Normative spatiotemporal gait parameters in older adults
John H. Hollman a,
*, Eric M. McDade b
, Ronald C. Petersen c
a
Department of Physical Medicine & Rehabilitation, Program in Physical Therapy, College of Medicine, Mayo Clinic, Rochester, MN, United States
b
Department of Neurology, University of Pittsburgh, Pittsburgh, PA, United States
c
Department of Neurology, College of Medicine, Mayo Clinic, Rochester, MN, United States
1. Introduction
Gait may be used to assess quality of life [1], health status [2], and
physical function [3] in older adults. Specific parameters may be
used to assess risk of dementia [4], risk of falling [5], and even risk of
early mortality [6]. Understanding how gait is associated with these
clinical phenomena is challenging, in part because dozens of gait
parameters can be measured and consensus on which are most
relevant is lacking. Moreover, normative data for many parameters
arelackingand instudies thatreportreference values, thereisagreat
deal of variability among the data reported [2,7–11].
Verghese et al. [4] while studying the relationship between gait
and risk of dementia, identified three domains that characterize gait
performance in older adults. Pace differences characterized by gait
speed and stride length were associated with reduced executive
function; rhythm differences characterized by cadence, swing time
and stance time were associated with memory decline; and
variability differences characterized by stride length variability
distinguished subtypes of mild cognitive impairment. The variability
domain may also best predict falls in older adults [5]. Those three
domains of gait, however, constitute only a small proportion of gait
parameters that can be quantified. Additional parameters may yield
greater insight into how gait is associated with other clinical
phenomena.
In addition to understanding which parameters characterize
gait performance, classifying gait as being dysfunctional necessitates
that a clinician understands what ‘‘normal’’ gait is. While
normal gait is not readily defined, normative and population-based
data provide insight about what one might expect the magnitude
of certain gait parameters to be. Normative data represent
descriptive or prescriptive measurements as they ought to be;
hence normative studies examine parameters in healthy participants.
In contrast, population-based data represent measures as
they are in a population; hence population-based studies examine
parameters in participants oftentimes regardless of health status.
Gait speed is the most often reported reference value for gait
performance. Reported values of mean gait speed in adults aged
70–79 range from approximately 90–130 cm/s [2,7–11]. Normative
studies [2,9–11] typically report higher values than population-based
studies [7,8], presumably because the normative
studies describe gait in healthy individuals whereas populationGait
& Posture 34 (2011) 111–118
A R T I C L E I N F O
Article history:
Received 14 September 2010
Received in revised form 9 March 2011
Accepted 22 March 2011
Keywords:
Gait
Aging
Sex characteristics
Reference values
A B S T R A C T
While factor analyses have characterized pace, rhythm and variability as factors that explain variance in
gait performance in older adults, comprehensive analyses incorporating many gait parameters have not
been undertaken and normative data for many of those parameters are lacking. The purposes of this
study were to conduct a factor analysis on nearly two dozen spatiotemporal gait parameters and to
contribute to the normative database of gait parameters from healthy, able-bodied men and women over
the age of 70. Data were extracted from 294 participants enrolled in the Mayo Clinic Study of Aging.
Spatiotemporal gait data were obtained as participants completed two walks across a 5.6-m electronic
walkway (GAITRite1
). Five primary domains of spatiotemporal gait performance were identified: a
‘‘rhythm’’ domain was characterized by cadence and temporal parameters such as stride time; a ‘‘phase’’
domain was characterized by temporophasic parameters that constitute distinct divisions of the gait
cycle; a ‘‘variability’’ domain encompassed gait cycle and step variability parameters; a ‘‘pace’’ domain
was characterized by parameters that included gait speed, step length and stride length; and a ‘‘base of
support’’ domain was characterized by step width and step width variability. Several domains differed
between men and women and differed across age groups. Reference values of 23 gait parameters are
presented which researchers or clinicians can use for assessing and interpreting gait dysfunction in aging
persons.
ß 2011 Elsevier B.V. All rights reserved.
* Corresponding author at: College of Medicine, Mayo Clinic, Department of
Physical Medicine & Rehabilitation, Program in Physical Therapy, Mayo School of
Health Sciences, Siebens 11-04, 200 First Street SW, Rochester, MN, 55905, United
States. Tel.: +1 507 284 9547; fax: +1 507 284 0656.
E-mail address: hollman.john@mayo.edu (J.H. Hollman).
Contents lists available at ScienceDirect
Gait & Posture
journal homepage: www.elsevier.com/locate/gaitpost
0966-6362/$ – see front matter ß 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2011.03.024
based studies include participants who may have pathological
conditions affecting their gait performance. While reference values
for other spatiotemporal parameters have been reported, the
magnitudes of those measurements, similar to gait speed, are quite
variable. Normative data for many parameters—particularly those
quantifying variability—are lacking.
The purpose of this study was twofold. First, we conducted a
factor analysis on nearly 2 dozen gait parameters to examine the
spatiotemporal domains of gait that researchers and clinicians may
consider measuring to enhance their understanding of gait
performance. Factor analysis organizes multiple observations into
communalities that correlate with a lesser number of unobserved
1,750 eligible participants enrolled in
Mayo Clinic Study of Aging who
completed gait assessment
Excluded: Dementia associated with probable Alzheimer’s
disease, Lewy body dementia, or mild cognitive impairment
(n=278)
1,472 eligible participants
Excluded: Stroke, Parkinson’s disease or other neuromuscular
disease (n=45)
1,427 eligible participants
Excluded: Depressive/anxiety disorders, alcoholism, bipolar
disorders or paranoid schizophrenia (n=99)
1,328 eligible participants
Excluded: Diabetes (n=260)
Final Sample: n=294
• 108 men (age=79±5 years, height=175±6 cm, weight=85±13 kg, BMI=27.7±3.9 kg/m2
)
• 186 women (age=79±5 years, height=160±6 cm, weight=67±12 kg, BMI=26.1±4.3 kg/m2
)
1,068 eligible participants
Excluded: Cancer (n=496)
572 eligible participants
Excluded: Myocardial infarction, congestive heart failure,
blood clots, atrial fibrillation, coronary artery bypass graft,
angioplasty, pacemaker, other heart surgeries (n=164)
408 eligible participants
Excluded: lumbar spine surgery or leg pain when walking
(n=114)
Fig. 1. The flow sheet describes participant selection, with exclusion criteria.
J.H. Hollman et al. / Gait & Posture 34 (2011) 111–118112
thematic constructs, thus allowing an investigator to partition a
large number of parameters into a lesser number that characterize
distinct domains of the parameters being measured. Second, we
sought to contribute to the normative database of gait parameters in
older, able-bodied adults. Establishing normative data may provide
clinicians and researchers values against which measurements can
be compared for assessing and interpreting gait dysfunction.
2. Methods
2.1. Participants
Data were extracted from 1750 ambulatory participants aged 70+ who were
enrolled in the Mayo Clinic Study of Aging, a population-based study of aging [12].
Participants were recruited via stratified sampling in which Olmsted County,
Minnesota, residents from within four strata (men aged 70–79 and 80–89, and
women aged 70–79 and 80–89) were randomly selected and invited to participate in
the study. Most participants were Caucasian (93%), retired (94%), lived in their own
dwelling (85%) and were married (59%). Data from participants who presented with
morbidities that affect gait were excluded from the analysis (Fig. 1). After exclusion,
108 male participants and 186 female participants remained, each of whom rated
their health status as good to excellent on a 5-point ordinal scale. The study was
approved by the Mayo Foundation Institutional Review Board and all subjects
provided informed consent prior to participating in the study.
2.2. Instrumentation
Gait performance was measured with GAITRite1
instrumentation (CIR Systems
Inc., Havertown, PA) consisting of an electronic walkway 5.6 m in length and 0.9 m in
width. The walkway encapsulated multiple sensor pads placed on 1.27 cm centers
that were activated under pressure at footfall and deactivated at toe-off, enabling the
systemtocapturethe relativearrangementoffootfallsas a functionoftime. Data were
sampled at 80 Hz and processed using GAITRite1
Platinum software.
2.3. Procedures
Subjects completed two walks across the walkway, initiating and terminating
their walks 1 m fore and aft of the walkway to minimize acceleration effects. Data
from both walks were combined and considered as a single test. Subjects were
instructed to walk at their normal pace and were not permitted to use gait aids
during testing. Testing was conducted within a clinical research center.
We collected data from 16 spatiotemporal and temporophasic parameters that
quantify mean values (Table 1). Additionally, we quantified variability in step
length, stride length, step width, step time, stride time, stance time, swing time,
stride speed and step width. We used the coefficient of variation (%CV) to reflect
variability for each of the parameters with the exception of step width, for which we
used the SD. Step width variability was expressed as the SD because the mean step
width is relatively small in magnitude and therefore the %CV (the SD normalized to
the magnitude of the mean) was excessively large in several cases and skewed the
data, making parametric statistical analyses difficult to interpret.
2.4. Data analysis
Data were analyzed with SPSS 15.0 software (SPSS Inc., Chicago, IL). Principal
components factor analysis with varimax rotation was used to examine factors with
eigenvalues exceeding 1.0 that characterized gait performance. Parameters with
correlation loadings of 0.5 or higher were interpreted as being significant
contributors to the factor. Once the factor analysis was completed, then data
under each domain were analyzed with two-way multivariate analysis of variance
(MANOVA), comparing gender and age group (ages 70–74, 75–79, 80–84, and 85+)
as independent variables (a = 0.05), and post hoc Bonferroni corrections for
multiple comparisons for significant MANOVAs.
3. Results
3.1. Factor analysis
Five factors accounted for 83.3% of the variance in gait
performance (Table 2). The first factor accounted for 25.8% of
the variance in gait performance and loaded highly on temporal
parameters quantifying cadence, step time, stride time, swing
time, stance time and single limb support time. We labeled this
factor a ‘‘rhythm’’ factor. The second factor accounted for 20.0% of
the variance and loaded highly on temporophasic divisions of the
gait cycle comprising swing, stance, single limb support, and
double limb support as well as double limb support time. We
labeled this factor a ‘‘phase’’ factor. The third factor accounted for
19.1% of the variance and loaded highly on the variability
parameters, excluding step width variability. We labeled this
Table 1
Operational definitions of gait parameters.
Gait parameter Operational definition
Spatial parameters
Step length (cm) Anterior-posterior distance from the heel of one footprint to the heel of the opposite footprint
Stride length (cm) Anterior-posterior distance between heels of two consecutive footprints of the same foot (left to
left, right to right); two steps (e.g., a right step followed by a left step) comprise one stride or one gait
cycle
Step width (cm) Lateral distance from heel center of one footprint to the line of progression formed by two
consecutive footprints of the opposite foot
Temporal parameters
Cadence (steps/min) Number of steps per minute, sometimes referred to as step rate
Step time (s) Time elapsed from initial contact of one foot to initial contact of the opposite foot
Stride time (s) Time elapsed between the initial contacts of two consecutive footfalls of the same foot
Stance time (s) The stance phase is the weight bearing portion of each gait cycle initiated at heel contact and ending
at toe off of the same foot; stance time is the time elapsed between the initial contact and the last
contact of a single footfall
Swing time (s) The swing phase is initiated with toe off and ends with initial contact of the same foot; swing time is
the time elapsed between the last contact of the current footfall to the initial contact of the next
footfall of the same foot
Single support time (s) Single support occurs when only one foot is in contact with the ground; single support time is the
time elapsed between the last contact of the opposite footfall to the initial contact of the next
footfall of the same foot
Double support time (s) Double support occurs when both feet are in contact with the ground simultaneously; double
support time is the sum of the time elapsed during two periods of double support in the gait cycle
Temporophasic parameters
Stance time (%GC) Stance time normalized to stride time
Swing time (%GC) Swing time normalized to stride time
Single support time (%GC) Single support time normalized to stride time
Double support time (%GC) Double support time normalized to stride time
Spatiotemporal parameters
Gait speed (cm/s) Calculated by dividing the distance walked by the ambulation time
Stride speed (cm/s) Calculated by dividing stride length by the stride time
Note: cm = centimeters; s = seconds; %GC = % gait cycle.
J.H. Hollman et al. / Gait & Posture 34 (2011) 111–118 113
Table 2
Factor loadings of gait parameters on five factors rotated and extracted by factor analysis.
Gait parameter Rhythm Phases Variability Pace Base of support
Gait speed (cm/s) À0.463 0.368 À0.087 0.779 0.088
Cadence (steps/min) À0.940 0.021 0.228 0.016 0.009
Mean step time (s) 0.955 À0.080 À0.093 À0.027 À0.091
Mean step length 0.135 0.335 À0.173 0.900 0.055
Mean stride time (s) 0.970 À0.063 À0.143 À0.035 À0.116
Mean stride length (cm) 0.123 0.380 À0.211 0.874 0.022
Mean step width (cm) 0.202 À0.012 À0.080 À0.123 À0.841
Mean swing (%GC) À0.015 0.910 À0.007 0.245 0.043
Mean swing time (s) 0.838 0.447 À0.120 0.123 À0.067
Mean stance (%GC) 0.028 À0.894 0.010 À0.199 0.032
Mean stance time (s) 0.898 À0.356 À0.135 À0.109 À0.100
Single support (%GC) 0.159 0.759 À0.339 0.145 0.101
Single support time (s) 0.826 0.397 À0.263 0.094 À0.018
Double support (%GC) À0.088 À0.921 0.217 À0.233 À0.068
Double support time (s) 0.541 À0.756 0.037 À0.243 À0.140
Step length variability (%CV) À0.374 À0.112 0.593 À0.213 À0.084
Step time variability (%CV) À0.322 À0.024 0.695 À0.108 À0.055
Stride length variability (%CV) À0.100 À0.036 0.883 À0.093 0.041
Stride time variability (%CV) À0.130 À0.117 0.921 À0.018 0.136
Swing time variability (%CV) À0.273 À0.167 0.832 0.009 0.079
Stance time variability (%CV) 0.008 À0.059 0.702 À0.042 0.379
Stride speed variability (%CV) 0.188 À0.113 0.526 À0.036 0.058
Step width variability SD (cm) À0.035 0.136 0.166 À0.027 0.778
% of variance 25.8 20.0 19.1 11.6 6.9
Note: Bolded correlation coefficients are interpreted as significant contributors to the identified factor.
Table 3
Normative data (Mean Æ SD) for parameters constituting rhythm, phases, variability, pace and base of support domains of gait performance by gender and age.
Parameter Men (N = 108) Women (N = 186)
70–74 75–79 80–84 85+ 70–74 75–79 80–84 85+
N = 27 N = 30 N = 37 N = 14 N = 33 N = 77 N = 43 N = 33
Rhythm
Cadence (steps/min)*
102 Æ 8 106 Æ 10 103 Æ 8 102 Æ 11 113 Æ 20 114 Æ 13 110 Æ 9 108 Æ 10
Step time (s)y
0.59 Æ 0.05 0.56 Æ 0.05 0.59 Æ 0.04 0.59 Æ 0.08 0.53 Æ 0.06 0.53 Æ 0.06 0.55 Æ 0.05 0.56 Æ 0.05
Stride time (s)z
1.18 Æ 0.08 1.13 Æ 0.09 1.16 Æ 0.08 1.19 Æ 0.14 1.06 Æ 0.13 1.06 Æ 0.12 1.10 Æ 0.09 1.12 Æ 0.11
Swing time (s)§
0.43 Æ 0.03 0.41 Æ 0.03 0.42 Æ 0.04 0.42 Æ 0.05 0.39 Æ 0.05 0.38 Æ 0.05 0.39 Æ 0.04 0.40 Æ 0.04
Stance time (s)ô
0.75 Æ 0.07 0.72 Æ 0.06 0.74 Æ 0.06 0.78 Æ 0.11 0.68 Æ 0.10 0.67 Æ 0.08 0.71 Æ 0.07 0.72 Æ 0.09
Single support time (s)#
0.44 Æ 0.03 0.42 Æ 0.03 0.42 Æ 0.04 0.42 Æ 0.04 0.39 Æ 0.06 0.38 Æ 0.06 0.39 Æ 0.04 0.40 Æ 0.04
Phases
Swing (%GC) 36.6 Æ 1.5 36.7 Æ 1.5 36.6 Æ 2.8 35.1 Æ 2.69 36.6 Æ 2.6 36.1 Æ 3.0 35.5 Æ 2.5 35.7 Æ 2.6
Stance (%GC) 63.2 Æ 2.1 64.0 Æ 2.5 63.8 Æ 2.7 64.9 Æ 2.7 63.3 Æ 3.1 63.9 Æ 3.0 64.5 Æ 2.6 64.5 Æ 2.5
Single support (%GC) 37.1 Æ 1.8 37.0 Æ 1.7 36.5 Æ 2.2 35.2 Æ 2.1 37.0 Æ 3.20 35.8 Æ 4.8 35.6 Æ 2.4 35.7 Æ 2.8
Double support (%GC) 26.3 Æ 3.0 26.5 Æ 2.3 27.4 Æ 4.7 30.3 Æ 3.5 27.14 Æ 4.0 28.4 Æ 6.4 29.0 Æ 4.6 28.7 Æ 4.8
Double support time (s)**
0.31 Æ 0.05 0.30 Æ 0.04 0.32 Æ 0.06 0.36 Æ 0.08 0.29 Æ 0.06 0.30 Æ 0.06 0.32 Æ 0.06 0.32 Æ 0.08
Variability
Step length (%CV) 4.6 Æ 6.7 5.4 Æ 2.7 5.1 Æ 2.8 5.8 Æ 3.4 7.7 Æ 11.6 5.7 Æ 7.2 5.9 Æ 2.7 6.2 Æ 2.4
Step time (%CV) 5.2 Æ 6.9 4.1 Æ 2.9 4.7 Æ 2.0 5.0 Æ 1.7 7.1 Æ 8.6 5.9 Æ 6.4 5.5 Æ 2.6 5.5 Æ 2.6
Stride length (%CV) 2.9 Æ 1.1 4.2 Æ 4.6 3.8 Æ 2.0 5.7 Æ 3.2 4.1 Æ 4.7 4.7 Æ 5.6 4.3 Æ 2.1 5.2 Æ 5.2
Stride time (%CV) 3.5 Æ 2.2 3.8 Æ 4.4 3.3 Æ 1.9 5.3 Æ 3.7 3.9 Æ 3.9 5.2 Æ 6.9 4.4 Æ 1.9 4.9 Æ 5.3
Swing time (%CV) 4.5 Æ 2.2 4.5 Æ 7.8 5.1 Æ 2.2 8.6 Æ 11.4 6.2 Æ 10.5 8.5 Æ 9.5 6.2 Æ 2.2 8.0 Æ 9.7
Stance time (%CV) 4.9 Æ 4.6 5.9 Æ 7.0 4.7 Æ 3.0 5.3 Æ 2.9 5.7 Æ 4.4 5.3 Æ 5.1 5.3 Æ 2.9 5.6 Æ 4.1
Stride speed (%CV) 5.0 Æ 2.9 5.5 Æ 4.4 5.5 Æ 3.2 7.3 Æ 3.8 5.6 Æ 3.4 5.5 Æ 2.7 6.8 Æ 3.2 6.9 Æ 3.3
Pace
Gait speed (cm/s)yy
117 Æ 16 122 Æ 15 112 Æ 17 101 Æ 22 116 Æ 20 112 Æ 17 101 Æ 15 98 Æ 20
Step length (cm)zz
69 Æ 8 68 Æ 7 65 Æ 8 59 Æ 10 61 Æ 9 59 Æ 7 55 Æ 7 54 Æ 9
Stride length (cm)§§
139 Æ 14 137 Æ 12 131 Æ 17 119 Æ 21 123 Æ 17 118 Æ 15 111 Æ 14 109 Æ 18
Base of support
Step width (cm)ôô
9.7 Æ 3.0 8.9 Æ 5.2 11.2 Æ 4.0 9.9 Æ 4.8 7.0 Æ 3.5 7.7 Æ 4.0 7.9 Æ 4.1 9.1 Æ 2.6
Step width SD (cm) 3.1 Æ 2.2 2.9 Æ 1.9 3.3 Æ 2.3 2.8 Æ 1.2 3.4 Æ 2.4 3.2 Æ 2.5 3.6 Æ 3.1 3.0 Æ 1.1
*
Mean difference in cadence between genders is 8.3 steps/min (p < 0.001, 95%CI = 5.4–11.1 steps/min).
y
Mean difference in step time between genders is 0.04 s (p < 0.001, 95%CI = 0.03–0.06 s).
z
Mean difference in stride time between genders is 0.09 s (p < 0.001, 95%CI = 0.06–0.11 s).
§
Mean difference in swing time between genders is 0.03 s (p < 0.001, 95% CI = 0.02–0.04 s).
ô
Mean difference in stance time between genders is 0.05 s (p = 0.004, 95%CI = 0.03–0.07 s).
#
Mean difference in single limb support time between genders is 0.04 s (p < 0.001, 95%CI = 0.03–0.05 s).
**
Mean difference in double support time between genders is 0.02 s (p = 0.05, 95%CI = 0.00–0.03 s).
yy
Mean difference in gait speed between genders is 7 cm/s (p = 0.002, 95%CI = 2.7–11.5 cm/s).
zz
Mean difference in step length between genders is 9.0 cm (p < 0.001, 95%CI = 7.0–11.0 cm); mean difference in stride length between genders is 17.5 cm (p < 0.001,
95%CI = 13.6–21.5 cm).
ôô
Mean difference in step width between genders is 2.2 cm (p < 0.001, 95%CI = 1.2–3.1 cm).
J.H. Hollman et al. / Gait & Posture 34 (2011) 111–118114
factor a ‘‘variability’’ factor. The fourth factor accounted for 11.6%
of the variance and loaded highly on gait speed, step length and
stride length. We labeled this factor a ‘‘pace’’ factor. The last factor
accounted for 6.9% of the variance in gait performance and loaded
highly on step width and step width variability. We labeled this
factor a ‘‘base of support’’ factor.
3.2. Gender and age group comparisons
Descriptive data for parameters constituting the five domains of
gait performance, by gender and age, are presented in Table 3.
3.2.1. Rhythm
The gender main effect was significant (F6,281 = 7.030,
p < 0.001), though neither the age group main effect
(F18,849 = 1.612, p = 0.051) nor gender  age group interaction
(F18,849 = 0.749, p = 0.783) were statistically significant. Men
walked at a lower cadence and with greater step times, stride
times, swing times, stance times, and single limb support times
than did women. Men walked at approximately 103 steps/min and
had mean step and stride times of 0.58 and 1.16 s, respectively;
mean swing and stance times of 0.42 and 0.74 s, respectively; and
mean single limb support times of 0.43 s. Women, in contrast,
walked at approximately 112 steps/min and had shorter mean step
and stride times of 0.54 s and 1.08 s, respectively; shorter mean
swing and stance times of 0.39 s and 0.69 s, respectively; and
shorter mean single limb support times of 0.39 s.
3.2.2. Gait cycle phases
In contrast to the rhythm domain, for the phasic parameters
both the gender (F5,282 = 4.981, p < 0.001) and age group main
effects (F15,852 = 1.750, p = 0.037) were significant, though the
gender  age group interaction was not (F15,852 = 0.956, p = 0.500).
On further analysis, only double limb support time differed
between genders (F1,286 = 4.796, p = 0.029). Double support time
was approximately 0.32 s in men and 0.30 s in women. Age group
effects (Fig. 2) were only present in the double limb support phase
(F3,286 = 3.199, p = 0.024) and double limb support time parameters
(F3,286 = 6.334, p < 0.001). Participants in the 85+ age group
spent longer periods in double limb support.
3.2.3. Variability
Neither the gender (F7,278 = 0.651, p = 0.713) nor age group
main effects (F21,840 = 1.466, p = 0.081) were statistically significant
for the variability measurements, nor was the gender  age
group interaction (F21,840 = 1.010, p = 0.448).
3.2.4. Pace
For the pace domain of gait performance, the gender
(F3,284 = 34.937, p < 0.001) and age group main effects
(F9,858 = 4.798, p < 0.001) were statistically significant, though the
gender  age group interaction was not (F9,858 = 1.076, p = 0.378).
Men walked faster and with longer step and stride lengths than did
women. Men walked at approximately 115 cm/s and had mean step
lengths of 66 cm and stride lengths of 133 cm. Women walked at
approximately 108 cm/s and had mean step lengths of 57 cm and
stride lengths of 115 cm. When normalized to height, however, the
gender difference in gait speed was negated (F1,286 = 1.516,
p = 0.219). Gait speed, step lengths and stride lengths were greater
in the70–74 and75–79thaninthe 80–84and85+age groups(Fig. 2).
3.2.5. Base of support
Similar to the rhythm domain analysis, the gender main effect
was statistically significant in the base of support domain
0
5
10
15
20
25
30
35
85-8980-8475-7970-74
Age Group (years)
DoubleSupport(%GC)
Male
Female
0
0.1
0.2
0.3
0.4
0.5
85-8980-8475-7970-74
Age Group (years)
DoubleSupportTime(s)
Male
Female
0
20
40
60
80
100
120
140
85-8980-8475-7970-74
Age Group (years)
GaitSpeed(cm/s)
Male
Female
0
20
40
60
80
100
120
140
160
85-8980-8475-7970-74
Age Group (years)
StrideLength(cm)
Male
Female
*
*
**
**
**
***
*** ***
*
***
**
BA
DC
Fig. 2. The phase and pace domains of gait performance had significant age group effects. The proportion of the gait cycle spent in double limb support (A) was greater in the
85+ year age group than in the 70–74 year age group. Double limb support time (B) was greater in the 85+ year age group than in the 70–74 year age group and in the 75–79
year age group. Gait speed (C) in the 80–84 and 85+ year age groups was slower than in the 70–74 and 75–79 year age groups. Stride lengths (D) and step lengths (not
illustrated because step length data are one-half the length of stride length data but show identical statistical comparisons) in the 85+ year age group were shorter than in the
70–74, 75–79 and 80–84 year age groups, and were shorter in the 80–84 year age group than in the 70–74 year age group. Note: * denotes Bonferroni-adjusted p < 0.05, **
denotes Bonferroni-adjusted p < 0.01, *** denotes Bonferroni-adjusted p < 0.001.
J.H. Hollman et al. / Gait & Posture 34 (2011) 111–118 115
(F2,284 = 8.669, p < 0.001), though neither the age group main
effect (F6,570 = 1.897, p = 0.079) nor gender  age group interaction
(F6,570 = 0.940, p = 0.466) were significant. Mean step width in men
was approximately 10.0 cm whereas mean step width in women
was approximately 7.9 cm. Step width variability did not differ
between men and women.
4. Discussion
While several studies have provided reference values for gait
[2,7–11], we believe the present study represents one of the most
comprehensive analyses of normative spatiotemporal gait data in
older adults. The breadth of gait parameters analyzed exceeds that
of previous studies and, secondarily, the extent to which
morbidities were controlled and participants excluded assured
that data represent gait performance in able-bodied participants.
Normative data for nearly 2 dozen spatiotemporal gait parameters
comprising five domains of gait performance in older adults are
provided from which results of other studies or clinical measures
can be compared.
The findings expand on the factors that Verghese et al.
described [4]. First, while Verghese et al. identified a rhythm
factor characterized by cadence and swing time, our study suggests
that several additional temporal parameters including step time,
stride time, stance time and single limb support time also load
highly on the factor. Second, while Verghese et al. identified a
variability factor that loaded highly on stride length variability, our
study suggests several parameters contribute to a variability
domain of gait performance. Gait cycle variability parameters
including variability in stride length, stride time, stance time,
swing time and stride speed all load highly on the factor, as do step
variability parameters including step length and step time
variability. Third, both studies identified a pace factor characterized
by gait speed, step length and stride length. We believe two
additional factors, however, also characterize gait performance: a
phasic factor characterized by temporophasic proportions of the
gait cycle spent in stance, swing and single and double limb
support; and a base of support factor characterized by step width
and step width variability.
Parameters in the pace domain have been studied most
extensively among the domains identified in the present study
and provide for readily available comparisons. Gait speed, for
example, has been recommended as a ‘‘vital sign’’ for physical
performance in older persons [1,3] and 10 cm/s decreases in gait
speed are associated with higher falls risk in older persons [13]. It is
important that researchers and clinicians understand normative
values for pace parameters in older adults. Mean gait speed
(110 Æ 19 cm/s, 95%CI = 107.8–112.2 cm/s) and stride length measurements
(122 Æ 19 cm, 95%CI = 119.8–124.2 cm) reported in our
study are largely consistent with other studies reporting reference
values for preferred gait speed in older adults. For example, Oh-Park
et al. [10], using similar methods and instrumentation to collect gait
data and using similarly aged subjects, reported an overall mean gait
speed of 106 cm/s and a mean stride length of 121 cm in 304 robust,
healthy older individuals. Callisaya et al. [8], also using similar
methods and instrumentation, reported slightly greater mean gait
speed (116 cm/s in men and 111 cm/s in women) and stride length
measurements (approximately 130 cm in men and 115 cm in
women) in a population-based sample of 411 older adults; those
differences, however, may be accounted for because they included
subjects aged 60–69 that were not included in our study or in the OhPark
study. Given the relative similarities between studies, we believe
our findings represent valid target values of pace parameters in
persons over the age of 70 for assessing abnormal gait.
Regarding gender differences in the pace domain, most studies
indicate that gait speed and stride length in men exceeds that of
women [8,10,11]. Our findings were similar. Gait speed in men
exceeded that in women by approximately 7 cm/s (95%CI = 2.7–
11.5 cm/s) and stride length in men exceed that in women by
approximately 18 cm (95%CI = 13.6–21.5 cm). When gait speed
was normalized to height, however, the difference between
genders was no longer significant, indicating that gait speed is
more a function of height than it is gender. Interestingly, when
stride length and cadence were also normalized to height, the
gender differences remained. Men tended to walk with longer
strides but with lower cadences.
In cross-sectional comparisons, most studies indicate that gait
speed and stride length decrease with age. Himann et al. [14]
reported that gait speed decreases 12–16% per decade after the age
of 70. Oh-Park et al. [10] reported similar reductions in gait speed
and stride length in a population-based sample of people over the
age of 70, although those reductions were minimized in a
subsample characterized as robust, healthy individuals. Our
cross-sectional findings suggest that parameters in the pace
domain of gait are equivalent among persons in the 70–74 and 75–
79 year age groups, but decrease significantly past the age of 80
(Fig. 2). The functional impact becomes evident when one realizes
that walking at 120 cm/s is necessary to navigate typical urban
crosswalks [15]. Among participants aged 70–79 years, 42%
walked at a speed that exceeded 120 cm/s. In the 80–84 and
85+ year age groups, however, those proportions dropped to 18%
and 15%, respectively. These data suggest that, even in good health,
some decline in gait performance is inevitable in older persons and
that decline is likely to produce functional deficits.
A rhythm factor accounted for 25.8% of the variance in gait
performance. It loaded highly on cadence and temporal parameters
including step, stride, swing, stance and single limb support time.
Rhythm measurements reported in the present study are largely
consistent with other studies reporting reference values in older
adults. The mean cadence (109 Æ 13 steps/min), for example, is
slightly greater but similar to the cadence of 105.2 steps/min that OhPark
et al. reported [10]. Mean cadences for men (103 steps/min) and
women (112 steps/min) are similar to values (107.2 and 114.6 steps/
min, respectively) that Callisaya et al. reported [8]. Likewise, the mean
swing time (0.40 Æ 0.05 s) is similar to the 0.43-s swing time that OhPark
et al. reported [10]; and step times (0.58 Æ 0.05 s in men,
0.54 Æ 0.06 s in women) are similar to step times (0.56 and 0.53 s,
respectively) that Callisaya et al. reported [8]. Quantifying rhythm
and understanding normative values for parameters representing the
rhythm domain are important because changes in rhythm may be
associated with increased risk of dementia and memory decline [4].
A variability factor accounted for 19.1% of the variance in gait
performance in our study. Comparing variability values with other
studies is hindered because of inconsistencies between studies that
use the CV versus the SD to report variability. When our data are
expressed as the SD, stride length variability (4.4 Æ 4.2%CV, or
5.0 Æ 4.4-cm SD) and swing time variability data (6.6 Æ 9.3%CV, or
0.03 Æ 0.03-cm SD) are very similar to the 4.6-cm and 0.03-cm
variability values, respectively, that Verghese et al. reported [4].
Likewise, stance time variability (5.3 Æ 4.5%CV, or 0.037 Æ 0.030-s SD)
is nearly identical to the 0.038-s stance time variability that Brach et al.
reported [16]. Our step variability data demonstrate similar consistencies
with other studies. Step length variability (5.9 Æ 6.0%CV, or
3.3 Æ 2.7-cm SD) is similar in magnitude to the 6.4%CV step length
variability that Brach et al. reported [17] and to the 2.8-cm variability
that Callisaya reported [8]. Step time variability (5.6 Æ 5.3%CV, or
0.03 Æ 0.02-s SD) is similar in magnitude to the 0.02-s step time
variability data that Callisaya reported. Given the relative similarities
between studies, we believe our findings represent valid target
variability parameters for assessing abnormal gait.
Increased variability reflects a loss of automaticity in walking
that presumably makes persons more susceptible to falling.
J.H. Hollman et al. / Gait & Posture 34 (2011) 111–118116
Variability parameters have been used to quantify falls risk, to
assess risk of dementia and, under dual task conditions, to examine
executive functioning in aging adults [4,5,18,19]. Variability
parameters used to examine these capacities have included
variability in stride speed [20,21], stride time [18], stride length
[4], step time [22], step length [22], swing time [4,23] and step
width [22,24]. Few studies have examined multiple modes of
variability collectively, presumably because investigators assumed
that each parameter, be it a marker of gait cycle variability, step
variability or base of support variability, represents a similar
construct. Some researchers have challenged that assumption,
reporting that stride time variability and stride length variability,
for example, respond differently to speed of walking and to dual
task walking [25]. Findings from our study suggest that each of the
gait cycle and step variability parameters, apart from stride width
variability, represent one variability construct.
A phase factor representing temporophasic divisions of the gait
cycle was a fourth factor identified in our study. While a phase
factor was not previously identified in a factor analysis [4],
considering gait phases is relevant. A 10% decrease in the swing
phase proportion of the gait cycle, for example, may be associated
with higher falls risk in older persons [13]. In our study, the stance
phase represented 63.9% (95%CI = 63.6–64.2%GC) and the swing
phase represented 36.1% of the gait cycle (95%CI = 35.8–36.4%GC),
values that are consistent with the stance (63%) and swing (37%)
phase proportions in healthy older men established by Murray
et al. [26]. Double limb support occupied 27.9% (95%CI = 27.3–
28.5%GC) and single limb support occupied 36.2% of the gait cycle
(95%CI = 35.8–36.6%GC). To our knowledge, normative values for
all of these gait cycle phases have not been established for older
adults. The values differ slightly from norms established by the
Rancho Los Amigos (RLA) National Rehabilitation Center (stance
phase 62%, swing phase 38%, double support 24% and single
support 38% [27]), but that likely represents the fact that older
participants in our study walked more slowly (110 cm/s) than
presumably younger subjects represented in the RLA data, who
walked at an average of 136 cm/s.
Last, a base of support factor characterized by step width and
step width variability accounted for 6.9% of the variability in gait
performance. Mean step width was 10.0 cm in men (95%CI = 9.5–
10.5 cm) and 7.9 cm in women (95%CI = 7.5–8.3 cm), magnitudes
similar to population-based step width data reported elsewhere
[8]. Mean step width variability, which did not differ between
genders, was 3.2 cm (95%CI = 2.9–3.5 cm), approximately 1 cm
greater than that reported elsewhere [8]. The influence of the base
of support factor associated with aging is equivocal. Step width
variability may distinguish older from younger adults [22] and
may be associated with peripheral sensory impairments that
contribute to gait dysfunction [28]. Others have characterized
increased step width as an adaptation by aging persons who are
fearful of falling, but reported that step width itself is not a
predictor of falling [29]. In contrast, reduced and increased step
width variability have been associated with falls histories [24].
While the association of step width and step width variability with
aging and fall risk is not certain, findings of the present study
indicate that examining a base of support factor is warranted.
Several limitations in the study should be considered. First,
subjects walked at self-selected speeds, not at maximum speeds.
Interpretations of the findings should not be generalized to one’s
ability to walk faster than at self-selected speeds. Second, this
study represents a secondary analysis of a population-based cohort
study designed to investigate the prevalence, incidence and risk
factors for dementia [12]. The primary study screened a variety of
cardiovascular, neurological and other medical conditions like
diabetes and cancer, but data on specific orthopedic or rheumatologic
conditions that might influence gait were not collected. While
the findings represent normative data for able-bodied older adults,
the results may mis-represent gait parameters in older adults in
whom specific orthopedic or rheumatologic conditions are
definitively absent. Last, data for the study were collected from
an average of 14 Æ 4 steps (7 strides) per participant. While there is
evidence that measuring pace and rhythm parameters is highly
reliable with even small numbers of strides, reliability for other gait
domains, particularly variability, may be enhanced by collecting data
from hundreds of strides [30]. Validity of that data may be
questioned.
5. Conclusion
Data for 23 spatiotemporal gait parameters were collected from
subjects aged 70+. Normative values are presented which
researchers or clinicians can use for assessing and interpreting
gait dysfunction. Based on factor analysis, five domains of gait
performance were identified. A rhythm domain was characterized
by cadence and temporal parameters (e.g., stride time). A phase
domain was characterized by temporophasic divisions of the gait
cycle. A variability domain encompassed gait cycle variability
characterized by variability in stride measurements, and step
variability characterized by variability in step measurements. A
pace domain was characterized by gait speed, step length and
stride length. A base of support domain was characterized by step
width and step width variability. Several gender and age group
differences were present among the domains. We recommend that
data representing each of these domains be collected when gait
analyses are conducted.
Acknowledgments
This work was supported by grants from the National Institute
on Aging (P50 AG016574, U01 AG006786 and R01 AG034676), and
by the Robert H. and Clarice Smith and Abigail Van Buren
Alzheimer’s Disease Research Program. Study sponsors had no
involvement in the study design; in the collection, analysis or
interpretation of data; in the writing of the manuscript; nor in the
decision to submit the manuscript for publication. The authors
thank Matthew R. Miller for his assistance with data collection and
organization.
Conflict of interest
None of the authors has a financial or personal relationship with
people or organizations that could inappropriately influence this
work.
References
[1] Ferrucci L, Baldasseroni S, Bandinelli S, de Alfieri W, Cartei A, Calvani D, et al.
Disease severity and health-related quality of life across different chronic
conditions. J Am Geriatr Soc 2000;48:1490–5.
[2] Cesari M, Kritchevsky SB, Penninx BWHJ, Nicklas BJ, Simonsick EM, Newman
AB, et al. Prognostic value of usual gait speed in well-functioning older
people—results from the health, aging and body composition study. J Am
Geriatr Soc 2005;53:1675–80.
[3] Studenski S, Perera S, Wallace D, Chandler JM, Duncan PW, Rooney E, et al.
Physical performance measures in the clinical setting. J Am Geriatr Soc
2003;51:314–22.
[4] Verghese J, Wang C, Lipton RB, Holtzer R, Xue X. Quantitative gait dysfunction
and risk of cognitive decline and dementia. J Neurol Neurosurg Psychiatry
2007;78:929–35.
[5] Maki BE. Gait changes in older adults: predictors of falls or indicators of fear. J
Am Geriatr Soc 1997;45:313–20.
[6] Studenski S, Perera S, Patel K, Rosano C, Faulkner K, Inzitari M, et al. Gait speed
and survival in older adults. JAMA 2011;305:50–8.
[7] Bohannon RW. Population representative gait speed and its determinants. J
Geriatr Phys Ther 2008;31:49–52.
J.H. Hollman et al. / Gait & Posture 34 (2011) 111–118 117
[8] Callisaya ML, Blizzard L, Schmidt MD, McGinley JL, Srikanth VK. Ageing and
gait variability—a population-based study of older people. Age Ageing
2010;39:191–7.
[9] Oberg T, Karsznia A, Oberg K. Basic gait parameters: reference data for normal
subjects, 10–79 years of age. J Rehabil Res Dev 1993;30:210–23.
[10] Oh-Park M, Holtzer R, Xue X, Verghese J. Conventional and robust quantitative
gait norms in community-dwelling older adults. J Am Geriatr Soc
2010;58:1512–8.
[11] Samson MM, Crowe A, de Vreede PL, Dessens JA, Duursma SA, Verhaar HJ.
Differences in gait parameters at a preferred walking speed in healthy
subjects due to age, height and body weight. Aging Clin Exp Res
2001;13:16–21.
[12] Roberts RO, Geda YE, Knopman DS, Cha RH, Pankratz VS, Boeve BF, et al. The
Mayo clinic study of aging: design and sampling, participation, baseline
measures and sample characteristics. Neuroepidemiology 2008;30:58–69.
[13] Verghese J, Holtzer R, Lipton RB, Wang C. Quantitative gait markers and
incident fall risk in older adults. J Gerontol A Biol Sci Med Sci
2009;64:896–901.
[14] Himann JE, Cunningham DA, Rechnitzer PA, Paterson DH. Age-related changes
in speed of walking. Med Sci Sports Exerc 1988;20:161–6.
[15] Langlois JA, Keyl PM, Guralnik JM, Foley DJ, Marottoli RA, Wallace RB. Characteristics
of older pedestrians who have difficulty crossing the street. Am J
Public Health 1997;87:393–7.
[16] Brach JS, Perera S, Studenski S, Katz M, Hall C, Verghese J. Meaningful
change in measures of gait variability in older adults. Gait Posture
2010;31:175–9.
[17] Brach JS, Berthold R, Craik R, VanSwearingen JM, Newman AB. Gait variability
in community-dwelling older adults. J Am Geriatr Soc 2001;49:1646–50.
[18] Hausdorff JM, Rios DA, Edelberg HK. Gait variability and fall risk in community-living
older adults: a 1-year prospective study. Arch Phys Med Rehabil
2001;82:1050–6.
[19] Sheridan PL, Solomont J, Kowall N, Hausdorff JM. Influence of executive
function on locomotor function: divided attention increases gait variability
in Alzheimer’s disease. J Am Geriatr Soc 2003;51:1633–7.
[20] Beauchet O, Kressig RW, Najafi B, Aminian K, Dubost V, Mourey F. Age-related
decline of gait control under a dual-task condition. J Am Geriatr Soc
2003;51:1187–8.
[21] Hollman JH, Kovash FM, Kubik JJ, Linbo RA. Age-related differences in spatiotemporal
markers of gait stability during dual task walking. Gait Posture
2007;26:113–9.
[22] Owings TM, Grabiner MD. Step width variability, but not step length variability
or step time variability, discriminates gait of healthy young and older
adults during treadmill locomotion. J Biomech 2004;37:935–8.
[23] Hausdorff JM, Schweiger A, Herman T, Yogev-Seligmann G, Giladi N. Dual-task
decrements in gait: contributing factors among healthy older adults. J Gerontol
A Biol Sci Med Sci 2008;63:1335–43.
[24] Brach JS, Berlin JE, VanSwearingen JM, Newman AB, Studenski SA. Too much or
too little step width variability is associated with a fall history in older persons
who walk at or near normal gait speed. J Neuroeng Rehabil 2005;2:21.
[25] Beauchet O, Dubost V, Gonthier R, Kressig RW. Dual-task-related gait changes
in transitionally frail older adults: the type of the walking-associated cognitive
task matters. Gerontology 2005;51:48–52.
[26] Murray MP, Kory RC, Clarkson BH. Walking patterns in healthy old men. J
Gerontol 1969;24:169–78.
[27] Department of Pathokinesiology and Physical Therapy. Observational Gait
Analysis Handbook. Downey, CA: Los Amigos Research and Education Institute,
Inc., Rancho Los Amigos National Rehabilitation Center; 2001.
[28] Brach JS, Studenski S, Perera S, VanSwearingen JM, Newman AB. Stance time
and step width variability have unique contributing impairments in older
persons. Gait Posture 2008;27:431–9.
[29] Chamberlin ME, Fulwider BD, Sanders SL, Medeiros JM. Does fear of falling
influence spatial and temporal gait parameters in elderly persons beyond
changes associated with normal aging? J Gerontol A Biol Sci Med Sci
2005;60:1163–7.
[30] Hollman JH, Childs KB, McNeil ML, Mueller AC, Quilter CM, Youdas JW.
Number of strides required for reliable measurements of pace, rhythm and
variability parameters of gait during normal and dual task walking in older
individuals. Gait Posture 2010;32:23–8.
J.H. Hollman et al. / Gait & Posture 34 (2011) 111–118118