1. Original Study
Is Trunk Posture in Walking a Better Marker than Gait Speed
in Predicting Decline in Function and Subsequent Frailty?
Reshma A. Merchant MBBS, MRCP a,b
, Subhasis Banerji PhD c
, Gautam Singh MSc c
,
Effie Chew MBBS, MRCP a,b
, Chueh L. Poh PhD d
, Sarah C. Tapawan MRCPCH c
,
Yan R. Guo BSc c
, Yu W. Pang Dip c
, Mridula Sharma PhD e
, Ravi Kambadur PhD f
,
Stacey Tay MBBS, MMed (Paeds), FRCPCH c,g,
*
a
Department of Medicine, National University Health System, Singapore
b
Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
c
Department of Paediatrics, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
d
Division of Molecular Genetics and Cell Biology, School of Chemical and Biomedical Engineering, Nanyang Technological University, Singapore
e
Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
f
Division of Molecular Genetics and Cell Biology, School of Biological Sciences, Nanyang Technological University, Singapore
g
Department of Paediatrics, KTP-National University Children’s Medical Institute, National University Health System, Singapore
Keywords:
Sarcopenia
frailty
gait speed
walking speed
posture adaptation
gait analysis
falls
a b s t r a c t
Background: Many recent guidelines and consensus on sarcopenia have incorporated gait speed and grip
strength as diagnostic criteria without addressing early posture changes adopted to maintain gait speed
before weakness is clinically evident.
Objectives: Older adults are known to compensate well for declining physiological reserve through
environmental modification and posture adaptation. This study aimed to analyze and identify significant
posture adaptation in older adults that is required to maintain gait speed in the face of increasing
vulnerability. This would be a useful guide for early posture correction exercise interventions to prevent
further decline, in addition to traditional gait, balance, and strength training.
Design: A community-based cross-sectional study.
Setting and Participants: The participants comprised 90 healthy community-dwelling Chinese men
between the ages of 60 and 80 years and 20 Chinese adults between the ages of 21 and 50 years within
the normal BMI range as a comparison group.
Measurements: All the participants underwent handgrip strength testing, 6-minute walk, timed up-and-
go (TUG), and motion analysis for gait characteristics. Low function was characterized by slow walking
speed (<1.0 m/s) and/or slow TUG (>10 seconds), whereas low strength was determined by hand grip
dynamometer testing (<26 kg). The degree of frailty was classified using the Canadian Study for Health
and Ageing Clinical Frailty Scale (CSHA-CFS) to differentiate between healthy and vulnerable older adults.
Results: As expected, the vulnerable older adults had lower functional performance and strength
compared with the healthy older adults group. However, a significant number demonstrated posture
adaptations in walking in all 3 groups, including those who maintained a good walking speed (>1.0 m/s).
The extent of such adaptation was larger in the vulnerable group as compared with the healthy group.
Conclusion: Although gait speed is a robust parameter for screening older adults for sarcopenia and
frailty, our data suggest that identifying trunk posture adaptation before the onset of decline in gait
speed will help in planning interventions in the at-risk community-dwelling older adults even before
gait speed declines.
Ó 2015 AMDA e The Society for Post-Acute and Long-Term Care Medicine.
The authors declare no conflicts of interest.
This work was supported by a grant (NRF 2008 NRF-CRP 001e30) from the
Competitive Research Program of the National Research Foundation, Singapore.
* Address correspondence to A/Prof Stacey Tay, Department of Paediatrics, Na-
tional University Health System, 1E Kent Ridge Road, NUHS Tower Block, Level 12,
Singapore 119228.
E-mail address: stacey_tay@nuhs.edu.sg (S. Tay).
JAMDA
journal homepage: www.jamda.com
http://dx.doi.org/10.1016/j.jamda.2015.08.008
1525-8610/Ó 2015 AMDA e The Society for Post-Acute and Long-Term Care Medicine.
JAMDA xxx (2015) 1e6

2. With the aging of the population and the rising costs of health care,
many countries are refocusing health care policy on health promotion
and disability prevention among older people. It has been argued that
efforts aimed at identifying at-risk groups of older people so as to
provide early intervention and/or multidisciplinary case management
should be done at the level of general practice via adoption of a clinical
paradigm based on the concept of frailty, which fits well with the
biopsychosocial model of primary care. However, this ideal has
exposed the lack of frailty metrics that are appropriate for primary
care. Indeed, family physicians and community practitioners are still
in need of easy instruments for identifying and estimating frailty early.
Fried and Walston1
had hypothesized a “cycle of frailty” that was
consistent with clinical signs and symptoms. This hypothesis indi-
cated reduced levels of nutrition and activity, age-related musculo-
skeletal changes, and disease as being the possible precursors to loss
of muscle mass as seen in the onset of sarcopenia, which then pro-
gressed to decreased walking speed, strength, and power along with
respiratory and metabolic changes. In keeping with this hypothesis,
various researchers have tried to develop simplified protocols to
reliably identify frailty and associated sarcopenia, with the result that
a strong consensus has emerged around the use of walking or gait
speed as the most reliable and easy to administer marker.2e4
Gait
speed is now considered a strong predictor of a wide range of out-
comes in older adults, including falls and fractures, hospitalization,
cognitive decline, and mortality.5,6
Hence, many recent guidelines and
consensus definitions of sarcopenia have been based on gait speed,
but without addressing posture adaptation to maintain gait speed in
the face of weakness or joint stiffness.
With aging, older adults compensate for general decline through
environmental modification and posture adaptation. Lord et al7
demonstrated that to be dwelling in the community, one must
maintain a walking speed close to 1.14 m/s. Kang and Dingwell8
also
showed that the young and elderly both reported the same preferred
walking speed. Because the individuals in this study were all com-
munity dwelling, the obvious question was whether they were
adapting posture and gait to maintain a reasonable gait speed. It is
possible that those who developed early adaptations in posture were
more prone to early functional deficits and, hence, at a greater risk of
rapid progression to sarcopenia and frailty. These would represent,
more accurately, the “transition to frailty and sarcopenia” group in the
otherwise healthy, community-dwelling group of Chinese men. Such
an adaptation-mediated compensation for declining gait speed is not
very well addressed in the literature at the moment.
The objective of this prospective, cross-sectional community study
was, therefore, to identify early adaptations in posture during walking
that may precede actual decline in gait speed among healthy
community-dwelling Chinese men. A progression pathway from
posture adaptation to gait adaptation leading to decline in functional
measures such as gait speed and timed up-and-go (TUG) measures
was also hypothesized after data analysis.
Methods
The study team recruited 90 community-dwelling older Chinese
adults between the ages of 60 and 80 years with body mass index
(BMI) in the normal range (18.5e23.5 according to Asian standards).
An additional 20 adults were recruited in the 21 to 50 years age group
to act as a reference comparison. Their function was evaluated using
handgrip strength testing, 6-minute walk, TUG, and 6-camera motion
analysis for gait characteristics. Parameter cutoffs were initially
benchmarked using data from published literature by the Asian
Working Group on Sarcopenia for Chinese subjects9
and modified
based on data obtained from this study. Low function was character-
ized by slow walking speed (<1.0 m/s) and/or slow TUG
(>10 seconds), whereas low strength was determined by handgrip
dynamometer testing (<26 kg). The degree of frailty was classified
using the Canadian Study for Health and Ageing Clinical Frailty Scale
(CSHA-CFS)10
to differentiate between healthy and vulnerable elderly.
Rockwood et al10
proposed the CSHA-CFS, a global clinical measure
of fitness and frailty in elderly people (CSHA 1: Very fit; CSHA 2:
Without active disease but less fit than CSHA 1; CSHA 3: Well, with
treated and well-controlled comorbid disease; CSHA 4: Apparently
vulnerable, not dependent but complain of being “slowed up” and
have disease symptoms). The CSHA-CFS was derived from a 5-year
prospective cohort study. This 7-point scale was then further trans-
lated by Chan et al11
as the Chinese-Canadian physician version and
used to validate a telephone version. We used the CSHA-CFS physician
version algorithm to categorize our participants into 3 groups. Group 1
was healthy older adults (CSHA 1 and 2, n ¼ 41), group 2 was
intermediate-risk older adults (CSHA 3, n ¼ 33), and group 3 consisted
of the vulnerable older adults (CSHA 4, n ¼ 16). This was appropriate
because all the participants were community dwelling with no evi-
dence of significant functional impairment or dependence, which may
have placed them in the frailer categories of CSHA 5 to 7. None of the
recruited individuals remembered having a fall in the past 12 months.
Those with chronic obstructive pulmonary disease and cardiac pace-
makers were screened out, as were those on steroid medication and
growth hormones. Further, blood samples were drawn to record free
testosterone, growth hormone levels, and insulin-like growth factor
(IGF-1). This was done to further qualify the frailty classification, as all
3 blood markers are associated with frailty in published
literature.12e14
To analyze joint and limb segment kinematics, each participant
was fitted with 31 reflective anatomical (25-mm diameter) markers,
positioned according to the standard Plug-in Gait Marker Set, along
with 5 additional markers to track the head and 1 marker on the chin.
Participants were asked to walk along a walkway 6 meters in length
with 2 force plates, recording at least 3 trials per participant. Six
infrared cameras (Vicon MX system; Oxford Metrics, Oxford UK)
placed around the walkway recorded the coordinates of these
reflective markers at 200 Hz in 3 dimensions. The kinematics data
were smoothed using a Woltring filter with a mean squared error of
20. The software for data collection was the Vicon Workstation 5.1 and
the data analysis was performed with Vicon Polygon 3.1.
Ethics approval was obtained from Domain-Specific Review Board
of National Healthcare Group, Singapore. All participants provided
signed consent.
Data Analysis
At first, the strength and function data were used to verify that the
clinical frailty classification using the CSHA-CFS was accurate and was
reflected in the strength and function measures. We analyzed the
walking speed as measured in both 6-minute walk and camera-based
motion analysis study separately. The latter was performed in a lab-
oratory and involved only a few footfalls. It has been reported to be a
good measure for assessing frailty. However, the community-dwelling
participants in this study tended not to walk at their normal pace in
the gait laboratory and appeared to walk more normally (and faster)
while doing a 6-minute walk in the gym. Both the measures were
analyzed, as we did not know the extent of gait difficulties that might
be present in the individuals recruited from the community.
The joint and limb kinematics of each participant were analyzed
using the motion analysis data. For trunk posture, markers placed at
C7 on spinous process on the back, shoulders (acromion process),
sternum, anterior superior iliac spine (ASIS), and posterior superior
iliac spine were studied to identify posture variations in pitch and roll
during walking. The Plug-in Gait kinematic model from the Vicon
Polygon software was used to measure shoulder offset in the X-di-
rection (direction of walking) during 2 consecutive sets of left and
R.A. Merchant et al. / JAMDA xxx (2015) 1e62

3. right heel strikes. The mean of the 4 values was then considered as the
participant’s shoulder offset value for purposes of analysis and clas-
sification. Sway was similarly measured in the X-direction (ante-
roposterior movement) and Y-direction (mediolateral movement),
whereas hip elevation was measured using Z-direction (vertical
movement component) displacement of the ASIS markers. Ankle
dorsiflexion, ankle plantar flexion, knee and hip flexion angles were
similarly generated using the kinematic limb segment model.
Statistical analysis using Minitab (Minitab Inc., State College, PA)
software tools was carried out for the healthy, intermediate, and
vulnerable groups to study general distribution, trends, and statistical
significance of differences in posture and gait adaptation measures
between groups.
Results
Validating CSHA Groups Using Strength and Function Data
The strength, function, and temporal gait data were first evaluated
to validate the categorization of individuals based on frailty scale. Grip
strength, walking speed, and TUG measures were compiled, which
showed that the performance of some participants in the older adults
age group were comparable with the younger age group on one hand,
whereas others were significantly lower, as expected. Although the
testosterone and growth hormone values did not show any clear
trend, IGF-1 in the vulnerable group was lower than the other groups.
Table 1 gives the demographics, strength, function, and IGF-1 profiles
of the groups.
Cutoffs of strength and function were calculated as 2 SDs below the
mean of the control age group of 21 to 30 years. These cutoffs were
comparable to those reported by the Asian Working Group on
Sarcopenia. Walking speed (<1.0 m/s), grip strength (<26 kg), and
TUG (>10 seconds) were identified as benchmarks to further classify
those in the healthy older adults (CSHA 1 and 2) and intermediate-risk
older adults (CSHA 3) groups.
Overall, the vulnerable olderadult group (CSHA 4) performed worse
in the function tests as compared with the healthy group (Table 1). Grip
strength mean values were, however, comparable. All 3 older adult
groups had various degrees of mix between those who performed
poorly on strength and function and those who performed well.
Study of Overall Posture and Gait Kinematics
The vulnerable older adult group had significantly altered posture
when compared with the healthy older adult group while walking.
This difference was less between the intermediate-risk group and the
healthy older adults and significant to a lesser extent. The postural
parameters tracked during the gait cycle included trunk rotation, hip
flexion, and hip rotation. The parameters in the swing phase of gait
included knee flexion, dorsi- and plantar flexion, and anteroposterior
and mediolateral sway of the hip. The individual analyses are as
follows.
Stage-wise increasing trunk rotation with increased vulnerability
The posture was deemed to be significantly altered in groups 1 and
2 if the shoulder offset (representing trunk rotation) was equal to or
more than the mean value for group 3 (24.30 mm), as group 3 was the
clinically vulnerable group and had a large majority of subjects who
showed maximum posture adaptation. The degree of trunk rotation,
as seen in the shoulder offset, increased with increasing vulnerability
(Figure 1).
The trunk was rotated toward the left side, which was the
preferred stabilizing lower limb in 70.7% of group 1, 87.9% of group 2,
and 94.0% of group 3 participants (Table 2). There was a progressively
decreasing number of participants who maintained gait speed greater
than 1.0 m/s from group 1 to group 3. However, an increasing number
of such “healthy walkers” displayed significant trunk rotation from
group 1 to group 3 (Table 2). An example of such progressive adap-
tation is shown in Figure 2. Trunk posture adaptation correlated with
deficits of strength, function, and walking speed in each of the groups.
The data show that it may be difficult to separate those who are
clinically healthy and those who may be transitioning to frailty based
on gait speed alone, especially in community-dwelling older adults.
The trunk rotation measures provide a clearer picture of progressive
deterioration, allowing earlier detection. Although there is increased
hip flexion in Figure 2 (B and C), the actual angle at the hip during
walking was not significantly different between the groups. The
greater difference was in the trunk rotation, and all those with trunk
rotation displayed the bent-forward posture. The pitch and roll of the
upper body as measured by movement of the C7 marker with respect
to center of mass was also found to have no significant difference. This
Table 1
Demographic Data of Individuals With Strength, Function, Gait, and Blood Parameters
Data Particulars Young Adults, 21e50 y
n ¼ 20
Mean (SD)
Older Adults, 60e80 y
Group 1, CSHA 1 and 2
n ¼ 41
Mean (SD)
Group 2, CSHA 3
n ¼ 33
Mean (SD)
Group 3, CSHA 4
n ¼ 16
Mean (SD)
No. of individuals
Age, y 33.85 (10.49) 66 (4.97) 66.33 (3.96) 70.56 (4.55)
BMI, kg/m2
21.47 (2.49) 22.98 (1.89) 23.45 (1.74) 23.03 (3.45)
Grip strength, kg 35 (7.27) 28.9 (5.93) 29.88 (5.11) 29.62 (4.85)
Gait speed: motion analysis, m/s 1.21 (0.14) 1.13 (0.17) 1.14 (0.17) 1.03 (0.25)
Gait speed: 6-min walk, m/s 1.70 (0.29) 1.42 (0.25) 1.43 (0.23) 1.27 (0.33)
TUG, s 7.84 (1.15) 8.57 (1.96) 8.89 (1.44) 10.03 (3.03)
Toe-off as percentage of stance phase of gait cycle, % 58.61 (2.35) 59.85 (2.15) 60.18 (1.98) 60.45 (2.45)
Stride length, m 1.27 (0.18) 1.24 (0.11) 1.25 (0.10) 1.18 (0.14)
IGF-1, ng/mL 181.95 (47.85) 132.07 (46.87) 130.24 (46.93) 129.19 (42.73)
Fig. 1. The trunk rotation adaptation within groups 1, 2, and 3 increased progressively
from the healthy (group 1) to the vulnerable (group 3). Differences in the shoulder
offset were particularly significant between groups 1 and 4.
R.A. Merchant et al. / JAMDA xxx (2015) 1e6 3

4. may be because none of the participants, all living and functioning in
the community, were so frail as to have significant hip flexion, hip
hitching, or increased upper body mediolateral movement during
walking.
Adaptations in gait kinematics associated with trunk adaptation
Although the hip and knee flexion in the older adult age group
overall during the swing phase of gait was less than the young age
groups, they did not differ significantly between healthy and vulner-
able older adults.
Individuals with significant trunk rotation toward the left side
among the vulnerable older adult group demonstrated reduced hip
elevation and hip rotation during the swing phase of gait on the left
side as compared with healthy older adults, as also increased dorsi-
flexion in swing phase and reduced plantar flexion at toe-off stage of
gait as seen in the motion analysis (Table 3).
Although the healthy older adults showed a difference in hip
elevation between left and right side of approximately 2 mm, in the
vulnerable, the left elevation was on an average approximately 7 mm
less than the right. In the vulnerable group, hip rotation measured
primarily negative values (backward rotation) and the hip never
rotated past the zero-angle mark, whereas the healthy group rotated
the hip on both sides (positive and negative) of the zero mark. The
maximum plantar flexion on the left side during toe-off was reduced
overall in the vulnerable group (although not significant), with a
corresponding increase in dorsiflexion during swing (as a possible
compensation for the reduced hip elevation). The sway in group 3 was
reduced in the anteroposterior direction, but with a greater percent-
age increase in backward sway. The mediolateral sway in this group
was reduced as well, in particular the sway toward the left side. On
revisiting the gait analysis data, it was found that the vulnerable group
also had increased toe-off percentage in stance phase and decreased
stride length (Table 1).
Discussion
To our knowledge, this is the first study that has explicitly shown
that posture adaptation precedes decline in gait speed. Most of the
literature on sarcopenia and frailty focuses around gait speed, as it is
well known that slow gait speed leads to falls and fractures, hospi-
talization, cognitive decline, and subsequent mortality. Most in-
terventions for frailty and sarcopenia have focused mainly on
resistance, endurance, and balance training without specific attention
to correcting posture. The findings from our study support the need
for posture correction rehabilitation for all older adults who display
posture adaptation, even though they may have good gait speed at the
time of evaluation. Such intervention may then affect the occurrence
of falls, as well as reduce the increased backward sway shown by the
vulnerable group.
A posture adaptation study with elderly in quiet standing was
published by Kirby et al.15
Kirby et al15
found that the elderly in-
dividuals placed the left foot approximately 30 to 40 mm behind the
right foot so as to stabilize in quiet standing. Such an adaptation
resulted in minimizing anteroposterior and mediolateral sway. The
Table 2
Trunk Rotation (Shoulder Offset) Values and Percentage Distribution Between Groups
Group Mean Shoulder
Offset, mm
Mean (SD)
Overall % Individuals
With Trunk Rotation
% of Individuals
With Gait Speed >1 m/s
% of Individuals With Gait Speed >1 m/s
AND Significant Trunk Rotation
1 (CSHA 1, 2) 9.03 (10.07) 70.7 95.12 24.39
2 (CSHA 3) 15.11 (15.74) 87.9 93.93 39.39
3 (CSHA 4) 24.30 (8.31) 94.0 62.50 67.00
Fig. 2. (A, B) Individuals who have walking speed >1.0 m/s. (A) Individual from group 1 and shows no posture adaptation. (B) Individual from group 2 who showed posture
adaptation while maintaining a good walking speed. (C) Participant from group 3 with a walking speed <1.0 m/s and an altered posture.
R.A. Merchant et al. / JAMDA xxx (2015) 1e64

5. results of this study show a more stage-wise transition from a mean
9.03-mm shoulder offset in the healthy elderly (comparable with that
in younger individuals) during walking to 15.11 mm in the interme-
diate group to 24.30 mm in the vulnerable elderly group (Table 1). The
range of offset in group 2 was the largest, arguably characterizing a
group that is in transition. The percentages of individuals adapting in
each group was also progressively higher, from 70.7% in the healthy
group up to 94.0% in the vulnerable group. Although grip strength in
the dominant hand, walking speed, and TUG trended toward lower
values in the vulnerable group, each group had a mix of those who
performed as well as the healthy group. In particular, more than half of
the vulnerable group maintained a gait speed faster than 1 m/s, of
whom 67% had adapted posture. All of the slow walkers in the 3
groups showed posture adaptation, with most stabilizing on the left
side. This seems to suggest that posture adaptation during walking
precedes decline in gait speed and is a better marker to identify early
those at risk of functional decline and frailty. This adaptation could be
due to a hip strategy that might facilitate step initiation with the
dominant right leg.16
Change in the walking posture was also associ-
ated with standing posture with individuals displaying a distinct
retropulsion stance when not walking.17
An example of such a case
from the intermediate-risk group is shown in Figure 3. Diseases
causing backward disequilibrium in older adults (such as Parkinson
disease, multiple systems atrophy, strokes, and normal-pressure hy-
drocephalus) are known to be associated with higher risk of falls,
highlighting the importance of retropulsion in both normal aging as
well as in disease states.18
The decline in hip rotation, hip elevation, and plantar flexion at the
ankle with increased dorsiflexion during swing phase on the left side
suggests a progression to adaptation in traditional gait characteristics.
Although the reduced hip movement may ultimately manifest as or
exacerbate existing hip stiffness and weakness, the reduced ankle
plantar flexion is known to affect the swing phase and stride length in
gait. It is likely that the deficits in hip movement and stride length
would then combine to affect traditional measures of frailty, such as
gait speed and TUG.
Based on this information, one may hypothesize that posture and
gait adaptations precede actual decline in walking speed, as the older
adults adjust to increasing weakness and joint stiffness until the scope
of such adaptation is exhausted. The progression may be represented
by the following stages:
Increasing tendency to stabilize due to age and weakness
Adaptation with trunk rotation toward one side (most stabilize
on the left side)
Reduced hip rotation angle and reduced hip elevation on sta-
bilizing side
Resultant adaptation with a reduced plantar flexion at toe-off
and increased ankle dorsiflexion during swing of left leg to
compensate for poor hip elevation
Progressive reduction in stride length due to a shorter swing
phase and higher toe-off percentage in stance phase
Reduced gait speed and poor TUG due to reduced stride length
and increased hip stiffness combined
A future longitudinal study may be used to investigate such a
progression hypothesis. Although gait speed is measured often as part
of screening for frailty, there are varied opinions on exercise in-
terventions to improve failing gait ability.19e21
This data analysis
suggests that specific exercises addressing biomechanical alignment,
improved range of motion of the hip, strengthening of the gluteus
medius, and strengthening of the muscles to improve plantar flexion
and posture correction exercises to reduce trunk rotation and retro-
pulsion, among others, would benefit those older adults in the com-
munity to delay the decline in gait speed and help them remain active
and healthy much longer.
Conclusion
Decline in gait speed is widely used as a reliable marker for onset of
frailty and associated sarcopenia. This study indicates that trunk
posture adaptation precedes decline in gait speed. A significant
number of Chinese male older adults who maintained good gait speed
demonstrated posture adaptations in walking, whereas all slow
walkers had posture adaptation. It is useful to track trunk posture
adaptation during walking to identify the at-risk older adults earlier,
even before gait speed declines.
Early exercise interventions concentrated around posture correc-
tion, hip and ankle range of motion, and strengthening of specific
muscles on the stabilizing side may then help delay such vulnerable
community-dwelling elders from progressing to frailty, keeping them
independent longer.
Further prospective studies are needed to see if healthy older adults
with trunk posture adaptation and normal gait speed develop impaired
gait speed earlier than those without such adaptation and if this can
be prevented with posture correction and range of motion exercises.
Fig. 3. Example of a participant from group 2, considered intermediate risk, who had
healthy gait speed (1.33 m/s) but displayed posture adaptation in walking (A) and
retropulsion while standing (B) with increased sway asymmetry as shown in Table 2.
Table 3
Hip and Ankle Movement and Sway Characteristics on Stabilizing Left Side for the Older Adults Indicating Adaptation in Gait in Group 3 as Compared With Group 1
Group Hip Rotation, Degrees Hip Elevation
Difference, mm
Dorsiflexion in
Swing Phase, Degrees
Plantarflexion in
Toe-off Phase, Degrees
Sway, mm
Left (Right) À (Left) Left Left Left Back
Group 1 2 2.2 6 11 16.3 9.5
Group 3 À3 7.1 9 6 10.1 13
R.A. Merchant et al. / JAMDA xxx (2015) 1e6 5

6. References
1. Fried LP, Walston J. Frailty and failure to thrive. In: Hazzard WR, Blass JP,
Ettinger Jr WH, editors. Principles of Geriatric Medicine and Gerontology. 4th
ed. New York: McGraw Hill; 1998. p. 1387e1402.
2. Syddall HE, Westbury LD, Cooper C, Sayer AA. Self-reported walking speed: A
useful marker of physical performance among community-dwelling older
people? J Am Med Dir Assoc 2015;16:323e328.
3. Guedes RC, Dias RC, Pereira LSM, et al. Influence of dual task and frailty on gait
parameters of older community dwelling individuals. Braz J Phys Ther 2014;18:
445e452.
4. Parentoni AN, Mendonça VA, Dos Santos KD, et al. Gait speed as a predictor of
respiratory muscle function, strength and frailty syndrome in community
dwelling elderly people. J Frailty Aging 2015;4:64e68.
5. Afilalo J, Eisenberg MJ, Morin JF, et al. Gait speed as an incremental predictor of
mortality and major morbidity in elderly patients undergoing cardiac surgery.
J Am Coll Cardiol 2010;56:1668e1676.
6. Quach L, Galica AM, Jones RN, et al. The non-linear relationship between
gait speed and falls: The maintenance of balance, independent living,
intellect and zest in the elderly Boston study. J Am Geriatr Soc 2011;59:
1069e1073.
7. Lord SE, McPherson K, McNaughton HK, et al. Community ambulation after
stroke: How important and obtainable is it and what measures appear pre-
dictive? Arch Phys Med Rehabil 2004;85:234e239.
8. Kang HG, Dingwell JB. Effects of walking speed, strength and range of
motion on gait stability in healthy older adults. J Biomech 2008;41:
2899e2905.
9. Chen LK, Liu LK, Woo J, et al. Sarcopenia in Asia: Consensus report of the Asian
Working Group for Sarcopenia. J Am Med Dir Assoc 2014;15:95e101.
10. Rockwood K, Song X, McKnight C, et al. A global clinical measure of fitness and
frailty in elderly people. CMAJ 2005;173:489e495.
11. Chan DC, Tsou HH, Chen CY, Chen CY. Validation of the Chinese-Canadian study
of health and ageing clinical frailty scale (CSHA-CFS) telephone version. Arch
Gerontol Geriatr 2010;50:e74ee80.
12. Yang CW, Li TC, Li CI, et al. Insulin-like growth factor-1 and its binding protein-
3 polymorphisms predict circulating igf-1 level and appendicular skeletal
muscle mass in Chinese elderly. J Am Med Dir Assoc 2015;16:365e370.
13. Hyde Z, Flicker L, Almeida OP, et al. Low free testosterone predicts frailty in
older men: The health in men study. J Clin Endocrinol Metab 2010;95:
3165e3172.
14. Hennessey JV, Chromiak JA, DellaVentura S, et al. Growth hormone adminis-
tration and exercise effects on muscle fiber type and diameter in moderately
frail older people. J Am Geriatr Soc 2001;49:852e858.
15. Kirby RL, Price NA, MacLeod DA. The influence of foot position on standing
balance. J Biomech 1987;20:423e427.
16. Berger L, Chuzel M, Buisson G, Rougier P. Undisturbed upright stance control in
the elderly: Part 1. Age-related changes in undisturbed upright stance control.
J Mot Behav 2005;37:348e358.
17. Morley JE, Rolland Y, Tolson D, Vellas B. Increasing awareness of the factors
producing falls: The Mini Falls Assessment. J Am Med Dir Assoc 2012;13:
87e90.
18. Manckoundia P, Mourey F, Perennou D, Pfitzenmeyer P. Backward disequilib-
rium in elderly subjects. Clin Interv Aging 2008;3:667e672.
19. Van Swearingen JM, Perera S, Brach JS, et al. Impact of exercise to
improve gait efficiency on activity and participation in older adults with
mobility limitations: A randomized controlled trial. Phys Ther 2011;91:
1740e1751.
20. Valenzuela T. Efficacy of progressive resistance training interventions in older
adults in nursing homes: A systematic review. J Am Med Dir Assoc 2012;13:
418e428.
21. Cadore EL, Rodriguez-Manas L, Sinclair A, Izquierdo M. Effects of different
exercise interventions on risk of falls, gait ability and balance in physically frail
older adults: A systematic review. Rejuvenation Res 2013;16:105e114.
R.A. Merchant et al. / JAMDA xxx (2015) 1e66