Background/Purpose:   Systemic sclerosis (SSc) is a rapidly evolving field with multiple potential therapeutic agents under development and several active clinical trials focused on treating diffuse cutaneous SSc.   When weighing management decisions for therapy and considering clinical trial involvement, estimating the likelihood of skin thickness progression over the next year will be helpful. Objectives:  To derive a clinical risk prediction tool to predict skin thickness progression over one year.

Methods:   We used a US single-center cohort of prospectively followed early diffuse SSc patients seen for an initial visit from Jan. 1, 1980 to Dec. 31, 2014.   To reflect current early diffuse SSc clinical trial design, early was defined as <3 years after the first non-Raynaud symptoms and diffuse skin as proximal to the elbows or knees.   Eligible patients had ≥ 2 modified Rodnan skin scores (mRSS) within one year of the first SSc center visit, for ≥ 3 mRSS over one year.  The outcome was mRSS progression at any time over one year of follow-up, defined as an increase of ≥5 points and a 25% increase in the mRSS from baseline.   SSc autoantibodies were confirmed by the gold standard methods of immunodiffusion or immunoprecipitation.  Skin change was assessed in multiple ways, including Receiver Operating Curve (ROC) analysis  to determine a risk cut-off point for mRSS.  Multivariable logistic regression modeling was used.  Beta-estimates were rounded to the nearest 0.5, summed and ROC analysis performed.

Results:   Among 317 eligible patients the mean age was 50.8 ± 13.5 years, 74% female and 92% Caucasian.  The median disease duration from first SSc symptom was 0.90 years (IQR 0.63, 1.46).  The mean first visit mRSS was 23±11.  25% (n=79) were scl-70 positive, 56% (n=178) RNA-polymerase III positive (RNAP) positive, 54 (17%) other SSc-associated antibody and 2% (n=6) missing.    In total, 163 patients (51%) developed skin progression over one year.  The mean time to peak mRSS among progressors was 0.51±0.23 years from the first visit.   ROC analysis showed mRSS ≤27 to best predict skin progression.  A 4-variable model was developed to predict skin progression over one year (Table 1) with a AUC = 0.81 (95% CI 0.75, 0.88).  When β-estimates were rounded and summed as in Table 1, a 3-level risk stratification model was created with rate of skin progression depicted in Table 2.   The AUC for this model remained excellent at 0.80 (0.75, 0.85). 

Conclusion:    We have derived an accurate 3-level clinical risk stratification tool for skin thickness progression over one year of follow-up in early diffuse SSc patients.