Introduction and Context
Haemochromatosis arthropathy (HA) is the distinctive joint disease that occurs in many people with hereditary haemochromatosis (HH), the commonest genetic iron overload disorder in populations of European ancestry. HA often targets the metacarpophalangeal (MCP) joints, distal interphalangeal (DIP) joints and weight-bearing joints such as the ankle and hip and can be mistaken for osteoarthritis (OA) or calcium pyrophosphate deposition disease (CPPD). Until now, there have been no formal, internationally developed classification criteria to identify HA for research purposes—an omission that has hampered cohort definition, comparative studies and clinical trials.
The European Alliance of Associations for Rheumatology (EULAR) task force published the first multicentre, evidence-based EULAR 2025 classification criteria for haemochromatosis arthropathy (Kiely et al., Ann Rheum Dis. 2025). The international effort responds to three practical needs: (1) a standard way to define HA for research and epidemiology, (2) improved ability to distinguish HA from primary generalized OA and CPPD in patients with iron loading, and (3) the availability of robust multicentre data that allow formal derivation of data-driven criteria.
Key background sources that informed the project include long-standing reviews on hereditary haemochromatosis pathophysiology and clinical presentation (Pietrangelo A., NEJM 2004; Adams & Barton, Lancet 2007) and the EULAR standardized task force process used to develop the criteria.
New Guideline Highlights
– The product is a point-based classification model specifically intended to classify HA in patients who are HFE C282Y homozygous with clinical joint symptoms and evidence of systemic iron loading. The criteria were not intended as diagnostic rules for routine clinical care but rather to standardize inclusion in research studies.
– The derivation cohort comprised 154 patients with haemochromatosis arthropathy and 120 patients with disease mimics (primary generalized osteoarthritis or CPPD), recruited across multiple international centres.
– The final model uses 8 variables spanning age at symptom onset, characteristic clinical findings, joint-level radiographic features (MCP, DIP and ankle), and a history of hip or ankle arthroplasty or surgery. Individual variables are assigned points that sum to a maximum of 11.
– A threshold score of ≥5 out of 11 classifies a patient as having haemochromatosis arthropathy for research purposes; this cut-off yields 93.3% specificity and 71.4% sensitivity in the derivation cohort.
– The task force emphasizes that these criteria require external validation in independent cohorts prior to broad adoption for research beyond populations similar to the derivation set.
Updated Recommendations and Key Changes
This is the first formal set of classification criteria specific to HA; therefore, rather than updating a prior version, the EULAR document establishes a standard. Key advances compared with prior practice are:
– Formalization: Prior to these criteria, researchers used heterogeneous composite definitions (e.g., C282Y genotype + joint symptoms + variable imaging or biochemical thresholds). The EULAR criteria standardize item selection, scoring and a validated threshold.
– Data-driven item selection: Candidate items were screened by systematic review, Delphi consensus and then empirically tested in a well-characterized derivation cohort. The final item set balances statistical discrimination and face validity (disease Gestalt).
– Joint- and surgery-focused items: The model explicitly includes radiographic features at MCP, DIP and ankle joints and prior hip/ankle surgery—variables that reflect the typical joint pattern and long-term burden of HA but are often underused in research definitions.
Topic-by-Topic Recommendations
Overall purpose and scope
– Intended use: These criteria are intended to classify patients with haemochromatosis arthropathy for research inclusion and cohort definition. They are not diagnostic guidelines for routine clinical decision-making and do not replace clinical judgment.
– Target population: Adults with known HFE C282Y homozygosity, joint pain and evidence of iron loading (e.g., elevated serum ferritin/transferrin saturation or liver iron—recruitment required these baseline features). The derivation cohort included only C282Y homozygotes; applicability to other genotypes (compound heterozygotes or non-HFE iron overload) remains to be tested.
Figure 1The iron fist. The iron fist is a physical sign due to incomplete full flexion of the second and/or third MCP joints. When making a fist, the PIP joints of the second and/or third fingers are raised above the level of the PIP joints of the fourth and fifth fingers, and the fingertips of the second and third fingers may not touch the palm. PIP, proximal interphalangeal, MCP, metacarpophalangeal.
Figure 2 Hook osteophytes and joint space narrowing. Hook osteophytes at MCP 3 (A), attached to the radial aspect of the metacarpal head, and at the ankle joint (B), attached to the anterior distal tibia (see red arrows). In these images, MCP 2 and 3 and the ankle joints all have grade 2 joint space narrowing. MCP, metacarpophalangeal.
Figure 3 Subchondral cysts and joint space narrowing. Subchondral cysts of ankle (A) and MCP joints (B) (see black arrows). In this image, MCP joints 2 and 3 have a joint space narrowing grade 3, and there is a hook osteophyte at MCP 3. MCP, metacarpophalangeal.
| Variable | Score |
|---|---|
| Joint symptom onset before 50 y of age | 2 |
| Iron fist | 1 |
| Absence of any DIP joint soft-tissue swelling and bony swelling and deformity | 2 |
| Hip or ankle surgery | Ankle only: 2 Hip only: 1 Ankle and hip: 2 |
| MCP joint 2 or 3 or 4 or 5 tenderness, clinical examination | 1 |
| X-ray: any MCP joint—joint space narrowing grade 3 or surgery or subchondral cyst | 1 |
| X-ray: ankle joint—joint space narrowing grade 3 or subchondral cysta | 1 |
| X-ray: hook osteophytes MCP joint 2 or 3 or ankle jointa | MCP 2/3 only: 1 Ankle only: 2 MCP 2/3 and ankle: 2 |
| Classification criteria met | Total score ≥5 |
Table 1 Classification criteria for haemochromatosis arthropathy
DIP, distal interphalangeal, MCP, metacarpophalangeal.
To be used in people with joint pain, the C282Y homozygous HFE mutation and evidence of iron overload. Specificity 93.3%; sensitivity 71.4%.
a Item not scored if ankle surgery either side.
| Variable | Definition |
|---|---|
| Joint symptom onset before 50 y of age | Age of onset of persisting pain, stiffness or swelling in any joint. Exclude transient joint symptoms, eg, due to trauma |
| Iron fist | Score 1 whether unilateral or bilateral |
| Absence of any DIP joint soft-tissue swelling and bony swelling and deformity | All DIP joints clinically normal with no palpable synovial or bony swelling or deformity. Intention to exclude nodal osteoarthritis or psoriatic arthritis. DIP joint deformity following trauma is allowed |
| Hip or ankle surgery | Joint fusion and arthroplasty For ankle score 2 whether unilateral or bilateral surgery. For hip score 1 whether unilateral or bilateral surgery For ankle and hip surgery score 2 whether unilateral or bilateral surgery at each joint site Surgery should have been performed for primary arthropathy, do not score if surgery was performed as a consequence of trauma |
| MCP joint 2 or 3 or 4 or 5 tenderness, clinical examination | Score 1 if tenderness on clinical examination of any of the second to fifth MCP joints of either hand |
| X-ray: any MCP joint – joint space narrowing grade 3 or surgery or subchondral cyst | Score 1 if any MCP joint has grade 3 joint space narrowing, or surgical fusion or arthroplasty or subchondral cyst/s (see reference) |
| X-ray: ankle joint – joint space narrowing grade 3 or subchondral cyst⁎ | Ankle joint space narrowing grade 3 or subchondral cyst/s in any of 5 regions: lateral view, anterior, and posterior tibiotalar space, Mortise view, medial and superior tibiotalar and talofibular joint space (see reference) Score 1 whether unilateral or bilateral, and whether either joint space narrowing grade 3 or subchondral cyst or both. *This item is not scored if either ankle has had surgery. Only score for primary ankle arthritis; do not score if ankle arthritis is a consequence of trauma. |
| X-ray: hook osteophytes⁎ MCP joint 2 or 3 or ankle joint | MCP joint: osteophyte attached to the radial aspect of the second or third metacarpal head, characteristically large and hook-like in appearance. Score 1 if present in one or more second or third MCP joints (see reference [25]) Ankle joint: osteophyte attached to the distal tibia seen on lateral view, either anterior or posterior. Score 2 if present in either or both ankles (see reference [24]) *This item is not scored at the ankle if either ankle has had surgery. A score of 1 is still given for MCP hook osteophytes in a person with ankle surgery |
Essential structural elements of the criteria model
– Item domains: The 8-item score draws from three domains:
– Age at symptom onset (early onset favors HA in the selected population);
– Clinical features and joint distribution (notably MCP and ankle involvement);
– Radiographic features at key joints (MCP, DIP, ankle) and a history of hip or ankle surgery attributable to severe joint disease.
– Scoring and threshold: Each item contributes points to a total possible of 11. The cut-point ≥5 was chosen by the task force because it achieves high specificity (93.3%) while retaining acceptable sensitivity (71.4%) in the derivation cohort—appropriate for classification purposes where specificity is often prioritized to ensure homogenous cohorts.
How to apply the criteria in research
– Prerequisites: Confirm HFE C282Y homozygosity and evidence of iron loading. Exclude alternative diagnoses that fully explain the joint disease (e.g., advanced inflammatory arthritis unrelated to iron overload).
– Data collection: Use standardized clinical assessment (joint counts, history of joint symptoms and surgeries) and radiographs of the MCPs, DIP and ankles (or imaging modality specified by study protocol) to score items.
– Interpretation: Patients who meet the prerequisites and score ≥5 should be considered to have HA for inclusion in research cohorts studying the arthropathy of hereditary haemochromatosis.
Special populations and caveats
– Age and demographics: The derivation cohort characteristics (age distribution, ancestry) influenced item selection; investigators should be cautious when applying criteria to very young adults, older adults with advanced OA, or non-European ancestry groups until validation data are available.
– Non-C282Y genotypes and secondary iron overload: Because only C282Y homozygotes were included in derivation, the criteria are not validated for other genetic variants (e.g., H63D compound heterozygotes) or for patients with secondary iron loading (e.g., transfusion-related iron overload).
– Overlap with CPPD: Chondrocalcinosis and CPPD frequently coexist with HA and can confound evaluation. The derivation sample explicitly included CPPD as a mimic group to improve discriminatory performance; nevertheless, careful phenotyping is needed in research settings.
Expert Commentary and Insights
Consensus process and expert rationale
– The task force combined systematic review, multi-round Delphi consensus and empirical testing in real-world clinical cohorts. This mixed-methods approach ensured that items were both evidence-based and consonant with clinician experience (face validity).
– The prioritization of high specificity reflects the task force’s view that classification criteria should produce well-defined research groups with low contamination by OA or CPPD—important for mechanistic studies and trials of targeted interventions.
Areas of agreement
– There was broad agreement that genotype confirmation (C282Y homozygosity) and evidence of iron loading are essential prerequisites for a research definition of HA.
– The prominence of MCP joint involvement and the inclusion of hip/ankle surgery reflect long-recognized clinical signals of HA severity and topography.
Controversies and open questions
– Diagnostic vs classification use: Some clinicians may be tempted to apply criteria diagnostically. The task force reiterated that criteria are for research classification; real-world diagnosis should remain a clinical synthesis of history, imaging, laboratory iron studies and response to therapies such as phlebotomy or joint-directed care.
– Imaging modalities and thresholds: The criteria rely largely on plain radiography items. The role of MRI, ultrasound or CT (which may detect subclinical changes) remains an open research question. Future validation studies might test imaging variants or add quantitative measures.
– External validity: The derivation cohort size (154 HA cases, 120 mimics) is solid for an uncommon condition, but external cohorts with broader geographic/ethnic diversity and inclusion of other HFE genotypes are needed.
Practical Implications for Clinical Practice and Research
For clinicians
– Diagnosis and management of an individual patient with joint pain and iron overload should not rely solely on these criteria. However, clinicians can use the items as a checklist to appraise whether a patient’s pattern of joint involvement is typical of HA, and to communicate with researchers about case ascertainment.
For researchers
– Standardized enrollment: The criteria provide a reproducible approach to define cases for studies of HA pathogenesis, natural history, imaging biomarkers and therapeutic trials.
– Trial enrichment: High specificity at the chosen threshold makes the criteria suitable for recruiting homogeneous patient populations to trials of disease-modifying interventions for HA-related joint disease.
For epidemiologists and health services planners
– Better case definition will improve estimates of HA prevalence, disease burden, healthcare utilization (e.g., hip/ankle arthroplasty rates) and cost analyses in people with hereditary haemochromatosis.
Practical Example (Vignette)
John, a 52-year-old man with known HFE C282Y homozygosity and elevated ferritin, presents with chronic pain and stiffness of his 2nd and 3rd MCP joints, progressive ankle pain on walking and a prior right hip replacement for non-inflammatory arthritis at age 49. His hand radiographs show characteristic MCP bony changes. Applying the EULAR HA classification items (after scoring the 8 variables) yields a total of 6 points—above the ≥5 threshold—so in a research context John would be classified as having haemochromatosis arthropathy. Clinically, his care would still integrate iron management, symptomatic joint therapies and rehabilitation; classification facilitates enrollment in HA-focused studies.
Limitations and Research Agenda
– Validation required: The authors and task force stress that external validation is required before widespread research adoption in populations that differ from the derivation cohort.
– Broader genotypes and secondary iron overload: Future work must examine applicability to H63D carriers, compound heterozygotes and secondary iron-loading states.
– Imaging refinement: Comparative studies should test whether adding MRI or ultrasound parameters can improve sensitivity without compromising specificity.
– Longitudinal performance: Prospective studies are needed to assess how well classification predicts outcomes such as progressive joint damage, need for arthroplasty, pain trajectories and functional decline.
Conclusions
The EULAR 2025 classification criteria for haemochromatosis arthropathy represent a major step forward: for the first time, investigators have a consensus, empirically derived, reproducible instrument to classify HA in research settings. The point-based model (8 variables, 11-point scale; threshold ≥5) favors specificity to generate homogenous cohorts. Clinicians should appreciate the distinction between classification and diagnosis; researchers should prioritize external validation and extension to broader genotypes and imaging strategies. With additional validation, these criteria are likely to accelerate research into the mechanisms, natural history and treatment of joint disease in hereditary haemochromatosis.
References
– Kiely PD, Finzel S, Farisogullari B, Carroll GJ, McCarthy G, Stack J, Parisi S, Porto G, Richette P, Nagy G, Weidl M, Rosenthal A, Guggenbuhl P, Banaszkiewicz KJ, Engelhardt S, Shearman JD, Mitchell D, Barker J, Brueton V, Butzeck B, Coathup P, Don H, Dowsett J, Duncan M, Dunleavy T, Fish I, Hoggarth A, McKinnon M, Minter J, Osborne T, Smith M, Wright C, Machado PM. EULAR 2025 classification criteria for haemochromatosis arthropathy. Ann Rheum Dis. 2025 Nov 4:S0003-4967(25)04434-6. doi: 10.1016/j.ard.2025.10.003 IF: 20.6 Q1 . Epub ahead of print. PMID: 41193334 IF: 20.6 Q1 .
– Pietrangelo A. Hereditary hemochromatosis — a new look at an old disease. N Engl J Med. 2004 Jun 10;350(23):2383–97. doi:10.1056/NEJMra031573 IF: 78.5 Q1 .
– Adams PC, Barton JC. Haemochromatosis. Lancet. 2007 Oct 6;370(9586):1855–60. doi:10.1016/S0140-6736(07)61930-5
(Readers should consult the full EULAR publication for detailed item definitions, scoring table and methodological appendices when applying the criteria in research settings.)
