Background Cognitive decline is one of the most marked symptoms of Huntington's Disease (HD). Hence, sensitive assessments that can measure subtle signs of declining cognitive function and serve as endpoints in interventional designs are needed. Currently, the most commonly used measure is the cognitive battery from the Unified Huntington's Disease Rating Scale (UHDRS), comprising a version of the Stroop colour-word-test, a verbal fluency test, as well as the Symbol Digit Modalities Test.

Aims We wanted to investigate how precisely the tests of the UHDRS cognitive battery measure disease severity and progression in the cognitive domain. In the longitudinal REGISTRY cohort we analysed the performance of these cognitive tests over four consecutive years.

Materials and Methods In a dataset of 2703 Subjects from 16 countries several statistical measures were applied. All cognitive tests including two additional interference score variants for the Stroop test were correlated to other disease markers in HD. Mean differences from baseline were computed for three consecutive years and tested for statistical significance. A mixed model was run to estimate the yearly change in per cent for each of the test scores.

Results Correlations to other disease markers were low to moderate. All tests failed to show significant decline after 1 year and most also failed to show significant decline after 2 years. Yearly change in per cent as calculated by the mixed model was not significant for Stroop word reading, the two Stroop interference scores and the verbal fluency test. For all other tests, percentages of change were small but significant.

Conclusions The UHDRS cognitive battery seems to be insensitive to subtle change in the cognitive domain and may not be suitable as a sole endpoint in interventional studies.