Timed Up-and-Go Dual-Task Tests for Early Detection of Dementia Disorder

Abstract: Dementia constitutes an important and growing public health concern. There is a need for new, simple, and inexpensive methods to detect dementia disorders early in the disease progression. For this purpose, dual-tasking, i.e., simultaneous performance of two tasks, has been proposed.The overall aim of this thesis was to explore if Timed Up-and-Go (TUG) dual-task (TUGdt) tests can be used for early detection of dementia disorder. Cross-sectional and longitudinal designs were used. Participants were recruited when undergoing memory assessment at memory clinics (patients) and through advertisements (controls). The TUGdt tests involved TUG combined with the cognitive tasks a) naming animals (TUGdt NA) and b) reciting months in reverse order (TUGdt MB). The tests were video recorded. Test outcomes were calculated using time scores and/or verbal performances. Additionally, the data collection comprised clinical tests and medical record reviews. Paper I included 90 patients who had carried out lumbar puncture as part of the memory assessment. By Spearman’s rank correlation, the TUGdt NA test outcomes “number of animals” and “animals/10 s” correlated negatively to the cerebrospinal fluid biomarkers t-tau and p-tau, suggesting that neurodegeneration is associated with dual-task performance. In Paper II, 298 patients and 166 controls participated. Logistic regression models showed that “animals/10 s” and “months/10 s” discriminated significantly between dementia, mild cognitive impairment (MCI), subjective cognitive impairment (SCI), and controls. Thus, TUGdt testing could be useful in diagnostic assessments. Paper III involved 172 patients, initially diagnosed with MCI or SCI, for whom diagnostic information was available after 2.5 years. Logistic regression showed inverse associations between “animals/10 s” and dementia incidence, particularly for patients <72 years (median age). For these younger patients, the predictive capacity of “animals/10 s” was excellent. Hence, TUGdt NA has potential for predicting dementia from SCI or MCI, particularly among younger patients. Paper IV included 166 controls for presenting TUGdt reference values in age- and sex-specific groups, and 43 controls for test-retest reliability. Reference values were calculated with quantile regression and may be useful in clinic and research. Intra-class correlation coefficients showed excellent reliability for time scores, while the other test outcomes were poor to good. “Animals/10 s” showed fair to good reliability despite being a ratio of other variables, which negatively affects reliability. In summary, TUGdt NA has the potential to be used for early detection of dementia disorder, and the test outcome “animals/10 s” merits further evaluation.