Our work describes the clinical characteristics of individuals of a symptomatic low-income HOA (LIHOA) non-caucasian cohort. In addition to reporting data from underrepresented populations, we also aimed to determine whether social disparities in patients living in the same region have a clinical impact in HOA. Self-declared monthly family income, occupation, and level of literacy were used as surrogates for classifying patients as pertaining to more or less favoured social environments. In an attempt to collect data from different social strata in our region, we invited rheumatologists working in both public and private practice to indicate patients to be included in our sample. It turns out that this strategy led us to include over 80% of individuals declaring ≥ 300.00 US$ monthly family earnings, including a slight majority earning > 900.00 US$. Although these higher values are far below earnings declared in affluent populations [28], individuals of the LIHOA cohort do not represent the majority of the population living in our region. Indeed, using official data, our mean 2022 Growth domestic product (GDP) per capta was R$ 1023.00 (Brazilian currency), which is roughly <200US$ [28]. However, our strategy allowed us to compare the clinical picture among individuals of different socioeconomic status living in the same region. This can also be illustrated by the fact that most patients declared more than 9SY of formal education with a third having a university degree. By comparison, official data show that only 19.2% of the population living in our region have a university degree [29]. We also have recently published data on juvenile idiopathic arthritis and orthopaedic surgeries, in which patients were recruited consecutively, showing a high prevalence of low-income patients thus reflecting our official data [30, 31]. We ended up finding that individuals classified as low-income had more impact in hand function, showcasing lower pinch and grip strength values and higher FIHOA scores, whereas pain levels did not differ among both groups. Interestingly, neither having a blue or white collar occupation nor the level of literacy influenced symptom severity and/or function in this population. We believe these are the first data showing that income, but not occupation, in patients sharing the same environment, has a significant impact in the function of patients with HOA.

Similar to other cohorts [3, 11, 17], the vast majority of our patients were female, middle-aged, with a long-term HOA disease and moderate to severe pain at movement, with mild to virtually no pain at rest. Comorbidities also mirrored data from other cohorts [17] with hypertension, metabolic syndrome, and dyslipidemia being the most common, followed by osteoporosis, thyroid disease, mood disorders, and diabetes. Interestingly, most patients had concomitant OA in joints other than the hands, most commonly in the spine, followed by the knee. Indeed, in the DIGICOD study, reporting a cohort of HOA patients in France, 25.8% of the patients fulfilled criteria for knee OA, which is similar to the 29.4% frequency of knee OA in our study [17].

It is also worth noting that 20.5% of the patients had bunions that were attributed to OA, which was also as common as the 18.6% of patients that had rizarthrosis. We are not aware of previous studies in HOA patients that assessed the prevalence of concomitant OA in other joints. However, in the DIGICOD cohort, presence of spinal (cervical and/or lumbar) and knee pain, which could be linked to OA, were common musculoskeletal complaints [17].

Functional impairment can be considered mild in our study, given the median values of 8 and 19 in the FIHOA and Cochin questionnaires, respectively. This is very similar to the baseline 10 score in the FIHOA questionnaire reported in the EHOA trial evaluating etanercept as a treatment for hand osteoarthritis [32]. Interestingly, patients in the recently developed DIGICOD cohort had higher FIHOA scores, meaning greater functional impairment, as compared to our results. Given that we did not specifically determined the percentage of patients with erosive HOA in our sample, we can only speculate whether the inclusion of 45.8% individuals with erosive HOA led to higher FIHOA scores in the DIGICOD cohort as compared to our data [17]. Values obtained with the SF-12 questionnaire also mirrored data reported in the HERO study, showcasing a substantial deficit in both mental and physical domains, as compared to healthy individuals [33]. Similar results were reported in the EHOA clinical trial, which included patients with erosive osteoarthritis, showing baseline Physical Component SF-36 values of 42.9, which are lower than the 50 score threshold considered as normal [34, 35]. These data are intriguing, as only 5.8% of our patients declared having a diagnosis of depression and only 11.7% reported having a generalized anxiety disorder. In conjunction, these results suggest a significant mental burden in patients with HOA. Whether this can be specifically attributed to the HOA itself should be further explored.

The low-income subgroup of our cohort presented worse hand function and lower performance in pinch and grip strengths. Those with lower income were also more frequently classified as having blue-collar jobs, with 59% vs. only 21% in the higher-income group. However, there was no statistically significant difference regarding pain, function, imaging or patient reported outcome parameters regardless of patients being classified of having a blue or white collar occupation. We should also consider that this classification does not necessarily reflect other daily-life activities that may impact pain or function in HOA. Moreover, the low number of individuals in each subgroup might have also impacted the analysis. A large majority of patients were not using medications to treat their HOA. This might illustrate a major unmet need regarding pharmacological treatments for pain relief in this disease.

Our study has several limitations, which includes being a single-center study and the cross-sectional analysis. The fact that it was initiated shortly prior to the COVID-19 pandemic severely affected the inclusion of patients. We also did not focus the erosion pattern, as we intended to recruit symptomatic HOA patients. Also, our strategy to enhance the number of individuals of higher income, aiming to find differences regarding social disparities in the same sample, has biased the recruitment by reducing the number of low-income individuals included, which does not reflect the majority of our population. Another limitation is the fact that we did not adjust the statistical evaluation for potential confounders.