Intervertebral Stiffness of the Spine Is Increased by Evoked Contraction of Transversus Abdominis and the Diaphragm: In Vivo Porcine Studies

Transcription

1 Intervertebral Stiffness of the Spine Is Increased by Evoked Contraction of Transversus Abdominis and the Diaphragm: In Vivo Porcine Studies SPINE Volume 28, Number 23, pp , Lippincott Williams & Wilkins, Inc. Paul Hodges, PhD,* Allison Kaigle Holm, PhD, Sten Holm, PhD, Lars Ekström, BS, Andrew Cresswell, PhD, Tommy Hansson, PhD, and Alf Thorstensson, PhD Study Design. In vivo porcine study of intervertebral kinematics. Objectives. This study investigated the effect of transversus abdominis and diaphragm activity, and increased intra-abdominal pressure on intervertebral kinematics in porcine lumbar spines. Background. Studies of trunk muscle recruitment in humans suggest that diaphragm and transversus abdominis activity, and the associated intra-abdominal pressure contribute to the control of intervertebral motion. However, this has not been tested in vivo. Methods. Relative intervertebral motion of the L3 and L4 vertebrae and the stiffness at L4 were measured in response to displacements of the L4 vertebra imposed via a device fixed to the L4 vertebral body. In separate trials, diaphragm and transversus abdominis activity was evoked by stimulation of the phrenic nerves and via electrodes threaded through the abdominal wall. Results. When intra-abdominal pressure was increased by diaphragm or transversus abdominis stimulation, the relative intervertebral displacement of the L3 and L4 vertebrae was reduced and the stiffness of L4 was increased for caudal displacements. There was no change in either parameter for rostral displacements. In separate trials, the diaphragm crurae and the fascial attachments of transversus abdominis were cut, but intra-abdominal pressure was increased. In these trials, the reduction in intervertebral motion was similar to trials with intact attachments for caudal motion, but was increased for rostral trials. Conclusions. The results of these studies indicate that elevated intra-abdominal pressure, and contraction of diaphragm and transversus abdominis provide a mechanical contribution to the control of spinal intervertebral stiffness. Furthermore, the effect is modified by the muscular attachments to the spine. [Key words: spinal stiffness, intra-abdominal pressure, diaphragm, transversus abdominis, porcine] Spine 2003;28: From the *Department of Physiotherapy, The University of Queensland, Brisbane, Australia, Department of Orthopaedics, Sahlgrenska University Hospital, Göteborg, Sweden, Department of Neuroscience, Karolinska Institutet, Stockholm, and Department of Sport and Health Sciences, University College of Physical Education and Sports, Stockholm, Sweden. Financial assistance was provided by the National Health and Medical Research Council of Australia, and the Swedish Research Council. The manuscript submitted does not contain information about medical device(s)/drug(s). Institutional funds were received to support this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript. Address correspondence to Dr. Paul Hodges, Department of Physiotherapy, The University of Queensland, Qld 4072, Australia; p.hodges@shrs.uq.edu.au The osseoligamentous spine is inherently unstable and dependent on muscle contraction to maintain stability. 1,2 Although contraction of all muscles of the trunk may contribute to its stability, there has been considerable debate in the literature regarding the effect of increased intra-abdominal pressure (IAP). It has been argued that increased IAP may tension the spine as a result of forces against the pelvic floor and diaphragm. 3,4 In addition, others have argued that IAP may increase the intervertebral stiffness of the spine. 5,6 However, there has been little agreement in the literature from in vitro, in vivo, and modeling studies. 6 8 Recent in vivo evidence in humans suggests that increased pressure, without associated contraction of the abdominal or paraspinal muscles, produces a small extension moment at the lumbar spine. 9 In these experiments, IAP was increased by the electrical stimulation of the phrenic nerves to evoke contraction of the diaphragm. In addition, there is preliminary evidence that pressure increased in this manner may increase the posteroanterior stiffness of the spine. 5 These studies have been limited to measurement of motion of large regions of the spine. No study has investigated the mechanical effect of IAP or contraction of the muscles associated with its production on intervertebral control in vivo. Recent neurophysiologic studies have argued for the importance of contraction of transversus abdominis (TrA) and the diaphragm and the associated increase in IAP for the control of intervertebral motion. Studies of voluntary limb and trunk movements 13 and external loading 14 have indicated that electromyographic (EMG) activity of TrA and/or the diaphragm occurs irrespective of the direction of the internal and external forces. This activity is consistent with a contribution of IAP and activity of TrA and the diaphragm to control of intervertebral stiffness. 12,15 However, the hypothesis requires confirmation by biomechanical data. This issue is of functional importance, as studies have indicated that the EMG activity of TrA may be delayed in people with chronic low back pain 16 and when back pain is induced by injection of hypertonic saline into the paraspinal muscles. 17 Furthermore, tonic EMG activity of TrA and the diaphragm and IAP are reduced when respiratory demand is increased by hypercapnoea. 18 The aims of this experiment were, firstly, to determine whether selective contraction of the diaphragm or TrA, and the associated increase in IAP, in the absence of other abdominal or paraspinal muscle activity, affects interver- 2594

2 Intervertebral Stiffness of the Spine Hodges et al 2595 tebral motion in the lumbar spine in an in vivo porcine model. If the intervertebral motion was decreased, a second aim was to determine whether this effect was specific to the direction of movement. Finally, if the intervertebral motion is decreased, to determine whether the effect of contraction of the diaphragm and TrA on the spine was due to the increase in IAP or tension of the attachments of these muscles to the spine via the diaphragmatic crurae or middle layer of the thoracolumbar fascia, respectively. Methods Animals and Surgical Procedures. Eleven adolescent domestic pigs (Swedish landrace), 4 months of age and weighing 50 to 60 kg, were used in the study. Approval for the study was obtained from the Animal Research Ethics Committee. The animals were sedated by an intramuscular injection of ketamine (Ketalar, 30 mg/kg body weight, Parke-Davis, Gwent, UK) and, after 10 minutes, were anesthetized with intravenously administered -chloralose (Chloralose, 100 mg/kg body weight, Sigma Chemical, St. Louis, MO). Additional maintenance doses were given each 10 to 15 minutes. The animal was tracheotomized to facilitate intubation and placed on a ventilator. A venous catheter was installed through which Ringer solution was continuously infused. Animals were placed on an adjustable operating table and a thin-film strain gauge (Gaeltec, UK), in a thin flexible catheter, was introduced into the stomach via the mouth to measure IAP. With the animal in right side-lying, stimulation electrodes were attached to the left phrenic nerve. An incision was made at the level of the seventh or eighth rib and a 10 cm section of the rib was removed to provide adequate exposure of the thoracic cavity. The phrenic nerve was identified inferior to the heart and an incision was made through the pleura to free the nerve. Two connectors (test leads) were attached directly to the prenic nerve. The skin was sutured to close the wound, and the procedure was repeated for the right phrenic nerve. The pig was then placed in prone for preparation of the spinal apparatus. A midline longitudinal incision was made dorsally from L2 to L5. The supraspinous ligament and posterior layer of the thoracolumbar fascia were removed to provide adequate exposure, and the tips of the spinous processes of L3 and L4 were removed to provide a flat surface for pin insertion. Although removal of these structures may have affected the mechanical stability of the spine, this would be constant for control and experimental trials. Holes were drilled via the pedicle into the vertebral body of L4 on each side into which pedicular screws (6 mm diameter, stainless steel, self-tapping) were firmly screwed to provide attachment for the motor to invoke movement of the vertebra. A cross-bar was firmly attached between the pedicle screws (Figure 1A). The cross-bar was attached to the motor via a vertical lever and lightweight-rod (Figure 1A). Interosseous pins (2.4 mm diameter, 120 mm long; Smith and Nephew Orthopaedics, Tuttlingen, Germany) were inserted approximately 20 mm into the spinous processes of L3 and L4 in the midsagittal plane for attachment of movement sensors to measure the intervertebral motion of the spine. The pins were angled at approximately 25 and 45 deg so that they were away from the device that attached the motor to L4 and did not contact each other during the motion experiments (Figure 1A). Finally, stimulation electrodes were inserted into the TrA Figure 1. Experimental set-up. A, Pigs were positioned in prone with the servocontrolled motor situated posteriorly. The phrenic wires were inserted via the lateral chest wall with the pig in side-lying position and are obscured from view. B, Orientation of motion sensors placed in the spinous processes of the L3 and L4 vertebrae. Trajectory data for a single trial and the orientation of the external references frame are shown as dashed lines. Note the relatively linear vertebral displacement in the Z-Y plane. C, Illustration of the effect of the caudal and rostral displacement of the L4 vertebra. Note the opposite relative displacement at the vertebrae above and below L4 (i.e., with caudal displacement the L3 L4 undergoes relative flexion while the L4 L5 segment is extended). with ultrasound guidance. Electrodes were inserted 150 and 200 mm lateral to the midline and were placed longitudinally to cross the L3 and L4 segments (Figure 1A). The stimulating wires were made from polyurethane-coated copper wire (0.20 mm diameter) and had 10 mm of insulation removed. Wires were inserted using a curved suture needle (7 cm curvature diameter) that was modified to allow attachment of the stimulating electrode to the suture wire. The location of the needle was identified with ultrasound imaging (7 MHz linear transducer, Aspen Model, Acuson), and the needle was passed ~3 cm through TrA before exiting to the surface. Animals were positioned with the thorax and pelvis on firm supports, leaving the abdomen unsupported (Figure 1A) to minimize the movement of the spine in response to increased pressure in the abdominal cavity. The hind limbs were flexed beside the body to place the spine in a relatively neutral position, and the animal was firmly secured to the table at the pelvic and low thoracic levels with belts. Stimulation Protocol. Electrical stimulation via the bipolar electrodes was provided from a constant voltage source (S9 Stimulator, Grass Instruments, Quincy, MA). Contractions of the diaphragm and TrA were evoked electrically with repetitive stimuli (square pulse voltage: 6 ms duration, 40 Hz frequency) of an intensity to produce a pressure increase of ~5 cm H 2 O (~0.5 kpa) in all conditions. This pressure increase was chosen as higher pressures induced movement of the trunk, which would place the spine in a different position to the control trials, and this pressure increase is within the range of pressures

3 2596 Spine Volume 28 Number Figure 2. Specificity of stimulation of transversus abdominis and anatomy of the vertebral attachments of the crural diaphragm. Mean data (SD) of change in muscle thickness as a proportion of the thickness of the relaxed muscle measured from ultrasound images with and without stimulation (n 11). * P (A) and representative ultrasound images taken with the transducer placed transversely on the abdominal wall at the L4 level (B) are shown. Note the relative increase in thickness of TrA with minimal change in thickness of obliquus internus (OI) and externus (OE) abdominis. C, Shows the anatomy of the crurae. Note the extensive attachment of the right crux to a midline ridge on the vertebral bodies of the lumbar vertebrae (extending to L6 and the sacrum) and the smaller left crux deep to the tendon of the right crux, also attaching to the midline ridge. recorded in humans during function (e.g., refs ). Before commencement of the trials, the position of the exposed area of the stimulation electrodes inserted into TrA was adjusted to optimize the isolation of the contraction of this muscle from the adjacent oblique muscles (Figures 2A and 2B). Contraction of the abdominal muscles was visualized using a transverse ultrasound image with the transducer placed between the medial and lateral stimulating electrodes (i.e., approximately the level of the L3 L4 intervertebral disc). As the increase in muscle thickness with contraction can be identified with ultrasound imaging, 22 the electrodes were moved (by pulling on the ends of the wires) until a change in thickness with stimulation could be identified in TrA, but not the oblique muscles. This was possible in all animals. Evaluation of the changes in thickness of the abdominal muscles with ultrasound imaging indicated that the electrical stimulation applied to the electrodes inserted into TrA produced a relatively isolated contraction of TrA (Figures 2A and 2B). The thickness of this muscle increased by 83 25% from the resting state (one-way analysis of variance [ANOVA], P 0.001). In contrast, there was no significant increase in the width of the other abdominal muscles (OE: P 0.89, OI: P 0.45). The stimulating electrodes were taped in the optimal configuration. Contraction of the diaphragm was confirmed from compound motor unit action potentials recorded via a pair of EMG surface electrodes placed longitudinally at approximately the 6 8 intercostal space. The location of the diaphragm was identified with ultrasound imaging to confirm the placement of the electrodes. Stimulation intensity was set for each muscle to achieve a pressure increase of ~5cm H 2 0(i.e., diaphragm: 1 5V, TrA: 5 90V). Three configurations of electrically evoked contraction were tested: 1) diaphragm contraction (bilateral); 2) bilateral TrA contraction; and 3) unilateral contraction of TrA (left [n 10] or right side [n 1]). Sagittal Plane Intervertebral Kinematics and Stiffness. In order to measure the intervertebral kinematics and stiffness of the spine, a torque was applied by a servocontrolled motor to the L4 vertebrae via a lever system attached to the pedicle screws and cross-bar (see above, Figure 1A). The motor was attached to a vertical lever via a multiaxial joint that was placed ~20 cm from the cross bar. The motor was set to displace the distal end of the lever 2 cm caudally and rostrally (Figure 1C) from the start position, at which zero force was recorded. The torque was applied at a frequency of 0.5 Hz for 4 cycles. Force required to achieve the imposed displacement was recorded with a force transducer placed between the actuator and the rod for connection to the lever system. Three-dimensional kinematics of the L3 and L4 vertebrae were measured with an electromagnetic motion analysis system (Motion Star, Ascension Technology Corp., Burlington, VT) with a resolution of 0.8 mm and 0.1. Motion sensors were attached to the L3 and L4 spinous processes via intraosseous pins (Figures 1A and 1B). Preliminary testing confirmed that the metal pins did not interfere with system. Data were recorded in trials with no electrical stimulation and during the three stimulation configurations. Control trials were conducted at the start of the trial and after every second stimulation trial. Force, movement, and pressure data were sampled at 100 Hz using a computerized data acquisition system (MotionMonitor, Innovative Sports Training, Chicago, IL).

4 Intervertebral Stiffness of the Spine Hodges et al 2597 Additional Experiments. Two additional experiments of sagittal plane motion were conducted on a subset of pigs. The aims of these series were: 1) if intervertebral motion was affected by diaphragm stimulation, to investigate whether the effect of diaphragm contraction on the spine was due to the resultant IAP increase or tension in the crural fibers of the diaphragm; and 2) if intervertebral motion was affected by TrA stimulation, to investigate the relative effect of IAP and tension in the middle layer of the thoracolumbar fascia. Porcine crurae of the diaphragm are shown in Figure 2C and attach to the anterior aspect of the vertebral bodies of L1 L6. To investigate the relative effect of IAP or crural tension from diaphragm contraction, a series of trials were conducted in five pigs. Firstly, the intersegmental kinematics were measured during evoked contraction of the diaphragm with both the abdominal cavity and crurae intact. In a second trial, intersegmental kinematics were measured with the abdominal cavity opened via a lateral incision (~10 cm) so that diaphragm contraction tensioned the crurae with no or minimal increase in IAP. Thirdly, the crurae were cut via a small incision (~5 cm) anterior to the left transverse processes, the abdominal incision was sutured, and a trial was conducted in which phrenic stimulation produced an increase in IAP (~5 cm H 2 0) but without crural tension. Control trials (no stimulation) were conducted before each trial to evaluate whether passive stiffness of the spine was affected by the interventions. To test the relative importance of the increase in IAP or tension in the thoracolumbar fascia, sequential trials were conducted in five pigs. Following a trial with TrA stimulation, the attachment of the left TrA to the transverse processes of L2 L5 was cut via a small posterolateral incision (~10 cm), and a trial was conducted with bilateral TrA stimulation. A further cut was made through the right attachment of TrA, followed by bilateral TrA stimulation. In all trials, the stimulation intensity was set to produce an increase in pressure of ~5 cm H 2 0. Trials without stimulation were conducted after each intervention. Data Analysis. Preliminary evaluation of the data indicated that the displacements of L4 and L3 were approximately linear (see raw data presented as dotted lines in Figure 1B). As there was limited angular displacement, all intervertebral kinematics were quantified as the relative difference in linear displacement of the sensors attached to L3 and L4. Linear displacement was quantified as the net displacement vector in all three planes. A decrease in intervertebral motion, despite identical induced motion of L4, would indicate an increase in intervertebral stiffness. The amplitude of peak motion in each direction was calculated. Stiffness was also calculated from the force-displacement properties of the L4 vertebra. In contrast to intervertebral kinematics, this measure quantifies the force required to move the vertebrae, irrespective of its relationship to adjacent vertebrae. Force was calculated from the actuator and displacement was quantified as the actual motion of the vertebrae (i.e., the displacement vector). For the measurement of spinal stiffness, the linear region of the force displacement curve was identified and the slope (for rostral and caudal trials) was used as a measure of stiffness. This was found to lie between ~1 and ~7 N of applied force (i.e., torque: ~0.2 and ~1.4 Nm). The difference in intervertebral kinematics (relative displacement of L3 and L4) and L4 stiffness between the control trials and each of the 3 stimulation configurations was compared for rostral and caudal displacement conditions using repeated measures ANOVAs. Post hoc testing was done with the Figure 3. Intervertebral kinematics (L3 L4) and L4 stiffness with displacement of the L4 vertebra for control trials and trials with phrenic stimulation. Representative raw data are show for a single trial. Motion of the L4 and L3 vertebrae in each plane (see Figure 1B for definition) (A), and the relative difference in the length of the vectors of L3 and L4 (B) are shown. In A, note the large displacement of L4 and lesser motion of L3 in the y and z planes and the increased of motion of L3 (i.e., associated with reduced relative intervertebral motion between L3 and L4) following phrenic stimulation. In B, bars underneath the data indicate the rostral displacement trials. Note the reduction of net relative intervertebral motion between L3 and L4 for caudal displacement with phrenic stimulation. Force-displacement plots for control and phrenic stimulation trials are shown in C. Note the increased slope for caudal displacement during phrenic stimulation, indicating increased stiffness. Duncan multiple range test. Analysis of the additional trials was done using repeated measures ANOVAs. Alpha-level was set at Results Intervertebral Kinematics Application of force to the L4 vertebra was designed to displace the sensor attached to L4 by ~1 cm in each direction (2 cm of actuator movement) in each direction from an initial position at which the applied force was zero. In control trials with no electrical muscle stimulation, this displacement of L4 also moved the L3 vertebra (Figure 3A). The net displacement of L3 (length of the vector in all three planes) was 7.2 (1.2) mm less than that of L4 in the caudal direction and 5.2 (1.8) mm less for rostral displacement. If the L3 L4 articulation was immobile, there would be no relative displacement between the vertebrae. Diaphragm Contraction When contraction of the diaphragm was evoked by electrical stimulation of the phrenic nerves, gastric pressure was increased by 5.1 (2.6) cm H 2 0, as intended. With force applied to the L4 vertebra in the caudal direction during this contraction, intervertebral motion between

5 2598 Spine Volume 28 Number Figure 4. Mean (SD) data for L3 L4 intervertebral kinematics and stiffness of L4. In A, the relative peak intervertebral motion for control trials and the 3 stimulation conditions are shown, expressed as a proportion of the value recorded for control trials (n 11). In B, stiffness is quantified as the slopes of the regression line fitted to the forcedisplacement data and expressed as a proportion of the value recorded during control trials (n 11). In C, data are presented for trials with the diaphragm intact and with manipulation of IAP and the crurae (n 5), and in D with TrA intact and with disruption of the vertebral TrA attachments (n 5). * P 0.05 between values bridged by horizontal brackets. the L3 and L4 vertebrae was decreased by ~10% to 6.6 (0.5) mm (P 0.05) (see reduced relative displacement marked with arrows in Figure 3B and Figure 4A), compared to control trials without stimulation, indicating increased intervertebral stiffness. In contrast, there was no difference when force was applied in the rostral direction (5.4 [0.5] mm; P 0.13; Figures 3B and 4A). Stimulation of the phrenic nerves also changed the forcedisplacement properties of L4. The slope of the regression line fitted to the force-displacement data of L4 (Figure 3C, upper panel) in the caudal direction was increased by 12 (6)% (Figure 4B), indicating that more force was required to move the vertebrae compared to the control trials (P 0.05). In contrast, the stiffness of L4 was decreased by ~14 (7)% when force was applied in the rostral direction (P 0.05; Figure 4B). However, unlike other parameters, there was interspecimen variation in this parameter, and the reduction in slope of the force-displacement properties was not identified for all trials (see Figure 3C, lower panel). In the additional trials in 5 pigs, the IAP increase from diaphragm contraction was decreased to 17 8% ( 1 cm H 2 0) of that achieved in the main trial after placing an incision along the abdominal wall. In this condition, when the diaphragm was stimulated, but IAP increase was minimized ( 2 IAP, crura intact in Figure 4C), the intervertebral motion to caudal displacement was not different from the control trials (P 0.18) and was greater than trials with the abdomen intact (P 0.02; IAP, crura intact in Figure 4C). When the crurae were cut at the L2 level and the abdomen was resutured ( IAP, no crura in Figure 4C), the increase in IAP from diaphragm contraction was 87 5% of that obtained before the incision. In this condition, diaphragm contraction again decreased the intervertebral motion when force was applied caudally (P 0.03) and was not different to the trials with the crurae and abdominal wall intact (P 0.18; IAP, crura intact in Figure 4C). When force was applied in the rostral direction with the crura intact, there was no difference between trials with and without increased IAP (P 0.25, Figure 4C). However, in contrast to the trials with the crurae intact, when IAP was increased, but with the crurae cut ( IAP, no crura in Figure 4C), contraction of the diaphragm increased the relative intervertebral motion in the rostral direction (P 0.03). That is, cutting the diaphragm crurae meant that elevated IAP produced an increase in intervertebral displacement in response to force applied in the rostral direction that was not observed when the crurae were intact. Transversus Abdominis Contraction. When the intensity of electrical stimulation of TrA was set to increase IAP to a similar level to that achieved in the diaphragm trials (i.e., 3.4 (1.6) cm H 2 O, P 0.2), bilateral contraction of this muscle decreased the relative intervertebral motion between L3 and L4, compared to control trials, in the caudal direction by 13.2 (6)% to 6.2 (0.5) mm (P 0.05; Figure 4A). There was no difference between the decrease in intervertebral motion achieved with the diaphragm or TrA contractions, when the increase in IAP was elevated to approximately the same pressure (P 0.19). Similar to contraction of the diaphragm, TrA contraction did not produce a significant change from control values of intervertebral motion when force was applied in the rostral direction (Figure 4A). However, the relative intervertebral motion was less in this direction with TrA contraction than when the diaphragm (P 0.05) was stimulated, and there was a trend for intervertebral stiffness to be

6 Intervertebral Stiffness of the Spine Hodges et al 2599 reduced compared to the control trials (P 0.1). Similar to the data for the diaphragm contraction, the slope of the regression line fitted to the force-displacement data of L4 in the caudal direction was increased by 16 (5)% with bilateral stimulation of TrA (P 0.01; Figure 4B), indicating that more force was required to move the vertebrae the set distance compared to the control trials. There was no change when force was applied in the rostral direction (Figure 4B). When TrA was electrically stimulated unilaterally (left: n 10, right: n 1) but the IAP increase was maintained at a similar level (3.7 [1.8] cm H 2 O, P 0.2) the contraction did not affect the intervertebral kinematics in either direction (Figure 4A). That is, the reduction in intervertebral motion that was achieved by bilateral contraction of TrA was not reproduced by unilateral contraction of the muscle. Unlike the data for intervertebral kinematics, with caudad displacement, the forcedisplacement data were consistent with an increase in intervertebral stiffness for trials in which TrA was stimulated unilaterally (Figure 4B). The slope of the forcedisplacement line was increased by 15 (5)% compared to control trials (P 0.02) and was not different to trials with bilateral stimulation of the muscle (P 0.8). Additional trials were undertaken in five pigs to compare the relative contribution of IAP and tension in the attachment of the muscle to the vertebral column via the middle layer of the thoracolumbar fascia by cutting the attachment of TrA from the lumbar transverse processes on the left and then right side. When the L4 vertebrae was displaced caudally, the reduction in intervertebral motion with bilateral stimulation was not different between trials with intact muscle ( TrA intact in Figure 4D) and with the unilateral ( TrA cut left insertion in Figure 4D) and bilateral ( TrA cut bilateral insertion in Figure 4D) separation of the vertebral attachments (P 0.25). However, for trials with rostral displacement of L4, the relative intervertebral motion was increased with both unilateral and bilateral separation of the vertebral attachment (P 0.05), and there was no difference between the two interventions (P 0.25; Figure 4D). Thus, when IAP is increased by TrA activity, without attachments to the transverse processes, there is no difference in the stiffening effect on caudal displacement, but in the rostral direction the stiffness reduced. Discussion The results of this study provide the first in vivo evidence that raised IAP and contraction of the diaphragm and TrA increase intervertebral stiffness. Furthermore, the data confirm that the mechanical effect of their contraction is dependent on both IAP and their attachments to the vertebral column. The measures used to quantify intervertebral kinematic behavior provide an indirect measure of intervertebral stiffness. Measurement of intervertebral kinematics is based on the premise that greater relative difference in motion of the vertebrae will occur if the intervening joint is flexible. Thus, any reduction in relative intervertebral motion would indicate an increase in stiffness. This is analogous to the quantification of the transmission of vibration across joints and bone fractures. 23 The advantage of this measure is that it provides a discrete index of stiffness at an intervertebral joint. In this study, the motion induced by the loading was complex. As demonstrated in Figure 1C, the imposed caudal displacement of L4 resulted in a relative flexion (and distraction) motion between L3 and L4, whereas rostral displacement produced a relative extension (and compression) motion. Thus, the results can be interpreted in view of this specific physiologic motion. The possibility that the changes in motion are due to changes in boundary conditions also requires consideration. The second measure of intervertebral control involved quantification of the slope of the regression line fitted to the force-displacement data of L4 as a measure of stiffness. Similar measures have been used to quantify the stiffness of the spine to posteroanterior displacements in humans 24 and shown to be a valid and reliable measure of spinal stiffness. 25 In this study, the displacement was induced by caudal and rostral displacement of the L4 vertebra. This measure is a less sensitive measure of stiffness of a single intervertebral segment as it may be influenced by other factors such as the stiffness at segments above and below the level of force application. In the present study, this was complex as the imposed L4 displacement produced opposite directions of motion at the segments above and below L4. Thus, the displacement did not represent a functional motion and emphasis was placed on the changes in intervertebral kinematics. An additional issue is whether the pig is an adequate model of the human spine. Several authors have compared the anatomy of human and quadruped spines and confirmed that they are comparable with few exceptions. For instance, pigs have an additional lumbar vertebrae and longer transverse processes. We dissected several pigs to investigate the anatomy of diaphragm crurae. This structure is more extensive than in the human, with attachment of the right crura via a thick tendon to a ridge on the anterior vertebral bodies and extending to L6 (Figure 2C). The left crurae is smaller and deep to these fibers. In humans, this structure is smaller and attaches to the upper lumbar vertebrae (right: L3, left: L2), 29 although fibers extend to the sacrum as the anterior longitudinal ligament. 30 Despite these variations, the effects of IAP and contraction of the diaphragm and TrA are likely to be similar. It must be acknowledged that coordination of activity of these muscles may vary between pigs and bipedal humans. Effect of Diaphragm Contraction and Increased IAP on Intervertebral Stiffness Original hypotheses of the effect of IAP on the spine suggested that it may tension the spine by exerting opposing forces on the pelvic floor and diaphragm. 3,31,32 On this basis, it was argued that an extension moment from IAP might reduce paraspinal activity and decrease

7 2600 Spine Volume 28 Number spinal loading. Indirect support was derived from studies demonstrating a relationship between IAP and loads applied to the spine. 13,20 However, others have not shown reduced paraspinal activity with increased IAP. 33,34 Moreover, modeling studies provide contradictory evidence. 6,35 37 For example, it is suggested that the moment arm and surface area of the diaphragm 6 and the IAP amplitude generated during functional tasks 38 are too small to generate a significant extensor moment. It has been difficult to evaluate the affect of IAP in vivo, as it is difficult to isolate the influence of IAP from the associated abdominal muscle activity. Recently, this problem was solved by elevation of IAP via electrical stimulation of the phrenic nerves in humans. 9 This study confirmed that IAP produces an extensor moment, although its amplitude was relatively small in that situation. This is consistent with modeling data, which predicted an extensor moment of similar amplitude to the in vivo data. 4 Alternatively, it has been argued that IAP may influence intervertebral stiffness directly by tensioning the spine or stiffening the abdominal contents, or indirectly by increasing the hoop tension in the abdominal muscles and their fascias. 6,9,12,39 In the present study, electrical stimulation of the diaphragm increased IAP without abdominal or paraspinal activity. The results indicate decreased intervertebral motion (increased stiffness) with caudal displacements but no change with rostral displacements. As the caudal displacements flexed the L3 L4 segment, and the IAP increase from phrenic nerve stimulation resisted this motion, the present data are consistent with a tensioning of the spine by IAP. As the change in intervertebral kinematics was observed with small pressures (~5 cm H 2 O) that are at the lower end of the range observed in human function (e.g.,5 220 cm H 2 O), 11,20,40 43 this argues for a significant contribution of IAP to intervertebral stiffness. The change in force-displacement properties with diaphragm stimulation was consistent with this finding. It is important to consider that the intervertebral flexion is likely to be also associated with a distraction component, which might also be affected by muscle contraction. Failure of IAP to change intervertebral stiffness with rostral displacement is of interest, as tonic diaphragm activity and sustained increases in IAP have been reported in tasks in which the perturbing force on the spine alternates between flexor and extensor moments. 10,44 However, these moments were not restricted to sagittal plane and IAP and diaphragm activity may increase intervertebral stiffness in other directions. For instance, IAP has been shown to contribute mechanically to the control of moments in the frontal plane. 8 An important consideration for the present results is the attachment of the diaphragm to the vertebrae via the crurae. As mentioned above, the porcine crurae are more extensive than in humans and may be anticipated to have a greater affect on the spine. However, the influence of crural tension has not been tested previously. The present data argue that although the crurae may not directly increase the stiffness of the spine, their absence results in increased intervertebral motion to rostral displacement. That is, the presence of the crurae appears to overcome the effect of IAP to increase intervertebral motion (i.e., reduce intervertebral stiffness) for displacements that extend the segment. Further studies are required to investigate the morphology of the crurae in humans and their potential to affect spinal stiffness. Effect of TrA Contraction and IAP on Intervertebral Stiffness It has been argued that the flexor moment from abdominal contraction counteracts any extensor moment produced by IAP and, thus, offset any beneficial action of IAP. 6,36 However, if IAP is increased by contraction of muscles with predominantly transversely orientated muscle fibers, such as TrA, this would be less critical. According to a model by Daggfeldt and Thorstensson, 4 fibers of the abdominal muscles with an inclination to the vertical of greater than 55 can contribute to unloading. Consistent with this, proposal studies have shown that TrA is closely associated with modulation of IAP. 10,13,14,19 The results of this study provide the first in vivo evidence that contraction of TrA can modulate intervertebral kinematics and stiffness. These data argue that TrA contraction increases lumbar stiffness and decreases relative intervertebral displacement, but only in response to caudal displacements. As this corresponds to the task of the simulated intervertebral flexion at L3 L4, this is consistent with the hypothesis that increased IAP tensions the spine and produces and extension moment. 4,9 Similar to the diaphragm, this is surprising, as studies of TrA indicate that this muscle is active in a manner that is not specific to the direction of applied force. 10,12,13 However, there was a trend for intervertebral motion to be reduced with motion in the rostral direction, and studies of motion in other planes may provide further insight. The TrA may influence the spine by increasing IAP or tensioning the thoracolumbar fascia. 4,8,12 The relative importance of these was assessed in this study by cutting the attachments of TrA to the lumbar transverse processes. Although this intervention did not change the effect of TrA on intervertebral kinematics for caudal displacement (i.e., the effect in this direction was due to IAP), when rostral displacement was applied, intervertebral motion was increased by TrA contraction. Thus, although the fascial attachment may not stiffen the spine, IAP may decrease the stiffness to rostral displacement if the fascial attachment is absent. Furthermore, the data indicate that unilateral contraction of TrA does not change relative intervertebral motion. Thus, the mechanical effect of TrA is dependent on bilateral contraction. Implications This study confirms that IAP and diaphragm and TrA contraction may provide mechanical stability to the lumbar intervertebral joints. This confirms that tonic and phasic activity of these muscles observed during functional tasks provides an integral component of spinal control. Further-

8 Intervertebral Stiffness of the Spine Hodges et al 2601 more, the data suggest that deficits in the postural activity of TrA (as identified in low back pain) 16 and the diaphragm (as occurs when respiratory demand is increased) 18 may lead to impaired stability of the lumbar spine. Key Points Intervertebral motion was evaluated with and without electrically stimulated activity of transversus abdominis and the diaphragm in pigs. Stimulation of both muscles and increased intraabdominal pressure reduced intervertebral motion with specific directions of movement. The effect of intra-abdominal pressure was modified when the muscle attachments of TrA and the diaphragm were cut. References 1. Lucas DB, Bresler B. Stability of the ligamentous spine. Technical Report esr. 11 No. 40. Biomechanics Laboratory; University of California, Berkeley and San Francisco; Panjabi MM. The stabilizing system of the spine. Part I. Function, dysfunction, adaptation, and enhancement. J Spinal Disord 1992;5: Bartelink DL. The role of intra-abdominal pressure in relieving the pressure on the lumbar vertebral discs. J Bone Joint Surg Br 1957;39B: Daggfeldt K, Thorstensson A. The role of intra-abdominal pressure in spinal unloading. J Biomech 1997;30: Hodges PW, Eriksson AEM, Shirley D, et al. Intra-abdominal pressure can directly increase stiffness of the lumbar spine. J Biomech, in press. 6. McGill SM, Norman RW. Reassessment of the role of intra-abdominal pressure in spinal compression. Ergonomics 1987;30: Cholewicki J, Juluru K, McGill SM. Intra-abdominal pressure mechanism for stabilizing the lumbar spine. J Biomech 1999;32: Tesh KM, ShawDunn J, Evans JH. The abdominal muscles and vertebral stability. Spine 1987;12: Hodges PW, Cresswell AG, Daggfeldt K, et al. In vivo measurement of the effect of intra-abdominal pressure on the human spine. J Biomech 2001;34: Hodges P, Gandevia S. Changes in intra-abdominal pressure during postural and respiratory activation of the human diaphragm. J Appl Physiol 2000; 89: Hodges PW, Butler JE, McKenzie D, et al. Contraction of the human diaphragm during postural adjustments. J Physiol (Lond) 1997;505: Hodges PW, Richardson CA. Feedforward contraction of transversus abdominis in not influenced by the direction of arm movement. Exp Brain Res 1997;114: Cresswell AG, Grundstrom H, Thorstensson A. Observations on intraabdominal pressure and patterns of abdominal intra-muscular activity in man. Acta Physiol Scand 1992;144: 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: Hodges PW, Cresswell AG, Thorstensson A. Preparatory trunk motion accompanies rapid upper limb movement. Exp Brain Res 1999;124: Hodges PW, Richardson CA. Inefficient muscular stabilization of the lumbar spine associated with low back pain: a motor control evaluation of transversus abdominis. Spine 1996;21: Hodges P, Moseley G, Gabrielsson A, et al. Acute experimental pain changes postural recruitment of the trunk muscles in pain-free humans. Exp Brain Res 2003;151: Hodges PW, Heijnen I, Gandevia SC. Reduced postural activity of the diaphragm in humans when respiratory demand is increased. J Physiol (Lond) 2001;537: Cresswell AG. Responses of intra-abdominal pressure and abdominal muscle activity during dynamic trunk loading in man. Eur J Appl Physiol 1993; 66: Grillner S, Nilsson J, Thorstensson A. Intraabdominal pressure changes during natural movements in man. Acta Physiol Scand 1978;103: Hemborg B. Intraabdominal pressure and trunk muscle activity during lifting [doctoral dissertation]. Department of Physical Therapy; University of Lund; Hodges PW, Pengel L, Herbert R, et al. Ultrasound measurement of changes in muscle geometry over a range of isometric contraction intensities. Proc Int Soc Biomech 2001;18: Buyruk H, Stam H, Snijders C., Measurement of sacroiliac joint stiffness in peripartum pelvic pain patients with Doppler imaging of vibrations (DIV). Eur J Obstet Gynecol Reprod Biol 1999;83: Lee M, Svensson NL. Measurement of stiffness during simulated spinal physiotherapy. Clin Phys Physiol Meas 1990;11: Latimer J, Goodsel MM, Lee M, et al. Evaluation of a new device for measuring responses to posteroanterior forces in a patient population. Part 1: Reliability testing. Phys Ther 1996;76: Smit TH. The use of a quadruped as an in vivo model for the study of the spine biomechanical considerations. Eur Spine J 2002;11: Wilke H-J, Kettler A, Claes LE. Are sheep spines a valid biomechanical model for human spines? Spine 1997;22: Yingling V, Callaghan J, McGill S. The porcine cervical spine as a model of the lumbar spine: an anatomical, geometric, and functional comparison. J Spinal Disord 1999;12: Williams PL, Warwick R, Dyson M, et al. Grays Anatomy. London: Churchill Livingstone; Bogduk N. Clinical Anatomy of the Lumbar Spine and Sacrum. London: Churchill Livingstone; Keith A. Man s posture: its evolution and disorders. Br Med J 1923;1: Morris JM, Lucas DM, Bresler B. Role of the trunk in stability of the spine. J Bone Joint Surg 1961;43A: Cholewicki J, Juluru K, Radebold A, et al. Lumbar spine stability can be augmented with an abdominal belt and/or increased intra-abdominal pressure. Eur Spine J 1999;8: Krag MH, Byrne KB, Gilbertson LG, et al. Failure of intraabdominal pressurization to reduce erector spinae loads during lifting tasks. Proc North Am Congress Biomech 1986; Bearn JG. The significance of the activity of the abdominal muscles in weight lifting. Acta Anat 1961;45: Nachemson AL, Andersson GBJ, Schultz AB. Valsalva maneuver biomechanics: effects on lumbar trunk loads of elevated intra-abdominal pressures. Spine 1986;11: Örtengren R, Andersson GBJ. Electromyographic studies of trunk muscles with special reference to the functional anatomy of the lumbar spine. Spine 1977;2: Davis PR, Troup JDG. Pressures in the trunk cavity when pulling, pushing and lifting. Ergonomics 1964;7: Aspden RM. Intra-abdominal pressure and its role in spinal mechanics. Clin Biomech 1987;2: Cresswell AG, Blake PL, Thorstensson A. The effect of an abdominal muscle training program on the intra-abdominal pressure. Scand J Rehabil Med 1993;26: Harman EA, Frykman PN, Clagett ER, et al. Intra-abdominal, intra-thoracic pressures during lifting and jumping. Med Sci Sports Exercise 1988;20: Hemborg B, Moritz U, Löwing H. Intra-abdominal pressure and trunk muscle activity during lifting. IV. The causal factors of the intra-abdominal pressure rise. Scand J Rehabil Med 1985;17: Marras W, Joynt RL, King AI. The force velocity relation and intraabdominal pressure during lifting activities. Ergonomics 1985;28: Hodges P, Gandevia S. Activation of the human diaphragm during a repetitive postural task. J Physiol (Lond) 2000;522: