European multicentre study to define disease activity criteria for systemic sclerosis. II. Identification of disease activity variables and development of preliminary activity indexes

Objective-To develop criteria for disease activity in systemic sclerosis (S Sc) that are valid, reliable, and easy to use. Methods-Investigators from 19 European centres completed a standardised cli nical chart for a consecutive number of patients with SSc. Three protocol m anagement members blindly evaluated each chart and assigned a disease activ ity score on a semiquantitative scale of 0-10. Two of them, in addition, ga ve a blinded, qualitative evaluation of disease activity ("inactive to mode rately active" or "active to very active" disease). Both these evaluations were found to be reliable. A final disease activity score and qualitative e valuation of disease activity were arrived at by consensus for each patient ; the former represented the gold standard for subsequent analyses. The cor relations between individual items in the chart and this gold standard were then analysed. Results-A total of 290 patients with SSc (117 with diffuse SSc (dSSc) and 1 73 with limited SSc (lSSc)) were enrolled in the study. The items (includin g Delta -factors-that is, worsening according to the patient report) that w ere found to correlate with the gold standard on multiple regression were u sed to construct three separate 10-point indices of disease activity: (a) D elta -cardiopulmonary (4.0), Delta -skin (3.0), Delta -vascular (2.0), and Delta -articular/muscular (1.0) for patients with dSSc; (b) Delta -skin (2. 5), erythrocyte sedimentation rate (ESR) >30 mm/1st h (2.5), Delta -cardiop ulmonary (1.5), Delta -vascular (1.0), arthritis (1.0), hypocomplementaemia (1.0), and scleredema (0.5) for lSSc; (c) Delta -cardiopulmonary (2.0), De lta -skin (2.0), ESR >30 mm/1st h (1.5), total skin score >20 (1.0), hypoco mplementaemia (1.0), scleredema (0.5), digital necrosis (0.5), Delta -vascu lar (0.5), arthritis (0.5), TLCO <80% (0.5) for all patients with SSc. The three indexes were validated by the jackknife technique. Finally, receiver operating characteristic curves were constructed in order to define the val ue of the index with the best discriminant capacity for "active to very act ive" patients. Conclusions-Three feasible, reliable, and valid preliminary indices to defi ne disease activity in SSc were constructed.