Critically Appraised Topics (CATs)

CAT 3. Effective interventions for preventing or reducing falls Appraised by Kevin Galbraith, December 2016 Update due 2019

Clinical question Search terms Among people with Parkinson’s, what are the most (Parkinson Disease/ OR Parkinson$) AND effective interventions for preventing or reducing falls? (Randomised Controlled Trials as topic/ OR randomi$) AND (Accidental Falls/ OR fall$) AND prevent$. Background Search strategy Among people with Parkinson’s, falls are both Ovid Medline, and adapted for Cochrane Library, frequent and recurrent, with 45–68% and CINAHL. All searches 2010 to August 2016, annually, and two-thirds of these falling English language. recurrently.1 Falls, impairment, and postural instability increase the risk of mortality and Evidence morbidity. Falls in PD can lead to fractures, From 31 articles, 10 addressed fall incidence or increased dependency, nursing home admission and frequency as a clinically relevant outcome. One 2 poor quality of life. This CAT sought to identify additional phase 2 trial, which investigated the the most effective interventions for preventing or effect of rivastigmine on gait stability,3 was added reducing falls among people with Parkinson’s. after discussion with the Evidence Based Practice Themed Working Group, in view of wide publicity.

Clinical bottom line Summary of evidence Evidence was found from well-conducted The search conducted in this study yielded 10 randomised controlled trials (RCTs) that intervention studies with fall reduction as an outcome: progressive resistance strength training and one meta-analysis of exercise and motor training;4 movement strategy training (both combined with three RCTs evaluating training specifically;5-7 fall prevention education), and are effective one RCT evaluating two specific physiotherapy in preventing falls. Rivastigmine has been assessed interventions (resistance strength training and but has not been subject to a phase 3 trial, and movement strategy training), both combined with cannot be recommended. There is a continued fall prevention education;8 one systematic review9 need for well-conducted RCTs in fall prevention, and one trial10 evaluating sunlight exposure; and three and for meta-analyses that can detect possible studies evaluating Tai chi - a systematic review,11 12 real effects from existing underpowered studies. a randomised controlled trial and a commentary on a trial published in Chinese.13 Two trials were added after discussion among the Evidence-Based Practice Themed Working Group: a phase 2 trial of rivastigmine3, and a trial of treadmill training enhanced with virtual reality (VR) technology23. Exercise and motor training Sunlight exposure Evidence was from one meta-analysis.4 It included Evidence was from one systematic review,9 which only two trials of interventions to reduce falls, and included only one study in patients with Parkinson’s.10 failed to demonstrate a reduction in falling. One of This study has been retracted, so neither it, nor the the trials did not include highly challenging balance systematic review were critically appraised. exercises,14 and did not aim to demonstrate a reduction in falls; the other trial failed to give a clear Tai Chi 15 description of the balance training intervention. Evidence was from one systematic review, one RCT, and one commentary on a trial published in Chinese, Balance training described here in chronological order. Evidence was from three RCTs that included balance

Critically Appraised Topics (CATs) Topics Appraised Critically training in the intervention. No trials of only balance The systematic review of Tai Chi concluded from training were found.5-7 Two lacked sufficient power a search up to 2008 that there was insufficient to show real effects, and relied on self-reporting evidence that it is an effective intervention for of falls.6-7 One was well-conducted overall, but Parkinson’s.11 One included RCT found Tai Chi to suffered significant attrition: 25 participants (22%) be superior to conventional exercise in terms of undertook a modified program to account for pain the Unified PD Rating Scale UPDRS, p=0.026 and and coexisting conditions, and 24 participants frequency of falls p=0.009.16 A second RCT found (21%) discontinued the exercise program.5 This no effect on locomotor ability compared with qigong trial identified a lower fall-rate among a lower meditation.17 This examined surrogate outcomes severity subgroup. Though pre-planned, such (gait velocity, stride length) rather than the more a subgroup analysis should be interpreted with clinically relevant fall frequency. Two other trials also caution, particularly as no Bonferroni adjustment employed surrogate outcomes. One was a crossover was made for multiple analyses. Two trials excluded trial, which investigated the effect of Tai Chi versus patients with cognitive impairment, reducing their wait list controls on UPDRS, Parkinson’s Disease generalisability to the wider population.5-7 One Questionnaire 39 (PDQ39), and the Timed Up and did not employ blinding of assessors to the group Go (TUG) test.18 After 12 weeks, there were no allocation.6 significant differences for any outcome. The other was a non-randomised trial comparing Tai Chi, a Resistance strength training and combined exercise program, and no treatment.19 movement strategy training After 8 weeks, the UPDRS, Schwab and England Evidence was from one RCT, in which 210 patients Activity of Daily Living, and quality of life scores with Parkinson’s were randomised to receive either were higher for Tai Chi than for no exercise. No progressive resistance strength training plus fall significant difference was found between Tai Chi prevention education N=70, or movement strategy and the combined exercise program. Other included training plus fall prevention education N=69, or the studies were of poor quality. control intervention of life skills information N=71.8 The strength-training group had 84.9% fewer falls The RCT published in 2012 however, found the than controls (incidence rate ratio [IRR] = 0.151, 95% incidence of falls to be lower with an individualised CI 0.071-0.322, p<0.001. The movement strategy programme of Tai Chi than with stretching (incidence training group had 61.5% fewer falls than controls (IRR rate ratio, 0.33; 95% CI, 0.16 to 0.71) or resistance = 0.385, 95% CI 0.184-0.808, p=0.012. This was a training (incidence rate ratio, 0.47; 95% CI, 0.21 to well conducted RCT. Attention was paid to allocation 1.00), though the latter did not achieve significance.20 concealment, assessors were blind to group allocation, After follow-up for 3 months, patients who received and the assessments were standardised. Analyses Tai Chi had fewer falls than those in the stretching were conducted according to the intention-to-treat group (incidence rate ratio, 0.31; 95% CI, 0.14 to principle. Falls were self-recorded on fall calendars 0.67; p=0.003) and those in the resistance-training over a long follow-up period of 12 months. The group (incidence rate ratio, 0.40; 95% CI, 0.18 to 0.88; study excluded patients with cognitive impairment, p=0.02). The study was well conducted. Assessors and severe Parkinson’s. This could impair its external were blind to group allocation, analysis was conducted validity in the wider population. according to the intention-to-treat principle, and losses to follow up were minimal. Falls were self-reported Virtual-reality-enhanced however, and patients with cognitive impairment (itself treadmill training a risk factor for falls21) were excluded. A multicentre, single-blinded randomised controlled The commentary13 described a single-blind RCT trial was conducted among adults aged 60−90 published in 2014,22 of Tai Chi (three sessions years with a history of two or more falls in the per week for 12 weeks) versus no Tai Chi, among previous 6 months23. They were randomly allocated Parkinson’s patients drawn from hospital and to either treadmill training enhanced by virtual local communities in China. The risk of falls in the reality (VR) technology (n=154) or treadmill intervention group was significantly lower than training alone (n=148). The VR component involved among controls (relative risk 0.44, 95% CI 0.22 to digital capture of foot movements and projection 0.89). Falls were self-reported using a falls diary, but of these onto a visual simulation, complete with checked by monthly phone calls. The commentator obstacles, multiple pathways and distractors

Critically Appraised Topics (CATs) Topics Appraised Critically noted that the intervention was not clearly described, necessitating continual adjustment of steps. and that replication in other settings might require an Randomisation was stratified in each clinical site by experienced instructor, which might be difficult in some sex and by 3 subgroups: (1) history of idiopathic countries. It was helpfully calculated that for every four falls (n=109); (2) mild cognitive impairment people (95% CI 2 to 18) who undertake Tai Chi, one (n=43); and Parkinson’s (n=130). Among the person avoids a fall who otherwise would have fallen. participants with Parkinson’s, the fall incidence after training was lower in the treadmill training plus Rivastigmine VR group than in the treadmill training group (IRR Evidence from a single phase 2 pilot trial, in which 0·45, 95% CI 0·24−0·86; p=0·015). After adjusting 130 patients with moderate Parkinson’s were for disease severity, the IRR was 0·47 (95% CI randomised to receive the anticholinesterase 0·25−0·89; p=0·021). The other subgroups did not inhibitor rivastigmine (n=65) or placebo (n=65).3 experience a similar between-group reduction of The primary outcome was a surrogate marker of fall incidence. The following should be noted: fall risk (reduction in step time variability). After 32 • As a subgroup analysis, this finding should be treated weeks, compared to the placebo group, patients on with caution. However, the analysis was pre- rivastigmine had improved on normal walking (ratio planned, it was one of a small number of subgroup of geometric means 0.72, 95% CI 0.58 to 0.88) and analyses, and the magnitude of effect was large. on a simple dual task (0.79, 95% CI 0.62 to 0.99). On a complex dual task, improvements did not differ • The study was not powered to detect differences significantly (0.81, 95% CI 0.60 to 1.09). Sensitivity in effect between study groups. As the authors analyses to account for missing data resulted in an point out, we can therefore only hypothesise upper limit of the 95% CI for simple walking of 0.97 that the benefit was greatest among those with (just below the null value of 1.0), and the difference Parkinson’s, and such a hypothesis will need to be for the simple dual task failed to reach significance. tested in further studies. No correction was made for multiple testing, rendering the findings for the simple dual task susceptible to Type I error. More participants in the • Falls were self-reported. The study was therefore rivastigmine group correctly guessed their allocation susceptible to recall bias among participants, than would be expected by chance, signaling possible aware that they were allocated to the ‘training failure of blinding, and potential ascertainment bias. plus VR’ arm of the study rather than treadmill Participants in the rivastigmine group suffered more training alone. gastrointestinal side effects (31% versus 5% for nausea, and 17% versus 5% for vomiting). Among • Overall, this study provides moderate strength secondary outcomes, patients in the rivastigmine of evidence that treadmill training plus VR is group suffered fewer falls (adjusted IRR 0.55; 95% CI effective in preventing falls among people 0.38 to 0.81), but there was no significant difference with Parkinson’s. for disease severity or quality of life. Conclusion 8. Morris ME, Menz HB, McGinley JL, et al. A Randomized Controlled Trial to Reduce Falls in People In summary, the best evidence gleaned by this CAT With Parkinson’s Disease. Neurorehabilitation & Neural was for (i) progressive resistance strength training in Repair 2015;29(8):777-85. combination with fall prevention education, (ii) Tai Chi, and (iii) virtual-reality-enhanced treadmill training. 9. Iwamoto J, Takeda T, Matsumoto H. Sunlight Common methodological problems among studies exposure is important for preventing hip fractures were the use of surrogate outcomes, exclusion of in patients with Alzheimer’s disease, Parkinson’s Parkinson’s patients with cognitive impairment, and disease, or . Acta Neurologica Scandinavica lack of power to demonstrate a real effect (Type II 2012;125(4):279-84. error). There is a continued need for well-conducted 10. Sato Y, Iwamoto J, Honda Y. Amelioration of RCTs, and for meta-analyses that can detect possible and hypovitaminosis D by sunlight real effects from existing underpowered studies.

Critically Appraised Topics (CATs) Topics Appraised Critically exposure in Parkinson’s disease. Parkinsonism and References Related Disorders 2011;17(1):22-26. 1. Canning CG, Paul SS, Nieuwboer A. Prevention of falls 11. Lee MS, Lam P, Ernst E. Effectiveness of tai chi for in Parkinson’s disease: a review of fall risk factors and Parkinson’s disease: a critical review. Parkinsonism & the role of physical interventions. Neurodegenerative Related Disorders 2008;14(8):589-94. Disease Management 2014;4(3):203-21. 12. Li F, Harmer P, Fitzgerald K, et al. Tai chi and postural 2. Wood BH, Bilclough JA, Bowron A, et al. Incidence and stability in patients with Parkinson’s disease. New prediction of falls in Parkinson’s disease: a prospective England Journal of Medicine 2012;366(6):511-19. multidisciplinary study. Journal of Neurology, 13. Scianni A. Tai Chi improves balance and prevents Neurosurgery & Psychiatry 2002;72(6):721-5. falls in people with Parkinson’s disease. J Physiother 3. Henderson EJ, Lord SR, Brodie MA, et al. Rivastigmine 2015;61(1):44. for gait stability in patients with Parkinson’s disease 14. Nieuwboer A, Kwakkel G, Rochester L, et al. Cueing (ReSPonD): a randomised, double-blind, placebo- training in the home improves gait-related mobility controlled, phase 2 trial. The Lancet Neurology in Parkinson’s disease: the RESCUE trial. J Neurol 2016;15(3):249-58. Neurosurg Psychiatry 2007;78(2):134-40. 4. Allen NE, Sherrington C, Paul SS, et al. Balance and 15. Ashburn A, Fazakarley L, Ballinger C, et al. A falls in Parkinson’s disease: a meta-analysis of the randomised controlled trial of a home based effect of exercise and motor training. Movement exercise programme to reduce the risk of falling Disorders 2011;26(9):1605-15. among people with Parkinson’s disease. Journal 5. Canning CG, Sherrington C, Lord SR, et al. Exercise of Neurology, Neurosurgery & Psychiatry for falls prevention in Parkinson disease: a randomized 2007;78(7):678-84. controlled trial. Neurology 2015;84(3):304-12. 16. Marjama-Lyons J, Smith L, Mylar B, et al. Tai Chi 6. Goodwin VA, Richards SH, Henley W, et al. An exercise and reduced rate of falling in Parkinson’s disease: intervention to prevent falls in people with Parkinson’s A single-blinded pilot study. Movement Disorders; disease: a pragmatic randomised controlled trial. 2002. WILEY-LISS DIV JOHN WILEY & SONS INC, Journal of Neurology, Neurosurgery & Psychiatry 605 THIRD AVE, NEW YORK, NY 10158-0012 USA. 2011;82(11):1232-8. 17. Hass CJ, Waddell DE, Wolf SL, et al. The influence 7. Wong-Yu IS, Mak MK. Task- and Context-Specific of tai chi training on locomotor ability in parkinson’s Balance Training Program Enhances Dynamic disease. Annual Meeting of American Society of Balance and Functional Performance in Parkinsonian Biomechanics; 2006. Nonfallers: A Randomized Controlled Trial With Six- 18. Purchas M, MacMahon D. The effects of Tai Chi Month Follow-Up. Archives of Physical Medicine & training on general wellbeing and motor performance Rehabilitation 2015;96(12):2103-11. in patients with Parkinson’s disease (PD): A pilot study. Movement Disorders; 2007. WILEY-LISS DIV JOHN WILEY & SONS INC, 111 RIVER ST, HOBOKEN, NJ 07030 USA. 19. Cheon S, Sung H, Ha M, et al. Tai Chi for physical and mental health of patients with Parkinson’s disease. Proceeding of the first international conference of tai chi for health Seoul, Korea; 2006. 20. Li F, Fitzgerald K. Postural stability in Parkinson’s disease patients after Tai Chi training: A randomized controlled trial. Parkinsonism and Related Disorders 2012;18:S155. 21. van der Marck MA, Klok MP, Okun MS, et al. Consensus-based clinical practice recommendations for the examination and management of falls in Critically Appraised Topics (CATs) Topics Appraised Critically patients with Parkinson’s disease. Parkinsonism and Related Disorders 2014;20(4):360-69. 22. Gao Q, Leung A, Yang Y, et al. Effects of Tai Chi on balance and fall prevention in Parkinson’s disease: a randomized controlled trial. Clinical rehabilitation 2014;28(8):748-53. 23. Mirelman A, Rochester L, Maidan I, et al. Addition of a non-immersive virtual reality component to treadmill training to reduce fall risk in older adults (V-TIME): a randomised controlled trial. Lancet. 2016;388(10050):1170-1182.

This Critically Appraised Topic, the others in the series, and the related clinical summary are the work of the Evidence-Based Practice Theme Working Group: • Lynn Rochester (Chair) • Kevin Galbraith • Donald Grosset • Mike Samuel • Katherine Deane • Debbie Davies • Julia Johnson • Carl Clarke • Janine Barnes • Alison Yarnall

The UK Parkinson’s Excellence Network is the driving force for improving Parkinson’s care, connecting and equipping professionals to provide the services people affected by the condition want to see. The tools, education and data it provides are crucial for better services and professional development. The network links key professionals and people affected by Parkinson’s, bringing new opportunities to learn from each other and work together for change.

Visit parkinsons.org.uk/excellencenetwork

Parkinson’s UK is the operating name of the Parkinson’s Disease Society of the United Kingdom. A charity registered in England and Wales (258197) and in Scotland (SC037554). © Parkinson’s UK 6/2017 (CS2734)