Background Research has shown increased TNF levels in heart failure (HF). Randomized controlled trials (RCT) have evaluated the effects of TNF inhibitors (TNFi) in these patients and failed to show benefit, and possibly even harm. This raised concerns about the use of TNFi in patients with RA and concomitant HF.

Objectives The aim of this study was to explore the current available evidence to determine whether TNFi was associated with an increased risk of cardiac failure using data from both the RCTs of TNFi in HF as well as observational data from TNFi use in RA.

Methods Two systematic reviews were performed: (1) Placebo controlled trials of TNFi therapy as a treatment for HF. (2) Studies of TNFi therapy in RA where data on HF as a specific outcome was published. Searches were performed using Ovid to search PUBMED and EMBASE. The TNFi therapies included in the literature search were adalimumab, etanercept, infliximab, certolizumab pegol and golimumab. Search (1) identified placebo RCTs of TNFi in HF published up until December 2012. Search (2) included RCTs and observational cohort studies and was limited to studies reporting rates for cardiac failure. Observational studies were required to be prospectively designed and with a comparator arm of non-biologic treated patients.

Results Four RCTs were identified that have studied the effects of TNFi upon HF. Infliximab at 10mg/kg associated with a statistically significant increased risk of HF or death. None of the other RCTs showed a statistically significant difference. Meta-analysis showed a pooled hazard estimate of HF or death of 1.08 (95%CI 0.97, 1.20) with minimal heterogeneity (I² < 0.0%). No RCTs of TNFi therapies in RA were identified that reported on HF as an outcome. The only available data available were from published results from five separate observational cohort studies. The pooled estimate of risk of HF from these studies was 0.98 (95%CI 0.85, 1.12), however with substantial heterogeneity (I² 71.7%). The studies differed in design, recruitment, length of follow up and outcome ascertainment. Substantial conflict was apparent, with point estimates varying both sides of unity.

Conclusions Despite animal and human studies highlighting the importance of TNF-α in HF, RCTs of TNFi failed to show any significant benefit and while the meta-analysis does not confirm a statistically significant risk of harm, the presence of a dose response is concerning. The published data in RA remains mixed, with significant heterogeneity in both study design and study results. Further research is urgently needed to clarify the risk of TNFi in patients with concomitant HF. The large observational registries around the world are in an ideal position to take this question further.

Disclosure of Interest None Declared