• Rheumatology Guidelines
• Featured Meeting - Specialty page
• Rheumatological Autoimmune Disorders

Proposed new classification criteria developed for systemic sclerosis (scleroderma; SSc) have improved sensitivity and specificity compared with the 1980 American College of Rheumatology (ACR) SSc criteria and should allow for more patients to be classified as having SSc. The proposed classification criteria are preliminary and need to be reviewed by the ACR and the European League Against Rheumatism (EULAR), but the authors do not anticipate they will be altered from the findings presented here.

As reported by Janet E. Pope, MD, St. Joseph's Health Care and University of Western Ontario, London, Ontario, Canada, the 1980 Preliminary Criteria for the Classification of Systemic Sclerosis failed to classify a significant proportion of patients with early SSc and patients with the limited subtype of the disease, who experienced clinicians believed should be classified as having SSc [Pope JE et al. ACR 2012 Poster L3].

A committee to develop new criteria was established jointly by ACR and EULAR. The committee used an 8-step process that included first using an Internet survey of more than 100 potential criteria for SSc sent to multiple experts and narrowing the number down significantly, using a Delphi technique to further reduce the number of criteria, and testing the validity of 23 items selected in existing databases of SSc cases and controls from North America and Europe (further decreasing the number of items to 17). Twenty cases that represented the spectrum of SSc (low probability to high probability) were then used. The cases were ranked by experts, using conjoint analysis to assign weights of importance to 17 preliminary items.

According to Prof. Pope, it was agreed that the 1980 major criteria still worked well in classifying sclerodactyly that was continuous and proximal to the metacarpophalangeal joints (ie, the former major criterion for SSc that is still in the new proposed classification). A provisional threshold was established to classify definite SSc based on the sum of the weights of the 17 items. Experts collected serial cases of new SSc and prevalent SSc, and controls from multiple sites where the sensitivity and specificity of the final criteria were tested and validated. To test the provisional algorithm, data on the 17 items were collected from 605 cases and controls (possible mimickers) in North America and Europe.

From the collected data, the threshold was refined in a subset of 25 cases within the range of borderline probability of SSc. Experts were then asked to determine whether each case had “definite SSc” or not, which led to a new threshold.

The final items with the proposed weights of the classification system are presented in Table 1. A cutoff of ≥9 indicates SSc. These criteria will be vetted through the ACR and EULAR committees and then finalized. Items are added (maximum score in each category) so they are slightly similar to the recent rheumatoid arthritis criteria where weighted items are added to exceed a threshold score or not.

These 17 items were reduced to 9 during a face-to-face meeting of the steering committee, while maintaining adequate sensitivity and specificity, as tested in a random sample of cases and controls (n=200) from North America and Europe. In the validation cohort of 405 cases, a score of ≥9 showed a sensitivity of 91% and a specificity of 92%, whereas the 1980 criteria had a sensitivity of 75% and a specificity of 72% in this cohort (Table 2).

• © 2012 MD Conference Express®