Heterogeneity among the included studies was assessed by the chi squared test and I2 statistics. When the value of I2 was greater than or equal to 50%, it was considered as statistically signi ficant. To assess the possible sources of heterogeneity among the included studies, subgroup analysis based on tofacitinib doses and type of therapy and meta regression with two covariates were conducted. Sensitivity analysis was also conducted to determine the robustness of the overall values and the change in I2 statistics when any of the included study was withdrawn from the analysis. Risk of bias of individual studies was evaluated with the Cochrane risk of bias tool. On the other hand, publication or disclosure bias was assessed with fun nel plots.

However, tests for funnel plot asymmetry were not conducted as recommended in a meta analysis that included less than ten studies. All the statistical analyses were conducted by the Open MetaAnalyst software. Results Search result Based on the predetermined inclusion criteria, from the retrieved 43 publication, only eight double blind ran domized clinical trials were included in the meta analysis. Except one study that was conducted in Japan at multiple sites all the included studies recruited patients with rheumatoid arthritis from more than one country. As shown in Table 1, five of the included studies compared the efficacy, safety and tolerability of tofacitinib in combination with back ground methotrexate against placebo with background methotrexate regimen. While the re maining three studies compared tofacitinib mono therapy against placebo.

In all the included studies concomitant medication with stable doses of low dose corticosteroids, nonsteroidal anti inflammatory drugs and selective cyclooxygenase 2 inhibitors were allowed for all treatment groups. From all the selected studies a total of 2,513 patients with rheumatoid arthritis have received one of the four doses of tofacitinib BID with or without methotrexate. While 1,770 patients have received placebo or placebo with background metho trexate. On the other hand, the risk of bias assessment among the included studies did not demonstrate the presences of biases in randomization, blinding and se lective reporting. Efficacy As presented in Figure 2, the odds of tofacitinib treated patients who met the criteria for an ACR 20 response was more than 4 times higher than placebo treated patients.

More over, to the exception of one study, in all the included studies the ACR 20 response rates for patients receiving all tofacitinib dosages 3 mg BID was signi ficantly greater than Carfilzomib those who received placebo. Never theless, the subgroup odds ratios in the subgroups of tofacitinib 10 mg and 15 mg was higher than tofacitinib 3 mg and 5 mg treated groups. Heterogeneity testing has unveiled the presence of sig nificant inconsistency among the included studies.