3 Studies that were published in English. The study selection was conducted in two stages. First, by reviewing the abstracts of all the retrieved literature, they were categorized as eligible for full document review and ineligible for full document review . Se condly, the whole document of all the articles categorized as eligible for full document review were reviewed and categorized as eligible for meta analysis and ineligible for meta analysis . Data extraction After developing a data extraction template, data extrac tion was conducted with standard Excel spreadsheets.

From the included studies the following information were extracted name of the first author, year of publication, study design, phase of the trial, duration of therapy, dose, sample size, name of drug used as background regimen, ACR20 response rates, least squares means standard errors or standard deviations for changes in la boratory test results and ACR 20 core component scores, number of patients who experienced adverse events, number of patients who discontinued medication due to adverse events, number of patients with alanine amino transferase levels that were greater than one times the upper limit of the normal range, number of patients with aspartate aminotransferase levels that were greater than one times the upper limit of the normal range and incidences of infections. Operational definitions In the included studies ACR 20 was defined as at least a 20% reduction from baseline in the number of both tender and swollen joints and at least a 20% improvement in three or more of the five remaining ACR core set measures and 6 or more swollen joints and either an erythro cyte sedimentation rate above ULN or a C reactive protein level 7 mg/liter.

Data synthesis statistical analysis For continues variables where SEs were reported instead of SDs, values for SDs were computed by multiplying the SEs with the square root of sample size. Similarly, when the value of serum creatinine was reported as umol/L, it was converted to mg/dl by dividing the values to 88. 4. The efficacy, safety and tolerability of tofacitinib 3, 5, 10, and 15 mg BID alone or in combination with background methotrexate relative to placebo or placebo with back ground methotrexate in the treatment of rheumatoid arthritis were determined by using the random effects model. The OR and the 95% CI for the number of patients with at least a 20% improvement in ACR 20, ALT 1 X ULN, AST 1 X ULN, GSK-3 adverse events, infec tions and discontinued treatment due to adverse events were computed with Mantel Haenszel method. The SMD and 95% CIs for the mean changes in laboratory test results and Health Assessment Questionnaire Disability Index were computed using the inverse variance method.