oral methotrexate for inhibiting radiographic progression. Methotrexate + abatacept had

fewer WDAE than several treatments.

Conclusions: Triple therapy and most regimens combining biologic DMARDs with

methotrexate were similarly effective in controlling disease activity and all generally well

tolerated in both methotrexate-naive and methotrexate-exposed patients. Given cost

considerations, these findings support a therapeutic trial of low-cost triple therapy prior to

using biologic DMARDs.

Page 2 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

3

INTRODUCTION

Methotrexate (MTX)-based treatments form the core of rheumatoid arthritis (RA)

treatment. MTX is recommended as the first disease-modifying anti-rheumatic drug

(DMARD) for most patients with RA, and MTX co-prescription is recommended when

using biologic DMARDs (bDMARDs) or the recently approved tofacitinib.1 2 Combining

MTX with other conventional synthetic DMARDs (csDMARDs), however, is more

controversial. A trial of csDMARD combination therapy prior to biologic therapy is not

currently recommended by either major rheumatology guideline, although each provides

the option.1 2 Understanding the comparative benefits and harms of these treatments is

essential to inform decision-making, as bDMARD therapy and tofacitinib costs 10-20

times that of MTX and most csDMARDs.

Network (mixed treatment) meta-analyses are a natural avenue of comparative

effectiveness research, as they combine all direct and indirect evidence to estimate

treatment effects between all treatments of interest.3 If treatments A and B are in the same

study, there is direct evidence linking A and B. If they are compared in separate studies

to a common comparator C, then the A-C and B-C studies allow an indirect comparison

of A and B. Longer chains of indirect comparisons (A-B, B-C, C-D) are also possible.

Considering indirect evidence is critical if a treatment decision must be made and the

treatments have not been directly compared in a head-to-head trial. Indirect evidence is

also important to consider when head-to-head trials are available, as it adds to the entire

body of evidence and may help refine the precision in estimation of the treatment effect.3

A previous Cochrane network meta-analysis examined the relative effects of

different biologic therapies through indirect comparisons, and found some differences

between agents.4 Our review expands on this, by including combination therapy with

MTX + csDMARDs. A previous Cochrane traditional (non-network) meta-analysis did

not find an additional overall benefit with combination therapy over MTX alone.5 By

including indirect evidence we expand the evidence base to draw from. For example,

three trials have been recently published that have compared combination therapy with

MTX + sulphasalazine + hydroxychloroquine versus MTX + anti-TNF therapy.6-8 The

inclusion of these trials in a network meta-analysis adds indirect evidence on the relative

effects of MTX + sulphasalazine + hydroxychloroquine compared to all other treatments

in the network.

METHODS

This is an abridged version of a Cochrane systematic review. The protocol and

Cochrane review can be accessed from The Cochrane Library.9[placeholder for reference

to review]

Inclusion criteria

Study type

Page 3 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

4

We included randomised or quasi-randomised clinical trials of at least 12 weeks

duration that contained any intervention of interest (defined in detail below) and could be

linked within the network through a shared comparator. For example, if we identified a

trial comparing MTX to hydroxychloroquine, the trial was included if another trial was

available that compared hydroxychloroquine to any other treatment of interest. Trials

were divided into groups, based on patients’ prior exposure to MTX: either MTX-naïve

or MTX-inadequate response (IR). Trials that required all patients to have failed anti-

TNF therapy were excluded.

Participants

We included studies of adults (age > 18 years) with RA, according to 1958, 1987

or 2010 classification criteria.10-12

Interventions

The interventions of interest were: oral MTX; parenteral (intramuscular or

subcutaneous) MTX; MTX in combination with csDMARD therapy including anti-

malarials (hydroxychloroquine/chloroquine), sulphasalazine, leflunomide, cyclosporine,

intra-muscular gold and azathioprine; MTX in combination with bDMARD therapy

(adalimumab, certolizumab, etanercept, golimumab, infliximab, abatacept, rituximab,

tocilizumab); methotrexate in combination with tofacitinib. No dose restriction was

applied to csDMARDs, given the variability of dosing in clinical practice. Biologic

DMARDs and tofacitinib were limited to currently recommended doses or dose

equivalents.

Outcomes

The major outcomes of the review were: 1) American College of Rheumatology

(ACR)-50 response, a composite measure of improvement in disease activity;13 2)

Radiographic progression, measured by Larsen, Sharp or modified Larsen/Sharp scores;14

and 3) Withdrawals due to adverse events, including death. Multiple secondary outcomes

were evaluated and are reported in the unabridged Cochrane review.[placeholder for

reference]

Search methods

An electronic database search was performed in MEDLINE, EMBASE and

Cochrane CENTRAL from inception to August 13, 2014. The search strategies contained

subject headings and keywords for "rheumatoid arthritis", "methotrexate" and

"randomised controlled trial". The search strategy is available in the unabridged

Cochrane review.[placeholder for reference] We also searched the trial registries

ClinicalTrials.gov (http://clinicaltrials.gov/) and the International Clinical Trials Registry

Platform (http://apps.who.int/trialsearch/) using the search term "rheumatoid arthritis

AND methotrexate". Finally, we performed hand-searches for abstracts from American

College of Rheumatology (2009-2013) and European League Against Rheumatism

Page 4 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

5

(2009-2014) conferences and reviewed existing Cochrane reviews. All languages were

included.

Study selection

Two review authors (GH, ChB) independently screened articles for inclusion by

title or abstract and full-text if necessary. Disagreements were resolved by consensus or

discussion with a third review author (ClB).

Data extraction and handling of missing data

Three review authors working in pairs (GH, ChB, DD) abstracted relevant data

from included studies into an Excel spreadsheet. Trial characteristics and baseline patient

characteristics were extracted by one author (GH) and confirmed by a second (ChB or

DD); outcome data were extracted independently, with disagreements resolved through

discussion. For all trials, we also sought data from clinical trial registries

(www.clinicaltrials.gov, http://www.who.int/ictrp/en/), US Food and Drug

Administration, European Medicines Agency and drug manufacturer websites. For

continuous measures, if standard deviations (sd) were not available, they were calculated

from available variance data (e.g. – standard errors) if possible. For toxicity outcomes, if

the drug exposure was not available, it was calculated; withdrawals were assumed to

occur at a constant rate, unless specific information was available to permit a more

accurate calculation. If data were presented only in graphical format, numbers were

extracted digitally by 2 independent reviewers (GH, ChB) and averaged or corrected if an

obvious discrepancy was apparent.

Time point of outcome assessment

For parallel group RCTs, we used end-of-trial data for all outcomes. For crossover

trials or trials comparing different DMARD sequences we used data at the time of the

initial switch in treatment for efficacy outcomes (ACR50 response and radiographic

progression). For WDAE of crossover and strategy trials, we abstracted data from the

longest follow-up available, provided the AE was assigned to the treatment received at

the time of the AE (on-treatment data). In certain trials, a switch in treatment was allowed

or required for patients who failed to achieve a certain response at a given time (usually

12-16 weeks). Patients who receive early escape therapy are typically treated as treatment

failures at all future time points. For early-escape trials we used end-of-trial (carried

forward) data for ACR50 and radiographic progression. For WDAE we used on-

treatment data from the longest follow-up available, similar to the approach for crossover

trials.

Risk of bias of studies

The methodological quality of included trials was assessed using the Cochrane

Collaboration's tool for assessing risk of bias independently by 2 review authors, or by

one author (GH) if his assessment agreed with a published Cochrane review. Studies

Page 5 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

6

were graded as having a "low risk", "high risk" or "unclear risk" of bias across the seven

specified domains.15 The risk of bias was assessed separately for the 3 major outcomes

for domains where the risk of bias could differ. The domains “blinding of participants”

and “blinding of outcome assessment” were assessed separately for radiographic

progression; “incomplete outcome data” was assessed separately for each of the 3 major

outcomes. For each of the 3 major outcomes we also judged an overall risk of bias.

Data analysis

For the primary analyses, trials with a high risk of bias for that outcome were

excluded. Treatment effects were evaluated using odds ratios (ORs) for ACR50 response

and WDAE, and a standardised mean difference (SMD) for radiographic progression. For

WDAE, we used the total exposure (in patient-months) in each arm as the denominator,

to allow for differences in exposure between arms in early escape and crossover trials.

Random-effects Bayesian network meta-analyses were fitted for each outcome

measure.16 17

We reported the posterior median and 95% credible interval (CrI) for all

treatments effects relative to oral MTX. We considered an effect to be significant if the

95% CrI excluded the null effect. Uninformative prior probability distributions were used

for all parameters. We used 10,000 burn-in iterations followed by 10,000 monitoring

iterations. Convergence was assessed by running 3 chains, inspecting the sampling

history plots and calculating Gelman–Rubin–Brooks (GBR) statistics.18 Model fit was

assessed using residual deviance and Deviance Information Criterion (DIC). All data

analyses were performed using R statistical software version 3.1.2 (www.r-project.org)

with rjags package version 3-14 running Just Another Gibbs Sampler (JAGS) version

3.4.0.19

Presentation of results and assessment of quality of evidence

The treatment effect for each outcome was converted into an absolute response

using an assumed (baseline) risk for oral MTX. The assumed risk was the point estimate

from a Bayesian random-effects model of the oral MTX arms across the trials for that

analysis. We used recently published GRADE (Grading of Recommendations

Assessment, Development and Evaluation) guidance to assess the quality of evidence

from a network meta-analysis.20 This approach considers the quality of both the direct

and indirect evidence as well as the likelihood of ‘intransitivity’, which exists if there is

heterogeneity in the trials that form the different comparisons within the indirect

evidence. We performed ‘node-splitting’ to separate the indirect evidence from the direct

evidence to inform these evaluations.21

Meta-regression and sensitivity analyses

We performed meta-regression for several pre-specified trial-level characteristics.

Selected post-hoc sensitivity analyses were added, including fixed-effect models for the

major outcome ‘radiographic progression’, as there were few trials available to estimate a

random-effect. Full details of these analyses are provided in the unabridged review.

Page 6 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Confidential: For Review Only

7

RESULTS

Search results and description of included studies

From 8584 unique records, we identified 150 trials including over 34,000 patients meeting our inclusion criteria (Figure 1).

MTX dosing varied across studies and was variably reported (Table 1). Most trials enrolled patients with high disease activity (median

baseline swollen joint count 15.3), with a similar distribution across medication classes (Table 1).

Table 1: Summary of trial characteristics Overall risk of bias for ACR50 response,

% of studies (n pts)

Medication

Studies (N) Patients (n) Year published,

median (range)

Early escape

design, % of

studies (n pts)

Trial duration*,

wks, median

(range)

MTX dose

>15mg/wk**, %

of studies (n pts)

Disease

duration, yrs,

median

(range)

Swollen joint

count, median

(range)

Low Unclear High

MTX + bDMARDs/tofacitinib

MTX + Etanercept 10 2448 2007 (1999-2014) 0% 38 (12-52) 50% (n=1833) 2 (0.5-13) 13.9 (11-24) 10% (n=424) 70% (n=1965) 20%

(n=59)

MTX + Infliximab 12 2745 2006 (2000-2013) 8% (n=264) 28 (13-54) 50% (n=1990) 7.6 (0.4-10.5) 15 (5-21.5) 25% (n=837) 58% (n=1824) 17%

(n=84)

MTX + Adalimumab 14 3527 2009 (2003-2014) 29% (n=998) 24 (12-104) 50% (n=1689) 0.7 (0.1-11.7) 14.4 (8.7-22.5) 14% (n=1204) 79% (n=2258) 7%

(n=65)

MTX + Rituximab 4 1262 2008 (2004-2012) 25% (n=342) 24 (24-104) 50% (n=683) 8.6 (0.9-11.5) 20.9 (20.2-21.6) 25% (n=80) 75% (n=1182) 0%

MTX + Abatacept 9 3377 2011 (2005-2014) 0% 24 (17-52) 67% (n=3014) 6.9 (0.5-9.3) 18.6 (10-22.4) 56% (n=2261) 44% (n=1116) 0%

MTX + Tocilizumab 10 4859 2012 (2006-2014) 50% (n=3671) 24 (16-52) 60% (n=2729) 6.6 (0.4-9.2) 14.8 (6.4-20.1) 10% (n=553) 60% (n=2765) 30%

(n=1541)

MTX + Certolizumab 6 1804 2012 (2008-2014) 67% (n=1245) 24 (24-52) 17% (n=243) 6 (0.3-9.6) 21 (16.4-22.5) 0% 17% (n=316) 83%

(n=1488)

MTX + Golimumab 6 1640 2010 (2008-2013) 83% (n=1570) 24 (16-52) 50% (n=1132) 6.9 (3.2-8.7) 13.5 (11.6-15.4) 0% 100% (n=1640) 0%

MTX + Tofacitinib 3 676 2012 (2011-2013) 67% (n=621) 24 (12-24) 67% (n=621) 9 (8.4-9.1) 14.7 (14.1-14.9) 0% 67% (n=195) 33%

(n=481)

SUBTOTAL 74 22338 2010 (1999-2014) 30% (n=8711) 24 (12-104) 51% (n=13934) 6.4 (0.1-13) 16.6 (5-24) 18% (n=5359) 64% (n=13261) 19%

(n=3718)

Comparative effectiveness trials

Head to head bDMARDs/

tofacitinib

4 1658 2010 (2006-2014) 25% (n=501) 27 (26-104) 50% (n=1077) 7.8 (1.8-11.3) 16.6 (15.9-20.6) 25% (n=431) 50% (n=1147) 25%

(n=80)

MTX + bDMARDs vs. MTX +

csDMARDs

4 1382 2012 (2012-2013) 0% 63 (16-104) 25% (n=353) 0.5 (0.3-5.2) 12 (11.2-12.8) 50% (n=1108) 0% 50%

(n=274)

SUBTOTAL 8 3040 2012 (2006-2014) 12% (n=501) 27 (16-104) 38% (n=1430) 5.2 (0.3-11.3) 15.9 (11.2-20.6) 38% (n=1539) 25% (n=1147) 38%

(n=354)

MTX + csDMARDs

MTX + azathioprine 1 209 1992 0% 24 0% 8.6 17.3 0% 100% (n=209) 0%

MTX + hydroxychloroquine/

chloroquine

7 452 2005 (1993-2012) 0% 26 (12-52) 0% 4 (0.3-7.7) 10.7 (8.2-27.3) 14% (n=82) 14% (n=40) 71%

(n=330)

MTX + sulphasalazine 6 639 2000 (1994-2007) 0% 52 (24-76) 0% 0.4 (0.2-5) 16.7 (9.8-22.6) 0% 67% (n=515) 33%

(n=124)

MTX + cyclosporine 9 1100 2003 (1995-2008) 11% (n=120) 48 (16-104) 22% (n=64) 1.1 (0.2-10.3) 13.6 (11-19) 0% 89% (n=1076) 11%

(n=24)

Page 7 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


8

MTX + sulphasalazine +

hydroxychloroquine

4 503 2005 (1996-2013) 0% 91 (13-104) 0% 4.4 (0.5-8.6) 17.1 (9.5-29.8) 0% 75% (n=463) 25%

(n=40)

MTX + leflunomide 2 729 2004 (2002-2006) 0% 30 (24-36) 50% (n=263) 11.6 18 0% 50% (n=263) 50%

(n=466)

MTX + IM Gold 1 65 2005 0% 48 100% (n=65) 3.2 11 0% 100% (n=65) 0%

SUBTOTAL 30 3697 2003 (1992-2013) 3% (n=120) 48 (12-104) 13% (n=392) 1.2 (0.2-11.6) 14.5 (8.2-29.8) 3% (n=82) 63% (n=2631) 33%

(n=984)

MTX vs csDMARD monotherapy

Placebo 5 324 1985 (1985-1993) 20% (n=52) 14 (12-18) 0% 9.5 (4.8-14) 27.5 (24-30.9) 0% 100% (n=324) 0%

Azathioprine 5 257 1991 (1987-1994) 0% 24 (24-52) 0% 12 (8.7-13.9) 14.6 (9.5-21.9) 0% 40% (n=106) 60%

(n=151)

IM Gold 5 489 1991 (1988-2001) 0% 48 (26-52) 20% (n=99) 4 (1.2-6) 14 (13.9-15.2) 0% 60% (n=249) 40%

(n=240)

Sulphasalazine 2 211 1998 (1995-2002) 0% 24 0% 4 (1.4-6.6) 9.3 0% 50% (n=126) 50%

(n=85)

Cyclosporine 2 203 1999 (1998-2000) 0% 65 (26-104) 50% (n=100) 3.8 (2.2-5.5) 13.1 (12.2-14) 0% 0% 100%

(n=203)

Leflunomide 15 3218 2002 (1999-2012) 13% (n=567) 24 (12-52) 27% (n=927) 3.9 (0.5-6.8) 11.8 (8.2-16.5) 0% 53% (n=1598) 47%

(n=1620)

Hydroxychloroquine 2 409 2006 (2000-2012) 0% 24 0% 1.5 (1-2.1) NA 0% 0% 100%

(n=409)

sc vs. oral MTX 2 467 2009 (2008-2010) 50% (n=383) 24 100% (n=467) 0.2 15 0% 50% (n=383) 50%

(n=84)

SUBTOTAL 38 5578 2000 (1985-2012) 11% (n=1002) 24 (12-104) 21% (n=1593) 4.5 (0.2-14) 14 (8.2-30.9) 0% 53% (n=2786) 47%

(n=2792)

TOTAL 150 34653 2007 (1985-2014) 19% (n=10334) 24 (12-104) 35% (n=17349) 5 (0.1-14) 15.3 (5-30.9) 11% (n=6980) 59% (n=19825) 30%

(n=7848)

Abbreviations: DMARD, disease-modifying anti-rheumatic drug; bDMARD, biologic DMARD; csDMARD, conventional synthetic DMARD; tsDMARD, targeted synthetic DMARD; IM, intra-muscular; MTX, methotrexate; pts, patients; sc, subcutaneous; SJC, swollen joint count

Trials are grouped by comparator and sorted by the year of the first trial within each class, for illustrative purposes. Patient characteristics, including the number of patients, only include the

arms considered in the review. *Trial duration for efficacy outcomes; some studies had longer follow-up for safety outcomes **Studies where the dose of MTX could be confirmed as >= 15

mg/wk. In some studies, MTX was dosed across a range of values that included 15 mg/wk but the actual dose was not provided.

Page 8 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


9

In the MTX-naïve network (Figure 2A), most comparisons of MTX + bDMARDs were to MTX, with no head-to-head

comparisons between different bDMARDs. Trials evaluating csDMARD therapy were generally smaller but more inter-connected

than trials of bDMARDs. In the MTX-IR network (Figure 2B), connections between MTX + bDMARDs and MTX were large in size

(more patients), whereas the connections between csDMARDs were few and small in size. Four head-to-head trials of biologic therapy

formed links between several biologic therapies and all four trials that compared MTX + biologic therapy to MTX + csDMARDs were

included in this network. The network diagrams show all trials; the actual numbers of trials for each outcome varied and are reported

below.

Methodological quality of included studies

The risk of bias of the trials varied considerably across each domain (Figure 3). The overall risk of bias was high in 30%, 22%,

and 17% of trials for ACR50 response, radiographic progression, and WDAE respectively. These trials were excluded from the

primary analysis.

Effects of interventions: major outcomes

MTX-naïve trials

ACR50: Twenty-six trials with 9513 patients were included in this analysis. The combination of MTX + the bDMARDs abatacept,

adalimumab, etanercept, infliximab, rituximab and tocilizumab (8 mg/kg) were superior to oral MTX (Table 2). MTX + golimumab

and MTX + tocilizumab (4 mg/kg) were not significantly different from oral MTX. There was no data available for certolizumab or

tofacitinib. MTX+ sulphasalazine (SSZ) + hydroxychloroquine (HCQ) was the only csDMARD combination that was significantly

superior to oral MTX (Table 2). This comparison was based on indirect evidence, and was judged to be ‘moderate’ quality. In pairwise

comparisons, there was no significant difference between MTX+SSZ+HCQ and MTX + any bDMARD (Appendix A, Table 1).

Page 9 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


10

Table 2: Summary of findings Intervention Assumed risk Corresponding risk Treatment effect relative

to oral methotrexate

(95%CrI)

Quality of the evidence

MTX-naïve

ACR50 (26 studies; 9513 patients)

MTX + abatacept (IV) 399 per 1000 550 per 1000 (401 to 689) OR 1.8 (1.0 to 3.3) High

MTX + adalimumab 399 per 1000 583 per 1000 (504 to 655) OR 2.1 (1.5 to 2.9) High

IM/sc MTX + adalimumab 399 per 1000 595 per 1000 (352 to 801) OR 2.2 (0.82 to 6.0) Moderate (imprecision)

MTX + etanercept 399 per 1000 665 per 1000 (576 to 754) OR 3.0 (2.1 to 4.6) Moderate (indirectness)

MTX + golimumab (sc) 399 per 1000 472 per 1000 (314 to 631) OR 1.3 (0.69 to 2.6) Moderate (study limitations)

MTX + infliximab 399 per 1000 575 per 1000 (460 to 717) OR 2.0 (1.3 to 3.8) High

MTX + rituximab 399 per 1000 618 per 1000 (466 to 744) OR 2.4 (1.3 to 4.4) High

MTX + tocilizumab (4 mg/kg) 399 per 1000 521 per 1000 (388 to 653) OR 1.6 (0.96 to 2.8) Moderate (study limitations)

MTX + tocilizumab (8 mg/kg) 399 per 1000 558 per 1000 (421 to 686) OR 1.9 (1.1 to 3.3) Moderate (study limitations)

MTX + cyclosporine 399 per 1000 539 per 1000 (379 to 699) OR 1.8 (0.92 to 3.5) Low (indirectness, imprecision,

study limitations)

IM/sc MTX + cyclosporine 399 per 1000 504 per 1000 (220 to 785) OR 1.5 (0.42 to 5.5)

Low (imprecision, study

limitations)

MTX + hydroxychloroquine/ chloroquine 399 per 1000 357 per 1000 (135 to 657) OR 0.84 (0.24 to 2.9) Moderate (imprecision)

MTX + sulphasalazine 399 per 1000 429 per 1000 (225 to 650) OR 1.1 (0.44 to 2.8)

Low (indirectness, imprecision,

study limitations)

MTX + sulphasalazine + hydroxychloroquine 399 per 1000 607 per 1000 (444 to 762) OR 2.3 (1.2 to 4.8) Moderate (study limitations)

IM/sc MTX 399 per 1000 426 per 1000 (283 to 584) OR 1.1 (0.59 to 2.1) Moderate (study limitations)

Radiographic progression (15 studies; 6540

patients)

MTX + abatacept (IV)

Mean change 2.6 points on

Sharp-VdH scale over 1 year

Mean change 1.3 points (-

1.7 to 4.3) SMD -0.20 (-0.65 to 0.25)

Moderate (imprecision)

MTX + adalimumab Mean change 2.6 points on Sharp-VdH scale over 1 year

Mean change 0.16 points (-1.9 to 2.4) SMD -0.37 (-0.68 to -0.03)

High

MTX + etanercept




1.4 to 2.2) SMD -0.36 (-0.61 to -0.07)

High

MTX + golimumab (sc) Mean change 2.6 points on Sharp-VdH scale over 1 year

Mean change 1.8 points (-1.2 to 4.9) SMD -0.13 (-0.57 to 0.34)

Low (study limitations, imprecision)

MTX + infliximab



Mean change -0.22 points

(-3.2 to 2.7) SMD -0.43 (-0.87 to 0.002)


MTX + rituximab Mean change 2.6 points on Sharp-VdH scale over 1 year

Mean change 0.13 points (-3 to 3.1) SMD -0.38 (-0.84 to 0.07)


MTX + tocilizumab (4 mg/kg)



Mean change 1 points (-2

to 4.1) SMD -0.24 (-0.70 to 0.22)

Low (study limitations,

imprecision)

MTX + tocilizumab (8 mg/kg)




2.7 to 3.3) SMD -0.36 (-0.80 to 0.10)


imprecision)

MTX + cyclosporine




0.91 to 3.6) SMD -0.21 (-0.53 to 0.14)


imprecision)

Page 10 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


11

MTX + sulphasalazine + hydroxychloroquine




2.5 to 7.9) SMD -0.02 (-0.78 to 0.79)


Withdrawals due to adverse events (35

studies; 10250 pt-years)



IM/sc MTX + adalimumab 76 per 1000 59 per 1000 (4.9 to 402) OR 0.77 (0.06 to 8.2) Moderate (imprecision)

MTX + etanercept 76 per 1000 61 per 1000 (35 to 119) OR 0.80 (0.45 to 1.6) High

MTX + golimumab (sc) 76 per 1000 160 per 1000 (52 to 430) OR 2.3 (0.67 to 9.2)


imprecision)

MTX + infliximab 76 per 1000 171 per 1000 (71 to 377) OR 2.5 (0.94 to 7.4) Moderate (imprecision)


MTX + tocilizumab (4 mg/kg) 76 per 1000 98 per 1000 (37 to 239) OR 1.3 (0.47 to 3.8) Low (study limitations)

MTX + tocilizumab (8 mg/kg) 76 per 1000 159 per 1000 (65 to 345) OR 2.3 (0.85 to 6.4) Low (study limitations)

MTX + azathioprine 76 per 1000 341 per 1000 (122 to 688) OR 6.3 (1.7 to 27) Moderate (indirectness)

MTX + cyclosporine 76 per 1000 80 per 1000 (30 to 166) OR 1.1 (0.38 to 2.4) Moderate (indirectness)

IM/sc MTX + cyclosporine 76 per 1000 429 per 1000 (79 to 895) OR 9.2 (1.0 to 104) Very low (extreme imprecision)

MTX + hydroxychloroquine/ chloroquine 76 per 1000 98 per 1000 (30 to 292) OR 1.3 (0.38 to 5.0) Low (imprecision)

MTX + sulphasalazine 76 per 1000 100 per 1000 (52 to 194) OR 1.4 (0.68 to 2.9) Moderate (indirectness)

MTX + sulphasalazine + hydroxychloroquine 76 per 1000 53 per 1000 (22 to 111) OR 0.68 (0.28 to 1.5) High

IM/sc MTX 76 per 1000 130 per 1000 (42 to 330) OR 1.8 (0.54 to 6.0) Moderate (imprecision)

MTX-inadequate response

ACR50 (42 studies; 11550 patients)


MTX + abatacept (sc) 126 per 1000 381 per 1000 (284 to 500) OR 4.3 (2.7 to 6.9) High


MTX + etanercept 126 per 1000 647 per 1000 (455 to 804) OR 13 (5.8 to 28) Moderate (study limitations)

MTX + golimumab (sc) 126 per 1000 397 per 1000 (270 to 546) OR 4.6 (2.6 to 8.3)

Moderate (study limitations,

indirectness)

MTX + golimumab (IV) 126 per 1000 339 per 1000 (199 to 527) OR 3.5 (1.7 to 7.7) Moderate (study limitations)



MTX + tocilizumab (4 mg/kg) 126 per 1000 270 per 1000 (169 to 399) OR 2.6 (1.4 to 4.6) High


MTX + tofacitinib 126 per 1000 445 per 1000 (322 to 571) OR 5.5 (3.3 to 9.2) High

MTX + hydroxychloroquine/ chloroquine 126 per 1000 596 per 1000 (231 to 864) OR 10 (2.1 to 44) Low (high imprecision)

MTX + IM Gold 126 per 1000 646 per 1000 (198 to 985) OR 13 (1.7 to 446) Very low (extreme imprecision)

MTX + leflunomide 126 per 1000 448 per 1000 (241 to 709) OR 5.6 (2.2 to 17) Moderate (imprecision)

MTX + sulphasalazine 126 per 1000 289 per 1000 (66 to 677) OR 2.8 (0.49 to 15) Low (high imprecision)

MTX + sulphasalazine + hydroxychloroquine 126 per 1000 614 per 1000 (380 to 803) OR 11 (4.3 to 28) Moderate (study limitations)

Radiographic progression (10 studies; 3238

patients)

MTX + abatacept (IV)




5.9 to 8.8) SMD -0.30 (-1.4 to 0.85)


Page 11 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


12

MTX + abatacept (sc)




9.6 to 11) SMD -0.48 (-2.0 to 1.2)


MTX + adalimumab Mean change 3.3 points on Sharp-VdH scale over 1 year

Mean change 0.51 points (-6.4 to 8) SMD -0.44 (-1.5 to 0.72)

Moderate (study limitations, imprecision)

MTX + etanercept



Mean change -0.49 points

(-12 to 11) SMD -0.60 (-2.4 to 1.2)


MTX + golimumab (sc)




2.8 to 7.7) SMD -0.14 (-0.96 to 0.67)

Low (study limitations, inconsistency, indirectness,

imprecision)

MTX + golimumab (IV) Mean change 3.3 points on Sharp-VdH scale over 1 year


Low (study limitations, imprecision)

MTX + infliximab



Mean change -1.1 points (-

8.3 to 6.4) SMD -0.69 (-1.8 to 0.47)


imprecision)

MTX + sulphasalazine + hydroxychloroquine Mean change 3.3 points on Sharp-VdH scale over 1 year


Low (indirectness, imprecision)

Withdrawals due to adverse events (35

studies; 9882 pt-years)


MTX + abatacept (sc) 74 per 1000 40 per 1000 (21 to 74) OR 0.52 (0.27 to 1.00) Moderate (indirectness)


MTX + certolizumab 74 per 1000 103 per 1000 (56 to 183) OR 1.4 (0.74 to 2.8)


indirectness)

MTX + etanercept 74 per 1000 95 per 1000 (43 to 190) OR 1.3 (0.56 to 2.9) Moderate (study limitations)

MTX + golimumab (sc) 74 per 1000 74 per 1000 (29 to 179) OR 1.00 (0.37 to 2.7) Low (study limitations,

indirectness)

MTX + golimumab (IV) 74 per 1000 99 per 1000 (29 to 368) OR 1.4 (0.37 to 7.3)


imprecision)


MTX + rituximab 74 per 1000 144 per 1000 (48 to 396) OR 2.1 (0.63 to 8.2) Moderate (imprecision)



MTX + tofacitinib 74 per 1000 90 per 1000 (51 to 144) OR 1.2 (0.67 to 2.1) High

MTX + cyclosporine 74 per 1000 198 per 1000 (82 to 446) OR 3.1 (1.1 to 10) Low (indirectness, imprecision)

MTX + IM Gold 74 per 1000 163 per 1000 (30 to 927) OR 2.4 (0.39 to 159) Very low (extreme imprecision)

MTX + leflunomide 74 per 1000 134 per 1000 (55 to 288) OR 1.9 (0.73 to 5.0) Moderate (imprecision)

MTX + sulphasalazine + hydroxychloroquine 74 per 1000 131 per 1000 (67 to 243) OR 1.9 (0.89 to 4.0) Moderate (imprecision)

Abbreviations: IM, intra-muscular; IV, intravenous; MTX, methotrexate; OR, odds ratio; sc, subcutaneous; SMD, standardized mean difference

Page 12 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

13

Radiographic progression: Fifteen trials with 6540 patients were included in this

analysis. MTX + etanercept and MTX + adalimumab were the only treatments that were

significantly superior to oral MTX (Table 2). The sizes of the effects for all interventions

relative to oral MTX were small. The expected radiographic progression was 2.6 points

over 1 year with oral MTX, versus 0.2 points with MTX + etanercept or MTX+

adalimumab, below the minimal clinically important difference of 5 units on the Sharp-

vDh scale.22 There were no significant differences between treatments in pair-wise

comparisons (Appendix A, Table 2). In post-hoc sensitivity analyses using fixed-effects

models, the point estimates were nearly identical to the random-effects model, but the

credible intervals were not as wide, resulting in several bDMARDs (+MTX) that were

superior to oral MTX (Appendix A, Table 3). MTX+SSZ+HCQ was the only csDMARD

combination with outcome data available, and was not superior to oral MTX in either the

random-effects or fixed-effect models.

WDAE: Thirty-five trials with a total follow-up of 10250 patient-years were included in

this analysis. The combinations of MTX + azathioprine and IM/sc MTX + cyclosporine

were inferior to oral MTX (more WDAE) and several other treatments (Tables 2 and

Appendix A, Table 4). There were no differences in pairwise comparisons between

different bDMARDs (+MTX). MTX+SSZ+HCQ was associated with fewer WDAE than

MTX + infliximab (Appendix A, Table 4).

MTX-IR patients

ACR50: Forty-two trials with 11550 patients were included in this analysis. Several

treatments were significantly superior to oral MTX for ACR50 response, including

csDMARD combinations (MTX+SSZ+HCQ, MTX+HCQ, MTX + leflunomide, MTX +

Intramuscular gold), MTX + all bDMARDs with available evidence, and MTX +

tofacitinib (Table 2). There was no evidence for certolizumab, as the available trials were

judged at high risk of bias. In pair-wise comparisons, MTX + etanercept was superior to

the combination of MTX + most bDMARDs and MTX+SSZ+HCQ was superior to MTX

+ the bDMARDs abatacept, infliximab, and tocilizumab (4 mg/kg) (Appendix A, Table

5). The quality of the evidence for MTX+SSZ+HCQ versus MTX was judged to be

moderate as there were some minor inconsistencies in the findings of the 2 trials that

compared MTX+SSZ+HCQ to MTX+ETN,6 8 and because the study design of one of the

trials was judged to indirectly address the comparison of interest.8 This trial randomised

patients at baseline to a step-up to triple therapy versus a step-up to MTX+ETN, only if

an inadequate response to MTX was found after 6 months.8

Radiographic progression: Ten trials with 3238 patients were included in this analysis.

There were no significant differences between any treatment and oral MTX (Table 2 and

Appendix A, Table 6). Similar to the MTX-naïve analysis, differences were observed in

the post-hoc fixed-effect model. MTX + abatacept (iv and sc), adalimumab, etanercept,

IV (not sc) golimumab and infliximab were superior to oral MTX (Appendix A, Table 7).

The point estimate favoured MTX+SSZ+HCQ in the comparison of MTX+SSZ+HCQ

versus oral MTX in a fixed-effects model, but the result was not significant [SMD: -0.40

(95%CrI: -0.84 to -0.04)].

Page 13 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

14

Withdrawals due to adverse events (WDAE): Thirty-five trials with a total follow-up of

9882 patient-years were included in this analysis. MTX + CyA was the only treatment

with a significantly higher rate of WDAE relative to oral MTX. In pairwise comparisons,

MTX + sc abatacept and MTX + iv abatacept were associated with fewer WDAE than

several treatments, including MTX + bDMARDs and MTX+SSZ+HCQ (Appendix A,

Table 11).

Minor outcomes

Results for the minor outcomes are reported in detail in the unabridged

review.[placeholder for reference] MTX+SSZ+HCQ and MTX+SSZ had a significantly

higher odds of total GI events (excluding oral and liver toxicity) in the MTX-naïve

analysis [OR: 2.3 (95%CrI: 1.2 to 4.4) and 2.0 (95%CrI: 1.2 to 3.2)], and MTX +

tocilizumab (8 mg/kg) had a significantly higher odds of leukopenia in the MTX-IR

analysis [OR: 18 (95%CrI: 1.1 to 272)].

Meta-regression and sensitivity analyses

In meta-regression analyses, we found a significant relationship between the OR

for ACR50 response and certain study-level covariates (Appendix B, Table 1), but the

adjusted treatment effects were largely unchanged (Appendix B, Figures 1 and 2). When

all studies (both MTX-naïve and MTX-IR) were included in the same network meta-

analysis and the network assignment was specified with a meta-regression covariate, the

OR of MTX-IR trials were 2.13 times higher (95% CrI: 1.74-2.65).

When studies with partial MTX exposure were excluded from the MTX-naïve

analysis, MTX+SSZ+HCQ was not significantly superior to oral MTX for ACR50

response (Appendix B, Figure 1). The point estimate, however, was slightly higher than

the main analysis and higher than any other treatment. When trials at high risk of bias

were included in the MTX-IR analysis, MTX + certolizumab and MTX + sc/IM MTX

were also superior to oral MTX (Appendix B, Figure 2). There was little change in the

ACR50 response at different time-points of assessment (6 or 12 months), or when data

from the time of the switch in therapy was used for early-escape trials (results presented

in unabridged Cochrane review).[placeholder for reference]

DISCUSSION

Summary of main results

Our systematic review and network meta-analysis compared MTX and all

currently used DMARD combinations with MTX. The main new finding from our review

was that MTX+SSZ+HCQ (‘triple therapy’) was superior to oral MTX and similar to

MTX + biologic therapy for ACR50 response, in both MTX-naïve and MTX-IR

populations. Other csDMARD combinations were also superior to continuing MTX after

an inadequate response to MTX, although the magnitude of effect or quality evidence

Page 14 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

15

was graded lower than for triple therapy. Most bDMARDs, in combination with MTX,

were superior to oral MTX for ACR50 response in both MTX-naïve and MTX-IR

populations, although exceptions existed. Most treatments were well tolerated with only

MTX + azathioprine and MTX + cyclosporine having significantly more WDAE than

oral MTX. Only MTX + adalimumab and MTX + etanercept were superior to oral MTX

for inhibiting joint damage, but the effect was small and observed only in MTX-naïve

trials.

Completeness and applicability of evidence

We did not evaluate the effect of glucocorticoids, which are known to have a

disease-modifying effect, particularly in early RA.23 We did not exclude trials with

corticosteroids, however; our findings therefore relate to the effects of DMARD therapy

independent of a corticosteroid effect. Our results should not be generalised to patients

who have had an inadequate response to biologic therapy, as we did not include these

trials. Our study was also not designed to directly compare treatment strategies.

Specifically, we did not directly compare the approach of starting with MTX

monotherapy in MTX-naïve patients and progressing to triple therapy versus the strategy

of starting with triple therapy directly. The estimates of absolute risk with each treatment

can help inform this decision. Based on the included trials, ~40% of patients naïve to

MTX are expected to have an ACR50 response to oral MTX, compared to 60% for triple

therapy. Patients may accept this difference in risk and choose MTX monotherapy as

initial treatment, reserving combination therapy if they fail to respond adequately. Triple

therapy was also associated with an increase in total gastrointestinal events in MTX-naïve

patients, which may influence patients’ decisions.

Strengths and weaknesses of the review

This review included 150 trials with over 34000 patients. We employed a rigorous

approach to trial identification and outcome abstraction, and thus have confidence that

the results encompass the best RCT evidence of the comparative benefits and harms for

the treatments of interest.

The extent to which indirect evidence is considered can affect the results of a

network meta-analysis.24 With our search strategy we included all direct and first-order

indirect evidence between the treatments of interest; we did not attempt to capture all

second-order indirect evidence. The contribution of the indirect evidence to the overall

estimate from the network meta-analysis decreases quite rapidly as the ‘order’ of the

comparison increases.24 In addition, most of the treatments that would potentially form

second-order indirect evidence were csDMARD monotherapy for which there are few

trials that are generally small and have rarely measured ACR responses.23 We therefore

expect minimal impact from the exclusion of these trials.

An ‘early escape’ design was common in trials of MTX + bDMARDs and MTX +

tofacitinib, particularly in more recent trials. While this allows trials to ethically include a

placebo arm, it presents challenges in interpreting and synthesizing the results. The

Page 15 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

16

proportion of patients remaining on the control treatment at trial end can be very low.

We chose to extract end-of-trial data for the efficacy outcomes, where patients who

require rescue treatment are considered a treatment failure. We included a sensitivity

analysis for ACR50 response using pre-rescue data and found few differences in

treatment effects. Synthesizing adverse events in early escape trials is also challenging.

Patients who crossover from placebo to active therapy often represent substantial patient-

years of exposure; excluding these patients may obscure important safety signals. We

therefore choose to summarize all toxicity data with exposure-adjusted estimates, using

the on-treatment data from early escape trials. This could potentially bias the estimates,

as patients who crossover may differ in certain ways from patients assigned to the

original treatment. We felt this to be less of a potential bias than excluding the patients

who crossed over. Some trials also only reported exposure-adjusted data and we

otherwise would have excluded these trials.

Through meta-regression, we demonstrated that treatments in MTX-IR trials were

associated with odds ratios for ACR50 response that were twice as high as those in MTX-

naïve trials. Thus, prior MTX response is a strong effect-modifier of the clinical response

and pooling studies in MTX-naive and MTX-IR patients will yield biased estimates that

are difficult to relate to clinical practice. This supports our decision to analyze MTX-

naïve and MTX-IR trials separately.

Agreements and disagreements with other studies or reviews

Multiple network meta-analyses of biologic therapy in RA, including a Cochrane

review, have been performed.4 25

This is the first review to our knowledge that has

systematically compared all MTX-based csDMARD and bDMARD/tofacitinib treatment

approaches. A recently published network meta-analysis by Graudal and colleagues

compared combination DMARD therapy with csDMARDs and bDMARDs for

radiographic outcomes.26 Overall, Graudal found that one csDMARD + one bDMARD

was not superior to combination therapy with 2 or 3 csDMARDs for radiographic

outcomes. There are several important differences with our study. First, Graudal grouped

csDMARD combination therapy according to the number of medications (2 or 3),

whereas each biologic agent was considered separately. Grouping DMARD combinations

that are commonly used with those that are rarely used (e.g. – combinations with

bucillamine or auranofin) adds heterogeneity to the estimates and makes it difficult to

apply the results to clinical practice. Second, trials in DMARD-naïve and DMARD-IR

patients were grouped which, as we demonstrated, may bias the estimated treatment

effects. Third, we evaluated a range of outcomes beyond radiographic outcomes,

covering multiple domains relevant to decision making. The number of trials included in

our review (150) was also much greater than the 39 trials that Graudal included. Finally,

Graudal included trials where corticosteroids were part of the intervention (i.e.- applied

differently between arms). The results, therefore address a different research question.

Other traditional (non-network) systematic reviews have evaluated csDMARD

combination therapy.5 27 28

The reviews differed in their outcomes considered and

inclusion criteria, particularly around the inclusion and exclusion of interventions with

corticosteroids. Combined withdrawal due to inefficacy or adverse events was used as the

Page 16 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

17

primary outcome for 2 of the reviews,5 28

as it is commonly reported. Trials are not

designed around this outcome, however, and it does not allow a separation of benefits and

harms necessary to inform clinical decisions.

In contrast to other systematic reviews, our review evaluated the risk of bias

separately for different outcomes, which is the approach recommended by GRADE.29 We

also used recently published GRADE guidance for grading the quality of the evidence.20

While this approach requires subjective decisions, it should increase the transparency of

these choices.

Implications for practice

On the basis of all available direct and indirect evidence, our results suggest that

triple therapy is effective in both MTX-naïve and MTX-IR patients, and not different to

MTX + bDMARDs, for controlling disease activity. Given cost considerations, it would

be difficult to justify the use of MTX + bDMARDs prior to an adequate trial of

combination therapy with MTX + csDMARDs (preferably triple therapy).

What is already known on this topic

• Meta-analyses have demonstrated that most biologic DMARDs combined with

methotrexate are superior to methotrexate alone for controlling disease activity,

but the benefits of combining methotrexate with conventional synthetic DMARDs

are uncertain.

• A previous Cochrane network meta-analysis of biologic treatments for

rheumatoid arthritis showed some differences between agents, but did not

compare biologic therapy to combination therapy with conventional synthetic

DMARDs

• Understanding the comparative benefits and harms of these treatments is

essential, given that biologic therapy costs 10-20 times that of most conventional

synthetic DMARDs.

What this study adds

• Through a Cochrane network meta-analysis of 150 trials including over 34,000

patients, ‘triple therapy’ was superior to methotrexate alone and not significantly

different from methotrexate + any biologic DMARD for controlling disease

activity.

• Given cost considerations, these findings support a therapeutic trial of low-cost

triple therapy prior to using biologic DMARDs.

Page 17 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

18

Copyright/license statement

“The Corresponding Author has the right to grant on behalf of all authors and does

grant on behalf of all authors, a worldwide licence to the Publishers and its licensees in

perpetuity, in all forms, formats and media (whether known now or created in the future),

to i) publish, reproduce, distribute, display and store the Contribution, ii) translate the

Contribution into other languages, create adaptations, reprints, include within

collections and create summaries, extracts and/or, abstracts of the Contribution, iii)

create any other derivative work(s) based on the Contribution, iv) to exploit all

subsidiary rights in the Contribution, v) the inclusion of electronic links from the

Contribution to third party material where-ever it may be located; and, vi) licence any

third party to do any or all of the above.”

Competing interests: All authors have completed the ICMJE uniform disclosure form

at www.icmje.org/coi_disclosure.pdf and declare: no support from any organisation for

the submitted work; GH has received fellowship funding supported by the Canadian

Rheumatology Association/ The Arthritis Society/ UCB Pharma, honoraria and travel

expenses from Abbott, honoraria from UCB Pharma and has participated in an advisory

board meeting for Amgen; ChB holds the Canadian Rheumatology Association/ The

Arthritis Society Clinician Investigator Salary Award and is a Canadian Institutes of

Health Research New Investigator (Community-Based Primary Healthcare). In the past

two years she has participated in advisory boards for Roche and UCB, received honoraria

from UCB and Amgen, and an unrestricted travel grant from Celgene; GT has nothing to

disclose; DM is supported by a Canada Research Chair in Health Systems and Services

Research, and Arthur J.E. Child Chair in Rheumatology. In the past year she has received

honoraria from Abbvie for a seminar; DD has nothing to disclose; ClB has received grant

support from Janssen, Pfizer, Amgen, Abbott/Abbvie Canada, BMS, Celgene, Eli Lilly,

Fresenius Kabi, Hoffman La Roche, Sanofi, UCB and Calea, has acted as a consultant for

AstraZeneca, Abbott/Abbvie Canada and BMS, and has participated in advisory board

meetings for Janssen, Pfizer, Amgen, and AstraZeneca; no other relationships or

activities that could appear to have influenced the submitted work.

Contributors: GH developed the concept for the study, wrote the protocol, participated

in all stages of the study including literature search and data abstraction. He conducted all

analyses and wrote and revised the manuscript. He is the guarantor. ChB edited and

revised the protocol and manuscript, assisted with the literature search and data

abstraction. GT edited and revised the protocol and manuscript, and assisted with

developing the concept, the analyses and interpretation of the data. DM edited and

revised the protocol and manuscript and assisted with developing the concept for the

study. DD edited and revised the manuscript and assisted with the data abstraction. ClB

edited and revised the protocol and manuscript, and assisted with developing the concept

for the study. All authors had full access to all of the data (including statistical reports

and tables) in the study and can take responsibility for the integrity of the data and the

accuracy of the data analysis.

Transparency declaration: The lead author affirms that this manuscript is an honest,

accurate, and transparent account of the study being reported; that no important aspects of

Page 18 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

19

the study have been omitted; and that any discrepancies from the study as planned (and, if

relevant, registered) have been explained.

Ethics approval: Not required

Funding: None

Data sharing: Statistical models available with permission of corresponding author on

request.

Page 19 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

20

REFERENCES

1. Smolen JS, Landewe R, Breedveld FC, et al. EULAR recommendations for the

management of rheumatoid arthritis with synthetic and biological disease-

modifying antirheumatic drugs: 2013 update. Ann Rheum Dis 2014;73:492-

509

2. Singh JA, Furst DE, Bharat A, et al. 2012 update of the 2008 American College

of Rheumatology recommendations for the use of disease-modifying

antirheumatic drugs and biologic agents in the treatment of rheumatoid

arthritis. Arthritis Care Res (Hoboken) 2012;64:625-39

3. Jansen JP, Fleurence R, Devine B, et al. Interpreting indirect treatment

comparisons and network meta-analysis for health-care decision making:

report of the ISPOR Task Force on Indirect Treatment Comparisons Good

Research Practices: part 1. Value Health 2011;14:417-28

4. Singh JA, Christensen R, Wells GA, et al. Biologics for rheumatoid arthritis: an

overview of Cochrane reviews. Cochrane Database Syst Rev

2009;4:CD007848

5. Katchamart W, Trudeau J, Phumethum V, Bombardier C. Methotrexate

monotherapy versus methotrexate combination therapy with non-biologic

disease modifying anti-rheumatic drugs for rheumatoid arthritis. Cochrane

Database Syst Rev 2010;4:CD008495

6. O'Dell JR, Mikuls TR, Taylor TH, et al. Therapies for active rheumatoid

arthritis after methotrexate failure. New Engl J Med 2013;369:307-18

7. van Vollenhoven RF, Geborek P, Forslind K, et al. Conventional combination

treatment versus biological treatment in methotrexate-refractory early

rheumatoid arthritis: 2 year follow-up of the randomised, non-blinded,

parallel-group Swefot trial. Lancet 2012;379:1712-20

8. Moreland LW, O'Dell JR, Paulus HE, et al. A randomized comparative

effectiveness study of oral triple therapy versus etanercept plus

methotrexate in early aggressive rheumatoid arthritis: The treatment of

early aggressive rheumatoid arthritis trial. Arthritis Rheum 2012;64:2824-35

9. Hazlewood G, Barnabe C, Tomlinson G, Marshall D, Bombardier C.

Methotrexate monotherapy and methotrexate combination therapy with

traditional and biologic disease modifying anti- rheumatic drugs for

rheumatoid arthritis: A network meta- analysis (Protocol). Cochrane

Database Syst Rev 2013:CD010227

10. Aletaha D, Neogi T, Silman AJ, et al. 2010 rheumatoid arthritis classification

criteria: an American College of Rheumatology/European League Against

Rheumatism collaborative initiative. Ann Rheum Dis 2010;69:1580-8

11. Arnett FC, Edworthy SM, Bloch DA, et al. The American Rheumatism

Association 1987 revised criteria for the classification of rheumatoid

arthritis. Arthritis Rheum 1988;31:315-24

12. Ropes MW, Bennett GA, Cobb S, Jacox R, Jessar RA. 1958 Revision of

diagnostic criteria for rheumatoid arthritis. Bull Rheum Dis 1958;9:175-6

Page 20 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

21

13. Felson DT, Anderson JJ, Boers M, et al. American College of Rheumatology.

Preliminary definition of improvement in rheumatoid arthritis. Arthritis

Rheum 1995;38:727-35

14. Ory PA. Interpreting radiographic data in rheumatoid arthritis. Ann Rheum

Dis 2003;62:597-604

15. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool

for assessing risk of bias in randomised trials. BMJ 2011;343:d5928

16. Dias S, Sutton AJ, Ades AE, Welton NJ. Evidence synthesis for decision making

2: a generalized linear modeling framework for pairwise and network meta-

analysis of randomized controlled trials. Med Decis Making 2013;33:607-17

17. Ades AE, Sculpher M, Sutton A, et al. Bayesian methods for evidence synthesis

in cost-effectiveness analysis. Pharmacoeconomics 2006;24:1-19

18. Brooks S, Gelman A. General Methods for Monitoring Convergence of

Iterative Simulations. J Comput Graph Stat 1998;7:434-55

19. Plummer M. rjags: Bayesian graphical models using MCMC. R package

version 3-14. Secondary rjags: Bayesian graphical models using MCMC. R

package version 3-14 2014. http://CRAN.R-project.org/package=rjags.

20. Puhan MA, Schunemann HJ, Murad MH, et al. A GRADE Working Group

approach for rating the quality of treatment effect estimates from network

meta-analysis. BMJ 2014;349:g5630

21. Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency in mixed

treatment comparison meta-analysis. Stat Med 2010;29:932-44

22. Bruynesteyn K, van der Heijde D, Boers M, et al. Determination of the

minimal clinically important difference in rheumatoid arthritis joint damage

of the Sharp/van der Heijde and Larsen/Scott scoring methods by clinical

experts and comparison with the smallest detectable difference. Arthritis

Rheum 2002;46:913-20

23. Gaujoux-Viala C, Nam J, Ramiro S, et al. Efficacy of conventional synthetic

disease-modifying antirheumatic drugs, glucocorticoids and tofacitinib: a

systematic literature review informing the 2013 update of the EULAR

recommendations for management of rheumatoid arthritis. Ann Rheum Dis

2014;73:510-5

24. Hawkins N, Scott DA, Woods B. How far do you go? Efficient searching for

indirect evidence. Med Decis Making 2009;29:273-81

25. Thorlund K, Druyts E, Avina-Zubieta JA, Wu P, Mills EJ. Why the findings of

published multiple treatment comparison meta-analyses of biologic

treatments for rheumatoid arthritis are different: an overview of recurrent

methodological shortcomings. Ann Rheum Dis 2013;72:1524-35

26. Graudal N, Hubeck-Graudal T, Tarp S, Christensen R, Jurgens G. Effect of

combination therapy on joint destruction in rheumatoid arthritis: a network

meta-analysis of randomized controlled trials. PLoS ONE 2014;9:e106408

27. Graudal N, Jurgens G. Similar effects of disease-modifying antirheumatic

drugs, glucocorticoids, and biologic agents on radiographic progression in

rheumatoid arthritis: Meta-analysis of 70 randomized placebo-controlled or

drug-controlled studies, including 112 comparisons. Arthritis Rheum

2010;62:2852-63

Page 21 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

22

28. Choy EHS, Smith C, Dore CJ, Scott DL. A meta-analysis of the efficacy and

toxicity of combining disease-modifying anti-rheumatic drugs in rheumatoid

arthritis based on patient withdrawal. Rheumatology (Oxford)

2005;44:1414-21

29. Guyatt GH, Oxman AD, Vist G, et al. GRADE guidelines: 4. Rating the quality of

evidence--study limitations (risk of bias). J Clin Epidemiol 2011;64:407-15

Page 22 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

Search flowchart

361x270mm (72 x 72 DPI)

Page 23 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

Networks of included studies for MTX-naïve (A) and MTX-inadequate response populations (B). Each line represents a direct comparison between two treatments from one or more trials. The line thickness is

directly proportional to the total sample size of all trials for that comparison (line length has no meaning).

Biologic/targeted synthetic DMARDs are shown on the left of each network and conventional synthetic DMARDs on the right. Treatments on the innermost circle (green hashed line) are treatments of interest,

whereas treatments on the outermost circle (red hashed line) are other treatments that form links between treatments of interest. Comparisons to MTX are shown in blue. Two trials were included in both analyses.

Page 24 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

564x282mm (72 x 72 DPI)

Page 25 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


1

Appendix A – Tables of pair-wise comparisons for all major outcomes

Table 1: Pair-wise comparisons for ACR50 response: MTX-naïve

Medication MTX MTX+ABAT (IV) MTX+ADA IM/sc MTX+ADA MTX+ETN MTX+GOL (sc) MTX+IFX MTX+RTX

MTX+TCZ

(4 mg/kg)

MTX+TCZ

(8 mg/kg) MTX+CyA IM/sc MTX+CyA MTX+HCQ/CQ MTX+SSZ MTX+SSZ+HCQ

MTX+ABAT (IV)

1.8 (1.0 to 3.3)

98% --

MTX+ADA

2.1 (1.5 to 2.9)

>99%

1.1 (0.58 to 2.2)

68% --

IM/sc MTX+ADA

2.2 (0.82 to 6.0)

95%

1.2 (0.37 to 3.8)

64%

1.0 (0.38 to 3.0)

55% --

MTX+ETN

3.0 (2.1 to 4.6)

>99%

1.6 (0.81 to 3.4)

93%

1.4 (0.88 to 2.5)

94%

1.3 (0.47 to 4.0)

74% --

MTX+GOL (sc)

1.3 (0.69 to 2.6)

82%

0.73 (0.30 to 1.8)

23%

0.64 (0.31 to

1.3)

11%

0.61 (0.18 to 2.0)

19%

0.45 (0.20 to 0.95)

2% --

MTX+IFX

2.0 (1.3 to 3.8)

>99%

1.1 (0.53 to 2.7)

62%

0.96 (0.57 to

2.0)

44%

0.91 (0.32 to 3.1)

42%

0.67 (0.37 to 1.4)

12%

1.5 (0.68 to 3.9)

85% --

MTX+RTX

2.4 (1.3 to 4.4)

>99%

1.3 (0.56 to 3.1)

77%

1.2 (0.58 to 2.3)

68%

1.1 (0.34 to 3.5)

57%

0.81 (0.38 to 1.6)

26%

1.8 (0.75 to 4.4)

92%

1.2 (0.48 to 2.5)

68% --

MTX+TCZ

(4 mg/kg)

1.6 (0.96 to 2.8)

97%

0.89 (0.40 to 2.0)

37%

0.78 (0.42 to

1.5)

18%

0.73 (0.24 to 2.4)

27%

0.55 (0.27 to 1.1)

3%

1.2 (0.52 to 2.9)

69%

0.80 (0.35 to

1.6)

25%

0.67 (0.30 to

1.6)

15% --

MTX+TCZ

(8 mg/kg)

1.9 (1.1 to 3.3)

99%

1.0 (0.46 to 2.3)

54%

0.90 (0.48 to

1.7)

35%

0.85 (0.27 to 2.7)

38%

0.63 (0.31 to 1.2)

7%

1.4 (0.60 to 3.3)

80%

0.93 (0.39 to

1.8)

42%

0.77 (0.35 to

1.8)

24%

1.2 (0.64 to 2.1)

72% --

MTX+CyA

1.8 (0.92 to 3.5)

96%

0.96 (0.40 to 2.3)

46%

0.84 (0.41 to

1.8)

30%

0.80 (0.24 to 2.6)

35%

0.58 (0.27 to 1.3)

9%

1.3 (0.52 to 3.4)

72%

0.86 (0.35 to

1.9)

35%

0.72 (0.30 to

1.8)

22%

1.1 (0.46 to 2.5)

57%

0.93 (0.40 to 2.2)

42% --

IM/sc MTX+CyA

1.5 (0.42 to 5.5)

76%

0.83 (0.21 to 3.4)

39%

0.73 (0.20 to

2.7)

31%

0.69 (0.18 to 2.6)

29%

0.51 (0.13 to 1.9)

15%

1.2 (0.28 to 4.7)

59%

0.75 (0.17 to

2.8)

33%

0.64 (0.15 to

2.6)

25%

0.94 (0.23 to 3.7)

47%

0.82 (0.20 to 3.3)

38%

0.88 (0.19 to

3.7)

43% --

MTX+HCQ/CQ

0.84 (0.24 to

2.9)

37%

0.45 (0.11 to 1.8)

12%

0.40 (0.11 to

1.4)

7%

0.37 (0.08 to 1.8)

10%

0.27 (0.08 to 0.94)

2%

0.61 (0.15 to

2.5)

23%

0.40 (0.10 to

1.5)

8%

0.33 (0.08 to

1.4)

6%

0.51 (0.13 to 2.0)

14%

0.43 (0.11 to 1.7)

10%

0.47 (0.11 to

1.9)

13%

0.54 (0.09 to 3.1)

22% --

MTX+SSZ

1.1 (0.44 to 2.8)

61%

0.61 (0.21 to 1.9)

17%

0.54 (0.20 to

1.4)

10%

0.49 (0.13 to 2.0)

15%

0.38 (0.14 to 0.94)

2%

0.84 (0.28 to

2.6)

37%

0.55 (0.18 to

1.5)

12%

0.47 (0.15 to

1.4)

8%

0.69 (0.23 to 2.0)

23%

0.60 (0.20 to 1.8)

16%

0.65 (0.20 to

2.0)

21%

0.72 (0.16 to 3.5)

33%

1.4 (0.41 to 4.7)

70% --

MTX+SSZ+HCQ

2.3 (1.2 to 4.8)

99%

1.3 (0.52 to 3.3)

71%

1.1 (0.53 to 2.5)

61%

1.0 (0.33 to 3.7)

53%

0.78 (0.43 to 1.5)

20%

1.8 (0.70 to 4.7)

88%

1.1 (0.47 to 2.6)

63%

0.96 (0.40 to

2.5)

46%

1.4 (0.61 to 3.6)

81%

1.2 (0.52 to 3.1)

69%

1.3 (0.49 to 3.5)

73%

1.5 (0.37 to 6.7)

73%

2.8 (0.93 to 9.1)

97%

2.1 (0.85 to 5.3)

95% --

IM/sc MTX

1.1 (0.59 to 2.1)

65%

0.61 (0.25 to 1.4)

11%

0.53 (0.26 to

1.1)

4%

0.50 (0.24 to 1.1)

4%

0.37 (0.17 to 0.77)

1%

0.83 (0.34 to

2.1)

33%

0.55 (0.21 to

1.2)

6%

0.46 (0.19 to

1.1)

4%

0.68 (0.30 to 1.6)

16%

0.59 (0.25 to 1.4)

9%

0.64 (0.24 to

1.6)

15%

0.73 (0.24 to 2.2)

27%

1.4 (0.34 to 5.6)

66%

1.0 (0.33 to 3.0)

51%

0.47 (0.18 to 1.2)

5%

Treatment effects are presented as the median Odds Ratio (OR) with 95% credible intervals and the probability of superiority (higher odds of ACR50 response) for the row versus column. Shaded cells

reflect comparisons in which the credible interval excluded the null value. Abbreviations: ABAT, abatacept; ADA, adalimumab; CQ, chloroquine; CyA, cyclosporine A; ETN, etanercept; GOL, golimumab; HCQ, hydroxychloroquine; IFX, infliximab; IM, intra-muscular; IV,

intravenous; MTX, methotrexate; RTX, rituximab; sc, subcutaneous; SSZ, sulphasalazine; TCZ, tocilizumab; TOFA, tofacitinib

Page 26 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


2

Table 2: Pair-wise comparisons for radiographic progression (change from baseline, random-effects model): MTX-naïve

Medication MTX MTX+ABAT (IV) MTX+ADA MTX+ETN MTX+GOL (sc) MTX+IFX MTX+RTX

MTX+TCZ

(4 mg/kg)

MTX+TCZ

(8 mg/kg) MTX+CyA

MTX+ABAT (IV)

-0.20 (-0.65 to 0.25)

86% --

MTX+ADA

-0.37 (-0.68 to -0.03)

98%

-0.17 (-0.71 to 0.39)

77% --

MTX+ETN

-0.36 (-0.61 to -0.07)

99%

-0.16 (-0.67 to 0.38)

77%

0.01 (-0.41 to 0.44)

48% --

MTX+GOL (sc)

-0.13 (-0.57 to 0.34)

75%

0.08 (-0.56 to 0.72)

38%

0.25 (-0.33 to 0.79)

16%

0.24 (-0.31 to 0.75)

16% --

MTX+IFX

-0.43 (-0.87 to 0.00)

97%

-0.23 (-0.87 to 0.40)

82%

-0.06 (-0.61 to 0.47)

61%

-0.06 (-0.62 to 0.42)

63%

-0.31 (-0.94 to 0.32)

87% --

MTX+RTX

-0.38 (-0.84 to 0.07)

96%

-0.17 (-0.82 to 0.45)

75%

-0.01 (-0.58 to 0.54)

51%

-0.01 (-0.58 to 0.49)

52%

-0.25 (-0.90 to 0.39)

83%

0.05 (-0.58 to 0.69)

42% --

MTX+TCZ

(4 mg/kg)

-0.24 (-0.70 to 0.22)

88%

-0.04 (-0.67 to 0.60)

55%

0.13 (-0.44 to 0.69)

29%

0.12 (-0.44 to 0.64)

29%

-0.11 (-0.77 to 0.54)

67%

0.19 (-0.44 to 0.83)

23%

0.14 (-0.50 to 0.80)

30% --

MTX+TCZ

(8 mg/kg)

-0.36 (-0.80 to 0.10)

95%

-0.15 (-0.80 to 0.48)

73%

0.01 (-0.56 to 0.56)

48%

0.01 (-0.54 to 0.51)

49%

-0.23 (-0.87 to 0.40)

81%

0.07 (-0.55 to 0.71)

39%

0.02 (-0.61 to 0.67)

47%

-0.12 (-0.56 to 0.32)

75% --

MTX+CyA

-0.21 (-0.53 to 0.14)

91%

-0.01 (-0.56 to 0.56)

52%

0.16 (-0.30 to 0.62)

21%

0.15 (-0.29 to 0.58)

20%

-0.09 (-0.65 to 0.49)

64%

0.22 (-0.31 to 0.79)

17%

0.17 (-0.38 to 0.76)

23%

0.03 (-0.53 to 0.61)

45%

0.15 (-0.40 to 0.71)

26% --

MTX+SSZ+HCQ

-0.02 (-0.78 to 0.79)

52%

0.18 (-0.69 to 1.1)

33%

0.35 (-0.48 to 1.2)

19%

0.34 (-0.38 to 1.1)

17%

0.11 (-0.77 to 1.0)

40%

0.41 (-0.45 to 1.3)

16%

0.36 (-0.51 to 1.3)

20%

0.23 (-0.64 to 1.1)

30%

0.34 (-0.53 to 1.3)

21%

0.19 (-0.64 to 1.1)

31%

Treatment effects are presented as the median standardized mean difference (smd) with 95% credible intervals and the probability of superiority (less radiographic progression) for the row versus

column. Shaded cells reflect comparisons in which the credible interval excluded the null value.

Abbreviations: ABAT, abatacept; ADA, adalimumab; CyA, cyclosporine A; ETN, etanercept; GOL, golimumab; HCQ, hydroxychloroquine; IFX, infliximab; IV, intravenous; MTX, methotrexate; RTX, rituximab; SSZ, sulphasalazine; TCZ, tocilizumab

Page 27 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


3

Table 3: Pair-wise comparisons for radiographic progression (change from baseline, fixed-effects model): MTX-naïve

Medication MTX MTX+ABAT (IV) MTX+ADA MTX+ETN MTX+GOL (sc) MTX+IFX MTX+RTX

MTX+TCZ

(4 mg/kg)

MTX+TCZ

(8 mg/kg) MTX+CyA

MTX+ABAT (IV)

-0.20 (-0.38 to -0.02)

99% --

MTX+ADA

-0.40 (-0.53 to -0.26)

>99%

-0.19 (-0.42 to 0.03)

95% --

MTX+ETN

-0.40 (-0.50 to -0.29)

>99%

-0.19 (-0.40 to 0.02)

96%

-0.00 (-0.18 to 0.18)

50% --

MTX+GOL (sc)

-0.13 (-0.35 to 0.09)

88%

0.07 (-0.21 to 0.36)

31%

0.27 (0.01 to 0.52)

2%

0.27 (0.02 to 0.51)

2% --

MTX+IFX

-0.43 (-0.59 to -0.26)

>99%

-0.23 (-0.47 to 0.02)

96%

-0.03 (-0.25 to 0.18)

61%

-0.03 (-0.23 to 0.17)

62%

-0.30 (-0.57 to -0.02)

98% --

MTX+RTX

-0.39 (-0.57 to -0.20)

>99%

-0.19 (-0.44 to 0.07)

92%

0.01 (-0.21 to 0.24)

48%

0.01 (-0.21 to 0.23)

47%

-0.26 (-0.55 to 0.03)

96%

0.04 (-0.20 to 0.29)

38% --

MTX+TCZ

(4 mg/kg)

-0.24 (-0.45 to -0.03)

99%

-0.04 (-0.32 to 0.23)

61%

0.16 (-0.10 to 0.40)

11%

0.16 (-0.08 to 0.39)

10%

-0.11 (-0.42 to 0.19)

76%

0.19 (-0.08 to 0.45)

9%

0.15 (-0.14 to 0.42)

15% --

MTX+TCZ

(8 mg/kg)

-0.36 (-0.55 to -0.17)

>99%

-0.15 (-0.42 to 0.11)

87%

0.04 (-0.20 to 0.28)

37%

0.04 (-0.18 to 0.26)

36%

-0.23 (-0.52 to 0.06)

94%

0.07 (-0.18 to 0.32)

29%

0.03 (-0.24 to 0.29)

40%

-0.12 (-0.26 to 0.03)

94% --

MTX+CyA

-0.23 (-0.39 to -0.07)

>99%

-0.03 (-0.26 to 0.21)

59%

0.17 (-0.04 to 0.37)

6%

0.17 (-0.02 to 0.36)

4%

-0.10 (-0.37 to 0.17)

77%

0.20 (-0.03 to 0.42)

5%

0.16 (-0.09 to 0.40)

10%

0.01 (-0.25 to 0.27)

47%

0.13 (-0.12 to 0.37)

16% --

MTX+SSZ+HCQ

-0.05 (-0.68 to 0.57)

57%

0.15 (-0.50 to 0.80)

33%

0.34 (-0.29 to 0.98)

15%

0.34 (-0.27 to 0.96)

14%

0.08 (-0.59 to 0.74)

41%

0.37 (-0.27 to 1.0)

13%

0.33 (-0.32 to 0.98)

16%

0.19 (-0.47 to 0.85)

29%

0.30 (-0.35 to 0.95)

18%

0.18 (-0.47 to 0.82)

29%

Treatment effects are presented as the median standardized mean difference (smd) with 95% credible intervals and the probability of superiority (less radiographic progression) for the row versus column. Shaded cells reflect comparisons in which the credible interval excluded the null value.

Abbreviations: ABAT, abatacept; ADA, adalimumab; CyA, cyclosporine A; ETN, etanercept; GOL, golimumab; HCQ, hydroxychloroquine; IFX, infliximab; IV, intravenous; MTX, methotrexate;

RTX, rituximab; SSZ, sulphasalazine; TCZ, tocilizumab

Page 28 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


4

Table 4: Pair-wise comparisons for withdrawals due to adverse events: MTX-naïve

Medication MTX MTX+ABAT (IV) MTX+ADA IM/sc MTX+ADA MTX+ETN MTX+GOL (sc) MTX+IFX MTX+RTX

MTX+TCZ

(4 mg/kg)

MTX+TCZ

(8 mg/kg) MTX+AZA MTX+CyA IM/sc MTX+CyA MTX+HCQ/CQ MTX+SSZ MTX+SSZ+HCQ

MTX+ABAT (IV)

0.73 (0.21 to 2.4)

72% --

MTX+ADA

1.2 (0.62 to 2.2)

24%

1.7 (0.41 to 6.6)

20% --

IM/sc MTX+ADA

0.77 (0.06 to 8.2)

59%

1.1 (0.07 to 14)

48%

0.63 (0.05 to 7.0)

65% --

MTX+ETN

0.80 (0.45 to 1.6)

78%

1.1 (0.30 to 4.8)

44%

0.65 (0.29 to 1.8)

84%

1.1 (0.09 to 15)

48% --

MTX+GOL (sc)

2.3 (0.67 to 9.2)

8%

3.2 (0.57 to 21)

8%

1.9 (0.49 to 9.0)

16%

3.0 (0.22 to 59)

19%

2.9 (0.65 to 12)

6% --

MTX+IFX

2.5 (0.94 to 7.4)

3%

3.5 (0.74 to 19)

5%

2.1 (0.67 to 7.5)

10%

3.3 (0.27 to 53)

17%

3.2 (0.90 to 11)

3%

1.1 (0.20 to 5.9)

46% --

MTX+RTX

0.85 (0.22 to 3.0)

61%

1.2 (0.20 to 6.9)

42%

0.70 (0.16 to 3.0)

72%

1.1 (0.08 to 18)

48%

1.1 (0.23 to 4.3)

47%

0.36 (0.05 to 2.2)

88%

0.33 (0.06 to 1.7)

92% --

MTX+TCZ

(4 mg/kg)

1.3 (0.47 to 3.8)

25%

1.8 (0.37 to 9.7)

19%

1.1 (0.34 to 3.8)

44%

1.7 (0.14 to 27)

33%

1.7 (0.46 to 5.4)

17%

0.57 (0.10 to 2.9)

78%

0.53 (0.12 to 2.2)

84%

1.6 (0.30 to 8.6)

26% --

MTX+TCZ

(8 mg/kg)

2.3 (0.85 to 6.4)

4%

3.2 (0.67 to 16)

6%

1.9 (0.60 to 6.5)

11%

3.0 (0.23 to 47)

19%

2.9 (0.80 to 9.1)

4%

0.99 (0.19 to 4.9)

51%

0.92 (0.21 to 3.7)

55%

2.7 (0.55 to 14)

9%

1.7 (0.59 to 5.1)

12% --

MTX+AZA

6.3 (1.7 to 27)

1%

8.9 (1.4 to 59)

1%

5.2 (1.3 to 25)

1%

8.5 (0.57 to 154)

6%

7.8 (1.7 to 38)

1%

2.7 (0.41 to 19)

14%

2.4 (0.45 to 15)

13%

7.6 (1.2 to 51)

2%

4.8 (0.88 to 28)

3%

2.8 (0.53 to 16)

10% --

MTX+CyA

1.1 (0.38 to 2.4)

43%

1.5 (0.28 to 6.3)

28%

0.86 (0.27 to 2.5)

63%

1.3 (0.11 to 19)

40%

1.3 (0.36 to 3.5)

28%

0.45 (0.08 to 2.0)

87%

0.42 (0.09 to 1.5)

93%

1.2 (0.24 to 5.7)

38%

0.80 (0.18 to 2.8)

66%

0.46 (0.11 to 1.6)

91%

0.17 (0.03 to 0.77)

99% --

IM/sc MTX+CyA

9.2 (1.0 to 104)

2%

13 (1.1 to 180)

2%

7.6 (0.81 to 96)

4%

12 (0.87 to 290)

3%

11 (1.1 to 133)

2%

4.0 (0.28 to 57)

15%

3.6 (0.32 to 48)

15%

11 (0.89 to 169)

3%

6.9 (0.63 to 94)

6%

4.0 (0.36 to 54)

13%

1.4 (0.10 to 23)

40%

8.9 (0.90 to 120)

3% --

MTX+HCQ/CQ

1.3 (0.38 to 5.0)

33%

1.9 (0.34 to 11)

24%

1.1 (0.27 to 4.9)

45%

1.7 (0.12 to 32)

34%

1.6 (0.42 to 6.6)

23%

0.56 (0.09 to 3.6)

74%

0.52 (0.11 to 2.8)

80%

1.6 (0.27 to 10)

30%

0.99 (0.20 to 5.5)

51%

0.57 (0.12 to 3.1)

76%

0.21 (0.03 to 1.4)

95%

1.3 (0.30 to 7.1)

37%

0.14 (0.01 to 2.0)

93% --

MTX+SSZ

1.4 (0.68 to 2.9)

19%

1.9 (0.48 to 8.3)

17%

1.1 (0.45 to 3.2)

41%

1.8 (0.15 to 26)

32%

1.7 (0.66 to 4.2)

12%

0.57 (0.13 to 2.6)

79%

0.53 (0.15 to 1.9)

86%

1.6 (0.39 to 7.6)

25%

1.0 (0.30 to 3.8)

49%

0.58 (0.18 to 2.1)

83%

0.21 (0.04 to 1.00)

98%

1.3 (0.45 to 4.9)

31%

0.15 (0.01 to 1.5)

95%

1.0 (0.28 to 3.7)

50% --

MTX+SSZ+HCQ

0.68 (0.28 to 1.5)

83%

0.93 (0.21 to 4.2)

54%

0.55 (0.19 to 1.6)

89%

0.89 (0.08 to 13)

54%

0.85 (0.31 to 1.9)

66%

0.29 (0.06 to 1.3)

95%

0.27 (0.07 to 0.96)

98%

0.80 (0.17 to 3.7)

62%

0.51 (0.13 to 1.8)

87%

0.30 (0.07 to 1.0)

97%

0.11 (0.02 to 0.50)

>99%

0.64 (0.19 to 2.4)

79%

0.07 (0.01 to 0.75)

99%

0.50 (0.12 to 1.9)

84%

0.50 (0.17 to 1.3)

92% --

IM/sc MTX

1.8 (0.54 to 6.0)

14%

2.5 (0.44 to 14)

12%

1.5 (0.40 to 6.0)

25%

2.4 (0.32 to 22)

19%

2.3 (0.53 to 8.3)

10%

0.77 (0.12 to 4.4)

62%

0.72 (0.14 to 3.3)

67%

2.2 (0.36 to 13)

17%

1.4 (0.27 to 6.5)

32%

0.79 (0.16 to 3.7)

64%

0.28 (0.04 to 1.7)

92%

1.7 (0.42 to 8.3)

20%

0.20 (0.02 to 1.2)

96%

1.4 (0.22 to 7.6)

36%

1.3 (0.32 to 5.3)

33%

2.7 (0.63 to 12)

8%

Treatment effects are presented as the median Odds Ratio (OR) with 95% credible intervals and the probability of superiority (lower odds of withdrawal) for the row versus column. Shaded cells reflect

comparisons in which the credible interval excluded the null value. Abbreviations: ABAT, abatacept; ADA, adalimumab; AZA, azathioprine; CTZ, certolizumab; CQ, chloroquine; CyA, cyclosporine A; ETN, etanercept; GOL, golimumab; HCQ, hydroxychloroquine;

IFX, infliximab; IM, intra-muscular; IV, intravenous; LEF, leflunomide; MTX, methotrexate; RTX, rituximab; sc, subcutaneous; SSZ, sulphasalazine; TCZ, tocilizumab

Page 29 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


5

Table 5: Pair-wise comparisons for ACR50 response: MTX-inadequate response

Medication MTX MTX+ABAT (IV) MTX+ABAT (sc) MTX+ADA MTX+ETN MTX+GOL (sc) MTX+GOL (IV) MTX+IFX MTX+RTX

MTX+TCZ

(4 mg/kg)

MTX+TCZ

(8 mg/kg) MTX+TOFA MTX+HCQ/CQ MTX+IMGold MTX+LEF MTX+SSZ

MTX+ABAT (IV)

3.9 (2.8 to 5.5)

>99% --

MTX+ABAT (sc)

4.3 (2.7 to 6.9)

>99%

1.1 (0.75 to 1.6)

71% --

MTX+ADA

4.7 (3.4 to 6.5)

>99%

1.2 (0.78 to 1.8)

81%

1.1 (0.67 to 1.7)

66% --

MTX+ETN

13 (5.8 to 28)

>99%

3.3 (1.4 to 7.8)

>99%

3.0 (1.2 to 7.5)

99%

2.8 (1.2 to 6.5)

99% --

MTX+GOL (sc)

4.6 (2.6 to 8.3)

>99%

1.2 (0.60 to 2.3)

70%

1.1 (0.51 to 2.3)

57%

0.97 (0.51 to 1.9)

47%

0.36 (0.14 to 0.96)

2% --

MTX+GOL (IV)

3.5 (1.7 to 7.7)

>99%

0.92 (0.41 to 2.1)

41%

0.83 (0.35 to 2.0)

33%

0.76 (0.34 to 1.8)

25%

0.28 (0.09 to 0.84)

1%

0.77 (0.30 to 2.0)

29% --

MTX+IFX

3.3 (2.3 to 4.8)

>99%

0.86 (0.53 to 1.3)

25%

0.79 (0.43 to 1.3)

17%

0.72 (0.43 to 1.2)

8%

0.26 (0.11 to 0.64)

<1%

0.73 (0.36 to 1.4)

18%

0.93 (0.39 to 2.1)

43% --

MTX+RTX

3.7 (2.2 to 6.6)

>99%

0.96 (0.51 to 1.8)

45%

0.87 (0.43 to 1.8)

34%

0.80 (0.43 to 1.5)

24%

0.29 (0.11 to 0.78)

1%

0.82 (0.36 to 1.9)

30%

1.1 (0.41 to 2.7)

55%

1.1 (0.59 to 2.2)

64% --

MTX+TCZ

(4 mg/kg)

2.6 (1.4 to 4.6)

>99%

0.67 (0.33 to 1.3)

11%

0.61 (0.27 to 1.2)

8%

0.56 (0.28 to 1.1)

4%

0.20 (0.08 to 0.55)

<1%

0.56 (0.24 to 1.3)

8%

0.72 (0.27 to 1.8)

23%

0.78 (0.38 to 1.6)

23%

0.69 (0.30 to 1.5)

17% --

MTX+TCZ

(8 mg/kg)

4.2 (2.4 to 7.0)

>99%

1.1 (0.55 to 2.0)

61%

0.99 (0.46 to 1.9)

49%

0.91 (0.47 to 1.6)

39%

0.32 (0.12 to 0.85)

1%

0.92 (0.40 to 2.0)

42%

1.2 (0.44 to 2.8)

64%

1.3 (0.65 to 2.4)

77%

1.1 (0.50 to 2.3)

62%

1.6 (0.95 to 2.8)

96% --

MTX+TOFA

5.5 (3.3 to 9.2)

>99%

1.4 (0.77 to 2.6)

88%

1.3 (0.66 to 2.5)

79%

1.2 (0.69 to 2.0)

75%

0.43 (0.17 to 1.1)

4%

1.2 (0.55 to 2.6)

69%

1.6 (0.61 to 3.8)

84%

1.7 (0.89 to 3.2)

95%

1.5 (0.69 to 3.1)

85%

2.1 (0.99 to 4.8)

97%

1.3 (0.65 to 2.9)

78% --

MTX+HCQ/CQ

10 (2.1 to 44)

>99%

2.7 (0.51 to 12)

88%

2.4 (0.45 to 11)

85%

2.2 (0.43 to 9.8)

83%

0.77 (0.20 to 2.8)

35%

2.2 (0.40 to 11)

82%

2.8 (0.48 to 15)

88%

3.1 (0.60 to 14)

91%

2.8 (0.50 to 13)

88%

3.9 (0.73 to 19)

95%

2.4 (0.47 to 12)

86%

1.9 (0.35 to 8.6)

77% --

MTX+IMGold

13 (1.7 to 446)

99%

3.3 (0.42 to 112)

85%

3.0 (0.37 to 98)

83%

2.7 (0.35 to 97)

81%

1.0 (0.11 to 34)

50%

2.8 (0.35 to 103)

81%

3.6 (0.41 to 119)

86%

3.9 (0.50 to 135)

88%

3.4 (0.41 to 112)

86%

5.0 (0.61 to 170)

92%

3.1 (0.38 to 95)

83%

2.3 (0.28 to 82)

76%

1.4 (0.11 to 53)

58% --

MTX+LEF

5.6 (2.2 to 17)

>99%

1.5 (0.53 to 4.6)

76%

1.3 (0.45 to 4.3)

69%

1.2 (0.45 to 3.9)

64%

0.44 (0.13 to 1.8)

12%

1.2 (0.40 to 4.3)

64%

1.6 (0.46 to 6.0)

78%

1.7 (0.61 to 5.4)

83%

1.5 (0.50 to 5.1)

77%

2.2 (0.74 to 7.6)

92%

1.3 (0.47 to 4.6)

70%

1.0 (0.34 to 3.4)

52%

0.57 (0.10 to 4.2)

27%

0.44 (0.01 to 4.7)

27% --

MTX+SSZ

2.8 (0.49 to 15)

87%

0.73 (0.12 to 3.9)

34%

0.66 (0.10 to 3.7)

30%

0.60 (0.10 to 3.3)

26%

0.21 (0.05 to 0.95)

2%

0.61 (0.09 to 3.5)

28%

0.78 (0.11 to 4.9)

40%

0.83 (0.14 to 4.5)

41%

0.76 (0.12 to 4.3)

36%

1.1 (0.17 to 6.1)

53%

0.65 (0.11 to 3.7)

31%

0.49 (0.08 to 2.7)

21%

0.28 (0.06 to 1.0)

3%

0.19 (0.00 to 2.8)

13%

0.49 (0.06 to 3.4)

24% --

MTX+SSZ+HCQ

11 (4.3 to 28)

>99%

2.9 (1.0 to 7.8)

98%

2.6 (0.87 to 7.4)

96%

2.4 (0.86 to 6.5)

95%

0.87 (0.51 to 1.4)

28%

2.4 (0.77 to 7.1)

94%

3.1 (0.91 to 10)

96%

3.3 (1.2 to 9.1)

99%

3.0 (0.94 to 8.7)

97%

4.3 (1.4 to 13)

99%

2.7 (0.90 to 7.8)

96%

2.0 (0.67 to 5.9)

90%

1.1 (0.34 to 3.8)

57%

0.84 (0.03 to 8.1)

45%

1.9 (0.44 to 7.5)

82%

4.0 (0.99 to 18)

97%

Treatment effects are presented as the median Odds Ratio (OR) with 95% credible intervals and the probability of superiority (higher odds of ACR50 response) for the row versus column. Shaded cells

reflect comparisons in which the credible interval excluded the null value.

Abbreviations: ABAT, abatacept; ADA, adalimumab; AZA, azathioprine; CTZ, certolizumab; CQ, chloroquine; CyA, cyclosporine A; ETN, etanercept; GOL, golimumab; HCQ, hydroxychloroquine; IFX, infliximab; IM, intra-muscular; IV, intravenous; LEF, leflunomide; MTX, methotrexate; RTX, rituximab; sc, subcutaneous; SSZ, sulphasalazine; TCZ, tocilizumab; TOFA, tofacitinib

Page 30 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


6

Table 6: Pair-wise comparisons for radiographic progression (change from baseline, random-effects model): MTX-inadequate response

Medication MTX MTX+ABAT (IV) MTX+ABAT (sc) MTX+ADA MTX+ETN MTX+GOL (sc) MTX+GOL (IV) MTX+IFX

MTX+ABAT (IV)

-0.30 (-1.44 to 0.85)

84% --

MTX+ABAT (sc)

-0.48 (-2.03 to 1.2)

86%

-0.18 (-2.11 to 1.8)

67% --

MTX+ADA

-0.44 (-1.53 to 0.72)

90%

-0.14 (-1.71 to 1.5)

67%

0.04 (-1.11 to 1.2)

43% --

MTX+ETN

-0.60 (-2.41 to 1.2)

87%

-0.30 (-2.46 to 1.8)

72%

-0.11 (-2.59 to 2.3)

58%

-0.14 (-2.34 to 2.0)

61% --

MTX+GOL (sc)

-0.14 (-0.96 to 0.67)

76%

0.15 (-1.27 to 1.6)

32%

0.33 (-1.52 to 2.1)

21%

0.29 (-1.12 to 1.7)

19%

0.45 (-1.50 to 2.4)

20% --

MTX+GOL (IV)

-0.44 (-1.55 to 0.73)

89%

-0.15 (-1.76 to 1.5)

67%

0.04 (-1.99 to 2.0)

46%

0.00 (-1.62 to 1.5)

50%

0.15 (-1.98 to 2.3)

38%

-0.30 (-1.67 to 1.1)

80% --

MTX+IFX

-0.69 (-1.83 to 0.47)

94%

-0.40 (-2.04 to 1.2)

82%

-0.21 (-2.21 to 1.7)

67%

-0.24 (-1.89 to 1.3)

72%

-0.10 (-1.50 to 1.3)

62%

-0.55 (-1.97 to 0.88)

88%

-0.25 (-1.88 to 1.3)

73% --

MTX+SSZ+HCQ

-0.41 (-2.02 to 1.2)

82%

-0.12 (-2.14 to 1.9)

60%

0.07 (-2.24 to 2.4)

45%

0.04 (-1.96 to 2.0)

47%

0.19 (-0.62 to 0.99)

18%

-0.27 (-2.11 to 1.5)

71%

0.03 (-1.95 to 2.0)

47%

0.28 (-0.85 to 1.4)

17%

Treatment effects are presented as the median standardized mean difference (smd) with 95% credible intervals and the probability of superiority (less radiographic progression) for the row versus

column. Abbreviations: ABAT, abatacept; ADA, adalimumab; ETN, etanercept; GOL, golimumab; HCQ, hydroxychloroquine; IV, intravenous; MTX, methotrexate; sc, subcutaneous; SSZ, sulphasalazine

Page 31 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


7

Table 7: Pair-wise comparisons for radiographic progression (change from baseline, fixed-effect model): MTX-inadequate response

Medication MTX MTX+ABAT (IV) MTX+ABAT (sc) MTX+ADA MTX+ETN MTX+GOL (sc) MTX+GOL (IV) MTX+IFX

MTX+ABAT (IV)

-0.29 (-0.49 to -0.09)

>99% --

MTX+ABAT (sc)

-0.48 (-0.75 to -0.21)

>99%

-0.18 (-0.53 to 0.15)

86% --

MTX+ADA

-0.44 (-0.66 to -0.23)

>99%

-0.15 (-0.45 to 0.14)

84%

0.04 (-0.13 to 0.21)

34% --

MTX+ETN

-0.57 (-1.05 to -0.10)

99%

-0.28 (-0.81 to 0.23)

86%

-0.10 (-0.63 to 0.45)

63%

-0.13 (-0.65 to 0.39)

69% --

MTX+GOL (sc)

-0.13 (-0.34 to 0.07)

90%

0.16 (-0.13 to 0.45)

14%

0.35 (0.00 to 0.69)

2%

0.31 (0.01 to 0.60)

2%

0.44 (-0.07 to 0.96)

5% --

MTX+GOL (IV)

-0.44 (-0.64 to -0.24)

>99%

-0.15 (-0.43 to 0.14)

84%

0.04 (-0.30 to 0.38)

41%

0.01 (-0.29 to 0.30)

48%

0.14 (-0.38 to 0.66)

31%

-0.30 (-0.59 to -0.02)

98% --

MTX+IFX

-0.68 (-1.03 to -0.34)

>99%

-0.39 (-0.80 to 0.00)

97%

-0.21 (-0.65 to 0.23)

82%

-0.24 (-0.64 to 0.16)

88%

-0.11 (-0.44 to 0.21)

75%

-0.55 (-0.95 to -0.15)

>99%

-0.25 (-0.66 to 0.15)

89% --

MTX+SSZ+HCQ

-0.40 (-0.84 to 0.04)

96%

-0.10 (-0.60 to 0.38)

66%

0.08 (-0.43 to 0.60)

39%

0.04 (-0.43 to 0.53)

43%

0.17 (-0.01 to 0.36)

3%

-0.26 (-0.75 to 0.21)

86%

0.04 (-0.45 to 0.53)

44%

0.29 (0.02 to 0.56)

2%

Treatment effects are presented as the median standardized mean difference (smd) with 95% credible intervals and the probability of superiority (less radiographic progression) for the row versus column. Shaded cells reflect comparisons in which the credible interval excluded the null value.

Page 32 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


8

Table 8: Pair-wise comparisons for withdrawals due to adverse events: MTX-inadequate response

Medication MTX MTX+ABAT (IV) MTX+ABAT (sc) MTX+ADA MTX+CTZ MTX+ETN MTX+GOL (sc) MTX+GOL (IV) MTX+IFX MTX+RTX

MTX+TCZ

(4 mg/kg)

MTX+TCZ

(8 mg/kg) MTX+TOFA MTX+CyA MTX+IMGold MTX+LEF

MTX+ABAT (IV)

0.75 (0.43 to 1.3)

88% --

MTX+ABAT (sc)

0.52 (0.27 to 1.00)

98%

0.70 (0.40 to 1.2)

90% --

MTX+ADA

1.4 (0.90 to 2.2)

6%

1.9 (1.0 to 3.5)

2%

2.8 (1.4 to 5.1)

<1% --

MTX+CTZ

1.4 (0.74 to 2.8)

16%

2.0 (0.85 to 4.4)

6%

2.7 (1.1 to 6.7)

2%

1.0 (0.47 to 2.2)

50% --

MTX+ETN

1.3 (0.56 to 2.9)

24%

1.8 (0.67 to 4.5)

12%

2.5 (0.85 to 7.3)

5%

0.92 (0.35 to 2.3)

56%

0.93 (0.31 to 2.6)

56% --

MTX+GOL (sc)

1.00 (0.37 to 2.7)

50%

1.4 (0.46 to 4.0)

28%

1.9 (0.59 to 6.0)

13%

0.71 (0.24 to 2.1)

73%

0.74 (0.21 to 2.3)

72%

0.77 (0.21 to 2.6)

67% --

MTX+GOL (IV)

1.4 (0.37 to 7.3)

32%

1.9 (0.44 to 11)

20%

2.7 (0.58 to 15)

10%

0.97 (0.24 to 5.3)

52%

0.98 (0.22 to 5.4)

51%

1.0 (0.22 to 6.2)

48%

1.4 (0.28 to 9.3)

36% --

MTX+IFX

1.6 (0.99 to 2.8)

3%

2.2 (1.1 to 4.7)

1%

3.1 (1.4 to 7.5)

<1%

1.1 (0.60 to 2.3)

34%

1.2 (0.50 to 2.6)

36%

1.3 (0.55 to 2.9)

30%

1.6 (0.54 to 5.2)

19%

1.2 (0.22 to 4.8)

41% --

MTX+RTX

2.1 (0.63 to 8.2)

12%

2.8 (0.73 to 13)

7%

4.2 (0.98 to 21)

3%

1.5 (0.40 to 6.5)

27%

1.5 (0.37 to 6.9)

29%

1.6 (0.37 to 7.1)

26%

2.1 (0.40 to 12)

19%

1.5 (0.22 to 11)

33%

1.3 (0.33 to 5.6)

36% --

MTX+TCZ

(4 mg/kg)

1.6 (0.93 to 3.1)

5%

2.2 (0.98 to 5.0)

3%

3.2 (1.3 to 7.8)

1%

1.1 (0.55 to 2.5)

35%

1.1 (0.47 to 3.1)

39%

1.3 (0.45 to 3.4)

33%

1.6 (0.49 to 5.8)

22%

1.2 (0.19 to 5.4)

43%

0.99 (0.47 to 2.2)

51%

0.76 (0.19 to 3.1)

65% --

MTX+TCZ

(8 mg/kg)

1.7 (0.99 to 3.0)

3%

2.3 (1.1 to 5.0)

2%

3.3 (1.4 to 7.9)

<1%

1.2 (0.61 to 2.5)

32%

1.2 (0.50 to 3.0)

35%

1.3 (0.50 to 3.4)

31%

1.7 (0.53 to 5.9)

20%

1.2 (0.20 to 5.3)

41%

1.0 (0.49 to 2.2)

46%

0.78 (0.20 to 3.0)

63%

1.0 (0.63 to 1.7)

42% --

MTX+TOFA

1.2 (0.67 to 2.1)

23%

1.7 (0.79 to 3.4)

8%

2.4 (1.1 to 5.2)

2%

0.86 (0.49 to 1.5)

71%

0.87 (0.35 to 2.0)

63%

0.93 (0.35 to 2.5)

55%

1.2 (0.39 to 3.9)

38%

0.89 (0.16 to 3.8)

56%

0.76 (0.33 to 1.5)

79%

0.58 (0.13 to 2.1)

79%

0.75 (0.33 to 1.7)

76%

0.73 (0.33 to 1.5)

80% --

MTX+CyA

3.1 (1.1 to 10)

2%

4.1 (1.3 to 16)

1%

5.8 (1.7 to 25)

<1%

2.2 (0.70 to 8.1)

9%

2.3 (0.61 to 8.3)

10%

2.4 (0.62 to 9.7)

11%

3.2 (0.73 to 13)

6%

2.2 (0.32 to 12)

18%

2.0 (0.60 to 6.7)

15%

1.5 (0.26 to 7.6)

35%

1.9 (0.57 to 7.1)

16%

1.8 (0.57 to 7.2)

16%

2.5 (0.79 to 9.7)

6% --

MTX+IMGold

2.4 (0.39 to 159)

20%

3.4 (0.50 to 239)

11%

4.7 (0.65 to 350)

6%

1.7 (0.27 to 137)

32%

1.7 (0.24 to 107)

32%

1.9 (0.25 to 131)

29%

2.5 (0.31 to 169)

21%

1.9 (0.15 to 115)

33%

1.5 (0.22 to 97)

36%

1.2 (0.13 to 101)

45%

1.5 (0.23 to 109)

36%

1.4 (0.22 to 108)

38%

2.0 (0.29 to 140)

28%

0.77 (0.10 to 48)

58% --

MTX+LEF

1.9 (0.73 to 5.0)

9%

2.6 (0.84 to 7.8)

5%

3.6 (1.1 to 12)

2%

1.4 (0.47 to 3.9)

30%

1.3 (0.40 to 4.4)

31%

1.4 (0.39 to 5.0)

27%

1.9 (0.45 to 8.0)

20%

1.4 (0.19 to 7.0)

35%

1.2 (0.38 to 3.5)

40%

0.90 (0.17 to 4.1)

56%

1.2 (0.37 to 3.6)

40%

1.1 (0.38 to 3.5)

42%

1.6 (0.52 to 4.8)

21%

0.60 (0.13 to 2.6)

74%

0.78 (0.01 to 5.9)

59% --

MTX+SSZ+HCQ

1.9 (0.89 to 4.0)

4%

2.5 (1.0 to 6.5)

2%

3.6 (1.3 to 10.0)

1%

1.3 (0.57 to 3.2)

26%

1.3 (0.50 to 3.7)

29%

1.4 (0.73 to 2.8)

15%

1.8 (0.54 to 6.9)

15%

1.3 (0.22 to 6.5)

36%

1.1 (0.58 to 2.3)

32%

0.88 (0.19 to 3.8)

57%

1.2 (0.46 to 2.8)

38%

1.1 (0.45 to 2.7)

40%

1.5 (0.65 to 4.0)

17%

0.60 (0.15 to 2.3)

77%

0.78 (0.01 to 5.8)

59%

0.98 (0.30 to 3.4)

51%

Treatment effects are presented as the median Odds Ratio (OR) with 95% credible intervals and the probability of superiority (lower odds of withdrawal) for the row versus column. Shaded cells reflect

comparisons in which the credible interval excluded the null value.

Abbreviations: ABAT, abatacept; ADA, adalimumab; CTZ, certolizumab; CyA, cyclosporine A; ETN, etanercept; GOL, golimumab; HCQ, hydroxychloroquine; IFX, infliximab; IM, intra-muscular; IV, intravenous; LEF, leflunomide; MTX, methotrexate; RTX, rituximab; sc, subcutaneous; SSZ, sulphasalazine; TCZ, tocilizumab; TOFA, tofacitinib

Page 33 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


1

Appendix B – Meta-regression and sensitivity analyses

Table 1: Meta-regression for ACR50 response MODEL FIT

Beta coefficient,

median (CrI)

Interpretation Unadjusted analysis Adjusted

analysis

DIC Between-study

standard

deviation

Total residual

deviance

(number of parameters)

DIC Between-study

standard

deviation

Total residual

deviance

(number of parameters)

MTX-naïve

MTX response rate -1.8 (-4.2 to 0.67) Decrease in OR of 0.84 times (0.66 to 1.07) for every

10% increase in the response rate for MTX

546.1 0.19 57.6 (58) 548.3 0.22 58.0 (58)

Disease duration (years) 0.005 (-0.10 to 0.11) Increase in OR of 1.01 times (0.90 to 1.11) for every

year of disease duration

546.1 0.19 57.6 (58) 546.8 0.25 57.1 (58)

Duration of trial (weeks) -0.0019 (-0.01 to

0.008)

Decrease in OR of 0.97 times (0.87 to 1.1) for every

12 weeks of trial duration

546.1 0.19 57.6 (58) 546.1 0.23 57.1 (58)

MTX dose >= 15 mg/week -0.34 (-0.79 to 0.03) Decrease in OR of 0.71 times (0.46 to 1.0) for trials

where the dose of MTX is >= 15 mg/wk

546.1 0.19 57.6 (58) 543.9 0.18 55.9 (58)

Year of publication of trial 0.047 (0.009 to

0.092)

Increase in OR of 1.05 times (1.01 to 1.10) for each

year later of publication (range of years 2000-2014)

546.1 0.19 57.6 (58) 542.4 0.11 55.3 (58)

Swollen joint count -0.05 (-0.11 to 0.01) Decrease in OR of 0.96 times (0.90 to 1.01) for every

1 additional swollen joint at baseline

376.7 0.31 38.2 (38) 374.8 0.19 38.4 (38)

DAS-28 0.51 (-0.74 to 1.7) Increase in OR of 1.7 times (0.48 to 5.3) for every 1

additional point increase in DAS28 at baseline

343.5 0.13 32.9 (34) 344.9 0.16 32.4 (34)

MTX-inadequate response

MTX response rate -5.1 (-9.1 to -2.3) Decrease in OR of 0.60 times (0.40 to 0.80) for every 10% increase in the response rate for MTX

765.1 0.26 95.7 (91) 770.0 0.26 96.8 (91)

Disease duration (years) 0.11 (0.02 to 0.20) Increase in OR of 1.11 times (1.02 to 1.21) for every

year of disease duration

723.5 0.28 90.3 (85) 720.3 0.21 90.2 (85)

Duration of trial (weeks) 0.002 (-0.02 to 0.02) Increase in OR of 1.02 times (0.83 to 1.24) for every 12 weeks of trial duration

765.1 0.26 95.7 (91) 763.6 0.28 95.2 (91)

MTX dose >= 15 mg/week -0.21 (-0.60 to 0.22) Decrease in OR of 0.81 times (0.55 to 1.24) for trials

where the dose of MTX is >= 15 mg/wk

765.1 0.26 95.7 (91) 763.5 0.28 94.0 (91)

Year of publication of trial -0.05 (-0.11 to 0.02) Decrease in OR of 0.95 times (0.90 to 1.02) for each year later of publication (range of years 1999-2014)

765.1 0.26 95.7 (91) 762.3 0.23 96.3 (91)

Swollen joint count 0.02 (-0.05 to 0.08) Increase in OR of 1.02 times (0.95 to 1.09) for every

1 additional swollen joint at baseline

709.0 0.27 89.8 (85) 711.5 0.29 88.7 (85)

DAS-28 0.27 (-0.26 to 0.85) Increase in OR of 1.31 times (0.77 to 2.34) for every 1 additional point increase in DAS28 at baseline

514.2 0.27 64.5 (60) 515.9 0.26 65.0 (60)

Abbreviations: CrI, credible interval; DIC, deviance information criterion; MTX, methotrexate; OR, odds ratio

Model fit between the unadjusted and adjusted analyses can be compared through the values for DIC, between study standard deviation and residual deviance (lower values means improved fit)

Page 34 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

2

Figure 1: Selected meta-regression and sensitivity analyses for ACR50 response in MTX-naïve trials.

Abbreviations: ABAT, abatacept; ADA, adalimumab; CQ, chloroquine; CyA, cyclosporine; ETN,

etanercept; GOL, golimumab; HCQ, hydroxychloroquine; IFX, infliximab; IM, intra-muscular; IV,

intravenous; LEF, leflunomide; MTX, methotrexate; RTX, rituximab; sc, subcutaneous; SSZ,

sulphasalazine; TCZ, tocilizumab

Page 35 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

3

Figure 2: Selected meta-regression and sensitivity analyses for ACR50 response in MTX-inadequate

response trials

Abbreviations: ABAT, abatacept; ADA, adalimumab; CTZ, certolizumab; CQ, chloroquine; ETN,

etanercept; GOL, golimumab; HCQ, hydroxychloroquine; IFX, infliximab; IA, IM, intra-muscular; IV,

intravenous; LEF, leflunomide; MTX, methotrexate; RTX, rituximab; sc, subcutaneous; SSZ,

sulphasalazine; TCZ, tocilizumab; TOFA, tofacitinib

Page 36 of 36


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Documents

Confidential: For Review Only · Date Submitted by the Author: 25-Mar-2015 Complete List of Authors: Hazlewood, Glen; University of Calgary, Medicine ... Arthritis, Rheumatoid (MeSH),