Figure 7 displays a forest plot of the results of the mixed treatment comparisons for overall withdrawals in randomized controlled trials of biologic DMARDs, including: abatacept, adalimumab, anakinra, certolizumab, etanercept, golimumab, infliximab, rituximab, and tocilizumab. The relative withdrawal rate of each pairwise comparison is presented as an odds ratio and 95 percent credible interval. The drug comparisons are listed on the left-hand side (e.g. Abatacept vs. Adalimumab) and there are a total of 72 pairwise comparisons presented. In the center of the figure, the odds ratios are presented as boxes along the x-axis ranging from a scale of 0.01 to 100. Each box has a line running through it representing the width of the credible interval. On the right-hand side of the graph the odds ratios and 95 percent credible intervals are listed down the page corresponding to the box and line graphic in the center and the drug comparison on the right-hand side. A vertical line running the center of the page for an odds ratio equal to one represents the line of no difference. For each comparison, if the credible interval crosses the line of no difference, then a conclusion of no statistically significant difference between the two drugs is supported. Odds ratios greater than one indicate the second drug listed in the comparison has a lower withdrawal rate, while ratios less than one indicate the first drug listed in the comparison has a lower withdrawal rate. Most comparisons crossed this line of no difference. Differences reaching statistical significance include the following comparisons: adalimumab versus etanercept, OR 1.98 and 95 percent credible interval of 1.03 to 4.61; anakinra versus etanercept, OR 3.11 and 95 percent credible interval of 1.42 to 8.67; anakinra versus rituximab, OR 3.05 and 95 percent credible interval of 1.28 to 9.39; certolizumab versus abatacept, OR 0.19 and 95 percent credible interval of 0.07 to 0.41; certolizumab versus adalimumab, OR 0.17 and 95 percent credible interval of 0.06 to 0.35; certolizumab versus anakinra, OR 0.11 and 95 percent credible interval of 0.03 to 0.25; certolizumab versus etanercept, OR 0.29 and 95 percent credible interval of 0.13 to 0.80; certolizumab versus golimumab, OR 0.13 and 95 percent credible interval of 0.05 to 0.53; certolizumab versus infliximab, OR 0.12 and 95 percent credible interval of 0.05 to 0.32; certolizumab versus rituximab, OR 0.28 and 95 percent credible interval of 0.12 to 0.87; certolizumab versus tocilizumab, OR 0.10 and 95 percent credible interval of 0.04 to 0.32; infliximab versus etanercept, OR 2.65 and 95 percent credible interval of 1.12 to 5.54; rituximab versus infliximab, OR 0.45 and 95 percent credible interval of 0.17 to 1.00; rituximab versus tocilizumab, OR 0.32 and 95 percent credible interval of 0.13 to 0.99; and tocilizumab versus etanercept, OR 3.06 and 95 percent credible interval of 1.12 to 6.97.
Figure 7 displays a forest plot of the results of the mixed treatment comparisons for overall withdrawals in randomized controlled trials of biologic DMARDs, including: abatacept, adalimumab, anakinra, certolizumab, etanercept, golimumab, infliximab, rituximab, and tocilizumab. The relative withdrawal rate of each pairwise comparison is presented as an odds ratio and 95 percent credible interval. The drug comparisons are listed on the left-hand side (e.g. Abatacept vs. Adalimumab) and there are a total of 72 pairwise comparisons presented. In the center of the figure, the odds ratios are presented as boxes along the x-axis ranging from a scale of 0.01 to 100. Each box has a line running through it representing the width of the credible interval. On the right-hand side of the graph the odds ratios and 95 percent credible intervals are listed down the page corresponding to the box and line graphic in the center and the drug comparison on the right-hand side. A vertical line running the center of the page for an odds ratio equal to one represents the line of no difference. For each comparison, if the credible interval crosses the line of no difference, then a conclusion of no statistically significant difference between the two drugs is supported. Odds ratios greater than one indicate the second drug listed in the comparison has a lower withdrawal rate, while ratios less than one indicate the first drug listed in the comparison has a lower withdrawal rate. Most comparisons crossed this line of no difference. Differences reaching statistical significance include the following comparisons: adalimumab versus etanercept, OR 1.98 and 95 percent credible interval of 1.03 to 4.61; anakinra versus etanercept, OR 3.11 and 95 percent credible interval of 1.42 to 8.67; anakinra versus rituximab, OR 3.05 and 95 percent credible interval of 1.28 to 9.39; certolizumab versus abatacept, OR 0.19 and 95 percent credible interval of 0.07 to 0.41; certolizumab versus adalimumab, OR 0.17 and 95 percent credible interval of 0.06 to 0.35; certolizumab versus anakinra, OR 0.11 and 95 percent credible interval of 0.03 to 0.25; certolizumab versus etanercept, OR 0.29 and 95 percent credible interval of 0.13 to 0.80; certolizumab versus golimumab, OR 0.13 and 95 percent credible interval of 0.05 to 0.53; certolizumab versus infliximab, OR 0.12 and 95 percent credible interval of 0.05 to 0.32; certolizumab versus rituximab, OR 0.28 and 95 percent credible interval of 0.12 to 0.87; certolizumab versus tocilizumab, OR 0.10 and 95 percent credible interval of 0.04 to 0.32; infliximab versus etanercept, OR 2.65 and 95 percent credible interval of 1.12 to 5.54; rituximab versus infliximab, OR 0.45 and 95 percent credible interval of 0.17 to 1.00; rituximab versus tocilizumab, OR 0.32 and 95 percent credible interval of 0.13 to 0.99; and tocilizumab versus etanercept, OR 3.06 and 95 percent credible interval of 1.12 to 6.97.

Figure 7Mixed treatment comparisons for overall withdrawals in randomized controlled trials of biologic DMARDs

From: Results

Cover of Drug Therapy for Rheumatoid Arthritis in Adults: An Update
Drug Therapy for Rheumatoid Arthritis in Adults: An Update [Internet].
Comparative Effectiveness Reviews, No. 55.
Donahue KE, Jonas DE, Hansen RA, et al.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.