The PREdictor of MAlnutrition in Systemic Sclerosis (PREMASS) Score: A Combined Index to Predict 12 Months Onset of Malnutrition in Systemic Sclerosis

Abstract

Objective: Malnutrition is a severe complication in Systemic Sclerosis (SSc) and it is associated with significant mortality. Notwithstanding, there is no defined screening or clinical pathway for patients, which is hampering effective management and limiting the opportunity for early intervention. Here we aim to identify a combined index predictive of malnutrition at 12 months using clinical data and specific serum adipokines. Methods: This was an international, multicentre observational study involving 159 SSc patients in two independent discovery (n = 98) and validation (n = 61) cohorts. Besides routine clinical and serum data at baseline and 12 months, Malnutrition Universal Screening Tool (MUST) score and serum concentration of leptin and adiponectin were measured for each participant at baseline. The endpoint of malnutrition was defined according to European Society of Clinical Nutrition and Metabolism (ESPEN) recommendation. Significant parameters from univariate analysis were tested in logistic regression analysis to identify the predictive index of malnutrition in the derivation cohort. Results: The onset of malnutrition at 12 months correlated with adiponectin, leptin and their ratio (A/L), MUST, clinical subset, disease duration, Scl70 and Forced Vital Capaciy (FVC). Logistic regression analysis defined the formula: -2.13 + (A/L^*0.45) + (Scl70^*0.28) as the best PREdictor of MAlnutrition in SSc (PREMASS) (AUC = 0.96; 95% CI 0.93, 0.99). PREMASS < -1.46 had a positive predictive value (PPV) > 62% and negative predictive value (NPV) > 97% for malnutrition at 12 months. Conclusion: PREMASS is a feasible index which has shown very good performance in two independent cohorts for predicting malnutrition at 12 months in SSc. The implementation of PREMASS could aid both in clinical management and clinical trial stratification/enrichment to target malnutrition in SSc.

Keywords: adipokines; autoimmune disease; malnutrition; outcome research; systemic sclerosis.

Figure 1

Linear regression analysis shows the association between body mass index (BMI) and adiponectin to leptin ratio (A/L ratio) at baseline for discovery cohort (A) and validation cohort (B). A significant inverse correlation was observed between A/L and BMI in both cohorts. Red symbols show progressors (patients developing malnutrition at 12 months). The x axis (A/L ratio) was stretched to a logarithmic scale (log10) in both panels. Dotted lines represent the BMI cut-off according to age (BMI <20 kg/m² if age <70 years or <22 kg/m² if age >70 years) for the identification of future malnutrition at 12 months according to the European Society for Clinical Nutrition and Metabolism (ESPEN) definition.

Figure 2

In the derivation cohort, resulting from the combination of the discovery cohort and the validation cohort, 15.7% of patients (n = 25, A) developed malnutrition at 12 months (progressors). Of note, 29% of patients having a low to moderate MUST score experienced malnutrition at 12 months and, on the other side, 77% of patients having a high risk of malnutrition (MUST ≥ 2) did not develop malnutrition at 12 months (B). (C) shows the performance of MUST in predicting malnutrition at 12 months in the derivation cohort.

Figure 3

Receiver operating characteristics (ROC) curve for the variables associated with malnutrition in the derivation cohort [adiponectin, leptin, MUST, A/L, FVC, Scl70, clinical subset, and disease duration. AUC and 95% CI are reported for each variable in the legend. MUST, Malnutrition Universal Screening Tool; A/L, adiponectin to leptin ratio; FVC, Forcev Vital Capacity; Scl70, antitopoisomerase I antibodies.

Figure 4

Receiver operating characteristics (ROC) curve (A) demonstrates that the best performance in predicting malnutrition at 12 months in the derivation cohort was represented by a combined index based on adiponectin to leptin ratio (A/L) and antitopoisomerase I antibodies (Scl70). Using the formula malnutrition = −2.13 + (A/L*0.45) + (Scl70*0.28), we validated the PREdictor of MAlnutrition in Systemic Sclerosis (PREMASS) index. (B) Shows the AUC, sensitivity, specificity, negative predictive value (NPV) and positive predictive value (PPV) for the discovery cohort, validation cohort, and derivation cohort. The percentage of progressors and non-progressors in the derivation cohort (C) are shown according to the cut-off (−1.46) of PREMASS optimized for sensitivity and specificity.