Work 32 (2009) 351–360 351

DOI 10.3233/WOR-2009-0833
IOS Press

Effect of backpack load placement on posture

and spinal curvature in prepubescent children
Heather M. Brackley, Joan M. Stevenson ∗ and Jessica C. Selinger
School of Kinesiology and Health Studies, Queen’s University, Kingston, Ontario, Canada

Received 16 January 2008

Accepted 24 May 2008

Abstract. Parents, educators and researchers have expressed concern about the long term impacts of children carrying excessive
loads in their backpacks on a daily basis. Although many researchers have investigated appropriate weight limits for children’s
packs, little research has been conducted on the design of children’s backpacks. The purpose of this study was to evaluate
the changes in children’s trunk forward lean (TFL), cranio-vertebral angle (CVA) and spinal lordosis angle (LA) that occurred
with high, medium and low load locations during standing and walking. Ten-year-old children (n = 15) completed a repeated
measures designed study while carrying 15% of each child’s body weight in a typical backpack with only shoulder straps. A
special instrumented backpack (IBP) was designed that allowed the weight to be placed in the proper location and continuously
measure changes in spinal curvature. TFL and CVA postures were captured on digital video at five intervals including: standing
without a backpack prior to a 1000 m walk; standing with a backpack at the beginning and end of a 1000 m walk; and walking
with a backpack at the beginning and end of a 1000 m walk. Results indicated that significant changes occurred in TFL and CVA
when the backpack was loaded to 15% body weight. The low load placement in the backpack produced fewer changes in CVA
from the initial standing baseline measure than the high and mid placements. When all measures were assessed collectively, there
were fewer changes in LA in the low load placement. These findings indicate that future backpack designs should place loads
lower on the spine in order to minimize children’s postural adaptations.

Keywords: Load carriage, backpack design, children, posture, gait, biomechanics

1. Introduction In an attempt to determine a safe weight limit for chil-

Within developed nations, backpack use amongst
schoolchildren has become the most popular means of
transporting belongings to and from school. Howev-
er, there is a growing public concern that overloaded
children’s and adolescent’s backpacks may lead to the
development of back pain and other musculoskeletal
injuries. Some cross-sectional studies link backpack
use with back pain [10,17,18,24,31] and other studies
indicate that carrying excessive weight [29,32] or ex-
periencing fatigue [10,23] while carrying a backpack
may be linked with back pain and rucksack palsy.
∗ Address for correspondence: Dr. Joan M Stevenson, School of
Kinesiology and Health Studies, Room 158, PEC, Queen’s Univer-
sity, Kingston, Ontario, Canada, K7L 3N6. Tel.: +1 613 533 6288;
Fax: +1 613 533 2009; E-mail: stevensj@post.queensu.ca.
dren’s backpacks, several studies have examined the
effects of increasing backpack loads on physiological
parameters, including heart rate, VO 2 , blood pressure
and respiration [14,16,19,20,22],as well as biomechan-
ical parameters, including trunk forward lean, head on
neck posture and gait [7,12–15,20]. Based on a crit-
ical review, which took into account both physiologi-
cal and biomechanical findings, Brackley and Steven-
son [6] recommend that backpacks should weight no
more than 10–15% of a child’s body weight. This con-
clusion is consistent with recommendations set forth
from various health organizations, such as, the America
Occupational Therapy Association [2] and the Ontario
Chiropractic Association [3].
Although the current accepted weight limit appears
to be rather well established, the applicability of this

1051-9815/09/$17.00  2009 – IOS Press and the authors. All rights reserved
352 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children
generic limit is questionable given the lack of investi-
gation and reporting of children’s backpack design [6].
This scarcity of research is concerning, particularly in
light of the fact that abundant research exists into larger
adult rucksack designs for military and hiking purpos-
es. Research into these adult packs strongly suggests
that improved pack design and fit greatly influences the
user’s performance, comfort and health [28]. These
findings highlight the fact that preventing back pain
and injury may be best accomplished not only through
children’s compliance with suggested load limits, but
also though industry’s improvement of current pack
designs.
One of the first and most crucial pack design deci-
sions is load placement. The location of the load must
be established before other aspects of the design, such
as frame sheets, shoulder straps and hip belts, can be
applied. Without knowing where to place the load’s
center of mass to minimize postural adaptations and fa-
tigue, altering other design features is premature. Only
two studies were found that investigated the effects of
load placement on children’s posture [8,11]. Grimmer
at al. [11] examined the effects of different loads and
load placement on horizontal displacement of various
body segments in the sagittal plane. It was found that
when the load was placed low on the back (L3) horizon-
tal displacements were minimized, indicating that less
postural adjustment was needed when supporting the
load. Contrary to studies on children, adult rucksack
studies have found that high load placement minimizes
postural adaptations, indicating that adult findings can-
not be directly transferred to children. Frank et al. [8]
also investigated load placement in children’s back-
packs. Using biomechanical modeling, it was found
that lumbar and shoulder reaction forces were mini-
mized under low load placement. However, contrary to
what Grimmer et al. [11] found, posture was adversely
affected; low load placement resulted in greater postu-
ral adaptations then high load placement. The contrast-
ing results of these two studies indicate a need to fur-
ther investigate optimal load placement for children’s
backpacks.
Furthermore, theses two studies only examined pos-
tural variables in a rested standing posture, which is
not consistent with the typical use of a backpack. Chil-
dren use backpacks to transport loads; therefore, they
are frequently walking while donning a pack. Walking
requires that the center of gravity be shifted forward,
outside the base of support, while maintaining dynamic
stability and control. Given these fluctuations in the
center of gravity and given that prolonged stresses dur-
ing walking may lead to fatigue, it seems more appro-
priate to evaluate postural changes throughout a gait
cycle.
The purpose of the present study was to examine
the effect of load placement (high, mid, and low) on
posture, specifically trunk forward lean (TFL) posture,
head on neck (CVA) postures and lordosis angle (LA)
for standing and walking in pre-pubescent children.
2. Methods
2.1. Instrumentation
In order to assess spinal curvature, a specialized
backpack needed to be designed and validated. The in-
strumented backpack (IBP) was designed to be a stand-
alone system that was housed in a custom designed
backpack, based on the technology developed by Orloff
and Rapp [25] (Fig. 1a). To assess changes in spinal
curvature, sixteen light resistance spring-loaded slide
potentiometers with 100 mm of travel (Panasonic, mod-
el EVA-JQLR15B14) protruded out of the backpack
frame toward the user’s spine (Fig. 1b). The output
from the potentiometers was multiplexed to 8 chan-
nels and converted to digital form with 0.4 mm resolu-
tion. All data were recorded onto an 8-channel AD128
Valitec Analog/Digital Datalogger V5.0 (Valitec Inc,
Dayton, OH). Data acquisition was controlled by a ra-
dio frequency remote control that enabled the system to
be turned on and off in the testing area. All data were
collected at 31.25 Hz and downloaded to a personal
computer following each testing session.
To determine the validity and reliability of the indi-
vidual potentiometers and the system as a whole, three
sub-studies were completed. In the first study, accura-
cy and repeatability was found for the individual rods
by displacing each rod 20, 40, 60, 80 and 100 mm.
Each displacement was completed three times and the
average difference from expected and the coefficients
of variation (COV) were found. The average differ-
ence from the expected displacement was −1.46 mm
representing a COV was 0.15% across all rods at all
distances. The second study evaluated IBP creep with
data collection every 5 minutes for 30 minutes with
rods at various lengths. No creep existed as the differ-
ence between measured and IBP displacements aver-
aged 0.98 mm ± 0.87 mm. To examine its repeatability
on a curved surface, the IBP was placed five times onto
a block of Styrofoam that was carved into a curved sur-
face. Repeatability was assessed with COV values and
 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children
Fig. 1. Pictures of the Instrumented Back Pack (IBP). a) Uninstalled instrumentation components; b) External side view of the IBP.
accuracy was compared graphically. The COV varied
from 1.4% on flat or gently curving surfaces and 4.9%
on rapidly curving surfaces. This higher COV error
was caused by the fact that only rods’ endpoint edges
made contact with the hard Styrofoam surface. Since
the spine has more gently curving surfaces, one would
anticipate errors in same range as the flat surface at
1.4%.
2.2. Subjects
All subjects were recruited from the same grade five
class at a local public school. Sixteen self-reported
healthy students returned complete parental consent
forms. One student did not meet the inclusion crite-
ria because her body size was not adequate for high,
mid, and low load adjustments. The remaining 15 stu-
dents (10 females and 5 males) completed all three test-
ing conditions. No significant differences in body an-
thropometrics were found between males and females.
Summary anthropometrics and demographic data on
the subjects can be seen in Table 1.
2.3. Procedure
Every child was asked to complete a backpack and
activity questionnaire based on the survey created by
Grimmer and Williams [11]. The questionnaire as-
sessed backpack use, backpack preferences, weight of
load carriage, time of load carriage, activity levels, and
experienced pain. Anthropometrics including: stand-
ing height, weight, vertical back length, surface back
length, waist circumference, and leg lengths were col-
lected prior to the first day of testing.
353
The study methodology was based on a self-
controlled repeated measures design. On three differ-
ent days each child carried 15% of his/her body weight
with the center of mass of the pack located high, mid
or low back. The loads were carried in the IBP. TFL
and CVA were measured using two-dimensional video
and LA was measured using the IBP. Each student was
provided with their own black study t-shirt to increase
consistency between subjects and provide better con-
trast for joint markers. A strip of packing tape was
placed down the back of each shirt to prevent the rods
of the IBP from catching on the subject’s clothing.
On each test day, randomized for high, medium and
low load locations, five data samples were evaluated.
The first sample was a baseline of unloaded standing
posture, which was used to normalize subsequent data
so that each child’s adaptation to the backpack condi-
tion could be compared statistically. The second sam-
ple was a loaded standing posture, which allowed an
assessment of adaptation due to the backpack alone.
Then, each child was accompanied by an adult escort
during the walking phase in order to keep attention on
the task and ensure that markers adhered to the clothing
or skin. Each child walked 100 m around an interior
course in the school before walking through the test-
ing area, for the third data sample, where simultaneous
video and curvature data were collected. Each child
then walked for 1000 m to simulate a typical distance
traveled by children of that age. This distance is an
approximate average of the distances used by Hong
and Cheng [13]. Ratings of perceived exertion (RPE)
were collected every 250 m. At the completion of the
1000 m walk, video and spinal curvatures were taken
again during walking, followed by a final 100 m loop
354 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children

Table 1
Subject anthropometrics
Variables Boys Girls Combined
(n = 5) (n = 10) (n = 15)
Height (cm) 143.8 ± 4.9 146.1 ± 6.1 145.7 ± 5.7
Weight (kg) 37.8 ± 5.1 42.3 ± 10.2 41.8 ± 9.0
Surface Back Length (C7 – L5) 38.7 ± 2.0 40.2 ± 3.7 39.9 ± 3.3
Vertical Back Length (C7 – L5) 37.8 ± 1.7 38.7 ± 3.0 38.5 ± 2.6
Waist Circumference (at Navel) (cm) 67.4 ± 3.8 73.5 ± 8.6 72.1 ± 7.8
Leg Length (GT to floor) (cm) 74.2 ± 5.0 77.7 ± 4.9 76.7 ± 5.1
BMI (kg/m2 ) 18.3 ± 1.8 19.7 ± 3.9 19.6 ± 3.4

before the final standing posture was measured. For UNLOADED STANDING BASELINE
simplicity, these samples will be referred to as: (1) (SB)
standing baseline (SB), (2) initial standing (IS), (3) ini- Video of sagittal plane
tial walking (IW), (4) final walking (FW) and (5) final Spinal curvature by suspension
standing (FS). All standing trials were recorded for five
seconds and evaluated for 1 second at 30 Hz. All walk-
ing trials were evaluated for 2 complete strides. Upon INITIAL STANDING (IS)
completion of the daily testing, the child was asked to Video of sagittal plane
complete a body map diagram with the Wong-Baker Spinal curvature by IBP
Faces pain scale [35] to determine any areas of dis-
comfort. A summary of daily testing can be seen in
Fig. 2.
100m Walk
2.4. Data processing
INITIAL WALKING (IW)
Video data were digitized using Peak MOTUS v.3 Video of sagittal plane
(Peak Performance Technology, Centennial, CO). The Spinal curvature by IBP
ankle, hip, C7 spinal process, base of neck, tragus of the
ear, and top and bottom of the pack were digitized for
one second of each standing trial and two strides of each 1000m
walking trial. TFL was used to evaluate trunk posture
and was defined as the angle between the vertical and
a line passing from the hip to the base of the neck. FINAL WALKING (FW)
CVA was used to evaluate head-on-neck posture and Video of sagittal plane
was defined as the angle between the horizontal and Spinal curvature by IBP
a line passing from the spinal process of C7 and the
tragus of the ear. These angles were averaged within
each testing condition. 100m Walk
The IBP was used to assess the LA. IBP data were
decoded and converted from voltage to displacement.
The data were checked in Excel to ensure that L5 and FINAL STANDING (FS)
rod 16 were aligned. The data were then imported into Video of sagittal plane
MatLab and the spinal curvature profile was rotated Spinal curvature by IBP
and aligned with the trunk to prevent TFL angle from Perceived discomfort scale
confounding spinal curvature data. The data were then
fitted with a 4th order polynomial. Fig. 2. Schedule of data collection procedures within a testing day.
Following normalization of the data, ensemble aver-
ages, that provide the average curve shape, were calcu-
lated for each load placement and each testing condi- culated based on the Cobb angle system for lumbar lor-
tion. To evaluate the differences in the curves, the angle dosis with sagittal view x-rays [5]. Because locations
around the inflection point of the lumbar curve was cal- of bony landmarks were not available, the greatest rates
 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children 355

a) b) c)
Apply Cobb Complete analysis
angle with curve horizontal
technique to to avoid
curve LA indiscriminant slope
a
LA

d) (x-n, ymin)
Max slope tangent line
B = (n2 + (ymin - yapex)2)0.5

C = ymax - y min Apex of

LA lumbar
curve x = 0

A = (n2 + (ymax - yapex)2)0.5

(x-n, ymax) Min slope tangent line
n units

Fig. 3. a) Cobb angle system for lordosis angle; b) LA on curvature data collected; c) Rotate curvature data 90◦ to prevent indiscriminate slopes;
d) Mathematical equations used to calculate lordosis angle.

of change within the lumbar curve were the points used
to determine the tangent line to the lumbar curve. The
angle between the two tangents was termed the lordosis
angle (LA). This methodology can be seen in Fig. 3
and is similar to that used by Willner and Johnson [34]
with pantograph data.
2.5. Statistical analysis
Two way repeated measures ANOVAs (initial/final
x load placement) were completed for standing and
walking postures, for TFL, CVA, LA, and perceived
discomfort. Post hoc analyses for significant results
were completed using paired t-tests with a Bonferroni
correction.
all subjects and indicated that between-subject variabil-
ity was high in all postural measures thus causing high
standard deviations. However, within-subject variabil-
ity was small by comparison and showed logical pat-
terns of postural change. Average trunk forward lean
(TFL) significantly increased (p < 0.001) from 1.7 ◦
± 2.6◦ without the pack to 8.5 ◦ ± 1.9◦ with the pack
loaded to 15% BW. Cranio-vertebral angle (CVA) sig-
nificantly decreased (p < 0.01) from 45.8 ◦ ± 9.3◦ to
38.1◦ ± 7.6◦ indicating a forward head-on-neck pos-
ture when the pack is worn. Lordosis angle (LA) did
not significantly change when the pack was applied.
3.1. Effect of load placement

3. Results The changes in posture (TFL, CVA, and LA) caused

by various load placements (high, mid, and low) were
The standing baseline (SB) with no backpack was evaluated for initial and final standing (IS, FS) and
compared to initial standing (IS) with backpack across walking (IW, FW) samples.
356 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children
Fig. 4. Effect of load placement on the changes in mean trunk forward
lean (TFL), cranio-vertebral angle (CVA) and lordosis angle (LA)
from unloaded standing to final standing. Significant difference be-
tween high and low load placement for CVA was found. (∗ indicates
p < 0.05).
3.1.1. Trunk Forward Lean (TFL)
The changes from the standing baseline (SB) in TFL
were positive for all load placements, indicating a for-
ward trunk lean. However, no significant differences
were found between load placements for TFL within
any sample, standing (IS, FS) or walking (IW, FW).
3.1.2. Cranio-vertebral Angle (CVA)
The changes in CVA from standing baseline (SB) to
initial standing (IS) were negative, indicating a forward
head-on-neck posture. For initial standing (IS) there
were no significance differences between the high, mid,
and low load placements (P > 0.05). However, during
final standing, CVA under the low load placement was
significantly less than the high load placement. The
mean changes from standing baseline (SB) for CVA
with high, mid, and low load placements in the FS
testing condition were −12.52 ◦ ± 7.80◦ , −8.39◦ ±
11.58◦, and −3.79 ◦ ± 11.24◦, respectively (Fig. 4). No
significant differences in CVA were found between load
placements in either initial or final walking conditions.
3.1.3. Lordosis Angle (LA)
The changes from standing baseline (SB) to initial
standing (IS) and final standing (FS) in LA were neg-
ative, indicating a flattening of the lumbar curve. The
mean changes in the LA from SB with high, mid, and
low load placements in the initial standing (IS) condi-
tions were not significant (p > 0.05). However, fol-
lowing the 1000 m walk, a trend towards high and low
placement being better then mid was evident (p = 0.071
Fig. 5. Main effect of 1000 m walk on trunk forward lean (TFL),
cranio-vertebral angle (CVA) and lordosis angle (LA) for all load
placements in standing. (∗ indicates p < 0.05).
and p = 0.087, respectively), although significance dif-
ferences were not found between load placements. The
difference in LA from standing baseline to final stand-
ing (FS) was 5.18 ◦ ± 19.23◦, 22.91◦ ± 14.25◦ and
10.52◦ ± 18.23◦ for high, mid, and low load placement
respectively (Fig. 4). No significant differences were
found in LA between load placements in either initial
or final walking testing conditions.
3.2. Effect of 1000 m walk
Main postural effects of the 1000 m walk were exam-
ined by collapsing all load placements (high, mid, and
low) for both standing (Fig. 5: IS vs. FS) and walking
postures (Fig. 6: IW vs. FW).
3.2.1. Trunk Forward Lean (TFL)
The mean difference in TFL from standing base-
line (SB) over all three load placements was 6.82 ◦ ±
3.13◦ for the initial standing (IS) condition. This angle
significantly increased following the walk to 8.24 ◦ ±
3.88◦ in the final standing (FS) condition (p = 0.03)
(Fig. 5). Conversely, no significant differences in TFL
were found between initial and final walking conditions
(p > 0.05) (Fig. 6).
3.2.2. Cranio-verterbral Angle (CVA)
No significant differences between initial and final
standing postures were found for CVA (Fig. 5). How-
ever, the mean CVA significantly (p = 0.041) decreased
following the 1000 m walk for walking postures. The
mean change from SB for CVA was −12.40 ◦ ± −9.14◦
and −14.5 ◦ ± 10.65◦ in the initial walking (IW) and
final walking (FW) conditions, respectively (Fig. 6).
 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children
Fig. 6. Main effect of 1000 m walk on trunk forward lean (TFL),
cranio-vertebral angle (CVA) and lordosis angle (LA) for all load
placements in walking. (∗ indicates p < 0.05).
3.2.3. Lordosis Angle (LA)
The mean change in LA from the unloaded standing
for the initial standing (IS) condition was −4.65 ◦ ±
12.71◦. This angle significantly increased following
the 1000 m walk (FS) (p < 0.001), creating a hyper-
lordotic curve. The mean change in LA from initial
to final standing was 12.87 ◦ ± 18.56◦ (Fig. 5). Sim-
ilarly, under walking conditions, the mean difference
from unloaded standing for LA following the 1000m
walk (FW) was 10.15 ◦ ± 18.16◦, which is significantly
greater than the initial walking (IW) value of −3.79 ◦
± 14.76◦ (Fig. 6).
3.3. Interaction effects between 1000 m walk and load
placement
3.3.1. Trunk Forward Lean (TFL) and
Cranio-vertebral Angle (CVA)
No interaction effects between load placement and
the 1000 m walk were found for the TFL and CVA
outcome measures in either standing or walking condi-
tions.
3.3.2. Lordosis Angle (LA)
In standing postures a significant increase was found
in LA with mid pack placement only (p = 0.03). How-
ever, in walking postures a significant increase was
found in LA with high pack placement (p = 0.02). No
significant differences were noted in LA with low load
placement in either standing or walking postures.
357
3.4. Perceived exertion and discomfort
Ratings of perceived exertion increased linearly with
no significant differences between load placements.
Using a 10 point scale, average exertion ratings signif-
icantly increased (p < 0.05) from 1.58 after 250 m to
4.03 after 1000 m. However, very few students reported
discomfort with the use of the instrumented backpack
following the 1000 m walk. No differences were found
between load placements in the perceived discomfort
data.
4. Discussion
Regardless of load placement, wearing the pack with
a 15% BW load caused significant increases (p < 0.05)
in TFL and CVA when comparing the unloaded stand-
ing baseline (SB) to initial standing (IS). These findings
support the results of many load carriage studies [7,12,
14,26]; however, other studies did not find a significant
increase until loads were increased to 20% BW [13,15].
These postural deviations may perhaps indicate that the
previously recommended load limit of 15% BW could
be too heavy for ten-year-old children. These changes
were present without any activity or substantial time
of wearing the pack. Shifts in neck alignment may
result in imbalanced muscle performance along with
strain on cervical joints and soft tissue. After compar-
ing radiographic angles and CVAs of adults, Pirunsan
et al. [27] asserted that changes in the position of the
cervical spine may result in additional stresses on the
structures of the upper and lower cervical spine. Fur-
thermore, other investigators have found links between
subjects with smaller CVAs (more forward head on-
on neck lean) and headaches [33]. The results of the
current study cannot however state conclusively that
the 15% weight limit is too high. Although postural
adaptations were evident, the children did not perceive
significant pain or discomfort.
As the purpose of backpacks is to transport loads,
both walking and a significant exposure time provide a
more realistic test of the 15% load than a rested stand-
ing evaluation. It was evident that walking with the
backpack elicited greater trunk forward lean (TFL) and
head poke (CVA) when compared to initial standing
postures. Following the 1000 m walk, TFL postures
did not significantly increase during walking; however,
there was a significant decrease in CVA. This may be
explained by the ceiling effect on TFL previously sug-
gested by Goodgold et al. [9]. When walking, system
358 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children
(pack + person) centre of gravity (C of G) must fall
outside the base of support for a person to accomplish
forward movement, and therefore increased TFL is re-
quired to shift the system C of G forward and balance
the backwards moment created by the posterior load.
The maximum TFL for an individual may be required
even in a rested state to satisfy these requirements,
thereby preventing the child from increasing TFL to
compensate for fatigue. If this is true, following the
1000 m walk, postural accommodation to balance the
load may occur by altering the head-on-neck posture
and subsequently decreasing the CVA. An addition-
al compensation strategy used by the children was to
increase lordotic curvature of the spine, which would
shift the trunk C of G forward.
Standing postures following the 1000 m walk (FS)
indicated TFL significantly increased (p < 0.05) from
the SB posture; however, CVA did not significantly
change. This is logical based on the same ceiling effect
theory stated above. In standing the required TFL is
less because the system C of G only needs to remain
within the base of support. Therefore, TFL can be in-
creased with the increased demands with fatigue. These
changes in TFL are sufficient to maintain equilibrium
and CVA does not need to be altered.
In addition to the above postural changes, significant
increase (p < 0.01) in hyperlordotic spinal postures was
found in the final standing (IS) posture. The changes in
LA for both walking and standing postures indicate that
spinal postural changes are present after the 1000 m
walk. These findings are similar to that of Orloff and
Rapp [25] who found a significant change in spinal cur-
vature following a 15 minute walk with young women.
It is hypothesized that the hyperlordotic curvature in
the lumbar spine is associated with tension on the ante-
rior aspect of the vertebral column and compression on
the posterior aspect, with uneven distribution of forces
on the disc and vertebral body. This posture could lead
to musculoskeletal discomfort and possibly increased
risk of injury because of increased compressive force
on the zygopophysial joints and stresses to the inferior
margins of the articular surfaces [1].
Although the level of physiological fatigue cannot
be determined, the increase in ratings of perceived ex-
ertion as the amount of activity increased can be equat-
ed to the self-report of fatigue used by Negrini and
Carabalona [23]. These researchers found that report-
ing fatigue while carrying a backpack corresponded to
an odds ratio of 3.8 for experiencing pain while carry-
ing the backpack and 5.6 for experiencing at least one
episode of back pain. The changes found following the
1000 m walk in this study may explain these findings.
Comparisons across load placements indicate that
low load placement may minimize changes in posture
and spinal curvature. Although there were no dif-
ferences between load placements for TFL in any of
the testing conditions, CVA results indicated low load
placement was better than high for standing postures
following a 1000 m walk. LA results demonstrated a
trend indicating high and low might be better than mid
load placement. Examining the interaction effects of
load placement and the 1000 m walk indicated that LA
was affected the most by high placement after activity
for walking and for mid placement for standing pos-
tures. The only load location that did not affect LA in
some manner was the low load placement.
Low load placement (at the spinal level of L3) was
recommended by Grimmer et al. [11] when looking at
horizontal displacements of various anatomical land-
marks with only a rested standing posture and 10%
BW loads. Additionally, Frank et al. [8] found that the
forces on the shoulder and back were minimized with
low load placement; however, a postural analysis found
less changes with high placement. Mackie et al. [21]
found that load weight, use of a waist belt and longer
strap lengths resulted in less shoulder pressure, thus
advocating lower load placement. Other researchers
have advocated higher load placements as the forward
lean is less than with low placements [4]; however, this
study was completed with rucksacks in adults where
waist belts were also used and may not be transferable
to children’s backpacks without a waist belt.
Although there are now several studies that suggest
a low placement is best, moderating factors in pack
design such as fit, sternum straps, frame sheets or hip
belts may make other designs advantageous. The re-
sults of an isolated load placement study provide a
starting point for future backpack designs and help to
understand the interactions between backpack use and
postural and spinal curvature changes.
5. Conclusions
Altering the load placement of backpacks was found
to affect the body posture (TFL and CVA) and spinal
curvature (LA) of children. Low load placement ap-
pears to minimize these postural changes as compared
to mid or high load placement, indicating that back-
packs that rest lower on the spine may be best suited for
children. However, future work needs to be complet-
ed before design changes can be recommended with
confidence. The findings of this study are limited in
 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children
that a lack of statistical significance was found between
load placements for TFL in all testing conditions and
in LA and CVA for most testing conditions. This indi-
cates that a follow-up study should be conducted that
includes more than 15 subjects. Furthermore, these re-
sults are restricted to only one factor of backpack de-
sign namely, load placement. Other features of pack
design may affect these results, such as adding sternum
straps, frame sheets or waist belts. The finding that
low load placement minimizes postural adaptations in
children provides the foundation for developing a safer
backpack for children.
Acknowledgements
The authors would like to thank the Ontario Chiro-
practic Association for funding this project. Thanks
are also extended to Alexander Perry (MSc), and Susan
Reid (MSc, PEng) of the Queen’s Ergonomics Research
Group for construction of the Instrumented Back Pack.
Final thanks to the children in grade 5 of the Kingston
Rideau Public School.
References
[1] M.A. Adams, N. Bogduk, A.K. Burton and P. Dolan, The
Biomechanics of Back Pain, Elsevier Science Ltd., 2002.
[2] Backpack Awareness. American Occupational Therapy Asso-
ciation. 2002a. http://www.aota.org/backpacks/docs/schltips.
pdf.
[3] Backpack Safety. Ontario Chiropractic Association. 2002b.
http://www.chiropractic.on.ca/PacLite.htm.
[4] D. Bloom and A.P. Woodhull-McNeal, Postural adjustments
while standing with two types of loaded backpack, Ergonomics
30 (1987), 425–430.
[5] D.G. Borenstien, S.W. Wiesel and S.D. Boden, Low Back Pain
Medical Diagnosis and Comprehensive Management, (2nd
ed.), W.B. Saunders Company, 1995.
[6] H.M. Brackley and J.M. Stevenson, Are Children’s Backpack
Weight Limits Enough? A critical review of the relevant
literature, Spine 29 (2004), 2184–2190.
[7] W. Chansirinukor, D. Wilson, K. Grimmer and B. Dansie,
Effects of backpacks on students: measurement of cervical
and shoulder posture, Aust J Physiother 47 (2001), 110–116.
[8] E. Frank, J.M. Stevenson and P. Stothart, The Effect of Load
Placement on Static Posture and Reaction Forces in Youth,
2003, S21.
[9] S. Goodgold, K. Mohr, A. Samant, T. Parke, T. Burns and L.
Gardner, Effects of backpack load and task demand on trunk
forward lean: Pilot findings on two boys, Work 18 (2002),
213–220.
[10] K. Grimmer and M. Williams, Gender-age environmental as-
sociates of adolescent low back pain, Appl Ergon 31 (2000),
343–360.
359
[11] K. Grimmer, B. Dansie, S. Milanese, U. Pirunsan and P. Trott,
Adolescent standing postural response to backpack loads:
a randomised controlled experimental study, BMC Muscu-
loskelet Disord 3 (2002), 10.
[12] K.A. Grimmer, M.T. Williams and T.K. Gill, The associations
between adolescent head-on-neck posture, backpack weight,
and anthropometric features, Spine 24 (1999), 2262–2267.
[13] Y. Hong and C.K. Cheung, Gait and posture responses to
backpack load during level walking in children, Gait Posture
17 (2003), 28–33.
[14] Y. Hong and G.P. Brueggemann, Changes in gait patterns in
10-year-old boys with increasing loads when walking on a
treadmill, Gait Posture 11 (2000), 254–259.
[15] Y. Hong and J.X. Li, Movement Kinematics of Treadmill
Walking Under Load Carriage in 6-year-old children – A pre-
liminary report. Proceedings of XIX International Symposium
on Biomechanics in Sports, J.R. Blackwell Ltd., 2001, 174–
176.
[16] Y. Hong, J.X. Li, A.S. Wong and P.D. Robinson, Effects of load
carriage on heart rate, blood pressure and energy expenditure
in children, Ergonomics 43 (2000), 717–727.
[17] S.R. Iyer, An ergonomic study of chronic musculoskeletal pain
in schoolchildren, Indian J Pediatr 68 (2001), 937–941.
[18] S.R. Iyer, Backpacks and musculoskeletal pain: do children
with idiopathic scoliosis face a greater risk? J Sch Health 72
(2002), 270–271.
[19] J.P. Lai and A.Y. Jones, The effect of shoulder-girdle loading
by a school bag on lung volumes in Chinese primary school
children, Early Hum Dev 62 (2001), 79–86.
[20] J.X. Li and Y. Hong, Changes of Trunk Position and Breathing
Pattern in Children Walking Under Conditions of Load Car-
riage. Proceedings of XIX International Symposium on Biome-
chanics in Sports, J.R. Blackwell Ltd., 2001, 177–179.
[21] H.W. Mackie, J.M. Stevenson, S.A. Reid and S.J. Legg, Kinet-
ic demands on the shoulder during different simulated school
load carriage conditions, Applied Ergonomics 36 (2005), 188–
206.
[22] G. Merati, S. Negrini, P. Sarchi, F. Mauro and A. Veic-
steinas, Cardio-respiratory adjustments and cost of locomo-
tion in school children during backpack walking: the Italian
Backpack Study, Eur J Appl Physiol 85 (2001), 41–48.
[23] S. Negrini and R. Carabalona, Backpacks on! Schoolchil-
dren’s perceptions of load, associations with back pain and
factors determining the load, Spine 27 (2002), 187–195.
[24] S. Negrini, R. Carabalona, R. Pinochi, R. Malengo and P.
Sibilla, Backpack and back pain in school children: is there a
direct relationship? J Bone Joint Surg Br 80-B (1998), 247.
[25] H.A. Orloff and C.M. Rapp, The effects of load carriage on
spinal curvature and posture, Spine 29 (2004), 1325–1329.
[26] D.D. Pascoe, D.E. Pascoe, Y.T. Wang, D.M. Shim and C.K.
Kim, Influence of carrying book bags on gait cycle and posture
of youths, Ergonomics 40 (1997), 631–641.
[27] U. Pirunsan, P. Trott and K. Grimmer, Cervical posture in the
sagittal plane: A comparison of three methods of measure-
ment. In: Proceedings, Biennial Conference of the Manip-
ulative Physiotherapy Association of Australia, Melbourne,
November 1997, 145–150.
[28] S.A. Reid, J.M. Stevenson and W. Whiteside, Biomechani-
cal assessment of lateral stiffness elements in the suspension
system of a rucksack, Ergonomics 47 (2004), 1255–1271.
[29] G.J. Sheir-Neiss, R.W. Kruse, T. Rahman, L.P. Jacobson and
J.A. Pelli, The Association of Backpack Use and Back Pain in
Adolescents, Spine 28 (2003), 922–930.
360 H.M. Brackley et al. / Effect of backpack load placement on posture and spinal curvature in prepubescent children
[30] J.M. Stevenson, L.L. Bossi, J.T. Bryant, R.P. Pelot, S.A. Reid
and E.L. Morin, A suite of objective biomechanical measure-
ment tools for personal load carriage system assessment, Er-
gonomics 47 (2004), 1160–1179.
[31] B. Troussier, P. Davoine, R. de Gaudemaris, J. Fauconnier and
X. Phelip, Back pain in school children: A study among 1178
pupils, Scand J Rehabil Med 26 (1994), 143–146.
[32] P. Viry, C. Creveuil and C. Marcelli, Nonspecific back pain
in children. A search for associated factors in 14-year-old
schoolchildren, Rev Rhum Engl Ed 66(7–9) (1999), 381–388.
[33] D.H. Watson and P.H. Trott, Cervical headache: an investiga-
tion of natural head posture and upper cervical flexor muscle
performance, Cephalalgia 13(4) (1993), 272–284.
[34] S. Willner and B. Johnson, Thoracic kyphosis and lumbar
lordosis during the growth period in children, Acta Paediatr
Scand 72 (1983), 873–878.
[35] A.S. Wong and Y. Hong, Walking pattern analysis of primary
school children during load carriages on treadmill 492, Med
Sci Sports Exerc 29 (1997), 85.