BACKGROUND & AIMS:
It is not clear whether combination therapy with immunomodulators affects the immunogenicity of tumor necrosis factor (TNF) antagonists in patients with inflammatory bowel disease. We performed a meta-analysis to quantify the effects of combined immunomodulator therapy on the presence of antibodies against TNF antagonists (antidrug antibodies [ADAs]) and trough levels of anti-TNF agents.
We systematically searched publication databases for studies that reported prevalence of ADAs in patients who received anti-TNF agents. Raw data from studies that met the inclusion criteria were pooled to determine effect estimates. We performed subgroup and metaregression analyses to determine the level of heterogeneity among study outcomes.
We analyzed findings from 35 studies that met inclusion criteria (results reported from 6790 patients with inflammatory bowel disease). The pooled risk ratio for formation of ADAs in patients receiving combined therapy with immunomodulators, versus that of patients receiving anti-TNF monotherapy, was 0.49 (95% confidence interval, 0.41-0.59; P < .001). However, the pooled analysis did not demonstrate a significant difference in trough levels of anti-TNF agents between patients with versus without concurrent use of immunomodulators (standardized mean difference, 0.11; 95% confidence interval, 0.19-0.41; P = .47). Subgroup analyses of patients treated with different TNF antagonists revealed no difference in the formation of ADAs (P = .50 for interaction); the protective effect of immunomodulators did not differ with type of drug patients were given (methotrexate vs thiopurines), or assay for ADA. We observed heterogeneity only among studies of patients with ulcerative colitis (I2 = 76%). Funnel plot and Egger test analyses indicated publication bias in the studies (P = .001).
In a meta-analysis of published studies, we associated combined treatment with immunomodulators with reduced risk of formation of antibodies against TNF antagonists in patients with inflammatory bowel disease.