Significant cognitive impairment has been found in 20-30% of patients with clinically isolated syndromes suggestive of multiple sclerosis. In this study we aimed to assess the prognostic value of the presence of cognitive impairment for the conversion to multiple sclerosis in patients with clinically isolated syndromes. All patients with clinically isolated syndromes consecutively referred to our centre since 2002 and who had been followed-up for at least one year underwent cognitive assessment through the Rao's Battery and the Stroop test. Possible predictors of conversion to clinically definite multiple sclerosis were evaluated through the Kaplan Meier curves and Cox regression analysis. A total of 56 patients (41 women; age 33.2 +/- 8.5 years; expanded disability scale score 1.2 +/- 0.7) were recruited. At baseline, 32 patients (57%) fulfilled McDonald's criteria for dissemination in space. During the follow-up (3.5 +/- 2.3 years), 26 patients (46%) converted to a diagnosis of multiple sclerosis. In particular, 64% of patients failing >or= 2 tests and 88% of patients failing >or= 3 tests converted to multiple sclerosis. In the Cox regression model, the failure of at least three tests (HR 3.3; 95% CI 1.4-8.1; p = 0.003) and the presence of McDonald's dissemination in space at baseline (HR 3.8; 95% CI 1.5-9.7; p = 0.005), were found to be predictors for conversion to multiple sclerosis. We conclude that cognitive impairment is detectable in a sizable proportion of patients with clinically isolated syndromes. In these subjects cognitive impairment has a prognostic value in predicting conversion to multiple sclerosis and may therefore play a role in therapeutic decision making.