Reliability of a cognitive endpoint for use in a multiple sclerosis pharmaceutical trial

Objective Determine reliability and basic psychometric properties of a composite cognitive endpoint, MS-COG, for monitoring change in cognitive function in MS drug trials. Background 50% of MS patients have cognitive impairment that impacts ability to work and quality of life. We selected neuropsychological tests based on sensitivity to MS cognitive impairment, availability of alternate forms, cross-cultural utility, and feasibility for multicenter trials, and assessed the reliability and validity of a composite endpoint, MS-COG. Design/methods Administered SRT, BVMT-R, PASAT, and SDMT to 60 MS patients at 4 US centers twice over 45 days, along with symptom inventories by patients and informants. Results The MS-COG had test-retest reliability of 0.91. Processing Speed and Memory indices had reliabilities of 0.89 and 0.86, with modest practice effects. Reliability was high for the RR MS and SP MS subgroups as well, with correlations of.90 and.93, respectively for MS-COG. Overall, 42% of subjects obtained MS-COG scores in the impaired range, with SP MS subjects performing 0.8 SD below RR MS subjects. Impairment correlated well (r = 0.37 to 0.40) with informant reports but was inconsistent with patient report, with the least reliable assessments by those with greater symptom severity. Conclusions The MS-COG is a reliable, repeatable measure of MS cognitive functioning that is sensitive to cognitive impairment in SP MS and RR MS patients and feasible for multicenter clinical trials. Further development is warranted.