The aim of this study was to obtain measurements of pressures and stresses in the intervertebral discs of scoliotic patients in order to try to prove the hypothesis that asymmetrical loading was present in the scoliotic spine and hence could realistically be involved in curve progression. We were also very interested in the stress environment of the intervertebral disc in scoliosis in general. We considered that recording from discs prior to excision during anterior scoliosis reconstructive surgery involved minimal additional patient risk. However, in anaesthetised patients with additional muscle relaxation (an essential aspect of the anaesthetic procedure) no loading due to muscle activity is present. We therefore were unsure if our approach would show any meaningful results.

Previous authors have theorised about the biomechanical role of asymmetrical loading in progression of the scoliotic deformity [1-3]. Non-operative treatments such as bracing are designed to counteract these loads [4-6]. The majority of studies have focused on the musculature as the origin of this loading asymmetry. Electromyographic measurements have demonstrated differences in muscle activity between the convex and concave sides of the spine [7-9]. Muscle biopsies additionally find a significantly lower percentage of Type I fibers in the multifidus muscle on the concave side, particularly at the curve apex and also in the superficial muscles above and below the apex [10-13].

The cellular mechanisms by which abnormal loading may generate deformity has been shown previously in cell culture and in animal models. Cells of bone and disc are very responsive to mechanical stress[14,15], with high loads inhibiting longitudinal growth [16-18]. An asymmetric load may thus lead to asymmetrical longitudinal growth and hence wedging of the vertebral bodies and intervertebral disc[19]. Scoliotic-like deformities have also been produced in otherwise healthy animals by applying asymmetrical loads across the spine [19-23]. These loads differentially affect longitudinal bone growth. The disc remodels and becomes wedged and distorted. Thus asymmetrical loading could lead to permanent changes in vertebral bodies and discs and hence contribute to progression of the scoliotic deformity.

Most of the information on loads acting on the scoliotic spine is derived from measurements of changes in spinal morphology and from modelling[3,24]. There have however been no direct measures of loading asymmetry. We have therefore used stress profilometry [25] to directly measure pressure and stress profiles across scoliotic discs during anterior reconstructive surgery. This technique involves introducing a miniature pressure transducer, side mounted on a needle, across the disc transversely. Two linear profiles of stresses are measuring with the transducer orientated parallel (termed horizontal) and perpendicular (termed vertical) to the endplate.

Table 1 Details of Scoliotic patients examined.

Table 2 Details of non-Scoliotic Patients (Control) examined.

The transducer was calibrated up to 1 MPa using a pressure vessel(a tube of circular cross section) with a known cross sectional area filled with water to which known loads were applied by a servocontrolled hydraulic materials testing machine (Dartec, Dartec Ltd. Stourbridge, United Kingdon). The rate of application of load during calibration was approximately 0.2 MPa/s. Over the range 0–1 MPa the voltage recorded was found to be linear with respect to pressure applied. Re-calibrations of the transducer showed that it was very stable and showed negligible drift over time. Before surgical use, the pressure transducer was sterilised using Perisafe (Antec International, Sudbury, UK) according to standard hospital protocols.

Figure 1 Photograph of intraoperative measurement set up. Rostral is to the left, caudal to the right. Three electrodes can be seen inserted through the convex annulus of three adjacent discs. The pressure transducer is seen inserted into the middle disc.

Figure 2 Part of intraoperative representative radiograph showing electrodes and pressure transducer inserted from the convex side of the scoliotic curve (same patient as figure 1).

Figure 3 Diagram of position of needle mounted pressure transducer in scoliotic disc during intraoperative pressure measurements. Detail of pressure transducer tip included to show plane of transducer membrane parallel to axis of introducer needle.

Chart software (ADInstruments, Oxford, UK) was used to convert the voltages obtained into stresses/pressures in accordance with the measured calibration coefficient. The data were then exported to Microsoft Excel for further analysis.

A peak in the vertical stress recorded in the concave annulus (far left side of the profile) was a common feature of the profiles in the scoliotic discs in this study, a typical example being shown in figure 4. This shows the vertical and horizontal stress profiles recorded during anterior scoliosis surgery in the L1/2 disc of patient 3, a 13 year old neuromuscular patient. When stress peak is defined as a stress value of at least twice that recorded in the anatomical nucleus, 13/25 scoliotic discs have this feature.

Figure 4 Typical profile of measurements across a single disc with a hydrostatic region. These results showing the typical features of a vertical stress peak in the concave annulus and a hydrostatic region towards the convexity were recorded in the L1/L2 disc (more ...)

Previous pressure profiles measured in non-scoliotic, non-degenerate discs of similar age to the scoliotic patients show large hydrostatic central areas (equal vertical and horizontal pressures)[25,26] In many scoliotic discs we found that over significant areas, horizontal and vertical stresses recorded from the centre of the disc were of different magnitude indicating non-hydrostatic behaviour. In order to quantify this, we defined a hydrostatic region as one in which over at least two consecutive readings, the horizontal and vertical stresses were within 0.01 MPa. Infact, in 11/25 discs there were no regions of hydrostatic pressure. A disc with a hydrostatic region is illustrated in figure 4, the hydrostatic pressure region is shown at consecutive readings 13,18 and 23 mm from the concave disc boundary and can be seen to be shifted towards the convex side of the nucleus. An example of a disc with no hydrostatic regions is shown in figure 5 (Patient 2 L1/L2, 17 year old with idiopathic scoliosis) which also shows the pattern of high vertical stresses in the concavity compared to other regions of the disc.

Figure 5 Typical profile of measurements across a single disc with no hydrostatic region. These results showing a vertical stress peak at the concavity but no hydrostatic region were recorded in the L1/L2 disc of a 17 year old idiopathic scoliotic patient (Patient (more ...)

In 5 cases, peaks in the convex annulus also occurred. On only 2 occasions was this higher than the maximum vertical stress in the concave annulus in the same disc.

Figure 6 Hydrostatic pressure levels measured in all scoliotic discs where a hydrostatic pressure region was recorded. Results are shown versus disc level relative to the apical disc.

Hydrostatic disc pressures in the apical disc were not obviously higher overall or higher than at adjacent levels and the apical discs had more hydrostatic nuclei 6/7 (86%) compared with 10/18 non-apical discs (56%).

We found a slight trend towards higher mean pressure progressing caudally for absolute disc level, shown in figure 7. This graph shows the mean pressure at each disc level for all the scoliotic patients. Mean pressures were >0.1 MPa at all levels and tended to be similar in magnitude apart from the rostral 2 discs and most caudal disc. There was also no effect of age on levels of pressure measured (not shown).

Figure 7 Variation of the mean hydrostatic pressure level with absolute disc level. Results shown for all scoliotic discs where a hydrostatic pressure region was recorded.

Figure 8 The maximum vertical stress measured in the concave annulus in scoliotic discs. Results are shown for all discs in which a stress peak was recorded versus disc level relative to the apical disc.

The stress change across each of the scoliotic study discs is shown in Figure 9 relative to apical disc level; this difference gives a measure of load asymmetry.

Figure 9 Difference in vertical stress between the concave and convex annulus in scoliotic discs. Results are shown for all discs where recordings were made in both the concave and convex annulus versus disc level relative to the apical disc.

Due to the low numbers of patients who both were having anterior surgery for non-scoliotic conditions and who consented to having pressure measurements performed during the time period of the study, we only have pressure profiles from 10 discs for comparison.

Table 2 shows clinical details for the two patients with back pain and two patients with kyphosis in the study.

Patient P10 who was 26 years old showed a pressure profile with higher stresses towards the posterior annulus, a hydrostatic region and a trough of horizontal pressure in the anterior annulus. In both discs of patient 11, stresses were essentially zero throughout apart from an increase in vertical stress in both and a horizontal stress peak in the anterior annulus of the L5/S1 disc with no hydrostatic regions.

In the kyphotic patients, patient 8 who had quite a flexible curve clinically showed profiles with pressure peaks mainly in the anterior annulus with an under pressurised central region. The T11/12 and L1/2 levels showed hydrostatic nuclei. Patient 9 who was a kyphotic with a curve that was noted to be very stiff intra-operatively had very narrow discs and made introduction of the pressure probe difficult. Only a few recordings per disc (mean of 2.75) were made and so profile shape could not be accurately determined.

For the patients with back pain, hydrostatic regions were seen in the L4/5 disc of patient 10 (0.214 MPa). The central regions of both discs in patient 11 showed negative pressures very close to zero level (L4/5 and L5/S1, means of -0.023 and 0.040 MPa respectively) and the horizontal and vertical pressures were very close in value over these regions. In the kyphotic patients, only the T11/12 and L1/2 discs of patient 8 showed hydrostatic regions with a mean of 0.030 ± 0.022 MPa. For patient 9, there were no hydrostatic regions however in order to indicate the general size of the stresses present, the mean vertical stress was 0.327 ± 0.33 MPa and horizontal was 0.627 ± 0.46 MPa over all 4 recorded discs.

All the non-scoliotic discs showed some stress difference across the disc. It was very variable in magnitude ranging from .057 to 1.33 MPa with a mean of 0.35 ± 0.39 MPa (See figure 10).

Figure 10 Difference in vertical stress between the anterior and posterior annulus in non-scoliotic patients. Results are shown versus absolute disc level for each disc.

Because of the small number of non-scoliotic patients in this study, with the permission of the author (Mr Yonezawa) we have also compared our results with intradiscal pressures measured in Japanese subjects with back pain prior to laser nucleotomy[29]. In this publication, intra-discal pressure measurements were made on patients under general anaesthesia and hence are comparable to our patients in terms of levels of spinal loading due to similar levels of muscular tension. This published series has older patients than those of the present study but neither we nor Yonezawa et al. found any effect of age on intradiscal pressure[29]We only used data from discs that we thought had hydrostatic nuclei from Yonezawa's study. We determined this by defining a hydrostatic nucleus as one in which the vertical and horizontal pressures differed by less than 0.01 MPa. For all those patients with hydrostatic nuclei (13/18) the mean disc pressure was 0.058 ± 0.03 MPa and for those age matched to the study patients the mean disc pressure was 0.070 ± 0.030 MPa.

In our study, the mean value of hydrostatic pressure for the scoliotic patients was 0.25 ± 0.10 MPa, > 3 fold higher than that measured in Yonezawa's patients.

Figure 11 shows how the mean hydrostatic pressure measured in the scoliotic discs compared with pressures measured during surgery for other spinal pathologies. It is apparent that the pressure levels measured in scoliotic patients are considerably greater than those measured in other anaesthetized patients undergoing anterior spinal surgery

Figure 11 Mean hydrostatic pressure measured during surgery in scoliotic discs compared to other conditions. Details of the scoliotic and study back-pain and kyphotic patients are given in tables 1 and 2 respectively. The other results of disc pressures in patients (more ...)

Rather surprisingly, the readings obtained in the scoliotic discs were much higher in magnitude than we expected. The mean pressure reading in those discs with a hydrostatic nucleus (14/25) was 0.25 ± 0.10 MPa which was on average >3 fold greater than in the available control data (Figure 11).

Despite the lack of muscle activity, our results also showed evidence of asymmetrical loading in the scoliotic spine. Stresses were higher in the concave than in the convex annulus in 18/25 discs (Figure 9). We did not however find increased hydrostatic pressure (Figure 6) or peak stress levels (Figure 8) in the apical disc compared with its neighbours and in all but one disc, stress change was lowest in the apical disc compared with adjacent levels.

The pressures we measured in discs with hydrostatic regions in the scoliotic patients were considerably higher than those measured in patients with kyphosis and with chronic low back pain, which were on average low and of similar magnitude to that reported by Yonezawa[29]. In addition, the mean intradiscal pressures measured in scoliotic patients (0.25 MPa; Fig 11) were also higher than pressures measured in healthy awake volunteers in similar postures, 0.12 MPa in the L4/5 disc of a healthy orthopaedic surgeon[43] and a mean of 0.15 MPa in a group of 22–29 y old Japanese volunteers with no disc degeneration[31]. These pressures indicate overloading of the scoliotic disc and hence if present chronically could cause relative dehydration and also lead to the non-hydrostatic behaviour seen in some discs.

The finding of high pressures and stresses in recumbent, anaesthetised scoliotic patients with minimal loading due to muscle activity was unexpected and its origins are unclear.

The internal mechanical environment of the intervertebral disc is complex. The factors influencing disc pressure/stress at any point in the disc will arise from both intrinsic factors, viz. disc swelling pressure and matrix organisation [44]and extrinsic factors including muscle action, body weight and ligamentous tethering[43]. Swelling pressure arises from the balance between tissue composition, particularly glycosaminoglycan concentration, and the opposing tension imposed by the collagen network [44]and while changes in glycosaminoglycan concentrations have been reported across scoliotic discs[37,45], these would tend to rather affect the swelling pressure profile than increase swelling pressure levels.

With regard to externally imposed forces, body weight and muscle forces, though altered in scoliotic patients[46], should play no role in these anaesthetized recumbent patients. The patients in this study were also all well supported inferiorly by an evacuated bean-bag which should have reduced externally imposed loading to a minimum. However, the surgical positioning of a patient with a curved, axially rotated spine might increase torsion in the deformed segments of the spine. The 7 surgical patients were all in the lateral position and in vitro tests show that imposition of rotation, flexion or extension in axially loaded spines can lead to a pressure rise in the disc [47,48].

Ligamentous tethering or changes in annulus organisation[49]could be another factor pre-stressing the disc and causing higher pressures however we feel that this is unlikely to explain the magnitude of change seen. Recent studies of the lumbar fascia have shown that it can transmit loads and has contractile properties and hence is an intriguing candidate for the origin of these forces[50,51]. Further studies of this structure in scoliosis would therefore be of interest.

In 18/24 discs, higher stresses were found on the concave side of the curve compared to the convex indicating asymmetrical loading. This is obviously not explainable by asymmetrical muscle loading since the spinal muscles were relaxed in these patients. The simple presence of concave annular stress peaks, due to the possible previously mentioned factors, could lead to a stress gradient across the disc. However, even in some discs without a defined stress peak, asymmetrical loading was present.

This study has found that stresses in scoliotic discs are abnormal. Scoliotic discs in recumbent anaesthetised, muscle relaxed patients have higher nuclear hydrostatic pressures than those measured in non-scoliotic discs in this study and reported in the literature. In 18/24 discs from patients, a stress gradient from concave to convex sides of the disc was measured, with some very high differential stresses (c. 1.0 MPa, Figure 9), indicating asymmetrical loading. In addition, the stress profiles seen in the scoliotic discs were very different to normal discs, showing similarities found in previous studies of degenerate discs [25-27].

Without the presence of muscle activity, these findings are very surprising and intriguing. Previous studies have implied that asymmetrical muscle loading is the origin of asymmetrical loading which via changes in cellular activity and hence asymmetrical growth causes a "vicous cycle" of progression in an otherwise physiologically normal spine[57,58]. The results shown indicate the presence of abnormal and asymmetric stresses in the scoliotic disc but in this case not due directly to muscle action. If they arise from remodelling of disc and ligaments or are due to alterations in thoracolumbar fascial mechanics they are likely to be carried through to normal daily postures. This could influence cellular behaviour and growth and hence generate progression of the scoliotic deformity.

AM carried out the pressure measurements in vivo, analysed the data and drafted the manuscript. DJ and MS were responsible for setting up and validating the pressure measurements in vitro and providing technical help. JF provided the clinical overview and expertise. JU and JF conceived the study, were involved in the study design and co-ordination and helped draft the manuscript. All authors read and approved the final manuscript.

This study was supported by Fondation Yves Cotrel and by the Arthritis Research Campaign (U0511).