Relationship between mild cognitive impairment and falls in older people with and without Parkinson's disease: 1-Year Prospective Cohort Study

Abstract

We conducted a 12-month longitudinal cohort study of 102 older people without dementia (52 Parkinson's Disease [PD]; 50 age- and sex-matched controls) to determine (1) if mild cognitive impairment predicts falls in older people with or without PD and (2) how baseline falls, a history of freezing and Hoehn and Yahr stage affected the association between cognitive impairment and multiple falls in PD patients. Cognitive impairment was defined as the sum of impairments on the caregiver-rated Clinical Dementia Rating Scale (CCDRSum > 0). Overall the mean age (SD) was 71.5 (4.7) years, 42% were women, 26% had fallen and 14% had cognitive impairment at baseline. Thirty-one percent (15/52) of PD patients vs 12% (6/50) of controls fell more than once during 12-month follow-up, p = 0.04. When combined in a 2-predictor model for the entire cohort, the adjusted odds ratios [aOR] for falling were significantly increased for cognitive impairment at baseline (aOR: 4.8, 95% CI: 1.3–18.2) and prior falls (aOR: 7.4, 95% CI: 2.4–22.3). The overall accuracy of the model was 82%, with low sensitivity of 19% but high specificity of 99%. In the PD subgroup, the overall accuracy of the same prediction model was 79%, with much better sensitivity of 73% but a lower specificity of 81%. We conclude that mild cognitive impairment might contribute to falls risk beyond conventional risk factors in older people with and without PD.