THU0393 The predictor of malnutrition in systemic sclerosis (PREMASS) score: a validated combined index predictive of future weight loss in systemic sclerosis

Background Malnutrition and severe gastrointestinal dysfunction are the cause of mortality in 4%–15% of systemic sclerosis (SSc) patients whereas overall gastrointestinal involvement is observed in 75%–90% of cases.1 Hence, a reliable tool for stratification of risk for malnutrition would be of great value in the clinical management of SSc. Objectives Here we set out to identify a combined index predictive of significant weight loss at 12 months employing Malnutrition Universal Screening Tool (MUST) and serum adiponectin to leptin ratio (A/L) already used in other conditions. Methods This was an international, multicentre, longitudinal study employing 180 SSc patients in two independent cohorts: a study cohort (110 consecutive SSc patients) enrolled from University of Messina (60) and University of Padova,50 and a validation cohort (70) at the University of Leeds. Serum A/L ratio was measured by ELISA. MUST score, which includes BMI and weight loss reported by the patient in the last 3–6 months, was calculated as described: 0=no, 1=mild,>2= moderate/severe risk of malnutrition. End point of the study was weight loss >10% of baseline weight at 12 months. Results The two cohorts showed no significant differences in demographic and clinical features. Overall, median BMI decreased over time in both study and validation cohorts (23.5 vs 22.35 and 23.44 vs 22.49, respectively; p 10% wt despite having “no” or “mild” MUST scores. Logistic regression analysis identified the combination of BMI and A/L as the best PREdictor of MAlnutrition in Systemic Sclerosis (PREMASS). The formula 12.18-(0.63*BMI)+(1.51*A/L) predicted the end point with AUC=0.91 (95% CI:0.77–0.84). A PREMASS score >0.23 showed 91.3% sensitivity (95% CI:79.79–100) and 80.46% specificity (95% CI:72.13–88.79) for >10% wt loss with an overall 55.26% positive predictive value (PPV) (95% CI:39.45–71.07) and 97.22% negative predictive value (NPV) (95% CI:93.43–100) and a relative risk (RR) of 19.90 (95% CI:4.93–80.37). In the validation cohort, PREMASS showed 76.47% sensitivity (95% CI:56.31–96.63) and 75.47% specificity (95% CI:63.89–87.06) with an overall 50% PPV (95% CI:30.78–69.22) and 90.91% NPV (95% CI:82.41–99.4) and a RR of 5.5 (95% CI: 2–15.10). Conclusions PREMASS is the first validated index for weight loss risk stratification in the following 12 months in SSc. Prediction of future weight loss in SSc could aid both in clinical management and stratification/enrichment in clinical trials. Disclosure of Interest None declared