There is an increased need to develop female-specific injury criteria and anthropomorphic test devices (dummies) for military and automotive environments, especially as women take occupational roles traditionally reserved for men. Although some exhaustive reviews on the biomechanics and injuries of the human spine have appeared in clinical and bioengineering literatures, focus has been largely ignored on the difference between male and female cervical spine responses and characteristics. Current neck injury criteria for automotive dummies for assessing crashworthiness and occupant safety are obtained from animal and human cadaver experiments, computational modeling, and human volunteer studies. They are also used in the military. Since the average human female spines are smaller than average male spines, metrics specific to the female population may be derived using simple geometric scaling, based on the assumption that male and female spines are geometrically scalable. However, as described in this technical brief, studies have shown that the biomechanical responses between males and females do not obey strict geometric similitude. Anatomical differences in terms of the structural component geometry are also different between the two cervical spines. Postural, physiological, and motion responses under automotive scenarios are also different. This technical brief, focused on such nonuniform differences, underscores the need to conduct female spine-specific evaluations/experiments to derive injury criteria for this important group of the population.
Introduction
Numerous biomechanical studies have been reported using the human cadaver cervical spines [1–5]. Subfailure (incremental/stepwise and quasistatic) and failure (static and dynamic) loads have been applied to postmortem human subject (PMHS) specimen models [6–21]. Male and female volunteer studies have been conducted to determine their geometry and bone density [22–29]. Anatomical studies have included the documentation of injuries and determination of morphologies [30–34]. Together, these studies have enhanced the understanding of geometrical, structural, and mechanical responses of the cervical spine and its components for medical and bioengineering applications, and advance human safety in athletic, automotive, and military environments [1].
Injury criteria in the safety field are essentially scaled from one to the other to account for differences between male and female groups of the population [35,36]. The scaling process is based on pure geometrical considerations with the principal assumption that the female cervical spine is a geometrically scaled-down version of the male spine. The neck circumference is used to scale forces, moments, and interaction-based metrics to obtain women-specific injury criteria. Since the average female neck circumference is smaller than the average male, this type of scaling predicts lower loads in the average women than the average man. However, the relationship is unknown between the neck circumference and anatomically based osteo-ligamentous support, and/or, potential fundamental biomechanical material response difference between men and women spines. The validity of the pure geometric scaling assumption and its application to women-specific injury criteria have not been adequately addressed. This technical brief, through a review and synthesis of some literatures, examines differences between PMHS and volunteer biomechanical responses, and the facet, vertebra- and disk-related geometrical characteristics between male and female cervical spines. These are organized into Biomechanical Studies and Anatomical Studies in the following paragraphs.
Biomechanical Studies
Functional Unit Responses Under Stepwise and Failure Loading.
Nondestructive pure moment flexion and extension tests were conducted using the same experimental design to compare male and female responses of PMHS upper and lower cervical spine segments [10,14]. The total moment of 3.5 N·m in each loading mode was applied in steps of 0.5 N·m. They were followed by the failure tests in flexion or extension, applied at a rate of 90 N·m/s, and the failure was defined as decrease in the measured sagittal moment concomitant with increasing segmental angular rotation. The mean ages of the female and male specimens were in the fifth and sixth decades (female: 50.8 + 8.8 yr, range: 33–66 yr, and male: 66 + 7.2 yr, range: 51–74 yr) with sample sizes of 52 and 41 segments. For the upper cervical region, tests were conducted using female and male specimens from OC–C2 and OC–C3 segments. For the lower spine, they were conducted with female specimens using C3–C4, C5–C6, and C7–T1 segments, and with male specimens using C4–C5 and C6–C7 segments. Failure bending moments in extension and flexion were lower in females than males for the upper cervical segments. Statistically significant differences were not apparent (p > 0.05) in the failure moment and angulation when all subaxial segmental responses were grouped. However, the range of motions of male OC–C2 segments was significantly less than female segments (p = 0.011). The total range of motion over the applied load range of ± 3.5 N·m was greater in females in both regions. These segment-based data demonstrated significant differences in response to the same loading in some biomechanical metrics between male and female spines.
Head–Neck Complex Responses Under Inertial Loading.
Dynamic tests simulating the responses of the cervical spine from rear-end motor vehicle impacts also showed differences in kinematics between PMHS male and female specimens [37,38]. These studies used the entire human head–neck complex, including the effects of passive musculature, and applied inertial loads at increasing velocities. The data were obtained from nine specimens (four males and five females) with similar mean ages: 62 ± 15 yr for male and 55 ± 17 yr for female specimens, approximately representing similar structural morphology for both sexes. The lower neck shear force was approximately twice greater in male specimens (384 N versus 636 N), contributing to the kinetic aspects of the extension-induced dynamic loading rate effects on PMHS head–neck complexes. The rostral segment at the C2–C3 level responded with flexion, while the caudal segments sustained extension regardless of change in velocity. C2–C7 segmental motions were greater (1.2–2.6 times, p = 0.002–0.049) in female than male head–neck complexes at all subaxial levels with the exception of the midcervical spinal region, i.e., at the C3–C4 level [39]. In addition to these differences in segmental kinematics, facet joint motions in the ventral and dorsal regions were significantly (p = 0.004 and p = 0.011) dependent on sex [38]. Thus, both sagittal plane kinematics and load-related responses of head–neck complexes demonstrated sex specificity under inertial loading to intact PMHS cervical spines.
Kinematic Responses From Volunteer Studies.
Acknowledging that it is not ethical to conduct human volunteer studies by applying insults into the injury producing range, cervical spine kinematic responses were found to be different between men and women under static and inertial rear impact loading. Spinal range of motions in female subjects were greater than male subjects [40–43]. In recent studies, kinematics were analyzed from three different test series: (a) 12 men and eight women exposed to sled velocities of 2.3 m/s and 2.8 m/s, (b) four men and two women exposed to 1.6 m/s, and (c) four women and nine men subjected to voluntary quasi-static head–neck bending [44–46]. Analysis of sequential X-ray images from the first two series indicated that women responded with a peak flexion of the head relative to T1 in the S-shape phase, while the head-T1 was in extension at the time of peak head flexion. In contrast, men responded with flexion in both phases. Women had larger flexion in upper and larger extension in lower segments. Analysis of data from the latter two series indicated that, in the peak S-shape phase, women responded with greater extension angulations from C4–C7 than the maximum voluntary retraction, C6–C7 rearward displacements exceeded maximum voluntary extension, and C5–C6 rotation was the greatest during peak extension, and it exceeded the voluntary extension, while C5–6 and C6–7 segments in both male and female subjects showed larger rearward displacements during the extension phase than the maximum voluntary extension. Other volunteer studies have reported significantly greater (p < 0.05) head–neck kinematics in female than male subjects, based on the data obtained from accelerometers and high-speed video images [47,48]. In addition, the duration of injury symptoms experienced by women was also significantly (p < 0.05) longer than men. A 2016 study reported spinal alignment differences in a simulated automotive-seated posture using three male and five female human volunteers using an upright magnetic resonance imaging system [49]. One male and four female cervical spines were less lordotic, while two male and one female spines were lordotic. The most recent study by the same group of researchers investigated spinal alignment patterns of eight females and seven males in a seated posture using multidimensional scaling [50]. These two studies reported that female subjects are more likely to have nonlordotic curvature than male subjects. As pre-alignment affects injuries and injury mechanisms due to contact or head impact loading, changes in female subject posture lead to different transmission of the external loading along the cervical column compared to male subjects [51]. Thus, volunteer studies have shown the existence of kinematic, physiological, and postural differences between men and women for automotive applications.
Anatomical Studies
Facet Joint‐Related Geometry.
Studies included the determination of the posterior column morphology from human cadaver cervical spines [52]. Six specimens (three males, mean age 60 yr, and three females, mean age 76 yr) were used to quantify the anatomy of facet capsules from the occiput to T1 levels using cryomicrotomy techniques [32]. The total gap in the upper spine was lower (p < 0.0001) than the gap in the lower spine. The gaps at the ventral and dorsal regions were lower in the upper spine than lower spine (p < 0.0001 and p = 0.0004). Further, the gap in the dorsal region for males was lower (p = 0.0523, reaching significance) than the gap for females. The overall mean facet cartilage thickness was lesser (p = 0.0111) in females than males in the upper spine and lower spine (p = 0.0077). This change in the bony, cartilage and joint anatomy offered an explanation to the increased susceptibility of neck pain in women in rear impacts, especially emanating from the posterior facet joint complex. The authors also supplemented clinical studies of facet joint involvement [23,53]. Another study analyzed facet joint angles from 250 males and 173 females using computed tomography (CT) scans of 18+ yr-old subjects by grouping the data into 18–29, 30–44, 45–59 and 60+ yr bins [26]. The C2 body depths were larger in men than women in the corresponding age groups (p < 0.0001). Men had larger facet angles at C2, C3, and C6 than women (p < 0.01). The increase from the 18–29 age group to the 60+age group was greater in men. Retroversion angles were larger in men at the C2–C3 (p < 0.01) and smaller at the C5–C6 (p < 0.01) levels. These data show sex differences at different intervertebral levels across the osteo-ligamentous cervical column, suggesting the importance of including local anatomical features for scaling relationships between adult men and women.
Vertebra‐ and Disk‐Related Geometries.
Lateral radiographs were analyzed for vertebral body geometry from 30 male and 31 female volunteers with ages ranging from 18 to 24 yr [24]. Analysis was based on the stature and sex on the height, depth, and area of C3–C7 bodies. Stature-based differences in these parameters were not statistically different. However, C3, C6, and C7 depths and areas, C4 depth, and C7 height in women were smaller than corresponding dimensions in men (p < 0.05). For males, the Ponderal index for C7 height and depth; and head weight for C3, C5, and C6 heights and depths, and C4 height; and for females, the Ponderal index for C7 height and head weight for C6 depth reached significance (p < 0.05) with other parameters showing no significant difference. This demonstrates that cervical spine vertebral geometry is not uniformly distributed between men and women at all segmental levels of the subaxial spine. The authors suggested the need for future work to determine sex dependence in the vertebral body size of similar stature between the two populations. A study extracting the cervical spine geometry using lateral X-rays from 69 male and 51 female subjects (20–80 yr) showed that C2–C3 and C6–C7 disk heights were lower (p < 0.05) in women than men, regardless of age [25]. Another study measured vertebral height and postero–anterior displacement of C1/C2 using radiographs from 35 male (mean 32.2 ± 10 yr) and 100 female (mean 32.4 ± 12 yr) subjects [54]. The height and displacement significantly (p < 0.05) decreased and increased, respectively, with age in female compared to male subjects [54].
A more recent study investigated differences in geometries of vertebrae between men and women by controlling for linear body measurements using quantitated CT scans [28,29,55]. From an overall set of 130 healthy volunteers ranging from 18 to 40 yr of age, independent size-matched subgroups were identified based on sitting height (87–90 cm) and head circumference (55–57 cm) [55]. Vertebral body depth and width, vertebral width, interfacet width, disk-facet depth, and C1–C5 column height data were analyzed. In addition, the support area was calculated. It was defined as one-half of the product of the interfacet width and disk-facet depth. Fourteen male and nine female volunteers were in the sitting height group, and 13 male and 15 female volunteers were in the head circumference subgroup. Sitting height and head circumference were not significantly different between men and women for the two respective subgroups. In the sitting height-matched subgroup, the vertebra width, disk-facet depth, and cervical column height were greater (p = 0.0010, p = 0.0329 and p = 0.0238) in men than women, while other metrics did not show significant sex-based differences. However, in the head circumference-matched subgroup, all the vertebral dimensions were greater (p ranging from 0.0012 to 0.0236) by 7–18% in males. Thus, even when controlling for metrics such as stature, expressed as sitting height or head circumference, female cervical spines demonstrated differences at different vertebral levels for different physical parameters from male cervical spines. It is also noted that, real-world case analysis has shown that even with similar body sizes, females are more likely to sustain injuries from rear impacts [56].
Discussion
Although reviews on the biomechanics and injuries of the spine have appeared in clinical and bioengineering literatures for more than three decades, focus on the differences between male and female cervical spine responses and geometry have largely been ignored [3,4,57–61]. Clinicians from operative perspectives generally do not distinguish between male and female spines, and fixation devices including cervical arthroplasty are not sex specific [1]. Current automotive dummy designs are also not female specific [62]. Current injury criteria, including those for the neck, in the automotive field and other environments evaluated using different types of dummies, are based on the geometric similitude and dimensional analysis [36,63,64]. The methodology requires a linear parameter for scaling metrics from the male to the female population. Scaling was needed because of the lack of demographics-based experimental tolerance data [65]. The data for age groups and sex were not available during the development of Federal Motor Vehicle Safety Standards, FMVSS-208, in the late 1990 s [66]. This continues to be true although the research has been done since 1998, the year in which the neck injury criteria and scaling methods were announced in the Notice for Proposed Rule Making by the US National Highway Safety Administration. At the time of the notice and adopted later in the present federal automotive standards, the mean neck circumference of the two populations was chosen, because it was a simpler measurement to record.
As described in this technical brief, static biomechanical responses at the segmental level, between upper and lower cervical spines show sex differences, some reaching statistical significance. Segmental angulations under dynamic (inertial) loading are also significantly different between level and sex. Cervical spine curvature representing postural effects are more lordotic in men than women for automotive applications. Human cervical spine vertebral geometry is nonuniformly distributed between females and males at all segments. Posterior component geometry including soft tissues also have sex bias. While the discussed literatures have shown statistical significances, it is worth noting that, albeit somewhat speculative, other studies may be used to draw similar inferences with the acknowledgment that they were not designed to investigate either the role of sex or the ability of the sample size to provide statistically significant biomechanical results. A case in point is experiments conducted in the late 1980–1990 s wherein injuries due to contact loadings were found to be concentrated at one vertebral level in males (n = 4) in contrast to multiple level and hard and soft tissue component involvements in female cervical spines (n = 5) spines [17]. Thus, from biomechanical loading, anatomical, and geometrical considerations, the human cervical spine shows sex differences. The segmented nature coupled with the heterogeneous components of the spine should be given consideration while establishing injury criteria for females. The simple one-dimensional neck circumference-based scaling process, adopted in the current specification of automotive neck injury criteria, ignores these nonuniform, local, regional, and overall differences between male and female cervical spine anatomy and response characteristics. An improved process would be to accommodate such differences, and this technical brief alerts the military, automotive, and other safety engineering communities to design controlled studies to achieve this goal. To account for these multifactorial differences, a need exists for conducting tests using female spines and use those data in establishing female-specific injury criteria. This process eliminates the use of results from the male population for conversion and injury criteria specification to the female population.
The early development of the Toyota Total Human Model for Safety (THUMS) small female model was done by scaling the midsize male version [67,68]. However, CT scans from a 38-yr old female with a stature of 154 cm and total body mass of 52 kg was used to create a newer version 4.0 [69]. Analysis under rear impacts indicated that differences in the peak capsular strains between female and male models were due to their “anatomical difference in terms of the neck stiffness.” While the female model responded with S-shape, the male model was in extension. The Global Human Body Modeling Consortium (GHBMC) reported on the development of the small-size female model by using CT and magnetic resonance images from a 24-yr old female, with a stature and total body mass of 149.9 cm and 48.1 kg, respectively [70]. Full features of the neck representing the female population are not completely mimicked in these single subject-based THUMS and GHBMC models. Detailed female-specific finite element models based on female anatomy, geometry, and female-spine tested experiments for calibration/verification and validation purposes are also necessary to develop computational models. As stated, conducting female-specific spine experimental evaluations in this area are needed first to achieve these goals.
Acknowledgment
This research was supported in part by Department of Veterans Affairs Medical Research, Medical College of Wisconsin, Office of the Assistant Secretary of Defense for Health Affairs, through the Broad Agency Announcement under Award No. W81XWH-16-1-0010, Cooperative Agreement W81XWH-12-0041, and the Biomechanics Product Team led by the Johns Hopkins Applied Physics Laboratory for the WIAMan Project under contract number N00024-13-D-6400. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the Department of Defense. NY and FAP are part-time employees of the Zablocki VA Medical Center, Milwaukee, Wisconsin.