PT  - JOURNAL ARTICLE
AU  - Ariane L. Herrick
AU  - Mark Lunt
AU  - Nina Whidby
AU  - Holly Ennis
AU  - Alan Silman
AU  - Neil MacHugh
AU  - Christopher P. Denton
TI  - Observational Study of Treatment Outcome in Early Diffuse Cutaneous Systemic Sclerosis
AID  - 10.3899/jrheum.090668
DP  - 2009 Dec 01
TA  - The Journal of Rheumatology
PG  - jrheum.090668
4099  - http://www.jrheum.org/content/early/2009/11/25/jrheum.090668.short
4100  - http://www.jrheum.org/content/early/2009/11/25/jrheum.090668.full
AB  - Objective Randomized clinical trials in early diffuse cutaneous systemic sclerosis (dcSSc) are challenging. We used an observational approach to estimate the relative effectiveness of different current treatment approaches, capturing entry and outcome data in a standardized way. Methods Patients with dcSSc within 3 years of the onset of skin thickening were included. Standardized entry and followup data were collected in relation to the first disease-modifying treatment at baseline and 4-6 weeks, then 3, 6, 12, 18, 24, 30, and 36 months. The 5 different protocols were (1) intravenous cyclophosphamide followed by mycophenolate mofetil (MMF); (2) antithymocyte globulin followed by MMF; (3) MMF alone; (4) no disease-modifying treatment; (5) other immunosuppressant treatment. The primary outcome measure was the modified Rodnan skin score (mRSS). Inverse probability of treatment weights were used to allow for differing patient characteristics between groups. Results The study included 147 patients from 12 centers. Numbers of patients starting on Protocols 1 to 5 were 29, 25, 61, 19, and 13, respectively. mRSS decreased over time from 24 (IQ 19–32) at baseline to 15.5 (IQ 9–24.5) at 3 years. Although there were differences in the magnitude of the change for different protocols, there were no significant differences between protocols in the rate of change of mRSS over time (p = 0.43). When inverse probability weights were applied, the results remained nonsignificant (p = 0.41). Conclusion Using this observational approach, there were no obvious differences in outcome between groups after allowing as far as possible for baseline differences in treatment allocations.