Supplementary Material for: Performance of a modified version of the Charlson Comorbidity Index in predicting Multiple Sclerosis disability accrual

Background: The natural history of Multiple Sclerosis (MS) is highly heterogeneous and almost unpredictable since several factors may affect the disease course including comorbidities. The aims of this study are to predict the risk of disability worsening and disease progression at first patient’s visit by using a modified version of the Charlson Comorbidity Index (mCCI). Methods: the mCCI was obtained by incorporating the grade of pyramidal functional system scores extracted by the Expanded Disability Status Scale (EDSS) into the original CCI version. The risk of reaching EDSS 4, EDSS 6 and secondary MS progression (SPMS) associated to mCCI classes was calculated by carrying out multivariable Cox regression models and it was reported as hazard ratios (HRs) and 95% confidence intervals (95% CIs). The accuracy of mCCI for the recognition of individuals who reached the study milestones was estimated by building the receiving operator curves and the optimal cut-off values were estimated. Results: a total of n=622 individuals were enrolled (72.7% women; median age 30.8 years [24-40]). Compared with patients with a mCCI equal to zero, the HRs for those with a mCCI comprised between 1 and 2 at first visit were 1.53 (1.1-2.1), 2.17 (1.48-2.96) and 1.57 (1.16-2.1) for the reaching of EDSS 4, EDSS 6 and SPMS, respectively. Moreover, individuals with a mCCI equal or higher than 3 were at even higher risk of reaching EDSS 6 (HR= 2.34; [1.44-3.8]) and SPMS conversion (HR= 2.38; [1.29-4.01]). The mCCI cut-off value of 3 reached a sensitivity and specificity of 88.1% and 77.8%, respectively, for the recognition of EDSS 4 while the mCCI cut-off of 4 reached a sensitivity of 83.1% and a specificity of 80.7% for the recognition of EDSS 6 and a sensitivity and a specificity of 76.8% and 87.5%, respectively, for the recognition of SPMS conversion. Conclusion: mCCI appeared a simple and fast tool for the prediction of MS prognosis since the first patient’s visit and its best cut-off values showed higher sensitivity and specificity for the recognition of patients who undergo disability worsening and SPMS conversion.