Nineteen RCTs (n=2,025, range 37 to 263) were identified. The quality score ranged from 3 to 10; only three RCTs scored lower than 7. Ten studies had appropriate randomisation and adequate blinding for both patients and care providers. For the remaining studies, the method of randomisation was either not explicitly described or was unclear.
In DMARD-N patients there was no significant difference in patient withdrawal due to lack of efficacy of methotrexate combination versus monotherapy (five studies). Patient withdrawal due to toxicity was significantly increased with methotrexate combination therapy (RR 1.72, 95% CI 1.04 to 2.83). Data for ACR response was available for three RCTs with combination arms that used ciclosporin (two RCTs) or doxycycline (one RCT). The only significant result was for ACR 70 response in one of the ciclosporin trials (RR 2.41, 95% CI 1.07 to 5.44), which favoured the methotrexate combination arm. The outcomes for EULAR response or remission were not significant (two RCTS).
In MTX-IR patients, significantly fewer patients withdrew in the methotrexate combination group compared with the monotherapy group due to lack of efficacy (RR 0.42, 95% CI 0.21 to 0.84; three studies). Patient withdrawal due to toxicity was significantly increased with methotrexate combination therapy (RR 1.89, 95% CI 1.05 to 3.41).
Combination therapy was significantly more effective than methotrexate monotherapy for ACR response (four studies): RR 2.51 (95% CI 1.92 to 3.28) for ACR 20; RR 4.54 (95% CI 2.51 to 8.20) for ACR 50; and RR 5.59 (95% CI 2.08 to 15.01) for ACR 70 response. There were no data for ACR remission or EULAR response.
In non-MTX-IR patients there was a significant benefit for patient withdrawal due to lack of efficacy of methotrexate combination versus monotherapy (RR 0.37, 95% CI 0.16 to 0.87; five studies). Data for ACR response were available for two RCTs, where combination therapy was only significantly more effective than methotrexate monotherapy for ACR 20 response (RR 1.85, 95% CI 1.21 to 2.83). The outcome for EULAR response was not significant (one RCT). There were no data on ACR remission.
In 17 of the 19 RCTs, combination therapy had higher withdrawal due to adverse reactions than monotherapy. The differences were significant only for ciclosporin (RR 1.88, 95% CI 1.02 to 3.50) and azathioprine (RR 5.18, CI 95% 1.58 to 16.96).
Combined withdrawal due to lack of efficacy:
There was no significant difference in withdrawals for both efficacy and safety for DMARD-N patients, MTX-IR patients or non-MTX-IR patients (13 trials). There was significant heterogeneity for non-MTX-IR patients (I2=57.4%), with one important outlier. The outlier RCT of non-MTX-IR patients with a combination of methotrexate with sulfasalazine and hydroxychloroquine gave a better efficacy/toxicity ratio than methotrexate alone (RR 0.30, 95% CI 0.14 to 0.65).
Results for toxicity were also reported in the paper.