Background/Purpose: Clinical guidelines are needed to help clinicians provide optimal medical treatment and advise patients about the potential hazards associated with certain drugs. Our objective was to compare the number of serious adverse events (SAEs) for the biologics available for inflammatory arthritis (i.e., rheumatoid arthritis, psoriatic arthritis, and spondylarthritis), enabling a national consensus on safety associated with using these drugs.
Methods: A national guideline panel consisting of clinical experts and methodologists conducted systematic literature searches, identifying randomized controlled trials (RCTs) and inviting all pharmaceutical companies
marketing the biologics in question. Eligible RCTs included patients with
rheumatoid arthritis, psoriatic arthritis, or spondylarthritis, where approved
biologics in standard dose were compared with another biologic or placebo.
One reviewer extracted data on the number of patients with an SAE from
included trials, and a second reviewer confirmed data, which also entailed
inviting the respective pharmaceutical companies to verify extracted data
regarding their own drug(s). The network meta-analysis was based on
mixed-effects logistic regression (modeled in SAS) [1] combining statistical
inference from both direct and indirect comparisons of the treatment effects of
among the biologics. Results were reported as odds ratios (OR [95%CI]). For sensitivity, we explored trial duration using weeks as a covariate in the model.
Results: From the 94 identified RCTs complying with our eligibility
criteria, 7 did not report data on SAEs. Thus, the meta-analysis included 87
trials (27,333 patients) comprising 85 placebo and 90 biologic trial arms:
abatacept (8), adalimumab (22), anakinra (2), certolizumab (8), etanercept
(15), golimumab (8), infliximab (14), rituximab (5), and tocilizumab (8). The
odds for SAEs were statistically higher (P 0.05) for certolizumab and
tocilizumab compared with the placebo (1.60 [1.19;2.16]; P 0.0022 and
1.33 [1.03;1.70]; P 0.028 respectively). Certolizumab was statistically
more likely to result in SAEs compared with all of the following: golimumab
(2.02[1.26;3.25]; P 0.0042), etanercept (1.70[1.15;2.51]; P 0.0084),
rituximab (1.68[1.06;2.66]; P 0.027), abatacept (1.53 [1.05;2.25]; P
0.028), and adalimumab (1.44[1.02;2.02]; P 0.037). Further, tocilizumab
was statistically more likely to result in SAEs than golimumab (1.67[1.07;
2.62]; P 0.025). All other comparisons showed no statistically significant
differences (P 0.05).
Conclusion: This network meta-analysis of RCTs provides empirical
evidence that certolizumab and tocilizumab both present an increased likelihood
of SAEs compared with placebo. Supported by a recent Cochrane
review [2] the Danish guideline panel concluded that certolizumab was more
likely to cause SAEs compared with several other biologics and thus made a
weak recommendation against its use.