Background and objective: We explored which clinical and biochemical variables predict conversion from clinically isolated syndrome (CIS) to clinically definite multiple sclerosis (CDMS) in a large international cohort. Methods: Thirty-three centres provided serum samples from 1047 CIS cases with at least two years’ follow-up. Age, sex, clinical presentation, T2-hyperintense lesions, cerebrospinal fluid (CSF) oligoclonal bands (OCBs), CSF IgG index, CSF cell count, serum 25-hydroxyvitamin D3 (25-OH-D), cotinine and IgG titres against Epstein-Barr nuclear antigen 1 (EBNA-1) and cytomegalovirus were tested for association with risk of CDMS. Results: At median follow-up of 4.31 years, 623 CIS cases converted to CDMS. Predictors of conversion in multivariable analyses were OCB (HR = 2.18, 95% CI = 1.71–2.77, p < 0.001), number of T2 lesions (two to nine lesions vs 0/1 lesions: HR = 1.97, 95% CI = 1.52–2.55, p < 0.001; >9 lesions vs 0/1 lesions: HR = 2.74, 95% CI = 2.04–3.68, p < 0.001) and age at CIS (HR per year inversely increase = 0.98, 95% CI = 0.98–0.99, p < 0.001). Lower 25-OH-D levels were associated with CDMS in univariable analysis, but this was attenuated in the multivariable model. OCB positivity was associated with higher EBNA-1 IgG titres. Conclusions: We validated MRI lesion load, OCB and age at CIS as the strongest independent predictors of conversion to CDMS in this multicentre setting. A role for vitamin D is suggested but requires further investigation.