Gait & Posture 37 (2013) 126–130 Contents lists available at SciVerse ScienceDirect Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost Gait parameter risk factors for falls under simple and dual task conditions in cognitively impaired older people Morag E Taylor a,b,**, Kim Delbaere a,c, A Stefanie Mikolaizak a, Stephen R Lord a,c, Jacqueline C.T Close a,b,* a b c Falls and Balance Research Group, Neuroscience Research Australia, Sydney, Australia Prince of Wales Clinical School, University of New South Wales, Sydney, Australia School of Public Health and Community Medicine, University of New South Wales, Sydney, Australia A R T I C L E I N F O A B S T R A C T Article history: Received 13 December 2011 Received in revised form 25 May 2012 Accepted 26 June 2012 Impaired gait may contribute to the increased rate of falls in cognitively impaired older people We investigated whether gait under simple and dual task conditions could predict falls in this group The study sample consisted of 64 community dwelling older people with mild to moderate cognitive impairment Participants walked at their preferred speed under three conditions: (a) simple walking, (b) walking while carrying a glass of water and (c) walking while counting backwards from 30 Spatiotemporal gait parameters were measured using the GAITRite1 mat Falls were recorded prospectively for 12 months with the assistance of carers Twenty-two (35%) people fell two or more times in the 12 month follow-up period There was a significant main effect of gait condition and a significant main effect of faller status for mean value measures (velocity, stride length, double support time and stride width) and for variability measures (swing time variability and stride length variability) Examination of individual gait parameters indicated that the multiple fallers walked more slowly, had shorter stride length, spent longer time in double support, had a wider support width and showed more variability in stride length and swing time (p < 0.05) There was no significant interaction between gait condition and faller status for any of the gait variables In conclusion, dual task activities adversely affect gait in cognitively impaired older people Multiple fallers performed worse in each gait condition but the addition of a functional or cognitive secondary task provided no added benefit in discriminating fallers from non-fallers with cognitive impairment ß 2012 Elsevier B.V All rights reserved Keywords: Gait Dementia Cognition Accidental falls Dual task Introduction Cognitive impairment and dementia are known risk factors for falls [1], and the incidence of falls in older people with cognitive impairment (60% per annum) is approximately double that of their cognitively intact counterparts Underlying mechanisms for this increased risk are poorly understood and gait dysfunction may play an important role Whilst considered primarily an automated motor process, there is evidence to support the involvement of higher level cognitive processes in the gait cycle [2,3] The simultaneous performance of an attention-demanding task while walking often results in changes in gait patterns, which is interpreted as interference * Corresponding author at: Neuroscience Research Australia, Barker Street, Randwick, NSW 2031, Australia Tel.: +61 293991055; fax: +61 293991204 ** Co-corresponding author at: Neuroscience Research Australia, Barker Street, Randwick, NSW 2031, Australia Tel.: +61 293991060; fax: +61 293991204 E-mail addresses: m.taylor@neura.edu.au (M.E Taylor), j.close@neura.edu.au (Jacqueline C.T Close) 0966-6362/$ – see front matter ß 2012 Elsevier B.V All rights reserved http://dx.doi.org/10.1016/j.gaitpost.2012.06.024 caused by competing demands for limited resources [2,3] Gait changes related to the addition of a secondary task have been associated with falls in cognitively intact older adults [4–8] Gait disorders including increased gait variability are common in people with cognitive impairment and dementia [9–16] and dual task activities have been shown to detrimentally affect gait in people with these conditions [10–14,16–19] Three studies have examined gait and prospective falls in older people with dementia under simple walking conditions In older people with dementia living in residential care, velocity, cadence, stride length, double support time variability, heel-to-heel base of support variability and stride length variability were important in discriminating between fallers and non-fallers [20–22] To date, no studies have examined whether dual task gait abnormalities are associated with falls in community dwelling cognitively impaired older people or whether the addition of the secondary task provides additional benefit in discriminating fallers from non-fallers with this condition The current study, therefore, investigated whether poor gait performance under simple, functional and cognitive dual task conditions increases the risk M.E Taylor et al / Gait & Posture 37 (2013) 126–130 of falls in community dwelling older adults with cognitive impairment Methods 2.1 Participants Sixty-four cognitively impaired participants were recruited from a number of routine health service settings, community services and advertisements in the local press Participation was dependent on (i) having an age 60 years or above, (ii) living in the community and (iii) having an identified and willing ‘‘person responsible’’ with at least 3.5 h of face to face contact per week Cognitive impairment was defined as a Mini-Mental State Examination (MMSE) < 24, Addenbrooke’s Cognitive Examination – revised (ACE-R) < 83 [23] or where a specialist clinician had made a diagnosis of cognitive impairment or dementia Exclusion criteria included recent stroke (within 18 months), progressive neurodegenerative disorders (excluding dementia), insufficient English to complete the tasks or known end stage illness The study was approved by the South East Sydney Human Research Ethics Committee and consent was obtained from all participants and their person responsible prior to assessment 2.2 Assessment Participants and their person responsible were initially assessed in their home environment with information obtained on demographics, medical and medication history and functional performance Consenting participants attended the gait laboratory for the gait assessment Temporal and spatial gait parameters were measured with the GAITRite1 mat (CIR Systems Inc., Clifton, NJ, USA) The standard GAITRite1 system is a 460 cm mat with an active area of 2.23 m2 containing 13,824 pressure sensors arranged in a grid pattern with a spatial resolution of 1.27 cm and a sampling frequency of 80 Hz Gait parameters were obtained from GAITRite Gold Software Version 3.3 Participants performed six trials on the GAITRite1 mat, wearing comfortable footwear, at their preferred walking speed and with their usual walking aid in a well lit and quiet room Three gait conditions were assessed, each with two trials; (i) simple walking, (ii) walking while carrying a glass of water in the participants preferred hand (filled to 10 mm from rim) (Functional dual task) and (iii) walking while counting backwards from 30 (Cognitive dual task) Two participants who mobilised with a walking frame did not perform the functional dual task Participants started one and a half metres before the GAITRite1 mat and were instructed to walk beyond the mat completing a further one and half metres They were instructed to concentrate equally on their walking and functional/cognitive activity under dual task conditions Two research assistants were present for all gait assessments Footfall data were recorded by the computerised walkway with manual correction of footfall labelling completed where necessary Seven gait variables were calculated using the mean of two trials; velocity (cm/s), cadence (steps/min), stride length (cm), double support time (s), stride width (cm), coefficient of variation of stride length (%) and coefficient of variation of swing time (%) The coefficient of variation (CoV) is a measure of variability and is expressed as the percentage of the ratio of the standard deviation to the mean Dual task cost was calculated for each gait variable under both dual task conditions These measures were chosen as they have previously been reported in the literature in relation to gait in cognitively impaired people [3,9,15,17,18,20,21] 127 2.3 Falls follow-up Monthly falls calendars and reply paid envelopes were given to the participants/ carers to ensure accurate data collection with regard to falls The fall definition articulated to the participant and carer was, ‘‘In the past month, have you had any falls including a slip or a trip in which you lost your balance and landed on the floor or ground or lower level?’’ [24] If a participant/carer failed to return a calendar, a telephone call was made to the carer/person responsible to obtain the participant’s falls data A multiple faller was defined as someone who fell at least twice during the 12 month follow-up period 2.4 Statistical analysis Data were analysed using SPSS 18.0 for Windows (SPSS, Inc., Chicago, IL) Coefficient of variation (CoV = SD/mean  100) was calculated for stride length (stride length variability) and swing time (swing time variability); the CoV for each trial was calculated and the mean of two trials is reported For variables with skewed distributions, data were log normalised Data on double support time, swing time, stride length variability and swing time variability were normalised using logarithmic transformations for all conditions Extreme scores were censored at a level of 3SD from the mean and were included in the analyses i.e for one case for the following variables: simple double support time, functional dual task swing time variability and stride length variability, cognitive dual task double support time and stride length variability; and for two cases for functional dual task swing time variability All parametric analyses were performed with the transformed data Univariate analyses were carried out using independent sample t-tests and Chi squared tests for cross-tabulation tables to compare non-multiple fallers and multiple fallers’ demographic and medical characteristics Data for each gait variable were analysed using repeated measures multivariate ANOVA with gait condition (simple vs functional vs cognitive) as the within-subject factor and faller status (0–1 vs 2+) as the between-subject factor Spatiotemporal gait measures and gait variability measures were analysed separately DTC was calculated for each variable for both functional and cognitive dual task conditions: Dual Task Cost = ([dual task À simple]/simple)  100 Results Falls data were available for 63 of the participants; one participant was lost to follow-up Thirty-four (54%) participants reported falls in the 12 month follow-up period, with 22 (35%) reporting multiple falls (!2) The demographic and medical characteristics of the non-multiple fallers and multiple fallers are shown in Table The spatiotemporal gait parameters under the three conditions and the dual task cost for non-multiple fallers and multiple fallers are displayed in Table 3.1 Effects of dual task condition Thirty-three participants (51.6%) made an iteration error while counting backwards and walking Five (7.8%) of 64 were unable to Table Participant characteristics: displaying p values for univariate analysis Characteristic Non-multiple fallers (