AUTHORS
Annalisa Castagna, Antonino Aparo2, Fabiana Busti3, Rossella Gottardo4, Giacomo Marchi5, Claudia Bozzini6, Alice Vianello7, Fabio Chesini8, Misha Fatima9, Peter Pramstaller10, Domenico Girelli, MD, PHD11
AFFILIATIONS
1 University of Verona, Verona, N/A, Italy
2 University of Verona, Verona, N/A, Italy
3 University of Verona, Verona, N/A, Italy
4 University of Verona, Verona, Italy, University of Verona, Verona, Italy
5 Azienda Ospedaliera Universitaria Integrata Verona, Verona, Italy
6 Azienda Ospedaliera Universitaria Integrata Verona, Verona, Italy
7 Azienda Ospedaliera Universitaria Integrata Verona, Verona, Italy
8 University of Verona, Verona, N/A, Italy
9 University of Verona, Verona, N/A, Italy
10 Eurac Research, Bolzano, Italy, Eurac Research, Bolzano, Italy
11 University of Verona, Verona, N/A, Italy
ABSTRACT
Introduction: Serum ferritin is the cornerstone biomarker for assessing iron stores and diagnosing iron deficiency (ID). Nevertheless, the ferritin thresholds currently used in clinical practice, including the WHO cutoff of <15 μg/L in healthy adults, have been derived largely from historical consensus rather than from physiological evidence (WHO 2020). Recently, Addo et al. identified physiologically derived ferritin thresholds associated with the onset of iron-deficient erythropoiesis, suggesting that this transition may develop at ferritin concentrations substantially higher than current diagnostic cutoffs (Addo OY et al., Br J Haematol. 2026; Addo OY et al., Lancet Glob Health. 2025). We sought to determine whether physiologically derived ferritin thresholds recently reported in blood donors and other populations could be replicated in an independent European population-based cohort using an alternative analytical approach.
Subjects and methods: We analyzed data from the Cooperative Health Research in South Tyrol (CHRIS) study (Pattaro C et al., J Transl Med. 2015; Lundin R et al., Int J Epidemiol. 2025), a population-based cohort aimed at investigating the acquired and genetic basis of common chronic conditions associated with aging. Blood samples of a subset of CHRIS participants were analyzed for biomarkers of iron metabolism, including ferritin, transferrin saturation (TSAT), soluble transferrin receptor (sTfR), hepcidin, and erythroferrone (ERFE). Individuals aged 18-75 years with C-Reactive Protein (CRP) ≤0.5 mg/dL were included, yielding an analytical cohort of 3,036 participants (46.8% men) stratified into women <50 years (n=997), women ≥50 years (n=617), and men (n=1,422). Ferritin was evaluated as the primary predictor of iron status against sTfR, hemoglobin, serum iron, TSAT, transferrin, hepcidin, and erythroferrone. Restricted cubic spline models were used to identify physiological transition points corresponding to the shift from iron-replete to iron-deficient states.
Results: Ferritin showed the most consistent threshold-like behavior among all evaluated biomarkers. In women younger than 50 years, transition points showed striking concordance across multiple indicators of iron-deficient erythropoiesis, clustering between approximately 26 and 30 μg/L. Ferritin thresholds were 26.2 μg/L for sTfR (95% CI 24.7-28.7; p<0.001), 28.1 μg/L for hemoglobin (95% CI 25.5-31.9; p<0.001), 26.4 μg/L for serum iron (95% CI 23.8-29.9; p<0.001), and 30.3 μg/L for TSAT, although the latter showed greater uncertainty (95% CI 26.4-74.4; p<0.001). In contrast, relationships between ferritin and hepcidin or transferrin were continuous and did not identify discrete transition points. Ferritin-ERFE modeling suggested a transition threshold of 19.4 μg/L (95% CI 14.2-90.9), broadly consistent with the thresholds identified using other erythropoietic biomarkers. However, the overall association was not statistically significant (p=0.862), and this finding should therefore be regarded as exploratory. In women aged ≥50 years, ferritin transition thresholds shifted toward higher values. Consistent thresholds were observed for sTfR (52.3 μg/L, 95% CI 46.0-64.9), serum iron (46.5 μg/L, 95% CI 40.6-62.1), TSAT (52.6 μg/L, 95% CI 46.4-95.4), and ERFE (51.5 μg/L; 95% CI 44.2-65.7; p<0.001). Thresholds identified using hemoglobin (60.1 μg/L; 95% CI 48.2-182.5; p<0.001) and transferrin (64.8 μg/L; 95% CI 55.1-193.0; p<0.001) showed wider confidence intervals. In men, transition points occurred at substantially higher ferritin concentrations than in women and were more heterogeneous across biomarkers, precluding a clear physiological interpretation.
Discussion and conclusion: Using a distinct analytical approach in a large population-based cohort, we reproduced the physiologically derived ferritin thresholds recently described by Addo et al., particularly in women of reproductive age. Across multiple complementary iron metabolism biomarkers, the transition toward iron-deficient erythropoiesis consistently occurred at ferritin concentrations around 26-30 μg/L, well above the ferritin threshold currently recommended for diagnosing ID in healthy adults. These findings provide independent support for a physiology-based definition of iron deficiency and strengthen the rationale for re-evaluating the serum ferritin thresholds currently used in clinical practice and research.
DISCLOSURES
Domenico Girelli, MD, PHD: Sanofi - Consultancy (Includes expert testimony); Kedrion - Consultancy (Includes expert testimony); CLS Behring - Consultancy (Includes expert testimony); Takeda - Consultancy (Includes expert testimony)