Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article


Doctor of Philosophy


Health and Rehabilitation Sciences

Collaborative Specialization

Musculoskeletal Health Research


Battié, Michele C.


Objective: To synthesize current knowledge on the association of endplate structural defects (EPSD) with back pain (BP), improve EPSD measurement, and investigate EPSD prevalence, distribution, and association with age and body mass index (BMI).

Methods: In study 1, a systematic review was conducted on five databases for studies reporting on the association between EPSD and BP. Studies 2 and 3 used CTs and mCTs of 19 embalmed cadavers to examine the diagnostic accuracy of common EPSD assessment methods, and to develop and validate a novel method. Study 4 used the novel method on 200 adult males’ MRI to estimate EPSD prevalence, distribution, and association with age and BMI.

Results: Data from the 26 studies (11,027 subjects) on the association of EPSD with BP included in the systematic review were not pooled due to heterogeneity (I2=73%) relating to measurements and nomenclature, except for erosion, sclerosis and Schmorl’s nodes (OR:1.53-1326). The common EPSD assessment methods had a sensitivity of 70.9%-79.5% and specificity of 57.5%-79.1% and certain phenotypes were absent or misclassified (e.g., wavy/irregular and erosion). A novel method was therefore developed, consisting of definitions and atlases of six EPSD phenotypes with good inter-rater reliability for EPSD presence (K=0.65-0.68) and improved sensitivity (71.0%-79.0%) and specificity (77.0%-87.0%). Inter-rater reliability for specific phenotypes was fair (K=0.52-0.55). Using the novel method, there was a high prevalence of EPSDs (45.6%), with erosion (17.6%) being the most common phenotype. EPSD occurred more in the upper lumbar regions (c2= 41.68) and on the caudal endplates (c2=9.28) and were associated with greater age (OR:1.02, 95%CI:1.01-1.03) but not BMI (OR:1.00, 95%CI:0.98-1.03). Furthermore, age was associated with focal defects (OR:1.02, 95%CI:1.00-1.03) and erosion (OR:1.03, 95%CI: 1.01-1.04), while BMI was only associated with corner defects (OR:1.15, 95%CI:1.03-1.30).

Conclusion: The lack of standardized methods has impeded the understanding of EPSD and BP. Using a novel standardized assessment method, a developmental origin of Schmorl’s nodes is supported, and focal and erosive defects appear degenerative in nature, while corner defects appear to have a biomechanical origin. The project has opened a new avenue for measurement and further understanding of EPSD and their etiology and clinical significance.

Summary for Lay Audience

Recent advances in medical imaging have allowed a clearer view of the endplate, a thin bony structure with rich blood and nerve supply, that is located at the interface between intervertebral discs and vertebrae of the spine. The endplate is now attracting attention as a possible source or contributor to back pain. Endplate structural defects (EPSD) are common, yet little is known of their causes and clinical consequences, and related measurements are varied and underdeveloped. Toward advancing the understanding of EPSD to help resolve questions about their occurrence, causes and clinical significance, we first investigated the importance of EPSD to back pain. We found that they are associated with pain, but problems due to underdeveloped methods of assessment make this finding uncertain, particularly for various types of EPSD. We, therefore, assessed the reliability and validity of common endplate assessment methods, which led to the development of an improved EPSD assessment method. The newly developed EPSD assessment method and data from a large population-based study of adult males (n=200) were used to assess EPSD prevalence, distribution and association with age and body mass index. We found that EPSD were very common in the adult male population, but occurrence varied by type, with erosion being the most common EPSD phenotype. EPSD occurred more at the upper lumbar levels and on the endplate below rather than above the intervertebral disc. EPSD were related to older age but not body mass index. However, findings differed by type of EPSD. Schmorl’s nodes were not related to aging and may be developmental, while focal and erosive defects were related to older age, and corner defects to the body mass index. The project has provided a means to better identify EPSD to advance research on their prevalence, causes and consequences.

Available for download on Sunday, March 03, 2024