POS1247 HIERARCHICAL ANALYSIS FOR IDENTIFICATION OF CLINICAL PHENOTYPES OF CALCIUM PYROPHOSPHATE DEPOSITION DISEASE

Le résumé fourni par la source

Background: Calcium pyrophosphate deposition (CPPD) is a heterogeneous condition, ranging from asymptomatic crystal deposition to acute or chronic arthritis. It is frequently misdiagnosed, as it mimics other musculoskeletal disorders, leading to its historical labels such as "pseudogout" or "pseudo-rheumatoid arthritis". The diverse clinical presentations and variable disease progression pose significant challenges for both clinical assessment and therapeutic management. In 2011 the EULAR task force proposed 4 phenotypes based on expert opinion, 1) asymptomatic CPPD, 2) osteoarthritis (OA) with CPPD, 3) acute CPP crystal arthritis, 4) chronic CPP crystal inflammatory arthritis. However, these phenotypes show substantial overlap, for example a patient with OA and CPPD on imaging could be classified as either type 1 or type 2. Further, they were never validated in real-world patient cohorts. Objectives: The objective of this study is to identify, for the first time using rigorous methodology, clinical phenotypes of CPPD disease by integrating real-world data from two datasets, rather than relying on expert opinion alone. Methods: Data from the COLCHICORT and CHRONIC CPPD studies were used. The comparability of the populations from the 2 studies was assessed with Student's t-test or Wilcoxon-Mann-Whitney test according to normality from Shapiro-Wilk test for quantitative data, and Chi-squared or Fisher's exact tests according to the Cochran's rule for qualitative data. Multiple Correspondence Analysis (MCA) was performed to graphically evaluate the association between the modalities of the qualitative variables, the few quantitative variables having been discretized. Therefore, the principal components allow for visualisation of patients/variables (dimension reduction) and identification of significative associations among variables (modalities). Hierarchical Clustering on Principal Components (HCPC) explaining at least 90% of variability was applied to identify specific patterns of patients (homogeneous groups/clusters). The number of clusters was determined taking into account the differences between clusters (variance between clusters) via the dendrogram of inertia gain and cluster interpretability. A bivariate analysis was performed to characterize the clusters in order to complete the interpretation given by the projection of individuals in the MCA maps, and Chi-squared or Fisher's exact tests according to the Cochran's rule were used. Results: A total of 227 patients were included in the analysis, with 98 patients from the COLCHICORT study and 129 patients from the CHRONIC CPPD study. Due to missing data, 94 patients were excluded, and the final analysis was conducted on 133 patients with complete datasets. MCA was performed to identify the variables contributing most to overall variability, followed by hierarchical clustering on principal components derived from the MCA. This analysis revealed four distinct patient clusters (Figure 1). The distribution of patients across these clusters was as follows: Cluster 1 (45.9%), Cluster 2 (16.5%), Cluster 3 (18.0%), and Cluster 4 (19.5%). The characteristics of each cluster are summarized below: •Cluster 1 accounts for patients with predominantly monoarticular involvement, more frequent recurrent acute flares than persistent arthritis, older age at symptom onset (>75 years), female prevalence, with elevated CRP levels. •Cluster 2 comprises patients with prevalent monoarticular involvement, with prevalent persistent arthritis and recurrent acute flares, symptoms onset > 60 years, female predominance, and elevated CRP in approximately half of the patients. •Cluster 3 comprises patients with polyarticular involvement, more frequent persistent arthritis compared to recurrent acute flares, older at symptom onset (>75 years), female prevalence, with lower CRP levels, and involvement of MCP and wrist with structural damage, typical of CPPD. •Cluster 4 represents patients with polyarticular involvement, more frequent spine and large joints involvement, predominantly recurrent acute flares rather than persistent arthritis, younger age at symptom onset (<75 years), male predominance, with elevated CRP levels. Figure 1 Conclusion: This cluster analysis identified four distinct phenotypes of CPPD disease (Figure 2). Two phenotypes demonstrated monoarticular involvement: cluster 1, corresponding to the EULAR-defined "acute CPP crystal arthritis", and cluster 2, which does not fully align with any EULAR-defined phenotype. The other two phenotypes featured polyarticular involvement: cluster 3 consistent with the EULAR phenotype "chronic CPP crystal inflammatory arthritis", and cluster 4 resembling polymyalgia rheumatica, which is not described by the EULAR. In conclusion, by using for the first time a robust, data-driven methodology based on real-world data two clusters similar to the EULAR phenotypes "acute CPP crystal arthritis", and "chronic CPP crystal inflammatory arthritis" were confirmed, but in addition two more clusters that do not correspond to any existing EULAR phenotypes were also identified. Unlike the EULAR phenotypes, which include asymptomatic subjects, our study focused exclusively on symptomatic individuals, this explains the absence of the "asymptomatic deposition" in our findings. This study has some limitations, primarily the small sample size and its retrospective design. Future research should aim to validate these findings in larger prospective cohorts or registries to further refine CPPD phenotypes. The identification of different phenotypes in this extremely heterogeneous condition could c Figure 2 REFERENCES: NIL . Acknowledgements: NIL . Disclosure of Interests: None declared . © The Authors 2025. This abstract is an open access article published in Annals of Rheumatic Diseases under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Neither EULAR nor the publisher make any representation as to the accuracy of the content. The authors are solely responsible for the content in their abstract including accuracy of the facts, statements, results, conclusion, citing resources etc.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.