Iron deficiency (ID) affects 55–80% of heart failure (HF) patients regardless of left ventricular ejection fraction (LVEF). ID has been linked to reduced exercise capacity, HF hospitalisation, and worse prognosis, independently of anaemia or other comorbidities.1–3

Clinical practice guidelines recommended periodically screening ID in HF patients and correcting with intravenous ferric carboxymaltose (FCM).4,5 Based on FAIR-HF6 and CONFIRM-HF7 trials, FCM is indicated in HF patients with LVEF <45% to improve HF symptoms, exercise capacity, and quality of life.6,7 Recently, the AFFIRM-HF8 trial extended the benefit to patients recently hospitalised for HF and LVEF <50% to reduce the risk of HF hospitalization.7 Thus, the evidence in HF with reduced LVEF (HfrEF) is strong, but there is a lack of evidence in HF with mildly reduced LVEF (HfmrEF) and preserved LVEF (HfpEF), considering that the characteristics and outcomes of HF patients are different according to LVEF.9,10

Recent studies showed that ID is linked to a higher risk for worsening HF and functional deterioration in all-range LVEF, and its correction could improve exercise capacity and functional outcomes.1,11,12 This study aims to evaluate the effectiveness and safety of ID correction with FCM at 3-month follow-up in a cohort of HF outpatients across the whole spectrum of LVEF.

We performed a retrospective, unicentric, and observational registry, including all consecutive HF outpatients who received FCM to correct ID between 2015 and 2017 in the HF unit, according to current recommendations.4,13 We included all-range LVEF patients because ID was identified as a relevant parameter in all HF patients,1,14 waiting for the results of the Effect of IV Iron in Patients With Heart Failure With Preserved Ejection Fraction trial (FAIR-HfpEF; NCT03074591). The study complied with the Declaration of Helsinki and was approved by the Hospital Ethics Committee of Hospital Universitario Clínico San Carlos. Due to the retrospective character analysis, obtaining informed consent was unnecessary. Data were recorded in an electronic data database with anonymised patient data. Only the principal investigators can consult the total data of the included patients.

All patients have a chronic history of clinical HF, and the classification of HF was made according to clinical guidelines.5 All the analysis was made according to LVEF: HfrEF (<40%), HfmrEF (41–49%), and HfpEF (≥50%). In case of HfpEF, due to the heterogenity of this population, all of them have HF symptoms, a LVEF ≥50% and raised natriuretic peptides.

Symptomatic patients received FCM when ID was detected, defined as ferritin <100ng/mL (absolute deficiency) or ferritin 100–299ng/mL with transferrin saturation <20% (functional deficiency). Anaemia was defined as haemoglobin (Hb) <12g/dL (female) and <13g/dL (male).4,5 According to the studies with FCM, patients with Hb >15mg/dL were excluded, with no other exclusion criteria. FCM was administered in saline 0.9% (50–150mL) for 15min (maximum 1000mg of FCM a week) in the daily hospital with nurse supervision and 30min of monitoring post-infusion. The dose of FCM was calculated based on weight and Hb: (a) Hb 10–14g/dL and weight <70kg, 1000mg; (b) Hb <10g/dL and weight <70kg, 1500mg; (c) Hb <14g/dL and weight ≥70kg, 2000mg; (d) Hb ≥14g/dL, 500mg

We analysed different variables before FCM infusion: (a) medical history, vital signs, New York Heart Association (NYHA) functional class, HF aetiology, and LVEF (Simpson method by echocardiogram); (b) blood test: hematic and ferric parameters, N-terminal pro-B-type natriuretic peptide (NT-proBNP), renal function, and electrolytes; (c) aspects related to FCM: dose and adverse effects; (d) in the patients that voluntary wanted to answer, the health-related quality of life (HRQoL) with the EQ-5D-3L test15 and the Spanish version of Minnesota Living with Heart Failure Questionnaire (MLHFQ).16,17 We repeated clinical, analytical, and HRQoL tests at 3 months, according to clinical practice guidelines.4,18 At 3 months, all patients were answered about their subjective perception of clinical improvement after FCM administration.

Quantitative variables are expressed as mean (standard deviation) or median [interquartile range] in nonparametric data, and qualitative variables as numbers and percentages. The Student's t-test, or the sum of Wilcoxon ranges in nonparametric data, was used to compare continuous quantitative variables. Categorical variables were compared with the chi-square test and Fisher's exact test. A significance level of 0.05 (bilateral) was established for all statistical tests. The statistical analysis was performed with SPSS 21.0 (IBM, United States).

We included 235 patients (51.5% female) with a mean age of 73.5 years (SD, 10.7), 50.8% older than 75 years. There were 96 patients with HFrEF (40.8%), 41 with HFmrEF (17.4%), and 98 with HFpEF (41.7%). There were more females in HFmrEF and HFpEF groups, and patients with HFpEF had more anaemia and higher blood pressure (Table 1). There were no differences between groups in HF aetiology, NYHA, or type of ID (absolute or functional). Also, there were no differences in ferric analytical parameters except in transferrin, which was higher in HFpEF. Patients with HFrEF also have significantly higher basal NTproBNP levels (P=.02), being quite similar in HFmrEF and HFpEF groups (Table 2).

Two hundred and six patients (87.7%) answered the tests before FCM administration. There were differences in basal EQ-5D index between groups (Table 1 of the supplementary data), with worse punctuations in the HFpEF group (0.7±0.2 in HFrEF, 0.7±0.3 in HFmrEF, 0.6±0.2 in HFpEF; P=.01), especially in the question of activities, with fewer patients without any problem to do activities in HFpEF (56.6% in HFrEF, 58.3% HFmrEF and 35.6% in HFpEF; P=.049).

The mean dose of FCM was 947.0mg (SD, 154.2), with <50% of patients receiving the correct dose according to recommendations, especially in HFpEF patients (P=.004). One patient (0.4%) with HFrEF presented a local exanthema resolved with oral antihistaminics and corticosteroids. There were no other adverse effects.

The mean follow-up time after FCM infusion was 89.8 days (SD, 2.9), with HFrEF and HFpEF patients improving ferric parameters and haemoglobin (Table 3). In HFmrEF, the percentage of patients with anaemia was similar to before FCM administration, and haemoglobin levels did not significantly change (12.9 vs 13.0; P=.95). NTproBNP levels decreased in HFrEF and HFpEF patients, and the NYHA functional class improved in HFrEF and HFmrEF ones (Fig. 1).

Fig. 1.

Changes in analytical parameters after FCM administration according to LVEF. FCM: ferric carboxymaltose; HFmrEF: heart failure with a mildly reduced left ventricular ejection fraction; HfpEF: heart failure with a preserved left ventricular ejection fraction; HfrEF: heart failure with a reduced left ventricular ejection fraction; LVEF, left ventricular ejection fraction; TSAT, transferrin saturation.

(0.22MB).

Only 52 patients (22.1%) completed the HRQoL tests at 3 months (23 in HFrEF, 11 in HFmrEF, and 18 in HFpEF). There was no improvement in any test in HFmrEF patients. In HFrEF, there was only an improvement in question 6 in the Minnesota test (P=.009). HFpEF patients only improve in question 2 (P=.002) and 5 (P=.005) of the Minnesota test and the physical dimension in MLHFQ (P=.031).

The main findings of our study are: (a) FCM is effective in correcting ID regardless of LVEF; (b) HFpEF patients were less likely to improve functional class and report clinical improvement; (c) NT-proBNP is reduced at 3 months of FCM administration, even in the HFpEF group; (d) administration of FCM is safe in all patients with HF.

ID affects more than half of the patients with HF, regardless of LVEF.3,19 In the CARENFER study, the prevalence was a little higher in HFpEF patients than in HFmrEF (57.5% vs 47.4%) or HFrEF ones (44.3%),20 and also Martens et al. reported a higher prevalence in HFpEF than in HFmrEF (64% vs 61%) and HFrEF (50%).11 ID has been linked to impaired functional capacity and worse exercise capacity in all HF patients, and probably also in clinical outcomes, especially HF hospitalization.11 The myocardium of HF patients with ID has a lower concentration of iron than no HF patients,21 affecting cardiomyocyte contraction and relaxation,22 explaining the symptoms of HF patients with ID. Different studies observed mitochondrial dysfunction in ID patients, with lower mRNA levels of mitochondrial ferritin and other proteins linked to iron regulation and endothelial function, impacting prognosis, especially in HFrEF patients.1,3,23,24 In our study, less than 5% of the patients were asymptomatic, and the most symptomatic ones were HFpEF, which probably have more diastolic dysfunction and comorbidities (i.e. anaemia), affecting the functional class. Ferric parameters were quite similar between groups in our study, except for anaemia, which is more frequent in HFpEF patients.

FCM is indicated to correct ID in chronic symptomatic outpatients with LVEF <45% to improve HF symptoms, exercise capacity, and quality of life,5 with constated evidence of myocardial iron repletion after FCM administration.25 The evidence with FCM is limited in chronic patients with HFmrEF and HFpEF. López-Vilella et al., in a study with 565 outpatients with HFpEF, showed that FCM administration improves all analytical parameters,26 similar to ours. However, the benefit of FCM in functional class improvement is controversial in patients with LVEF >40%.1,27,14 A study showed a slight improvement in NYHA class in HFpEF after FCM administration, but it did not seem clinically relevant (from a median of 2.4 to 1.9).26 Mollace et al. also showed an improvement in exercise capacity after FCM, and it was more likely in HFpEF patients with more affected diastolic function due to enhancing endothelial function and reducing the production of oxygen radical species.28 In our study, HFpEF patients, who have worse basal NYHA and HRQoL test punctuations, did not significantly improve NYHA class or refer clinical improvement after FCM infusion compared to the other groups. The different phenotypes of HFpEF patients, with a high prevalence of other comorbidities, such as anaemia of pulmonary hypertension, can interfere with improving exercise capacity.29,30 Also, sometimes there is a limitation in HFpEF diagnosis, although all our patients have HF symptoms, LVEF ≥50% and raised natriuretic peptides.

Previous studies showed that NYHA improvement is more likely in more symptomatic patients, those with absolute deficiency, and when the correct dose was administred.31,32 In our cohort, HFmrEF patients were more likely to improve NYHA class by at least one point and self-reporting clinical improvement than in the other groups, suggesting that HFmrEF and HFrEF patients behave more similarly in this aspect than HFpEF. Interestingly, this is the only group not reducing NTproBNP levels, probably due to the small number of patients in this group. Other studies, including HF patients with LVEF >40%, had not seen this NTproBNP improvement with FCM.26 The reduction of NTproBNP could be related to improving cardiomyocyte function after iron replacement, with an enhancement of diastolic function and, as previously reported in other studies, a potential improvement of ventricular function in those with LVEF >50%.33

Identifying and conveniently correcting ID is mandatory in HF patients to improve functional class and reduce HF hospitalisation with a good security profile. Despite clinical evidence, especially in HFrEF, only 27% of patients are checked to ID, and only 25% receive FCM.19 Also, it is not unusual for those with ID diagnosis to use lower doses due to logistical and economic reasons, with the consequent necessity of new infusions during the follow-up.31,32 In our study, the correct doses occur in only half of the patients, especially in HFpEF (33.7%), which means not properly replacing iron in the myocardium and probably, not improving diastolic function and functional class. Transferrin saturation seems the most potent parameter to detect and classify ID in all-range HF patients, considering that ferritin levels can be interfered with by inflammation and concomitant comorbidities, such as infection or cancer. In the CARENFER study, the authors confirmed that transferrin saturation is the more powerful parameter to detect ID, especially in the acute setting.20

In conclusion, ID is frequent in HF patients regardless of LVEF and should be checked periodically. Correcting ID with FCM is a safe and effective option in all patients, improving ferric parameters. NTproBNP was reduced in HFrEF and HFpEF patients, and NYHA class improvement is less likely in HFpEF patients than in the rest of the groups.

Vifor Pharma had an economic contribution to statistical analysis performance.

Study design: A. Esteban-Fernández, R. Bover. Data collection: A. Esteban-Fernández, M. Pérez-Serrano, R. Bover. Data review and statistical analysis: A. Esteban-Fernández, R. Bover. Preparation of the manuscript: A. Esteban-Fernández. Review, edit, and acceptance of the manuscript: all authors.

A. Esteban-Fernández and R. Bover received payments for scientific conferences from Vifor Pharma. The rest of the authors indicate no conflicts of interest.