Metabolic syndrome is not associated with erosive hand osteoarthritis: a cross-sectional study using data from the PROCOAC cohort

To delineate the phenotype of erosive hand osteoarthritis (EHOA) in a Spanish population and assess its correlation with metabolic syndrome. We conducted a cross-sectional study using baseline data from the Prospective Cohort of Osteoarthritis from A Coruña (PROCOAC). Demographic and clinical variables, obtained through questionnaires, clinical examinations, and patient analytics, were compared among individuals with hand OA, with and without EHOA. We performed appropriate univariate and multivariate stepwise regression analyses using SPSS v28. Among 1039 subjects diagnosed with hand OA, 303 exhibited EHOA. Multivariate logistic regression analysis revealed associations with inflamed joints, nodular hand OA, and total AUSCAN. Furthermore, the association with a lower prevalence of knee OA remained significant. The influence of metabolic syndrome (MetS) on EHOA patients was analyzed by including MetS as a covariate in the model. It was observed that MetS does not significantly impact the presence of EHOA, maintaining the effect size of other factors. In conclusion, in the PROCOAC cohort, EHOA is associated with nodular hand OA, inflammatory hand OA, and a higher total AUSCAN. However, EHOA is linked to a lower prevalence of knee OA. Importantly, in our cohort, no relationship was found between EHOA and MetS.


Measurements
All X-rays from patients with hand OA underwent rigorous review to confirm the erosive phenotype 31 , conducted by both a radiologist and a trained rheumatologist specialist in a blinded manner.Erosive hand OA (EHOA) was radiographically defined by subchondral erosion, cortical destruction, and subsequent reparative changes in interphalangeal (IP) joints, potentially involving bony ankylosis in at least one hand.Additionally, the total number of IP joints with erosions per patient was counted and recorded.The reliability of all X-rays was meticulously assessed.Evaluation of right and left-hand joints, including wrists, was conducted using anteroposterior images.Magnified views of the entire hand were employed to confirm the presence of typical erosions 4 .
For this study, baseline data included the following variables: age, gender, smoking habit (categorized as never smoker, current smoker, or former smoker), body mass index (BMI), abdominal perimeter, personal history of hypertension, type 2 diabetes, dyslipidemia, psoriasis, osteoporosis, and osteopenia.Definitions for conditions included systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg for hypertension 32 , fasting plasma glucose ≥ 126 mg/dL, glycated hemoglobin ≥ 6.5%, or classic symptoms of hyperglycemia for diabetes 33 , total cholesterol ≥ 220 mg/dL, low-density lipoprotein (LDL) ≥ 150, high-density lipoprotein (HDL) < 40 mg/dL in men and < 50 mg/dL in women, and triglycerides > 150 mg/dL for dyslipidemia.Osteoporosis was defined as a bone mineral density (BMD) ≤ − 2.5 SD in the femoral neck or lumbar spine, and osteopenia as BMD between − 1.5 and − 2.5 SD 34 .Metabolic Syndrome status was assessed following ALAD (Asociación Latinoamericana de Diabetes) 2010 criteria 35 , considering criteria such as waist circumference ≥ 94 cm in men and ≥ 88 cm in women, and meeting at least 2 of 4 specified criteria.Nodular hand OA was defined as Heberden and/or Bouchard nodes plus underlying interphalangeal damage.Thumb base OA, inflamed IP joints (characterized by swelling, heat, pain, and redness), and fasting blood levels of glucose, LDL, HDL, triglycerides, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), rheumatoid factor (RF), and anticitrulline antibodies (ACPA) were also collected at baseline.Additionally, the AUStralian CANadian index (AUSCAN Osteoarthritis Hand index NRS 3.1) questionnaire was assessed in a subset of 575 hand OA patients, providing a total score that considers pain, stiffness, and function, with each section evaluated on a scale from 0 to 100 36,37 .

Statistical analysis
The statistical analysis encompassed a descriptive examination of the studied variables.Continuous variables were presented as mean ± standard deviation (SD) and median, while qualitative variables were articulated in absolute values (n) and percentages.
Mean comparisons between two groups were conducted utilizing the Student's t-test or Mann-Whitney test, depending on appropriateness, subsequent to normality verification through the Kolmogorov-Smirnov test.Associations of qualitative variables were scrutinized using the Pearson Chi-Square test.
Univariate and multivariate logistic regression models were employed to investigate factors linked to the presence of erosive osteoarthritis (OA) of the hand.Construction of the multivariable models adhered to the backward selection method, where the Wald statistic served as the elimination criterion in each step, retaining variables with a significance level of p < 0.05 or clinical relevance.The impact of MetS on the probability of presenting EHOA was assessed by adjusting for variables selected in the final model.The effect of each factor was expressed through estimation of the odds ratio (OR) and its 95% confidence interval (CI).The area under the receiver operating characteristic (ROC) curve (AUC) quantified the discrimination performance of the EHOA predictive models.Sensitivity, specificity, and predictive values were reported with their 95% confidence interval.
The relationship between age and the risk of EHOA was explored using multivariable-adjusted cubic splines curves, incorporating five knots in the model.
All analyses were conducted using statistical software, specifically SPSS v.28 and R v.4.2.The "pROC" package facilitated ROC analysis, while restricted cubic splines regression was implemented with the "rms" package.Significance levels reported were two-sided, and the threshold for statistical significance was defined as p < 0.05.

Results
A total of 1039 patients had hand OA with a mean age of 62.95 years and a predominance of women (82.58%).Table 1 illustrates the characteristics of our study population.Within this cohort, 303 patients (29.16%) had EHOA.
Regarding MetS, only hypertension and abdominal circumference showed a statistically significant relationship in univariate analysis, but these results were not sustained in multivariate analysis (Table 3).www.nature.com/scientificreports/ In analyzing the influence of MetS on EHOA patients by including MetS as a covariate in the proposed model, it was observed that MetS did not significantly influence the presence of EHOA, maintaining the size of the effect of other factors (Table 4).Although not an independent factor for predicting EHOA in the analyzed sample, a non-statistically significant trend of increased EHOA risk was observed among patients with MetS (OR = 1.54; 95%CI 0.89-2.68)(Table 4 and Fig. 1).
Moreover, multivariable-adjusted restricted cubic splines demonstrated a non-significant relationship between age and EHOA (p for overall trend = 0.335) (Supplementary Fig. 1).Upon thorough data review and considering a threshold at age 63 to explore a possible change in the association between age and EHOA risk and its interaction with MetS, a reversed trend was observed among those over 63, indicating a higher prevalence of MetS among EHOA (31.91 vs. 26.04%),although these differences were not significant in either group (Supplementary Table 1 and Fig. 2).
Analyzing the association between age and MetS independently revealed a significant association, with a higher prevalence of metabolic syndrome in those over 63 years of age (Fig. 2).Among patients ≥ 63 years, there were more with non-EHOA than with EHOA, while EHOA predominated among the youngest.Analyzing age  and MetS within each erosive group (non-EHOA vs. EHOA) showed a significant association between age and MetS in both groups, with a greater presence of MetS among those ≥ 63 years of age (Fig. 2).The number of erosions, total erosions on both hands, and the average number of erosions between the hands were recorded.The mean number of total erosions between both hands was around 3, both in patients with MetS and without MetS (p = 0.327) (Table 5).

Discussion
One objective of this study was to describe the Erosive Hand Osteoarthritis (EHOA) phenotype in the Spanish PROCOAC cohort.Characterizing this phenotype revealed that, compared to non-EHOA, patients with EHOA are younger, exhibit a lower presence of knee Osteoarthritis (OA), a higher concurrent presence of clinically inflamed joints, and an increased prevalence of nodular hand OA.The heightened levels of pain, disability, and inflammation observed in EHOA patients, in addition to the increased presence of nodes and erosions, contribute to a greater clinical burden when compared to patients with non-erosive hand OA, aligning with prior findings 7,22,38,39 .
In our cohort, EHOA showed no association with Rheumatoid Factor (RF).These results are consistent with previous publications [40][41][42] , and similarly, there is no observed relationship between EHOA and Anti-Cyclic Citrullinated Peptide antibodies (ACPA), a novel finding not reported before in our knowledge.However, there are existing data about the connection between ACPA and inflammatory hand OA 43 .
The second aim was to elucidate the relationship between EHOA and Metabolic Syndrome (MetS).Univariate and multivariate analyses were conducted without revealing a significant association.Additionally, a predictive model for EHOA incorporating MetS showed no modification in the results.This is in concordance with findings from the Strand publication based on the Framingham cohort, which found no significant association between hand OA, including EHOA, and MetS 44 .Notably, there was a borderline association with hypertension, aligning with our cohort data.
The association of MetS with hand OA and EHOA has yielded conflicting results in the literature.While some studies suggest a potential link, others present contradictory evidence 25,26,27 .Our study adds to this complexity, emphasizing the need for further investigation.Notably, our analysis of EHOA patients did not demonstrate a significant association with MetS, consistent with findings from the Framingham cohort 44 .
The association of hand OA or EHOA with more generalized OA has been debated 5,21,45 .Our study indicates a lower prevalence of OA in other joints, specifically the knees, among patients with EHOA in the PROCOAC cohort.This could be attributed to the relatively lower age of EHOA patients and their lower BMI compared to non-EHOA patients.However, this finding contrasts with some studies showing a more common occurrence of hand OA in the context of generalized OA 20 .A recent publication shows the inverse associationof two loci between EHOA and knee OA 46 .In this paper they found 4 significant EHOA loci with high effect on EHOA risk.Two of them, rs1800801 and rs4496445 was associated with EHOA risk and with protection of knee OA.Apart from the manuscript of Haugen 18 and Marshall 13 , to our knowledge, none of the abrove-mentioned studies focused their analysses on the erosive subtype specifically, but rahter on hand OA in general.The study of Marshall et al. consisted of the description of subsets of hand OA in a prospective cohort of 1076 older adult patients, of which only 52 (4.8%) developed the erosive phenotype and they concluded that radiographic knee OA was not increased in EHOA patients.
While the present study focuses on EHOA, the analysis of other studies is predominantly centered on hand OA in general.Future research with large, prospective cohorts is imperative to delve deeper into the association between EHOA and other joint involvements.
This study has potential limitations, including its cross-sectional approach to hand OA patients, limiting the establishment of causality.Unmeasured confounders, such as diet, alcohol consumption, or socioeconomic status, could influence the results.Nevertheless, the proportion of EHOA in our cohort, at 20.3%, exceeds the mean compared with other studies.Other studies patients.

Conclusion
Within the PROCOAC cohort, Erosive Hand Osteoarthritis (EHOA) demonstrates associations with nodular hand OA, inflammatory hand OA, and elevated total AUSCAN scores.Notably, EHOA exhibits a lower prevalence of knee Osteoarthritis (OA) in our cohort.Additionally, we observed no discernible relationship between EHOA and Metabolic Syndrome (MetS) in our study.

Figure 1 .
Figure 1.ROC curve of EHOA predictive model without and with MetS The continuous line represents the data from the EHOA predictive model that includes the variables from Tables3 and 4(age at diagnosis, sex, hypertension, body mass index, knee OA, nodular hand OA, inflammatory hand OA and hip OA).This model has a sensitivity of 0.75, an specificity of 0.64, a negative predictive value of 0.83 and a positive predictive value of 0.53.The dashed line represents the model by adding metabolic syndrome (MetS).This model with MetS has a sensitivity of 0.75, an specificity of 0.63, a negative predictive value of 0.83 and a positive predictive value of 0.51.

Figure 2 .
Figure 2. Correlation between age and metabolic syndrome (MetS) in patients with and without Erosive Hand OA (EHOA).

Table 2 .
Descriptive analysis and comparison of hand OA patients according to EHOA.BMI, body mass index (Kg/m 2 ); HDL, high-density lipoprotein; MetS, metabolic syndrome; Knee OA, OA at least in one knee; Hip OA, OA at least in one hip; EHOA, erosive hand osteoarthritis at least in one hand; Inflammatory hand OA, OA at least in one hand; Nodular hand OA, OA at least in one hand; Thumb base OA, OA at least in one hand; ESR, erythrocyte sedimentation rate (mmHg); CRP, C-reactive protein (mg/dL); RF, rheumatoid factor; ACPA, anti-citrullinated protein antibodies.Significant values are in[bold].

Table 3 .
Erosive hand OA predictive model.Significant values are in [bold].

Table 4 .
Erosive Hand OA predictive model including MetS.BMI, body mass index (Kg/m 2 ); Knee OA, OA at least in one knee; Nodular hand OA, OA at least in one hand; Inflammatory hand OA, OA at least in one hand; Hip OA, OA at least in one hip; MetS, metabolic syndrome.Significant values are in [bold].

Table 5 .
Relation between number of erosions in hand OA and MetS.