Hyperferritinemia: Pathophysiology, Etiologies, and Diagnostic Approach

Abstract

Objectives:

Hyperferritinemia is common in adults, yet most lack true iron overload. Confusion among metabolic, inflammatory, genetic, and reactive causes drives under-investigation and over-treatment, compounded by SGLT2 inhibitors and, less certainly, GLP-1 receptor agonists. We synthesize the hepcidin-ferroportin axis as the unifying mechanism and a transferrin saturation (TSAT)-guided pathway for workup.

Methods:

Narrative review with structured PubMed, Embase, and Cochrane searches through April 2026, written to address the SANRA quality domains.

Results:

Ferritin above 10 000 μg/L flags hemophagocytic lymphohistiocytosis (HLH), although the often-cited 90%/96% performance is pediatric; adult HLH-2004 or HScore ≥ 169 reach 82%–95% sensitivity and 60%–94% specificity, lower in ICU populations. The Valenti consensus on metabolic hyperferritinemia is partially validated against clinical outcomes in cohorts applying its grading. A TSAT-guided pathway directs workup toward HFE genotyping, hepatic MRI, or metabolic evaluation; soluble transferrin receptor (sTfR) and the sTfR/log-ferritin index complement ferritin in distinguishing iron deficiency from anemia of inflammation, though sTfR is itself influenced by inflammation.

Conclusions:

The hepcidin-ferroportin axis unifies adult hyperferritinemia under TSAT-guided evaluation, though the algorithm does not capture diabetes risk in C282Y homozygosity. SGLT2 inhibitors alter ferritin interpretation; GLP-1 receptor agonists may do so, though dedicated human evidence remains insufficient. Both warrant documentation at review.

1 |. Introduction

Hyperferritinemia is common in adult primary care and in-patient practice, with population-based prevalence approxmately 20% of men and 3% to 17% of women, depending on threshold and reference population [1]. Only about 10% of cases reflect true tissue iron overload; the remainder are metabolic, inflammatory, hepatic, or reactive, with ferritin acting as an acute-phase protein rather than an index of body iron stores [1]. Many patients with ferritin above 1000 μg/L are under-investigated [1], and values above ~600 μg/L track with increased cardiovascular and all-cause mortality in general-population cohorts [2]. Misclassifying a reactive elevation as iron overload drives unnecessary phlebotomy, imaging, and genetic testing; overlooking a true overload syndrome forfeits definitive treatment.

Three developments motivate an updated synthesis. Valenti and colleagues published a consensus definition of metabolic hyperferritinemia with a three-grade severity system [3]. The 2024 update to the HLH-2004 diagnostic criteria was derived in a pediatric familial-HLH case–control sample; adult external validation remains partial [4]. Sodium-glucose cotransporter 2 (SGLT2) inhibitors and glucagon-like peptide-1 (GLP-1) receptor agonists modulate the hepcidin-ferroportin axis and shift iron-status biomarkers in diabetes, heart failure, chronic kidney disease, and metabolic dysfunction-associated steatotic liver disease (MASLD) [5–7]. To our knowledge, none of these developments have been integrated into a single review for the generalist who orders ferritin testing. Table 1 lays out the differential by mechanism and clinical context, and Table 2 summarizes the principal studies that inform diagnosis and management.

TABLE 1 |.

Differential diagnosis of adult hyperferritinemia by mechanism and clinical context.

Condition	Ferritin, μg/L	TSAT	Key clinical features	Evidence (cohort, n; refs)
Metabolic hyperferritinemia (MHF)	300–1000 (grade 3 > 1000)	Normal, often < 30%	Sex-specific elevation + ≥ 1 cardiometabolic risk factor; 3 grades track cardiometabolic/hepatic outcomes; no phlebotomy.	Valenti consensus; NHANES III n = 12 779; PERSONS MAFLD n = 864; Michigan MASLD n = 7333 [3, 8, 9]
Dysmetabolic iron overload syndrome (DIOS)	> 1000	Normal/mildly ↑ (<45%)	MHF subset; mild-moderate hepatic iron (Perls/T2; 60–150 μmol/g); central adiposity, metabolic syndrome, steatohepatitis.	Mendler 1999, French insulin-resistance cohort n = 161 [10]
Acute infection/systemic inflammation	200–1000 (higher in sepsis)	Low (< 20%)	Acute-phase rise: IL-6/STAT3 hepcidin induction + macrophage ferritin synthesis; normalizes with recovery.	Sandnes 2021 overview; hepcidin mechanism [1, 11]
Heavy or chronic alcohol use	Usually < 1000	Normal or ↑	40%−70% of chronic drinkers; ethanol ↓ hepcidin, ↑ ferritin; liver disease ↓ transferrin, ↑ TSAT; ferritin ↓ ~50% by ~2 weeks abstinence.	Sandnes 2021 overview [1]
Porphyria cutanea tarda (PCT)	Elevated; hepatic iron overload	Normal or ↑	Photosensitive rash, liver dysfunction; 80% acquired; alcohol/smoking/hepatitis C trigger hepatic iron; phlebotomy to <25 μg/L ↓ rash.	Sandnes 2021 overview [1]
Chronic kidney disease/hemodialysis	200–800 (up to ~700 on proactive iron)	Low-normal; functional iron deficiency at TSAT < 20%	↓ Hepcidin clearance + inflammation + IV iron ↑ ferritin; PIVOTAL: proactive IV iron, 700 μg/L cap, incident HD.	PIVOTAL RCT n = 2141 incident HD [12]
Malignancy (solid and hematologic)	500–3000 (higher with cytokine storm/AML)	Low–normal	Tumor IL-6 +macrophage activation sustain hepcidin/ferritin; trend > single threshold (Section 8).	Referral cohorts + mechanism reviews [1, 11]
Cancer-associated (AML, DLBCL, HCC)	AML 4-figure; DLBCL ≥ 500; HCC ≥ 1000 with TSAT ≥ 45%	AML/DLBCL low–normal; HCC ≥ 45%	AML: baseline ferritin predicts OS; DLBCL: ≥ 500 μg/L, worse beyond IPI; HCC: cirrhosis/MASH/HFE marker, not tumor.	Bertoli 2019; Sadighi 2024; Kim 2020; Mitchell 2025 [13–16]
Adult-onset Still’s disease (AOSD)	> 5× ULN; mean ~4752	Variable; not diagnostic	Fever, salmon rash, arthralgia, leukocytosis; glycosylated ferritin ≤ 20% ↑ specificity (assay-limited); Yamaguchi criteria.	Fautrel 2001 n = 78; Yamaguchi 1992 [17, 18]
Hemophagocytic lymphohistiocytosis (HLH)/MAS	> 10000; ≥ 25000 narrows differential	Variable; not diagnostic	HLH-2024: 5/7 + ferritin ≥ 500; adults: HLH-2004 ≥ 4 or HScore ≥ 169 (82%−95% sens; 60%−94% spec); pediatric 90/96 not replicated; malignancy-HLH: OHI best (sCD25 > 3900 + ferritin > 1000).	Allen CE 2008 n = 330; Lachmann 2025; Knaak 2020; Fauter 2022 [4, 19–23]
HFE-related hereditary hemochromatosis (HFE-HC)	> 1000 in expressed disease	↑, ≥ 45%, often > 60%	Biallelic C282Y predominates in northern-European ancestry; TSAT ↑ years before ferritin; phlebotomy target 50–100 μg/L.	HealthIron n = 31 192; HEIRS n = 99 711; UK Biobank [24–26]
Ferroportin disease, SLC40A1 loss-of-function (type 4A)	500–2000	Normal/mildly ↑ (<45%)	Macrophage iron retention: ↑ ferritin, preserved TSAT; relative phlebotomy intolerance; may mimic MHF.	Uguen 2023 (50 variants/308 patients); EASL non-HFE registry n = 458 (18, 67, 68)
SLC40A1 gain-of-function (type 4B; hepcidin-resistant)	> 1000 in expressed disease	↑, ≥ 45%	Hepcidin-resistant ferroportin → parenchymal loading; HFE-HC phenotype, phlebotomy-responsive; Q248H common in 5%−20% of African ancestry, modest effect.	Uguen 2023; EASL non-HFE registry [27–29]
Juvenile hemochromatosis (HJVor HAMP)	> 1000; severe loading by 2nd–3rd decade	↑, ≥ 45%	Biallelic HJV/HAMP loss-of-function abolishes hepcidin signaling; severe early parenchymal overload; cardiomyopathy, hypogonadism.	Piperno 2023; HealthIron [24, 30]
Hereditary hyperferritinemia-cataract syndrome (HHCS)	Markedly ↑ (> 1000), isolated	Normal	FTL 5’-UTR mutation → constitutive L-ferritin without iron loading; bilateral early cataract; phlebotomy harmful (iatrogenic iron deficiency).	Piperno 2023 [30]

Note: Ferritin ranges are typical values across published cohorts and are provided for orientation, not as diagnostic cutoffs; within-condition variability is substantial. Transferrin saturation (TSAT) thresholds follow current EASL guidance (≥45% triggers iron-overload workup).

Abbreviations: AML, acute myeloid leukemia; AOSD, adult-onset Still’s disease; CKD, chronic kidney disease; DIOS, dysmetabolic iron overload syndrome; DLBCL, diffuse large B-cell lymphoma; EASL, European Association for the Study of the Liver; FTH1, ferritin heavy chain 1 gene; FTL, ferritin light chain gene; HAMP, hepcidin antimicrobial peptide gene; HCC, hepatocellular carcinoma; HD, hemodialysis; HEIRS, Hemochromatosis and Iron Overload Screening study; HFE, HFE gene (homeostatic iron regulator); HFE-HC, HFE-related hereditary hemochromatosis; HHCS, hereditary hyperferritinemia-cataract syndrome; HJV, hemojuvelin gene; HLH, hemophagocytic lymphohistiocytosis; HScore, hemophagocytic syndrome score; ICU, intensive care unit; IL-6, interleukin-6; IPI, International Prognostic Index; IV, intravenous; MAFLD, metabolic dysfunction-associated fatty liver disease; MAS, macrophage activation syndrome; MASH, metabolic dysfunction-associated steatohepatitis; MASLD, metabolic dysfunction-associated steatotic liver disease; MetSyn, metabolic syndrome; MHF, metabolic hyperferritinemia; MRI, magnetic resonance imaging; NHANES, National Health and Nutrition Examination Survey; OHI, Optimized HLH Inflammatory index; OS, overall survival; PCT, porphyria cutanea tarda; PERSONS, prospective Chinese MAFLD biopsy cohort (Hangzhou); PIVOTAL, Proactive IV Iron Therapy in Hemodialysis Patients trial; RCT, randomized controlled trial; sCD25, soluble interleukin-2 receptor alpha (CD25); sens, sensitivity; SLC40A1, solute carrier family 40 member 1 (ferroportin gene); spec, specificity; STAT3, signal transducer and activator of transcription 3; TSAT, transferrin saturation; ULN, upper limit of normal; UTR, untranslated region.

TABLE 2 |.

Landmark studies informing the diagnosis and management of adult hyperferritinemia.

Topic/marker/intervention	Study	Population/cohort	Key finding (effect size)	Limitations
TSAT thresholds for HFE genotyping	Zoller 2022 [31]	EASL consensus; European HH cohorts + HEIRS-era data	Genotype if TSAT ≥ 45% + ferritin > 200 (premenopausal women), or TSAT > 50% + ferritin > 300 μg/L (men, postmenopausal); repeat single TSAT (high within-person variability).	Guideline from observational data; GRADE varies by recommendation.
Ferritin > 1000 μg/L cirrhosis risk in HFE-HC	Allen KJ 2008, HealthIron [24]	Australian community cohort, n = 31 192 genotyped	C282Y homozygotes with ferritin > 1000 μg/L: high iron-overload morbidity/cirrhosis; complications in up to 40% of men/13% of women [32].	Genotype ascertainment; few biopsy-confirmed cirrhosis events; European ancestry.
Ferritin > 10 000 μg/L as HLH threshold (pediatric benchmark)	Allen CE 2008 [21]	US pediatric center, n = 330 (ferritin > 500; 10 HLH)	Ferritin > 10 000 μg/L: 90% sens/96% spec for HLH in the pediatric derivation.	Pediatric, single-center; not replicated in adults; selected case mix limits transportability/predictive values.
Adult HLH composite-criteria validation	Knaak 2020; Queffeulou 2026; Lachmann 2025 [20, 22, 33]	Charité ICU 40 HLH/2623; 13-cohort pooled ~2885	HLH-2004 (≥ 4/7) and HScore (≥ 169): 82%–95% sens/60%–94% spec; specificity lowest in unselected ICU.	No prospective adult validation; sepsis/hepatic dysfunction/malignancy meet criteria; single-ferritin thresholds 9000–13 400 μg/L.
2024 update to HLH-2004 criteria	Henter 2024 [4]	Pediatric familial-HLH case–control, n = 366/1064	5/7 features; retain ferritin ≥ 500 μg/L; 96%–99% derivation accuracy; NK-cell function removed.	Pediatric familial-HLH only; no adult validation; not for secondary-HLH estimation.
MHF consensus and severity grading	Valenti 2023; Liu 2024 [3, 8]	NHANES III n = 12 779 + PERSONS biopsy-proven MAFLD n = 864	MHF = sex-specific ferritin ↑ + ≥ 1 cardiometabolic factor; 3 grades track mortality, fibrosis, albuminuria, sarcopenia; advanced-fibrosis OR 1.43 NHANES/1.92 PERSONS; all-cause mortality HR 1.28, grade 2/3.	US + East Asian only; biopsy arm single-country; expert-derived thresholds.
Dysmetabolic iron overload syndrome (DIOS)	Mendler 1999 [10]	French insulin-resistance referral, n = 161	Original DIOS: MHF subset; mild–moderate hepatic iron; ferritin > 1000 μg/L; normal/mildly ↑ TSAT; no HFE overload (Perls ≥ 2 without HFE genotype).	Single-center referral; biopsy ascertainment; pre-MASLD era.
AOSD ferritin signature	Fautrel 2001 [17]	French multicenter, n = 78 suspected AOSD	Mean ferritin 4752 μg/L; glycosylated fraction < 20% (vs. 50%–80%); glycosylated ≤ 20% + ferritin ≥ 5× ULN: 70.5% sens/83.2% spec.	Glycosylated assay not routine; single-country; AOSD–HLH overlap remains.
HFE C282Y prevalence (multi-ethnic primary care)	Adams 2005, HEIRS [25]	n = 99 711 adults across US/Canada	Pooled C282Y homozygotes: ~0.30% overall (0.44% non-Hispanic whites); supports TSAT/ferritin screening.	Cross-sectional; penetrance not an endpoint; ancestry affects transportability.
HFE penetrance in modern population data	Lucas Mitchell 2024, UK Biobank [26]	UK Biobank prospective; HH diagnosis to age 80	Clinical HH penetrance modest; baseline TSAT/ferritin track progression; cumulative HH by 80: 56.4% men (40.5% women); male mortality HR 1.29 (1.12–1.48).	Healthier than general UK; European ancestry; observational.
EMPEROR iron-axis biomarkers (empagliflozin)	Ferreira 2025; Zannad 2022 [34, 35]	Pooled EMPEROR-Reduced + Preserved	Empagliflozin: erythroferrone > 40% ↑; hepcidin/ferritin ↓; concordant TSAT/TIBC/sTfR shifts; TfR1 ~20% ↑ (FDR 2.5 × 10−10); Hb +0.6–0.9 g/dL at 12 weeks.	HF population; ferritin not on Olink panel; observational substudy.
SGLT2i iron handling in CKD (CREDENCE)	Koshino 2025 [36]	CREDENCE post hoc, n = 2416 CKD + T2D	Canagliflozin: TIBC ↑; ferritin ↓ ~11.5% over 12 months; Hb gains; cardiorenal benefit preserved regardless of iron status.	Post hoc; iron outcomes secondary; no SGLT2i class comparison.
GLP-1 RA and ferritin in HH with diabetes	Bain 2023 [6]	Danish HH registry, n = 439 HH + T2D	GLP-1 RA users: lower ferritin across 6529 measurements; association needs replication; mechanism/within-class consistency unknown.	Single registry; no trial; HH + T2D only; no within-class comparison.
Proactive IV iron in hemodialysis (PIVOTAL)	Macdougall 2019 [12]	UK RCT, n = 2141 incident HD	Proactive high-dose IV iron (target ferritin ≤ 700, TSAT ≤ 40%) ↓ death/MI/stroke/HF hospitalization: HR 0.85 (0.73–1.00), p = 0.04; less ESA/transfusion.	HD only; open-label; 700 μg/L cap above pre-PIVOTAL practice.
KDIGO 2026 anemia in CKD	Babitt 2026 [37]	KDIGO consensus + PIVOTAL/post-PIVOTAL data	TSAT + ferritin = iron-status backbone; treat at TSAT ≤ 30% + ferritin ≤ 500 μg/L (hemodialysis), or TSAT < 40% + ferritin < 100 μg/L/TSAT < 25% + ferritin 100–300 μg/L (non-dialysis CKD); TSAT ≤ 20% = iron-restricted erythropoiesis; ferritin caps in post-PIVOTAL range.	Consensus, not a primary cohort; pragmatic thresholds.
BSH investigation of raised ferritin	Cullis 2018 [38]	BSH expert consensus + synthesis	Sequence: TSAT + ferritin → inflammatory markers → HFE genotyping at threshold; ambiguous cases: MRI/sTfR; phlebotomy only for confirmed overload.	Consensus; UK orientation; limited sTfR availability.
ASH 2024 education on hemochromatosis	Girelli 2024 [39]	ASH educational synthesis	TSAT = HH hallmark; sequence: TSAT → HFE genotyping → MRI; supports generalist TSAT-guided algorithm.	Educational, not a guideline; no new cohort.
BIOIRON hemochromatosis classification	Girelli 2022 [40]	BIOIRON expert consensus	HC1–HC5 classification: TSAT > 45% (both sexes) = entry biochemical phenotype; preserves TSAT triage across HFE/non-HFE forms.	Consensus; HC4 (ferroportin disease) separated by mechanism, not TSAT.
sTfR and sTfR/log-ferritin (Thomas) index	Skikne 2011 [38, 41–44]	US/Finland multicenter n = 237; + Lahtiharju 2025 n = 6610	sTfR/log-ferritin index outperforms ferritin or sTfR alone for ACD versus ACD + iron deficiency; ~84%–92% marrow-iron concordance; incremental value mainly with active inflammation.	Method-specific reference ranges; not point-of-care; inter-platform variability; value not uniform.
Real-world phlebotomy registry	Le Gac 2022 [45]	French phlebotomy registry, n = 1059	C282Y homozygotes: 24.4%; remainder metabolic (80% obese, 59% MetSyn); TSAT separates HFE (74.3%) from non-HFE (41.1%).	Single-country; observational; phenotyping depends on documentation.
SGLT2i (dapagliflozin) iron biomarkers in HFrEF	Docherty 2022, DAPA-HF [46]	HFrEF (EF ≤ 40%), NYHA II–IV; n = 3009 with iron biomarkers	Dapagliflozin: ferritin/hepcidin/TSAT ↓, sTfR ↑ over 12 months; benefit by iron status: HR 0.74 (0.58–0.92) deficient versus 0.81 (0.63–1.03) replete; p-interaction 0.59.	Post hoc; iron indices exploratory; no tissue-iron (MRI).
Clinical screening score for ferroportin disease	Landemaine 2022 [29]	French rare-disease network, n = 1306 probands	SLC40A1-variant score (sex, age, no hypertension/diabetes, ferritin, LIC): AUROC 0.83 (0.78–0.88); 93.6% sensitivity at cut-off ≥ 9.5.	Single-country, European; needs external validation; requires LIC.

Abbreviations: ACD, anemia of chronic disease; AOSD, adult-onset Still’s disease; ASH, American Society of Hematology; AUROC, area under the receiver operating characteristic curve; BIOIRON, International Society for Iron in Biology and Medicine; BSH, British Society for Hematology; CI, confidence interval; CKD, chronic kidney disease; CREDENCE, Canagliflozin and Renal Events in Diabetes with Established Nephropathy Clinical Evaluation; DAPA-HF, Dapagliflozin and Prevention of Adverse-outcomes in Heart Failure; DIOS, dysmetabolic iron overload syndrome; EASL, European Association for the Study of the Liver; EF, ejection fraction; EMPEROR, Empagliflozin Outcome Trial in Patients with Chronic Heart Failure; ESA, erythropoiesis-stimulating agent; FDR, false discovery rate; GLP-1 RA, glucagon-like peptide 1 receptor agonist; GRADE, Grading of Recommendations Assessment, Development and Evaluation; Hb, hemoglobin; HC, hemochromatosis; HD, hemodialysis; HEIRS, Hemochromatosis and Iron Overload Screening study; HF, heart failure; HFE, homeostatic iron regulator gene; HFrEF, heart failure with reduced ejection fraction; HH, hereditary hemochromatosis; HLH, hemophagocytic lymphohistiocytosis; HR, hazard ratio; HScore, hemophagocytic syndrome score; ICU, intensive care unit; IDA, iron deficiency anemia; IR, insulin resistance; IV, intravenous; KDIGO, Kidney Disease Improving Global Outcomes; LIC, liver iron concentration; LOF, loss-of-function; MAFLD, metabolic dysfunction-associated fatty liver disease; MASLD, metabolic dysfunction-associated steatotic liver disease; MetSyn, metabolic syndrome; MHF, metabolic hyperferritinemia; MI, myocardial infarction; MRI, magnetic resonance imaging; NHANES, National Health and Nutrition Examination Survey; NK, natural killer; NYHA, New York Heart Association; OR, odds ratio; PERSONS, prospective Chinese MAFLD biopsy cohort (Hangzhou); PIVOTAL, Proactive IV Iron Therapy in Hemodialysis Patients trial; RCT, randomized controlled trial; sens, sensitivity; SGLT2, sodium-glucose cotransporter 2; SLC40A1, solute carrier family 40 member 1; spec, specificity; sTfR, soluble transferrin receptor; T2D, type 2 diabetes; TfR1, transferrin receptor protein 1; TIBC, total iron-binding capacity; TSAT, transferrin saturation; ULN, upper limit of normal.

The hepcidin-ferroportin axis is the convergent node at which metabolic, inflammatory, genetic, and pharmacologic perturbations act, so building the diagnostic approach around it lets one pathway span these etiologies. A transferrin saturation (TSAT)-guided evaluation directs further workup toward HFE genotyping, hepatic MRI, or metabolic evaluation, depending on whether TSAT is elevated, low, or near normal.

The review follows generalist reasoning: ferritin biology (Section 3), genetic iron overload (Section 4), inflammatory and reactive hyperferritinemia (Section 5), metabolic hyperferritinemia and dysmetabolic iron overload syndrome (Section 6), chronic kidney disease (Section 7), cancer-associated hyperferritinemia (Section 8), emerging metabolic therapeutics (Section 9), the TSAT-guided pathway (Section 10), conclusions (Section 11), and knowledge gaps (Section 12). Iron-loading anemias, transfusional iron overload, and thalassemia are reviewed elsewhere; iron-deficiency anemia is invoked only as the contrasting diagnosis disambiguated by soluble transferrin receptor testing.

2 |. Methods

This narrative review of adult hyperferritinemia is written to address the Scale for the Assessment of Narrative Review Articles (SANRA) quality domains [47]. PubMed, Embase, and the Cochrane Library were searched from inception through January 2026, with high-relevance sources identified during revision added through April 2026, using combinations of “hyperferritinemia,” “ferritin,” “hepcidin,” “ferroportin,” “iron overload,” “hereditary hemochromatosis,” “metabolic hyperferritinemia,” “dysmetabolic iron overload syndrome,” “hemophagocytic lymphohistiocytosis,” “adult-onset Still’s disease,” “chronic kidney disease,” “cancer,” “malignancy,” “soluble transferrin receptor,” “SGLT2 inhibitor,” and “GLP-1 receptor agonist.” Results were supplemented by reference-list backtracking from key consensus statements and society guidelines to capture older sources still carrying key quantitative claims.

Eligibility was decided by author consensus, with priority to primary cohort studies, randomized trials, systematic reviews and meta-analyses, and major society guidelines from 2010 onward. Older citations were retained where they remained the primary source for a definition, mechanism, or threshold still in use. Articles not in English, conference abstracts only, or limited to pediatric populations when adult literature was sufficient were excluded. No protocol was pre-registered and no dual independent screening was performed; selection reflects author judgment, and PRISMA or PRISMA-ScR compliance is not claimed.

Where contemporaneous evidence from different settings disagreed on a quantitative point, the range across cohorts is reported with population and design context. Figure 1 summarizes records identified, screened, and included. The reference list contains 80 sources.

FIGURE 1 |.

Study selection. This flow diagram summarizes records identified, screened, and included in a narrative review of adult hyperferritinemia designed to address the SANRA quality domains [47]. Counts are approximate and reflect structured author-consensus selection, not dual independent screening. The Level 7 by-topic row totals more than the 80 included articles because references may be cited in multiple sections. n, number of records; SANRA, Scale for the Assessment of Narrative Review Articles.

3 |. Ferritin Biology and the Hepcidin-Ferroportin Axis

Ferritin is a 24-subunit spherical protein, approximately 450 to 474 kDa, capable of sequestering up to 4500 iron atoms in a redox-inert ferric mineral core [48]. Circulating ferritin is released through unconventional vesicular pathways, including a multivesicular body-exosome route (Figure 2) [49]. Transcription of both subunits is upregulated by iron, cytokines (notably IL-6, IL-1β, TNF-α), and oxidative stress, so circulating ferritin reflects iron status, inflammation, and tissue injury rather than iron load alone [11].

FIGURE 2 |.

Unconventional secretion of ferritin via the TRIM16 multivesicular-body pathway. Systemic stressors, including inflammation (IL-6, IL-1β, TNF-α), oxidative stress, and tissue injury, upregulate FTH1 and FTL transcription (1). After translation and cytosolic assembly (2), the cargo-recognition protein TRIM16 captures ferritin (3) and packages it into ILVs within MVBs (4). MVB fusion with the plasma membrane releases ferritin-loaded exosomes extracellularly (5). This nonclassical route largely bypasses the endoplasmic-reticulum-to-Golgi pathway (6), accounts for a substantial fraction of circulating serum ferritin, and explains why plasma ferritin tracks inflammation and oxidative stress in addition to body iron stores. See Section 3. ER, endoplasmic reticulum; FTH1, ferritin heavy chain 1 gene; FTL, ferritin light chain gene; IL-1β, interleukin-1 beta; IL-6, interleukin-6; ILV, intraluminal vesicle; MVB, multivesicular body; ROS, reactive oxygen species; TNF-α, tumor necrosis factor alpha; TRIM16, tripartite motif-containing protein 16. [Correction added on 20 August 2026 after the first online publication: The Figure 2 caption has been corrected in this version.]

Systemic iron flux is governed by hepcidin, a hepatocyte-derived peptide that restrains iron entry into plasma by binding and degrading ferroportin [11]. Three regulatory inputs converge at the hepatocyte (Figure 3). Iron-sensing operates through bone morphogenetic protein 6 (BMP6) and hemojuvelin signaling through the SMAD1/5/8 axis [32]. Inflammatory induction is mediated by IL-6 via JAK1/2-STAT3, elevating hepcidin independently of iron stores during infection, autoimmune disease, malignancy, and metabolic-inflammatory states [11]. Erythropoietic suppression of hepcidin is mediated by erythroferrone, which is secreted by erythroid precursors under erythropoietin stimulation and sequesters BMP2/6 ligands to relieve hepcidin transcription and release iron for hemoglobin synthesis [50]. The balance among these inputs determines whether hepcidin is appropriately suppressed, appropriately elevated, or, in chronic inflammation, persistently inappropriate for the prevailing iron requirement.

FIGURE 3 |.

The hepcidin-ferroportin regulatory axis. Iron-status sensing regulates hepatocyte HAMP transcription through three inputs. Elevated iron stores stimulate LSECs to release BMP6, which binds the BMP receptor and HJV to activate SMAD1/5/8 signaling (1). Inflammation from infection, autoimmune disease, or malignancy induces IL-6 and activates JAK1/JAK2-STAT3 (2). EPO stimulation (7) drives erythroid precursors to secrete ERFE (8), which sequesters BMP6/BMP2 ligands and inhibits HAMP transcription (3). Secreted hepcidin (4) binds ferroportin (SLC40A1) on enterocytes, reticuloendothelial macrophages, and iron-storing hepatocytes, triggering ferroportin ubiquitination, internalization, and lysosomal degradation. This blocks intestinal iron absorption, macrophage iron recycling, and hepatic iron release 5. Iron is then partitioned into intracellular ferritin, raising intracellular stores and contributing to circulating serum ferritin (6). BMP, bone morphogenetic protein; EPO, erythropoietin; ERFE, erythroferrone; FTH1, ferritin heavy chain 1 gene; FTL, ferritin light chain gene; gp130, glycoprotein 130; HAMP, hepcidin antimicrobial peptide gene; HJV, hemojuvelin; IL-6, interleukin-6; IL-6Rα, interleukin-6 receptor alpha; JAK, Janus kinase; LSEC, liver sinusoidal endothelial cell; RBC, red blood cell; RES, reticuloendothelial system; SLC40A1, solute carrier family 40 member 1 (ferroportin gene); SMAD, mothers against decapentaplegic homolog; SMAD4, mothers against decapentaplegic homolog 4; STAT3, signal transducer and activator of transcription 3. [Correction added on 20 August 2026 after the first online publication: The Figure 3 caption has been corrected in this version.]

The clinical signature of chronic hepcidin elevation is low TSAT despite preserved or expanded tissue iron stores, with sustained compensatory ferritin transcription producing the elevated circulating ferritin that clinicians encounter [11].

Macrophages occupy a central position in this regulatory axis (Figure 3). Under inflammatory or infectious stress, hepcidin-driven ferroportin internalization halts iron export from the reticuloendothelial system. Intracellular ferritin is then rapidly synthesized to buffer the iron that can no longer be released, limiting microbial access at the cost of erythropoietic supply [11].

4 |. Genetic Causes of Iron Overload

Hereditary hemochromatosis (HH) is the prototypical genetic cause of adult iron overload, defined by primary failure of iron-regulatory signaling. Insufficient hepcidin signaling, or peripheral resistance to hepcidin, permits unrestricted dietary iron absorption and progressive parenchymal iron deposition [24].

In HFE-related HH, biallelic HFE mutation uncouples the hepatocyte from the iron signal normally conveyed through the transferrin receptor 2/hemojuvelin/BMP6 cascade. As a result, hepcidin remains inappropriately low while ferroportin on enterocytes and macrophages continues to export iron into the circulation. C282Y produces the more penetrant phenotype, with H63D contributing a milder compound-heterozygous expression. The biochemical signature is elevated TSAT, typically ≥ 45% and often above 60%, which precedes ferritin elevation and parenchymal iron deposition by years [31]. C282Y homozygosity accounts for most clinically expressed disease in populations of northern European ancestry; compound C282Y/H63D heterozygosity, or H63D homozygosity, contributes a smaller and more variable share [25].

4.1 |. Penetrance of HFE C282Y Homozygosity

Two complementary cohorts define penetrance. In the Australian HealthIron cohort of 31 192 adults, approximately 28% of homozygous men and 1% of homozygous women had documented iron-overload-related disease at phenotyping [24]; C282Y homozygosity is associated with iron-related complications in up to 40% of men and 13% of women [32]. In the UK Biobank, 12.1% of male C282Y homozygotes had baseline hemochromatosis diagnoses; cumulative incidence rose to 56.4% by age 80 (40.5% in females), with all-cause mortality hazard ratio 1.29 (95% CI: 1.12–1.48) versus participants without HFE variants [26]. HealthIron is a cross-sectional measure of iron-overload morbidity, whereas the UK Biobank figure is a cumulative incidence of clinical diagnosis. Penetrance is incomplete, biochemical and clinical expression diverge, and morbidity is substantially higher in men.

The HEIRS study of 99 711 primary-care participants across five North American centers reported a pooled C282Y homozygote prevalence of approximately 0.30%, with 0.44% in non-Hispanic whites and approximately 1 in 25 000 in participants of Asian ancestry (estimated from allele frequency) [25]. Clinically meaningful frequencies are essentially confined to populations of northern European descent.

4.2 |. Diagnostic Thresholds and Treatment

The European Association for the Study of the Liver guideline recommends HFE genotyping in any adult with a serum ferritin above 200 μg/L (women) or 300 μg/L (men) and a TSAT ≥ 45% (women) or > 50% (men) [31]; this review applies ≥ 45% to both sexes as a more sensitive operating threshold. Hepatic magnetic resonance $T_2^{*}$ imaging is recommended when ferritin exceeds approximately 1000 μg/L or when persistent elevation cannot be attributed to a non-iron cause [31]. The 1000 μg/L threshold marks the level above which the risk of advanced hepatic fibrosis or cirrhosis rises steeply in C282Y homozygotes [24]. Therapeutic phlebotomy targeting maintenance ferritin between 50 and 100 μg/L remains the standard of care for C282Y-homozygous patients with biochemical or clinical iron-overload disease [24, 31].

4.3 |. Non-HFE Hereditary Iron Overload

Non-HFE genetic causes are individually uncommon but define the workup after a negative HFE genotype. Ferroportin disease, historically classified as type 4 hemochromatosis, splits functionally by SLC40A1 variant type: loss-of-function variants yield macrophage iron retention with elevated ferritin, normal-to-mildly elevated TSAT, and relative phlebotomy intolerance, while gain-of-function or hepcidin-resistant variants mimic HFE-related disease [27].

In the EASL non-HFE registry and accompanying systematic literature review (95 prospective registry patients plus 363 literature-identified cases, 458 total), SLC40A1-associated disease split phenotypically into ferroportin disease (FD) and SLC40A1-related hemochromatosis (SLC40A1-HC). Within the SLC40A1 cohort, roughly two-thirds expressed the FD phenotype and one-third the SLC40A1-HC phenotype, with the latter associated with higher fibrosis risk; phlebotomy did not yield a measurable survival benefit [27].

The hereditary hyperferritinemia–cataract syndrome is a distinct entity that must not be confused with iron overload. Point mutations in the iron-responsive element of the FTL messenger RNA abolish iron-regulatory protein binding, driving constitutive L-ferritin translation without increasing total-body iron [30]. TSAT and hepatic iron concentration are normal, and phlebotomy provides no benefit, instead producing iatrogenic iron deficiency. The diagnosis rests on unexplained isolated ferritin elevation, bilateral early-onset cataracts, a positive family history, and confirmatory FTL sequencing.

5 |. Inflammatory and Reactive Hyperferritinemia

A subset of hyperferritinemic disorders is defined less by iron accumulation than by uncontrolled macrophage activation, in which ferritin indexes inflammatory burden rather than iron stores. The term hyperferritinemic syndrome captures the entities in which extreme ferritin elevation, cytokine storm, and overlapping immunomodulation converge: hemophagocytic lymphohistiocytosis (HLH) and the closely related macrophage activation syndrome, adult-onset Still’s disease (AOSD), septic shock with multiorgan failure, and catastrophic antiphospholipid syndrome. Category recognition matters because the iron-overload reflex (hemochromatosis workup, MRI, phlebotomy) is not recommended here; the relevant investigations are inflammatory, infectious, hematologic, and immunologic. Circulating ferritin tracks macrophage activation intensity, not iron pool size, and falls when the inflammatory driver is controlled rather than when iron is removed [4, 19]. A dedicated Blood review of adult-onset secondary HLH covers the expanding spectrum of precipitants, diagnostic challenges, and management in adults [51].

5.1 |. Hemophagocytic Lymphohistiocytosis

HLH divides into primary (familial) forms, arising from inherited defects in granule-mediated cytotoxicity and presenting chiefly in children, and secondary forms triggered by infection (Epstein–Barr virus is the prototype), malignancy (most often T-cell and natural killer-cell lymphoma), autoimmune disease (where macrophage activation syndrome is preferred), and, increasingly, T-cell-engaging immunotherapies including CAR-T and bispecific antibodies [4]. The 2024 update to the HLH-2004 diagnostic criteria [4] was derived in a pediatric case–control sample of 366 familial-HLH cases and 1064 pediatric controls; the principal change was the removal of natural killer cell function from the panel. The ferritin threshold remains ≥ 500 μg/L; diagnosis requires five of seven features in the familial derivation [4], whereas adult validation applies a lower cut-off of four of seven [20].

The often-cited 90% sensitivity/96% specificity for ferritin above 10 000 μg/L derives from a single-center pediatric study of 330 hyperferritinemic children [21], and adult cohorts have not reproduced this performance. Composite HLH criteria perform at approximately 82%–95% sensitivity and 60%–94% specificity across published adult cohorts [19, 22, 33, 52], with the lower specificity bound reflecting unselected intensive care populations where critically ill adults fulfill multiple criteria from competing causes. The 2025 NEJM review synthesizes adult HLH diagnostic performance and reaffirms that no single criteria set is uniformly applicable across clinical contexts [53]. An adult multicenter validation reported mean sensitivity 83.8% and specificity 87.8% for the revised HLH-2004 criteria (NK-cell activity removed; cut-off of 4) across 13 cohorts; the original HLH-2004 with NK-cell activity retained performed 86.5%/86.1% in the same comparison [20].

At ferritin ≥ 25 000 μg/L the differential narrows: in Fauter and colleagues’ adult multicenter cohort the four causes accounting for most cases were HLH, cytokine release syndromes, acute hepatitis, and infection [19]. In malignancy-associated HLH the Optimized HLH Inflammatory (OHI) composite of soluble CD25 above 3900 U/mL and ferritin above 1000 μg/L achieves 84% sensitivity and 81% specificity in an international cohort of 225 adults with hematologic malignancies, and additionally stratifies long-term mortality [23].

In critically ill adults with confirmed HLH, in-hospital mortality is consistently 40%–60% [22, 33]. Population-based English data reported 1-year survival of 30% in adults aged 75 or older versus 77% in children, with hematologic malignancy nearly tripling the adjusted hazard of death [54]. Malignancy-associated HLH carries the worst prognosis, with 500-day mortality exceeding 80% in patients meeting the OHI composite [23].

5.2 |. Adult-Onset Still’s Disease

AOSD is a systemic autoinflammatory disease in which extreme ferritin elevation is diagnostically suggestive. In Fautrel and colleagues’ prospective cohort of 78 patients, mean ferritin was 4752 μg/L; the glycosylated fraction (normally 50%–80%) falls below 20% in AOSD, and glycosylated ferritin ≤ 20% with total ferritin ≥ 5 × upper limit of normal (ULN) reached 70.5% sensitivity and 83.2% specificity [17]. Glycosylated ferritin assays are not routinely available. Diagnosis rests on the Yamaguchi criteria (96.2% sensitivity and 92.1% specificity in derivation) [18]. AOSD and adult secondary HLH overlap clinically; when both are considered, composite criteria and consultation are preferable to any single ferritin value. The 2022 EULAR/ACR points to consider for HLH/MAS reinforce that no single criteria set is sufficient across all clinical settings and recommend composite evaluation [55].

6 |. Metabolic Hyperferritinemia and Dysmetabolic Iron Overload Syndrome

Valenti and colleagues formalized metabolic hyperferritinemia (MHF) as ferritin elevation above sex-specific reference values in patients with at least one cardiometabolic risk factor and no alternative cause, with a three-grade severity system (< 550, 550–1000, > 1000 μg/L) [3]. The definition has been partially validated against clinical outcomes in two cohorts applying the consensus criteria [8] (NHANES III, n = 12 779, and PERSONS biopsy-proven Chinese MAFLD, n = 864). A US MASLD cohort provides a parallel ferritin-prognostic association without applying the grading [9], with elevation segregating with fibrogenic PNPLA3 and TM6SF2 alleles rather than HFE iron-overload alleles, supporting MHF in MASLD as a marker of the fibrotic milieu rather than true iron overload. Prevalence and grade distribution have been confirmed in a European general-population cohort [56] (Austrian Paracelsus 10 000, n = 8408). A first cardiovascular-endpoint validation has emerged from HFpEF [57] (n = 7498; MHF grade 2 or 3 versus non-MHF MACE hazard ratio 1.43, 95% CI: 1.26–1.62).

Mechanistically, MHF reflects the convergence of inflammation, hyperinsulinemia, and adipose-tissue dysfunction on the hepcidin-ferroportin axis [3]. The clinical signature is ferritin elevation against a paradoxically normal or elevated hepcidin, with mild or absent tissue iron expansion.

6.1 |. Dysmetabolic Iron Overload Syndrome

Dysmetabolic iron overload syndrome (DIOS) is the subset of MHF in which mild to moderate hepatic iron accumulation accompanies metabolic syndrome, ferritin elevation typically above 1000 μg/L, and normal or only mildly elevated TSAT. The entity was first characterized by Mendler and colleagues in a French referral cohort of 161 patients with insulin-resistance-associated hepatic iron overload showing Perls grade 2 or higher staining without HFE-genotype overload, alcohol exposure, or hematologic cause [10]. Hyperferritinemia consistent with MHF or DIOS has since been reported in approximately 20% to 35% of patients with MASLD [8, 9]. DIOS is distinguished from HFE-related HH, in which TSAT is typically 60% or higher [8]. The pathophysiology is best understood as hepcidin resistance rather than hepcidin deficiency.

6.2 |. Therapeutic Stance

The therapeutic position toward MHF and DIOS is conservative and grounded in the absence of benefit from venesection. A French randomized controlled trial assigned 274 patients with hyperferritinemic dysmetabolic syndrome to phlebotomy plus lifestyle counseling versus lifestyle counseling alone over 1 year [58]. Phlebotomy did not improve glycemic control or hepatic steatosis; HOMA-IR paradoxically worsened, and fatigue was reported in 25.3% of phlebotomized patients versus 2.3% of controls. Real-world data from a French national registry of 1059 patients undergoing therapeutic venesection for hyperferritinemia indicate that only 24.4% of recipients carried HFE C282Y homozygosity, with the remainder dominated by metabolic phenotypes (80% centrally obese; 59% meeting metabolic syndrome criteria), and TSAT cleanly separating HFE from non-HFE recipients (74.3% vs. 41.1%) [45]. The implication is that venesection is being delivered to a population in whom the pathophysiology is metabolic rather than iron-overload, and in whom randomized evidence does not support benefit. Current European hepatology and hematology guidance reserves T₂-weighted hepatic MRI and phlebotomy for patients with sustained ferritin elevation, persistently elevated TSAT, and HFE C282Y homozygosity or another confirmed iron-overload cause; uncomplicated MHF and DIOS are managed with weight reduction, treatment of insulin resistance, and metabolic syndrome management [3].

7 |. Chronic Kidney Disease and Hyperferritinemia

Chronic kidney disease is the setting in which serum ferritin is most consistently misinterpreted as a measure of body iron stores. As glomerular filtration rate falls, hepcidin clearance falls in parallel, and the chronic uremic milieu sustains the inflammatory and erythropoietin-deficient signals that elevate hepatic hepcidin [59]. Serum ferritin therefore rises in CKD as a consequence of cytokine-driven transcription and impaired iron mobilization rather than as a marker of expanded body iron stores and can do so in patients whose erythropoiesis is iron-restricted at the marrow level.

The randomized evidence in this setting comes from the Proactive IV Iron Therapy in Hemodialysis Patients (PIVOTAL) trial, which enrolled 2141 adults within 12 months of starting maintenance hemodialysis in the United Kingdom and compared two iron sucrose dosing strategies [12]. The high-dose proactive regimen administered 400 mg of intravenous iron sucrose monthly unless ferritin exceeded 700 μg/L or TSAT reached 40%; the reactive regimen administered up to 400 mg monthly only when ferritin fell below 200 μg/L or TSAT fell below 20%. Over a median follow-up of 2.1 years, the proactive arm met superiority for the primary composite of nonfatal myocardial infarction, nonfatal stroke, hospitalization for heart failure, or death (hazard ratio 0.85, 95% CI: 0.73–1.00, p = 0.04 for superiority) and required lower cumulative doses of erythropoiesis-stimulating agents and fewer red cell transfusions. PIVOTAL findings do not extrapolate without caveat to non-dialysis CKD, peritoneal dialysis, or other intravenous iron formulations.

Post-PIVOTAL guidelines have moved iron-status thresholds in two related directions. The 2021 KDIGO controversies conference accepted that historical ferritin upper bounds of 500 to 800 μg/L were exceeded safely in PIVOTAL and that caps in the 700 μg/L range, paired with TSAT guidance, remain reasonable defaults for proactive intravenous iron in hemodialysis [59]. The 2026 KDIGO Clinical Practice Guideline for the Management of Anemia in CKD continues to use paired TSAT and ferritin as the iron-status backbone. It sets treatment initiation thresholds of TSAT ≤ 30% with ferritin ≤ 500 μg/L for hemodialysis patients and TSAT < 40% with ferritin < 100 μg/L or TSAT < 25% with ferritin 100–300 μg/L for non-dialysis CKD, recognizing that single thresholds perform imperfectly in the inflammation-distorted CKD population [37].

The practical implications follow directly. Ferritin alone does not exclude iron-restricted erythropoiesis: a TSAT of 20% or less in CKD should be interpreted as functional iron deficiency regardless of ferritin. Conversely, isolated ferritin elevation in CKD frequently reflects the cytokine and clearance physiology described above rather than true iron overload and does not by itself justify iron restriction or hemochromatosis workup without corroborating TSAT, genotype, or imaging findings. The iron-mobilizing and hepcidin-suppressive effects of SGLT2 inhibitors in heart failure and CKD are addressed in Section 9 and carried into the diagnostic approach in Section 10.

8 |. Cancer-Associated Hyperferritinemia

Hyperferritinemia in adults with malignancy is overwhelmingly reactive rather than a marker of true iron overload. Tumor cells and tumor-associated macrophages release interleukin-6 locally, sustaining hepatic hepcidin secretion and driving intracellular ferritin synthesis through the interleukin-6/JAK/STAT3 and PI3K/mTOR pathways [60]. Tumor cells additionally release ferritin extracellularly through unconventional secretion and reprogram their own iron-handling machinery toward uptake and storage [61].

Several disease-specific patterns recur. In acute myeloid leukemia, baseline serum ferritin independently predicts overall survival across cytogenetic risk groups and age [13] (n = 525, FTH1/FTL expression). A meta-analysis of 16 studies (n = 8752, transfused patients excluded) reported pooled hazard ratios of 1.99 (95% CI: 1.48–2.66) for overall survival and 2.29 (95% CI: 1.73–3.05) for event-free survival with elevated baseline ferritin [14]. When ferritin in AML enters the four- or five-figure range, malignancy-associated HLH becomes a competing diagnosis with high short-term mortality, and the Section 5 HLH evaluation pathway applies [62]. In diffuse large B-cell lymphoma, a pre-treatment serum ferritin ≥ 500 μg/L identifies a poorer-prognosis subgroup beyond the International Prognostic Index [15]. In hepatocellular carcinoma, ferritin elevation usually reflects the underlying cirrhosis, metabolic dysfunction-associated steatohepatitis (MASH), or HFE-related iron overload rather than the tumor itself; an elevated TSAT (≥ 45%) prompts HFE evaluation and ferritin above approximately 1000 μg/L prompts hepatic iron quantification, per the Section 4 EASL framework [16].

When isolated ferritin elevation is the only abnormality in a patient with known or suspected cancer, the appropriate first step is the TSAT-guided algorithm developed in Section 10: a normal or low TSAT with elevated ferritin supports a reactive picture and does not warrant reflex HFE genotyping [38], while the sTfR/log-ferritin index can identify coexisting iron deficiency masked by inflammation. Ferritin trend is the most informative parameter; a value that continues to rise despite stable disease should prompt re-evaluation for malignancy-associated HLH.

9 |. Emerging Metabolic Therapeutics: SGLT2 Inhibitors and GLP-1 Receptor Agonists

Two drug classes now alter ferritin and iron-status interpretation in populations that constitute much of routine primary care: the SGLT2 inhibitors and the GLP-1 receptor agonists. The SGLT2 inhibitor effect is best characterized for empagliflozin and appears to extend across the class on canagliflozin and dapagliflozin data. The GLP-1 receptor agonist evidence base is comparatively thin.

9.1 |. SGLT2 Inhibitors and the Iron-Mobilization Axis

In a hypothesis-generating review, Packer [63] argued that the iron-deficiency markers seen with SGLT2 inhibition reflect mobilization of stored iron toward erythropoiesis rather than absolute iron deficit, helping reconcile the erythrocytosis observed in clinical trials with the changes in iron indices. A 12-week randomized trial of 52 patients with type 2 diabetes found that dapagliflozin reduced circulating hepcidin and ferritin while increasing erythroferrone, plasma transferrin, and mononuclear-cell transferrin receptor 1 and 2 expression [7]. The strongest mechanistic evidence comes from an analysis of erythropoiesis and iron mobilization across the EMPEROR program [34]. Empagliflozin increased erythroferrone by more than 40% and decreased hepcidin, with parallel decreases in serum iron and TSAT, increases in total iron-binding capacity and soluble transferrin receptor, and hemoglobin gains of approximately 0.6 to 0.9 g/dL at 12 weeks, which mediation analysis linked to the cardiorenal benefit. Ferreira and colleagues noted that in patients with significant baseline inflammation, erythroferrone-mediated hepcidin suppression may be blunted, raising the possibility of erythropoiesis-mediated iron depletion rather than pure mobilization in this subgroup [34]. Patients with baseline iron deficiency showed an attenuated erythrocytic response without loss of clinical benefit.

A pooled Olink Explore 1536 analysis of 1134 patients from EMPEROR-Reduced and EMPEROR-Preserved corroborated the mechanism at the circulating-protein level: empagliflozin produced a near 20% increase in transferrin receptor protein 1 (TfR1) with a false discovery rate of 2.5 × 10⁻¹⁰, alongside concordant increases in erythropoietin and insulin-like growth factor binding protein 1 [35]; ferritin was not on the Olink panel. A validation cohort confirmed the TfR1 increase as reproducible across both EMPEROR trials [64]. A subsequent mechanistic synthesis proposed SIRT1/HIF-2α–mediated renal and hepatic erythropoietin induction as the shared pathway linking SGLT2 inhibitor erythropoiesis to cardioprotection [65].

The effect is not heart-failure-specific. In a post hoc analysis of the CREDENCE trial of canagliflozin in chronic kidney disease and type 2 diabetes, canagliflozin increased total iron-binding capacity and decreased ferritin by approximately 11.5% over 12 months, with consistent hemoglobin gains and cardiorenal benefit regardless of baseline iron-deficiency status [36].

In the DAPA-HF trial (n = 3009 with iron biomarkers at baseline), dapagliflozin decreased ferritin, hepcidin, and TSAT while increasing soluble transferrin receptor and total iron-binding capacity over 12 months, consistent with the erythropoiesis-driven mobilization pattern observed across the EMPEROR program [46]. Clinical benefit was preserved regardless of baseline iron status (hazard ratio 0.74, 95% CI: 0.58–0.92 in iron-deficient versus 0.81, 95% CI: 0.63–1.03 in iron-replete; p-interaction 0.59) [46]. In the EMPATROPISM-FE post hoc analysis of non-diabetic patients with heart failure with reduced ejection fraction, empagliflozin increased myocardial iron content as estimated by cardiac magnetic resonance $T_2^{*}$ imaging while moving systemic iron markers toward a depletion pattern [66]. These findings converge on a model in which SGLT2 inhibition mobilizes iron from storage into erythropoiesis and into myocardial tissue rather than depleting total body iron. In a German Disease Analyzer cohort, adults with type 2 diabetes who initiated an SGLT2 inhibitor had a lower incidence of iron-deficiency anemia over follow-up than those who initiated a dipeptidyl peptidase 4 inhibitor [5]; prospective confirmation is required.

9.2 |. GLP-1 Receptor Agonists and Ferritin

Human iron-status data for GLP-1 receptor agonists remain limited. A Danish nationwide register-based study [6] reported that GLP-1 receptor agonist exposure was associated with lower blood ferritin in people with type 2 diabetes and HH (total HH cohort N = 4714; HH plus T2D N = 698; ferritin sub-cohort N = 439 contributing 6529 measurements). As of early 2026, no human trial has demonstrated a hepcidin-mediated effect of semaglutide, tirzepatide, or dulaglutide on iron status. The current primary mechanism evidence is from an Hfe-knockout murine model in which liraglutide treatment altered hepatic iron accumulation, hepcidin expression, and iron-handling protein abundance [67].

9.3 |. Clinical Interpretation

For the generalist, a falling ferritin in heart failure or CKD on an SGLT2 inhibitor should not be reflexively interpreted as new iron deficiency. The trial-tested biomarker in this population is TSAT below 20%, not ferritin alone, because SGLT2 inhibitors mobilize endogenous stores without depleting body iron [68]. Hemoglobin gains on therapy are expected and in large type 2 diabetes cohorts have not been linked to excess thrombotic events even when erythrocytosis develops [69]. When iron repletion is indicated, oral iron remains effective on background SGLT2 inhibitor therapy. An intravenous-iron cohort in systolic heart failure on SGLT2 inhibitors showed a thromboembolic-event excess in the combined-exposure group [70]; this contrasts with AFFIRM-AHF and IRONMAN and requires prospective confirmation. For GLP-1 receptor agonists, the human evidence is too limited for a specific rule; a falling ferritin more likely reflects weight-loss de-inflammation, and semaglutide’s new MASH indication places it in patients whose hepatic ferritin production is already perturbed. Because the two classes are frequently co-prescribed, both exposures should be documented so clinicians do not misread treatment-induced ferritin shifts as new iron pathology (Section 10).

10 |. Diagnostic Approach

10.1 |. The Diagnostic Algorithm

Transferrin saturation is used here as a categorical classifier of etiology, not as a quantitative estimate of hepatic iron. Saturation reflects the rate of iron flux through plasma; hepatic iron reflects cumulative deposition. Hepatic iron is therefore quantified by magnetic resonance rather than inferred from saturation. A saturation below 45% with a ferritin within the reference range carries a negative predictive value of 97% [32], which is the property this first branch relies on. In 44 809 HEIRS participants, ferritin and saturation together predicted C282Y homozygosity with an area under the receiver operating characteristic curve (AUROC) of 0.91 in men and 0.89 in women [71]. The composite pathway has not been prospectively validated and its overall sensitivity and specificity are unknown, though the branching sequence follows current EASL, ASH, BIOIRON, BSH, and KDIGO guidance [31, 37–40].

The algorithm in Figure 4 is intended for adults with unexplained ferritin elevation in primary care or general medicine, not as a substitute for specialist evaluation. The first branch is emergency triage. A ferritin above 10 000 μg/L in an adult with compatible features should prompt urgent evaluation for HLH; the often-quoted 90%/96% pediatric performance [21] has not been reproduced in adults. Adult composite criteria (HLH-2004 or HScore) outperform any single ferritin threshold (Section 5). Patients with features of AOSD should be evaluated by Yamaguchi 1992 criteria rather than ferritin alone [18]; a glycosylated ferritin fraction below 20% adds specificity where available.

FIGURE 4 |.

Stepwise diagnostic approach to adult hyperferritinemia. This algorithm outlines a triage framework utilizing TSAT as the primary directional screening test rather than a definitive marker of tissue iron load. Initial evaluation prioritizes urgent systemic conditions such as HLH composite criteria (HLH-2004 with at least four of seven features, or HScore ≥ 169) preferred over any single ferritin threshold, as isolated thresholds perform modestly in adults. However, ferritin > 10 000 μg/L with a compatible clinical syndrome should prompt composite-criteria evaluation for HLH. Throughout the diagnostic process, SGLT-2i and GLP-1RA must be assessed, as these medications may affect laboratory values [6, 36, 46]. Subsequent diagnostic branching relies on a TAST threshold of ≥ 45% applied to both sexes. A TSAT ≥ 45% prompts repeat testing to confirm persistent elevation before initiating HFE genotyping or phenotype-directed sequencing for variants like SLC40A1. Because the ≥ 45% TSAT threshold is less sensitive in women, a single normal (sub-threshold) TSAT in a woman with compatible features therefore warrants repeat measurement or HFE genotyping rather than redirection to the metabolic arm [32]. Conversely, TSAT is < 45% (normal) without concerning phenotypes directs the evaluation toward concurrent metabolic dysfunction, reactive etiologies or the possibility of iron deficiency despite of high ferritin. When tissue iron overload remains clinically plausible, particularly for ferritin level exceeding 10 000 μg/L, persistent unexplained elevations, or specific genetic profile, consider hepatic iron quantification via MRI R2* relaxometry. This evidence-informed framework is adapted from the Valenti 2023 consensus [3], the 2024 update to the HLH-2004 criteria [4], and EASL 2022 [31]. The algorithm has not been prospectively validated and its overall sensitivity and specificity remain unknown, though component test performance is cited in the text. The algorithm does not address diabetes risk in C282Y homozygosity; glycemic surveillance is preserved. A1c, glycated hemoglobin; AOSD, adult-onset Still’s disease; BP, blood pressure; CBC, complete blood count; CKD, chronic kidney disease; CMP, comprehensive metabolic panel; CRP, C-reactive protein; DIOS, dysmetabolic iron overload syndrome; EASL, European Association for the Study of the Liver; GLP-1 RA, glucagon-like peptide-1 receptor agonist; HAMP, hepcidin antimicrobial peptide gene; HFE, homeostatic iron regulator gene; HFE-HC, HFE-related hereditary hemochromatosis; HJV, hemojuvelin gene; HLH, hemophagocytic lymphohistiocytosis; HScore, hemophagocytic syndrome score; LIC, liver iron concentration; MHF, metabolic hyperferritinemia; MRI, magnetic resonance imaging; R2*, effective transverse relaxation rate; SGLT2i, sodium-glucose cotransporter 2 inhibitor; SLC40A1, solute carrier family 40 member 1 (ferroportin gene); sTfR, soluble transferrin receptor; TFR2, transferrin receptor 2 gene; TIBC, total iron-binding capacity; TSAT, transferrin saturation.

Outside emergency triage, the workup proceeds as shown in Figure 4. A TSAT at or above 45% with repeat labs directs further evaluation toward HFE genotyping for C282Y homozygosity or compound heterozygosity, with hepatic iron quantification by magnetic resonance imaging [72] as the downstream standard when ferritin exceeds approximately 1000 μg/L. Heavy or chronic alcohol use is a common reactive cause that can also raise TSAT and mimic HH; assess it before routing an elevated TSAT to genotyping [1]. A TSAT below 45% redirects evaluation toward metabolic causes (Valenti framework), inflammatory or reactive causes, or, in young patients with a family history of iron overload and normal TSAT, non-HFE genetic causes including FD (SLC40A1 loss-of-function) [28]. Persistent unexplained elevation in younger patients warrants referral for non-HFE genotyping [31].

The initial laboratory panel includes complete blood count, comprehensive metabolic panel, fasting lipids, hemoglobin A1c, and iron studies (serum iron, TIBC, TSAT, ferritin). C-reactive protein flags concurrent inflammation. Ferritin above approximately 1000 μg/L prompts hepatic iron quantification by magnetic resonance imaging when iron overload is plausible; abnormal hepatic enzymes instead prompt evaluation for hepatic injury and fibrosis [72]. Thresholds derive primarily from northern European ancestry cohorts; clinicians should interpret them in the context of patient background and ancestry-specific allele distributions, including the higher prevalence of the SLC40A1 Q248H variant in adults of African ancestry. Two patient-facing decisions follow. First, document SGLT2 inhibitor or GLP-1 receptor agonist exposure at every ferritin interpretation, since both can lower ferritin and confound its interpretation (Section 9). Second, phlebotomy should not begin without a confirmed iron-overload diagnosis; most contemporary venesection recipients lack true iron overload, and the algorithm above should be completed first (Section 6).

10.2 |. Test Performance and Its Limits

A clinical screening score incorporating sex, age, absence of hypertension or diabetes, ferritin level, and liver iron concentration achieved an AUROC of 0.83 (95% CI: 0.78–0.88) for identifying SLC40A1 pathogenic variants in a French referral cohort of 1306 probands, with 93.6% sensitivity but 49.5% specificity at a cut-off of 9.5 or higher [29]. It may guide when to pursue SLC40A1 sequencing when TSAT is normal.

TSAT has known limitations as a stand-alone test. In population screening, 33% of previously undiagnosed C282Y homozygotes fell below sex-specific thresholds (45% in women, 50% in men) at initial or repeat testing, more often in women, and fasting did not improve performance [73]; a value is best confirmed on repeat and interpreted alongside ferritin.

At a uniform 45% threshold, sensitivity is sex-dependent: TSAT at or above 45% detects approximately 94% of male but only 73% of female C282Y homozygotes [32]. A borderline male value of 45 to 50% still warrants HFE genotyping. A single normal value in such a woman warrants a repeat measurement or HFE genotyping rather than redirection to the metabolic arm. Its correlation with liver iron concentration is moderate overall (r ≈ 0.55) and weaker within C282Y homozygotes (r ≈ 0.29), and adiposity attenuates it further [74]. In a Canadian referral cohort, 58 of 917 previously untreated C282Y homozygotes (6.3%, 95% CI: 4.9%–8.1%) presented with normal TSAT and elevated ferritin; among the 14 who underwent liver biopsy, all 14 had hepatocyte iron grade 3 or 4, and 5.2% had cirrhosis [75], quantifying residual overload in the stratum the algorithm most under-flags.

A Danish prospective cohort followed C282Y homozygotes for liver, cardiac, and metabolic outcomes [76] (422 homozygotes within 132 542 total participants). Normal TSAT and normal ferritin together identified individuals at no significantly increased risk of incident liver disease or cardiovascular disease. In 570 adults with hyperferritinemia, TSAT and ferritin were the main predictors of MRI-measured liver iron content [77]. For incident diabetes, however, the same normal iron indices stratum carried a hazard ratio of 6.49 (95% CI: 2.09 to 20.18) versus participants without HFE variants; these diabetes estimates rest on few events with wide confidence intervals and warrant cautious interpretation.

In a UK Biobank analysis using polygenic instruments (n = 37 287), polygenic-instrumented TSAT showed β = 0.22 (95% CI: 0.19–0.26, p = 3.8 × 10⁻⁴⁶) with magnetic-resonance-estimated liver iron concentration, and undiagnosed C282Y homozygotes carried roughly double the liver iron of diagnosed and treated cases [78]. Outside HFE-related hemochromatosis the diagnostic value of TSAT narrows: in an Austrian liver-clinic cohort of 410 hyperferritinemia patients, low spleen iron on magnetic resonance R₂* (effective transverse relaxation rate), rather than TSAT, was significantly associated with pathogenic variants in non-HFE hemochromatosis genes identified by next-generation sequencing [79], supporting splenic iron imaging as an adjunct when non-HFE genotypes are suspected. The algorithm therefore supports triage for liver and cardiac end-organ disease but does not capture diabetes risk in C282Y homozygosity.

10.3 |. Assessing Coexisting Iron Deficiency

Soluble transferrin receptor (sTfR) and the sTfR/log-ferritin index complement ferritin in distinguishing iron deficiency from anemia of inflammation, though sTfR itself may be influenced by inflammatory states [80] (n = 1001 rheumatology cohort, sTfR-CRP rho = 0.44 vs. ferritin-CRP rho = 0.30). In a prospective multicenter cohort, the sTfR/log-ferritin index distinguished pure anemia of chronic disease from combined disease with concordance against bone-marrow iron staining of approximately 84% to 92% [41]. Contemporary BSH guidance and the 2025 JAMA adult iron-deficiency review recommend sTfR and the sTfR index as second-line tests when ferritin and TSAT cannot resolve a suspected iron-deficiency component [38, 42]. Incremental value is not uniform: in a single-center cohort of 6610 patients, ferritin alone reached an AUROC of approximately 88% to 89%, and sTfR added discrimination primarily in patients with active inflammation [43]. Availability is a practical limitation: sTfR is not point-of-care, and commercial immunoassays use non-interchangeable reference ranges that require standardization against WHO reference reagent 07/202 [44]. Their incremental value in MHF cohorts is not yet established.

11 |. Conclusions

Adult hyperferritinemia is best understood through the hepcidin-ferroportin axis, which ties together the genetic, inflammatory, metabolic, and reactive causes surveyed here and explains why most patients with elevated ferritin do not have true iron overload. TSAT functions as a directional screening test that triages further evaluation toward HFE genotyping and hepatic MRI when elevated, or toward a metabolic and inflammatory workup when normal or low, with glycemic surveillance preserved in C282Y homozygosity even when iron indices are reassuring.

Ferritin above 10 000 μg/L remains the central emergency cue for HLH; the 90%/96% performance originates in a pediatric cohort, and adult composite criteria perform less well. The 2024 update to the HLH-2004 criteria has been partially validated in adults for the NK-cell-removal component [20], with full external validation pending.

The Valenti definition of MHF is supported by outcome analyses in NHANES III and PERSONS [8], parallel prognostic data in MASLD [9], European prevalence data [56], and an HFpEF association with MACE (36, HR 1.43, 95% CI: 1.26 to 1.62), though broad external validation is still lacking. Cancer-associated hyperferritinemia is usually reactive within the same axis.

SGLT2 inhibitors alter ferritin interpretation; GLP-1 receptor agonists may do so, though dedicated human evidence remains insufficient. Prospective validation of TSAT in adult MHF cohorts remains the most decision-relevant evidence gap.

12 |. Knowledge Gaps and Future Directions

Several questions remain open, and the most decision-relevant is whether TSAT tracks MRI-confirmed hepatic iron in the modal Valenti-MHF patient who drives most of the contemporary workload; no post-2020 study has tested this directly (Sections 6 and 10). The Valenti severity grading itself still needs validation outside European referral series and against tissue endpoints, which prospective cohorts pairing the grades with serial hepatic-iron MRI and cardiometabolic outcomes could provide (Section 6).

The 2024 HLH-2004 update likewise awaits adult external validation (Sections 5 and 10); multicenter cohorts pairing the new criteria with HScore, the OHI composite, and adult-calibrated ferritin thresholds in intensive care and oncology would replace inferential estimates with primary data. For SLC40A1-associated FD, genotype-to-phenotype prediction is still undefined: Uguen and colleagues cataloged 50 disease-causing mutations across 308 patients [28], but no variant-level predictors of phenotype, fibrosis, or phlebotomy response exist, and registry studies combining functional annotation with longitudinal follow-up would fill the gap (Section 4).

The metabolic therapeutics raise two further gaps (Section 9). Human data on GLP-1 receptor agonists rest on a single Danish registry in HH with type 2 diabetes [6]; no clinical trial has yet shown ferritin changes and the primary mechanistic evidence remains murine, so cardiometabolic-trial substudies with ferritin, TSAT, and hepcidin as prespecified endpoints are needed. Beyond the myocardium, tissue iron redistribution under SGLT2 inhibitors is uncharacterized: EMPATROPISM-FE measured myocardial iron by cardiac MRI $T_2^{*}$ [66], but hepatic, splenic, and CNS stores are unstudied, and substudies extending $T_2^{*}$ or quantitative susceptibility mapping would show whether mobilization is heart-specific or systemic.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.