Background Brief cognitive assessments are of use both clinically and as outcome measures in clinical trials. To date, the most commonly employed of such measures has been the Mini-Mental State Examination but this has limited sensitivity to neurocognitive impairment in Huntington's disease (HD). A recently developed alternative to the Mini-Mental State Examination is the Montreal Cognitive Assessment (MoCA), which measures five indices of cognitive function, including language, memory, attention and executive function. Previous studies using the MoCA in HD have focused on direct comparison of these instruments; to date there has been no in-depth analysis of the sensitivity and specificity of the MoCA in HD.

Aims Our aim was to examine the sensitivity of the MoCA and its component subscales to both cognitive impairment and disease stage in HD. A further aim was to explore whether the instrument's sensitivity could be improved by the removal of items with limited sensitivity and the use of modified scoring methods.

Methods 59 manifest and premanifest HD patients were examined using the MoCA, in addition to the UHDRS. The association between total MoCA score and UHDRS indices of disease severity was examined. Following this, sensitivity of individual MoCA subscales to disease stage was explored using MANOVA and insensitive items were excluded. Finally, modified scoring criteria were composed and evaluated in terms of the strength of association between modified scores and disease stage.

Results Total MoCA scores were significantly associated with both disease stage (n=55, p=<0.01, f=18.87) and UHDRS measures. Several of the MoCA subscales were poorly associated with disease stage. Removing these items from the total MoCA score improved sensitivity to both disease stage (n=55, p<0.01, f=30.47) and UHDRS measures. Moreover, the use of modified scoring methods further improved association with disease stage and UHDRS outcome measures, as well as eliminating ceiling effects in premanifest HD.

Conclusions Although the MoCA is sensitive to neurocognitive function and disease stage in HD, the sensitivity of the instrument can be greatly improved by the exclusion of low sensitivity items and the use of expanded scoring criteria.