Bony Stress and Its Association With Intervertebral Disc Degeneration in the Lumbar Spine: a Review of Clinical and Basic Science Studies


 BackgroundMultiple components of the lumbar spine interact during the normal and pathological function. Bony stress in the lumbar spine is recognised as a factor in the development of pars defects and stress fractures, but its relationship with intervertebral disc (IVD) degeneration is not well understood. Therefore, a translational review encompassing basic science and clinical evidence was conducted to examine the relationship between bony stress in the lumbar spine and IVD degeneration.MethodsOnline databases Scopus, PubMed and MEDLINE via OVID were searched for relevant studies published between January 1980 – February 2020, using PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines. Two authors independently analysed the data, noting study characteristics and bias.ResultsThirty-two articles were included in the review: eight human studies, nine finite element analysis studies, three in-vivo biomechanical testing studies, and twelve in-vitro biomechanical testing studies. Of the 32 articles, 19 supported, four rejected and nine made no conclusion on the hypothesis that there is a positive associative relationship between IVD degeneration and bony stress. However, not enough evidence was available to confirm or reject a causal relationship.ConclusionMost studies suggest that the prevalence of IVD degeneration increases with the presence of bony stress, but whether a causal relationship exists is unclear. The literature recommends early diagnosis and clinical suspicion of IVD degeneration and bony stress. Longitudinal studies are required to explore causal relationships between IVD degeneration bony stress.

degeneration, we conducted a review of the basic science (in-vivo animal testing, computer modelling, and in-vitro biomechanical testing) and clinical data available in the published literature.

Eligibility Criteria
Only journal articles and book chapters in English, published between January 1980 and February 2020 were considered for this review. The inclusion criteria were: 1. Studies reporting clinical data on bony stress and IVD degeneration in the lumbar spine 2. Finite Element (FE) modelling studies making use of imaging data from living human subjects or human cadaveric material 3. Biomechanical in vitro testing studies making use of human and non-human cadaveric material 4. In vivo animal studies reporting on the association between induced vertebral damage and IVD degeneration The exclusion criteria were: 1. Studies reporting clinical data on the surgical treatment of bony stress fracture in the spine

Search Strategy
A review of the literature was undertaken following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (15). Electronic databases Scopus, PubMed, and MEDLINE, were searched using relevant keywords for journal articles and book chapters in English, published between January 1980 and February 2020. The search strategy, as presented schematically in Fig. 2, was intentionally designed to be of high sensitivity and low speci city to ensure that a minimum of potentially relevant studies was excluded based on poor keyword assignment or regional variations in terminology and spelling.

Study Selection
Abstract screening, full-text review for the eligibility criteria were used to include relevant studies. This process was conducted by two reviewers (DC and UC).

Data Extraction and Risk of Bias
Studies included in the nal search were read, analysed, and synthesised independently by two reviewers (DC and UC). A third reviewer (ADD) was available to resolve any disagreement between the two reviewers but was not needed. Due to the heterogeneity in the type of studies considered, the risk of bias was independently assessed for each study (as opposed to using a bias assessment tool). Data extracted from clinical studies varied depending on how the outcomes were reported and included speci c details about the cohort, outcomes assessed, and results. The primary outcome of interest was the evidence of association (or lack thereof) between IVD degeneration and bony stress in the lumbar spine, which was de ned as an increased (or decreased) prevalence of bony stress with IVD degeneration.

Data Presentation
Due to heterogeneity in the type of studies included in the review and the outcomes reported, a systematic review or meta-analysis of the data could not be undertaken. Therefore, data were collated into tables grouped by the type of study, and a narrative overview of the results was presented. Pictograms of the lumbar spine were used to present the synthesis of results from basic science and clinical studies separately

Results
The search identi ed 2810 potential studies with 2395 studies remaining after duplicates were removed. Following a review of the titles and abstracts of all 2395 studies, 73 studies were identi ed for potential inclusion. On examination of the full texts against the inclusion criteria, 32 were included in the nal review. Reference lists from these 32 studies were further reviewed for identifying relevant studies, but none were found. A owchart of the selection process is presented in Fig. 3.
The characteristics of each of the 32 studies are presented in Table 1. Eight clinical, nine FE analysis, three in-vivo biomechanical testing, and 12 in-vitro biomechanical testing studies are presented in this review. In the clinical studies, cohort sizes ranged from n = 22 to 642. The studies included retrospective reviews, cross sectional studies and prospective follow-up studies. Basic science studies (FE analysis, in-vivo biomechanical testing and in-vitro biomechanical testing) were mostly cross-sectional. Cohort sizes ranged from n = 15 to 24 for the in-vivo biomechanical testing studies and n = 6 to 600 for the in-vitro biomechanical testing studies. All FE analysis studies based their model on data from one human subject. Table 1 Characteristics of studies included in the review. Thirty-two studies were included in total: eight clinical, nine nite element analysis, three invivo biomechanical testing and 12 in-vitro biomechanical testing studies. Due to heterogeneity in the studies, data extracted differed between study categories, and therefore bias is reported in a narrative format. Risk of bias was assessed in each study. Limitations included small cohort (or sample) sizes, observer bias, publication bias and selection bias, as well as bias from more recent publications using data from older publications. However, no study was excluded based on the risk of bias. The majority of studies (59%) reported positive associations between bony stress and IVD degeneration; 13% reported a negative association and 28% did not conclude. The majority (63%) of human studies reported a positive association, none reported a negative association, and 37% did not conclude. The majority (56%) of FE modelling studies reported a positive association; 22% reported a negative association, and 22% did not conclude. None of the in-vivo biomechanical testing studies reported a positive association; 33% reported a negative association, and 67% did not conclude. The majority (75%) of in-vitro biomechanical testing studies reported a positive association; 8% reported a negative association, and 17% did not conclude.
The data from basic science and clinical studies were synthesised separately and presented spatially using pictograms of the lumbar spine ( Figs. 4 and 5). However, data from in-vivo biomechanical testing studies suggest that bony stress and IVD degeneration are negatively correlated. Animal studies allow for in-vivo examination of spine models, which introduces important concepts of bone remodelling and use of muscles and ligaments (46). Ryan et al. (2008) found that applying an asymmetrical load to pig spinal segments resulted in increased stress in the annulus brosus (AF) and vertebral bone adjacent to the AF when compared to a symmetrical load (34). Baranto et al. (2005)found that pigs with induced IVD degeneration (hole drilled through the endplate) were able to bear more compressive load and had altered stress distribution in the bone compared to those without IVD degeneration (32). The results suggest a negative correlation between bony stress and IVD degeneration, where bony stress does not occur in the presence of IVD degeneration, but rather bone remodelling occurs as a protective mechanism for the IVD. However, cohort sizes in the in-vivo studies ranged from 15 to 24; and with different animal models used the degree to which results may be compared across studies and are translatable to clinical settings remains unclear (31)(32)(33).

Discussion
The majority of the clinical studies provided evidence for a positive correlation between bony stress and IVD degeneration (shown in Fig. 5); none supported a negative correlation. Outcomes were consistent throughout the clinical studies that showed a positive correlation, with all studies using MRI to diagnose bony stress and IVD degeneration (6,7,(18)(19)(20)(21) (7), which may suggest the importance of diagnosing early signs of bony stress even if the healing occurs in subsequent follow-up scans. Bony stress may also be a contributing factor to the non-speci c LBP, due to its association with IVD degeneration or due to the microtrabecular injury alone or both. However, this hypothesis needs further clinical investigation.
This review noted several gaps in the literature on the association between bony stress and lumbar IVD degeneration. There is not enough evidence to suggest that IVD degeneration is a cause of bony stress in the lumbar spine or vice versa. Whether one is a pathological sequela of the other, or whether the two constitute a protective mechanism to mitigate further degenerative changes in the lumbar spine is unknown.
Reliability of the results may also be limited in some studies due to small sample sizes and di culty in replicating lumbar spinal loading.
Hence it is di cult to generalise the results to broader patient populations. There is evidence of potential sources of bias throughout the literature, including observer bias and publication bias regarding the non-signi cant association. The small number of clinical studies, lack of high-quality evidence cohort studies, and heterogeneity in the outcome measures for the included studies are limitations that should be taken into account when interpreting results discussed in this review. Most of the studies examined bony stress in the pars interarticularis, but not in other regions of the lumbar spine. It is unclear if this is because the pars interarticularis has historically been studied more, or it is easier to visualise and diagnose bony stress in the pars interarticularis. The results of these studies may differ if bony stress in other regions were also assessed.
To our knowledge, this is the rst review of the literature examining the relationship between bony stress and IVD degeneration in the lumbar spine. This review combines basic science (FE analysis, in-vivo biomechanical testing and in-vitro biomechanical testing studies) and clinical studies. Studies utilising FE analysis can control various parameters and attempt to understand the in uence of just one parameter on the outcomes. Cadaveric studies utilise human or animal spinal tissues for in-vitro testing to examine altered biomechanics in specimens with induced defects. In-vivo animal studies simulate pathological conditions or interventions to understand the biological and biomechanical response in somewhat more realistic conditions. This translational review was designed to be inclusive of different types of studies to synthesise all available evidence to understand better the pathophysiology of bony stress in the lumbar spine and its relationship with IVD degeneration.

Conclusion
As evident from the majority of clinical and basic science studies, this translational review suggests that there is a positive associative relationship between bony stress and IVD degeneration in the lumbar spine. Bony stress in the lumbar spine may cause LBP due to its association with IVD degeneration, microtrabecular injury/in ammation or both. The prevalence of IVD degeneration increases with bony stress in the lumbar spine. In adolescents, bony stress in the lumbar vertebra may heal with time; however, the prevalence of IVD degeneration is higher in such individuals even after bony stress has resolved. We, therefore, recommend a careful diagnosis of bony stress in younger patients presenting with or without IVD degeneration, so that the increased likelihood of early IVD degeneration in such patients can be effectively managed. However, there is not su cient evidence to suggest whether IVD degeneration is the cause or the consequence of bony stresses in the lumbar spine. This review has identi ed a need for further basic science and clinical research to elucidate the role of bony stress in causing degenerative changes in the lumbar spine and LBP.

Availability of Data and Materials
The datasets used in this study are available from the corresponding author on reasonable request.
DC, UC and ADD were all involved in the conception and planning of the methods of this study. DC conducted the literature search. DC and UC independently extracted data from the selected studies. DC wrote the original manuscript and designed the gures. DC and UC were both involved in editing the manuscript. DC, UC and ADD edited the nalised version of the manuscript. All authors have read and approved of the nal manuscript. Figure 1 T2 weighted mid-sagittal cut on magnetic resonance imaging (MRI) of a 12-year-old female patient from Sharma et al. (2017) showing oedema in the L4 right pedicle (red arrow) ( Figure 1A). MRI of both initial presentation ( Figure 1B) and 47-month follow up ( Figure 1C) shows the progression of L4-L5 disc degeneration (blue arrows). Other discs in the lumbar spine, including the L5-S1 disc that normally experiences the greatest axial load, are spared from degeneration (7).

Figure 4
Colour coordinated diagram showing positive associations between bony stress, intervertebral disc (IVD) degeneration and load in the lumbar spine segment as reported in various cadaveric, live animal, and nite element modelling studies. A colour represents each loading force, and its corresponding region of positive association is shown with the same coloured star or cross. A positive association was de ned as an increased prevalence of bony stress with IVD degeneration or vice versa. The grey star has no corresponding load.

Figure 5
Colour coordinated diagram showing positive associations between bony stress and intervertebral disc (IVD) degeneration in the lumbar motion segment based on data extracted from clinical studies. A positive association between nucleus pulposus degeneration and bony stress in the pars interarticularis is shown in red, and a positive association between annular degeneration and bony stress in the pars interarticularis and the pedicle is shown in green. A positive association was de ned as an increased prevalence of bony stress with IVD degeneration.