Adverse symptoms with anti-TNF-alpha therapy in inflammatory bowel disease: systematic review and duration-response meta-analysis.

BACKGROUND: Anti-tumor necrosis factor-alpha (TNF-α) agents have considerable advances in treating inflammatory bowel disease (IBD). These drugs carry possible risk of adverse symptoms, and no meta-analysis has examined this issue and the potential duration-response relationship. PURPOSE: The purpose of this study was to assess duration-response relationship between anti-TNF-α agents and risk of adverse symptoms from all available randomized control trials (RCTs) with placebo arms in IBD. METHODS: PubMed, OVID, and Cochrane Library were searched to January 2015. The RCTs comparing anti-TNF-α therapy with placebo in adults with IBD were eligible. We estimated pooled relative risks (RRs) of adverse symptoms for anti-TNF-α therapy and examined both non-linear and linear duration-response relations between therapy duration and significant related adverse symptoms. RESULTS: Twenty-three RCTs with 7325 patients were included. Adverse symptoms of headache, nausea/vomit, abdominal pain, fever, and arthralgia showed no significant relationship with anti-TNF-α therapy, respectively. Fatigue was significantly associated with anti-TNF-α therapy (RR 1.35, 95 % confidence interval (CI) 1.01-1.81), and subgroup analysis indicated that long therapy duration (>30 weeks) and combination without azathioprine (AZA) were two risk factors for the occurrence of fatigue (RR 1.74, 95 % CI 1.03-2.93; RR 1.65, 95 % CI 1.13-2.40). In the trials without AZA combination, there was a linear duration-response relationship between therapy duration and risk of fatigue (P = 0.0217), and duration of 35 weeks increased the risk of fatigue by 50 %. CONCLUSION: This meta-analysis suggested a promotive effect of anti-TNF-α therapy to the occurrence of fatigue, and for the anti-TNF-α therapy without AZA combination, a linear duration-response relationship existed between therapy duration and risk of fatigue.