- Email: [email protected]
Relative contribution of trunk muscles to the stability of the lumbar spine during isometric exertions
Clinical Biomechanics 17 (2002) 99–105 www.elsevier.com/locate/clinbiomech
Relative contribution of trunk muscles to the stability of the lumbar spine during isometric exertions Jacek Cholewicki *, James J. VanVliet IV Biomechanics Research Laboratory, Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, P.O. Box 208071, 333 Cedar Street, New Haven, CT 06520-8071, USA Received 20 September 2001; accepted 11 December 2001
Abstract Objective. To compare the relative contribution of various trunk muscles to the stability of the lumbar spine. Design. Quantification of spine stability with a biomechanical model. Background. Modern low back rehabilitation techniques focus on muscles that stabilize the lumbar spine. However, the relative contribution of various trunk muscles to spine stability is currently unknown. Methods. Eight male subjects performed isometric exertions in trunk flexion, extension, lateral bending, and axial rotation, and isometric exertions under vertical trunk loading and in a lifting hold. Each isometric trial was repeated three times at 20%, 40%, and 60% of the maximum trunk flexion force or with a load of 0%, 20%, 40%, and 60% of body weight for the latter two exertions. Surface EMG data from 12 major trunk muscles were used in the biomechanical model to estimate stability of the lumbar spine. A simulation of each trial was performed repeatedly with one of the 10 major trunk muscle groups removed from the model. Results. Relative contribution of each muscle to spine stability was significantly affected by the combination of loading magnitude and direction (3-way interaction). None of the removed muscles reduced spine stability by more than 30%. Conclusions. A single muscle cannot be identified as the most important for the stability of the lumbar spine. Rather, spine stability depends on the relative activation of all trunk muscles and other loading variables. Relevance This study will improve our understanding of individual trunk muscles’ contribution to overall stability of the lumbar spine. Ó 2002 Elsevier Science Ltd. All rights reserved. Keywords: Spine stability; Muscles; Lumbar spine
1. Introduction The importance of trunk muscles in providing stability to the lumbar spine is well established [1–6]. Under dynamic loading conditions, trunk muscles must be recruited in appropriate sequence and with appropriate strength of contraction to support loads and to maintain stability. Deficiency in timely muscle activation in response to sudden trunk loading has been documented among patients with low back pain (LBP) [7–10]. Such a deficiency or errors in motor control could lead to the loss of spine stability causing recurrent injuries to the lumbar spine [4,11].
*
Corresponding author. E-mail address: [email protected] (J. Cholewicki).
The above theories lead to the development of rehabilitation strategies that focus more on enhancing spine stability rather than on improving muscle strength and range of motion. Such treatments are designed to improve function of the muscles that are believed to govern spine stability and protect the spine from worsening trauma [12,13]. Unfortunately, there is no clear scientific evidence to suggest which of the trunk muscles are the most important or effective in their spine stabilizing function and which muscles such a therapy should target. Previous research identified several trunk muscles or muscle groups that may be important for spine stability, but these conclusions were based on a variety of inconsistent criteria. For example, deep inter-segmental trunk muscles (muscles with multiple attachments to the lumbar spine) were considered by many as ‘‘stabilizers’’ after Bergmark demonstrated that for a certain
0268-0033/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 8 - 0 0 3 3 ( 0 1 ) 0 0 1 1 8 - 8
100
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
activation level of inter-segmental muscles there existed a maximum level to which multi-segmental muscles (muscles that attach to the thorax and pelvis only) could be activated and beyond which spine buckling would occur [1]. In contrast, Crisco and Panjabi [2] concluded that large multi-segmental muscles were more effective in stabilizing the spine, although spine stability could not be achieved if any spine level was void of an intersegmental muscle fascicle attachment. Santaguida and McGill [14] suggested that psoas might have spine-stabilizing potential with its compressive force and otherwise minimal role in lumbar moment production. Because of its architectural features and increasing activation in response to increasing compressive spine loading, quadratus lumborum was identified as being an even better stabilizer of the spine than psoas [15]. Finally, transversus abdominus has recently been considered as one of the most important muscles controlling the stability of the lumbar spine [16] based on the work of Hodges and colleagues. These authors found that transversus abdominus is the first muscle to be activated in response to sudden trunk loading in healthy subjects and it lags behind other abdominal muscles in patients with LBP [8,9,17]. In addition, this muscle is consistently involved in generating intra-abdominal pressure [18], which has the potential to stabilize the spine [19]. Given the above data, spinal segmental stabilization therapy for low back pain focused on deep inter-segmental muscles such as the multifidus, internal oblique, and transversus abdominus with the understanding that these muscles are the major contributors to spine stability [12,13]. However, no systematic and consistent comparison of the stabilizing potential of various trunk muscles exists to date. Therefore, the purpose of this study was to compare the relative contribution of various trunk muscles to lumbar spine stability during a variety of isometric trunk exertions. We hypothesized that there is no one muscle or muscle group prevailing over other muscles in their spine stabilizing function under all loading conditions. On the contrary, the relative contribution of a muscle to spine stability will depend on the magnitude and direction of trunk loading.
2. Methods Eight male subjects (age 20, SD 2 years; height 184.0, SD 5.1 cm; weight 82.0, SD 10.3 kg) with no history of back pain volunteered to participate in this study by signing an informed consent form approved by an institutional Human Investigation Committee. Prior to testing, the measurements of the subjects’ height, weight, T9-L4/L5 length, shoulder to L4/L5 length, and shoulder width were taken. These measurements were used to calculate L4/L5 joint moments. Surface, disposable, bipolar, Ag–AgCl EMG electrodes (Graphic Controls,
Buffalo, NY, USA) were applied with a 3 cm center-tocenter spacing over the 6 major trunk muscles bilaterally: rectus abdominus (3 cm lateral to the umbilicus), external oblique (approximately 15 cm lateral to the umbilicus), internal oblique (approximately midway between the anterior superior iliac spine and symphysis pubis, above the inguinal ligament), latissimus dorsi (lateral to T9 over the muscle belly), thoracic erector spinae (5 cm lateral to T9 spinous process), and lumbar erector spinae (3 cm lateral to L3 spinous process). In our previous studies, this electrode placement proved to maximize signal-to-noise ratio with insignificant levels of cross talk [4]. For EMG normalization, the subjects elicited maximum voluntary contractions (MVC) of latissimus dorsi and other muscles in attempted sit-up, trunk extension, and lateral bending exercises. The baseline EMG activity was recorded with subjects lying completely relaxed in supine position. Following the MVC and EMG baseline trials, the subjects executed three repetitions at each load level of the six isometric exercises. Four exercises were the attempted exertions in trunk flexion, extension, left lateral bending, and clockwise twisting with the trunk held in a natural upright posture. The subjects were positioned in a semi-seated position in an apparatus that prevented pelvis and lower body motion. A cable, attached to a chest harness at approximately the T9 level, provided the resistance for the isometric trials. A force couple, created by two cables attached to the subjects’ shoulders, resisted attempted axial rotation trials. A force exerted by the subjects was measured with a transducer and was displayed on an oscilloscope along with target force levels. The target force levels for all four exertion directions were set at 20%, 40%, and 60% of the maximum isometric effort in trunk flexion. The other two exercises were a vertical loading on the trunk and a lifting hold. The vertical loading on the trunk was performed in the same apparatus as described above. The loading was accomplished with bagged lead shot evenly distributed among four pouches attached to the chest harness. The center of mass of the load was at approximately the T9 level. The isometric lifting hold was performed standing with feet placed shoulder width apart. Subjects were coached to flex strictly at the hips at 45° while holding a plastic crate filled with bagged lead shot. The trunk angle was measured with an inclinometer. The load mass for these two exercises was set at 0%, 20%, 40%, and 60% of subjects’ body mass. For every trial, 3 s of data were collected at a sampling rate of 1600 Hz. The EMG signals were band-pass limited between 20 and 500 Hz and differentially amplified (input impedance ¼ 100 GX; CMRR > 140 dB) prior to the A/D conversion. The EMG data were digitally rectified and averaged. The baseline EMG values were subtracted and the data were normalized to the maximum EMG activity attained during the MVC trials.
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
The normalized EMG data served as input to the biomechanical model developed for quantifying stability of the lumbar spine. The model was previously presented in detail [4] and therefore only a brief description will follow. The model consisted of a rigid pelvis and sacrum, 5 lumbar vertebrae separated by a lumped parameter, nonlinear disc and ligament equivalent for rotational joint stiffness about the 3 axes, rigid ribcage and 90 muscle fascicles. Three axes of rotation were assigned to each inter-vertebral joint between T12 and S1, for a total of 18 d.o.f. (6 joints 3 d.o.f. each). The moments and forces (after accounting for passive tissue contribution) necessary to balance the external load and upper body weight were partitioned between all 90 muscle fascicles with the assistance of EMG. For that purpose, the cross-bridge bond distribution moment (DM) model for obtaining muscle force and stiffness simultaneously [20] and the EMG Assisted Optimization approach to balance the moment equations [21,22] were used. Stability analysis was performed in accordance with the minimum potential energy principle. Average curvature of the surface of the system’s potential energy in a vicinity of the static equilibrium served as the relative stability index (SI). After quantifying stability of the lumbar spine for each trial, calculations of stability were repeated with the activation of each specific muscle group zeroed, thus effectively removed from the model. The percent decrease in SI due to removing a muscle group indicated the contribution of that muscle group to the overall stability of the lumbar spine. The following muscles were considered: rectus abdominus, external oblique, internal oblique, latissimus dorsi, iliocostalis lumborum, longissimus thoracis, lumbar erector spinae, multifidus, psoas, and quadratus lumborum. The relative contribution of these muscles to lumbar spine stability was compared using a three-factor, repeated measures A N O V A . Direction of isometric exertion, load level, and muscle removed constituted the three factors.
3. Results The isometric exercises and the load levels selected for this study resulted in a range of lumbar spine moments. Vertical loading trials demanded the smallest and the lifting trials the largest absolute moments about the L4– L5 inter-vertebral joint (Table 1). The remaining isometric exertions were designed to produce approximately similar absolute lumbar spine moments about the three joint axes. The resultant L4–L5 joint compression force was the highest during lifting trials, followed by trials in axial rotation, flexion, extension, lateral bending, and vertical loading (Table 1). However, stability estimated for the intact lumbar spine showed a different pattern. The SI was the highest during lifting
101
trials, followed by extension, axial rotation, lateral bending, flexion, and vertical loading trials (Table 1). The spine was the most vulnerable to instability during flexion trials. The biomechanical model predicted the highest number of buckling occurrences during the isometric flexion exertions and no buckling during the isometric trunk extension (Table 2). The above instability pattern held true for the simulations of an intact spine as well as a spine structurally compromised by the removal of one of the major trunk muscles (Table 2). The removal of longissimus thoracis resulted in the largest number of trials in which the spine buckled, followed by latissimus dorsi, iliocostalis lumborum, internal oblique and multifidus. However, the number of buckling occurrences represented only 4% of all different trials and scenarios simulated. Therefore, the comparison of SI values obtained in the remaining stable trials was necessary. Results of A N O V A returned a significant, second level interaction between the effects of exertion direction (direction), load level (load), and the muscle removed (muscle) on SI (Table 3). In other words, the contribution of different trunk muscle groups to lumbar spine stability depended on the combination of direction and load level of isometric exertions (Fig. 1(a)–(f)). The first general observation was that no single muscle group contributed more than 30% to the overall stability of the lumbar spine. The removal of lumbar erector spinae muscles from the simulation of lifting trials caused the largest reduction in SI, which dropped to approximately 70% of its intact value (Fig. 1(f)). The lumbar erector spine group also had the largest contribution to spine stability during extension, lateral bending and axial rotation exertions, but not during flexion or vertical loading trials. Generally, when muscles acted as agonists, they had a larger effect on SI than when they acted as antagonists (Fig. 1(a) and (b)). However, there were some exceptions. For example, the longissimus thoracis, a powerful trunk extensor muscle, had a greater detrimental effect on spine stability during flexion trials (16% reduction) than during extension trials (14% reduction) (Fig. 1(a) and (b)). No clear distinction was evident between inter-segmental and multi-segmental muscles in their contribution to spine stability. For example, lumbar erector spine, predominantly an inter-segmental muscle and longissimus thoracis, predominantly a multi-segmental muscle, were among the ones that had the greatest effect on SI. Conversely, psoas and rectus abdominus, interand multi-segmental muscles, both had equally little effect on spine stability (Fig. 1(a)–(f)). Similarly, no generalization could be made about muscle size. While the large trunk muscles such as longissimus thoracis or lumbar erector spine had the greatest effect on SI, rectus abdominus, an equally large muscle, had a negligible
102
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
Table 1 Spine moments generated during the isometric exercises and the resultant L4–L5 joint compression forces and the lumbar spine stability quantified with the stability index Load level
0
1
2
3
Moment about L4–L5 (Nm) Extension Flexion Lateral bending Axial rotation Vertical loading Lifting
1 (0) 1 (0) 1 (0) 1 (0) 1 (0) 75 (9)
29 (5) 32 (5) 30 (5) 27 (2) 1 (0) 99 (12)
55 (11) 57 (11) 56 (11) 50 (5) 1 (0) 119 (10)
82 (18) 84 (17) 82 (18) 71 (9) 1 (0) 142 (12)
L4–L5 Compression force (N) Extension Flexion Lateral bending Axial rotation Vertical loading Lifting
750 (374) 750 (374) 750 (374) 750 (374) 750 (374) 1478 (158)
1104 (87) 983 (241) 918 (113) 1204 (220) 790 (144) 1938 (177)
1468 (121) 1501 (431) 1364 (141) 1801 (235) 981 (295) 2276 (194)
1957 2072 1891 2205 1139 2626
Stability index (Nm/rad) Extension Flexion Lateral bending Axial rotation Vertical loading Lifting
172 172 172 172 172 422
323 161 207 279 162 521
416 243 315 413 191 604
549 309 435 485 222 690
(107) (107) (107) (107) (107) (63)
(38) (71) (46) (40) (43) (48)
(68) (75) (54) (64) (91) (58)
(210) (440) (272) (420) (345) (143)
(62) (83) (63) (100) (110) (49)
Means and SD in parenthesis.
Table 2 Frequency of buckling (instability) predicted by the model in an intact spine and with each individual muscle group removed. These trials represent 4% of the 5280 simulation trials performed Muscle group removed
Extension
Intact Rectus Abdominus External oblique Internal oblique Latissimus dorsi Iliocostalis lumborum Longissimus thoracis Lumbar erector spinae Multifidus Psoas Quadratus lumborum
0 0 0 0 0 0 0 0 0 0 0
Total
0
Flexion
Lateral bending
Axial rotation
Vertical loading
11 10 4 17 12 15 35 11 11 11 11
0 0 0 0 0 0 3 0 0 0 1
0 0 0 0 0 0 0 0 2 0 0
0 1 3 0 4 8 14 1 0 0 3
0 1 0 5 8 0 0 0 8 0 2
11 12 7 22 24 23 52 12 21 11 17
148
4
2
34
24
212
Table 3 Results of three-factor analysis of variance Source
d.o.f.
ðF Þ
ðP Þ
Exertion direction Load level Muscle removed Direction load Direction muscle Load muscle Direction load muscle
5 3 9 15 45 27 135
20.4 45.5 652.1 1.5 49.3 10.5 4.0
0.000 0.000 0.000 0.119 0.000 0.000 0.000
Degrees of freedom (d.o.f.), F -statistics ðF Þ, and significance ðP Þ.
Lifting
Total
effect on spine stability in all exertion directions (Fig. 1(a)–(f)). The contribution of individual muscles to spine stability with increasing load levels was either decreasing, increasing, or stayed approximately the same depending on the muscle and the direction of exertion (see, for example, internal oblique, iliocostalis lumborum, and psoas in flexion exertion, Fig. 1(b)). Although statistically significant (Table 3), the largest differences between different loading conditions did not exceed a 10% change in SI.
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
103
Fig. 1. Relative stability of the lumbar spine (percent of the intact stability index value) after the removal of each muscle group individually from the model during simulation of isometric trunk exertions in (a) extension, (b) flexion, (c) lateral bending to the left, (d) clockwise axial rotation, (e) vertical loading, and (f) lifting hold.
4. Discussion In this study, we compared the relative contribution of 10 major trunk muscle groups to the stability of the lumbar spine by removing each muscle systematically from the biomechanical model developed for quantifying spine stability. We purposely did not normalize each
trunk muscle to its size, moment arm, or activation because we wanted to report the contribution of each muscle as they function in vivo under various loading situations. The results support our hypothesis that no one muscle can be identified as the most important for the stability of the lumbar spine. Instead, the relative contribution of each muscle to spine stability depended
104
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
on trunk loading direction and magnitude. These results are not surprising, because the overall structural stability of the spine is a highly nonlinear function of many variables. In addition to the activation levels of all trunk muscles, spine stability is determined by ligamentous stiffness of the inter-vertebral joints, spine posture, and the magnitude, direction and end-conditions of external trunk loads [4]. Moreover, even if a given muscle fascicle does not cross all of the intervertebral joints, its activation level (stiffness) affects the stability at all other remaining joints [4]. Therefore, the relative contribution of a given muscle to spine stability will change not only with different loading conditions but also with different recruitment patterns of the other trunk muscles. Given the results of the present study, we suggest that the often-used classification of muscles into ‘‘local’’ (deep, inter-segmental) and ‘‘global’’ (superficial, multisegmental) systems [1], as the way to discriminate between the muscles responsible for inter-segmental stability and spine motion, is incorrect. All trunk muscles contribute to spine stability and their contribution depends on many state variables. Stability or instability should be thought of as a state of the entire spine system determined by the activation of all trunk muscles, passive joint properties, spine posture, and loading conditions. Perhaps low back rehabilitation should be approached from a global system point of view involving the entire lumbar spine musculature with its motor control to enhance spine stability under various loading conditions. There are several limitations related to the biomechanical model used in this study. We did not include transversus abdominus in the model and the internal oblique muscle was represented by only two fascicles. Transversus abdominus does not have the potential to produce moments around the lumbar spine directly [17,23]. The mechanical effect of this muscle on the spine is most likely mediated through the intra-abdominal pressure mechanism, which is not yet very well understood [17]. Given the size of the transversus, even if all of its mechanical effects were transmitted to the spine, we would not expect its relative contribution to spine stability to be disproportionably larger than other trunk muscles. However, we do acknowledge that this muscle may play an important spine stabilizing function in preparation to sudden movements, because it is the first muscle to be recruited in healthy subjects under such conditions [8,9,18]. The internal oblique muscle is believed to have a great potential for stabilizing the spine laterally, because of its attachments to each lumbar vertebrae via lumbodorsal fascia [24]. However, we do not suspect that this architecture would greatly increase the contribution of the internal oblique muscle to spine stability estimated in our study. First, the possibility of moment transmis-
sion through lumbodorsial fascia has been questioned in the past [25]. Second, the analysis of stability cannot look at just one direction, but must consider all of the d.o.f. simultaneously. Finally, our results did not indicate an overwhelmingly greater spine-stabilizing advantage of the inter-segmental muscles over the multisegmental muscles. The stiffening effect of the inter-vertebral joints due to the compressive forces arising from muscle contractions [26] was not considered in our biomechanical model. However, this would not greatly affect our results because we compared the relative contribution of muscles to spine stability. Specifically, each muscle is expected to add a spine compression force that is proportional to its size and activation level. Subsequently, a similar proportion will be contributed to the passive joint stiffness by each muscle. Because muscle stiffness is also proportional to its size and activation level [1,3,20], these factors are already taken into account when the relative contribution of each muscle to spine stability was computed. Therefore, the net effect of not accounting for increased inter-vertebral joint stiffness due to compression force was the underestimated overall stability of the lumbar spine, but not the percentage of each muscle’s contribution to spine stability.
5. Conclusions No one trunk muscle could be identified as contributing the most to spine stability under all different loading conditions. The relative contribution of a given muscle to spine stability depended significantly on loading magnitude and direction. Furthermore, dysfunction (removal) of any single trunk muscle did not reduce the overall spine stability by more than 30%. Therefore, rehabilitation exercises for enhancing spine stability should involve the entire spinal musculature and its motor control under various spine-loading conditions.
Acknowledgements This study was supported in part by the Whitaker Foundation Biomedical Engineering Research Grant and the Gaylord Rehabilitation Research Institute Grant.
References [1] Bergmark A. Stability of the lumbar spine. A study in mechanical engineering. Acta Orthop Scand Suppl 1989;230:1–54.
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105 [2] Crisco JJ 3rd, Panjabi MM. The intersegmental and multisegmental muscles of the lumbar spine. A biomechanical model comparing lateral stabilizing potential. Spine 1991;16:793–9. [3] Gardner-Morse M, Stokes IA, Laible JP. Role of muscles in lumbar spine stability in maximum extension efforts. J Orthop Res 1995;13:802–8. [4] Cholewicki J, McGill SM. Mechanical stability of the in vivo lumbar spine: implications for injury and chronic low back pain. Clin Biomech 1996;11:1–15. [5] Cholewicki J, Panjabi MM, Khachatryan A. Stabilizing function of trunk flexor-extensor muscles around a neutral spine posture. Spine 1997;22:2207–12. [6] Granata KP, Marras WS. Cost-benefit of muscle cocontraction in protecting against spinal instability. Spine 2000;25:1398–404. [7] Magnusson ML, Aleksiev A, Wilder DG, Pope MH, Spratt K, Lee SH, et al. Unexpected load and asymmetric posture as etiologic factors in low back pain. Eur Spine J 1996;5:23–35. [8] Hodges PW, Richardson CA. Delayed postural contraction of transversus abdominis in low back pain associated with movement of the lower limb. J Spinal Disor 1998;11:46–56. [9] Hodges PW, Richardson CA. Altered trunk muscle recruitment in people with low back pain with upper limb movement at different speeds. Arch Phys Med Rehabil 1999;80:1005–12. [10] Radebold A, Cholewicki J, Panjabi MM, Patel TC. Muscle response pattern to sudden trunk loading in healthy individuals and in patients with chronic low back pain. Spine 2000;25:947–54. [11] Panjabi MM. The stabilizing system of the spine. Part I. Function dysfunction, adaptation, and enhancement. J Spinal Disord 1992;5:383–9. [12] Richardson C, Jull G, Hodges P, Hides J. Therapeutic exercise for spinal segmental stabilization in low back pain. Scientific basis and clinical approach. London: Churchill Livingstone; 1999. [13] O’Sullivan PB. Lumbar segmental ‘instability’: clinical presentation and specific stabilizing exercise management. Man Ther 2000;5:2–12. [14] Santaguida PL, McGill SM. The psoas major muscle: a threedimensional geometric study. J Biomech 1995;28:339–45.
105
[15] McGill S, Juker D, Kropf P. Quantitative intramuscular myoelectric activity of quadratus lumborum during a wide variety of tasks. Clin Biomech 1996;11:170–2. [16] Pietrek M, Sheikhzadeh A, Nordin M, Hagins M. Biomechanical modeling of intra-abdominal pressure generation should include the transversus abdominis. J Biomech 2000;33:787–90. [17] Hodges PW. Is there a role for transversus abdominis in lumbopelvic stability? Man Ther 1999;4:74–86. [18] Cresswell AG, Oddsson L, Thorstensson A. The influence of sudden perturbations on trunk muscle activity and intra-abdominal pressure while standing. Exp Brain Res 1994;98:336–41. [19] Cholewicki J, Juluru K, Radebold A, Panjabi MM, McGill SM. Lumbar spine stability can be augmented with an abdominal belt and/or increased intra-abdominal pressure. Eur Spine J 1999;8:388–95. [20] Cholewicki J, McGill SM. Relationship between muscle force and stiffness in the whole mammalian muscle: a simulation study. J Biomech Eng 1995;117:339–42. [21] 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;28:321–31. [22] Cholewicki J, McGill SM. EMG assisted optimization: a hybrid approach for estimating muscle forces in an indeterminate biomechanical model. J Biomech 1994;27:1287–9. [23] McGill SM. A revised anatomical model of the abdominal musculature for torso flexion efforts. J Biomech 1996;29:973– 7. [24] Macintosh JE, Bogduk N, Gracovetsky S. Biomechanics of the thoracolumbar fascia. Clin Biomech 1987;2:78–83. [25] McGill SM, Norman RW. Potential of lumbodorsal fascia forces to generate back extension moments during squat lifts [see comments]. J Biomed Eng 1988;10:312–8. [26] Janevic J, Ashton-Miller JA, Schultz AB. Large compressive preloads decrease lumbar motion segment flexibility. J Orthop Res 1991;9:228–36.
Relative contribution of trunk muscles to the stability of the lumbar spine during isometric exertions Jacek Cholewicki *, James J. VanVliet IV Biomechanics Research Laboratory, Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, P.O. Box 208071, 333 Cedar Street, New Haven, CT 06520-8071, USA Received 20 September 2001; accepted 11 December 2001
Abstract Objective. To compare the relative contribution of various trunk muscles to the stability of the lumbar spine. Design. Quantification of spine stability with a biomechanical model. Background. Modern low back rehabilitation techniques focus on muscles that stabilize the lumbar spine. However, the relative contribution of various trunk muscles to spine stability is currently unknown. Methods. Eight male subjects performed isometric exertions in trunk flexion, extension, lateral bending, and axial rotation, and isometric exertions under vertical trunk loading and in a lifting hold. Each isometric trial was repeated three times at 20%, 40%, and 60% of the maximum trunk flexion force or with a load of 0%, 20%, 40%, and 60% of body weight for the latter two exertions. Surface EMG data from 12 major trunk muscles were used in the biomechanical model to estimate stability of the lumbar spine. A simulation of each trial was performed repeatedly with one of the 10 major trunk muscle groups removed from the model. Results. Relative contribution of each muscle to spine stability was significantly affected by the combination of loading magnitude and direction (3-way interaction). None of the removed muscles reduced spine stability by more than 30%. Conclusions. A single muscle cannot be identified as the most important for the stability of the lumbar spine. Rather, spine stability depends on the relative activation of all trunk muscles and other loading variables. Relevance This study will improve our understanding of individual trunk muscles’ contribution to overall stability of the lumbar spine. Ó 2002 Elsevier Science Ltd. All rights reserved. Keywords: Spine stability; Muscles; Lumbar spine
1. Introduction The importance of trunk muscles in providing stability to the lumbar spine is well established [1–6]. Under dynamic loading conditions, trunk muscles must be recruited in appropriate sequence and with appropriate strength of contraction to support loads and to maintain stability. Deficiency in timely muscle activation in response to sudden trunk loading has been documented among patients with low back pain (LBP) [7–10]. Such a deficiency or errors in motor control could lead to the loss of spine stability causing recurrent injuries to the lumbar spine [4,11].
*
Corresponding author. E-mail address: [email protected] (J. Cholewicki).
The above theories lead to the development of rehabilitation strategies that focus more on enhancing spine stability rather than on improving muscle strength and range of motion. Such treatments are designed to improve function of the muscles that are believed to govern spine stability and protect the spine from worsening trauma [12,13]. Unfortunately, there is no clear scientific evidence to suggest which of the trunk muscles are the most important or effective in their spine stabilizing function and which muscles such a therapy should target. Previous research identified several trunk muscles or muscle groups that may be important for spine stability, but these conclusions were based on a variety of inconsistent criteria. For example, deep inter-segmental trunk muscles (muscles with multiple attachments to the lumbar spine) were considered by many as ‘‘stabilizers’’ after Bergmark demonstrated that for a certain
0268-0033/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 8 - 0 0 3 3 ( 0 1 ) 0 0 1 1 8 - 8
100
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
activation level of inter-segmental muscles there existed a maximum level to which multi-segmental muscles (muscles that attach to the thorax and pelvis only) could be activated and beyond which spine buckling would occur [1]. In contrast, Crisco and Panjabi [2] concluded that large multi-segmental muscles were more effective in stabilizing the spine, although spine stability could not be achieved if any spine level was void of an intersegmental muscle fascicle attachment. Santaguida and McGill [14] suggested that psoas might have spine-stabilizing potential with its compressive force and otherwise minimal role in lumbar moment production. Because of its architectural features and increasing activation in response to increasing compressive spine loading, quadratus lumborum was identified as being an even better stabilizer of the spine than psoas [15]. Finally, transversus abdominus has recently been considered as one of the most important muscles controlling the stability of the lumbar spine [16] based on the work of Hodges and colleagues. These authors found that transversus abdominus is the first muscle to be activated in response to sudden trunk loading in healthy subjects and it lags behind other abdominal muscles in patients with LBP [8,9,17]. In addition, this muscle is consistently involved in generating intra-abdominal pressure [18], which has the potential to stabilize the spine [19]. Given the above data, spinal segmental stabilization therapy for low back pain focused on deep inter-segmental muscles such as the multifidus, internal oblique, and transversus abdominus with the understanding that these muscles are the major contributors to spine stability [12,13]. However, no systematic and consistent comparison of the stabilizing potential of various trunk muscles exists to date. Therefore, the purpose of this study was to compare the relative contribution of various trunk muscles to lumbar spine stability during a variety of isometric trunk exertions. We hypothesized that there is no one muscle or muscle group prevailing over other muscles in their spine stabilizing function under all loading conditions. On the contrary, the relative contribution of a muscle to spine stability will depend on the magnitude and direction of trunk loading.
2. Methods Eight male subjects (age 20, SD 2 years; height 184.0, SD 5.1 cm; weight 82.0, SD 10.3 kg) with no history of back pain volunteered to participate in this study by signing an informed consent form approved by an institutional Human Investigation Committee. Prior to testing, the measurements of the subjects’ height, weight, T9-L4/L5 length, shoulder to L4/L5 length, and shoulder width were taken. These measurements were used to calculate L4/L5 joint moments. Surface, disposable, bipolar, Ag–AgCl EMG electrodes (Graphic Controls,
Buffalo, NY, USA) were applied with a 3 cm center-tocenter spacing over the 6 major trunk muscles bilaterally: rectus abdominus (3 cm lateral to the umbilicus), external oblique (approximately 15 cm lateral to the umbilicus), internal oblique (approximately midway between the anterior superior iliac spine and symphysis pubis, above the inguinal ligament), latissimus dorsi (lateral to T9 over the muscle belly), thoracic erector spinae (5 cm lateral to T9 spinous process), and lumbar erector spinae (3 cm lateral to L3 spinous process). In our previous studies, this electrode placement proved to maximize signal-to-noise ratio with insignificant levels of cross talk [4]. For EMG normalization, the subjects elicited maximum voluntary contractions (MVC) of latissimus dorsi and other muscles in attempted sit-up, trunk extension, and lateral bending exercises. The baseline EMG activity was recorded with subjects lying completely relaxed in supine position. Following the MVC and EMG baseline trials, the subjects executed three repetitions at each load level of the six isometric exercises. Four exercises were the attempted exertions in trunk flexion, extension, left lateral bending, and clockwise twisting with the trunk held in a natural upright posture. The subjects were positioned in a semi-seated position in an apparatus that prevented pelvis and lower body motion. A cable, attached to a chest harness at approximately the T9 level, provided the resistance for the isometric trials. A force couple, created by two cables attached to the subjects’ shoulders, resisted attempted axial rotation trials. A force exerted by the subjects was measured with a transducer and was displayed on an oscilloscope along with target force levels. The target force levels for all four exertion directions were set at 20%, 40%, and 60% of the maximum isometric effort in trunk flexion. The other two exercises were a vertical loading on the trunk and a lifting hold. The vertical loading on the trunk was performed in the same apparatus as described above. The loading was accomplished with bagged lead shot evenly distributed among four pouches attached to the chest harness. The center of mass of the load was at approximately the T9 level. The isometric lifting hold was performed standing with feet placed shoulder width apart. Subjects were coached to flex strictly at the hips at 45° while holding a plastic crate filled with bagged lead shot. The trunk angle was measured with an inclinometer. The load mass for these two exercises was set at 0%, 20%, 40%, and 60% of subjects’ body mass. For every trial, 3 s of data were collected at a sampling rate of 1600 Hz. The EMG signals were band-pass limited between 20 and 500 Hz and differentially amplified (input impedance ¼ 100 GX; CMRR > 140 dB) prior to the A/D conversion. The EMG data were digitally rectified and averaged. The baseline EMG values were subtracted and the data were normalized to the maximum EMG activity attained during the MVC trials.
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
The normalized EMG data served as input to the biomechanical model developed for quantifying stability of the lumbar spine. The model was previously presented in detail [4] and therefore only a brief description will follow. The model consisted of a rigid pelvis and sacrum, 5 lumbar vertebrae separated by a lumped parameter, nonlinear disc and ligament equivalent for rotational joint stiffness about the 3 axes, rigid ribcage and 90 muscle fascicles. Three axes of rotation were assigned to each inter-vertebral joint between T12 and S1, for a total of 18 d.o.f. (6 joints 3 d.o.f. each). The moments and forces (after accounting for passive tissue contribution) necessary to balance the external load and upper body weight were partitioned between all 90 muscle fascicles with the assistance of EMG. For that purpose, the cross-bridge bond distribution moment (DM) model for obtaining muscle force and stiffness simultaneously [20] and the EMG Assisted Optimization approach to balance the moment equations [21,22] were used. Stability analysis was performed in accordance with the minimum potential energy principle. Average curvature of the surface of the system’s potential energy in a vicinity of the static equilibrium served as the relative stability index (SI). After quantifying stability of the lumbar spine for each trial, calculations of stability were repeated with the activation of each specific muscle group zeroed, thus effectively removed from the model. The percent decrease in SI due to removing a muscle group indicated the contribution of that muscle group to the overall stability of the lumbar spine. The following muscles were considered: rectus abdominus, external oblique, internal oblique, latissimus dorsi, iliocostalis lumborum, longissimus thoracis, lumbar erector spinae, multifidus, psoas, and quadratus lumborum. The relative contribution of these muscles to lumbar spine stability was compared using a three-factor, repeated measures A N O V A . Direction of isometric exertion, load level, and muscle removed constituted the three factors.
3. Results The isometric exercises and the load levels selected for this study resulted in a range of lumbar spine moments. Vertical loading trials demanded the smallest and the lifting trials the largest absolute moments about the L4– L5 inter-vertebral joint (Table 1). The remaining isometric exertions were designed to produce approximately similar absolute lumbar spine moments about the three joint axes. The resultant L4–L5 joint compression force was the highest during lifting trials, followed by trials in axial rotation, flexion, extension, lateral bending, and vertical loading (Table 1). However, stability estimated for the intact lumbar spine showed a different pattern. The SI was the highest during lifting
101
trials, followed by extension, axial rotation, lateral bending, flexion, and vertical loading trials (Table 1). The spine was the most vulnerable to instability during flexion trials. The biomechanical model predicted the highest number of buckling occurrences during the isometric flexion exertions and no buckling during the isometric trunk extension (Table 2). The above instability pattern held true for the simulations of an intact spine as well as a spine structurally compromised by the removal of one of the major trunk muscles (Table 2). The removal of longissimus thoracis resulted in the largest number of trials in which the spine buckled, followed by latissimus dorsi, iliocostalis lumborum, internal oblique and multifidus. However, the number of buckling occurrences represented only 4% of all different trials and scenarios simulated. Therefore, the comparison of SI values obtained in the remaining stable trials was necessary. Results of A N O V A returned a significant, second level interaction between the effects of exertion direction (direction), load level (load), and the muscle removed (muscle) on SI (Table 3). In other words, the contribution of different trunk muscle groups to lumbar spine stability depended on the combination of direction and load level of isometric exertions (Fig. 1(a)–(f)). The first general observation was that no single muscle group contributed more than 30% to the overall stability of the lumbar spine. The removal of lumbar erector spinae muscles from the simulation of lifting trials caused the largest reduction in SI, which dropped to approximately 70% of its intact value (Fig. 1(f)). The lumbar erector spine group also had the largest contribution to spine stability during extension, lateral bending and axial rotation exertions, but not during flexion or vertical loading trials. Generally, when muscles acted as agonists, they had a larger effect on SI than when they acted as antagonists (Fig. 1(a) and (b)). However, there were some exceptions. For example, the longissimus thoracis, a powerful trunk extensor muscle, had a greater detrimental effect on spine stability during flexion trials (16% reduction) than during extension trials (14% reduction) (Fig. 1(a) and (b)). No clear distinction was evident between inter-segmental and multi-segmental muscles in their contribution to spine stability. For example, lumbar erector spine, predominantly an inter-segmental muscle and longissimus thoracis, predominantly a multi-segmental muscle, were among the ones that had the greatest effect on SI. Conversely, psoas and rectus abdominus, interand multi-segmental muscles, both had equally little effect on spine stability (Fig. 1(a)–(f)). Similarly, no generalization could be made about muscle size. While the large trunk muscles such as longissimus thoracis or lumbar erector spine had the greatest effect on SI, rectus abdominus, an equally large muscle, had a negligible
102
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
Table 1 Spine moments generated during the isometric exercises and the resultant L4–L5 joint compression forces and the lumbar spine stability quantified with the stability index Load level
0
1
2
3
Moment about L4–L5 (Nm) Extension Flexion Lateral bending Axial rotation Vertical loading Lifting
1 (0) 1 (0) 1 (0) 1 (0) 1 (0) 75 (9)
29 (5) 32 (5) 30 (5) 27 (2) 1 (0) 99 (12)
55 (11) 57 (11) 56 (11) 50 (5) 1 (0) 119 (10)
82 (18) 84 (17) 82 (18) 71 (9) 1 (0) 142 (12)
L4–L5 Compression force (N) Extension Flexion Lateral bending Axial rotation Vertical loading Lifting
750 (374) 750 (374) 750 (374) 750 (374) 750 (374) 1478 (158)
1104 (87) 983 (241) 918 (113) 1204 (220) 790 (144) 1938 (177)
1468 (121) 1501 (431) 1364 (141) 1801 (235) 981 (295) 2276 (194)
1957 2072 1891 2205 1139 2626
Stability index (Nm/rad) Extension Flexion Lateral bending Axial rotation Vertical loading Lifting
172 172 172 172 172 422
323 161 207 279 162 521
416 243 315 413 191 604
549 309 435 485 222 690
(107) (107) (107) (107) (107) (63)
(38) (71) (46) (40) (43) (48)
(68) (75) (54) (64) (91) (58)
(210) (440) (272) (420) (345) (143)
(62) (83) (63) (100) (110) (49)
Means and SD in parenthesis.
Table 2 Frequency of buckling (instability) predicted by the model in an intact spine and with each individual muscle group removed. These trials represent 4% of the 5280 simulation trials performed Muscle group removed
Extension
Intact Rectus Abdominus External oblique Internal oblique Latissimus dorsi Iliocostalis lumborum Longissimus thoracis Lumbar erector spinae Multifidus Psoas Quadratus lumborum
0 0 0 0 0 0 0 0 0 0 0
Total
0
Flexion
Lateral bending
Axial rotation
Vertical loading
11 10 4 17 12 15 35 11 11 11 11
0 0 0 0 0 0 3 0 0 0 1
0 0 0 0 0 0 0 0 2 0 0
0 1 3 0 4 8 14 1 0 0 3
0 1 0 5 8 0 0 0 8 0 2
11 12 7 22 24 23 52 12 21 11 17
148
4
2
34
24
212
Table 3 Results of three-factor analysis of variance Source
d.o.f.
ðF Þ
ðP Þ
Exertion direction Load level Muscle removed Direction load Direction muscle Load muscle Direction load muscle
5 3 9 15 45 27 135
20.4 45.5 652.1 1.5 49.3 10.5 4.0
0.000 0.000 0.000 0.119 0.000 0.000 0.000
Degrees of freedom (d.o.f.), F -statistics ðF Þ, and significance ðP Þ.
Lifting
Total
effect on spine stability in all exertion directions (Fig. 1(a)–(f)). The contribution of individual muscles to spine stability with increasing load levels was either decreasing, increasing, or stayed approximately the same depending on the muscle and the direction of exertion (see, for example, internal oblique, iliocostalis lumborum, and psoas in flexion exertion, Fig. 1(b)). Although statistically significant (Table 3), the largest differences between different loading conditions did not exceed a 10% change in SI.
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
103
Fig. 1. Relative stability of the lumbar spine (percent of the intact stability index value) after the removal of each muscle group individually from the model during simulation of isometric trunk exertions in (a) extension, (b) flexion, (c) lateral bending to the left, (d) clockwise axial rotation, (e) vertical loading, and (f) lifting hold.
4. Discussion In this study, we compared the relative contribution of 10 major trunk muscle groups to the stability of the lumbar spine by removing each muscle systematically from the biomechanical model developed for quantifying spine stability. We purposely did not normalize each
trunk muscle to its size, moment arm, or activation because we wanted to report the contribution of each muscle as they function in vivo under various loading situations. The results support our hypothesis that no one muscle can be identified as the most important for the stability of the lumbar spine. Instead, the relative contribution of each muscle to spine stability depended
104
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105
on trunk loading direction and magnitude. These results are not surprising, because the overall structural stability of the spine is a highly nonlinear function of many variables. In addition to the activation levels of all trunk muscles, spine stability is determined by ligamentous stiffness of the inter-vertebral joints, spine posture, and the magnitude, direction and end-conditions of external trunk loads [4]. Moreover, even if a given muscle fascicle does not cross all of the intervertebral joints, its activation level (stiffness) affects the stability at all other remaining joints [4]. Therefore, the relative contribution of a given muscle to spine stability will change not only with different loading conditions but also with different recruitment patterns of the other trunk muscles. Given the results of the present study, we suggest that the often-used classification of muscles into ‘‘local’’ (deep, inter-segmental) and ‘‘global’’ (superficial, multisegmental) systems [1], as the way to discriminate between the muscles responsible for inter-segmental stability and spine motion, is incorrect. All trunk muscles contribute to spine stability and their contribution depends on many state variables. Stability or instability should be thought of as a state of the entire spine system determined by the activation of all trunk muscles, passive joint properties, spine posture, and loading conditions. Perhaps low back rehabilitation should be approached from a global system point of view involving the entire lumbar spine musculature with its motor control to enhance spine stability under various loading conditions. There are several limitations related to the biomechanical model used in this study. We did not include transversus abdominus in the model and the internal oblique muscle was represented by only two fascicles. Transversus abdominus does not have the potential to produce moments around the lumbar spine directly [17,23]. The mechanical effect of this muscle on the spine is most likely mediated through the intra-abdominal pressure mechanism, which is not yet very well understood [17]. Given the size of the transversus, even if all of its mechanical effects were transmitted to the spine, we would not expect its relative contribution to spine stability to be disproportionably larger than other trunk muscles. However, we do acknowledge that this muscle may play an important spine stabilizing function in preparation to sudden movements, because it is the first muscle to be recruited in healthy subjects under such conditions [8,9,18]. The internal oblique muscle is believed to have a great potential for stabilizing the spine laterally, because of its attachments to each lumbar vertebrae via lumbodorsal fascia [24]. However, we do not suspect that this architecture would greatly increase the contribution of the internal oblique muscle to spine stability estimated in our study. First, the possibility of moment transmis-
sion through lumbodorsial fascia has been questioned in the past [25]. Second, the analysis of stability cannot look at just one direction, but must consider all of the d.o.f. simultaneously. Finally, our results did not indicate an overwhelmingly greater spine-stabilizing advantage of the inter-segmental muscles over the multisegmental muscles. The stiffening effect of the inter-vertebral joints due to the compressive forces arising from muscle contractions [26] was not considered in our biomechanical model. However, this would not greatly affect our results because we compared the relative contribution of muscles to spine stability. Specifically, each muscle is expected to add a spine compression force that is proportional to its size and activation level. Subsequently, a similar proportion will be contributed to the passive joint stiffness by each muscle. Because muscle stiffness is also proportional to its size and activation level [1,3,20], these factors are already taken into account when the relative contribution of each muscle to spine stability was computed. Therefore, the net effect of not accounting for increased inter-vertebral joint stiffness due to compression force was the underestimated overall stability of the lumbar spine, but not the percentage of each muscle’s contribution to spine stability.
5. Conclusions No one trunk muscle could be identified as contributing the most to spine stability under all different loading conditions. The relative contribution of a given muscle to spine stability depended significantly on loading magnitude and direction. Furthermore, dysfunction (removal) of any single trunk muscle did not reduce the overall spine stability by more than 30%. Therefore, rehabilitation exercises for enhancing spine stability should involve the entire spinal musculature and its motor control under various spine-loading conditions.
Acknowledgements This study was supported in part by the Whitaker Foundation Biomedical Engineering Research Grant and the Gaylord Rehabilitation Research Institute Grant.
References [1] Bergmark A. Stability of the lumbar spine. A study in mechanical engineering. Acta Orthop Scand Suppl 1989;230:1–54.
J. Cholewicki, J.J. VanVliet IV / Clinical Biomechanics 17 (2002) 99–105 [2] Crisco JJ 3rd, Panjabi MM. The intersegmental and multisegmental muscles of the lumbar spine. A biomechanical model comparing lateral stabilizing potential. Spine 1991;16:793–9. [3] Gardner-Morse M, Stokes IA, Laible JP. Role of muscles in lumbar spine stability in maximum extension efforts. J Orthop Res 1995;13:802–8. [4] Cholewicki J, McGill SM. Mechanical stability of the in vivo lumbar spine: implications for injury and chronic low back pain. Clin Biomech 1996;11:1–15. [5] Cholewicki J, Panjabi MM, Khachatryan A. Stabilizing function of trunk flexor-extensor muscles around a neutral spine posture. Spine 1997;22:2207–12. [6] Granata KP, Marras WS. Cost-benefit of muscle cocontraction in protecting against spinal instability. Spine 2000;25:1398–404. [7] Magnusson ML, Aleksiev A, Wilder DG, Pope MH, Spratt K, Lee SH, et al. Unexpected load and asymmetric posture as etiologic factors in low back pain. Eur Spine J 1996;5:23–35. [8] Hodges PW, Richardson CA. Delayed postural contraction of transversus abdominis in low back pain associated with movement of the lower limb. J Spinal Disor 1998;11:46–56. [9] Hodges PW, Richardson CA. Altered trunk muscle recruitment in people with low back pain with upper limb movement at different speeds. Arch Phys Med Rehabil 1999;80:1005–12. [10] Radebold A, Cholewicki J, Panjabi MM, Patel TC. Muscle response pattern to sudden trunk loading in healthy individuals and in patients with chronic low back pain. Spine 2000;25:947–54. [11] Panjabi MM. The stabilizing system of the spine. Part I. Function dysfunction, adaptation, and enhancement. J Spinal Disord 1992;5:383–9. [12] Richardson C, Jull G, Hodges P, Hides J. Therapeutic exercise for spinal segmental stabilization in low back pain. Scientific basis and clinical approach. London: Churchill Livingstone; 1999. [13] O’Sullivan PB. Lumbar segmental ‘instability’: clinical presentation and specific stabilizing exercise management. Man Ther 2000;5:2–12. [14] Santaguida PL, McGill SM. The psoas major muscle: a threedimensional geometric study. J Biomech 1995;28:339–45.
105
[15] McGill S, Juker D, Kropf P. Quantitative intramuscular myoelectric activity of quadratus lumborum during a wide variety of tasks. Clin Biomech 1996;11:170–2. [16] Pietrek M, Sheikhzadeh A, Nordin M, Hagins M. Biomechanical modeling of intra-abdominal pressure generation should include the transversus abdominis. J Biomech 2000;33:787–90. [17] Hodges PW. Is there a role for transversus abdominis in lumbopelvic stability? Man Ther 1999;4:74–86. [18] Cresswell AG, Oddsson L, Thorstensson A. The influence of sudden perturbations on trunk muscle activity and intra-abdominal pressure while standing. Exp Brain Res 1994;98:336–41. [19] Cholewicki J, Juluru K, Radebold A, Panjabi MM, McGill SM. Lumbar spine stability can be augmented with an abdominal belt and/or increased intra-abdominal pressure. Eur Spine J 1999;8:388–95. [20] Cholewicki J, McGill SM. Relationship between muscle force and stiffness in the whole mammalian muscle: a simulation study. J Biomech Eng 1995;117:339–42. [21] 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;28:321–31. [22] Cholewicki J, McGill SM. EMG assisted optimization: a hybrid approach for estimating muscle forces in an indeterminate biomechanical model. J Biomech 1994;27:1287–9. [23] McGill SM. A revised anatomical model of the abdominal musculature for torso flexion efforts. J Biomech 1996;29:973– 7. [24] Macintosh JE, Bogduk N, Gracovetsky S. Biomechanics of the thoracolumbar fascia. Clin Biomech 1987;2:78–83. [25] McGill SM, Norman RW. Potential of lumbodorsal fascia forces to generate back extension moments during squat lifts [see comments]. J Biomed Eng 1988;10:312–8. [26] Janevic J, Ashton-Miller JA, Schultz AB. Large compressive preloads decrease lumbar motion segment flexibility. J Orthop Res 1991;9:228–36.