Prognostically distinct clinical patterns of systemic lupus erythematosus identified by cluster analysis

Lupus. 2009 Dec;18(14):1267-75. doi: 10.1177/0961203309345767. Epub 2009 Oct 27.

Abstract

The objective of this study was to evaluate the patterns of clinical manifestations and their mortality in a large cohort of Chinese patients with systemic lupus erythematosus. The cumulative clinical manifestations of a large group of Chinese systemic lupus erythematosus patients who fulfilled at least four American College of Rheumatology criteria for systemic lupus erythematosus were studied. Patients were divided into distinct groups by using the K-mean cluster analysis. Clinical features, prevalence of proliferative lupus nephritis (World Health Organization class III, IV), autoantibody profile, and treatment data were compared and the standardized mortality ratios were calculated for each cluster of patients. There were 1082 patients included in the study (mean age at systemic lupus erythematosus diagnosis 30.5 years; mean systemic lupus erythematosus duration 10.3 years). Three distinct groups of patients were identified. Cluster 1 (n = 347) was characterized predominantly by mucocutaneous manifestations (malar rash, discoid rash, photosensitivity, oral ulcer) and arthritis but having the lowest prevalence of serositis, hematologic manifestations (hemolytic anemia, leukopenia, and thrombocytopenia), and proliferative lupus nephritis. Patients in cluster 2 (n = 409) had mainly renal and hematological manifestations but having the lowest prevalence of mucocutaneous manifestations. Pulmonary and gastrointestinal manifestations were significantly more frequent in cluster 2 than the other clusters. Cluster 3 patients (n = 326) had the most heterogeneous features. Besides having a high prevalence of mucocutaneous manifestations, serositis and hematologic manifestations, renal involvement, and proliferative lupus nephritis was also most prevalent among the three clusters. Patients in cluster 2 had a much higher standardized mortality ratio [standardized mortality ratio 7.23 (6.7-7.7), p < 0.001] than those in cluster 3 [standardized mortality ratio 1.27 (1.1-1.5), p = 0.005] and cluster 1 [standardized mortality ratio 0.95 (0.5-1.7), p = 0.86]. In conclusion, patients with systemic lupus erythematosus could be clustered into prognostically distinct patterns of clinical manifestations.

MeSH terms

  • Adolescent
  • Adult
  • Age of Onset
  • Asian People / statistics & numerical data*
  • Autoantibodies / blood
  • Cause of Death
  • Cluster Analysis
  • Comorbidity
  • Female
  • Hong Kong / epidemiology
  • Humans
  • Immunosuppression Therapy
  • Lupus Erythematosus, Systemic / ethnology*
  • Lupus Erythematosus, Systemic / immunology
  • Lupus Erythematosus, Systemic / mortality*
  • Lupus Nephritis / ethnology*
  • Lupus Nephritis / immunology
  • Lupus Nephritis / mortality*
  • Male
  • Middle Aged
  • Multivariate Analysis
  • Prevalence
  • Prognosis
  • Risk Factors
  • Young Adult

Substances

  • Autoantibodies