Specimen-specific fracture risk curves of lumbar vertebrae under dynamic axial compression

Abstract

Underbody blast attacks of military vehicles by improvised explosives have resulted in high incidence of lumbar spine fractures below the thorocolumbar junction in military combatants. Fracture risk curves related to vertical loading at individual lumbar spinal levels can be used to assess the protective ability of new injury mitigation equipment. The objectives of this study were to derive fracture risk curves for the lumbar spine under high rate compression and identify how specimen-specific attributes and lumbar spinal level may influence fracture risk. In this study, we tested a sample of three-vertebra specimens encompassing all spinal levels between T12 to S1 in high-rate axial compression. Each specimen was tested with a non-injurious load, followed by a compressive force sufficient to induce vertebral body fracture. During testing, bone fracture was identified using measurements from acoustic emission sensors and changes in load cell readings. Following testing, the fractures were assessed using computed tomographic (CT) imaging. The CT images showed isolated fractures of trabecular bone, or fractures involving both cortical and trabecular bone. Results from the compressive force measurements in conjunction with a survival analysis demonstrated that the compressive force corresponding to fracture increased inferiorly as a function of lumbar spinal level. The axial rigidity (EA) measured at the mid-plane of the centre vertebra or the volumetric bone mineral density (vBMD) of the vertebral body trabecular bone most greatly influenced fracture risk. By including these covariates in the fracture risk curves, no other variables significantly affected fracture risk, including the lumbar spinal level. The fracture risk curves presented in this study may be used to assess the risk of injury at individual lumbar vertebra when exposed to dynamic axial compression.

Introduction

In contemporary armed conflicts, underbody blast (UBB) attacks on military vehicles by improvised explosive devices (IEDs) have presented a considerable threat due to the large forces imparted upon the vehicle occupants. Because the lumbar spine provides the primary load path from the pelvis and lower limbs to the torso, 35% of soldiers wounded-in-action from UBB sustain injuries at the lumbar spine, mostly in the form of a fractures or dislocation (Vasquez et al., 2018). Due to the proximity between the lumbar spine and nervous system, neurological deficits are commonly experienced due to these fractures, where 42.6% to more than 50% of individuals experience neurological deficits at the time of injury (Dai, 2002; Lehman et al., 2012), with ongoing pain observed in 44% of cases (Dai, 2002). Neurological trauma possess an additional concern for soldiers in military scenarios, as it may lead to a loss in mobility that limits their ability to egress from the vehicle. With the ongoing threat posed by UBB attacks on allied military vehicles, improved blast mitigation systems are needed to protect occupants against lumbar spine related injuries.

To assess the protective capability of a new military vehicle, experimental tests mimicking a UBB scenario have been performed using an anthropomorphic test dummy (ATD), such as the Hybrid III ATD (Humanetics Innovative Solutions, Farmington Hills, USA) (Backaitis and Mertz, 1993). The ATD provides force and acceleration measurements at different body regions, from which injury risk may be calculated using cadaveric or operational tolerance data in the form of injury criteria or human injury probability curves. The most common injury criterion used to assess spinal injury risk is the dynamic response index (DRI) (Anton, 1991; DeStefano, 1972). However, the DRI was originally developed for aircraft pilot seat ejection, which comprise slower loading rates than UBB scenarios (Bailey et al., 2015; Pintar et al., 2012) and do not accurately model interactions with the torso or the seat. Human injury probability curves that are specifically developed for dynamic vertical loading will address this issue and likely enhance injury risk assessment for new military vehicles.

In recent research, human injury probability curves were developed for the lumbar column under vertical loading representative of UBB (Stemper et al., 2017; Yoganandan et al, 2018, 2020a, 2020b). These studies provide novel fracture tolerance data for high-rate vertical loading of the lumbar spine and indicate specimen-specific differences in injury risk were dependent upon volumetric bone mineral density (vBMD) between L2 and L4. At present, however, it is unknown whether fractures sustained by vertebrae at different lumbar spinal levels follow an identical human injury probability curve to that generated for compression of an entire lumbar column. For example, level-specific differences in vertebral morphology or rigidity may cause the distribution of fractures to vary along the spine, as observed previously (Stemper et al., 2017). A greater understanding of variations in fracture risk along the spine will help focus development of injury mitigation strategies to specific body regions that are more susceptible to injury, particularly those at greater risk of neurological deficits related to traumatic spinal cord injury.

The aim of this study was to develop human injury probability curves for fracture at the different lumbar spinal levels under dynamic axial compression. It is known that the fracture tolerance of bone is influenced by subject- or site-specific differences due to their influence on bone material properties and morphology (Crawford et al., 2003; Eckstein et al., 2002; Schileo et al., 2008); however, these subject- and site-specific dependences have not been previously evaluated in conjunction with fracture probability curves of the spine (Stemper et al., 2017; Yoganandan et al., 2018). By including specimen-specific properties (i.e., geometry and material properties) and spinal level as covariates in the fracture risk model, it was hypothesised that fracture risk predictions with a high degree of confidence may be obtained for individual lumbar spine vertebra; thus, providing more detail compared to previous lumbar spine fracture risk curves that only predict injuries to this body region in its entirety (Stemper et al., 2017; Yoganandan et al, 2018, 2020a, 2020b).