Article Text

OP0059 Phenotypes Determined by Cluster Analysis and their Survival in the Prospective Eustar Cohort of Patients with Systemic Sclerosis
  1. V. Sobanski1,
  2. J. Giovannelli2,
  3. G. Riemekasten3,
  4. P. Airò4,
  5. S. Vettori5,
  6. F. Cozzi6,
  7. O. Distler7,
  8. Y. Allanore8,
  9. C.P. Denton9,
  10. D. Launay1,
  11. E. Hachulla1
  12. on behalf of EUSTAR co-authors
  1. 1Médecine Interne, Université Lille Nord De France
  2. 2Department of Epidemiology and Public Health, Lille, Lille, France
  3. 3Department of Rheumatology, Charité University Hospital, Berlin, Germany
  4. 45. UO Reumatologia e Immunologia Clinica, Spedali Civili, Brescia
  5. 5Rheumatology Unit, cond University of Naples, Naples
  6. 6Rheumatology Unit, Department of Medicine (DIMED), University of Padova, Padova, Italy
  7. 7Department of Rheumatology, University Hospital Zurich, Zurich, Switzerland
  8. 8Rheumatology A dpt, Paris 5 University, Cochin Hospital, Paris, France
  9. 9Centre for Rheumatology, Royal Free and University College London Medical School, Royal Free Campus, London, United Kingdom


Background Systemic sclerosis (SSc) is classically dichotomized in limited or diffuse cutaneous subsets associated with different prognosis. However, it appears that SSc is a highly heterogeneous disease, with probably more subtle clinical phenotypes.

Objectives The primary objective of our study was to identify and characterize homogeneous phenotypes in the EUSTAR SSc population using a cluster analysis. The secondary objective is to assess the survival in the different clusters.

Methods The participants were SSc patients of the EUSTAR prospective cohort. Inclusion criteria were: age ≥18 years, fulfilled American College of Rheumatology criteria for SSc, and had a calculable follow up. Hierarchical clustering based on the Ward method was performed using 25 clinically meaningful variables. The clinical relevance of the clusters was analysed by their characteristics and survival outcomes. Survival curves were plotted by the Kaplan-Meier method. Crud and adjusted (on the age at diagnosis and gender) Cox models were performed.

Results A total of 6.927 patients were analyzed. Cluster analysis identified 5 clusters (Table 1). The first cluster C1 (n=1,963) was mainly characterized by a low Rodnan skin score contrasting with a relatively similar percentage of anti topoisomerase 1 and anti-centromere Ab (28% and 40%) and a high percentage of lung fibrosis (48%); the second cluster C2 (n=1,186) by a large majority of limited SSc (89%) and anti-centromere (79%) and the lowest percentage of lung fibrosis (22%); the third cluster C3 (n=1,249) and the fourth cluster C4 (n=856) by a high percentage of diffuse SSc/anti topoisomerase 1 (72%/50% and 92%/77% respectively) and the highest proportion of male patients. Digital ulcers, renal crisis/anti RNA polymerase III antibodies, joint involvement were more frequently found in C4 than in the other clusters. Patients in the cluster C5 (n=1,673) were mainly female patients, older than in the other clusters, had a rather low mean Rodnan skin score, a high percentage of anti-topoisomerase 1 antibodies (46%) and a high percentage of lung fibrosis. Concerning survival, C2 had the best survival and C4 the worst. When compared to C1, C2 to C5 clusters had a significantly different prognosis (C2: adjusted hazard ratio [95% CI]: 0.70 [0.51-0.96]; C3: 2.27 [1.76-2.94]; C4: 4.52 [3.57-5.72]; and C5: 1.94 [1.51-2.44].

Conclusions This cluster analysis on the largest ever worldwide database of SSc patients showed 5 different clusters characterized by different sex ratio, maximum Rodnan skin scores, autoantibodies status, joint and organ involvement as well as by a different prognosis. Autoantibodies status seemed to play an important role for association with organ involvement. Although systemic sclerosis is a heterogeneous disease, this study shows that homogeneous groups beyond the classical limited/diffuse dichotomy can be delineated in this disease.

Disclosure of Interest None declared

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.