Workplace musculoskeletal disorders, such as low back pain and osteoarthritis in the knees can be physically and financially burdensome to members of the working class and will continue to have a negative impact on society as they age.1,2 One of the leading causes of workplace injuries is overexertion in tasks, such as manual materials handling (MMH), manufacturing, and patient transport.3 Workers in manual labor professions may perform repetitive, asymmetric, and physically taxing movements4 such as twisting, pushing, or bending while handling heavy loads5–7—putting them at risk for acute injury when exposed to extreme peak loading or chronic overuse injury with prolonged exposure to joint and tissue loading.8 Common approaches to characterize and mitigate injury risk in manual lifting have relied on metrics derived from measures of joint kinematics and kinetics. Factors such as varying the mass of a lifted object,9–11 the size of a lifted object,12 lifting speed,10 lifting technique,12,13 stance width and foot position,12,14 lowering versus lifting,15,16 and symmetry versus asymmetry (turning) all play significant roles in understanding their impact on musculoskeletal loading.10,11,13,15,17,18 In sum, external loads on the limb-joints during MMH tasks have been well characterized and helped provide best practices to guide lifting techniques that can help avoid injuries.
Characterizing internal loads on underlying musculoskeletal structures may provide a more direct assessment of injury risk than external measures based on inverse dynamics. Joint contact forces (JCFs), potential biomarkers for osteoarthritis onset due to structural changes in cartilage in ACL injuries,19–21 and correlated to low back injury risk22 could reveal key underlying changes in cartilage and bone health. JCFs quantify the internal loading at bone-to-bone interfaces necessary to support the forces produced by surrounding muscles, ligaments, and tendons that are necessary to counter external loads.23 Despite their potential utility, JCFs are impossible to measure noninvasively in humans, with only a few examples of direct in vivo measurements of JCFs.24,25 Furthermore, mechanical analyses that attempt to map net external forces and moments to internal joint loads are not straightforward and require estimates of how antagonistic muscle forces are coordinated.26 Fortunately, advances in computational biomechanics make it possible to utilize musculoskeletal modeling practices to reproduce MMH tasks in simulation27,28 and estimate JCFs.29–33 Electromyography (EMG)-informed solvers can reduce errors in JCF estimation from purely simulated muscle activation inputs by accounting for muscle co-activation,34,35 which is a critical piece of the puzzle given that muscle action typically accounts for well over half the joint contact load.36,37 These advanced computational tools make it possible to move “outside-in” and evaluate tissue-level mechanics in the context of highly dynamic movements—enabling monitoring of a new set of “local” biomarkers that could offer more precise injury prevention measures.
The purpose of this study was to extend limb-joint analyses and establish a framework to examine how external loads resolve as 3-dimensional JCFs inside limb joints during MMH tasks. To bridge this gap, we designed a comprehensive study protocol comprised of a diverse set of lifting tasks spanning a range of start-end heights and across a range of lift symmetries. In these conditions, we collected full-body motion data using high-speed motion capture, external ground reaction forces, and electromyography from a carefully selected set of muscles around the back and knee. We paired these experimental data with state-of-the-art computational modeling, simulation, and optimization techniques to evaluate lower back (L5/S1) and knee joint loads during lifts that varied starting and ending heights and lift symmetry, providing insight into potential “hot spots” for injury over a wide range of possible lifting situations. We hypothesized that asymmetrical,38,39 ascending,10 and below the waist lifts11,40 would induce the highest JCFs at both the L5/S1 and knee (H1–H3). We expected JCFs at the knee and L5/S1 joints would be proportional to each other,9,10,40–42 independent of the lifting task (H4)—reflecting a motor control strategy to avoid overloading the L5/S1 or knee by distributing loads across the limb joints. Studying the effects of MMH tasks on multijoint internal joint loading will enable a comprehensive mapping from task space to injury risk during industry-relevant lifting tasks. In doing so, we hope to provide a greater awareness of the susceptibility to eventual tissue and bone-damaging injuries in the workplace that can be used as a prophylactic tool to avoid prolonged exposure to critical loading in the workplace.
Methods
Participants
This study was approved by the Georgia Institute of Technology Institutional Review Board, and all participants provided signed consent prior to data collection. We enrolled 9 participants (7 males and 2 females; weight: 80.7 [15.0] kg; height: 178.7 [11.1] cm; age: 25.1 [2.9] y). Participants were excluded if they had a history of debilitating injuries from neurological, musculoskeletal, or cardiovascular conditions that prohibited them from successfully performing the manual labor tasks in this study.
Experimental Design
All participants performed 24 different lifts with varied lift start–end heights and degrees of twisting (Figure 1). Start–end height combinations included knee-to-waist (KW), waist-to-knee (WK), shoulder-to-waist (SW), and waist-to-shoulder (WS). Symmetric lifting had 0° of twisting, while asymmetric lifting could be with 90° or 180° of twisting, with turns centered around a neutral posture (ie, 180° lifts started and ended at ± 90° from neutral). Across conditions, participants lifted a 25 lb. (11.34 kg) dumbbell at their preferred lifting speed to fixed shelf positions at knee (17.8 cm [7 in]), waist (72.4 cm [28.5 in]), or shoulder (133.5 cm [52.5 in]) height.
—(A) Manual lifting conditions visualized by starting shelf height and target shelf position of the weight. During 0° (symmetric) lifts, a single shelf was oriented directly in front of participants and weight was lifted to-and-from various heights. For both 90° and 180° (asymmetric) lifting, tasks were performed starting from both sides of the body: left-starting lifts (clockwise rotation) and right-starting lifts (counterclockwise rotation). In 90° lifts, shelves were offset 45° from neutral. In 180° lifts, 2 shelves were placed on opposite sides of the participants. The top, blue arrows represent above-the-waist lifting conditions: SW in light blue and WS in dark blue. The bottom, red arrows represent below-the-waist lifting conditions: KW in dark red and WK in light red. (B) Placement of surface EMG sensors. Muscles were selected to capture activity from muscle groups’ salient to the targeted joints under study: lower back (L5/S1) and knee. Anterior and posterior muscle groups were collected for the torso and legs, but only the right leg was instrumented with sensors. EMG indicates electromyography; KW, knee-to-waist; SW, shoulder-to-waist; WK, waist-to-knee; WS, waist-to-shoulder. (Color figure online)
Citation: Journal of Applied Biomechanics 41, 2; 10.1123/jab.2023-0292
Participants wore a full-body reflective marker set to record segment positions (Vicon, 100 Hz). We collected data from 16 surface EMG sensors (Delsys, Natick, Massachusetts, 2000 Hz). EMG sensors were placed on the right leg muscles (tibialis anterior, lateral gastrocnemius, rectus femoris, vastus medialis, vastus lateralis, bicep femoris, and semitendinosus), and bilaterally on the torso muscles (erector spinae, latissimus dorsi, external obliques, and right-only rectus abdominis; Figure 1B). Maximum voluntary contractions (MVCs) were performed by each participant preceding the protocol to collect the maximum muscle activations for normalization.43,44 We used SENIAM to inform sensor placement and MVC collection.45 We collected ground reaction forces under each foot and each shelf (Bertec, 2000 Hz).
Data Analysis
Raw EMG signals were band-pass filtered (20–400 Hz, second-order Butterworth), full-wave rectified, low-pass filtered (6-Hz, fifth-order Butterworth), and half-wave rectified to create positively signed envelopes. Finally, the EMG signals were normalized using the peak MVC value across 3 MVC trials. We used the rate of change of ground reaction force from force plates on the shelves with a threshold value of 0.10 N/s to detect the start–end position of the dumbbell and define the initiation-termination time of each lift.
We implemented 2 OpenSim models (version 4.0 [SimTK]30) to simulate experimental lifting trials: Full-Body Running model46 for knee-joint analyses and Lifting Full Body model29 for L5/S1-joint analyses. The Full-Body Running model contains 12 segments with 92 trunk and lower limb musculotendon actuators, while the Lifting Full Body model contains 30 segments with 238 torso musculotendon actuators. The Full-Body Running and Lifting Full Body models have a high resolution of knee and back muscles, respectively, providing representative muscle activations and forces from muscles which we collected EMG. To accurately simulate our unique lifting conditions, we modified the range-of-motion constraints of joints within the arms, torso, pelvis, and legs to allot freedom for the model skeleton to match the participants’ movements. Additionally, to account for the external load contributions of the 25 lb. weight to internal musculoskeletal states, we added half the dumbbell mass to each hand using OpenSim WeldJoint.
Next, we fitted the 2 OpenSim models to participant-specific anthropometry taken from captured static pose data using the Scaling Tool. We used the Inverse Kinematics Tool to calculate joint angles. OpenSim’s Inverse Dynamics Tool computed joint kinetics for each participant incorporating the effects of external loads. We used the Muscle Analysis Tool to solve for musculotendon unit lengths and muscle moment arms.
Then, we used the Calibrated EMG-Informed Neuromusculoskeletal Modeling Toolbox (CEINMS)34 to perform optimization (ie, simulated annealing) and reduce error in joint and muscle-level property estimations. First, the CEINMS Hybrid mode optimized muscle activity over those muscles without experimental data inputs. Then, we used both the experimental and simulated muscle excitation signals, musculotendon lengths, and muscle moments to calibrate the model by reducing the error between estimated and experimental joint moments. The following objective function weighting parameters were inspired by those utilized in similar analyses by Molinaro et al18 to balance muscle activations and torque-tracking performance: torque tracking (α) = 10,000, muscle excitation minimization (β) = 10, and EMG tracking (γ) = 1000.18
Finally, we used the OpenSim Joint Reaction Analysis Tool, taking CEINMS simulated output, to calculate the JCFs. From-time series data, we computed peak and integrated values using custom MATLAB scripts. We divided JCFs into normal, shear, and total components. We calculated shear and total JCFs using the Euclidean norm of the anterior-posterior and mediolateral shear forces and the Euclidean norm of the normal and shear forces, respectively. We normalized JCFs by the product of participant mass and acceleration of gravity to minimize the effects of performance variance across participants. To avoid duplication of data in asymmetrical lifting conditions (90° and 180° turns), we compared the forces in clockwise (left-starting) and counterclockwise (right-starting) directed lifts and analyzed/presented the trial with the maximum JCF (ie, the “worst case”). We reported right knee JCFs, from the leg instrumented with EMG sensors.
Statistical Analysis
We employed a 2-way repeated-measures analysis of variance to test for significance of lift symmetry, direction, and start-to-end height on each component of the JCFs (H1–3). Subsequently, we performed a post hoc Bonferroni pairwise multivariate comparison test to determine statistical significance of differences in JCFs across lifting conditions with the threshold for significance set at α = .05 (Minitab–Penn State University). We used a linear regression model to compute correlation coefficients between the L5/S1 and knee JCFs (MATLAB; H4).
Results
Participants used different postures across lift conditions. Joint kinematics indicated asymmetric lifts had greater L5/S1 flexion (sagittal plane), lateral bending (frontal plane), and axial rotation (transverse plane) than symmetric lifts. Peak right knee flexion (sagittal plane) was less in asymmetric lifts than symmetric lifts across conditions (Supplementary Figure S1 [available online]). In addition, participants used different kinetic strategies47 to perform lifting tasks (Figure 2). The peak (Supplementary Figure S2 [available online]) and integrated (Supplementary Figure S3 [available online]) flexion moments about the L5/S1 and right knee joints were smaller in asymmetric lifts (90° and 180°) compared with symmetric lifts (0°). The peak (Supplementary Figure S2 [available online]) and integrated (Supplementary Figure S3 [available online]) lateral bending and axial rotation moments about the L5/S1 joint were greater in asymmetric lifts than symmetric lifts.
—The across-participant averaged time series of L5/S1 and knee kinetics of knee flexion (A), L5/S1 flexion (B), lateral bending (C), and axial rotation (D) during KW, WK, SW, and WS lifting conditions. The color opacity of each data line increases with degree turn of lift: from 0° (lightest) to 180° (darkest). The haze surrounding each average represents the standard error of the mean. A comparative analysis was performed across each asymmetric lift (90° and 180°) condition to select data from the starting side (right or left) that resulted in greater (worse) joint moments. Knee joint moments were reported from the right, instrumented leg. Positive joint moments for the L5/S1 and knee flexion components represented greater joint flexion. Axial rotation and lateral bending components of the L5/S1 joint were positive moving away from the starting side. KW indicates knee-to-waist; WK, waist-to-knee; SW, shoulder-to-waist; WS, waist-to-shoulder.
Citation: Journal of Applied Biomechanics 41, 2; 10.1123/jab.2023-0292
Muscle activations are an important variable in determining muscle forces, the primary contributor to JCFs.36,47 Across lifting conditions (Supplementary Figure S4 [available online]), the external obliques, vastus medialis, biceps femoris, and semitendinosus all had greater peak normalized muscle activations (EMG) in asymmetric (90° and 180°) versus symmetric lifts (0°). In contrast, the erector spinae, rectus femoris, and lateral gastrocnemius, all had greater peak normalized EMG in symmetric (0°) versus asymmetric lifts (90° and 180°). Latissimus dorsi and rectus abdominis muscles had negligible differences in their normalized EMG across lifts.
In the L5/S1 joint, asymmetric lifts (90°/180°) resulted in lower normal, shear, and total JCFs when compared with symmetric lifts (Figure 3A). Peak normal forces were 7.2% (P = .046) and 20.2% (P < .01) lower in asymmetric lifts at 90° and 180° versus symmetric lifts (0°), respectively. Peak shear forces were 20.3% (P = .001) and 14.3% (P = .032) lower for 90° and 180° lifts versus symmetric lifts (0°). Peak total forces were 7.3% (P = .032) and 20.6% (P < .01) lower for 90° and 180° lifts versus symmetric lifts (0°; Figure 4A, Table 1). Peak normal, shear, and total JCFs were not different in ascending (KW/WS) versus descending (WK/SW) lifts in the L5/S1 joint (Figure 5A). Peak normal, shear, and total JCFs were 21.0%, 49.3%, and 23.3% higher in below-the-waist (KW/WK) than above-the-waist (SW/WS) conditions (P < .01) in the L5/S1 joint (Figure 6A).
—The across-participant averaged time series for total JCFs in the L5/S1 joint (A) and knee (B). The 4 panels in each figure illustrate data of the lifting conditions: KW, WK, SW, and WS. The color opacity of each data line increases with degree turn of lift: from 0 (lightest) to 180° (darkest). The haze surrounding each average represents the standard error of the mean. JCFs indicates joint contact forces; KW, knee-to-waist; SW, shoulder-to-waist; WK, waist-to-knee; WS, waist-to-shoulder.
Citation: Journal of Applied Biomechanics 41, 2; 10.1123/jab.2023-0292
—The across-participant averaged peak normal, shear, and total JCFs in the knee and L5/S1 joint as explored in the first hypothesis (H1). The bars represent the JCF value averaged across the 4 lift conditions (KW, WK, SW, and WS) for each of the 3 lift degrees (0°, 90°, and 180°) in light to dark gray. Independent values for each lift condition are labeled as circles in the corresponding color (KW: red, WK: pink, SW: dark blue, and WS: light blue). Significantly different groups are denoted with an asterisk. Statistical significance across all conditions concluded when α = .05. JCFs indicates joint contact forces; KW, knee-to-waist; SW, shoulder-to-waist; WK, waist-to-knee; WS, waist-to-shoulder. (Color figure online)
Citation: Journal of Applied Biomechanics 41, 2; 10.1123/jab.2023-0292
Effect of Lifting Rotation and Start-to-End Height on Knee and Lower Back (L5/S1) Peak Total Joint Contact Forces (×BW)
| Rotation, ° | Start-to-end heights | |||
|---|---|---|---|---|
| KW | WK | SW | WS | |
| Knee | ||||
| 0° | 6.71 (2.10) | 7.81 (3.53) | 5.67 (1.34) | 5.16 (1.22) |
| 90° | 7.82 (2.28) | 8.09 (3.16) | 6.04 (1.70) | 5.54 (1.44) |
| 180° | 6.93 (2.22) | 7.49 (3.04) | 4.02 (1.03) | 4.08 (1.03) |
| Lower back (L5/S1) | ||||
| 0° | 10.86 (3.49) | 10.01 (3.40) | 8.55 (2.95) | 8.36 (2.71) |
| 90° | 10.04 (2.53) | 9.71 (2.62) | 7.81 (2.00) | 7.45 (1.52) |
| 180° | 9.29 (2.68) | 8.28 (2.38) | 6.16 (1.92) | 6.28 (1.57) |
Abbreviations: KW, knee-to-waist; SW, shoulder-to-waist; WK, waist-to-knee; WS, waist-to-shoulder; ×BW, ×body weight. Note: Knee and lower back (L5/S1) joint average peak joint contact forces (×BW) and SDs during KW, WK, SW, and WS lifting conditions and 0°, 90°, and 180° turn of lift. Values correspond to averages shown in Figure 4A and 4B.
—The across-participant averaged peak normal, shear, and total JCFs in the L5/S1 joint (A) and knee (B) as explored in the second hypothesis (H2). The bars represent the JCF value averaged across the 3 lift degrees (0°, 90°, and 180°) for both the ascension (KW and WS) and descension (WK and SW) lifting tasks in orange. Independent values for each lift degree are labeled as circles in the corresponding color. Significantly different groups are denoted with an asterisk. Statistical significance across all conditions concluded when α = .05. JCFs indicates joint contact forces; KW, knee-to-waist; SW, shoulder-to-waist; WK, waist-to-knee; WS, waist-to-shoulder.
Citation: Journal of Applied Biomechanics 41, 2; 10.1123/jab.2023-0292
—The across-participant averaged peak normal, shear, and total JCFs in the L5/S1 joint (A), and knee (B), as explored in the third hypothesis (H3). The bars represent the JCF value averaged across the 3 lift degrees (0°, 90°, and 180°) for both the below-waist (KW and WK) and above-waist (SW and WS) lifting tasks in orange. Independent values for each lift degree are labeled as circles in the corresponding color. Significantly different groups are denoted with an asterisk. Statistical significance across all conditions concluded when α = .05. JCFs indicates joint contact forces; KW, knee-to-waist; SW, shoulder-to-waist; WK, waist-to-knee; WS, waist-to-shoulder.
Citation: Journal of Applied Biomechanics 41, 2; 10.1123/jab.2023-0292
In the right knee, 180° asymmetric lifts resulted in lower normal, shear, and total JCFs when compared with 0° and 90° lifts (Figure 3B). Peak normal forces were 17.5% lower in 180° versus 90° lifts (P < .01), peak shear forces were 39.2% lower in 180° vs 0° lifts (P < .01), and peak total forces were 18.1% lower in 180° versus 90° lifts (P < .01; Figure 4B, Table 1). Peak shear forces in the knee were 42.1% greater in descension versus ascension lifting tasks (P < .01; Figure 5B). At the knee, peak normal, shear, and total JCFs were 29.6%, 72.8%, and 32.0% higher in below-the-waist versus above-the-waist lifts (P < .01; Figure 6B).
There was a positive linear correlation (R2 = .60, P < .01) between L5/S1 and knee and JCFs. That is, larger knee contact forces were associated with larger L5/S1 contact and vice versa (Figure 7).
—Linear regression comparing peak total (net) contact forces between the L5/S1 and knee joints as explored in the fourth hypothesis (H4). All lift conditions (KW, WK, SW, and WS) and degrees of turn (0°, 90°, and 180°) combinations for all 9 participants are included in the regression analysis. Each participant’s corresponding datapoints are represented by the smaller shapes, while the larger shapes represent across-participant averages for the conditions. JCFs indicates joint contact forces; KW, knee-to-waist; SW, shoulder-to-waist; WK, waist-to-knee; WS, waist-to-shoulder.
Citation: Journal of Applied Biomechanics 41, 2; 10.1123/jab.2023-0292
Discussion
The purpose of this study was to go beyond limb joint analyses and establish a framework to examine how external loads resolve as 3-dimensional JCFs inside limb joints during MMH tasks. Specifically, to gain insight into potential for injury risk from exposure to acute peak loading during repetitive, asymmetric, and physically taxing movements,4 we characterized both lower back (L5/S1) and knee joint loads in occupational MMH lifting tasks across 4 different starting positions and 3 degrees of twisting (Figure 1). We hypothesized that asymmetrical (H1), ascending (H2), and below the waist (H3) lifts would induce the highest JCFs at both the back and knee. Surprisingly, our data did not support H1 and H2 as asymmetric (180○) lifts had lower peak normal, shear, and total JCFs than 0○ and 90○ lifts (Figures 3 and 4) and lifting descension (WK/SW) tasks had higher peak shear JCFs only at the knee compared with ascension (KW/WS) tasks (Figures 3 and 5). In support of H3 and H4, we found that below-the-waist lifting tasks (KW/WK) had higher JCFs than above-the-waist (SW/WS) lifting tasks for both joints (Figure 6) and the knee and L5/S1 peak total JCFs had a positive linear correlation across the task space (Figure 7).
We anticipated asymmetrical lifting would generate more injurious JCFs than symmetrical lifting (H1) in the L5/S1 joint due to elevated shear forces caused by a shift in segmental centers of mass and increased external loading from the hands holding a weighted object.11 However, our results suggest that symmetric lifting actually induced higher peak normal, shear, and total JCFs than asymmetric lifting across conditions (Figures 3 and 4A). Instances of greater lumbar flexion (Supplementary Figure S1B [available online]) aligned with increases in peak JCFs (Figure 3)—a postural shift that causes increased external loading and concomitant increases in muscle activations to stabilize the body, resulting in higher JCFs. Indeed, peak muscle activations, primarily from the erector spinae (ES), and peak L5/S1 joint moments in the sagittal plane were both greater in symmetric lifting than asymmetric lifting (Supplementary Figures S2 and S3 [available online]). In addition to these findings, we believe local peak JCFs are likely caused by large moment arms created through constraints within the execution of each lifting task. Participants in the study were constrained to place 1 foot on each force plate, configuring symmetric lifts side-by-side and asymmetric lifts diagonally adjacent. These stance configurations likely affected how they transported the weight from the initial to terminal shelf position. In 0○ lifts, shelf placements caused participants to stand further away and extend the weight further from their center of mass. Thus, creating a larger moment arm about the L5/S1 joint, larger external moments about the lower back, and ultimately higher compressive and shear forces in the joint. In contrast, the asymmetric lifts (90° and 180°) did not present shelf obstructions, and participants could hold the weight closer to their torso throughout the lift. This technique yielded smaller moment arms between load and L5/S1 joint and resulted in lower compressive and shear forces in the joint.12
We predicted that ascending lifts would induce greater contact forces at the L5/S1 joint than descending lifts (H2) because we assumed the lumbar extension moment generated to counteract gravitational and inertial forces, both from the participant and the added mass being lifted, would be higher during ascending lifts10 (ie, accelerating upward against gravity would be more demanding). However, we found no significant differences in joint loading when comparing ascension versus descension (Figure 5A). We found no significant difference in lifting durations between ascension (2.78 s) and descension (2.76 s). The weight lifted likely did not influence the rate of the lifts in our study or cause inertial effects as large as reported in other studies, where participants lifted up to 40 kg.10,13
We expected that below-the-waist lifting conditions (KW and WK) would cause higher JCFs than above-the-waist (SW and WS) lifting at the L5/S1 joint (H3) due to greater flexion in the trunk, knees, and hips.40 In flexed postures, higher external moments are needed to compensate for poor mechanical advantage against body weight, shifts in centers of mass across segments,11 and the load of the weighted object in participants’ outstretched hands. Our results supported this rationale with below-waist lifts showing higher peak and integrated L5/S1 external moments (Supplementary Figures S2 and S3 [available online]) to go along with higher peak JCFs (Figures 3 and 6A). This confirms the intuition that lifts with higher degrees of trunk flexion places people in a vulnerable and weaker posture48 that leads to higher shear forces17 and increased injury risk.
We expected that adding asymmetrical twisting to a given lift would induce axial moments49 and increase shear forces in the knee leading to higher JCFs when compared with symmetrical lifting (H1). Contrary to our expectation, 90° and 180° turns in lifting had an inconsistent effect on knee joint loading. We propose 2 potential explanations: (1) constrained foot placements altered participants’ squatting stance width and/or (2) nonintuitive mechanics of the quadriceps in deep knee flexion. Escamilla et al14 found knee (tibiofemoral) compressive forces 16% greater in wider stances than narrow stances. Following this line of reasoning, we saw that, on average, participants had wider stances and greater contact forces in 90° asymmetrical lifts (Figure 4B) compared with 0° and 180° lifts. On the other hand, Nisell and Ekholm50 described a “wrapping effect” from the contraction of the quadriceps muscle pulling the patellar tendon posteriorly toward the femur (ie, intercondylar fossa) in very deep flexion, which can act to alleviate knee joint loading by providing an additional contact point for force distribution and transfer.51 In addition, literature suggests peak knee compressive joint forces occur around 90° knee flexion in squatting.51–53 This further supports our contention that the biomechanical technique that participants used when lifting symmetrically resulted in deeper knee flexion and incited the indirect benefits of the wrapping effect to reduce peak contact forces.
During squat descension, the knee extensor moment acts to control deceleration to levels below that induced by gravitational forces. In theory, this external mechanical demand should be less than that in ascent, when the extension moment must first overcome gravitational forces before it can accelerate the limbs—and as a result, we expected JCFs to be lower in descent versus ascent (H2). Surprisingly, we found little difference in JCFs between ascending and descending lifts at the knee. Only shear forces showed differences, with statistically greater contact forces appearing in descension versus ascension (Figure 5B), potentially due to the inertial effects of the added mass being lifted. These results agree with the findings of Dahlkvist et al54 who also observed similarities in knee joint forces between squats in descent and ascent. De Looze et al15 explain that similar muscle activity in ascent and descent lead to loading similarities but suggest that descending tasks are more predisposed to injury occurrence due to muscle lengthening during loading. Similarly, Van Rossom et al55 found that medial knee (tibiofemoral) peak and average shear forces are greater in stair descent than in stair ascent, which further solidifies the idea that the momentum of deceleration leads to greater JCFs. Taken together, we interpret these studies to suggest that lifting may be safer than lowering heavy weight if the goal is to reduce joint loading on the knees.
Our results support our hypothesis that below-the-waist lifting (KW and WK) would generate higher contact forces than above-the-waist (SW and WS) lifting for knee (H3; Figure 6B). Delisle et al56 found that minor knee flexion alleviates some stress from the L5/S1 joint in lifting. Indeed, we observed that above-the-waist lifting conditions in this study were performed with only slight knee flexion compared to below-the-waist lifts. In a nearly upright posture, activations and forces of the quadriceps (knee extensors), hamstrings, and gastrocnemius (knee flexor) are known to be relatively quiescent,54 yielding smaller knee JCFs than in below-the-waist lifting.
In this study, we allowed participants to lift without a predefined technique to assess the effect of natural repetitive lifting on internal joint loading and participants seemed to prefer movements that traded-off loads in lower back and knee. We found that, across all lifting tasks, peak JCFs between the L5/S1 and knee joints scaled linearly (Figure 7) perhaps reflecting a motor control principle to limit peak loading at any one joint. Studies have confirmed that posture is an important factor,10,40 which can influence load distribution across the body or off-loading away from a joint49 in lifting tasks. Lifting techniques, such as stoop (back lift) versus squat (leg lifting),13,15,48,57–59 induce postural constraints that can bias applied external loads and joint moments.49 For example, studies have shown that lifting form is user-specific and is dependent on the starting height of the movement, but also the trunk, knees, and hips become more kinematically coordinated in below-waist lifting; hence why often the preferred lifting technique is a blend between stooping and squatting.9,41 Along these lines, Splittstoesser et al60 conducted a lifting study while kneeling, fixing the role of the knee joint. Interestingly, they found that without contributions from the knee to perform the lift, moving loads to positions above the waist led to greater compressive forces in the back. Thus, highlighting the potential importance of coordinating lower limb joints to reduce loading in the lower back.40 Indeed, Bejjani et al42 concluded that both the back and the knees must bend about 60° and 90°, respectively, to reduce the average force distributed between both joints; thus posing the question if individuals might naturally opt for lifting postures which produce lower average JCFs.
There are several limitations which could have affected our experimental outcomes. We controlled the weight being lifted (11.34 kg), lifting conditions (starting-ending positions and degree of turn), and foot placement; however, the rate of lifting, consistency in reach distance from shelf, and stance width are all factors which were loosely regulated and may have contributed to variance in our results. We collected EMG from 14 muscles across the torso and legs, limiting CEINMS to simulate most muscle activity. Knee JCF results were focused on the right, EMG-instrumented leg, but we expect similar forces on the left knee. Finally, the lack of female representation in this study restricted us from investigating the effects of sex on JCFs.
This study investigated how JCFs in the knee and lower back (L5/S1) vary during manual lifting tasks. We found highest JCFs in symmetric (ie, load straight in front of the body), below the waist lifting tasks. Taken together, our results suggest that asymmetric lifting can actually reduce JCFs in the lower back and knee when compared to moving loads straight in front of the body. Interventions to reduce JCFs during heavy lifting should emphasize transporting weight above the waist or close to the body, while encouraging postures that minimize trunk and knee flexion (ie, small external moment arms), but to still use caution in asymmetric postures. Finally, we found that JCFs were proportionally distributed across knee and lower back joints, suggesting that preferred lifting techniques might be selected to balance JCFs across joints and avoid local overloading—a motor control principle indicating a trade-off that might be leveraged by assistive technology applied at 1 joint (eg, back) to protect another (eg, knee).
Acknowledgments
This work was supported by Sandia National Laboratories, a multimission laboratory managed and operated by the National Technology and Engineering Solutions of Sandia LLC, a wholly owned subsidiary of Honeywell International Inc for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525. This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government. We would like to formally thank Jason Wheeler for his instrumental involvement in the development of this study. We would also like to show our appreciation to all the participants in this study.
References
- 1.↑
Hartvigsen J, Hancock MJ, Kongsted A, et al. What low back pain is and why we need to pay attention. The Lancet. 2018;391(10137):2356–2367. doi:
- 2.↑
Silverstein B, Evanoff B. Musculoskeletal disorders. In: Levy BS, Wegman DH, Baron SL, Sokas RK, McStowe HL. eds. Occupational and Environmental Health: Recognizing and Preventing Disease and Injury. Oxford University Press; 2011:335–365.
- 3.↑
Wurzelbacher SJ, Meyers AR, Lampl MP, et al. Workers’ compensation claim counts and rates by injury event/exposure among state-insured private employers in Ohio, 2007–2017. J Saf Res. 2021;79:148–167. doi:
- 4.↑
Glover W, McGregor A, Sullivan C, Hague J. Work-related musculoskeletal disorders affecting members of the chartered society of physiotherapy. Physiotherapy. 2005;91(3):138–147. doi:
- 5.↑
Milhem M, Kalichman L, Ezra D, Alperovitch-Najenson D. Work-related musculoskeletal disorders among physical therapists: a comprehensive narrative review. Int J Occup Med Environ Health. 2016;29(5):735–747. doi:
- 6.
Schneider SP. Musculoskeletal injuries in construction: a review of the literature. Appl Occup Environ Hyg. 2001;16(11):1056–1064. doi:
- 7.↑
Klein BP, Jensen RC, Sanderson LM. Assessment of workers’ compensation claims for back strains/sprains. J Occup Environ Med. 1984;26(6):443–448. doi:
- 8.↑
McGill SM. Distribution of tissue loads in the low back during a variety of daily and rehabilitation tasks. J Rehabil Res Develop. 1997;34:448–458. PubMed ID: 9323648
- 9.↑
Schipplein OD, Trafimow JH, Andersson GBJ, Andriacchi TP. Relationship between moments at the L5/S1 level, hip and knee joint when lifting. J Biomech. 1990;23(9):907–912. doi:
- 10.↑
Vahdat I, Rostami M, Tabatabai Ghomsheh F, Khorramymehr S, Tanbakoosaz A. Effects of external loading on lumbar extension moment during squat lifting. Int J Occup Med Environ Health. 2017;10:896. doi:
- 11.↑
Jäger M, Luttmann A. The load on the lumbar spine during asymmetrical bi-manual materials handling. Ergonomics. 1992;35(7–8):783–805. doi:
- 12.↑
Kingma I, Bosch T, Bruins L, Van Dieën JH. Foot positioning instruction, initial vertical load position and lifting technique: effects on low back loading. Ergonomics. 2004;47(13):1365–1385. doi:
- 13.↑
Anderson CK, Chaffin DB. A biomechanical evaluation of five lifting techniques. Appl Ergon. 1986;17(1):2–8. doi:
- 14.↑
Escamilla RF, Fleisig GS, Lowry TM, Barrentine SW, Andrews JR. A three-dimensional biomechanical analysis of the squat during varying stance widths: Med Sci Sports Exerc. 2001;33(6):984–998. doi:
- 15.↑
De Looze MP, Toussaint HM, Van Dieën JH, Kemper HCG. Joint moments and muscle activity in the lower extremities and lower back in lifting and lowering tasks. J Biomech. 1993;26(9):1067–1076. doi:
- 16.↑
Davis KG, Marras WS, Waters TR. Evaluation of spinal loading during lowering and lifting. Clin Biomech. 1998;13(3):141–152. doi:
- 17.↑
Budihardjo I. Studies of compressive forces on L5/S1 during dynamic manual lifting. Eng Med. 2002;10:12063. doi:
- 18.↑
Molinaro DD, King AS, Young AJ. Biomechanical analysis of common solid waste collection throwing techniques using OpenSim and an EMG-assisted solver. J Biomech. 2020;104:109704. doi:
- 19.↑
Wellsandt E, Gardinier ES, Manal K, Axe MJ, Buchanan TS, Snyder-Mackler L. Decreased knee joint loading associated with early knee osteoarthritis after anterior cruciate ligament injury. Am J Sports Med. 2016;44(1):143–151. doi:
- 20.
Young People With Old Knees Research Team, Saxby DJ, Bryant AL, et al. Relationships between tibiofemoral contact forces and cartilage morphology at 2 to 3 years after single-bundle hamstring anterior cruciate ligament reconstruction and in healthy knees. Orthop J Sports Med. 2017;5(8):2506. doi:
- 21.↑
Gardinier ES, Manal K, Buchanan TS, Snyder-Mackler L. Clinically-relevant measures associated with altered contact forces in patients with anterior cruciate ligament deficiency. Clin Biomech. 2014;29(5):531–536. doi:
- 22.↑
Karwowski W, Caldwell M, Gaddie P. Relationships Between the NIOSH (1991) Lifting Index, Compressive and Shear Forces on the Lumbosacral Joint, and Low Back Injury Incidence Rate Based on Industrial Field Study. Vol 38. SAGE Publications Sage; 1994:654–657.
- 23.↑
Chen TLW, Wang Y, Wong DWC, Lam WK, Zhang M. Joint contact force and movement deceleration among badminton forward lunges: a musculoskeletal modelling study. Sports Biomech. 2022;21(10):1249–1261. doi:
- 24.↑
Nachemson A, Morris JM. In vivo measurements of intradiscal pressure: discometry, a method for the determination of pressure in the lower lumbar discs. JBJS. 1964;46(5):12. https://journals.lww.com/jbjsjournal/fulltext/1964/46050/in_vivo_measurements_of_intradiscal_pressure_.12.aspx
- 25.↑
Wilke HJ, Neef P, Hinz B, Seidel H, Claes L. Intradiscal pressure together with anthropometric data—a data set for the validation of models. Clin Biomech. 2001;16:S111–S126. doi:
- 26.↑
Fregly BJ, Besier TF, Lloyd DG, et al. Grand challenge competition to predict in vivo knee loads. J Orthop Res. 2012;30(4):503–513. doi:
- 27.↑
Bernard TM, Ayoub MM, Lin CJ. Evaluation of a biomechanical simulation model for sagittal plane lifting. Int J Ind Ergon. 1999;24(2):157–171. doi:
- 28.↑
Lin CJ, Ayoub MM, Bernard TM. Computer motion simulation for sagittal plane lifting activities. Int J Ind Ergon. 1999;24(2):141–155. doi:
- 29.↑
Beaucage-Gauvreau E, Robertson WSP, Brandon SCE, et al. Validation of an OpenSim full-body model with detailed lumbar spine for estimating lower lumbar spine loads during symmetric and asymmetric lifting tasks. Comput Methods Biomech Biomed Eng. 2019;22(5):451–464. doi:
- 30.↑
Delp SL, Anderson FC, Arnold AS, et al. OpenSim: Open-source software to create and analyze dynamic simulations of movement. IEEE Trans Biomed Eng. 2007;54(11):1940–1950. doi:
- 31.
Bedo BLS, Catelli DS, Lamontagne M, Santiago PRP. A custom musculoskeletal model for estimation of medial and lateral tibiofemoral contact forces during tasks with high knee and hip flexions. Comput Methods Biomech Biomed Eng. 2020;23(10):658–663. doi:
- 32.
Garg A, Chaffin DB, Freivalds A. Biomechanical stresses from manual load lifting: a static vs dynamic evaluation. AIIE Trans. 1982;14(4):272–281. doi:
- 33.↑
Cholewicki J, McGill SM, Norman RW. Comparison of muscle forces and joint load from an optimization and EMG assisted lumbar spine model: towards development of a hybrid approach. J Biomech. 1995;28(3):321–331. doi:
- 34.↑
Pizzolato C, Lloyd DG, Sartori M, et al. CEINMS: a toolbox to investigate the influence of different neural control solutions on the prediction of muscle excitation and joint moments during dynamic motor tasks. J Biomech. 2015;48(14):3929–3936. doi:
- 35.↑
Hoang HX, Diamond LE, Lloyd DG, Pizzolato C. A calibrated EMG-informed neuromusculoskeletal model can appropriately account for muscle co-contraction in the estimation of hip joint contact forces in people with hip osteoarthritis. J Biomech. 2019;83:134–142. doi:
- 36.↑
Uhlrich SD, Jackson RW, Seth A, Kolesar JA, Delp SL. Muscle coordination retraining inspired by musculoskeletal simulations reduces knee contact force. Sci Rep. 2022;12(1):9842. doi:
- 37.↑
McCain EM, Dalman MJ, Berno ME, et al. The influence of induced gait asymmetry on joint reaction forces. J Biomech. 2023;153:111581. doi:
- 38.↑
Gagnon M, Plamondon A, Gravel D. Pivoting with the load: an alternative for protecting the back in asymmetrical lifting. Spine. 1993;18(11):9010. https://journals.lww.com/spinejournal/fulltext/1993/09010/pivoting_with_the_load__an_alternative_for.17.aspx
- 39.↑
Skals S, Bláfoss R, De Zee M, Andersen LL, Andersen MS. Effects of load mass and position on the dynamic loading of the knees, shoulders and lumbar spine during lifting: a musculoskeletal modelling approach. Appl Ergon. 2021;96:103491. doi:
- 40.↑
Splittstoesser R, Davis K, Marras W. Trade-Offs Between Trunk Flexion, Hip Flexion, and Knee Angle in Lifting Below Waist Level. Vol. 44. SAGE Publications Sage; 2000:5–9.
- 41.↑
Davis KG, Splittstoesser RE, Marras WS. Kinematic contribution and synchronization of the trunk, hip, and knee during free-dynamic lifting. Occup Ergon. 2003;3(2):99–108. doi:
- 42.↑
Bejjani FJ, Gross CM, Pugh JW. Model for static lifting: relationship of loads on the spine and the knee. J Biomech. 1984;17(4):281–286. doi:
- 43.↑
Ng JKF, Kippers V, Parnianpour M, Richardson CA. EMG activity normalization for trunk muscles in subjects with and without back pain: Med Sci Sports Exerc. 2002;34(7):1082–1086. doi:
- 44.↑
Vera-Garcia FJ, Moreside JM, McGill SM. MVC techniques to normalize trunk muscle EMG in healthy women. J Electromyogr Kinesiol. 2010;20(1):10–16. doi:
- 45.↑
Stegeman D, Hermens H. Standards for surface electromyography: The European project Surface EMG for non-invasive assessment of muscles (SENIAM). Enschede: Roessingh Res Develop. 2007;10:8–12.
- 46.↑
Hamner SR, Seth A, Delp SL. Muscle contributions to propulsion and support during running. J Biomech. 2010;43(14):2709–2716. doi:
- 47.↑
Vigotsky AD, Zelik KE, Lake J, Hinrichs RN. Mechanical misconceptions: have we lost the “mechanics” in “sports biomechanics”? J Biomech. 2019;93:5. doi:
- 48.↑
Kumar S. Spinal compression at peak isometric and isokinetic exertions in simulated lifting in symmetric and asymmetric planes. Clin Biomech. 1996;11(5):281–289. doi:
- 49.↑
Gagnon M, Plamondon A, Gravel D, Lortie M. Knee movement strategies differentiate expert from novice workers in asymmetrical manual materials handling. J Biomech. 1996;29(11):1445–1453. doi:
- 50.↑
Nisell R, Ekholm J. Patellar forces during knee extension. Scand J Rehabil Med. 1985;17(2):63–74.
- 51.↑
Hartmann H, Wirth K, Klusemann M. Analysis of the load on the knee joint and vertebral column with changes in squatting depth and weight load. Sports Med. 2013;43:993–1008. doi:
- 52.
Wilk KE, Escamilla RF, Fleisig GS, Barrentine SW, Andrews JR, Boyd ML. A comparison of tibiofemoral joint forces and electromyographic activit during open and closed kinetic chain exercises. Am J Sports Med. 1996;24(4):518–527. doi:
- 53.↑
Huberti HH, Hayes WC. Patellofemoral contact pressures. The influence of q-angle and tendofemoral contact. J Bone Joint Surg Am. 1984;66(5):715–724. https://journals.lww.com/jbjsjournal/fulltext/1984/66050/patellofemoral_contact_pressures__the_influence_of.10.aspx
- 54.↑
Dahlkvist NJ, Mayo P, Seedhom BB. Forces during squatting and rising from a deep squat. Eng Med. 1982;11(2):69–76. doi:
- 55.↑
Van Rossom S, Smith CR, Thelen DG, Vanwanseele B, Van Assche D, Jonkers I. Knee joint loading in healthy adults during functional exercises: implications for rehabilitation guidelines. J Orthop Sports Phys Ther. 2018;48(3):162–173. doi:
- 56.↑
Delisle A, Gagnon M, Desjardins P. Knee flexion and base of support in asymmetrical handling: effects on the worker’s dynamic stability and the moments of the L5/S1 and knee joints. Clin Biomech. 1998;13(7):506–514. doi:
- 57.↑
Hwang S, Kim Y, Kim Y. Lower extremity joint kinetics and lumbar curvature during squat and stoop lifting. BMC Musculoskelet Disord. 2009;10(1):15. doi:
- 58.
Burgess-Limerick R, Abernethy B. Qualitatively different modes of manual lifting. Int J Ind Ergon. 1997;19(5):413–417. doi:
- 59.↑
Meyers JM, Fathallah FA, Janowitz I. Stooped and squatting postures in the workplace. Natl Ag Saf Database. 2015;10:2–7.
- 60.↑
Splittstoesser RE, Yang G, Knapik GG, et al. Spinal loading during manual materials handling in a kneeling posture. J Electromyogr Kinesiol. 2007;17(1):25–34. doi:
