Cluster analysis reveals eosinophilia and fibrosis as poor prognostic markers in 128 patients with eosinophilic fasciitis

Abstract

Background: Eosinophilic fasciitis (EF) is an extremely rare disease with polymorphic presentation and prognosis.

Objective: To further investigate EF features.

Methods: We performed a retrospective multicentre study of EF patients from 2013 to 2019, clustered patients using multivariate correspondence analysis, and sought prognosis factors.

Results: One hundred twenty-eight patients were included. Sixty-nine (50%) patients had skin sclerosis, and eosinophil count was increased in 71 (55%) patients. Multivariate correspondence analysis identified 3 clusters: a "mild," a "late-onset and hypereosinophilic," and a "fibrotic" cluster. Of 109 patients who followed up for more than 1 year, 49 (45%) presented a relapse, and 48 (44%) had sequelae. Multivariate analysis revealed that eosinophilia (hazard ratio = 1.56; P = .02) and fibrosis (hazard ratio = 4.02; P = .002) were predictive factors of relapse, whereas edema (odds ratio = 0.31; P = .03), relapse (odds ratio = 3.00; P = .04) and fibrosis (odds ratio = 1) were predictive factors of sequelae. Following relapse, treatment modifications consisted of an increase in glucocorticoids in 40 (82%) patients and the addition of methotrexate in 31 (63%) patients. These modifications led to clinical improvement and glucocorticoid withdrawal in 37 (76%) and 22 (45%) patients.

Limitations: Retrospective study.

Conclusion: EF patients can be divided into 3 homogenous clusters, which, along with fibrosis and eosinophilia, are prognosis factors of relapse and sequelae.

Keywords: cluster analysis; eosinophilic fasciitis; fibrosis eosinophils; relapse; sequelae.