Erosive hand osteoarthritis (EHOA) is a subset of hand osteoarthritis (HOA) that is defined by the presence of at least one erosion in interphalangeal (IP) joints on plain radiographs and occurs primarily in postmenopausal women (Punzi et al., 2004; Favero et al., 2022). The most affected joints in EHOA patients are the proximal (PIP) and distal interphalangeal (DIP) joints, typically sparing the metacarpophalangeal joints. A distinguishing feature of EHOA is the simultaneous involvement of multiple joints, unlike in typical HOA, which generally affects joints progressively and individually (Favero et al., 2022). The second and third fingers are most affected, often symmetrically, followed by the fourth and fifth fingers (Figure 1). The overall prevalence of EHOA is currently estimated at about 2.8%, though it may reach 10.2% among individuals with symptomatic OA (Kwok et al., 2011).

According to European Alliance of Associations for Rheumatology (EULAR), EHOA differs from classic HOA presents distinct features and characteristics from the classical HOA, and namely its acute onset and peculiar radiographic pattern characterised by central erosions in the subchondral bone, accompanied by cortical bone destruction and subsequent remodelling, which can lead to ankylosis of the affected joints (Punzi et al., 2004; Zhang et al., 2007; Marshall et al., 2018). Furthermore, EHOA is also characterised by clinical and radiographic signs of inflammation. Several studies have demonstrated inflammation of the synovial membrane in the interphalangeal joints using ultrasound and MRI (Haugen et al., 2016; Iagnocco et al., 2005). Additionally, biomarker analyses have detected elevated blood levels of pro-inflammatory cytokines (Beekhuizen et al., 2013; Punzi et al., 2012) and high-sensitivity C-reactive protein (hs-CRP) levels (Smith et al., 2012; Punzi et al., 2005). Notably, synovial inflammation in EHOA correlates closely with patients’ symptoms and the development of new bone erosions (Kortekaas et al., 2016; Mancarella et al., 2015).

Several metabolic and cardiovascular diseases have been associated with HOA, including osteoporosis (OP) (Kasher et al., 2020) and obesity (Kluzek et al., 2015; Brogren et al., 2024). However, the evidence linking metabolic syndrome (MetS) to OA remains inconsistent (Strand et al., 2018). While OA in general does not appear to be strongly associated with MetS, certain phenotypes, such as nodal OA and EHOA may have shown a stronger connection, suggesting a potentially distinct pathophysiological mechanism of this latter form (Brogren et al., 2024). Indeed, low-grade chronic inflammation, combined with altered biomechanical loading, appears to play a critical role in the pathogenesis of EHOA influencing both clinical presentation and disease progression.

Our study endeavoured to ascertain the specific impact of cardiovascular and metabolic bone comorbidities on the clinical course and radiographic findings—osteophytes, joint space narrowing, malalignment, erosions, sclerosis, and subchondral cysts—in a cohort of EHOA patients. Additionally, we aimed to investigate the relationship between cardiovascular comorbidities and the presence of osteoporosis using clinical and serological indices of joint disease activity, including pain Visual Analogue Scale (VAS), disease activity questionnaires, namely the Australian/Canadian Osteoarthritis Hand Index (AUSCAN) (Bellamy et al., 2002) and Dreiser Functional Index (DREISER) (Dreiser et al., 1995), tender (TJ) and swollen joints (SJ) count, along with inflammatory biomarkers.

In this retrospective cross-sectional study were included 87 patients (female 88.5%, mean age 63.17 ± 8.85 years) with a confirmed diagnosis of EHOA which were followed at the Hand Osteoarthritis Clinic of Padova University Hospital, Italy, between January 2015 and February 2025. For the present study, all clinical, laboratory, and radiographic variables were extracted from the index visit used as the cross-sectional timepoint. Inclusion criterion required the presence of at least one central erosion in a PIP or DIP joint on hand plain radiographs. Patients with erosions involving metacarpophalangeal joints were excluded from the study. All participants provided written informed consent. The study protocol was approved by the local Ethics Committee (approval number 446/AO/23, dated 12 February 2024) and the study was conducted in accordance with the principles of the Declaration of Helsinki.

Each patient underwent a comprehensive clinical evaluation, including a detailed joint examination conducted by an experienced rheumatologist to assess SJ and TJ counts. Patients completed the following validated tools/questionnaires for the assessment of hand osteoarthritis at a single timepoint visit:

VAS for hand pain, scored on a scale from 0 (no pain) to 10 (worst possible pain);

The following demographic, clinical, and laboratory variables were collected: age, sex, age at disease onset, comorbidities, ongoing pharmacological treatments, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) levels. We measured the following metabolic parameters: body mass index (BMI), blood pressure, and serum levels of uric acid, glucose, total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), and triglycerides. We also assessed the following variables of metabolic bone diseases: history of fragility fractures, family history of fractures, age of menopause, smoking status, bone mineral density (BMD) values, and serum and 24-h urinary levels of calcium and phosphate, parathyroid hormone (PTH), and 25-hydroxyvitamin D levels.

Eighty-seven EHOA patients [females 77 (88.5%), males 10 (11.5%)], mean age of 63.17 ± 8.85 years, were included. The clinical and demographic characteristics along with biological exams, including CRP, ESR, VAS scale AUSCAN/DREISER questionnaires scores are summarised in Table 1. Regarding the radiological assessment, bilateral hand X-rays were collected from 86 patients. For each patient, 22 joints were examined, with an overall 1.892 joints evaluated. Radiographic features of the study population are detailed in Supplementary Table S1. When assessing osteophytes in the DIP, PIP, and TMC joints, 23.54% had no osteophytes, while the remaining joints showed varying degrees, with a grade 3 achieved 25.61% overall, with DIP joints being more involved by the presence of grade 3 osteophytes (30.61%). The pattern of joint space narrowing was similar to that of osteophytes with only 10.08% of DIP, PIP, and TMC joints being unaffected, with grade 3 achieved in 26.23% of the joints overall. Grade 3 narrowing was most common in DIP joints (38.52%). Malalignment was present in 44.50% of DIP, PIP, and TMC joints, with the highest incidence in TMC joints (49.42%). Erosions were found in 40.49% of joints, most commonly in DIP joints (55.10%). Sclerosis was present in 45.85% of joints, more frequently in TMC joints (58.33%). Subchondral cysts were detected in 36.54% of joints, especially in DIP joints (46.06%). When analysing median total radiographic scores per patient, including osteophytes: 29.29 ± 11.90 (range: 0–58), joint space narrowing: 32.81 ± 9.38 (range: 0–58), malalignment: 8.00 ± 3.34 (range: 0–18), erosions: 7.28 ± 3.93 (range: 0–18), sclerosis: 8.23 ± 3.43 (range: 0–18), cysts: 6.57 ± 2.90 (range: 0–18). Cardiovascular and metabolic bone comorbidities exhibited by the patients are shown in Table 2. Overall, 75.86% of participants had at least one cardiovascular risk factor including hypercholesterolemia, hypertriglyceridemia, obesity, and arterial hypertension.

Osteoporosis (OP) (defined as a T-score ≤ −2.5 at the lumbar spine, femur, or femoral neck) was present in 18 patients (21% of the cohort), while osteopenia (T-score between −1 and −2.5) was observed in 33 patients (38%). Overall, 37 patients (42.5%) received supplementation with at least Vitamin D and/or calcium, and 13 patients (14.94%) were actively treated with amino bisphosphonates—specifically, 5 with alendronate, 1 with risedronate, 1 with neridronate, and 5 with zoledronic acid—or with teriparatide (1 patient). Twenty-two patients (25.29%) reported a family history of femoral or vertebral fractures, and 12 patients (13.79%) had a personal history of femoral or spinal fractures. Additionally, 37 patients (55%) were undergoing treatment with chondroprotective agents, including glucosamine sulfate, chondroitin sulfate, and/or hyaluronic acid, while 17 patients (19.5%) were on cyclical nonsteroidal anti-inflammatory drug (NSAID) therapy.

The CCI, calculated without age adjustment to estimate the burden of coexisting comorbidities, was generally low, with a maximum score of 5 out of 39 and an average score of 1.24 (±1.35).

This is one of the largest monocentric retrospective cross-sectional cohort studies investigating the impact of cardiovascular and metabolic bone diseases on radiographic damage and disease activity indices in EHOA patients. Our findings suggest that selected comorbidities are associated with a higher symptom burden and greater radiographic severity in patients with EHOA. This aligns with a recently published review (Li et al., 2024) that examined various comorbidities, including those related to cardiovascular and metabolic systems, highlighting growing evidence that these conditions can have a substantial impact on osteoarthritis disease course.

Although the overall comorbidity burden—as measured by the CCI—did not appear to significantly affect structural joint damage or clinical disease activity in our cohort, some specific comorbidities were selectively associated with radiographic features of EHOA, as well as with clinical outcomes such as disability and pain.

Notably, both the AUSCAN score—which assesses pain, stiffness, and physical function—and the DREISER score—which assesses functional impairment—were significantly associated with tender (TJ) and swollen (SJ) joints count along with pain severity as measured by the VAS scale, which itself correlated with SJ/TJ count. These findings underscore the close interplay between pain and functional disability, highlighting their combined impact on patients’ quality of life.

Furthermore, AUSCAN and DREISER scores were associated with specific radiographic features, such as osteophytosis and subchondral cysts, suggesting that distinct bone alterations contribute to functional impairment. These associations reinforce the idea that skeletal changes in EHOA are not only structural markers of disease progression but also directly contribute to the symptomatic burden experienced by patients.

Regarding metabolic biomarkers, and in line with the findings by Brogren et al. (2024) who reported an association between elevated LDL cholesterol and greater OA severity, our study found that total serum cholesterol showed a trend toward correlation with the number of TJ, although it did not reach statistical significance. Nevertheless, this suggests a potential link between dyslipidemia-related signalling pathway alterations and increased joint pain sensitivity.

Only nine patients in our cohort were classified as obese (BMI >30), but 23% had a BMI >25. Obesity and metabolic syndrome are known to be strongly associated with OA, not only through mechanical stress that accelerates cartilage wear, osteophyte formation, and subchondral bone changes, but also through systemic inflammation. Adipose tissue indeed secretes adipokines (e.g., leptin, resistin) and cytokines (e.g., TNF-α, IL-6) which can perpetuate local inflammation and cartilage degradation (Batushansky et al., 2022). Several adipokine signaling pathways implicated in OA have been elucidated, encompassing both well-established pathways such as the AMP-Activated Protein Kinase/Mammalian Target of Rapamycin (AMPK/mTOR) pathway, the Nuclear Factor-κB (NF-κB) signaling pathway, and the Janus Kinase (JAK)/Signal Transducer and Activator of Transcription (STAT) pathway, as well as emerging pathways, including C/EBP-β and ATF4/RANKL (Zhang et al., 2022). These findings pave the way for novel therapeutic strategies that extend beyond the traditional focus on mitigating cartilage degradation. Additionally, the synovial adipokines may originate either from the secretion of the infrapatellar fat pad or from the bloodstream that permeates the synovial membrane and enters the joint cavity. Its strong association with OA has been extensively documented in the literature (Collins et al., 2020). Although hand joints are not weight-bearing joints, our study found a significant correlation between higher BMI and joint space narrowing in the hands. This cannot be explained by mechanical overload alone and instead suggests a metabolic-inflammatory mechanism linking BMI to structural joint damage. These findings align with the concept of EHOA as a systemic, metabolically influenced condition, reinforcing the role of obesity as a modifier of joint health, even in non-weight-bearing joints.

Moreover, diabetes is known to be associated with OA through hyperglycemia-mediated activation of molecular pathways involved in cartilage degradation and inflammation (Aiello et al., 2016). On the other hand, OA itself can contribute to systemic inflammation, with increased levels of cytokines such as TNF-α and IL-1, which in turn can worsen insulin resistance and facilitate diabetes onset. Recent findings by Marshall et al. (2019) suggest that diabetes may be a risk factor for OA progression, and previous studies have reported a higher pain perception in patients with both EHOA and diabetes (Magnusson et al., 2015).

In our cohort, the proportion of diabetic patients was small, with generally well-controlled disease, therefore, we are unable to draw strong conclusions regarding a potential association between diabetes and EHOA.

Overall, despite the long disease duration, our cohort consisted of patients who had access to high-quality care and were selected from a specialized tertiary care center. Therefore, these patients typically presented with milder comorbidity profiles, including the metabolic ones, which were effectively managed. This selection process means that the cohort likely represents a group of ‘healthier’ patients in terms of comorbidity burden, potentially leading to an overestimation of favorable outcomes. Secondly, while CCI is widely used, it captures only a limited range of comorbidities and does not account for their severity or duration. As a result, its utility for a comprehensive assessment of comorbidities in EHOA is constrained.

In terms of bone health, our findings corroborated previous studies highlighting an increased risk of OP in EHOA patients (Zoli et al., 2006; Tan et al., 2023; Mathieu et al., 2024) by emphasizing the role of bone fragility in worsening radiographic damage. Our study also revealed a novel association between FRAX-derived fracture risk and radiographic severity in EHOA. While T-scores at the femoral neck and lumbar spine correlated mainly with pain and functional indices (AUSCAN, DREISER), FRAX showed stronger and broader correlations with structural alterations such as osteophytes, joint space narrowing, erosions, and sclerosis. This discrepancy may be explained by the fact that FRAX incorporates not only bone density but also key clinical risk factors such as age, BMI, prior fractures, family history, smoking status, glucocorticoid exposure, and comorbidities (e.g., diabetes, rheumatoid arthritis). Additionally, patient selection bias may have influenced the final score. Therefore, FRAX may better reflect the cumulative systemic burden that contributes to bone fragility and joint degeneration.

Despite ongoing debate about the relationship between osteoporosis and OA, several studies, including one on over 200 EHOA patients (Mahajan et al., 2024) suggest a link between EHOA and increased fracture risk. Additionally, the concept of an ‘osteoporotic’ endotype of OA, characterized by bone fragility, has recently been proposed. This concept underscores that patients with erosive hand EHOA exhibit thinner bone mass and experience greater bone loss throughout the progression of the disease. Consequently, EHOA may be better conceptualized as a disorder of bone fragility, accompanied by an associated inflammatory burden, rather than simply a disease driven by cartilage degradation (McAlindon et al., 2021). Moreover, recent data from the QUALYOR cohort further supports that EHOA is a significant predictor of osteoporotic fractures in postmenopausal women, independent of traditional fracture risk factors (Auroux et al., 2025). From these findings we could speculate that the FRAX score could serve as an important prognostic tool for assessing EHOA damage and its progression.

These results support the hypothesis that EHOA is not merely a localised joint disease but part of a broader systemic metabolic bone phenotype, in which fracture risk and radiographic progression share common pathogenic pathways. These results should be interpreted as associative and hypothesis-generating. Future longitudinal studies are needed to clarify whether FRAX can predict radiographic progression over time, and whether targeting metabolic bone comorbidities can modify disease course.

The relationship between OA and OP remains a matter of debate, with some studies suggesting an inverse association (Foss and Byers, 1972) and others supporting a more direct link. Indeed, Güler-Yüksel et al. (2011) reported that EHOA progression over a 2-year period was associated with accelerated metacarpal BMD loss. Although we assessed BMD only in standard anatomical sites (lumbar spine, femur), our findings raise the question of whether targeted Dual-Energy X-ray Absorptiometry (DEXA) measurements at the hands might better capture local bone loss and its relationship with EHOA. In fact, Rovetta et al. (2003) found significantly higher serum levels of CTX-I, marker of bone resorption, in EHOA patients compared to those with non-erosive OA, suggesting that bone turnover may play a key role in EHOA pathogenesis. While CTX-I was not measured in our study, its inclusion may have provided invaluable insights into the interplay between cartilage degeneration and bone fragility. Further large-scale studies are needed to more precisely quantify the relationship between OP and erosive OA. Moreover, mechanistic research is essential to determine whether osteoporosis directly contributes to OA progression or merely shares common risk factors. Some studies have proposed that bisphosphonates, such as denosumab or clodronate, may have therapeutic potential in OA by modulating subchondral bone remodeling, reducing inflammation, and alleviating pain. However, robust clinical trials and larger studies are needed to confirm their efficacy and define clear indications in OA treatment.

Overall, the AUSCAN and DREISER scores effectively captured the burden of pain and functional disability in our cohort. These scores correlated directly with swollen and tender joints count, as well as perceived pain on the VAS scale, which in turn was associated with the occurrence of osteophytosis. These relationships suggest that analgesic therapy remains essential, alongside chondroprotective agents and joint support therapies, for controlling pain, particularly in patients with a longer disease duration—which also correlated with more severe radiographic changes in our patients.

We would remise not to mention some of the limitations of our study. Firstly, selection bias may affect the generalisability of our findings, as the cohort consisted of a population with access to a high level of care and milder comorbidity profiles. Secondly, although widely used, the CCI only captures a limited set of comorbidities and does not account for their severity or duration, thus limiting its utility for comprehensive comorbidity assessment in OA. Finally, although the relatively long disease duration in our cohort (mean 13.3 ± 8.8 years) allowed for the assessment of long-term structural damage, it also introduced confounders linked to concomitant comorbidities and varied treatment histories over time, thus complicating the attribution of cause-effect relationships. Given the exploratory retrospective design, no a priori sample size calculation was performed. Notably, the absence of longitudinal radiographic data limits our ability to assess the long-term progression of joint damage and bone loss in EHOA. Additionally, potential residual confounding due to treatment history could influence the results, as variations in pharmacological and non-pharmacological interventions may have impacted the disease course in some patients. Moreover, given the number of exploratory correlations performed and the absence of formal correction for multiple testing, the possibility of chance findings cannot be excluded, particularly for associations outside the main bone-related analyses.

Nevertheless, our findings suggest potential associations between metabolic and osteoporotic comorbidities and both symptom burden and radiographic severity in EHOA. These results support a multidisciplinary, personalised approach to EHOA management, emphasizing the need for early identification and management of relevant comorbidities. Further prospective, longitudinal studies are warranted to clarify these associations and guide more effective, patient-tailored therapeutic strategies.