Follower Load Influences Cervical Spine Kinematics and Kinetics during Simulated Head First Impact: an Ex Vivo Study

FIRST IMPACT: AN EX VIVO STUDY Christopher R. Dennison, Amy Saari, Qingan Zhu , Tim Nelson, , Philip Morley, Eyal Itshayek, Tom Oxland and Peter A. Cripton Orthopaedic and Injury Biomechanics Group; Department of Mechanical Engineering Department of Orthopaedics; International Collaboration on Repair Discoveries, University of British Columbia, Vancouver, Canada; Department of Neurosurgery. Hadassah Hebrew University Hospital. Jerusalem. Israel INTRODUCTION Cervical spine osteoligamentous and cord injuries result in significant financial cost to the injured and society. Spinal cord injuries can lead to severe disability and fatality. Understanding of the biomechanics of these injuries is based on decades of epidemiology, clinical studies, computational models, ex vivo experimental studies, and combinations thereof, that consider axial compression from both head-first impact and quasi-static loading. Early biomechanical studies used whole cadavers but lacked measurement of cervical kinetics and kinematics because the neck was surrounded by musculature. More recent studies used isolated cervical spine complexes that allow measurement of cervical kinetics and kinematics because musculature is removed. Most studies describe cables tethered to the head to control neck posture, prior to impact. However, these cables do not recreate lines of action of loads that neck musculature supplies in vivo and that has been shown, through recent modeling, to significantly increase injury risk during impact [1]. For quasi-static loading, ex vivo experimental work has shown that follower loads (guided at each level to simulate musculature) lead to cervical mechanics that match in vivo [2] and increase the load carrying capacity of the spine prior to buckling under axial compression [3]. We hypothesized that simulate approximate in vivo compression using follower load in ex vivo biomechanics experiments will improve the biofidelity of our tests. The objective of this work was to compare kinematics and kinetics, during head-first impact, of isolated cervical spine complexes instrumented with follower load to those without follower load. METHODS We simulated head-first impact using a drop-tower and cadaveric cervical spines. Six segments (Occ.-T1) were instrumented with follower load (FL group, mean age 71 yrs.) applied via tensioned bilateral cables passing through the flexion-extension centre of rotation at each vertebral level [2], while six were not (NFL group, mean age 78 yrs.). Occiputs were fixed to a biofidelic surrogate head and T1 was encased in a mounting cup. Vertebral motions (kinematics) and compressive head force and neck forces and moments at T1 (kinetics) were measured. Vertebral motions were calculated by tracking the motions of vertebra-fixed photo-reflective markers with high-speed video cameras (Phantom V9.0, Vision Research Inc., NJ) and using photogrammetry to calculate three-dimensional motion from camera images. Head forces were measured with an axial load cell (LC, Omega Engineering Inc., CT) while T1 forces and moments were measured with a six-axis load cell (MC3A, AMTI, MA). The kinematic and kinetic metrics reported are: time from head impact to injury, Tinj; time from head impact to peak compressive neck force, Tpeak; impact speed, Vimpact; pre-injury peak compressive neck force, Fn; and mount-cup impulse from impact to injury, Iinj. Tinj was determined as the time at which compressive neck forces dropped to 2/3 of the peak while sagittal moment increased [4]. Wilcoxon Rank-sum tests were used to test for statistical significance. RESULTS The mean Vimpact was 2.8 m/s in the NFL group and 2.9 m/s in the FL group. There was no significant difference in Fn, between the NFL and FL groups (p = 0.24) which had means of 1.9 kN and 1.4 kN, respectively. There were significant differences between the NFL and FL groups for both Tpeak (p = 0.004) and Tinj (p = 0.009). The mean Tpeak for the NFL group was 4.2 msec and for the FL group was 1.6 msec. The mean Tinj for the NFL group was 4.9 msec and for the FL group was 2.5. There was no significant difference in Iinj between the NFL group and FL group (p = 0.200), which had means of 3.9 Nsec and 2.1 Nsec, respectively. Motions of photo-reflective markers (Fig.1) indicated that overall hyperextension was experienced in all specimens (both FL and NFL group).