E Gervasoni1, D Cattaneo1, P Messina2, E Casati1, A Montesano1, E Bianchi2, E Beghi2. 1. LaRiCE: Gait and Balance Disorders Laboratory Department of Neurorehabilitation, Don Gnocchi Foundation I.R.C.C.S, Milan, Italy. 2. Laboratory of Neurological Disorders, IRCCS - Mario Negri Institute for Pharmacological Research Via La Masa, Milan, Italy.
Abstract
OBJECTIVES: Clinical predictors of falls in patients with Parkinson disease (PD) are fairly inaccurate. Stabilometric measures appear useful in investigating the relationship between balance, sensory disturbance, and falls. The aim of the study was to identify the best combination of clinical and stabilometric tests to predict falls prospectively. MATERIALS & METHODS: Fifty-three consecutive subjects with PD or parkinsonisms at risk of falls were included and followed for 6 months. Clinical variables were used as fall predictors: the Unified Parkinson Disease's Rating Scale (motor section) and the Longitudinal Aging study Amsterdam Physical Activity Questionnaire (LAPAQ). Variables from stabilometric platform underwent a principal component analysis. Multivariate logistic models were used to predict fallers using fall status (fallers: 1 + falls; recurrent fallers: 2 + falls) as dependent variable. RESULTS: Seven patients were lost to follow up, leaving 46 evaluable subjects. Of these, 32 (70%) were fallers and 22 (48%) were recurrent fallers. The only variable predicting fallers was the LAPAQ (odd ratio [OR] 0.99 (95% confidence interval [CI] 0.98-1.00); accuracy 71.7%; sensitivity 87.5%; specificity 35.7%). For recurrent fallers, Factor 2 (body sway velocity) (OR 2.37; 95% CI 1.01-5.58) and, in part, LAPAQ (OR 0.99; 95% CI 0.98-1.00) retained significance in the multivariate model, showing an accuracy of 76.9%, a sensitivity of 77.8%, and a specificity of 76.2%. CONCLUSIONS: A combination of clinical and instrumental tools is useful to identify fallers in PD or parkinsonisms. Body sway velocity and ability to perform the activities of daily living are the best predictors of recurrent falls.
OBJECTIVES: Clinical predictors of falls in patients with Parkinson disease (PD) are fairly inaccurate. Stabilometric measures appear useful in investigating the relationship between balance, sensory disturbance, and falls. The aim of the study was to identify the best combination of clinical and stabilometric tests to predict falls prospectively. MATERIALS & METHODS: Fifty-three consecutive subjects with PD or parkinsonisms at risk of falls were included and followed for 6 months. Clinical variables were used as fall predictors: the Unified Parkinson Disease's Rating Scale (motor section) and the Longitudinal Aging study Amsterdam Physical Activity Questionnaire (LAPAQ). Variables from stabilometric platform underwent a principal component analysis. Multivariate logistic models were used to predict fallers using fall status (fallers: 1 + falls; recurrent fallers: 2 + falls) as dependent variable. RESULTS: Seven patients were lost to follow up, leaving 46 evaluable subjects. Of these, 32 (70%) were fallers and 22 (48%) were recurrent fallers. The only variable predicting fallers was the LAPAQ (odd ratio [OR] 0.99 (95% confidence interval [CI] 0.98-1.00); accuracy 71.7%; sensitivity 87.5%; specificity 35.7%). For recurrent fallers, Factor 2 (body sway velocity) (OR 2.37; 95% CI 1.01-5.58) and, in part, LAPAQ (OR 0.99; 95% CI 0.98-1.00) retained significance in the multivariate model, showing an accuracy of 76.9%, a sensitivity of 77.8%, and a specificity of 76.2%. CONCLUSIONS: A combination of clinical and instrumental tools is useful to identify fallers in PD or parkinsonisms. Body sway velocity and ability to perform the activities of daily living are the best predictors of recurrent falls.