Accepted Manuscript Effect of Pilates exercise for improving balance in older adults: A systematic review with meta-analysis Anna Lucia Barker , PhD, Marie-Louise Bird , PhD, Jason Talevski , BHealth Sci (Public Health) PII:
S0003-9993(14)01285-4
DOI:
10.1016/j.apmr.2014.11.021
Reference:
YAPMR 56054
To appear in:
ARCHIVES OF PHYSICAL MEDICINE AND REHABILITATION
Received Date: 11 August 2014 Revised Date:
25 November 2014
Accepted Date: 26 November 2014
Please cite this article as: Barker AL, Bird M-L, Talevski J, Effect of Pilates exercise for improving balance in older adults: A systematic review with meta-analysis, ARCHIVES OF PHYSICAL MEDICINE AND REHABILITATION (2015), doi: 10.1016/j.apmr.2014.11.021. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Pilates to improve balance in older adults
ACCEPTED MANUSCRIPT
Effect of Pilates exercise for improving balance in older adults: A systematic review with meta-analysis
RI PT
Anna Lucia Barker1, PhD, Marie-Louise Bird2, PhD, Jason Talevski1, BHealth Sci (Public Health)
1
Health Services Research Unit, Center of Research Excellence in Patient Safety, Division of
SC
Health Services and Global Health Research, Department of Epidemiology and Preventive
Medicine, School of Public Health and Preventive Medicine, Monash University, Melbourne,
M AN U
VIC, Australia 2
Faculty of Health, School of Health Sciences, University of Tasmania, Launceston, TAS,
TE D
Australia
Corresponding author: Dr Anna Barker
Address: Monash University, Department of Epidemiology and Preventive Medicine, The
EP
Alfred Centre, Level 6, 99 Commercial Road, Melbourne VIC 3004, AUSTRALIA Phone (BH): +61 3 9903 0946
Phone (AH): +61 408 033 287
AC C
Emial:
[email protected] ACKNOWLEDGEMENTS We thank Meghan Ambrens and research assistants in the Falls and Bone Health team at Monash University for their assistance with activities associated with this review, including the reviewing of the literature in each database.
1
ACCEPTED MANUSCRIPT This manuscript has not been submitted or is not simultaneously being submitted elsewhere, and no portion of the data has been or will be published in proceedings or transactions of meetings or symposium volumes.
RI PT
No commercial party having a direct interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are
SC
associated.
No financial support or other benefits from commercial sources for the work reported on in the
M AN U
manuscript have been received which could create a potential conflict of interest or the appearance of a conflict of interest with regard to the work.
Reprint requests to Dr Anna Barker, Monash University, Department of Epidemiology and
TE D
Preventive Medicine, The Alfred Centre, Level 6, 99 Commercial Road, Melbourne VIC 3004,
AC C
EP
AUSTRALIA, e-mail:
[email protected] 2
ACCEPTED MANUSCRIPT 1
ABSTRACT
2 3
Objective: To investigate the effect of Pilates on balance and falls in older adults; and whether
5
programs tested in prior studies met best-practice recommendations for exercise to prevent falls.
6
Data Sources: MEDLINE, SPORTDiscus, CINAHL, PubMed, The Physiotherapy Evidence
7
Database and The Cochrane Library were searched from earliest record to July 2014.
8
Study Selection: Randomized and controlled clinical trials evaluating the effect of Pilates on
9
balance and falls in older adults.
SC
RI PT
4
Data Extraction: Two reviewers independently extracted demographic, intervention and
11
outcome data. Six studies were included in this review.
12
Data Synthesis: There is a lack of high-quality studies in this area. When compared to non-
13
active control groups, Pilates was shown to improve balance (SMD 0.84, 95% CI 0.44 to 1.23; 6
14
studies) and reduce the number of falls (SMD -2.03, 95% CI -2.66 to -1.40; 1 study). Three
15
studies provided sufficient detail to enable assessment of compliance with the recommendation
16
of exercises providing a moderate or high challenge to balance. In these studies, 2-36% of
17
exercises were assessed as providing a moderate or high challenge to balance. All studies
18
provided ≥2 hours of exercise per week and one provided >50 hours of exercise during the
19
study period.
20
Conclusion: The evidence suggests Pilates can improve balance, an important risk factor for
21
falls in older adults. However, there is limited data on the impact on falls. Effects may have been
22
over-estimated due to the low methodological quality of studies. Best-practice recommendations
23
were rarely applied in prior studies indicating greater effects may have been achieved if
24
recommendations were incorporated.
25
Key
AC C
EP
TE D
M AN U
10
Words:
Accidental
falls;
Aged;
Pilates-Based
Exercises;
Postural
Balance
1
ACCEPTED MANUSCRIPT 26
ABBREVIATIONS
27
CI – Confidence interval
30
RCT – Randomized controlled trial
31
SMD – Standardized mean difference
AC C
EP
TE D
M AN U
SC
29
RI PT
28
2
ACCEPTED MANUSCRIPT Falls in older adults continue to be a major source of disability, mortality and health care
33
utilization1. In the United States, one older adult dies following a fall every 29 minutes2. While
34
falls do occur in young3 and middle aged people4, their sequelae are most deleterious in older
35
adults. Thirty percent of falls in older adults have been reported to result in severe injuries,
36
substantially impacting on independence and risk of early death5. Balance impairment is a key
37
modifiable risk factor for falls6, 7. Therefore, exercises that aim to improve balance are a key
38
component of fall prevention programs in both clinical practice and the research literature8.
RI PT
32
SC
39
Pilates is a ‘mind-body exercise’ that has been used since early in the 20th century. It focuses on
41
improving strength, core stability, flexibility, muscle control, posture and breathing9. Pilates aims
42
to improve co-ordination and control of the core muscles of the trunk, which contribute to the
43
optimal lumbar-pelvic stabilization needed for daily activities and function10. Pilates exercises can
44
be categorized into two types: (1) mat exercises (which have commonalities with yoga and Tai
45
Chi) and; (2) equipment exercises (that use springs to achieve muscle strengthening). Many
46
Pilates exercises are performed in a standing position, with a narrow base of support without
47
hand support, challenging trunk muscle stability to maintain upright postures. An example of this
48
is standing leg pumps on the Wunda chair, or scooter on the reformer (Figure 1).
TE D
EP
49
M AN U
40
A recent systematic review11 of 16 studies concluded that Pilates exercise can improve dynamic
51
balance in healthy populations. However, this conclusion was based on findings from only two
52
of the 16 included studies. Additionally, only one of the two studies that reported on dynamic
53
balance was completed in older adults (≥65 years), therefore limiting the generalizability of this
54
finding to older adults. The review also reported a significant decrease in number of falls;
55
however this result was also concluded from only one study11.
AC C
50
56
3
ACCEPTED MANUSCRIPT In addition to this review of the effectiveness of Pilates exercise, an additional systematic review
58
including 54 randomized controlled trials (RCTs) on all forms of exercise designed to prevent
59
falls has been completed12. This review reported that exercise programs including balance
60
training achieved the greatest reductions in falls. As many Pilates exercises provide a challenge to
61
balance, Pilates may achieve similar effects on falls as reported by this review. Following this
62
initial review, an updated review was published13 and provided best-practice recommendations to
63
guide the use of exercise for falls prevention (Box 1). The authors recommended that exercises
64
for falls prevention should: provide a moderate or high challenge to balance; be undertaken for
65
at least 2 hours per week; and provide a total dose of more than 50 hours of exercise over the
66
study period. Programs with these characteristics achieved a 38% reduction in the rate of falls
67
(95% CI: 27 to 46) 13.
M AN U
SC
RI PT
57
68 69
Box 1: Best-practice recommendations for exercises for falls prevention13
TE D
1. Exercise must provide a moderate or high challenge to balance 2. Exercise must be of a sufficient dose to have an effect (≥2 hours per week and >50 hours over total study period)
EP
3. On-going exercise is necessary
4. Exercise should be targeted at the general community as well as those at high risk of
AC C
falling
5. Exercise may be undertaken in a group or home-based setting. 6. Walking training may be included in addition to balance training but high risk individuals should not be prescribed brisk walking programs
7. Strength training may be included in addition to, but not in place of balance training 8. Exercise providers should make referrals for other falls risk factors to be addressed 70
4
ACCEPTED MANUSCRIPT The primary aim of this systematic review was to evaluate the evidence for the effect of Pilates
72
on balance and falls in older adults. A secondary aim was to identify if Pilates programs tested in
73
prior studies met the best-practice recommendations for exercise to prevent falls13 in terms of
74
program content and dose of exercise.
AC C
EP
TE D
M AN U
SC
RI PT
71
5
ACCEPTED MANUSCRIPT 75
METHODS
76
This systematic review was performed according to the criteria of the Preferred Reporting Items
77
for Systematic Reviews and Meta-analyses14.
78
Search strategy
80
A systematic search of literature was conducted in July 2014. MEDLINE, SPORTDiscus and
81
CINAHL were searched to find published research. Supplementary searches were also
82
conducted in PubMed, the Physiotherapy Evidence Database (PEDro) and The Cochrane
83
Library. Comparative research studies published up to July 2014 were selected for evaluation. A
84
sensitive search strategy was developed using the terms ‘Pilates’ AND ‘Balance’ OR ‘Accidental
85
Falls’ OR ‘Postural Stability’, and was translated for each database as appropriate. The search
86
strategy was validated by identifying a ‘validation set’ of references based on studies obtained
87
from the MEDLINE search and ensuring that these were also included in searches in each of the
88
other databases.
TE D
89
M AN U
SC
RI PT
79
Inclusion criteria
91
Studies were included in this review if they were published in a peer-reviewed journal; written in
92
the English language; conducted as a RCT or controlled clinical trial in older adults (mean age of
93
participants ≥60 years); tested an exercise intervention described as ‘Pilates exercise’ (mat,
94
equipment or both); included a comparison group that participated in no exercise (including
95
non-active activities such as education); and reported on at least one performance-based measure
96
of balance (dynamic or static) or falls. Studies that recruited people living in the community or
97
residential aged care, including nursing homes and long term care, were included.
AC C
EP
90
98 99
Study selection
6
ACCEPTED MANUSCRIPT Two reviewers independently screened and excluded studies based on title and abstract. If it was
101
clear that the intervention tested was not Pilates, or that measures of balance or falls were not
102
reported, the study was excluded. Duplicate publications, where there was overlap of participants
103
across publications, were also excluded. For studies not excluded by this process, full text was
104
obtained and assessed independently by both reviewers against the inclusion criteria. Differences
105
of opinion regarding eligibility of studies were resolved by a third reviewer.
RI PT
100
106
Data extraction
108
Two reviewers independently extracted data from the included studies. Information on
109
population (mean age, sample size), setting (community or residential aged care), study design,
110
outcome measures (balance and/or falls) and intervention characteristics were recorded. The
111
intervention was described in terms of who delivered the exercise (Pilates instructor or Physical
112
therapist); the type of Pilates exercises included (mat, equipment or apparatus such as resistance
113
band and balls); mode (group or individual), frequency (hours per week) and duration of delivery
114
(length of intervention in weeks); total dosage of exercise (in hours) and adherence to exercise
115
sessions. This information was recorded based on information included in the publication. If this
116
information was not included in the publication, the study authors were contacted to provide
117
additional detail.
M AN U
TE D
EP
AC C
118
SC
107
119
Quality assessment and risk of bias
120
Two review authors independently assessed the methodological quality and risk of bias of
121
included studies. Uncertainty or disagreements were resolved by discussion with a third reviewer.
122
Study quality was assessed according to six criteria described in the Cochrane Handbook for
123
Systematic Reviews of Interventions14: (1) blinding of provider or patient; (2) concealment of
124
allocation; (3) blinding of outcome assessment; (4) co-intervention avoided, controlled or used
125
similarly across comparison groups; (5) loss to follow-up less than 20% and equally distributed 7
ACCEPTED MANUSCRIPT 126
between comparison groups; and (6) intention to treat analysis performed. Criteria were assessed
127
as ‘met’, ‘unclear’ or ‘not met’.
128
Risk of bias was assessed as ‘high’, ‘low’ or ‘unclear’ based on the domains included in the
130
Cochrane Collaboration’s tool for assessing risk of bias in clinical trials15. The tool covers six
131
domains of bias: selection bias, performance bias, detection bias, attrition bias, reporting bias and
132
other bias. If insufficient detail was reported in the trial, the judgment resulted in an ‘unclear’
133
risk of bias.
SC
RI PT
129
134
Data synthesis
136
Meta-analysis
137
A meta-analysis was conducted to assess intervention effects on balance and falls, using
138
unadjusted, pooled data and described as standardized mean differences (SMD) and 95%
139
confidence intervals (CIs). Using pooled data is effective for comparing data collected using
140
different outcome measures and can provide an improved estimate of the effect of the
141
intervention as a consequence of the larger number of total participants and reduction in random
142
error due to sampling differences14. A SMD of less than 0.2 was considered a small effect,
143
between 0.2 and 0.8 a moderate effect and greater than 0.8 a large effect16. A random effects
144
model was used as it was assumed that the true effect could vary between studies due to the
145
heterogeneity of study populations, intervention characteristics and measurement of outcomes.
146
For example, the effect size might be higher if participants were older; if the study used a more
147
intensive Pilates program or longer duration of the intervention; or if balance outcomes were
148
measured using a more sensitive assessment tool. Heterogeneity between studies was assessed
149
using the I2 statistic which shows the proportion of total variance that is explained by
150
heterogeneity17. Statistical heterogeneity was considered substantial if I2 was greater than 50%14.
151
Scale directions were aligned by adding negative values where required.
AC C
EP
TE D
M AN U
135
8
ACCEPTED MANUSCRIPT 152
Balance outcomes were categorized as static, dynamic or both static and dynamic. Dynamic
154
balance was defined as the ability to maintain equilibrium while moving through space either
155
with or without moving the feet (e.g. reaching forward while keeping the feet in full contact with
156
the floor as in the functional reach test; or standing up from a chair, walking 3 meters, turning
157
and returning to the chair to sit down as in the timed up and go test). Static balance was defined
158
as the ability to hold an upright position while keeping the feet in full contact with the floor (e.g.
159
maintaining an upright stance while the feet are together). To be included in the meta-analysis,
160
studies must have reported outcome measures known to be responsive for measuring change in
161
balance. The outcome measure from each study was chosen prior to analysis according to the
162
following order of priority: timed up and go test19/ 8-foot up and go test, Tinetti Performance
163
Oriented Mobility Assessment (POMA) test18, five timed sit to stand test20, four square step
164
test21, functional reach test22 and static balance tests (measures of hold times, sway, stability,
165
pressure in static stance positions). This list was determined by tabulating the frequency which
166
each test was reported in the included studies, and the more frequently reported tests were
167
prioritised for inclusion in the meta-analysis over the less frequently reported tests. This was
168
done to decrease heterogeneity of outcome measures used in the meta-analysis. Where a study
169
reported results for more than one of these outcomes, only the outcome of the highest priority
170
was used to minimize bias in the results, and dynamic measures were prioritised over static where
171
both were reported. Outcome data were excluded from the meta-analysis if there were significant
172
differences in baseline scores of the outcome of interest between the two groups to ensure
173
results were not confounded. A test of sub-group differences was performed to ensure it was
174
appropriate to combine dynamic and static measures in one meta-analysis. The meta-analysis was
175
performed using Review Manager (RevMan 5.3) software.
AC C
EP
TE D
M AN U
SC
RI PT
153
176 177
Compliance with best-practice recommendations for exercises for falls prevention 9
ACCEPTED MANUSCRIPT This review assessed whether the Pilates exercises programs tested in each study met the
179
following best-practice recommendations for exercise to prevent falls13: (1) exercises provided a
180
moderate or high challenge to balance (i.e. performed in standing and using a small base of
181
support while moving the center of mass over that base); and (2) programs were of sufficient
182
dose (≥2 hours per week and >50 total hours over the study period). These assessments were
183
performed independently by two physiotherapists experienced in the delivery of both Pilates
184
exercise, and exercises for the prevention of falls.
AC C
EP
TE D
M AN U
SC
RI PT
178
10
ACCEPTED MANUSCRIPT 185
RESULTS
186 187
Study selection
189
The 13 references from the validation set were identified in the search results from each
190
database. The electronic search identified 309 potential studies for screening of eligibility. Of
191
these, 285 were excluded after review of title and abstract as not meeting the inclusion criteria
192
and full text was obtained for the remaining 24 studies. Based on the reviewer’s decisions, six
193
studies met the inclusion criteria and were included in the review23-28 (Figure 2).
SC
RI PT
188
M AN U
194 195
Study and intervention characteristics
196
Of the six studies included, five were conducted as an RCT23-26, 28 and one as a controlled clinical
197
trial27. Four studies reported on dynamic balance outcomes23,
198
outcomes23, 25, and one reported on a combined static and balance outcome28. Only one study
199
reported data on falls25. The sample sizes in all included studies were small, ranging from 30 to
200
60 participants, and the majority of participants were female. One study delivered the Pilates
201
intervention in a residential aged care facility25, and the others in community-dwelling adults
202
(Table 1). The programs included various combinations of mat, equipment and other apparatus
203
exercises. Two studies included only mat exercises24, 26, with others including mat exercises as
204
well as resistance training (bands or weights), small apparatus or equipment in combination. The
205
duration of Pilates interventions ranged from 5 to 24 weeks and sessions were all delivered in a
206
group setting. Only two studies recorded exercise adherence, and both of these had participation
207
rates of greater than 80%23, 26 (Table 2).
24, 26, 27
AC C
EP
TE D
, two on static balance
208 209
Quality assessment and risk of bias
11
ACCEPTED MANUSCRIPT The methodological quality of the included studies was low. Common flaws included lack of
211
reporting of allocation concealment; blinding of participants, persons delivering the intervention
212
and those assessing outcomes; inclusion of a ‘group delivery’ co-intervention for intervention but
213
not control participants; and no reporting of intention-to-treat analysis. Studies were assessed as
214
having a high degree of bias with common sources of bias being selection bias, performance
215
bias, detection bias and attrition bias (Table 3).
RI PT
210
216
Data Synthesis
218
Meta-analysis
219
There was significant heterogeneity detected for the six studies included in the meta-analysis of
220
balance outcomes (I2=61%). When a random-effects analysis was applied, a large significant
221
effect was observed for Pilates exercise on balance (SMD 0.84, 95% CI 0.44 to 1.23). Effects
222
were consistent across both dynamic (SMD 0.62, 95% CI 0.15 to1.08), static (SMD 1.21, 95% CI
223
0.66 to 1.77) and combined static and dynamic (SMD 1.25, 95% CI 0.66 to 1.85) balance
224
outcomes (test for sub-group differences, p=0.15) (Figure 3).
225
TE D
M AN U
SC
217
Falls
227
Only one study included in this review reported data on falls25. This RCT observed a significant
228
reduction in the number of falls in the Pilates intervention participants (pre-intervention=1.87;
229
post=0.37) compared to the control group participants (pre-intervention=1.63; post=1.30)
230
during the 12-week follow-up period (SMD -2.03, 95% CI -2.66 to -1.40) (Figure 4). Data on
231
falls was collected during the 12-week study period using monthly calendars.
AC C
EP
226
232 233
Compliance with best-practice recommendations for exercises for falls prevention
234
All six studies provided ≥2 hours of exercise per week over the study period. However, only one
235
study26 provided >50 hours of exercise during the study period. Three studies23, 24, 28 provided 12
ACCEPTED MANUSCRIPT sufficient detail on the Pilates exercises to enable assessment of compliance with the best-
237
practice recommendation of providing a moderate or high challenge to balance. Of these three
238
studies, the percentage of exercises that were assessed as providing a moderate to high challenge
239
to balance were 2%24, 20%28 and 36%23 (Table 2). Examples of exercises assessed as providing a
240
moderate or high level challenge to balance included: scooter and standing leg side series on the
241
reformer; assisted squats and standing arm series on the trapeze; arm arcs and leg side series in
242
standing on the mat; and hamstring series I on the wunda chair.
AC C
EP
TE D
M AN U
SC
RI PT
236
13
ACCEPTED MANUSCRIPT 243
DISCUSSION
244 245
The focus of this review was to examine the effect of Pilates on balance and falls in older adults.
247
Until now, no pooled estimates have been available on the effect of Pilates exercise on these
248
outcomes. The meta-analysis identified that Pilates can have a large significant positive effect on
249
balance in older adults. This finding appeared to be consistent across the majority of the studies
250
included. The number of falls was also shown to be lower after a Pilates intervention, however
251
this was concluded from only one RCT25. The best-practice recommendations for exercise to
252
prevent falls13 were rarely applied in prior studies in terms of program content and dose of
253
exercise, although three of the included studies24,
254
recommendations being published. Based on these findings, Pilates exercise can be considered
255
an effective form of exercise to improve balance in older adults which is known to decrease the
256
risk of falling. Greater effects may be achieved if best-practice recommendations for exercise to
257
prevent falls are incorporated.
SC
M AN U 26, 27
were completed prior to these
TE D
258
RI PT
246
There are several exercises in the Pilates repertoire that can provide a moderate or high challenge
260
to balance as recommended by the best-practice guideline recommendations for exercise to
261
prevent falls13. Balance-challenging exercises include those that: are multi-sensory (e.g. performed
262
on different surfaces such as mats, wobble boards and discs, or moving surfaces such as a
263
Wunda chair foot pedal or reformer carriage); are performed in a standing position; utilize a foot
264
position that achieves a narrow base of support (e.g. standing on one leg such as in a scooter or
265
standing leg pump exercise, Figure 1); and do not involve the use of hands for support. Many of
266
the exercises tested in studies included in this review were performed in a seated or lying
267
position, and therefore are likely to have only provided a minimal challenge to balance. The
268
impact of improving trunk muscle strength for postural stability is not known, but may be one
AC C
EP
259
14
ACCEPTED MANUSCRIPT mechanism by which positive effects on balance were achieved by studies with exercises
270
performed predominately in seated or lying positions. While the programs tested in studies
271
included in this review achieved positive improvements in balance, greater effects may have been
272
achieved if programs included more exercises in a standing position with a narrow base of
273
support or unstable surface, and were given at a higher dosage.
274
RI PT
269
The large positive effects on balance reported in the meta-analysis were consistent across five of
276
the six included studies. The one study where a positive effect was not reported was the
277
randomized cross-over study by Bird et al23. The authors of this study reported that the wash-out
278
period of 6 weeks used in the cross-over design was unlikely to have been long enough to allow
279
balance to return to pre-intervention levels. People who received the intervention may have
280
gained balance confidence that meant in the subsequent wash-out period they may have been
281
more active and continued to challenge their balance in everyday activities. This could have
282
maintained benefits achieved23. No observable differences on balance were identified between
283
studies that included more balance challenging exercises, or that applied a higher dosage of
284
exercise. It is likely that there were too few studies to enable accurate assessment of the impact
285
of these program characteristics on balance, and due to the high risk of bias in several of the
286
included studies, differences observed may be attributable to differences in study quality rather
287
than program characteristics.
M AN U
TE D
EP
AC C
288
SC
275
289
Results from this review should be interpreted with a degree of caution. The methodological
290
quality of the studies included was low, and a high risk of bias was present across the majority of
291
studies. High quality randomized controlled trials, with concealed allocation, blinded outcome
292
assessment, that control for co-interventions such as a group delivery mode and intention-to-
293
analysis would advance the quality of work in this field. Importantly, the study with the lowest
294
risk of bias25 reported a large positive effect on balance, and also a reduction in the number of 15
ACCEPTED MANUSCRIPT falls, strengthening the conclusion that Pilates has an effect on both of these outcomes. In all of
296
the included studies, the Pilates interventions was delivered in a group session while control
297
participants were asked to continue with their ‘usual activity’, rather than attending any placebo
298
group activity. As such, there is a risk that the group delivery mode of the Pilates classes formed
299
a co-intervention and may have influenced the results. To ensure that the co-intervention of the
300
delivery mode is controlled, future studies should provide a placebo intervention for control
301
participants by way of group delivery to match the delivery mode of the Pilates intervention.
RI PT
295
SC
302
The inconsistency in balance outcome measures used in prior studies limited the ability to
304
meaningfully compare results from different studies. The different outcome measures chosen
305
measured different facets of balance control. The Systems Framework for Postural Control
306
describes six major components of postural control: i) constraints on the biomechanical system
307
(degrees of freedom, strength, limits of stability); ii) movement strategies (reactive, anticipatory,
308
voluntary); iii) sensory strategies (integration, reweighting); iv) orientation in space (perception of
309
gravity, verticality); v) dynamic control (gait, proactive); and vi) cognitive processing (attention,
310
learning)29. The meta-analysis included four different balance outcome measures: (1) the timed-
311
up-and go (8-foot up and go); (2) five-times-sit-to-stand test; (3) POMA developed by Tinetti;
312
and (4) the rank value stability index (where a person maintains upright standing position on an
313
unstable surface similar to the foam condition of the clinical test of sensory interaction in
314
balance). These outcome measures evaluate different components of balance included in the
315
Systems Framework for Postural Control as summarized in a recent review30. All four of the
316
outcome measures included in the meta-analysis evaluated strength and sensory integration, and
317
three measures (five-times-sit-to-stand, timed-up-and go, POMA) evaluated anticipatory control
318
and dynamic stability. This suggests that while there was variability in the outcome measures
319
included in the meta-analysis, there was sufficient commonality in the underlying constructs of
320
balance being measured to warrant inclusion in a single meta-analysis. In addition, the POMA
AC C
EP
TE D
M AN U
303
16
ACCEPTED MANUSCRIPT 321
and the rank value stability index evaluated static stability, and only the POMA evaluated
322
functional stability limits and reactive postural control. Future studies should aim to include a
323
suite of balance outcome measures that ensures each of six components of balance are evaluated
324
to enable an increased understanding of which components of balance Pilates has an effect on.
RI PT
325
The majority of the studies included in this review did not report data on falls. Although one
327
study reported a greater reduction in the number of falls in the intervention compared to the
328
control group25, the authors did not provide sufficient detail on the time period for which fall
329
data was collected in the pre-intervention period, making meaningful comparison with the post
330
intervention fall data problematic. Additionally, 12 weeks is a short follow-up time to see a
331
change in falls and should be interpreted with caution. Therefore, we cannot confidently
332
conclude that Pilates exercise has an effect on falls. Future research focusing of the impact of
333
Pilates on falls in older people should be completed.
M AN U
SC
326
TE D
334
The level of adherence to the intervention was not well reported in the included studies, with
336
only two studies including this information23, 26. Adherence is a critical measure of effectiveness,
337
as it represents acceptability of an intervention. Even if an intervention is effective, if it is not
338
accepted by the target population it is of little benefit. Because of this, it is unknown if the
339
effectiveness of the Pilates programs tested was limited by poor adherence in programs.
AC C
340
EP
335
341
Study Limitations
342
Our study has several limitations; most notably, the available evidence is limited in quality and
343
quantity. There is a risk that effects may have been over-estimated due to methodologic
344
weaknesses of the studies included. The content of the exercise programs was highly variable
345
across studies, and many contained exercises that would not be considered specifically Pilates in
346
origin (e.g. squats with weights). The outcome measures used were also variable. None of the 17
ACCEPTED MANUSCRIPT balance measures included in the meta-analysis evaluated the verticality or cognitive influence
348
components of balance. Furthermore, functional stability limits and reactive postural control
349
components were only evaluated in one study. As such, this review only considers a subset of
350
components that constitute balance. Only studies published in English were included. Inclusion
351
of studies published in all languages may have given a different result.
AC C
EP
TE D
M AN U
SC
RI PT
347
18
ACCEPTED MANUSCRIPT 352
CONCLUSIONS
353 354
Current literature suggests that Pilates is effective for improving balance in older adults, which
356
may reduce their risk of falling. In both research and practice, Pilates programs should apply the
357
best-practice recommendations for exercise to prevent falls to ensure maximum benefit. Further
358
high-quality studies are needed that include more balance outcomes, and fall and adherence
359
outcomes measured over longer time periods.
AC C
EP
TE D
M AN U
SC
RI PT
355
19
ACCEPTED MANUSCRIPT 360
REFERENCES
361 362
1.
Sartini M, Cristina ML, Spagnolo AM, Cremonesi P, Costaguta C, Monacelli F et al. The
364
epidemiology of domestic injurious falls in a community dwelling elderly population: an
365
outgrowing economic burden. Eur J Public Health 2010;20(5):604-6.
366
2.
367
2012; http://webappa.cdc.gov/sasweb/ncipc/nfilead2001.html.
368
3.
369
community dwelling adults: perceived cause, environmental factors and injury. BMC Public
370
Health 2005;5.
371
4.
372
prevention intervention before 60. Climacteric: the journal of the International Menopause
373
Society 2008;11(6):461-6.
374
5.
375
disproportionate to mechanism. J Trauma 2001;50(1):116-9.
376
6.
377
the Elderly: A Prospective Study of Risk Factors and Risk Profiles. American Journal of
378
Epidemiology 1996;143(11):1129-36.
379
7.
380
factors for recurrent falling in older persons. Journal of clinical epidemiology 2003;56(7):659-68.
381
8.
382
Interventions for preventing falls in older people living in the community. Cochrane Database of
383
Systematic Reviews 2012;9:CD007146.
384
9.
385
Complement Ther Med 2012;20(4):253-62.
RI PT
363
SC
Centers for Disease Control and Prevention. Falls among older adults: an overview.
M AN U
Talbot LA, Musiol RJ, Witham EK, Metter EJ. Falls in young, middle-aged and older
TE D
Nitz JC, Choy NL. Falling is not just for older women: support for pre-emptive
Sterling DA, O'Connor JA, Bonadies J. Geriatric falls: injury severity is high and
AC C
EP
Graafmans WC, Ooms ME, Hofstee HMA, Bezemer PD, Bouter LM, Lips P. Falls in
Stel VS, Smit JH, Pluijm SMF, Lips P. Balance and mobility performance as treatable risk
Gillespie LD, Robertson MC, Gillespie WJ, Sherrington C, Gates S, Clemson LM et al.
Wells C, Kolt GS, Bialocerkowski A. Defining Pilates exercise: a systematic review.
20
ACCEPTED MANUSCRIPT 386
10.
Queiroz BC, Cagliari MF, Amorim CF, Sacco IC. Muscle activation during four Pilates
387
core stability exercises in quadruped position. Archives of physical medicine and rehabilitation
388
2010;91(1):86-92.
389
11.
390
the effects of pilates method of exercise in healthy people. Archives of physical medicine and
391
rehabilitation 2011;92(12):2071-81.
392
12.
393
exercise for the prevention of falls: a systematic review and meta-analysis. J Am Geriatr Soc
394
2008;56(12):2234-43.
395
13.
396
older adults: an updated meta-analysis and best practice recommendations. N S W Public Health
397
Bull 2011;22(3-4):78-83.
398
14.
399
reviews of interventions. Chichester, England ; Hoboken, NJ: Wiley-Blackwell; 2008.
400
15.
401
Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ
402
2011;343:d5928.
403
16.
404
Erlbaum; 1988.
405
17.
406
analyses. BMJ 2003;327(7414):557-60.
407
18.
408
Journal of the American Geriatrics Society 1986; 34: 119-126.
409
19.
410
and Balance. Physical Therapy 1993; 73: 346-351.
RI PT
Cruz-Ferreira A, Fernandes J, Laranjo L, Bernardo LM, Silva A. A systematic review of
SC
Sherrington C, Whitney JC, Lord SR, Herbert RD, Cumming RG, Close JC. Effective
M AN U
Sherrington C, Tiedemann A, Fairhall N, Close JC, Lord SR. Exercise to prevent falls in
TE D
Higgins JPT, Green S, Cochrane Collaboration. Cochrane handbook for systematic
EP
Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD et al. The
AC C
Cohen J. Statistical Power Analysis in the Behavioral Sciences 2nd ed. Hillsdale, NJ:
Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-
Tinetti ME. Performance-Oriented Assessment of Mobility Problems in Elderly Patients.
Cohen H, Blatchly CA, Gombash LL. A Study of the Clinical-Test of Sensory Interaction
21
ACCEPTED MANUSCRIPT 411
20.
Buatois S, Miljkovic D, Manckoundia P, Gueguen R, Miget P, Vancon G, Perrin P,
412
Benetos A. Five times sit to stand test is a predictor of recurrent falls in healthy community-
413
living subjects aged 65 and older. Journal of the American Geriatrics Society 2008; 56: 1575-1577.
414
21.
415
multiple falling older adults. Archives of Physical Medicine and Rehabilitation 2002; 83: 1566-
416
1571.
417
22.
418
in a Sample of Elderly Male Veterans. Journals of Gerontology 1992; 47: M93-M98.
419
23.
420
balance in older adults after training with Pilates. Archives of physical medicine and
421
rehabilitation 2011.
422
24.
423
Evaluation of Pilates training on agility, functional mobility and cardiorespiratory fitness in
424
elderly women. African Journal for Physical, Health Education, Recreation and Dance
425
2013;19(2):505-12.
426
25.
427
exercise program for 65+year-old women to reduce falls. J Sport Sci Med 2011;10(1):105-11.
428
26.
429
aqua fitness training on older adults’ physical functioning and quality of life. Biomedical Human
430
Kinetics 2013;5:22-7.
431
27.
432
depression and balance associated with falling in the elderly. Procd Soc Behv 2013;70:1714-23.
433
28.
434
EH. Pilates method in personal autonomy, static balance and quality of life of elderly females.
435
Journal of Bodywork and Movement Therapies 2010;14(2):195-202.
RI PT
Dite W, Temple VA. A clinical test of stepping and change of direction to identify
SC
Duncan PW, Studenski S, Chandler J, Prescott B. Functional Reach - Predictive-Validity
M AN U
Bird M, Hill K, Fell J. A randomised controlled study investigating static and dynamic
TE D
Gildenhuys GM, Fourie, M., Shaw, I., Shaw, B.S., Toriola, A.L. and Witthuhn, J.
EP
Irez GB, Ozdemir RA, Evin R, Irez SG, Korkusuz F. Integrating Pilates exercise into an
AC C
Kovách MV, Plachy, J. K., Bognár, J., Balogh, Z. O., Barthalos, I. Effects of Pilates and
Mokhtari M, Nezakatalhossaini M, Esfarjani F. The effect of 12-week pilates exercises on
Siqueira Rodrigues BGd, Ali Cader S, Bento Torres NVO, Oliveira EMd, Martin Dantas
436
22
ACCEPTED MANUSCRIPT 437
29.
Horak FB. Postural orientation and equilibrium: what do we need to know about neural
438
control of balance to prevent falls? Age and ageing 2006;35 Suppl 2:ii7-ii11.
439
30.
440
Framework For Postural Control To Analyze The Components Of Balance Evaluated In
441
Standardized Balance Measures: A Scoping Review. Arch Phys Med Rehabil 2014.
AC C
EP
TE D
M AN U
SC
RI PT
Sibley KM, Beauchamp MK, Van Ooteghem K, Straus SE, Jaglal SB. Using The Systems
23
ACCEPTED MANUSCRIPT 442
FIGURE LEGENDS
443
Figure 1: Pilates equipment exercises for balance
446
Figure 2: Flow chart of exclusion process
447
Figure 3: Meta-analysis of balance outcomes
448
Figure 4: Meta-analysis of fall outcomes
AC C
EP
TE D
M AN U
SC
445
RI PT
444
24
ACCEPTED MANUSCRIPT
Table 1: Study and participant characteristics Female, %
Control intervention
Bird et al23
RCT
Community
32
78%
Usual activity
Gildenhuys et al24
RCT
Community
50
65.3 (5.0)
100%
Residential aged care
60
72.8 (6.7)
100%
RCT
Community
54
76%
Mokhtari27
CCT
Community
30
66.4 (6.2) NR
Usual activity (asked not to perform structured exercise) Usual activity (asked not to perform structured exercise) Usual activity
Irez et al25
RCT
Kovach et al26
SiqueiraRodrigues et al28
RCT
Community
52
66 (4.0)
Outcome measures Dynamic Static balance balance FSST Centre of TUG Test pressure/Medio -lateral sway FTSST N/A
RI PT
Age, mean (SD) 67.2 (6.6)
SC
Sample size
M AN U
Setting
TE D
Study design
EP
Study
100%
Usual activity
100%
Usual activity
N/A
8-Foot Up and Go Test FRT TUG Test Tinetti POMA test
Falls N/A
N/A
Rank Value Stability Index (MED-SP300) N/A
Monthly falls calendar N/A
N/A
N/A
Tinetti POMA test
N/A
AC C
RCT = Randomised controlled trial; CCT = Controlled clinical trial; SD = Standard deviation; FSST = Four square step test; FRT = Functional reach test; TUG = Timed up and go; FTSST = Five-times-sit-to-stand test; POMA = Performance Oriented Mobility Assessment; NR = Not reported; N/A = Not applicable.
ACCEPTED MANUSCRIPT
Table 2: Pilates exercise intervention characteristics
Irez et al25
Kovach et al26
week and
Challenge
≥2 hrs/
>50 hrs
no. of
balance (%)
week (hrs)
total (hrs)
Exercise adherence
Group
5 weeks
3 x 60 mins
36%
Y (2)
N (10)
>80%
Group
8 weeks
3 x 60 mins
2%
Y (3)
N (24)
NR
12 weeks
3 x 60 mins
UTA
Y (3)
N (36)
NR
24 weeks
3 x 60 mins
UTA
Y (3)
Y (72)
>80%
instructor instructor Pilates
Group
trained PT NR
Group
TE D
Pilates instructor Pilates
Group
12 weeks
3 x 60 mins
UTA
Y (3)
N (36)
NR
Group
8 weeks
2 x 60 mins
20%
Y (2)
N (16)
NR
trained PT
EP
Guideline recomendations13
sessions
Pilates
Sessions per
RI PT
mode
Chair
Trapeze
by
Duration
SC
Gildenhuys et al24
Rodrigues et al28
Delivery
M AN U
Siqueira-
Delivered
Pilates
Bird et al23
Mokhtari27
Reformer
Ball
Theraband
Pilates exercise type
Mat
Study
AC C
NR=Not reported; PT = Physical therapist; UTA = Unable to assess.
ACCEPTED MANUSCRIPT
Table 3: Summary of quality assessment and risk of bias assessment Risk of bias assessment
Quality assessment Allocation concealed
Outcome assessor blinding
Cointervention avoided
Loss to followup
ITT analysis
Selection bias
Performance bias
Detection bias
Attrition bias
Reporting bias
Other bias
Bird et al23
Not met
Met
Met
Not met
Met
Met
Low
High
Low
High
Low
High
Gildenhuys et al24
Not met
Not met
Not met
Not met
Unclear
Not met
High
High
High
High
Low
Low
Irez et al25
Not met
Not met
Met
Not met
Met
Not met
High
High
Low
Low
Low
Low
Kovach et al26
Not met
Not met
Not met
Not met
Unclear
Mokhtari27
Not met
Unclear
Not met
Not met
Unclear
SiqueiraRodrigues et al28
Not met
Not met
Not met
Not met
Met
SC
M AN U Not met
High
High
High
High
Low
Low
Not met
High
High
High
High
High
Low
Not met
High
High
High
High
High
Low
TE D EP AC C
ITT = intention to treat
RI PT
Patient/ therapist blinding
Study
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Figure 1: Pilates equipment exercises for balance
ACCEPTED MANUSCRIPT Figure 2: Flow chart of exclusion process
M AN U
Full text obtained (n=24)
SC
Excluded on title and abstract (n=285)
RI PT
Potentially relevant studies identified and screened for retrieval (n=309)
TE D
Excluded after reading full text (n=18) Not a comparative study of Pilates intervention (n=9) Population mean age not ≥60 years (n=6) Did not report on balance or fall outcomes (n=1) Active control group (n=2)
AC C
EP
Studies included in the meta-analysis (n=6)
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Figure 3: Meta-analysis of balance outcomes
Abbreviations: Std. = Standard; Total = Number of participants in the study group; SD = Standard deviation; IV = Inverse variance; TUG = Timed up and go; FTSST = Five-times-sit-to-stand test; 8FUG = 8-foot up and go; RSVI = Rank value stability index; POMA=Performance Oriented Mobility Assessment.
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Figure 4: Meta-analysis of fall outcomes
AC C
EP
TE D
Abbreviations: Std. = Standard; Total = Number of participants in the study group; SD = Standard deviation; IV = Inverse variance.