0759: Derivation and Validation of Four Patient Clusters in Still’s Disease, Results from GIRRCS AOSD-study Group and AIDA Network Still Disease Registry

Piero Ruscitti¹, Antonio Vitale², Ilenia Di Cola¹, Roberto giacomelli³ and Luca Cantarini⁴, ¹University of L'Aquila, L'Aquila, Italy, ²University of Siena, Siena, Italy, ³University of Rome Campus Biomedico, Rome, Italy, ⁴Department of Medical Sciences, Surgery and Neurosciences, Rheumatology Unit, University of Siena and Azienda Ospedaliero-Universitaria Senese [European Reference Network (ERN) for Rare Immunodeficiency, Autoinflammatory and Autoimmune Diseases (RITA) Center], Siena, Italy

Background/Purpose: Still's disease is a rare inflammatory disorder, these patients have a highly heterogeneous disease according to age of onset, clinical presentation, presence of life-threatening complications, and outcomes. Thus, we aimed at deriving and validating disease clusters in a multicentre, observational, prospective study to stratify these patients.

Methods: Patients included in GIRRCS AOSD-study group and AIDA Network Still Disease Registry were selected to be assessed in this analysis if clinical variables selected for cluster analysis were available [i.e., age, systemic score, erythrocyte sedimentation rate (ESR), C reactive protein (CRP), and ferritin]. A cluster analysis was performed to devise deeper levels of categorization. The k-means algorithm with Euclidean metric was performed, setting 100 random assignments to the cluster seeds. Z-scores were also provided to account for the different units of the selected variables. The Elbow plot was used to devise an adequate number of clusters, avoiding too large choice which could underwent to overfitting. Besides the cluster plot enabled the view of the separation among clusters using the first two principal components. After that a descriptive assessment of the clusters was performed.

Results: By combining clinical selected variables, the K-means clustering assessment provided 4 clusters based on means standardized according to z-scores on 349 patients. The "within" and "between" separation properties were also derived. Within cluster sum of squares (SS) by cluster was estimated: i. cluster 1: 251.69; ii. cluster 2: 283.22; iii. cluster 3: 159.98; iv. cluster 4: 323.76. SS_between/SS_total was derived to be 41.5%. The latter is not particularly high, but not affected by overfitting phenomena. In fact, the derived clusters showed reasonable "within" and "between" variability properties. After that, randomly sampling 50% of the original records, the same number of clusters was provided by the elbow plot with a similar SS_between/SS_total of 42.6%.

All clusters mainly presented fever, skin rash, and joint involvement but each one of these has some different clinical features from others. Cluster 1 ("Juvenile/Transitional") was composed by 115 patients distinguished by the lowest value of age and characterized by skin rash myalgia, sore throat, and splenomegaly. Cluster 2 ("Common") included 128 patients identified by the lowest levels of ESR, ferritin, and systemic score; multiorgan manifestations were less frequently observed. Cluster 3 ("Hyperferritinemic") comprised 31 patients categorised by the highest levels of CRP and ferritin, they were characterized by fever and joint involvement. Cluster 4 ("Catastrophic") contained 75 patients derived by the highest values of age and systemic score. Myalgia, sore throat, liver involvement, and life-threatening complications, leading to a high mortality rate, were observed in these patients.

Conclusion: Four patient clusters in Still's disease may be recognized by a multidimensional characterization. Each one of these has some different clinical features from others; cluster 4 was burdened by an increased rate of life-threatening complications and mortality, suggesting a more severe patient group.


P. Ruscitti: None; A. Vitale: None; I. Di Cola: None; R. giacomelli: None; L. Cantarini: None.