| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Reports |
I Kingma, PhD, is Assistant Professor, Institute of Fundamental and Clinical Human Movement Sciences, Faculty of Human Movement Sciences, Vrije Universiteit, van der Boechorststraat 9, 1081 BT Amsterdam, the Netherlands.
GS Faber, MSc, is currently a PhD student, and at the time of this study was an MSc student, Institute of Fundamental and Clinical Human Movement Sciences, Faculty of Human Movement Sciences, Vrije Universiteit
AJM Bakker, MD, is currently working as a human movement scientist in the Department of Training Physiology, Royal Dutch Army Health Services. Dr Bakker was an MSc student, Institute of Fundamental and Clinical Human Movement Sciences, Faculty of Human Movement Sciences, Vrije Universiteit, at the time of the study
JH van Dieën, PhD, is Professor of Biomechanics, Institute of Fundamental and Clinical Human Movement Sciences, Faculty of Human Movement Sciences, Vrije Universiteit
i_kingma{at}fbw.vu.nl. Address all correspondence to Dr Kingma
Submitted July 21, 2005;
Accepted March 2, 2006
 | Abstract |
|---|
Key Words: Back injuries • Biomechanics • Ergonomics
| Introduction |
|---|
 TopAbstract Introduction MethodResultsDiscussion and ConclusionsReferences |
|---|
With respect to lifting technique, the evidence is still conflicting. Reviews focusing on squat lifting (lifting with a straight back while bending the knees) versus stoop lifting (lifting with straight knees while bending the back) indicated that the evidence regarding the best technique (ie, the technique resulting in the lowest loading of the low back) is inconclusive.13–15 This inconsistency may be related to the influence of specific details of a lifting condition on the effects of a lifting technique. In a recent study,10 it was shown that effects of lifting technique on low back loading changed or could even reverse when the initial lifting height and foot position relative to the load were changed. Kinematic factors underlying the inconsistency of effects of lifting technique appeared to be the horizontal L5–S1 position relative to the load, the upper-body acceleration, and lumbar flexion.10 The decrease in low back loading attributable to the decrease in lumbar flexion in squat lifting relative to stoop lifting was often smaller than the increase in low back loading caused by the larger horizontal distance from L5–S1 to the load.10 Therefore, it seems that the horizontal distance from L5–S1 to the load is the most important factor determining how a lifting technique affects low back loading.
There are at least 3 strategies for reducing the horizontal distance between L5–S1 and the load. One strategy is to lift the load in between the feet. This strategy has been shown to reduce low back loading in both the stoop lift and the squat lift.10 However, placing both feet beside the load is difficult when lifting larger objects. In both stoop lifting and squat lifting, subjects did not lift the load in between the feet when they were not explicitly instructed to do so.10 One reason may be that lifting with both feet beside the load is rather unstable. A second strategy for bringing L5–S1 closer to the load is to place only one leg beside the load. One such technique is the straddle technique, in which the foot of the second leg is placed behind the load. To our knowledge, only one study16 investigated this technique, with only one specific object at one specific height; that study showed no advantage of the straddle technique over the stoop technique. A third strategy for bringing L5–S1 closer to the load is to combine placing one leg beside the load with bringing the knee of the leg that remains behind the load to the floor. This strategy could further reduce the horizontal distance between the load and L5–S1 as well as allow the maintenance of an upright trunk posture. This technique, which we refer to as the "kneeling technique," has not, to our knowledge, been investigated before. A potential disadvantage of this technique, especially when lifting wider loads, may be the development of substantial asymmetrical low back loading.
The aim of this study was to compare the straddle technique and the kneeling technique with stoop lifting and squat lifting with respect to the effects on kinematics, 3-dimensional (3-D) moments at the L5–S1 joint, and spinal compression and shear forces at this joint. We hypothesized that placing one leg beside the load to be lifted reduces spinal loading but induces asymmetrical loading when lifting wider loads.
| Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussion and ConclusionsReferences |
|---|
Experimental Design
The experiment consisted of 2 repetitions of lifting movements with 20-kg boxes (crates). The lifting movements differed in lifting technique (4 techniques), initial hand height (2 heights), and dimensions of the box used (2 box dimensions). The lifting techniques were: (1) a stoop technique (lifting with the knees extended), (2) a squat technique (bending the knees), (3) a straddle technique (lifting while placing the left foot on the left side of the box and the right foot behind the box), and (4) a kneeling technique (lifting while placing the left foot on the left side of the box and kneeling on the right knee behind the box). In all techniques except the stoop technique, subjects were instructed to hold the back as upright as possible. The 2 boxes were a narrow box (300 mm; a single crate mounted on a wooden board) and a wide box (600 mm; a double crate mounted on a wooden board). Both boxes were 200 mm deep and 270 mm high. For the stoop and squat lifting techniques, subjects were instructed to lift the narrow box between their feet, but they were instructed not to lift the wide box between their feet. The lifting techniques are illustrated in Figure 1.
|
The boxes had been placed on a shelf, suspended 50 mm above the surface of the force plate on which the subjects stood. All lifts were performed with an initial hand height of 290 mm (boxes were grabbed in a symmetrical way at their handles) and with an initial height of 50 mm (boxes were grabbed in a symmetrical way at their lower edges). Each lifting movement started with the subject in an upright standing posture. After the start of data collection, the subject stepped forward and, using the instructed technique, lifted the box to a height that allowed the subject to stand upright with slightly flexed arms. After the recording stopped, the subjects placed the box back on the shelf. The order of lifting technique, initial hand height, and box width was randomized over subjects. Subjects were free to select their preferred lifting speed.
Dynamic 3-D Linked Segment Model
A dynamic 3-D linked segment model was used to estimate net moments at the L5–S1 intervertebral disk. This model has been described in detail elsewhere17 and has been internally validated by comparing a top-down to a bottom-up calculation of net moments. In addition, model results have been compared to independent net moment estimations with an electromyography (EMG)-based model and a neural network-based model.7 The current model uses anthropometric data as described by McConville et al,18 combined with force-plate data (measured at 500 Hz with a custom-made force plate measuring 1.0 x 1.0 m) and kinematics from light-emitting diode markers on cuffs to follow the lower-body segments (feet with lower legs, upper legs, pelvis, and trunk) during movement. To optimize visibility, markers on the cuffs were attached to small metal plates mounted to the cuffs with a double-hinge joint.
Trajectories of the cuff markers were recorded at 100 Hz and synchronized with force-plate signals by use of an automated 3-D movement registration system (Optotrak*; SD of system accuracy, <0.05 mm) with 3 arrays of 3 cameras. Before the measurements were obtained for each subject, the force plate and the Optotrak system were calibrated and cuff markers were related to anatomical landmarks by making a short recording while pointing at each landmark19 with a pointer containing 6 markers. Marker data were low-pass filtered by use of a bidirectional second-order Butterworth filter at a cutoff frequency of 10 Hz. A global equation of motion (rather than a segment-by-segment calculation) was used as described by Hof20:
i is the angular velocity of segment i, d(...)/dt is the time derivative of the expression within parentheses, and Mg is the ground reaction moment measured by the force plate. Boldface type represents vectors in the equation. This moment is non-0 around the vertical axis only. The L5–S1 joint was chosen as the level of analysis because it is the lumbar joint that is expected to undergo the largest loads. The global equation of motion allowed the use of 1 instead of 2 force plates. Anatomical axes of the trunk and pelvis were defined as follows: positive x-axis (lateral flexion) forward; positive y-axis (flexion-extension) to the left; and positive z-axis (twisting) upward. Net moments were expressed in the pelvic axis system. The trunk movement relative to the pelvis was decomposed in the order y-x-z.
3-D EMG-Driven Trunk Model
Fourteen pairs of surface EMG electrodes were attached to the skin after abrasion and cleaning with alcohol (Ag-AgCl electrodes
; interelectrode distance, 20 mm). Electrodes were bilaterally attached ventrally over the rectus abdominis muscle (at the level of the umbilicus), the internal oblique muscle (just superior to the inguinal ligament), and the anterior (approximately 15 cm cranial of the anterior iliac spine) and lateral (midaxillary line, halfway between the iliac crest and the lowest edge of the rib cage) parts of the external oblique muscle. Dorsally, electrodes were attached over the iliocostalis lumborum muscle (6 cm lateral to L2) and over the longissimus thoracis pars lumborum (3 cm lateral to L1) and pars thoracis (4 cm lateral to T10) muscles.
Before the actual experiment, subjects performed 7 maximum isometric contractions of the trunk muscles, 3 times, as described by McGill.21 The EMG data were recorded (Porti-17TM
; input impedance, 1012 
; common mode rejection ratio, >90 dB), band-pass filtered (10–400 Hz), converted from analog to digital (22 bits at 1,000 Hz), and stored synchronized to Optotrak and force-plate data. Offline, EMG signals were high-pass filtered (20 Hz), full-wave rectified, and low-pass filtered at 2.25 Hz.22 The EMG data were normalized to maximum voluntary contractions and used as the input of an EMG-driven trunk muscle model. The model has been described in more detail elsewhere23,24 and consists of a compilation of anatomical data described by Stokes and Gardner-Morse25 for the back muscles and by McGill26 for the abdominal muscles. The transversus abdominis muscle and the psoas major muscle were excluded because it is unlikely that their activity can be estimated reliably from surface EMG data and because their moment-producing capacity is limited. The latissimus dorsi muscle was omitted because a reliable indication of its force would require modeling the shoulder in detail and because its capacity to generate an extensor moment at the lumbar spine is only very small.27 After exclusion of the above-mentioned muscles, the model consisted of 90 muscle slips crossing the L5–S1 joint. The model was scaled to individual body height. For muscle slips crossing the L4 and T12 levels, nodes were used as points about which these long muscles were wrapped. In this way, the muscles followed the lumbar curvature during motion.
After assigning each of the 90 muscle slips to 1 of the 14 EMG signals, muscle forces were estimated as the product of the assumed muscle maximum stress (a single value for all muscles, which was adjusted for each subject to obtain the best fit between net moments and muscle moments), normalized EMG amplitude, and correction factors for the instantaneous muscle length28 and contraction velocity29 that had been calculated with 3-D trunk lumbar angles. Finally, to obtain compression and shear forces at the L5–S1 intervertebral joint, muscle forces and net reaction forces were summed after being projected on the axis system connected to the L5–S1 disk. For convenience, shear forces pushing the trunk forward were indicated as positive, and absolute values were taken for lateral shear forces.
Data Analysis
From the time series of the net moment around the L5–S1 joint, peak values were calculated for the extending, lateral flexion, and torsion components of the net moment as well as for the total moment (ie, the vector sum of the 3 moment components). Furthermore, at the instant of peak total moment, trunk inclination, lumbar flexion, lateral flexion, torsion, and the horizontal distance from L5–S1 to the load center of mass were determined. Finally, time series of the forces at the L5–S1 joint, calculated with the EMG-assisted trunk model, were used to calculate peak compression forces and peak forward and lateral shear forces.
For the values described above, as well as for co-contraction (the flexor moment generated by the abdominal muscles at the instant of peak compression, expressed as a percentage of the net extensor moment), repeated-measures analyses of variance (ANOVAs) were applied (one ANOVA for each dependent variable) with lifting technique (4 levels), initial hand height (2 levels), and box width (2 levels) as independent variables. A significance level of P<.05 was used. Each lift was performed twice and all variables were averaged over those 2 lifts before statistical analyses were applied. Finally, differences between individual lifting techniques were tested with Tukey honestly significant difference post hoc tests.
| Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussion and ConclusionsReferences |
|---|
|
|
Spinal Forces
Although the overall pattern of compression forces over conditions and lifting techniques (Fig. 3) had the same appearance as the pattern of total moments, there were some relevant deviations. There was no main effect of lifting technique on L5–S1 compression forces. This finding was unlike the finding for total moments. The most substantial difference with the pattern found for net total moments was a relatively higher loading for the kneeling technique.
|
Forward shear loads showed substantial differences between lifting techniques only when subjects lifted the wide box from 290 mm. In that condition, the straddle and squat techniques resulted in higher forward shear forces than did the stoop technique, and the squat technique resulted in higher shear forces than did the kneeling technique.
Lateral shear forces were below 300 N for all lifting techniques in all 4 initial height and box width conditions. Consistent with the lateral flexion moments, lateral shear forces were higher for the kneeling technique than for the stoop and squat techniques in all 4 initial height and box width conditions. For the straddle technique, this was the case only when subjects lifted the wide box.
Trunk Motion
Lumbar flexion and trunk inclination were larger when subjects lifted the wide box than when they lifted the narrow box and were larger when subjects lifted from 50 mm than when they lifted from 290 mm (Table, Fig. 4). In all 4 initial height and box width conditions, the stoop technique resulted in more lumbar flexion and more trunk inclination than all other techniques. The kneeling technique resulted in less trunk inclination than the other techniques, but the difference with the straddle technique was not significant for lifting of the narrow box. Lumbar flexion did not differ significantly among the squat, straddle, and kneeling techniques.
|
Distance
In 9 of 12 paired comparisons, the horizontal distance from L5–S1 to the box was significantly smaller with the stoop lifting technique than with the other lifting techniques (Fig. 5). When subjects lifted the wide box, the straddle technique and the kneeling technique were more effective in reducing the horizontal distance from L5–S1 to the box than was the squat technique. However, when subjects lifted the narrow box, for which squat lifting was performed with the feet beside the box, the opposite was the case.
|
| Discussion and Conclusions |
|---|
|
TopAbstractIntroductionMethodResultsDiscussion and ConclusionsReferences |
|---|
Spinal Compressive Loading
Over all 4 hand height and load size conditions, the kneeling technique resulted in smaller net total moments than did the other 3 techniques. However, L5–S1 compression forces did not show this overall advantage for the kneeling technique. To facilitate the interpretation of differences in effects of lifting technique between net total moments and compression forces, co-contraction was analyzed. This analysis showed that co-contraction contributed to the absence of an overall advantage of the kneeling technique for L5–S1 compression forces. The level of co-contraction was higher for the kneeling technique than for the other 3 techniques, especially for lifting from 290 mm. This co-contraction caused a higher level of spinal compression with the kneeling technique than with the other techniques. This increased co-contraction in the kneeling technique may be related to the asymmetry of the posture. In addition, most subjects indicated that the kneeling technique felt less stable than the other techniques; this factor also may have enhanced the level of co-contraction.
The effects of interactions between lifting technique and both load size and initial hand position were much more prominent than the main effect of lifting technique. The lifting technique resulting in the highest total moments and compression forces and the lifting technique resulting in the lowest total moments and compression forces varied over the 4 box width and initial hand height combinations. The main reason for those interactions is that total moments and compression forces appeared to be much more dependent on lifting height and on box width in the squat and straddle techniques than in the stoop technique. This difference between lifting techniques in sensitivity of spine loading to height and width can be attributed in part to trunk inclination and L5–S1 distance. With respect to trunk inclination, Figure 4 shows that trunk inclination was higher with all lifting techniques when subjects lifted the wide box than when they lifted the narrow box and when they lifted with an initial hand position of 50 mm than when they lifted with an initial hand position of 290 mm. This increase in trunk inclination caused an increase in the moment arm of the trunk center of mass for the squat, straddle, and kneeling techniques but not for the stoop technique. The reason is that the trunk inclination in the stoop lifting technique was already over 90 degrees when subjects lifted the narrow box from an initial hand position of 290 mm, whereas this was not the case in the other lifting techniques.
An additional reason for the interaction of lifting technique with box width was that the squat technique was more sensitive to box width than were the other techniques. When subjects lifted the wide box, especially when the initial hand position was low, they needed to maintain clearance for the knees by adopting a posture with a large horizontal distance between the box and the pelvis, because this box width did not allow lifting between the knees. A comparable problem occurred when subjects lifted the wide box with the straddle technique; unlike the situation with lifting of the narrow box, the knee of the leg that was placed behind the load could not be rotated outward far enough to bring the knee beside the load. Thus, the required clearance of the knee of the leg that was placed behind the box forced the subjects to move the pelvis backward. Consequently, the moment arm of the load relative to the L5–S1 joint increased with wider loads when the straddle technique was used. As a consequence of the interactions described above, stoop lifting resulted in the highest total moments and compression forces when subjects lifted the narrow box at a 290-mm height, whereas squat lifting resulted in the highest total moments and compression forces when subjects lifted the wide box at a 50-mm height. When subjects lifted the narrow box at a 50-mm height, no significant differences between lifting techniques were found.
For comparisons of spinal loading over task conditions or over lifting techniques, effects on spinal loading should be compared with effects on the strength of the spinal structures. Compression forces in this study are within the range of values that can cause end plate fractures in vitro.31 Facilitated by the nonlinear relationship between compression force and population at risk, the effects of lifting condition and lifting technique, as observed in our study, can have substantial effects on the population at risk.31 As such, those effects are clinically significant. However, compressive strength may be affected by task conditions and lifting techniques through differences in posture. Unfortunately, little is known about the effects of spinal posture on strength. Especially in stoop lifting, high levels of lumbar flexion are reached. Adams et al32 reported a reduced compressive strength beyond 75% of the maximum in vitro flexion. However, according to Adams and Hutton,33 it is unlikely that such flexion is reached in vivo. Nevertheless, full in vivo lumbar flexion may result in substantial stresses on the posterior annulus and on vertebral ligaments,34 and these stresses could increase the risk of soft-tissue injury in or around intervertebral disks.
With respect to asymmetry in lumbar spine posture, both the kneeling technique and the straddle technique were found to result in somewhat more asymmetry than the stoop technique and the squat technique under most conditions. Epidemiologic work has shown that asymmetry in lifting is a separate risk factor for acute disk prolapse.35 However, we are unaware of experimental work showing reduced spinal strength under lumbar torsion or lateral flexion of the limited magnitude found in the present study. Lumbar torsion especially may even have been overestimated in the present study because the markers on the trunk were mounted at about T9 rather than T12.
Besides a slightly asymmetrical posture during peak loading, especially the kneeling technique resulted in substantial asymmetrical moment components (Fig. 2). As a result, lateral shear forces were higher with the kneeling technique than with the other techniques. However, the magnitude of the lateral shear forces was relatively low (Fig. 3), making it unlikely that those forces, by themselves, could harm the lumbar spine. In addition to the implications for lateral forces, the asymmetrical moments likely contributed to the higher level of co-contraction that was found with the kneeling technique.
Spine Shear Loading
Substantial forward shear forces, approximately between 1,100 and 1,700 N, were found at the L5–S1 joint. Those forces are in the range of values that have been reported to cause bony failure in vitro.36,37 The magnitude of the forces at the L5–S1 joint is in line with previous work.9,10,38,39 These forces are much higher than the forces reported for the L4–L5 joint,10,40 a difference that may be explained by the more forward inclined orientation of the L5–S1 disk.
Forward spinal shear loading in 3 of 4 box width and initial height conditions was unaffected by lifting technique. Likewise, Kingma et al10 reported no effect of lifting technique on L5–S1 shear forces. Kingma et al10 decomposed shear forces at the L5–S1 joint into muscular and net reaction force components. They showed that in lifts with more trunk flexion (such as the stoop technique), a larger net reaction shear force was compensated for by a smaller muscular shear force, resulting in total shear forces being unaffected by lifting technique. In the present study, only when subjects lifted the wide box from the 290-mm height were substantial differences between lifting techniques seen. In this condition, shear forces were lower with stoop lifting than with squat lifting and straddle lifting. This result may have been caused by the large difference in lumbar flexion between the stoop lifting technique and the other lifting techniques in this condition. The larger lumbar flexion with stoop lifting decreased the muscular component of the forward shear force relative to the L5–S1 joint. When subjects lifted the narrow box from the 290-mm height, lumbar flexion also was much larger with stoop lifting than with the other techniques. However, this condition did not result in lower shear forces with stoop lifting, because of the relatively large moments requiring large muscle forces.
Some limitations of the present study should be mentioned. First, asymmetrical load placement relative to the subject was not considered, whereas approximately half of industrial lifting tasks involve more than 10 to 15 degrees of asymmetry in load placement.41 Furthermore, we studied only a relatively small group of healthy young male subjects. The interactions between lifting technique and box width and initial hand height, as found in the present study, may not be the same in females.30,42 In addition, although the role of lifting technique in this issue has not been clarified, back loading patterns attributable to variations in task constraints have been shown to differ between people with back pain and healthy people.43
In conclusion, the present study showed that no single lifting technique can be advised for all task conditions. Especially with the squat lifting technique, low back loading varied markedly with box width and initial hand height. The straddle and kneeling techniques were introduced because they might be able to reduce the horizontal distance from the load to the low back. However, when subjects lifted a narrow box, the opposite appeared to be the case, so that the straddle and kneeling techniques did not reduce low back loading in comparison with squat lifting. A wide box did not allow placement of the feet beside the box during squat lifting. As a result, the straddle technique and the kneeling technique were successful in reducing the horizontal distance from the low back to the load in comparison with the squat lifting technique, thereby also reducing net total moments. However, this result was obtained at the cost of substantial asymmetrical load components and a substantial amount of co-contraction. Consequently, compression forces did not differ as much as net total moments between the squat technique and the straddle or kneeling technique when subjects lifted a wide box.
Notably, lifting condition (height and width) had more influence on compression forces than did lifting technique. When subjects lifted the wide box from 50 mm, the lowest compression force was almost 6,000 N (ie, when the stoop technique was used). In contrast, when subjects lifted the narrow box from 290 mm, the highest compression force was only about 5,000 N (again, when the stoop technique was used). Thus, the most important advice with regard to lifting is to avoid lifting wide objects from floor level. Second, when objects can be lifted between the feet, squat lifting is preferred from the perspective of compression forces. Third, when lifting a wide object from the floor cannot be avoided, squat lifting is not advised, and the kneeling technique may be preferred in order to limit compression forces without the need for full lumbar flexion, as in stoop lifting.
| Footnotes |
|---|
The study was approved by the ethics review board of the Faculty of Human Movement Sciences, Vrije Universiteit.
This work, in part, was presented at the annual meeting of the International Society of Occupational Ergonomics and Safety; June 27–28, 2005; Las Vegas, Nev.
* Northern Digital Inc,103 Randall Dr, Waterloo, Ontario, Canada. 
Sentry Medical Products, 17171 Murphy Ave, Irvine, CA 92714.
TMS, Zutphensestraat 57, Oldenzaal, the Netherlands.
| References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussion and ConclusionsReferences |
|---|
Related Article
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
