You are on page 1of 7

Journal of Biomechanics ()

Contents lists available at ScienceDirect

Journal of Biomechanics
journal homepage:

Effect of body weight on spinal loads in various

activities: A personalized biomechanical modeling approach
M. Hajihosseinali a, N. Arjmand a,n, A. Shirazi-Adl b
Department of Mechanical Engineering, Sharif University of Technology, Tehran 11155-9567, Iran
Division of Applied Mechanics, Department of Mechanical Engineering, cole Polytechnique, Montral, Qubec, Canada

art ic l e i nf o a b s t r a c t

Article history: Epidemiological studies are divided over the causative role of body weight (BW) in low back pain.
Accepted 23 November 2014 Biomechanical modeling is a valuable approach to examine the effect of changes in BW on spinal loads
and risk of back pain. Changes in BW have not been properly simulated by previous models as associated
Keywords: alterations in model inputs on the musculature and moment arm of gravity loads have been neglected.
Body weight A detailed, multi-joint, scalable model of the thoracolumbar spine is used to study the effect of BW
Biomechanical modeling (varying at ve levels, i.e., 51, 68, 85, 102, and 119 kg) on the L5S1 spinal loads during various static
Musculature scaling symmetric activities while scaling moment arms and physiological cross-sectional areas of muscles
Spinal loads using in vivo imaging data. The L5S1 loads substantially increased with BW especially in exed
postures. As BW increased from 51 to 119 kg, the L5-S1 compression increased in exed postures by
 80147% with no load in hands and by  4652% in load holding tasks. In obese individuals with body
mass index430 kg/m2 spinal loads further increased by up to 15% as lever arms for gravity loading at
the waistline (T12 through L5) increased by 2 cm (for BW 102 kg) and 4 cm (for BW 119 kg). With
changes in BW, spinal loads would have moderately altered ( o17%) had identical muscle parameters
been considered. Since scaling muscle parameters demands additional efforts in modeling, one could opt
for simulation of alterations only in BW while using some averaged musculature values.
& 2014 Elsevier Ltd. All rights reserved.

1. Introduction Among the foregoing personal factors, body weight (BW) has
been the subject of much debate and controversy. While some
Etiology of low back pain (LBP) is often complex and multifactorial. epidemiological studies indicate BW as a strong risk factor of LBP,
Excessive mechanical loads on the trunk, applied intermittently or intervertebral disc herniation and osteophytes, others consider it as
repetitively over time, are however believed to play a causative role in either only a possible weak or even a non-contributing risk factor
associated injuries, degeneration and LBP. Numerous epidemiological (Kaila-Kangas et al., 2003; Kostova and Koleva, 2001; see also
investigations relating occupational LBP to physical activities corrobo- Manchikanti, 2000 for a comprehensive review). Larger stature loss
rate the role of biomechanical risk factors (Hoogendoorn et al., 2000; during walking has been observed in obese individuals as compared
Manchikanti, 2000; Thiese et al., 2014; Van Nieuwenhuyse et al., to non-obese ones (Rodacki et al., 2005). Larger spinal shrinkage has
2004). While both task-related (e.g., external trunk loading, posture also been reported in obese persons when comparing standing
and lift characteristics) and personal-related (e.g., gender, age, body height with recumbent length (Yar, 2008). The measured increase in
weight and height) parameters have been indicated in onset or spinal shrinkage can in part be an indication of larger compressive
progress of LBP, much less attention has been focused on the loads in obese individuals. Postoperative implant subsidence has
personal-related factors. Task-related factors are suggested to be more often been observed in patients with higher body mass index (BMI)
important when designing prevention programs (Waters and Putz- (Behrbalk et al., 2013) presumably again due to associated larger
Anderson, 1996). The Revised NIOSH Lifting Equation (Waters et al., spinal loads.
1993) incorporates also only task-related variables (e.g., handled load Musculoskeletal biomechanical models are valuable tools in cross-
position, its asymmetry, and travel distance) when estimating the validation of LBP risk factors via comprehensive evaluation of nd-
recommended weight limit for individuals performing lifting activities. ings reported by epidemiological and experimental investigations.
They can hence shed light on associations between personal-related
factors and LBP. Earlier models have accounted only for changes in
the upper BW with no consideration for accompanying alterations
Corresponding author. Tel.: 98 21 66 16 56 84; fax: 98 21 66 00 00 21. in inputs on the anatomy of muscles (e.g., moment arm and cross-
E-mail address: (N. Arjmand). sectional area), mechanical properties of the ligamentous spine
0021-9290/& 2014 Elsevier Ltd. All rights reserved.

Please cite this article as: Hajihosseinali, M., et al., Effect of body weight on spinal loads in various activities: A personalized
biomechanical modeling approach. Journal of Biomechanics (2014),
2 M. Hajihosseinali et al. / Journal of Biomechanics ()

(e.g., bending stiffness), mechanical properties of muscles (e.g., maxi- 2. Method

mal contractile stress and passive stiffness) and moment arm of
gravity loads at different spinal levels. This is mainly because the 2.1. Thoracolumbar spine model
extent of variations in foregoing parameters as BW alters is either
currently unknown or difcult to be simulated. Lifting analysis tools A three-dimensional, six-joint, 18 rotational degrees of freedom, combined
stability- and kinematics-driven nonlinear model (Fig. 1) was developed to estimate
to quantify spinal loads such as the 3D Static Strength Prediction
muscle forces and spinal loads (Fig. 2) (see Hajihosseinali et al., 2014). The model
ProgramTM (3DSSPP) (University of Michigan Center for Ergonomics, consisted of a rigid pelvis, rigid thorax and ve rigid vertebral bodies with a total of
2014), the revised Hand-Calculation Back Compressive Force (HCBCF) six spherical intervertebral joints (T12L1 through L5S1), each having three non-
model (Merryweather et al., 2009), the polynomial equation of low linear rotational stiffness. Eighteen angular springs were used to model the nonlinear
back compression (McGill et al., 1996) and the regression models rotational stiffness of T12S1 motion segments (i.e., vertebrae, disc, facets and
ligaments) based on reported model and measurement results (Arjmand and Shirazi-
(Fathallah et al., 1999; Potvin, 1997) though all consider the BW as an Adl, 2006; Shirazi-Adl, 2006). The neutral upright posture under gravity loading was
input into their model but fail to scale foregoing parameters accord- estimated from the initial undeformed conguration of the spine according to a
ingly. In other words, these models incorporate the effect of changes previously reported optimization approach (El-Rich et al., 2004; Shirazi-Adl et al.,
in BW on the trunk loading similar to those of a task-related variable 2002, 2005). This geometry of the reference model (BW 68 kg) in neutral upright
posture was individualized, in terms of the head and arm positions, for a healthy
such as hand load. AnyBody Modeling System (Damsgaard et al.,
male (52 years,  174.5 cm, and  68 kg) (Arjmand et al., 2009, 2010). For a total BW
2006), however, uses a general linear scaling approach to adjust the of 68 kg an upper trunk weight of 344.4 N ( 52% of total BW (Potvin, 1997)) was
segment-xed insertion nodes of muscles and ligaments with BW. considered. Weights of upper arms (35.6 N), forearms/hands (29.3 N) and head
Linear increases in spinal loads with BW have been reported when (46 N) were applied in each loading case at their mass centers estimated based on
using this software to simulate a number of static activities (Han et in vivo measurements (Arjmand et al., 2009, 2010). The remaining upper trunk
weight (i.e., 344.435.629.346233.4 N) was distributed eccentrically from the T1
al., 2013). No detailed modeling study has yet considered the effect of
to the L5 (Arjmand and Shirazi-Adl., 2006; Pearsall, 1994) and applied off center on
changes in BW on the lever arm of gravity forces distributed along each vertebra via rigid links. For the simulation of lifting tasks with a weight in hands
the spine as this parameter likely differs when the added BW is due an external load was also applied at the position measured in vivo (Arjmand et al.,
to anterior distributed adiposity (belly fat) in obese population or by 2009, 2010).
For a given lifting activity, the measured sagittal thorax and pelvis rotations
musculature in athletes (Pryce, 2013).
(Arjmand et al., 2009, 2010) were applied onto the model while the total lumbar
The present work hence aims to study the effect of variations in rotation (the difference between thorax and pelvis rotations) was partitioned
BW via a detailed, multi-joint, scalable trunk model (Hajihosseinali between individual vertebrae (T12 to L5) based on relative values reported in the
et al., 2014) on the L5S1 disc compression and shear forces during literature (Arjmand and Shirazi-Adl, 2005). Under this deformed conguration, the
a number of static symmetric tasks. Each activity is simulated using net external moment (due to gravity and hand-held loads minus the passive
resistance) was calculated at each joint and partitioned among all muscles spanning
ve distinct BWs ranging from 51 to 119 kg while scaling the
that joint. Eighteen equilibrium equations in the form of r  f M were considered
musculature (moment arms, MA, and physiological cross-sectional where r, f, and M are moment arm of muscles with respect to the joint, unknown
areas, PCSA), for the rst time, using recent in vivo imaging data total forces in muscles, and net moments, respectively. As the unknown muscle
(Anderson et al., 2012). Unlike many models that consider BW as a forces (76 in total, Fig. 1) outnumber the available equality equations (18 in total, i.e.,
3 equilibrium equations at each of the 6 joints) an optimization algorithm that
concentrated force on the trunk, in the current model BW is
minimizes sum of cubed muscle stresses was used to resolve redundancy. Further-
distributed off-centered along the spine with weights of arms and more, for the system to be stable, the Hessain matrix of the system potential energy
head applied at their mass centers. Moreover, for the rst time, the (i.e. the second partial derivatives of the potential energy with respect to degrees of
effect of larger lever arms of gravity loads at waistline in obese freedom) were constrained to remain positive semi-denite in the optimization
individuals on the L5S1 loads is investigated. According to the algorithm (i.e., all 18 eigenvalues of the Hessain matrixZ0). The potential energy
was determined from the geometry (posture), external load, muscle forces/stiffness
measurement of larger spinal shrinkage and stature loss in obese
and passive torsional (spring) stiffness during the task under consideration. Stiffness
individuals, it is hypothesized that greater BW imposes much larger of each muscle, k, was assigned using the linear stiffnessforce relation (k qf/l)
loads on the lumbar spine. (Bergmark, 1989), where q is the muscle stiffness coefcient taken equal to 5 for all

Fig. 1. The sagittally-symmetric trunk model in the upright standing posture under gravity loading represented in both sagittal and coronal planes. 46 bilateral local muscles
attached to lumbar vertebrae (ICPL: iliocostalis lumborum pars lumborum, LGPL: longissimus thoracis pars lumborum, MF: multidus, QL: quadratus lumborum, and IP:
iliopsoas) and 30 bilateral global muscles attached to the thoracic spine and cage (global back muscles: ICPT: iliocostalis lumborum pars thoracic and LGPT: longissimus
thoracis pars thoracic. Global abdominal muscles: IO: internal oblique, EO: external oblique, and RA: rectus abdominus) are considered (axes are not to the same scale). To
simulate curved paths of global extensor muscles in exed postures (i.e., ICPT and LGPT), they were restrained not to penetrate and wrap around, if necessary, predened
vertebral points in between their insertions while considering the wrapping forces generated at contact points as additional external loads on the spine.

Please cite this article as: Hajihosseinali, M., et al., Effect of body weight on spinal loads in various activities: A personalized
biomechanical modeling approach. Journal of Biomechanics (2014),
M. Hajihosseinali et al. / Journal of Biomechanics () 3

Fig. 2. Schematics of the simulated tasks in upright and exed postures with and without hand load used to predict spinal loads while varying the body weight (BW).

7 Sagittal MA Frontal MA PCSA 20

Moment Arm at T12 (cm)


PCSA (cm2)
4 12
3 8
0 0
BW = 51 kg 68 kg 85 kg 102 kg 119 kg
Fig. 3. Variations in the sagittal moment arm, frontal moment arm and physiological cross-sectional area (PCSA) of the principal extensor muscles in the model (i.e., ICPT:
iliocostalis lumborum pars thoracic and LGPT: longissimus thoracis pars thoracic depicted in Fig. 1) at the T12 level as the body weight (BW) alters.

muscles (Granata and Wilson, 2001; Stokes and Gardner-Morse, 2001), f the muscle and body height were xed to those of our reference model (i.e., male, 52 years,
force and l the instantaneous muscle length (Hajihosseinali et al., 2014). Introduction and 174.5 cm) BW was varied for the simulation of each task and the regression
of the stability requirement into our optimization algorithm allows for the prediction equations were used to evaluate MA and PCSA of muscles. For instance, the sagittal
of antagonistic (abdominal) muscle forces which are usually neglected in conven- MA, frontal MA and PCSA of the ICPT and LGPT muscles at the T12 level increase
tional optimization-based models. Spinal loads were subsequently estimated in local according to the regression equations by, respectively,  14%, 5.8% and 77% as BW
directions using equilibrium of forces at deformed congurations. increases from 51 to 119 kg (Fig. 3). For all simulations and regardless of BW, the
upper body weight (head, neck, arms and thoracolumbar spine from T1 to L5) was
considered equal to  52% of BW (Potvin 1997) with weight of each segment
2.2. Simulated tasks modied proportionally. Effect of larger lever arms of gravity loads at the waistline
in obese individuals (Pryce, 2013) on the L5S1 loads was also investigated in
Various static symmetric activities in upright and exed postures were simu- models with BMI430 kg/m2 (i.e., models with BW 102 and 119 kg corresponding
lated. This included the relaxed upright standing, forward upper trunk exion of 10, to, respectively, 33.5 and 39.1 kg/m2). For these two cases, anterior lever arms for
20, 30, 40, 50, 60, 70, and 801 with arms in the gravity direction, holding a weight of gravity loads at T12 through L5 were increased by 2 and 4 cm, respectively, and
19.8 kg close to (25 cm to the S1 joint) and away from (55 cm to the S1 joint) the analyses were repeated for activities in upright and exed (trunk exion of 201, 501,
chest in the upright posture and nally holding 180 N while exing forward by 201, and 801 with and without 180 N in hands) postures.
501 and 801 with arms in the gravity direction (Fig. 2). To investigate the effect of BW
on the predicted spinal loads at the L5-S1 disc, each task was simulated using
the reference BW68 kg as well as four different BWs of 51, 85, 102 and 119 kg
(reduction of 25% as well as increases of 25%, 50% and 75% to the reference BW) 3. Results
while scaling the musculature as described below. The minimal (51 kg) and maximal
(119 kg) BWs approximately correspond to 5th percentile of North American females The L5S1 compression and shear loads substantially increased
and 95th percentile of North American males, respectively, (McDowell et al., 2008).
Analyses were also repeated without scaling muscle parameters (i.e., muscle
with BW especially in exed postures (Figs. 4 and 5). As BW increased
parameters kept constant as in the reference model with BW 68 kg) to assess the from 51 to 119 kg, the L5S1 compression increased by  80147% in
effect of such scaling. tasks with no load in hands, by  4652% in load holding exed tasks
and by  25% when 19.8 kg was held close to the chest in the upright
2.3. Scaling approach posture but decreased by  12% when 19.8 kg was held away from
the chest with an extended posture (Fig. 4). Likewise, As BW
Muscle parameters (MA and PCSA) were scaled according to the regression increased from 51 to 119 kg, the posterior-anterior L5S1 shear load
equations of Anderson et al. (2012) constructed based on quantitative computed increased by  84160% in tasks with no load in hands, by  4557%
tomography (QCT) of all major paraspinal and abdominal muscles from the T6 to in load holding exed tasks and by  21% when 19.8 kg was held
the L5 from a sample of 100 men and women aged 3687 years. These regression
equations (Anderson et al., 2012) relate MA and anatomical cross sectional area
close to the chest but decreased by  20% when 19.8 kg was held
(ACSA) at each spinal level to individual's sex, age, height and weight. Identical away from the chest (Fig. 5). In obese individuals as lever arms for
relative changes were assumed in ACSA and PCSA at different BWs. While sex, age waistline gravity loading increased by 2 cm (for BW102 kg) and

Please cite this article as: Hajihosseinali, M., et al., Effect of body weight on spinal loads in various activities: A personalized
biomechanical modeling approach. Journal of Biomechanics (2014),
4 M. Hajihosseinali et al. / Journal of Biomechanics ()

51 kg 68 kg 85 kg 102 kg 119 kg

L5-S1 Compression (N)

Upright 10 20 30 40 50 60 70 80 19.8 kg 19.8 kg 20 50 80
Close Away
Flexed Postures Flexed Postures + 180 N

Fig. 4. Predicted L5S1 compression loads (normal to disc mid-height plane) for different activities presented in Fig. 2 at ve different body weights varying from 51 to
119 kg.

Fig. 5. Predicted L5S1 posterioranterior shear loads (parallel to disc mid-height plane) for different activities presented in Fig. 2 at ve different body weights varying from
51 to 119 kg. All shear forces are oriented towards anterior direction.

Fig. 6. Effect of larger lever arms for gravity loading at the waistline (2 cm for BW 102 kg and 4 cm for BW 119 kg) on the L5S1 compression loads in the scaled model.

4 cm (for BW119 kg), spinal loads further increased by up to 15% which compression and shear loads decreased by 19% and 24%,
and 12% in upright and exed postures, respectively (Fig. 6). respectively).
Repeating analyses without scaling MAs and PCSAs in the
models with BW of 85, 102, and 119 kg increased compression
and shear loads in scaled models by less than 17% (Fig. 7); L5-S1 4. Discussion
compression and shear loads increased by  5%, 10%, and 13% in
average as compared to the respective scaled models with BWs of, This study investigated the effect of variations in BW on the spinal
respectively, 85, 102, and 119 kg. In contrast, spinal loads decreased loads using a detailed biomechanical modeling approach. As BW
by less than 7% in the non-scaled model with BW of 51 kg (except in altered (at ve different levels from 51 to 119 kg), musculature of the
lifting tasks of 19.8 kg away from the chest in the upright posture in model was scaled, for the rst time, according to available imaging

Please cite this article as: Hajihosseinali, M., et al., Effect of body weight on spinal loads in various activities: A personalized
biomechanical modeling approach. Journal of Biomechanics (2014),
M. Hajihosseinali et al. / Journal of Biomechanics () 5

Fig. 7. Predicted L5S1 compression loads (normal to disc mid-height plane) for different activities presented in Fig. 2 at ve different body weights varying from 51 to
119 kg but without (in contrast to Figs. 4 and 5) scaling muscle parameters.

data. Results indicated that BW had a marked impact on computed extensor muscles should increase by 2025% when lifting in exed
spinal loads, i.e., the hypothesis of the study was conrmed. Larger postures with 180 N in hands, for instance. These MAs should further
spinal shrinkage in obese individuals compared to non-obese ones increase by 4555% and 6590% to reduce compression loads in the
(Rodacki et al., 2005; Yar, 2008) could at least in part be due to larger model with BWs of, respectively, 102 and 119 kg to those in the
spinal compressive loads in obese individuals. Results however would reference model. None of the in vivo imaging studies however suggest
have only moderately altered (o17%) had identical muscle para- such large increases in the sagittal MAs of muscles as BW alters (e.g.,
meters been considered irrespective of changes in BW (Fig. 7). Since Seo et al., 2003 suggests an increase of  7%, 15.5%, and 25.5% in the
scaling muscle parameters demands additional efforts in modeling, L3L4 MA of erector spinae for male subjects as BW increases from
one could opt for simulation of alterations only in BW while using 68 kg to 85, 102 and 119 kg, respectively).
some averaged musculature values. Our optimization-based model itself has some limitations as
discussed elsewhere (Hajihosseinali et al., 2014). Intervertebral
4.1. Limitations translational degrees of freedom (DOFs) were neglected for simpli-
cation when introducing the spinal stability in our optimization
There are some limitations in both the scaling approach and algorithm. We have recently investigated the effect of neglec-
biomechanical model used. Only musculature of the model (i.e., ting these DOFs on model predictions using our nite element
sagittal and frontal MAs as well as PCSA of muscles) was scaled as kinematics-driven model (Ghezelbash et al., 2014) and found that it
BW altered whereas the likely effect on other model input para- had small to moderate effects on model predictions. Validity of the
meters such as mechanical properties of the ligamentous spine or reference model has however been veried by comparing its
sacral angle in upright posture were not considered due to the lack predictions for compressive loads under a number of lifting activities
of in vivo data. The effect of BMI on the standing lumbar lordosis has with the measured intradiscal pressures under identical tasks
been however found to be statistically insignicant (Youdas (Hajihosseinali et al., 2014). In the present study, only symmetric
et al., 2006). The QCT imaging-based regression equations used in activities were investigated despite presence of out-of-plane degrees
the model to scale muscle parameters (Anderson et al., 2012) are not of freedom in the model. This was done due to the stability
sensitive, at constant body height, to the cause of increase in BW as constraints that were accounted in all planes. Neglecting out-of-
identical muscle parameters are predicted irrespective of whether plane degrees of freedom from the model was veried, however, to
BW is increased by gaining fat deposits (as in obese individuals) or only slightly affect the model predictions for the tasks studied here.
by gaining muscle masses (as in athletes). In the former case, back Finally, identical kinematics as those measured in vivo and pre-
extensor muscles may not actually become larger while the regres- scribed into the reference model (BW 68 kg) were considered in all
sion equations predict larger MAs and PCSAs. In obese individuals, models.
therefore, the effect of BW on the spinal loads may indeed be greater
than that estimated in this study especially when also considering 4.2. Analysis of results
the larger lever arms of gravity loads at the lower trunk in obese
individuals. On the other hand, as BW increases in athletes by gain in The effect of alterations in BW on the predicted spinal loads
muscle mass this trend reverses as the effect of BW on the spinal varied both magnitude- and pattern-wise depending on the task
loads may indeed be smaller than that found in this study. simulated (Figs. 4 and 5). For tasks in the upright standing posture,
There are a number of other in vivo studies that have reported spinal loads were primarily inuenced by changes in BW (e.g.,
muscle parameters based on CT or MR imaging scans (e.g., Jorgensen maximum of 592 N increase in the L5S1 compression as BW
et al., 2001; Marras et al., 2001; Seo et al., 2003). We however chose to increased by 68 kg from 51 to 119 kg). In exed postures, however,
use the regression equations of Anderson et al. (2012) because of the apart from the direct load of increased BW, the additional exion
population used. This work has used not only a large sample of both moment resulted in considerable increase in muscle forces and
sexes (50 men and 49 women) having a wide range of age (3687 thus spinal loads (e.g., maximum of  1527 N increase in the L5S1
years) but, unlike other studies, has reported MA (both sagittal and compression as BW increased from 51 to 119 kg in the exed
frontal) and ACSA of all individual back and abdominal muscles at all posture of 801 with 180 N in hands). In contrast to all other tasks,
spinal levels from T6 to L5. The ndings of our study could be affected spinal loads decreased when BW increased from 51 kg under
had we selected alternative input data while our conclusion regarding 19.8 kg held away from the chest (Figs. 4 and 5). In this task,
the impact of BW on spinal loads would have remained unchanged. according to our in vivo measurements which were also prescribed
This is because our analyses indicate that in order to reduce the L5S1 into the model, the upper trunk shifted posteriorly (i.e., backward
compression loads in the model with BW of 85 kg to those predicted extension of the trunk) (see Fig. 2 for a schematic representation of
in the reference model with BW of 68 kg, the sagittal MAs of all back the posture taken in this task) thus moving the center of mass of

Please cite this article as: Hajihosseinali, M., et al., Effect of body weight on spinal loads in various activities: A personalized
biomechanical modeling approach. Journal of Biomechanics (2014),
6 M. Hajihosseinali et al. / Journal of Biomechanics ()

the upper trunk posteriorly to the L5S1. This helps exploit the Behrbalk, E., Uri, O., Parks, R.M., Musson, R., Soh, R.C., Boszczyk, B.M., 2013. Fusion
extension moment of the gravity to counterbalance the large and subsidence rate of stand alone anterior lumbar interbody fusion using
PEEK cage with recombinant human bone morphogenetic protein-2. Eur. Spine
exion moment of the external load held away from the chest. J. 22, 28692875.
This auxiliary extension moment along with the increases in MAs Bergmark, A., 1989. Stability of the lumbar spine a study in mechanical
as BW increased reduced extensor muscle forces and thus spinal engineering. Acta Orthop. Scand. Suppl. 230, 154.
Damsgaard, M., Rasmussen, J., Christensen, S.T., Surma, E., de Zee, M., 2006. Analysis
loads in this task. Moreover, while the spinal forces increased
of musculoskeletal systems in the AnyBody Modeling System. Simul. Model.
approximately linearly with BW in all tasks in exed postures their Pract. Theory 14, 11001111.
variations with BW was not linear in tasks in upright postures El-Rich, M., Shirazi-Adl, A., Arjmand, N., 2004. Muscle activity, internal loads and
(Figs. 4 and 5). stability of the human spine in standing postures: combined model-in vivo
studies. Spine 29, 26332642.
While the effect of BW on the spine passive properties is less Fathallah, F.A., Marras, W.S., Parnianpour, M., 1999. Regression models for predict-
evident, it is known that as BW increases the center of mass of the ing peak and continuous three-dimensional spinal loads during symmetric and
upper trunk shifts further forward (especially as the body height asymmetric lifting tasks. Hum. Factors 41, 373388.
Ghezelbash, F., Arjmand, N., Shirazi-Adl, A., 2014. Effect of intervertebral transla-
remains unchanged) (Pryce, 2013) resulting in even greater muscle
tional exibilities on estimations of trunk muscle forces, kinematics, loads, and
forces and spinal loads than those reported here. Our sensitivity stability. Comput. Methods Biomech. Biomed. Eng. ,
analyses indicate that in obese individuals with BMI 430 kg/m2 as 10255842.2014.961440 (in press).
lever arms for gravity loading at the waistline (i.e., T12 through L5) Granata, K.P., Wilson, S.E., 2001. Trunk posture and spinal stability. Clin. Biomech.
16, 650659.
increased by 2 cm (for BW102 kg) and 4 cm (for BW 119 kg) Hajihosseinali, M., Arjmand, N., Shirazi-Adl, A., Farahmand, F., Ghiasi, M.S., 2014.
spinal loads further increased by up to 15% (Fig. 6). On one hand, A novel stability and kinematics-driven trunk biomechanical model to estimate
this concords with the fact that up to 90% of spinal forces, in muscle and spinal forces. Med. Eng. Phys. 36, 12961304.
Han, K.S., Rohlmann, A., Zander, T., Taylor, W.R., 2013. Lumbar spinal loads vary
physically more demanding tasks, is generated by muscle forces
with body height and weight. Med. Eng. Phys. 35, 969977.
(Arjmand and Shirazi-Adl, 2005) making thus the proper scaling of Hoogendoorn, W.E., Bongers, P.M., de Vet, H.C., Douwes, M., Koes, B.W., Miedema,
muscle parameters more crucial when simulating heavier tasks. M.C., Arins, G.A., Bouter, L.M., 2000. Flexion and rotation of the trunk and
On the other hand, this estimated additional increase of up to 15% lifting at work are risk factors for low back pain: results of a prospective cohort
study. Spine 25, 30873092.
in spinal loads as gravity lever arms at the waistline increases by Jorgensen, M.J., Marras, W.S., Granata, K.P., Wiand, J.W., 2001. MRI-derived
24 cm occurs in upright postures and could grow had we moment-arms of the female and male spine loading muscles. Clin. Biomech.
considered greater portion of body mass (taken constant in our 16, 182193.
Kaila-Kangas, L., Leino-Arjas, P., Riihimki, H., Luukkonen, R., Kirjonen, J., 2003.
study to  52% of total BW) distributed in the trunk segments
Smoking and overweight as predictors of hospitalization for back disorders.
(Pryce, 2013), greater increases in lever arms of gravity loads, or Spine 28, 18601868.
assumed that back extensor muscle parameters (MA and PCSA) do Kostova, V., Koleva, M., 2001. Back disorders (low back pain, cervicobrachial and
not become larger as much in obese individuals. lumbosacral radicular syndromes) and some related risk factors. J. Neurol. Sci.
192, 1725.
In summary, BW had a marked impact on the predicted spinal Manchikanti, L., 2000. Epidemiology of low back pain. Pain Physician 3, 167192.
loads. The extent of such effect varied depending on the task Marras, W.S., Jorgensen, M.J., Granata, K.P., Wiand, B., 2001. Female and male trunk
simulated, i.e., larger for tasks in exed postures as compared to geometry: size and prediction of the spine loading trunk muscles derived from
MRI. Clin. Biomech. 16, 3846.
those in the upright posture. Larger compressive loads are experi-
McDowell, M.A., Fryar, C.D., Ogde, C.L., Flegal, K.M., 2008. Anthropometric reference
enced by the spine as BW increases that should be considered as data for children and adults: United States, 20032006, National Health
an additional risk factor for LBP. Depending on the required trade- Statistics Reports; no. 10. National Center for Health Statistics, Hyattsville, MD.
off between accuracy and computational simplicity, biomechanical McGill, S.M., Norman, R.W., Cholewicki, J., 1996. A simple polynomial that predicts
low-back compression during complex 3-D tasks. Ergonomics 39, 11071118.
model studies may use some averaged values for the muscle Merryweather, A.S., Loertscher, M.C., Bloswick, D.S., 2009. A revised back compres-
parameters irrespective of changes in the BW. sive force estimation model for ergonomic evaluation of lifting tasks. Work 34,
Pearsall, D.J., 1994. Segmental Inertial Properties of the Human Trunk as Deter-
mined from Computer Tomography and Magnetic Resonance Imagery (Ph.D.
Conict of interest statement thesis). Queen's University, Kingston, Ontario, Canada.
Potvin, J.R., 1997. Using NIOSH equation inputs to calculate lumbosacral compres-
We have no conict of interest to declare. sion forces. Ergonomics 40, 691707.
Pryce, R.T., 2013. Obesity and Loading during Lifting (Ph.D. thesis). University of
Manitoba, Manitoba, Canada.
Rodacki, A.L., Fowler, N.E., Provensi, C.L., Rodacki, Cde L., Dezan, V.H., 2005. Body
Acknowledgments mass as a factor in stature change. Clin. Biomech. 20, 799805.
Seo, A., Lee, J.H., Kusaka, Y., 2003. Estimation of trunk muscle parameters for a
biomechanical model by age, height and weight. J. Occup. Health 45, 197201.
This work is supported by Grants from Sharif University of Shirazi-Adl, A., Sadouk, S., Parnianpour, M., Pop, D., El-Rich, M., 2002. Muscle force
Technology (Tehran, Iran) and Institut de recherche Robert-Sauv evaluation and the role of posture in human lumbar spine under compression.
Eur. Spine J. 11, 519526.
en sant et en scurit du travail, IRSST (Montral, Canada).
Shirazi-Adl, A., El-Rich, M., Pop, D.G., Parnianpour, M., 2005. Spinal muscle forces,
internal loads and stability in standing under various postures and loads:
applications of kinematics-based algorithm. Eur. Spine J. 14, 381392.
References Shirazi-Adl, A., 2006. Analysis of large compression loads on lumbar spine in exion
and in torsion using a novel wrapping element. J. Biomech. 39, 267275.
Anderson, D.E., D'Agostino, J.M., Bruno, A.G., Manoharan, R.K., Bouxsein, M.L., 2012. Stokes, I.A., Gardner-Morse, M., 2001. Lumbar spinal muscle activation synergies
Regressions for estimating muscle parameters in the thoracic and lumbar trunk predicted by multi-criteria cost function. J. Biomech. 34, 733740.
for use in musculoskeletal modeling. J. Biomech. 45, 6675. Thiese, M.S., Hegmann, K.T., Wood, E.M., Garg, A., Moore, J.S., Kapellusch, J.M.,
Arjmand, N., Shirazi-Adl, A., 2005. Biomechanics of changes in lumbar posture in Foster, J., Greene, T., Stoddard, G., Biggs, J., BackWords Study Team, 2014. Low-
static lifting. Spine 30, 26372648. back pain ratings for lifetime, 1-month period, and point prevalences in a large
Arjmand, N., Shirazi-Adl, A., 2006. Model and in vivo studies on human trunk load occupational population. Hum. Factors 56, 8697.
partitioning and stability in isometric forward exions. J. Biomech. 39, 510521. University of Michigan Center for Ergonomics, 2014. 3D Static Strength Prediction
Arjmand, N., Gagnon, D., Plamondon, A., Shirazi-Adl, A., Larivire, C., 2009. Program User's Manual.
Comparison of muscle forces and spinal loads estimated by two biomechanical Van Nieuwenhuyse, A., Fatkhutdinova, L., Verbeke, G., Pirenne, D., Johannik, K., Somville,
models of the human trunk. Clin. Biomech. 24, 533541. P.R., Mairiaux, P., Moens, G.F., Masschelein, R., 2004. Risk factors for rst-ever low
Arjmand, N., Gagnon, D., Plamondon, A., Shirazi-Adl, A., Larivire, C., 2010. A back pain among workers in their rst employment. Occup. Med. 54, 513519.
comparative study of two trunk biomechanical models under symmetric and Waters, T.R., Putz-Anderson, V., Garg, A., Fine, L.J., 1993. Revised NIOSH equation for
asymmetric loadings. J. Biomech. 43, 485491. the design and evaluation of manual lifting tasks. Ergonomics 36, 749776.

Please cite this article as: Hajihosseinali, M., et al., Effect of body weight on spinal loads in various activities: A personalized
biomechanical modeling approach. Journal of Biomechanics (2014),
M. Hajihosseinali et al. / Journal of Biomechanics () 7

Waters, T.R., Putz-Anderson, V., 1996. Manual materials handling. In: Bhatta-charya, Youdas, J.W., Hollman, J.H., Krause, D.A., 2006. The effects of gender, age, and body
A., McGlothlin, J.D. (Eds.), Occupational Ergonomics: Theory and Applications. mass index on standing lumbar curvature in persons without current low back
Marcel Dekker, New York, pp. 329349. pain. Physiother. Theory Pract. 22, 229237.
Yar, T., 2008. Spinal shrinkage as a measure of spinal loading in male Saudi
university students and its relationship with body mass index. Saudi Med. J. 29,

Please cite this article as: Hajihosseinali, M., et al., Effect of body weight on spinal loads in various activities: A personalized
biomechanical modeling approach. Journal of Biomechanics (2014),