In a retrospective cohort study of 146 patients with chronic heart failure treated with intravenous iron, researchers found that classifying patients into three study-defined mechanistic iron-deficiency phenotypes was not significantly associated with one-year clinical outcomes. The investigation, conducted at Kütahya Health Sciences University Evliya Çelebi Training and Research Hospital in Türkiye, evaluated patients receiving a 1,000-milligram dose of ferric carboxymaltose in routine practice. Led by Emrah Kaya and colleagues, the researchers evaluated whether operational proxies for proposed biological pathways could predict risk after treatment, while also assessing whether post-infusion changes in transferrin saturation provided prognostic utility. Within the 365-day follow-up, 32 individuals, or 21.9 percent of the cohort, experienced the primary composite outcome of death, heart failure hospitalization, or emergency department care. Although marked hypoferremia showed a directional trend toward increased events, primary multivariable estimates remained statistically non-significant.
Iron deficiency is a common comorbidity in heart failure that impairs functional capacity, quality of life, and clinical outcomes. Although randomized trials established intravenous iron as a beneficial therapy that improves symptoms and reduces heart failure hospitalizations, subsequent large trials including HEART-FID and FAIR-HF2 failed to meet primary endpoints. This divergence has focused scrutiny on existing diagnostic standards. Clinical guidelines define iron deficiency through static circulating markers, requiring serum ferritin below 100 micrograms per liter, or 100 to 299 micrograms per liter when transferrin saturation is below 20 percent. However, circulating ferritin is strongly influenced by systemic inflammation and oxidative stress, often failing to mirror biologically available tissue iron. Conversely, transferrin saturation reflects circulating iron accessible for cellular uptake and correlates more closely with functional iron availability, prompting interest in alternative diagnostic and phenotyping strategies.
To capture underlying biology beyond static systemic indices, a conceptual framework described by Milton Packer identified three distinct mechanistic pathways in iron-deficient heart failure. Type 1 represents severe systemic hypoferremia that impairs circulating iron supply, leading to cardiomyocyte iron depletion. Type 2 involves functional iron deficiency associated with inflammation and physical deconditioning, driving skeletal muscle iron depletion despite modest reductions in circulating levels. Type 3 encompasses advanced cardiomyopathy characterized by disrupted intracellular iron handling within cardiomyocytes, causing tissue vulnerability even under mild systemic hypoferremia. In this mechanistic paradigm, transferrin saturation is proposed as the primary determinant of therapeutic response to intravenous iron. However, because the framework was formulated as a pathophysiological concept without a validated clinical scoring algorithm, its practical utility in predicting outcomes following intravenous iron therapy remained untested in clinical cohorts.
The cohort study evaluated 146 consecutive adult patients with chronic heart failure who received intravenous ferric carboxymaltose between January 19, 2022, and March 19, 2024. The mean age was 69.6 years, and 42.5 percent were female. Most patients exhibited reduced systolic function, with 73.3 percent classified with reduced ejection fraction, 5.5 percent with mildly reduced ejection fraction, and 21.2 percent with preserved ejection fraction. Advanced symptoms were prevalent, with 51.4 percent presenting in New York Heart Association functional class III or IV, and median N-terminal pro-B-type natriuretic peptide reached 1,262 picograms per milliliter. Guideline-directed medical therapy was widely utilized: 81.5 percent received renin-angiotensin inhibitors, 92.5 percent beta-blockers, 74.0 percent mineralocorticoid receptor antagonists, and 61.6 percent sodium-glucose cotransporter 2 inhibitors. Overall, 97.3 percent met guideline criteria for iron deficiency, while four patients were treated based on clinical judgement.
Because underlying biological mechanisms can coexist within the same patient, the researchers developed study-specific operational proxies and permitted non-mutually exclusive classifications. Type 1 was assigned to patients with baseline transferrin saturation of 16 percent or less, representing marked hypoferremia. Type 2 was designated when baseline transferrin saturation was 20 percent or less alongside functional impairment, defined by New York Heart Association class III or IV symptoms or documented deconditioning in class II. Type 3 was assigned to patients with reduced ejection fraction of 35 percent or lower and documented advanced heart failure. In total, 66 patients met Type 1 criteria, 52 met Type 2, and 62 met Type 3. Overlap was substantial: 61 patients met criteria for a single phenotype, 28 satisfied two, and 21 exhibited features of all three. Thirty-six patients fulfilled none of the phenotype definitions, forming the reference group.
During the one-year follow-up, the primary composite outcome occurred in 32 patients, representing an overall event rate of 21.9 percent. Among individual composite components, six patients died, 24 required hospitalization for worsening heart failure, and 26 experienced emergency department visits for worsening heart failure, with components not being mutually exclusive. The median duration from intravenous iron infusion to the first qualifying event was 102 days, with an interquartile range of 58 to 164 days. An additional 13 patients experienced their first event after day 365, including eight during the extended protocol follow-up. Event rates were 25.8 percent in Type 1 patients, 25.0 percent in Type 2, and 27.4 percent in Type 3, compared with 16.7 percent in the unexposed reference group. Patients experiencing events had lower ejection fractions, more advanced functional impairment, and higher natriuretic peptide levels.
Multivariable logistic regression adjusted for age, sex, left ventricular ejection fraction, New York Heart Association class III or IV, and anemia revealed no statistically significant association between any phenotype and the composite outcome. The adjusted odds ratio for Type 1 marked hypoferremia was 2.60, with a 95 percent confidence interval of 0.84 to 8.01 and a p-value of 0.096. For Type 2, the adjusted odds ratio was 0.40 with a confidence interval of 0.11 to 1.42, while Type 3 yielded an odds ratio of 0.66 with a confidence interval of 0.21 to 2.11. Conversely, New York Heart Association class III or IV demonstrated a borderline association, with an adjusted odds ratio of 3.41 and a confidence interval of 0.95 to 12.18. The apparent receiver operating characteristic area under the curve of 0.68 attenuated to 0.58 following 500 bootstrap resamples.
Sensitivity analyses reinforced the consistency of these patterns across alternative statistical approaches. In Cox proportional hazards regression for time to first event, the hazard ratio was 2.30 for Type 1, 0.43 for Type 2, and 0.69 for Type 3. Firth’s penalized-likelihood logistic regression yielded an odds ratio of 2.36 for Type 1, while adjusting for renal function and natriuretic peptides produced an odds ratio of 2.21. When testing alternative Type 1 definitions, a transferrin saturation cutoff of 15 percent or less achieved statistical significance with an odds ratio of 3.12 and a confidence interval of 1.01 to 9.64. Over the extended protocol-defined follow-up, which recorded 40 events across a median observation period of 464 days, Cox regression revealed a significant hazard ratio for Type 1 of 2.99 with a confidence interval of 1.19 to 7.51, reflecting disease severity rather than standalone predictive power.
The investigation also conducted exploratory analyses examining whether the change in transferrin saturation after intravenous iron predicted clinical outcomes among 139 patients with follow-up laboratory measurements. In this subset, mean transferrin saturation increased from 22.2 percent at baseline to 35.6 percent post-treatment, yielding a mean increase of 13.3 percentage points. Despite this biochemical response, multivariable logistic regression demonstrated no association between the change in transferrin saturation and the composite outcome, with an adjusted odds ratio of 1.00 per 10 percentage-point increase and a confidence interval of 0.80 to 1.24. Further adjustment for baseline transferrin saturation yielded an odds ratio of 0.97. Restricting the cohort to 87 patients with measurements taken between 28 and 365 days after infusion similarly showed no association, yielding an adjusted odds ratio of 1.04. Post-treatment absolute saturation was likewise not associated with outcomes, producing an adjusted odds ratio of 0.96.
The study team emphasized that these null findings for transferrin saturation changes cannot be interpreted as proof that biomarker responses lack prognostic value, citing critical methodological timing limitations. In routine practice, follow-up blood sampling occurred at an unstandardized median interval of 147 days after infusion. Consequently, in 22 of the 27 patients who suffered an event in the biomarker analysis set, the event took place prior to the post-treatment blood test. This sequence created reverse causation, as post-event values likely reflected acute decompensation, inflammation, or inpatient treatment rather than predicting subsequent risk. In an exploratory analysis restricted to 112 patients who were event-free at follow-up measurement and observed for one year thereafter, only nine events occurred, yielding an unadjusted hazard ratio of 1.07. Furthermore, excluding seven patients who died before follow-up testing introduced immortal-time bias, precluding definitive conclusions regarding response dynamics.
Analyses assessing subgroup interactions and discrimination confirmed that circulating iron parameters provided poor predictive discrimination for individual clinical outcomes. A model evaluating phenotype-by-change interactions demonstrated no statistically significant effect modification across phenotypes, yielding a non-significant likelihood-ratio test across three degrees of freedom with a p-value of 0.371. In receiver operating characteristic analyses prespecified with lower values denoting higher risk, baseline transferrin saturation achieved an area under the curve of 0.57, while post-treatment saturation achieved 0.55 and the change score yielded 0.50. Serum ferritin demonstrated an area under the curve of 0.44, suggesting an inverse directional trend characteristic of acute-phase reactant behavior, though its confidence interval encompassed 0.50. Incorporating the change in transferrin saturation into a clinical baseline model comprising age, sex, ejection fraction, functional class, and anemia did not alter discriminatory accuracy, leaving the area under the curve unchanged at 0.648.
Subject of Research: Prognostic relevance of mechanistic iron deficiency phenotypes and transferrin saturation dynamics in heart failure
Article Title: Intravenous IRON therapy outcomes according to iron deficiency type in HEART failure (IRON-HEART)
Article References: Kaya, E., Erkip, E. B., Küçük, B., Doğru, F., Güven, U., Furtuna, Ö., Barlas, D. A., Kahraman, F., Demir, M., Sönmez, G., Durmuş, H. İ., Astarcıoğlu, M. A., & Şen, T. (2026). Intravenous IRON therapy outcomes according to iron deficiency type in HEART failure (IRON-HEART). Frontiers in Cardiovascular Medicine, 13, 1944221. https://doi.org/10.3389/fcvm.2026.1944221
Image Credits: AI Generated
DOI: 10.3389/fcvm.2026.1944221
Keywords: heart failure, iron deficiency, intravenous iron, ferric carboxymaltose, transferrin saturation, ferritin, mechanistic phenotypes, hypoferremia, cardiomyopathy, cardiology, NYHA functional class, left ventricular ejection fraction, NT-proBNP, hospitalization, biomarker dynamics, cohort study, receiver operating characteristic, prognostic modeling
News Source: Ophelia Keating. (November 19, 2026). Mechanistic Iron Deficiency Phenotypes Show Weak Prognostic Value in Heart Failure. Scienmag.



