Purpose: Erosive hand osteoarthritis (EHOA) is often considered a more severe form of hand OA that includes pain and inflammation. However, more data are needed regarding the different hand OA phenotypes, especially erosive and non-erosive, to conclude if they represent two different subsets of hand OA or even different degrees of affectation of the same disease. The purpose of this work is to define a specific clinical pattern of patients with EHOA and identify putative biochemical markers associated with this phenotype. Methods: This cross-sectional study was conducted in the Prospective Cohort of Osteoarthritis A Coruna (PROCOAC). This cohort consists of 1136 subjects, from which 834 were diagnosed of hand OA following ACR criteria. After reviewing the x-rays, the cohort was split into patients with and without erosive hand OA. Then, both clinical and demographic data within each group were subsequently analyzed. A Univariate analysis comparing different variables between both groups, followed by a stepwise regression logistic regression analysis was performed to characterize the clinical pattern of EHOA. The identification of biochemical markers was carried out following a two-step workflow: in a first discovery phase, a proteomic approach based on peptide labelling with Isobaric tags for relative and absolute quantitation (iTRAQ) was employed using two different sets of sera (n=20) from EHOA or non-EHOA patients. In a second validation phase an independent set of 346 individual serum samples from the PROCOAC cohort (EHOA, n=127; non-EHOA, n= 219) was used to quantify serum concentrations of the extracellular matrix protein 1 (ECM1) by ELISA. The Mann-Whitney test was carried out in order to look for significant differences between EHOA and non-EHOA groups. All the statistical analyses were performed using SPSS software v.24. Results: The univariate analysis showed that EHOA patients were younger (p<0.001), smokers (p=0.011), with lower both body mass index (p=0.006) and arterial hypertension (p=0.002); increased frequency of inflammatory symptoms in hands (p<0.001), more presence of nodular hand OA (p=0.001) but lower frequency of other forms of hand OA, specially thumb-base (p=0.005) and metacarpophalangeal (p=0.036). In addition, these patients also showed a lower number of damaged joints (p<0.001), specially the knee (p<0.001). The stepwise logistic regression model (Table 1) confirmed the strong association of age (OR=0.963;95%CI=0.937-0.989;p=0.005), inflammatory symptoms (OR=4.575;95%CI=2.621-7.985;p<0.001), increased prevalence of nodal OA (OR=2.125;95%CI=1.118-4.037;p=0.021) and, in patients over 64 years old, MetS (OR=2.029;95%CI=1.183-3.478;p=0.010) with the erosive phenotype. In addition, erosive patients show a lower prevalence of knee OA (OR=0.416;95%CI=0.252-0.687;p=0.001) and to carry this phenotype at baseline does not confer a significant increased risk for radiographic knee OA progression over time). Regarding the proteomic analysis for the identification of potential biomarkers associated with the EHOA phenotype, a total of 257 different proteins were identified in the serum samples with more than two peptides and a total score ≥2 at 95% confidence. Thirty-six proteins were found quantitatively altered between those patients with or without EHOA. A significant reduction (p=0.029) of ECM1 concentrations was found in the serum from patients with the EHOA phenotype (0.154 ± 0.010 μg/ml vs 0.169 ± 0.007 μg/ml) (Figure 1). Conclusions: A specific clinical pattern of patients with EHOA has been defined. It has been associated with altered levels of 36 proteins in serum. Particularly, decreased amounts of ECM1 have been found in patients with this erosive phenotype.View Large Image Figure ViewerDownload Hi-res image Download (PPT)