Introduction

Low back pain (LBP) is a significant health problem worldwide.3 Eighty percent of all individuals will experience LBP at some point in their lives.47,49 One-third of the adults in the United States will develop some degrees of back pain each year,2 seven to fifteen percent of the US adults report some daily restriction of activities due to back pain within the previous year, and there are ~700,000 workers’ compensation claims for work-related back pain each year.48 LBP also causes substantial economic burden each year around the world.31 In the United States, it was estimated that annual cost was close to 20 billion dollars.44

The exact etiology of LBP is still unclear,20,22,37 however, previous studies have identified a number of risk factors that are associated with the occurrence of LBP, such as high spinal compression10,11 and shear loading,25,26 uneven ground condition14,15,19,35 and lumbar tissue creep.39,40 Previous studies also demonstrated that lumbar extensor muscle fatigue could reduce trunk and spinal stability4,13 and alter the neuromuscular coordination for lumbar tissues during trunk bending;6 however, the direct association between lumbar extensor muscle fatigue and LBP is still not well understood. Previously, studies have found that lumbar extensor muscle fatigue alters the onset and cessation of myoelectric silence during the performance of flexion–extension tasks;6 such change indicated a shift of loading from active muscles to lumbar passive tissues. From the spinal stability point of view, one study demonstrated that lumbar extensor muscle fatigue increased body sway during standing possibly due to decreased muscle proprioceptive acuity, impaired postural control and reduced trunk stability.4 Another study discovered elevated trunk muscle co-contraction after lumbar muscle fatigue.13 It was concluded that the increase of muscle co-contraction was used to compensate for the reduced stability and to increase trunk stiffness. A following study showed that lumbar extensor fatigue led to reduced spinal dynamic stability and reduced dynamic complexity during sagittal flexion and extension motion9 which could increase the chance of uncontrollable intervertebral disk movement and tissue injury when experience perturbation.

Lumbar extensor muscle fatigue could also change lifting kinematics. Previous studies have documented that lumbar extensor muscle fatigue introduced by repetitive lifting could reduce trunk stability and cause reduced knee and hip range of motion and increased peak lumbar flexion angle during lifting.42,43 Understanding trunk kinematics is essential for the assessment of LBP risks. Especially in flexed postures, trunk motion to a large extent determines the magnitude of spinal loading18,29,46) which is directly associated with the occurrence of LBP.23,24,27 A recent study found that changes in the motion synergy between different trunk motion segments (e.g., lumbar, pelvis) during lifting could also significantly influence the spinal loading.45 However, the effect of lumbar extensor muscle fatigue on trunk motion synergy during the performance of trunk flexion and extension tasks has not been investigated.

Previous studies have demonstrated that LBP patients generate different lumbar pelvic motion rhythms in comparison to non-symptomatic individuals.38,41 In clinical settings such difference could be used for patient screening and monitoring the rehabilitation progress of patients. Understanding the effect of lumbar muscle fatigue on the lumbar pelvic motion rhythm could help physicians improve the accuracy of diagnostic and rehabilitation assessment of LBP patients.

The objective of this current study was to understand the effect of lumbar extensor muscle fatigue on lumbar–pelvic motion rhythm during weight lifting and lowering motions. More specifically, we investigated the effect of lumbar extensor muscle fatigue on normalized lumbar–pelvic rotation ratio and the corresponding L5/S1 joint loading. Based on the existing evidence,45 it was hypothesized that, with the influence of lumbar extensor muscle fatigue, non-symptomatic subjects will demonstrate increased normalized lumbar–pelvic rotation ratio and elevated spinal loading during trunk flexion and extension motion.

Methods

Participants

Thirteen male subjects participated in this study with signed consent form. Their mean age, height, BMI and mass were 27.0 ± 2.5 years, 175.2 ± 5.0 cm, 22.3 ± 2.3, and 68.8 ± 9.8 kg, respectively. The experiment protocol received approval from the West Virginia University Institutional Review Board. Subjects with a history of low back surgery or experienced lower back pain within the past 12 months were excluded from the current study.

Instrumentation

A three-dimensional, magnetic field based motion tracking system (Motion Star, Ascension Technology Corporation, Burlington, VT, USA) was used to measure lumbar and pelvis kinematics in this experiment. Three motion sensors were attached to the skin over the spinous processes at C7, T12, and S1 levels respectively (Fig. 1a). Muscle electromyography (EMG) data were collected from the left and right side of paraspinal muscles from L3 (4 cm lateral from the L3 spinous process) and L4 levels (2 cm lateral from the L4 spinous process) using bipolar surface EMG electrodes (Bagnoli, Delsys, Boston, MA, USA) (Fig. 1a). The MotionMonitor software (MotionMonitor, Innovative Sports Training, Chicago, IL, USA) was used to collect kinematics and EMG data simultaneously with a sampling frequency of 1024 Hz. A custom made wood structure was used to standardize subject’s posture and reduce stress and strain on lower extremity muscles and ligaments during the performance of lumbar extensor muscle fatiguing tasks. This structure is the combination of a standing platform and an attached leg support. A nylon safety belt was used to secure subject’s pelvis to the leg support in order to minimize lower extremity motions.

Figure 1
figure 1

Demonstration of experiment setup and the calculation of pelvis and lumbar rotation angles

Full size image

Independent and Dependent Variables

In the current experiment, a single independent variable: lumbar extensor muscle fatigue (with or without) was investigated. A fatiguing protocol was adopted to generate low back extensor muscle fatigue (detailed in the ‘Protocol’ section) and the median frequencies of the erector spinae EMG signals were used to quantify the accumulation of muscle fatigue. Four dependent variables were involved: (1) maximum lumbar rotation angle, lumbar rotation angle was defined as the angular difference between the T12 and S1 motion sensors in the sagittal plane (Fig. 1b); (2) maximum pelvis rotation angle; pelvis rotation angle was defined as the relative rotation of S1 sensor from the upright posture in the sagittal plane (Fig. 1b); (3). normalized lumbar–pelvic rotation ratios during both trunk flexion and extension motions. For each trail, lumbar and pelvis motions were normalized to their own range of motion (i.e., 0% motion at upright standing posture and 100% motion at trunk full flexion posture) respectively. Detailed description of this normalization process is included in the ‘Data processing’ section; 4. moment on the L5/S1 joint.

Protocol

Upon arrival, subjects first signed an informed consent form, then a 5-min warm-up session was provided to allow subjects to stretch and warm up muscles. Next, the skin over the sampling locations was shaved and swabbed using alcohol and cotton pad. The motion sensors and EMG electrodes were then placed using double-sided tape following previous instructions.14,33 When data collection started, subjects first conducted five repetitions of sagittally symmetric weight lifting and lowering tasks using a 20-lbs box placed 40 cm from the center of ankles (Fig. 2a). The number of repetitions was chosen to both provide enough statistical power and to avoid unwanted muscle fatigue before the fatiguing protocol. Subjects were required to use stoop lifting technique with both arms maintained straight and move at their natural pace without abrupt changes in motion. At least 1-min rests were provided between each two repetitions. When finished, subjects moved to the custom made structure to conduct the fatiguing protocol with pelvis fully secured (Fig. 2b). The fatiguing protocol requires subjects to maintain a ~45 degree trunk flexion posture with back kept straight while holding the same 20-lbs box until exhausted. After the fatiguing protocol, another five repetitions of box lifting and lowering tasks were conducted without rest in between in order to avoid significant recovery of lumbar muscle fatigue. Additionally, fatigue measurement tasks were also performed four times during the data collection: (1) at the beginning of the data collection; (2) immediately after the first five repetitions of box lifting and lowering tasks; (3) immediately after the fatigue protocol; (4) immediately after the second five repetitions of box lifting and lowering tasks. The fatigue measurement task requires subjects to hold the 20-lbs box and maintain a ~45 degree trunk flexion posture for 6 s, with pelvis secured by the custom made apparatus. The median frequencies of erector spinae EMG signals from the fatigue measurement tasks were used to verify (1) activities prior to the fatiguing protocol did not generate significant muscle fatigue; (2) significant muscle fatigue was generated during the fatiguing protocol and (3) Muscle fatigue remained until the end of the data collection.

Figure 2
figure 2

(a) A snapshot that demonstrates the weight lifting and lowering task and (b) Subject secured in the custom-made structure and perform the fatiguing protocol

Full size image

Data Processing

The pelvis and lumbar rotation data were normalized from 0 (i.e. upright posture) to 1 (i.e. full flexion posture) with respect to their range of motion. The lifting and lowering motions were both divided into 10 sub-segments based on the normalized pelvis rotational angle and the normalized lumbar rotational angles at 10th, 20th, 30th, 40th, 50th, 60th, 70th, 80th, and 90th percentile of normalized pelvis rotational angle were compared. The dynamic external moment that imposed on the L5/S1 joint was estimated using a previously developed physical model16,17,34 which considers trunk flexion angle, instantaneous linear and angular accelerations; the upper body mass and center of mass were estimated based on subjects’ anthropometric measurements (body height, weight, trunk length, trunk width, etc.,) as well as trunk kinematic data.30

A custom Matlab (Matlab 2010, MathWorks, USA) program was developed to process the EMG and kinematics data. The EMG signals were first transferred into frequency domain using Fast Fourier Transformation (FFT) and then passed through a notch filter at 60 Hz and its aliases, a 500 Hz low-pass filter and a 10 Hz high-pass filter. The median frequencies of the EMG data collected during the fatigue measurement tasks were calculated using method described in the previous literature.21

Statistical Analysis

The effects of lumbar extensor muscle fatigue on all dependent variables were tested using paired t test in order to reduce the between-subject variance. Each repetition was treated as a single observation and was used for the paired difference analysis with subject being the blocking variable. Minitab (Minitab 15, Minitab Inc., State College, PA, USA) was used for statistical analyzes with significance level of p < 0.05.

Results

Extensor Muscle Fatigue

The lumbar extensor muscle fatigue was characterized by the decrease of EMG median frequency.21 Data from the fatigue measurement tasks demonstrated that no statistically significant lumbar extensor muscle fatigue was introduced after conducting the five repetitions of weight lifting tasks (average median frequency changes from 62.77 (3.53) Hz to 60.29 (2.45) Hz, p value = 0.56). On the other hand, the fatiguing protocol successfully generated lumbar extensor muscle fatigue, which is demonstrated by the significant reduction of EMG median frequency after the fatiguing protocol (average median frequency changes from 60.29 (2.45) Hz to 52.70 (1.37) Hz, p value = 0.004). Finally, no clear recovery of muscle fatigue after the second five repetitions of weight lifting and lowering tasks was observed (average median frequency changes from 52.70 (1.37) Hz to 52.63 (1.49) Hz, p value = 1.00). Collectively, these results showed that the goal of our experiment design was successfully achieved.

Lumbar and Pelvis Maximum Rotation

Results of the statistical analysis showed that during the performance of weight lifting tasks neither the maximum lumbar rotation angle (p value = 0.973) nor the maximum pelvis rotation angle (p value = 0.620) was significantly affected by muscle fatigue (Fig. 3).

Figure 3
figure 3

The maximum lumbar and pelvis rotation angles in normal (N) and fatigued (F) conditions. Bars indicate the corresponding standard error

Full size image

Normalized Lumbar–Pelvis Ratio

Lumbar extensor muscle fatigue significantly influenced normalized lumbar–pelvic coordination during weight lifting and lowering tasks (Figs. 4 and 5 respectively). It can be observed from Fig. 4 that during lifting motions, at a given percentage of pelvis rotation, subjects had larger lumbar rotation (note that the initial posture is at 100% of lumbar flexion) with the influence of lumbar muscle fatigue (except at the 90th percentile of normalized pelvis rotation). Similarly, during the weight lowering motion larger lumbar rotations were observed (except at the 10th, 20th, and 30th percentile of normalized pelvis rotation) as shown in Fig. 5.

Figure 4
figure 4

The comparison of normalized lumbar–pelvic coordination during lifting motion between normal (N) and fatigue (F) conditions. Asterisk marks represent that the two conditions are statistically different from each other and bars indicate the corresponding standard error

Full size image
Figure 5
figure 5

The comparison of normalized lumbar–pelvic coordination during lowering motion between normal (N) and fatigue (F) conditions. Asterisk marks represent that the two conditions are statistically different from each other and bars indicate the corresponding standard error

Full size image

Spinal Loading

To demonstrate the effect of lumbar extensor muscle fatigue on spinal loading during weight lifting and lowering motions. The L5/S1 joint moments during both lifting and lowering motions were calculated from 10th to 100th percentiles of lifting/lowering duration. Results of our analyzes showed that with the influence of lumbar extensor muscle fatigue subjects experienced significantly larger L5/S1 joint moment during both weight lifting and lowering motions (Figs. 6 and 7).

Figure 6
figure 6

The comparison of L5/S1 external moment during weight lifting task between normal (N) and fatigue (F) conditions. Asterisk marks represent that the two conditions are statistically different from each other and bars indicate the corresponding standard error

Full size image
Figure 7
figure 7

The comparison of L5/S1 external moments during weight lowering task between normal (N) and fatigue (F) conditions. Asterisk marks represent that the two conditions are statistically different from each other and bars indicate the corresponding standard error

Full size image

Discussions

The purpose of the current study was to quantify the influence of lumbar extensor muscle fatigue on the changes of lumbar and pelvis motion synergy during weight lifting and lowering tasks. Results of the current study showed that after lumbar muscle became fatigued subjects adopted larger lumbar–pelvic motion ratio during both weighting lifting and lowering, such changes in kinematics also resulted in larger spinal loading. These results supported our initial hypotheses.

Early studies suggested that spinal motion is achieved by the sequential contribution of both pelvis and lumbar segments.5,8 More specifically, the beginning phase of trunk extension motion is initiated by pelvis moment while the ending phase (i.e. close to upright posture) is achieved mainly by lumbar spine extension motion. It was also believed that the trunk flexion motion is just the inverse process of the extension motion.8 On the other hand, more recent studies concluded that pelvis and lumbar motion occurred simultaneously during trunk flexion and extension motion.7,12,32,36,45 Results of the current study demonstrated that despite differ in magnitude, both lumbar and pelvis segments moved simultaneously during weight lifting and lowering tasks, which agreed with findings in the recent literature. In addition, the current results also showed that after fatigue, the lumbar and pelvis motions remain simultaneous even though the lumbar–pelvic ratio was significantly increased.

To further validate the results of the current study, we compared the lumbar–pelvic (L/P) ratios observed in the current study with values reported from previous studies. To be consistent with previous literature, trunk motions during weight lifting and lowering were equally (based on trunk angle) divided into four intervals (0–25, 25–50, 50–75 and 75–100% of maximum trunk angle), for a lowering motion, 0% represents the upright stand posture while for a lifting motion 0% represents the flexed trunk posture where subject initiates the lift. The L/P ratios were calculated in each interval by dividing the changes of lumbar rotation by the changes of pelvis rotation.7,28,45 Based on the results of the current study, without the influence of muscle fatigue, during weight lifting motion, the L/P ratio increased from 0.84 in first quarter (0–25%) to 1.88 in the fourth quarter (75–100%) of trunk range of motion (ROM) (Fig. 8a). On the other hand, during weight lowering motion the L/P ratio decreased from 1.55 in the first quarter (0–25%) to 0.47 in fourth quarter (75–100%) of trunk ROM (Fig. 8b), such changes in general agreed with data reported in previous studies.7,12,28,45

Figure 8
figure 8

The comparison of L/P ratios recorded in the current study with previous literature: a weight lifting motion; b weight lowering motion. During the lowering motion, 0% represents the upright stand posture while during the lifting motion 0% represents the flexed trunk posture that subject initiate the lift

Full size image

The results of the current study indicated that with the influence of lumbar extensor muscle fatigue subjects adopted higher L/P ratio during both weight lifting and weight lowering motions. Previous study reported that lumbar extensor muscle fatigue causes the cessation of EMG activities occur earlier on these muscles during trunk flexion and extension motions.6 It means that lumbar muscles are activated in a smaller range of trunk motion, in turn, lumbar passive tissues including ligaments, tendons and the elastic component of lumbar muscles need to compensate for the lost of active muscle force.34 In other words, the lumbar passive tissues were further elongated to exert larger elastic forces in order to counterbalance the external loading. Because the strain on lumbar posterior tissue is highly determined by the magnitude of lumbar flexion,29 therefore larger lumbar and pelvis rotation ratios were observed during trunk motions.

Such transition of loading, from lumbar active muscles to passive tissues would also influence the spinal loading. The structure of lumbar spine determines that lumbar posterior ligaments have shorter moment arm to the center of the L5/S1 joint compare to lumbar extensor muscles, therefore, to counterbalance the same amount of external loading, the increase of loading on lumbar passive tissues could potentially result in higher spinal compression and shear forces on intervertebral disks.1 Such effect combined with the observed increase in spinal external loading (Figs. 6 and 7) could further elevate spinal compressive and shear loading which may increase the risk of LBP.10,11,25,26

Several limitations of the current study need to be noted. First, in the current experiment design, a single level of hand load was used, the potential interaction effect between hand load and muscle fatigue on lumbar–pelvic motion ratio was not explored. Second, only stoop lifting was tested in the current study. Lumbar muscle fatigue may have different influences on other lifting techniques therefore warrants further investigation. Finally, with only male subjects included in the current study, the influence of gender was not tested.

Conclusion

Lumbar muscle fatigue alters lumbar–pelvic motion ratio during the performance of both weight lifting and lowering tasks. Such change elevates spinal moment, which may lead to higher spinal compressive and shear loadings and the associated risk of LBP. Results of this study further confirmed lumbar muscle fatigue as an injury risk factor during the performance of material handling especially lifting tasks. In a clinical setting, the current results indicate that subjects experiencing lumbar muscle fatigue may demonstrate transient changes of lumbar–pelvic motion rhythm that similar to that observed among LBP patients.7,28 Therefore, lumbar muscle fatigue must be avoided when using lumbar–pelvic motion rhythms for patient diagnosis or rehabilitation assessment.