Spanish scleroderma risk score (RESCLESCORE) to predict 15-year all-cause mortality in scleroderma patients at the time of diagnosis based on the RESCLE cohort: Derivation and internal validation

Abstract

A few scores predicting the short-term risk of mortality in Systemic sclerosis (SSc) have been reported to date. Our study aimed to create a predictive 15-year all-cause mortality score at the time of the diagnosis of SSc. The study was based on the Spanish Scleroderma Registry (RESCLE). The cohort was split up in derivation (DC) and validation cohort (VC). A multivariate analysis to detect variables related to all-cause mortality within the first 15 years from SSc diagnosis was performed, assigning points to the rounded beta values to create the score (RESCLESCORE). 1935 SSc patients were included. The variables in the final model were as follows: age at diagnosis (+2 points > 65 years-old), male gender (+1 point), lcSSc subset (-1 point), mode of onset other than Raynaud's (+1 point), cancer (+1 point) and visceral involvement, such as ILD (+1 point), PAH (+1 point), heart (+1 point) and renal involvement (+2 points). Autoantibodies did not achieve statistical significance in the multivariate analysis. The 3 categories of risk to predict 15-year all-cause mortality at the time of diagnosis were as follows: low risk (5% vs. 7%, p = .189), intermediate risk (26.5% vs. 25.5%, p = .911) and high risk (47.8% vs. 59%, p = .316). The AUC was 0.799 (DC) vs. 0.778 (VC) (p = .530). In conclusion, the RESCLESCORE demonstrated an excellent ability to categorize SSc patients at the time of diagnosis in separate 15-year all-cause mortality risk strata at the time of diagnosis.

Keywords: Mortality; Prognosis; Score; Systemic sclerosis.