Over 52 million people experience low back pain annually, which negatively affects quality of life (American Academy of Orthopaedic Surgeons, 2008). Vertebral fractures can cause low back pain;​ in particular, ring apophysis fractures (RAFs) occur in young, healthy adolescents (typically athletes) and osteoporotic fractures occur in the elderly population. These fractures often occur without memorable trauma which suggests that these fractures form during everyday movements and activities of daily living (ADLs). These fractures are underdiagnosed because they are not suspected and not visible on a plain x-ray. Therefore, more research is required to determine how these fractures form and how they can be best detected. The purpose of this work was to establish an ex vivo model of nontraumatic vertebral fracture and demonstrate that these fractures can be initiated by repetitive low levels of loading.

Fifteen cervine and five cadaver lumbar 5-vertebra motion segments were tested. Each specimen was cyclically loaded in flexion, similar to the small movements that occur during ADLs. These experiments consisted of pinned-end, low-load, low-angle, eccentric cyclic flexion applied to each motion segment;​ selected specimens were then monotonically compressed. Biomechanical and radiographic methods were used to describe the changes in behavior during loading, including possible fracture development. Imaging included both plain x-ray and µCT scans. Biomechanical methods included analysis of load-displacement hysteresis loops, intervertebral angles (IVAs) and surface strain from 3D digital image correlation (3D-DIC).

The x-rays of specimens without monotonic compression showed no evidence of fracture after cyclic loading. MicroCT scans showed evidence of RAFs in all specimens which were also monotonically compressed. Analysis of the mechanical behavior of the motion segment during loading included computation of the hysteresis loop area, neutral zone length, and neutral zone slope using the load-displacement data. There were no significant differences between species or cycling frequency after the first 1,005 cycles of loading. Motion analysis permitted the computation of IVAs and their changes throughout the cyclic loading. During cyclic loading the IVAs changed noticeably, indicating that the motion segment had different kinematics compared to the start of the test. Finally, the surface strain on the vertebral bodies was quantified using 3D-DIC which showed areas of localized compression, indicating the region that a fracture may have occurred.

This research has developed a robust ex vivo model to simulate repetitive flexion/extension in lumbar 5-vertebra motion segments. Our results show that the biomechanical evidence of fracture precedes radiographic evidence, which agrees with clinical speculation that nontraumatic fractures are usually invisible on plain x-rays. Young cadaver specimens are thankfully rare, thus having a cost-effective proxy to young human lumbar spines is imperative for studying fracture formation in this population. By understanding the mechanics behind these fractures, this work will ultimately enable clinicians to better understand how these fractures develop, may inspire improved diagnostic techniques, and will ultimately prevent chronic low back pain secondary to nontraumatic vertebral fracture.