Subjects
All subjects gave informed consent prior to participation in
the study, which was approved by the hospital's institutional review
board. Fifteen healthy women (nine primigravida and six multigravida)
were studied in the second half of the last trimester of pregnancy,
between thirty-five and forty weeks of gestation. All but two were
studied again at least one year post partum to obtain comparative
data on nonpregnant women, with the assumption that changes that
occur during pregnancy disappear by one year post partum. The remaining
two subjects were studied prior to the onset of pregnancy instead
of one year post partum. All data was collected between June 1994 and
June 1998. The data sets for two additional subjects were not included
in this study due to the presence of excessive anterior abdominal
adipose tissue, which may have compromised marker placement over
the anterior superior iliac spines. The mean age of the fifteen
subjects was thirty-two years (range, twenty-five to thirty-eight years);
the mean height, 1.67 meters (range, 1.58 to 1.82 meters); and the
mean body mass, 63.1 kilograms (range, 42.7 to 90.5 kilograms).
A mean increase in body mass of 13.0 kilograms (standard deviation,
4.2 kilograms) and increase in body height of 1.2 centimeters (standard
deviation, 1.8 centimeters) was measured during pregnancy.
Data Collection and Processing
Reflective spherical markers were taped on the skin over a point
midway between the posterior superior iliac spines, over the anterior
superior iliac spines, on the lateral side of each knee at the joint center,
on each lateral malleolus, and between the distal ends of the second
and third metatarsals of each foot. In addition, reflective spherical
markers on the end of short wands were strapped around the thighs
and calves with Velcro and elastic straps. For two subjects, a triad
of markers was placed on the trunk (one distal to each clavicle, spaced
equidistant from the sternal notch, and one on the seventh cervical
vertebra). Lower-extremity kinematic and kinetic data (six to eight
strides) was collected with use of a three-dimensional motion-analysis
system with six sixty-hertz video cameras (Vicon; Oxford Metrics
Limited, Oxford, England) and one force platform (sampling rate
of 600 hertz; AMTI, Newton, Massachusetts) as the subjects performed
repeated walks at a self-selected pace across the room (approximately
twelve meters in distance).
Lower-extremity joint angles, net joint moments, and net joint
powers during the gait cycle were calculated with use of Vicon Clinical
Manager software (Oxford Metrics Limited). The following angles
were analyzed: pelvic tilt, hip flexion and extension, knee flexion
and extension, and ankle dorsiflexion and plantar flexion in the
sagittal plane; pelvic obliquity and hip abduction and adduction
in the coronal plane; and pelvic rotation and foot progression (the
longitudinal axis of the foot relative to the line of progression)
in the transverse plane. Internal joint moments, which are presented
throughout this paper, were calculated as equal and opposite in
magnitude to the externally applied moments measured by the force
platform and kinematic data. Internal joint moments represent the
body's response to external loads. Net joint moments are the sum
of all moments that tend to extend the joint and those that tend
to flex the joint and include components due not only to muscle
forces but also to other soft-tissue structures that cross the joint.
Net joint power, calculated as the scalar product of net joint moment
and joint angular velocity, indicates the rate at which a moment
is rotated. Positive and negative values represent net power generation
and absorption at the joint, respectively. Theoretically, power
generation is associated with concentric muscle contraction, and
power absorption is associated with eccentric muscle contraction.
However, due to biomechanical modeling limitations at the time of this
study, we could calculate only net power values, which represent
the net effect of all soft-tissue and skeletal-muscle structures
acting about the joint rather than the effect of each specific muscle acting
at the joint. Joint moments and powers were assessed for the hip,
knee, and ankle in the sagittal plane and for the hip in the coronal
plane. Moments and powers were compared between experimental conditions
both unnormalized and normalized to body mass to assess the effects
of increased body mass during pregnancy.
Trunk tilt, obliquity, and rotation were computed with use of
custom-written software for the two subjects from whom that data
was collected. These trunk variables were computed relative to a
laboratory coordinate system (aligned with the line of progression
during walking). Trunk tilt was measured as the angle of forward
or backward inclination of the trunk relative to vertical (trunk orientation
was defined by the line segment running from the midpoint of the
triad of upper-trunk markers to the midpoint of the triad of pelvic markers),
obliquity was measured as the angle of sideward inclination of the
trunk relative to vertical (defined with use of the same line segment
as was used for trunk tilt), and rotation was measured as the angle
of rotation of the upper trunk relative to the line of progression
(upper-trunk orientation was defined by the line segment running
from the midpoint of the triad of upper-trunk markers to the marker
over the seventh cervical vertebra).
The normalized dynamic base-of-support width was calculated as
the mean width between the ankle joint centers during double support
(measured in the direction perpendicular to the line of progression)
divided by the pelvic width (measured as the width between the anterior
superior iliac spines). The base-of-support width was normalized
to pelvic width in an attempt to account for any increases in the
distance between the centers of the hip joints during pregnancy.
Pelvic width rather than the actual distance between the hip joint
centers was measured because pelvic width could be directly measured,
whereas, in current hip joint models, the locations of the hip joint
centers are crudely approximated in reference to the locations of
the anterior superior iliac spines.
Statistical Analysis
Maxima and minima, ranges of motion, or mean values over a portion
of the gait cycle for the kinematic and kinetic data were compared
between conditions with use of paired t tests at a 95 percent significance
level. Post hoc power analysis was performed for
all variables for which significant differences were not detected. Calculations
were done to determine the difference in parameters that could be
detected between experimental conditions with at least 80 percent power
for the number of subjects available for study.
Time and Distance Parameters
No significant differences were found in walking velocity (p
= 0.071), stride length (p = 0.082), or cadence (p = 0.147) between
experimental conditions (Table I). Power analysis revealed that
small changes (differences of 4 to 8 percent in magnitude) in these parameters
could be detected with 80 percent power. Changes that were small
(1 and 2 percent of the gait cycle) but significant were found in
single (p = 0.019) and double (p = 0.020) support times between
the experimental conditions.
Kinematic and Kinetic Data
Ensemble-averaged (the mean of all trials for all subjects) sagittal-plane
kinematic and kinetic variables (Fig. 1) and coronal and transverse-plane
kinematic and kinetic variables (Fig. 2) for each condition were generated
to display the data. During pregnancy, significant increases in maximum
anterior pelvic tilt (p = 0.018), maximum hip flexion (p = 0.004),
and stance-phase hip adduction (p = 0.012) during walking were found
(Table II).
Compared with the condition one year post partum, maximum anterior
pelvic tilt during pregnancy increased a mean of 4 degees overal;
it increased by 5 degrees or more for nine subjects, was similar
for five subjects, and decreased by more than 5 degrees for only
one subject (Table III). Trunk tilt, obliquity, and rotation
angles throughout the gait cycle during pregnancy were similar to
those measured one year post partum for both subjects from whom
the data were collected. No statistical analyses were performed
on the trunk data due to the small sample size. No significant changes
in pelvic obliquity range of motion (p = 0.073), pelvic rotation
range of motion (p = 0.905), or foot progression angle during stance
(p = 0.525) were found. Post hoc power analysis
revealed that changes of 5 degrees or less (that is, a clinically
relevant level) could have been detected with at least 80 percent
power for all kinematic variables. For the unnormalized kinetic
data, pregnancy was associated with significant increases in maximum
hip extension moment (p = 0.034), time for reversal from a hip extension to
a hip flexion moment (p = 0.008), maximum hip power generation in
the sagittal plane (p = 0.013), maximum hip abduction moment (first
and second peak values during stance, p < 0.001), maximum hip
power generation in the coronal plane (p = 0.007), maximum ankle
plantar flexion moment (p = 0.005), and maximum ankle plantar flexion power
absorption (p = 0.003) (Table IV). After normalization to body weight,
no significant differences (p > 0.05) were found for any of these
kinetic parameters with two exceptions: normalized hip abduction
moment during stance (first peak only) was still significantly increased
(p = 0.022) during pregnancy, and normalized maximum ankle plantar
flexion moment was actually significantly decreased (p = 0.002)
during pregnancy. There was adequate power (80 percent or greater)
to detect clinically important differences for all kinetic variables
for which significant differences were not found (Table IV). For all
of these parameters, the difference that could be detected was smaller
than a clinically important difference.
Base of Support
Significant increases in both pelvic width (4.3 centimeters,
p < 0.001) and mean ankle separation width (2.4 centimeters,
p < 0.001) during double support were found during pregnancy
(Fig. 3).
However, the ratio of the base-of-support width to the pelvic width
(0.68 during pregnancy; 0.70 one year post partum) remained constant
(p = 0.319). Power analysis revealed adequate power (80 percent
or greater) to detect clinically important changes in these parameters.
(Differences of 2.2 centimeters for pelvic width, 1.7 centimeters for
ankle separation width, and 0.05 for normalized base-of-support
width could be detected.)
The results of this study indicate that, kinematically, gait
during pregnancy is remarkably unchanged. Velocity, stride length,
and cadence during the third trimester of pregnancy were similar
to those measured one year post partum, and only small deviations
in pelvic tilt and hip flexion, extension, and adduction were observed
during pregnancy. A so-called waddling gait, consisting of increases
in the normalized dynamic base of support, the external foot progression
angle, pelvic obliquity, and pelvic rotation, was not documented during
pregnancy in this study. The increase in the width of the pelvis
during pregnancy, which could be considered a primary deviation,
was apparently compensated for by an increase in maximum hip adduction
during stance. With increased pelvic width, an increased hip adduction
angle during single support keeps the foot centered under the body
to avoid a wide base of support. Walking with a wide base of support
results in large side-to-side excursions of the center of mass and
is energy-inefficient.
It is not possible to draw general conclusions from the kinematic
data about how back posture during gait is altered during pregnancy.
Although an overall significant increase in anterior tilt of the pelvis
during pregnancy was found (p = 0.018), the changes in pelvic tilt
varied among the pregnant subjects, suggesting that not all women
have the same postural alterations. Six of the fifteen subjects
had either similar or decreased anterior pelvic tilt during pregnancy.
Typically, increased lumbar lordosis is associated with increased
anterior pelvic tilt and occurs to keep the upper body in an upright
position. In the two subjects from whom trunk data was obtained,
trunk tilt was not altered during pregnancy (pelvic tilt was increased
in one and unchanged in the other), suggesting that lumbar lordosis
was increased for the woman who had increased anterior pelvic tilt.
However, the slight increase in body height (a mean of 1.2 centimeters)
found in this study suggests that lumbar lordosis was decreased.
This change in height is similar to the mean 1.0-centimeter increase
during pregnancy found by Snijders et al.22 in
a study of thirty-four women two weeks prior to delivery and then
two weeks afterward. This height increase was attributed to diminished
lumbar lordosis during pregnancy. Studies of static postural adaptations
during pregnancy have also revealed individual variation in response
to the added load5,18,22. Furthermore,
although low-back pain during pregnancy is speculated to be linked
to an increase in lumbar lordosis, several researchers have failed to
find a significant association (p > 0.05) between increased lumbar
lordosis during pregnancy and low-back pain5,12.
Additional studies, incorporating improved dynamic analysis of trunk
and pelvic motion in a greater number of pregnant subjects, are
necessary to identify the most common primary deviations and compensatory
trunk alignment strategies utilized by pregnant women.
There were several alterations in kinetic parameters during pregnancy
that appear to reflect compensations utilized to maintain normal
gait despite substantial increases in body mass and an anterior shift
in the center of gravity. These alterations include increases in
hip moment and power in the coronal and sagittal planes as well
as ankle moment and power in the sagittal plane. These changes in
kinetic parameters suggest an increased use of hip abductor, hip
extensor, and ankle plantar flexor muscle groups. Overuse of these
muscle groups during pregnancy may be a contributing factor to low-back,
pelvic, and hip pain as well as painful muscle cramps in the calf
or other parts of the lower extremity. Women who are inactive or have
low muscle strength, or both, may be particularly susceptible to
these overuse conditions during pregnancy. Studies should be done
to investigate the relationship between poor muscle strength and
the prevalence of musculoskeletal disorders during pregnancy.
Increased hip abductor power during pregnancy is consistent with
an increased use of hip abductor muscles to maintain normal gait
with increased body mass. In support of this theory is the finding that
when power generation in the coronal plane was normalized to body
mass it was not significantly different (p = 0.192) during pregnancy.
It has been suggested that hip power generation in the coronal plane
during late stance is a measure of the power of the stance-side
hip abductor muscles used to raise the unsupported side of the pelvis9. With increased body mass during
pregnancy, an increased load must be lifted by the hip abductor muscles.
Significant increases in double-support time (p = 0.020) and decreases
in single-support time (p = 0.019) during pregnancy may be fine-tuning
compensations to minimize the time spent in single-limb support
when this increased muscular effort is required to support an increased
body mass with only one limb.
The observed magnitude and timing changes in hip kinetic patterns
in the sagittal plane with pregnancy are consistent with increased
body mass and a shift toward a more anterior center-of-mass location.
The magnitude of the internal hip extension moment during early
stance is potentially increased during pregnancy by both an increased body
mass and an anterior shift in the center-of-mass position relative
to the hip joint center (increasing the body weight vector moment
arm about the hip). After normalization of maximum hip extension
moment and maximum hip power generation in the sagittal plane to
body mass, no significant differences (p = 0.646 and p = 0.250, respectively)
between pregnancy and one year post partum were detected. This provides
evidence that the magnitude of changes in hip extension moment and
power are due mainly to the effect of body mass rather than to a
shift in the position of the center of mass. In contrast, normalizing
for body mass had no effect on the delayed hip extension to flexion
moment reversal time during pregnancy, providing evidence that the
delayed reversal time is due solely to the anterior shift in the
center-of-mass position. It is surprising that normalization for
body mass alone (without accounting for anterior shift in the center-of-mass position)
explained the increases in hip flexion moment and power magnitudes.
It is not clear how the pregnant women were able to compensate for the
anterior shift in the center-of-mass position to prevent further
increases in hip extension moment over and above the elevated level
due to increased body mass alone.
Maximum ankle power absorption is a measure of the amount of
eccentric muscle work being done by the ankle plantar flexors to
control the rotation of the tibia over the foot as the ankle dorsiflexes during
stance. Maximum ankle plantar flexion moment occurs following midstance
during single support as active plantar flexion at the ankle occurs
to support and advance the body forward. Increased sagittal-plane
maximum ankle power absorption and maximum ankle plantar flexion moment
during gait are both consistent with increased use of the ankle
plantar flexor muscle group due to increased body weight during
pregnancy. Normalized for body weight, these parameters were similar
(maximum power absorption) or minimally reduced (maximum moment)
during pregnancy compared with one year post partum, suggesting
that the elevated levels were due to increased body weight. In addition
to being related to hormonal and metabolic factors, calf cramps during
pregnancy may be related to the increased functional demand placed
on the ankle plantar flexors.
When the results of this study are considered, the possibility
of imprecise pelvic marker placement should be taken into consideration
as a source of potential error. Although all measures were taken to
ensure that the anterior superior iliac spine markers were placed
on the pelvis to accurately reflect actual pelvic tilt and to serve
as reference points for calculations of the locations of the hip joint
centers, the abdominal contour in pregnancy may have compromised
this placement. Artifact caused by either anterior or lateral misplacement of
the anterior superior iliac spine markers is a possible explanation
for some of the significant gait deviations that were found at the
pelvis and hip during pregnancy. Women with excessive adipose tissue
obscuring the pelvic landmarks were excluded from the study to reduce
this error. Investigators performing future studies should consider
the use of alternative strategies for pelvic marker placement in
order to minimize this potential source of error.
In conclusion, despite major anatomical changes associated with
pregnancy, the kinematics of gait during pregnancy were found to
be remarkably unchanged. However, significant increases in hip extensor,
hip abductor, and ankle plantar flexor kinetic gait parameters were
found (p < 0.05). The data suggests an increased use of hip
extensor, hip abductor, and ankle plantar flexor muscles to compensate
for increases in body mass and changes in body-mass distribution
during pregnancy to keep speed, stride length, cadence, and joint
angles relatively unchanged. These compensations may result in overuse
injuries to the muscle groups about the pelvis, hip, and ankle,
contributing to low-back, pelvic, and hip pain; calf cramps; and
other painful lower-extremity musculoskeletal conditions associated
with pregnancy. The results of this study, when considered in conjunction
with those in other reports documenting improved clinical outcomes
for women who remain physically fit during pregnancy, support the
clinician's recommendation of appropriate exercise and conditioning
programs during pregnancy in order to avoid overuse injury to specific
muscle groups14,17,20. In addition,
pregnant women with established musculoskeletal problems may benefit
from appropriate pharmacological control of inflammation4 and rehabilitation to restore muscle
tone and strength.