Iron deficiency in heart failure: news and pitfalls in 2025

Iron deficiency (ID) has seen a vast array of scientific and clinical activity in heart failure (HF) over the last 25 years. The construct of ID in HF hinges on proxies of iron availability—ferritin, transferrin saturation (TSAT), and serum iron—and each carries biological and methodological liabilities. Emerging data indicate that criteria privileging iron availability over storage may better capture risk, particularly in inflammatory or amyloid phenotypes, and that choice of definition reshapes both prevalence and prognostic signal. This article summarizes recent articles on ID published in ESC Heart Failure and accompanies an update of our Virtual Issue on ID.

Diagnostic criteria of iron deficiency in heart failure

In a multicentre Japanese cohort of 763 patients with chronic HF (mean age 71 ± 13 years), the proposed criterion TSAT <20% plus serum iron ≤13 μmol/L identified fewer patients than guideline criteria based on ferritin with or without low TSAT, yet tracked independently with all-cause mortality [hazard ratio (HR) 2.01, 95% confidence interval (CI) 1.16–3.51], whereas the guideline definition did not (HR 1.32, 95% CI 0.76–2.28). Mortality rose steeply when serum iron was ≤10 μmol/L, underscoring the primacy of iron availability over ferritin in risk stratification and suggesting recalibration of diagnostic thresholds for Asian populations.¹ Because TSAT is the quotient of serum iron concentration (SIC) and serum transferrin concentration (STC), variation in STC can mask low SIC. In 4422 chronic HF patients (median age 75 years; 32% LVEF ≤40%), low STC clustered with inflammation and malnutrition (higher high-sensitivity C-reactive protein, ferritin; lower albumin) and with low SIC and haemoglobin. Strikingly, among those with the lowest STC, over one-third with SIC ≤13 μmol/L still had TSAT ≥20% and would be missed by standard criteria; conversely, TSAT <20% occurred despite SIC >13 μmol/L in the highest STC quartile. After multivariable adjustment, higher SIC (HR 0.87, 95% CI 0.81–0.95) and higher STC (HR 0.82, 95% CI 0.73–0.91) each associated with lower mortality, with SIC the stronger correlate of anaemia and death—implicating reliance on TSAT or ferritin alone as a source of misclassification.² Disease-specific data in light chain cardiac amyloidosis (AL-CM) reinforce this theme. Among 149 patients with advanced AL-CM, ID prevalence varied widely by definition: 42.3% (guideline), 40.9% (TSAT <20%), and 70.5% (serum iron <13 μmol/L). Only TSAT <20% and serum iron <13 μmol/L predicted all-cause mortality (adjusted HRs 2.49 and 2.24, respectively), while the guideline ferritin-anchored definition was non-informative. Given ferritin’s acute-phase behaviour in AL-CM, these findings argue for availability-based cut-offs when risk-stratifying amyloid cardiomyopathy and highlight how the chosen definition can double the apparent disease burden.³ Zhang and Li, however, cautioned that single-centre, retrospective design, and small sample size temper generalizability. These authors emphasized ferritin’s limitations in inflammatory states and advocated incorporation of hepcidin or soluble transferrin receptor to refine classification, alongside multicentre validation and interventional data to determine whether ID in AL-CM is a modifiable driver or a bystander.⁴

Incident heart failure and early iron-deficiency trajectories

Iron deficiency is more than a biomarker in HF. In fact, it is a biologically plausible driver of impaired energetics that may predate overt decompensation and evolve dynamically after diagnosis. Understanding when ID emerges, how it behaves in the first year, and how its definition alters who should be treated is essential for aligning screening and therapy with risk.

In community-dwelling adults aged ≥65 years from the Cardiovascular Health Study, Sharma et al. used iron status to predict new-onset HF independent of kidney function. Among 1006 participants without baseline HF [414 incident HF cases; 41% with chronic kidney disease (CKD)], iron status categories combined ferritin and TSAT. In weighted Cox models adjusted for demographics, HF risk factors, and estimated glomerular filtration rate, ID defined by low ferritin with low TSAT conferred higher incident HF risk vs iron replete (HR 1.47, 95% CI 1.02–2.11), while functional, mixed, high-iron, or non-classified patterns were not associated. The association between HF and ID did not differ by CKD status (interaction P = 0.2), suggesting that ID per se, rather than renal impairment, underlies risk. These findings support trials testing whether correcting ID can prevent HF in older adults.⁵ In another study, a prospective Stockholm cohort of new-onset HF, Cabrera et al. found that ID was common at diagnosis and remained dynamic during the first year. Of 547 patients [median age 70; 65% men; left ventricular ejection fraction (LVEF) ≤40% in 55%, 41–49% in 19%, HFpEF in 26%], baseline ID by guideline criteria was 34%. Among 368 without intravenous (IV) iron who had paired measurements at 12 months, 32% had ID; across the year, 19% had persistent ID, 13% developed ID, 11% resolved ID, and 57% never had ID, meaning nearly one-quarter changed category. Predictors of 1-year ID included baseline anaemia [odds ratio (OR) 3.91], HF hospitalization (OR 2.21), female sex (OR 2.04), HFpEF (OR 1.96), and diabetes (OR 1.92). Iron deficiency correlated with worse baseline quality of life (Minnesota Living with Heart Failure Questionnaire +7.4 points) but not at follow-up, highlighting fluctuation and treatment effects.⁶ A retrospective single-centre analysis from Innsbruck (Austria) underscored how the choice of ID definition reshapes prevalence estimates in acute vs chronic HF. Among 942 patients (329 acute HF, median age 81; 613 chronic HF, median age 64), overall ID prevalence was consistently higher in acute HF across general (74.8% vs 32.6%), gastroenterology (75.7% vs 34.7%), and cardiology (79.9% vs 47.3%) criteria (all P < 0.001). Subtype assignment varied markedly: absolute ID was most frequent with cardiology thresholds (acute 44.7%, chronic 34.1%) compared with gastroenterology (20.4%, 13.4%) and general (7.0%, 7.2%); functional ID was maximal with the general definition (acute 34.7%, chronic 13.1%). Reclassification was substantial: of 494 labelled absolute/combined ID by cardiology criteria, only 252 were concordant with the general definition, while 107 became no ID and 135 functional ID, with clear implications for eligibility and expected response to iron therapy.⁷

Functional effects of iron deficiency in heart failure

Beyond hospitalization risk, iron availability maps onto physical performance and cardiac mechanics. In a prospective cohort of 562 consecutively admitted HF patients (mean age 72 years), guideline-defined ID was common (58%), but the binary label did not predict performance on the Short Physical Performance Battery. Instead, TSAT treated as a continuous variable tracked independently with low physical function (OR 0.980 per 1% TSAT increase, P = 0.024). Notably, the TSAT–function relationship disappeared in those with diabetes mellitus, an interaction confirmed in propensity-matched analyses, implying that diabetic skeletal muscle and metabolic milieu may blunt the functional signal of iron lack and potentially the yield of iron repletion on performance outcomes.⁸ Complementing whole-body measures, atrial mechanics appear iron-responsive. In a placebo-controlled subanalysis of Myocardial-IRON (ambulatory HF with LVEF <50% and ID), ferric carboxymaltose (FCM) produced a significant 30-day improvement in left atrial longitudinal strain on cardiac magnetic resonance feature tracking (−12.0% vs −8.5% with placebo; between-group Δ −3.55%, P < 0.001), extending prior biventricular findings to the atrium. Although clinical endpoints were not captured in this small study, the rapid enhancement of reservoir function offers a mechanistic bridge between restored iron availability and improved haemodynamics, and motivates larger trials linking atrial strain to symptoms, rhythm burden, and outcomes after iron repletion.⁹

Effects of treating iron deficiency in heart failure

Against the backdrop of consistent epidemiologic links between ID and adverse outcomes in HF, the therapeutic question is whether replenishment modifies risk beyond improving symptoms. Indeed, contemporary evidence now converges on reductions in HF hospitalizations with IV iron, with mortality effects less certain, and emerging work begins to map mechanistic correlates and potential drug–drug interactions.

A comprehensive meta-analysis of 12 randomized trials in HF with reduced ejection fraction (HFrEF, n = 2381) showed that IV iron–carbohydrate complexes—predominantly FCM—halved hospitalizations for worsening HF (rate ratio 0.53, 95% CI 0.42–0.65) and reduced the composite of first HF hospitalization or death (0.75, 0.59–0.95), while all-cause mortality was neutral. Functional status and exercise capacity improved without excess adverse events, and no meaningful differences emerged across iron formulations. These data underpinned the 2021 ESC recommendation to consider FCM in symptomatic HFrEF with LVEF < 45% and ID.¹⁰ Importantly, another study has shown that potential interaction with contemporary guideline-directed therapies is modest. In a post hoc IRONMAN analysis (n = 1137; LVEF ≤45%), baseline SGLT2 inhibitor use (2.6%) was associated with a numerically greater 4-week haemoglobin rise after ferric derisomaltose vs usual care (between-group difference 1.0 g/dL, 95% CI 0.1–1.8) than in non-users (0.4 g/dL, 95% CI 0.3–1.6); the interaction was not statistically significant (P = 0.10) and no polycythaemia occurred. These findings are reassuring for combined use yet warrant confirmation in larger SGLT2-treated cohorts.¹¹ Overall, study-level syntheses consistently show hospitalization benefits but neutral mortality. A 10-trial meta-analysis (n = 3438) reported significant reductions in the composite of cardiovascular (CV) death or first HF hospitalization (RR 0.85, 0.77–0.95), first HF hospitalization (0.82, 0.67–0.99), and total HF hospitalizations (0.74, 0.60–0.91), with no effect on all-cause mortality (0.95, 0.83–1.09) or CV mortality alone (0.89, 0.75–1.05).¹² An updated meta-analysis of 14 RCTs (n = 6651) corroborated a reduction in first HF hospitalization or CV death (OR 0.73, 0.58–0.92), with favourable but non-significant trends for CV mortality, 1-year all-cause mortality, and LVEF; meta-regression suggested greater benefit at lower baseline LVEF, aligning with a disease-modifying effect concentrated in more systolic-impaired patients.¹³

Mechanistic imaging supports a biologic rationale for symptomatic gains. In a prospective T2* magnetic resonance imaging study of systolic HF, patients with ID had lower splenic, hepatic, and trending lower septal myocardial iron at baseline; after FCM, ferritin, TSAT, and haemoglobin increased, peak VO₂ improved, and iron content rose in the left ventricle (+25%), spleen (+46%), and liver (+18%), while the skeletal muscle and brain were unchanged. Improvements in exercise capacity correlated with haemoglobin and ferritin increments, linking restored iron availability to cardiac energetics rather than generalized tissue loading.¹⁴

Taken together, IV iron consistently improves functional capacity and reduces HF hospitalizations in iron-deficient HF—with signals strongest in more depressed LVEF—while hard mortality benefits remain unproven. The accumulating mechanistic and clinical evidence justifies routine screening for ID and targeted IV repletion in symptomatic HFrEF, as we await adequately powered trials to clarify survival effects and optimize integration with modern polytherapy.

Ongoing trials

With hospitalization benefits established and mechanistic signals emerging, the next unanswered question is whether iron repletion also modulates electrophysiologic risk in systolic HF. RESAFE-HF is designed to address this gap by linking iron correction with device-detected arrhythmia metrics in real-world care. This prospective, single-centre, open-label registry (NCT04974021) enrols adults with HFrEF and ID who receive IV FCM as part of routine practice, with assessments at baseline, 3, 6, and 12 months. The primary endpoint is an iron-response composite (haemoglobin ≥12 g/dL, ferritin ≥50 ng/L, TSAT >20%); secondary endpoints include unplanned HF hospitalizations, ventricular tachyarrhythmias captured by cardiac implantable electronic devices or Holter, echocardiography, functional capacity (VO₂ max, 6-min walk), biomarkers, and quality of life. Baseline data from 106 patients [median age 72 years; 84.9% male; 92.5% New York Heart Association (NYHA) Class II–III] show a high arrhythmic substrate: 17.9% received appropriate device therapy for ventricular tachyarrhythmia in the prior year and 75.5% had frequent, multiform premature ventricular contractions. If iron repletion attenuates arrhythmic burden alongside functional gains, this would broaden the rationale for routine screening and treatment beyond decongestion and exercise tolerance.¹⁵

Conclusions

Across contemporary studies, ID emerges as a modifiable, prognostically relevant target in HF. Definitions that privilege iron availability (TSAT, serum iron, and context of STC) better identify risk than ferritin-centric rules, especially in inflammatory and amyloid phenotypes. Intravenous iron reliably improves symptoms and reduces hospitalizations—benefits most evident with lower LVEF—while survival effects remain unproven. Routine screening and targeted IV repletion in symptomatic HFrEF are justified, alongside harmonization of diagnostic criteria. Priorities include outcome-driven thresholds, integration with modern polytherapy, and adequately powered trials in HFpEF, diabetes, and amyloidosis, including arrhythmic endpoints.

Declarations

Disclosure of Interest

S.v.H. has been a paid consultat to AstraZeneca, Bayer, Besins, Boehringer Ingelheim, BRAHMS, CSL Vifor, Edwards Lifesciences, MSD, Novartis,Novo Nordisk and Pharmacosmos. SvH reports research support from Amgen, AstraZeneca, Boehringer Ingelheim, Pharmacosmos, IMI and the German Center for Cardiovascular Research (DZHK).

Funding

All authors declare no funding for this contribution.