Effect of Intravenous Iron Replacement on Recurrent Heart Failure Hospitalizations and Cardiovascular Mortality in Patients with Heart Failure and Iron Deficiency: A Bayesian Meta-Analysis

Abstract Aims Iron deficiency is common in patients with heart failure (HF) and reduced ejection fraction (HFrEF) and is associated with a poor prognosis. Whether intravenous iron replacement improves recurrent HF hospitalizations and cardiovascular mortality of these patients is uncertain although several trials were conducted. Moreover, none of the trials were powered to assess the effect of intravenous iron in clinically important subgroups. Therefore, we conducted a Bayesian analysis to derive precise estimates of the effect of intravenous iron replacement on recurrent HF hospitalizations and cardiovascular mortality in iron-deficient HFrEF patients using consistent subgroup definitions across trials. Methods and results Individual participant data were used from the FAIR-HF (n = 459), CONFIRM-HF (n = 304) and AFFIRM-AHF (n = 1108) trials. These data were re-analysed following as closely as possible the approach taken in the analyses of IRONMAN (n = 1137), for which study level data were used. Definitions of outcomes and subgroups from the FAIR-HF, CONFIRM-HF and AFFIRM-AHF were matched with those used in IRONMAN. The primary endpoint was recurrent HF hospitalizations and cardiovascular mortality. The analysis of recurrent events was based on rate ratios (RR) derived from the Lin-Wei-Yang-Ying model, and the data were pooled using Bayesian random-effects meta-analysis. Compared with placebo, intravenous iron significantly reduced the rates of recurrent HF hospitalizations and cardiovascular mortality (RR 0.73, 95% credible interval CI 0.48–0.99; between-trial heterogeneity tau = 0.16). The pooled treatment effects did not provide evidence for any differential effects for subgroups based on sex (ratio of rate ratios RRR 1.49 95% CI 0.95–2.37, age 69.4 vs. ≥69.4 years) (RRR 0.68 0.40–1.15), ischaemic versus non-ischaemic aetiology of HF (RRR 0.73 0.42–1.33), transferrin saturation 20% vs. ≥20% (RRR 0.75 0.40–1.34), estimated glomerular filtration rate ≤60 versus 60 ml/min/1.73 m2 (RRR 0.97 0.56–1.68), haemoglobin 11.8 versus ≥11.8 (RRR 0.95 0.53–1.60), ferritin 35 versus ≥35 μg/L (RRR 1.26 0.72–2.48) and New York Heart Association class II versus III/IV (RRR 0.91 0.54–1.56). Conclusions Treatment of iron-deficient HFrEF patients with intravenous iron – namely with ferric carboxymaltose or ferric derisomaltose – results in significant reduction in recurrent HF hospitalizations and cardiovascular mortality. Results were nominally consistent across the subgroups studied, but for several of these subgroups uncertainty remains present.