The Effects of Lifting Techniques on the L5-S1 Joint: Lifting Different Loads from Ground Level

doi:10.21203/rs.3.rs-4864263/v1

Download PDF

Article

The Effects of Lifting Techniques on the L5-S1 Joint: Lifting Different Loads from Ground Level

https://doi.org/10.21203/rs.3.rs-4864263/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Manual material lifting is a common activity in daily life and industrial work conditions, significantly affecting the L5/S1 joint in the lower back. This study replaces lifting objects with handles on both sides, as found in literature, with lifting industrial cargo boxes from the bottom using both hands. Experiments involved 5 healthy male cargo workers lifting weights of 4, 8, 12, and 16 kg from ground level using squat and stoop lifting techniques. Body joint positions and ground reaction forces were captured without markers using the Microsoft Kinect v2 sensor. These data were transferred to a 3D human model in the Opensim biomechanical analysis program for inverse kinematic and dynamic analyses. The force and moment values at the L5/S1 joint for each subject, weight, and lifting technique were compared. In conclusion, it was determined that, depending on the load, the squat lifting technique increased the torque values at the L5-S1 joint by 7.1–9.9%, increased the compression force by 8.8–9.2%, and decreased the shear force by 24.5–25.3% compared to the stoop lifting technique.

musculoskeletal system

lifting techniques

opensim

kinectv2

Despite the technological developments in industrial works and the widespread use of automation, various works in many sectors are mostly done manually. From business processes in distribution and storage centers; If we consider the manual lifting, dropping and carrying processes of materials; for example, about 80% of industrial enterprises in Europe and the USA are still manually operated [1, 2]. Frequent and repetitive manual lifting activities and unsuitable body postures while performing the lifting movement may lead to musculoskeletal disorders (MSD). Low back pain (LBP) from lifting and lowering over-exertion is a major source of injury and illness for manual material handling workers. Approximately 40% of occupational diseases in the world and 38% of the USA are caused by low back diseases. These disorders are important reasons for not being able to continue work [3]. According to the Bureau of Labor Statistics (BLS) 2016 report, overexertion is the leading cause of non-fatal occupational incidents, accounting for 33% of total cases that require an average of 12 days to recover [4].

Lumbar spine during lifting activities many biomechanical models have been developed to predict spinal loading. Some of these are static models to estimate the L4/L5 and L5/S1 compression force [5, 6], electromyography (EMG) assisted dynamic spine models [7, 8], models using three- dimensional finite element analysis [9], dynamic models in which body segments are examined separately and three-dimensional geometric body models [10]. 3D musculoskeletal models make dynamic motion analysis easier than traditional methods. With dynamic analysis, whole body kinematics and kinetics are determined, and musculoskeletal forces are simulated.

Faber et al. [11] designed a biomechanical model for estimating L5/S1 moments. Koopman et al. [12] subjected lower back loading to mechanical evaluation. While these models are being designed, the location information of the L5/S1 joint and other body parts is needed. Motion capture systems are used to meet this need. The most widely used of these is the motion capture system made with markers placed on the skin at certain locations on the body. Position information obtained from beacons and sensors is used to estimate the position of body joints. However, there are various difficulties in the implementation of this system. First of all, the installation cost of the system is high, and special motion-tracking laboratories are needed. It makes it difficult to make measurements in the studies planned to be done in the field [13].

In subsequent studies, researchers investigated alternative methods to the motion tracking system using pointers and sensors. In some of these studies, the markerless motion capture method was used. In this method, the use of the Kinect v2 camera was preferred because of its low cost and ease of use compared to motion capture with pointers. As examples of these studies, do Carmo Vilas-Boas et al. [14] used a Kinect camera for gait analysis, Roozbahani et al. [15] used real-time simulations of human movements, Tamura et al. used a Kinect camera to measure the joint angles of the body trunk limbs during walking on a treadmill and outside walking on flat ground.

Xu and McGorry [16] used a 25-joint skeleton model of the Kinect camera without markers to estimate the position of the L5/S1 joint. However, significant errors were found in the L5/S1 joint position estimation compared to the commonly used pointer motion capture method. Inaccurate results may occur when Matthew et al. [17] use the original Kinect skeleton model output to perform L5/S1 moment evaluations. Therefore, it is worth developing a method that can reduce the estimation error of the L5/S1 location while maintaining the advantages of using the Kinect unsigned approach.

Kinect has the advantages of having a human body skeleton identification function, portability and ease of use. However, the authors did not use advanced biomechanical models with Kinect to analyze the L5/S1 joint position and dynamic motion during lifting motion. Therefore, the current research used Opensim software and Rajagopal's whole-body musculoskeletal model for dynamic analysis of the L5/S1 joint based on the Kinect unsigned skeletal model. It was investigated whether the measurements obtained from the Kinect skeletal model for the estimation of loads acting on the L5/S1 joint during symmetrical squat and stoop lifting tasks could produce results similar to previous studies by transferring them to a biomechanical model and dynamic analysis. Specifically, this study measured lifting postures using a Kinect and a ground reaction meter and imported joint position data from the Kinect into the biomechanical model in Opensim. The results obtained were evaluated together with the results of previous studies.

This study was conducted by five healthy male participants (age: 27 years (± 5) body mass index (BMI): 23.34 kg/m2 (± 1.14)) working in the cargo industry. All methods were carried out in accordance with relevant guidelines and regulations. All experimental protocols were approved by the Ethics Committee of Kütahya Dumlupınar University (Ethics Committee Decision No: 2021/13–16). Informed consent was obtained from all participants before the study. The participants were informed before the start of the study and the relevant movement techniques were supervised throughout the study.

2.1. Equipment

2.1.1. Microsoft Kinect v2

Kinect v2, which is used as markerless motion capture device in this study, was developed by Microsoft for use on the Xbox game console. It is the second version of Kinect, an affordable and enhanced version of the device with a depth sensor. Microsoft Kinect v2 estimates depth information based on time-of-flight (ToF) principle using infrared (IR) emitters and detector. If we look at the features of the Kinect v2 device; An RGB camera with a resolution of 1920x1080 pixels acquires the color information of the image, obtains an image with a depth resolution of 512x426 pixels with time of flight (ToF) technology and allows to obtain images at 30 frames per second. In addition, the device has a depth sensing feature of 60 degrees vertically and 70 degrees horizontally. It can detect 25 of the user's body joints in Fig. 1 with various programming languages and software programs (e.g. Microsoft Kinect SDK etc.). In the study, the Kinect V2 sensor was installed by placing it 2.5 m in front of the user and 1.5 m above the ground, respectively [18].

2.1.2. Ground Reaction Force Measurement

Zebris™ FDM-2 device will be used for ground reaction force measurement while each lifting movement is applied. This device, shown in Fig. 2, is 2120 mm long, 605 mm wide and 21 mm high, with a frequency sampling rate of 100–200 Hz with 15360 sensors on it and is used for gait and balance assessment. The data obtained from the device is recorded by turning it into a report via the Zebris software installed on the computer. The user will be asked to remove the boxes by climbing onto the platform. Static stance evaluation is obtained from data consisting of center of pressure changes (ellipse area (mm²)), total load on the right and left foot, and antero-posterior foot [19].

2.2. Musculoskeletal Model

In this study, the whole-body model created by Rajagopal et al. [20] was used. The dimensions of the skeletal model represent a 76 kg, 172 cm tall male, and the bone geometry is also suitable for these dimensions. There are 22 bodies consisting of joints in the structure of the model. Trunk parts, pelvis, femur (right and left), calcaneus, talus, patella, tibia/fibula, and toes representing the lower body and a combined head and trunk and right and left humerus, ulna, radius, and hand representing the upper body. The model has 37 degrees of freedom. The model has 20 degrees of freedom in the lower body (pelvis-6 and right leg 7, left leg 7), and 17 degrees of freedom (waist joint-3 and right arm 7, left arm 7 in the upper body). The coordinate system for each of the model body parts is aligned with the x-direction facing forward, the y-direction facing upward, and the z-direction to the right [19].

2.3. Experiment Procedure

The participants performed two standard lifting techniques: squat and stoop lifting. Necessary explanations about the stoop and squat lifting techniques were made verbally and visually, and preliminary studies were carried out before the experiments began. The participants were asked to perform lifting movements of 4, 8, 12 and 16 kg with 4 different weights. The dimensions of the lifted object were 40 cm x 40 cm x 35 cm (width x depth x height). During the squat lifting movement, the information was given that the knees should be fully flexed, the heels should be lifted in such a way that the contact with the ground was cut off, and the body should be close to the vertical position (Fig. 3a). During the stoop lift, the participant was instructed to extend the knees without bending and bend the torso to lift the box from the ground in an upright position (Fig. 3b) [21].

The cargo box was placed in the middle of the sagittal plane. The participants lifted the box on the ground (0 cm above the ground) by grasping it symmetrically from below the bottom with both hands and held it upright at waist level. The participants were instructed to adopt the squat and stoop lifting technique (knees and torso slightly bent during lifting). Because the lateral position of the pelvis might alter flexion and torsional torques, the participant was not permitted to shift their feet throughout the lifting [22]. In addition, the participant was asked to complete the lift in approximately two seconds (from the first lift until holding the standing box at waist level) to minimize the time effect of the L5/S1 joint on loading [50]. For each lifting condition, five repetitions were conducted (200 lifts total). The eight lifting conditions were assigned at random, and the participant was given two minutes to rest in a sitting position between each lifting condition.

In Fig. 4, kinematic data was collected using a single Kinect (v2) sensor. With this sensor, data was collected at a sampling rate of 30Hz during the lifting movement. In kinetic analysis, Zebris FDM 2 platform with 100Hz sampling rate to detect ground reaction forces (GRF) and moments (Medical GmbH · Germany) was used. The measured data was processed using a standard musculoskeletal modeling workflow implemented in OpenSim 4.0 [23, 24] to calculate joint angles and moments. For the skeletal-muscle model used in the study, Rajagopal et al. The musculoskeletal model he created was taken as a reference. Half the weight of the box is added to the mass of the right and left hands to add the effect of the box to the model.

First, the biomechanical model was scaled according to the participant's anthropometry. Body joint angles were calculated using the body joint position information from the Kinect sensor as a result of lifting movements using the inverse kinematics calculation method [20]. With the obtained data, the L5/S1 joint moments were calculated using OpenSim's Inverse Dynamic Tool.

2.4. Data Processing in Opensim

Position data of human body joints from Kinect v2 and ground reaction force (GRF) data from the Zebris FDM 2 platform were entered into the human model in OpenSim. The size data of the body parts of the skeletal-muscle model used were scaled according to participant anthropometry Using position information from participant static posture data, the scaling tool enabled the building of a subject-specific model by matching the size of body components as nearly as feasible. After scaling, the inverse kinematics tool in Fig. 5 was used to construct three-dimensional joint angles or body joint kinematics. In order to determine the L5/S1 joint moments, inverse dynamics were used [23, 24]. Then, static optimization was performed as an extension of inverse kinematics to decompose the net joint moments into individual muscle forces. Finally, joint reaction analysis is performed to calculate the resultant forces transferred between the coupled bodies derived from the loads acting on the model. lumbar Spinal loading was calculated by solving the dynamic equations with the input of gravitational acceleration and joint moment.

The participant performed the box at his own comfortable lifting speed. Figure 6 depicts the waist (L5/S1) in the sagittal plane during lifting, as seen from the lower extremity joint angles. Although different weights were lifted in the same lifting technique, no significant difference in range of motion (ROMs) was found. The ROMs for the same joint (waist), however, were significantly different between the two techniques.

In general, the stoop and squat lifting motion of the spine are associated with the 3-D lumbar spine at the L5/S1 joints. The graphs presented in Fig. 7 compare the extension moments occurring in the lumbar region (L5/S1) of the spine during the lifting of various weights (4, 8, 12 and 16 kg) using squat and stoop lifting techniques. In all the graphs, the moment values are highest at the beginning of the lifting process for both squat and stoop positions. Differences in maximum L5/S1 extension moments between squat and stoop lifting were negligible at 4 and 8 kg. In contrast, it was greater when lifting 12 and 16 kg in squat than in stoop lifting.

Joint reaction forces in the SI (superior-inferior) and AP (anterio-posterior) directions were analyzed for the L5/S1 joint. The resulting forces correspond to compression force and shear forces, respectively (Fig. 8–11). When examining the compression forces at the L5/S1 joint during lifting movements, the values were similar across all five subjects. The compression forces varied according to the lifting tasks. In the box lifting experiments with different weights, the compression forces depicted in Fig. 8 were obtained as 49.38 N/kg – 60.95 N/kg (3911 N − 4827 N) for squat lifting, and 44.05 N/kg – 55.78 N/kg (3489 N − 4418 N) for stoop lifting.

To compare the manual lifting movement experiments with studies in the literature, the obtained compression force data were evaluated as N values per unit kg. In our study, the L5/S1 joint compression forces for 12 kg and 16 kg loads were found to be 57.01–60.95 N/kg for squat lifting, decreasing by 9% and 8.5% to 51.83–55.78 N/kg for stoop lifting, respectively. Arx et al. (2021) found 49.6 N/kg for squat lifting, decreasing by 5% to 47 N/kg for stoop lifting; Beaucage-Gauvreau et al. (2019) found 63 N/kg for squat lifting, decreasing by 17% to 52 N/kg for stoop lifting; Kingma et al. (2016) found 51.6 N/kg for squat lifting, increasing by 17% to 60.4 N/kg for stoop lifting (Fig. 9) [25–27].

When we examined the shear forces at the L5/S1 joint during lifting movements, unlike the compression forces, the shear forces for stoop lifting were greater than those for squat lifting. The shear forces obtained were 9.23 N/kg – 15.64 N/kg (731 N − 1239 N) for squat lifting, and 12.25 N/kg – 19.48 N/kg (970 N − 1543 N) for stoop lifting (Fig. 10).

To compare the manual lifting movement experiments with studies in the literature, the obtained shear force data were evaluated as N values per unit kg. In our study, the L5/S1 joint shear forces for 12 kg and 16 kg loads were found to be 12.65–15.62 N/kg for squat lifting, increasing by 25% to 15.83–19.48 N/kg for stoop lifting, respectively. Arx et al. (2021) found 23.8 N/kg for squat lifting, decreasing by 11% to 21.1 N/kg for stoop lifting; Beaucage-Gauvreau et al. (2019) found 16.2 N/kg for squat lifting, increasing by 20% to 19.4 N/kg for stoop lifting; Kingma et al. (2016) found 19 N/kg for squat lifting, decreasing by 9.5% to 17.2 N/kg for stoop lifting (Fig. 11) [25–27].

This study examines the impact of manual material lifting on the L5/S1 joint, which is most at risk of injury in the lumbar region of individuals. In studies investigating the effect of manual lifting on the L5/S1 joint, it has been observed that lifting loads within objects such as boxes and crates typically involves lifting from handles located on the sides and at specific heights from the ground. However, our study considers the absence of handles on boxes used by a cargo company, and the lifting movement is performed by holding the bottom of the box in our biomechanical model. This approach aims to achieve realistic results with cargo workers. During symmetric lifting tasks, data captured through a markerless motion capture method were applied to an existing full-body OpenSim biomechanical model. The human skeletal dimensions in the model were scaled specifically for each participant. The L5/S1 joint loads were estimated through inverse kinematic and dynamic analyses. The loads affecting the L5-S1 joint were analyzed by calculating the compression force and shear force values for maximum torso angle postures.

When examining the compression force values at the L5-S1 joint, it was found that the compression force values for squat lifting were approximately 9% lower than those for stoop lifting (for 12kg and 16kg). Compared to analyses with handle-holding object lifting, the compression force values have decreased by approximately 17% and 5%, respectively, in the studies of Beaucage-Gauvreau et al. (2019) and Arx et al. (2021) [25, 27]. Regarding shear force values, it was observed that shear force values for squat lifting increased by approximately 25% compared to stoop lifting (for 12kg and 16kg). In the study by Beaucage-Gauvreau et al. (2019), an increase of approximately 20% was observed, while in the study by Arx et al. (2021), a decrease of approximately 11% was noted. It is believed that the differences are due to the lifting weights and the gripping distances of the object [25, 27].

This study has certain limitations. The muscle force in this study was estimated through simulation based on specific assumptions. The evaluation of the biomechanical model did not include spinal ligaments and intra-abdominal pressure. Future research could explore the lumbar loads between different lifting techniques, expanding the parameters of the participant performing the lifting movement and evaluating the lifted object's dimensions.

Acknowledgements

This study, derived from the doctoral thesis of Melih Canlıdinç (Supervisor: Dr. Mustafa GÜLEŞEN), has been conducted under the approval of the ethics committee, with findings and analyses performed within the framework of ethical guidelines (Ethics Committee Decision No: 2021/13-16).

Data Availability

The datasets generated and/or analyzed during the current study are not publicly available due to confidentiality agreements with the participants' employer, which restrict the sharing of data. However, the data may be available from the corresponding author on reasonable request and with permission from the employer.

de Koster, R., Le-Duc, T., and Roodbergen, K. J., Design and Control of Warehouse Order Picking: A Literature Review (2007).
Peerless Research Group, “October 2017 September 2018”, Logist. Manag. Mod. Mater. Handl., (October 2017), pp. 1–38 (2018).
Bureau of Labor Statistics, Industry Injury and Illness Data (2019).
Bureau of Labor Statistics, U., “2016 SURVEY OF OCCUPATIONAL INJURIES & ILLNESSES CHARTS PACKAGE” (2017).
Bruno, A. G., Bouxsein, M. L., and Anderson, D. E., “Development and validation of a musculoskeletal model of the fully articulated thoracolumbar spine and rib cage”, J. Biomech. Eng., 137(8), pp. 1–10 (2015).
Rajaee, M. A., Arjmand, N., Shirazi-Adl, A., Plamondon, A., and Schmidt, H., “Comparative evaluation of six quantitative lifting tools to estimate spine loads during static activities”, Appl. Ergon., 48, pp. 22–32 (2015).
Jia, B., Kim, S., and Nussbaum, M. A., “An EMG-based model to estimate lumbar muscle forces and spinal loads during complex, high-effort tasks: Development and application to residential construction using prefabricated walls”, Int. J. Ind. Ergon., 41(5), pp. 437–446 (2011).
McGill, S. M., Marshall, L., and Andersen, J., “Low back loads while walking and carrying: Comparing the load carried in one hand or in both hands”, Ergonomics, 56(2), pp. 293–302 (2013).
Bazrgari, B., Shirazi-Adl, A., and Arjmand, N., “Analysis of squat and stoop dynamic liftings: Muscle forces and internal spinal loads”, Eur. Spine J., 16(5), pp. 687–699 (2007).
Kingma, I., Bosch, T., Bruins, L., and van Dieën, J. H., “Foot positioning instruction, initial vertical load position and lifting technique: Effects on low back loading”, Ergonomics, 47(13), pp. 1365–1385 (2004).
Faber, G. S., Kingma, I., Chang, C. C., Dennerlein, J. T., and van Dieën, J. H., “Validation of a wearable system for 3D ambulatory L5/S1 moment assessment during manual lifting using instrumented shoes and an inertial sensor suit”, J. Biomech., 102, p. 109671 (2020).
Koopman, A. S., Kingma, I., Faber, G. S., Bornmann, J., and van Dieën, J. H., “Estimating the L5S1 flexion/extension moment in symmetrical lifting using a simplified ambulatory measurement system”, J. Biomech., 70, pp. 242–248 (2018).
Gholipour, A. and Arjmand, N., “Artificial neural networks to predict 3D spinal posture in reaching and lifting activities; Applications in biomechanical models”, J. Biomech., 49(13), pp. 2946–2952 (2016).
Vilas-Boas, M. do C., Choupina, H. M. P., Rocha, A. P., Fernandes, J. M., and Cunha, J. P. S., “Full-body motion assessment: Concurrent validation of two body tracking depth sensors versus a gold standard system during gait”, J. Biomech., 87, pp. 189–196 (2019).
Roozbahani, H., Alizadeh, M., Ustinov, S., and Handroos, H., “Development of a novel real-time simulation of human skeleton/muscles”, J. Biomech., 114 (2021).
Xu, X. and McGorry, R. W., “The validity of the first and second generation Microsoft Kinect^TM for identifying joint center locations during static postures”, Appl. Ergon., 49, pp. 47–54 (2015).
Matthew, R. P., Seko, S., Bajcsy, R., and Lotz, J., “Kinematic and Kinetic Validation of an Improved Depth Camera Motion Assessment System Using Rigid Bodies”, IEEE J. Biomed. Heal. informatics, 23(4), pp. 1784–1793 (2019).
Sarbolandi, H., Lefloch, D., and Kolb, A., “Kinect range sensing: Structured-light versus Time-of-Flight Kinect”, Comput. Vis. Image Underst., 139, pp. 1–20 (2015).
https://www.zebris.de/fileadmin/Editoren/zebris-PDF/zebris-Prospekte-EN/27_9_FDM_EN_150.pdf, “The zebris FDM System-gait and roll-off analysis in practice” 25/04/2022.
Rajagopal, A., Dembia, C. L., DeMers, M. S., Delp, D. D., Hicks, J. L., and Delp, S. L., “Full-Body Musculoskeletal Model for Muscle-Driven Simulation of Human Gait”, IEEE Trans. Biomed. Eng., 63(10), pp. 2068–2079 (2016).
Wang, Z., Wu, L., Sun, J., He, L., Wang, S., and Yang, L., “Squat, stoop, or semi-squat: A comparative experiment on lifting technique”, J. Huazhong Univ. Sci. Technol. - Med. Sci., 32(4), pp. 630–636 (2012).
Plamondon, A., Gagnon, M., and Gravel, D., “Moments at the L5/S1 joint during asymmetrical lifting: effects of different load trajectories and initial load positions”, Clin. Biomech., 10(3), pp. 128–136 (1995).
Seth, A., Hicks, J. L., Uchida, T. K., Habib, A., Dembia, C. L., Dunne, J. J., Ong, C. F., DeMers, M. , Rajagopal, A., Millard, M., Hamner, S. R., Arnold, E. M., Yong, J. R., Lakshmikanth, S. K., Sherman, M. A., Ku, J. P., and Delp, S. L., “OpenSim: Simulating musculoskeletal dynamics and neuromuscular control to study human and animal movement”, PLoS Comput. Biol., 14(7) (2018).
Delp, S. L., Anderson, F. C., Arnold, A. S., Loan, P., Habib, A., John, C. T., Guendelman, E., and Thelen, D. G., “OpenSim: Open-source software to create and analyze dynamic simulations of movement”, IEEE Trans. Biomed. Eng., 54(11), pp. 1940–1950 (2007).
Beaucage-Gauvreau, E., Robertson, W. S. P., Brandon, S. C. E., Fraser, R., Freeman, B. J. C., Graham, R. B., Thewlis, D., and Jones, C. F., “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. Engin., 22(5), pp. 451–464 (2019).
Kingma, I., Faber, G. S., and van Dieën, J. H., “Supporting the upper body with the hand on the thigh reduces back loading during lifting”, J. Biomech., 49(6), pp. 881–889 (2016).
von Arx, M., Liechti, M., Connolly, L., Bangerter, C., Meier, M. L., & Schmid, S., “From stoop to squat: A comprehensive analysis of lumbar loading among different lifting styles”, Frontiers in bioengineering and biotechnology, 9, 1070, (2021).

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

The Effects of Lifting Techniques on the L5-S1 Joint: Lifting Different Loads from Ground Level

Status:

Version 1

Abstract

Figures

1. Introduction

2. Methods

2.1. Equipment

2.1.1. Microsoft Kinect v2

2.1.2. Ground Reaction Force Measurement

2.2. Musculoskeletal Model

2.3. Experiment Procedure

2.4. Data Processing in Opensim

3. Results

4. Discussion and Conclusion

Declarations

References

Additional Declarations

Status:

Version 1