Introduction

Soft-tissue injuries have been reported in 23–70% of individuals involved in frontal impacts [1, 28, 34, 35, 54]. Despite this high injury prevalence, the underlying soft-tissue injury mechanisms are not fully understood [55]. The majority of previous frontal impact studies have investigated the response of volunteers [13, 17–20, 27, 36, 40, 57, 59, 60] and cadavers [9, 31, 39, 61] to simulated frontal impacts at accelerations up to 16 g. Among soft-tissue injuries, intervertebral disc injury is often cited as a source of chronic pain [29], via mechanisms of delayed healing or altered loading patterns causing accelerated facet-joint degeneration [8, 25, 37, 56]. Previous autopsy studies have documented disc herniation [28], annulus fiber ruptures [12, 28], and cartilage endplate separations [56] due to frontal impacts. Frontal impacts may cause subfailure injuries to the soft tissues of the cervical spine, leading to clinical instability and chronic pain [22].

In addition to clinical investigations, biomechanical studies have demonstrated injuries to the disc due to simulated frontal impact. Using a human cervicothoracic model at accelerations up to 15.6 g, Clemens et al. [9] observed fissures and avulsions of the intervertebral disc in 80% of the trials, most commonly at C5–C6. Crowell et al. [11] applied flexion-compression loading to three-vertebra human cervical spine segments and reported macroscopic annular tears and disc herniations; the disc was the most commonly injured soft-tissue component. Nonetheless, the disc strains were not calculated in either of the previous studies.

Despite previous evidence that has demonstrated disc injury due to actual and simulated frontal impacts, and various studies investigating injury mechanisms due to rear impact [46–48], the injury mechanisms of frontal impact remain poorly understood [55]. Continued biomechanical research may aid in the development of methods to detect, treat and prevent intervertebral disc injuries. Therefore, the goals of this study were to use the whole human cervical spine model with muscle force replication to: (1) determine the dynamic disc annular tissue strain and disc shear strain during simulated frontal impacts, (2) compare these values with limits measured during physiological loading, and (3) compare the present frontal impact data with disc strain data from a previously reported rear impact study.

Materials and methods

Specimen preparation

Six fresh-frozen human osteoligamentous whole cervical spine specimens were mounted in resin (Fibre Glass-Evercoat, Cincinnati, OH) at the occiput and T1, according to a predefined neutral posture, such that a line from the top of the dens to the lowest point on the posterior occiput was parallel to the occipital mount, while the T1 vertebra was tilted anteriorly by 24° [6]. The average specimen age was 71.3 years (range: 54–87 years), and there were three male and three female donors. The donors did not suffer from any disease that could have affected the osteoligamentous structures. To attach lightweight motion-tracking flags, a headless wood screw was drilled into the anterior aspects of C1 and C2 and into the lamina of C3 through C7. The flags consisted of 3 mm-diameter, hollow brass tubes, with two white, spherical, radio-opaque markers. A flag was fitted rigidly onto each wood screw, and additional flags were attached to the occipital and T1 mounts. This constituted the whole cervical spine (WCS) model prepared for intact flexibility testing.

To prepare a specimen for frontal impact simulation, a surrogate head (mass 3.3 kg and sagittal plane moment of inertia 0.016 kg m2) was attached to the occipital mount. The head and spine were stabilized using the compressive muscle force replication (MFR) system (Fig. 1). The MFR system was symmetric about the mid-sagittal plane and consisted of anterior, posterior and lateral cables attached to preloaded springs anchored to the base. The stiffness coefficients of the anterior, lateral and posterior tension springs were 4.0 N/mm, 4.0 N/mm and 8.0 N/mm, respectively. Two anterior cables, each consisting of two strands, originated at the occipital mount, ran through separate guideposts at C4, through pulleys within the T1 mount and were each connected to a separate spring. The preload in each anterior spring was 15 N. To apply the posterior MFR, small holes were drilled bilaterally into each lamina (C3 through C7) and wire loops were inserted into the spinal canal through the laminae and tightly secured above each vertebral spinous process. The two posterior MFR cables originated from the occipital mount, ran through the wire loops, through a pulley at the T1 mount and each were connected to a spring, preloaded to 15 N. The lateral cables originated from C0, C2, C4, and C6, passed alternately along lateral guide rods, and were each connected to a preloaded 30 N spring. The lateral guide rods were inserted in the frontal plane into the vertebral bodies of C2 through C7, such that the guide rods at C3 through C7 were positioned at the approximate centers of rotation of C2–C3 through C6–C7, respectively. With this MFR arrangement the compressive preloads at each intervertebral level were: 120 N (C0–C1, C1–C2); 180 N (C2–C3, C3–C4); 240 N (C4–C5, C5–C6) and 300 N (C6–C7, C7–T1). A C0–C2 flexion limiter was secured to the occipital mount and to the C2 spinous process that allowed approximately 30° of sagittal rotation, consistent with the in vivo rotation of the normal cervical spine [16, 41]. This constituted the WCS+MFR model, ready for simulated frontal impact (Fig. 1). A lateral X-ray of the WCS+MFR model in the neutral posture, together with radio-opaque calibration markers positioned in the midsagittal plane, was taken and digitally scanned (Adobe Photoshop version 6.01, San Jose, CA).

Fig. 1
figure 1

The biofidelic whole cervical spine model (occiput–T1 vertebra) with muscle force replication (WCS+MFR). The surrogate head was attached to the occiput. Anterior, posterior and bilateral MFR cables stabilized the head. The extension head stop is not shown. For more details, please see text

Full size image

Intact flexibility testing

The intact WCS model underwent standard flexibility testing to determine the static physiological disc deformations [44]. Pure non-constraining flexion and extension moments up to a maximum of 1.5 Nm were applied to the occipital mount in four equal steps. Kinematic data were recorded on the third loading cycle, following two preconditioning cycles using a high-speed digital camera (MotionPro, Redlake, MASD, San Diego, CA).

Frontal impact simulation and monitoring

Frontal impact simulation was performed using a previously developed bench-top sled apparatus [45]. The incremental trauma protocol was used to apply frontal impact loading to the WCS+MFR model at nominal T1 maximum horizontal accelerations of 4 g, 6 g, 8 g and 10 g (1 g = 9.81 m/s2). A head stop was used to restrict head extension during the rebound phase to within physiologic limits and to prevent extension injuries. The high-speed digital camera recorded the spinal motions at 500 frames/s.

Intervertebral disc geometry and deformation

Endplate coordinate system

The anterosuperior (point 1) and posterosuperior (point 4) corners of the lower vertebral body of each functional spinal unit from C2–C3 to C6–C7 were selected on the lateral X-ray (Fig. 2). Two additional points (points 2 and 3) were selected, such that the point 2 was midway between points 1 and 4, and point 3 was midway between points 2 and 4. The endplate coordinate system y-z had its origin at point 4, its positive z-axis was oriented anteriorly through point 1, and its positive y-axis was orthogonal to the z-axis and oriented superiorly.

Fig. 2
figure 2

Functional spinal unit with motion tracking flags and the points on the endplates used to calculate intervertebral disc deformation. Two coordinate systems are also shown. The endplate coordinate system y-z was fixed to the lower vertebral body. The anterior (2–A) and middle (3–B) annular tissues are shown by dashed lines, while the posterior annular tissue (4–C) is shown by a solid line. The points (1′, 2′ and 4′) were selected on the inferior surface of the upper vertebral body that were directly above points 1, 2 and 4, respectively, and were used to calculate the disc shear strain at the anterior, middle and posterior disc regions. The ground coordinate system h-v defined flexion as positive rotation

Full size image

Disc annular tissue strain

The superficial anterior annular tissue of the cervical disc and the deepest alar tissues of the anterior longitudinal ligament have been described in three dimensions as running posteroinferiorly and laterally from the upper to the lower vertebral body [38], at approximately 30° to the z-axis. The thin, superficial posterior annular tissue has been described as having a longitudinal orientation [38]. These tissues were most likely to be loaded due to flexion during frontal impact simulations, and thus were mathematically constructed to develop a sagittal plane disc model based on the lateral X-ray. The anterior and middle annular tissues had origins at points 2 and 3, respectively, and insertions at the inferior surface of the upper vertebral body (points A and B, respectively), and were oriented at 30° to the z-axis (Fig. 2). The posterior annular tissue was oriented vertically and had its origin at point 4 and insertion at the posteroinferior corner of the upper vertebral body (point C).

Geometrical rigid body relationships between the flag markers and the annular tissue origins and insertions were digitized for each specimen from the scanned X-rays and were superimposed onto the first frame of the high-speed movie. Custom Matlab motion-tracking software, written with sub-pixel accuracy, computed the vertebral body rotations and flag marker translations at each subsequent frame in the ground coordinate system h-v (Fig. 2). These data, together with the geometrical rigid body relationships, were used to track the selected points and to calculate the translation of the annular tissue origins and insertions in the ground coordinate system h-v throughout frontal impact simulation. To compute the dynamic annular tissue strains, the change in tissue length was expressed as a percentage of the original tissue length in neutral posture. The physiological tissue strain was defined as the peak value obtained during the flexibility testing. The average error in the system used to calculate disc deformations was 0.3 mm (SD 0.2 mm), and, consequently, the average strain error was approximately 3.0% (SD 2.0%) [51].

Disc shear strain

Three additional points (1′, 2′ and 4′) were selected on the inferior surface of the upper vertebral body such that 1′, 2′ and 4′ were directly superior (positive y-coordinate) to points 1, 2 and 4, respectively (Fig. 2). These points, representative of the anterior, middle and posterior disc regions, were tracked throughout simulated frontal impact using the custom motion-tracking software and the rigid body relationships between the flag markers and the points. The points were used to calculate the disc shear strain (γ) at each of the three locations:

$$ \gamma = \arctan {\left( {\frac{{\Delta z}} {{y_{0} }}} \right)} $$
(1)

where Δz represented the z-axis translation of 1′, 2′ and 4′ relative to 1, 2 and 4, respectively. The original disc heights (y0) were obtained as the y-axis differences between 1′, 2′ and 4′ and 1, 2 and 4, on the neutral posture lateral X-ray. Corresponding physiological limits were defined as the maximum values obtained during the flexibility testing.

Data analyses

Disc strain data were low-pass digitally filtered at a 30 Hz cut-off frequency. For each frontal impact simulation, the peak anterior, middle and posterior disc annular tissue strains and disc shear strains were determined for each intervertebral level (C2–C3 to C6–C7) during the total intervertebral flexion time period (i.e., from the T1 horizontal acceleration onset to return to the neutral posture, after maximum intervertebral flexion). Single factor, repeated measures analysis of variance (ANOVA) (p<0.05) and Bonferroni post hoc tests (Minitab Release 13, State College, PA, USA) were used to compare peak annular tissue strain and disc shear strain during simulated frontal impact, with the corresponding physiological limits determined from the flexibility testing.

Results

The average measured peak T1 horizontal accelerations during the frontal impact simulations were 3.9 g, 5.7 g, 8.5 g, and 10.0 g (ΔV: 8.4 kph, 11.4 kph, 13.4 kph, and 13.8 kph), corresponding to the nominal maximum accelerations of 4 g, 6 g, 8 g, and 10 g, respectively. The average T1 horizontal acceleration pulse duration was 107.3 ms, and this compared favorably with values measured during real-life collisions, ranging between approximately 70 ms and 100 ms [5].

An example of disc strains during frontal impact

The disc kinematics varied among intervertebral levels and specimens. An example of disc strains as functions of time is provided by the C2–C3 disc of specimen No. 1 during the 6 g frontal impact (Fig. 3). Two phases, based on intervertebral rotation, were defined. Phase I began with the onset of T1 horizontal acceleration (time 0 ms) and concluded with peak intervertebral flexion, while phase II spanned from peak intervertebral flexion to the return of neutral posture. The anterior annular tissue strain and posterior disc shear strain increased and peaked in phase I, and returned towards zero in phase II. The posterior annular tissue strain peaked in phase II, following maximum intervertebral flexion.

Fig. 3
figure 3

An example of disc strains during frontal impact showing anterior and posterior disc annular tissue strain (2–A and 4–C, Fig. 2), posterior region disc shear strain, and intervertebral flexion. The results are of specimen No. 1 at level C2–C3 during the 6 g frontal impact

Full size image

Average behavior of six specimens

The dynamic disc annular tissue strains and disc shear strains are presented in Tables 1 and 2, where exact means of neutral tissue length, neutral disc height, and strain values with corresponding standard deviations are provided. The data are also presented as three-dimensional graphs together with the corresponding physiological limits. These provide an overview of their variance by anatomical location, intervertebral level and impact acceleration (Figs. 4 and 5).

Table 1 Peak disc annular tissue strains. Average (SD) neutral tissue lengths (mm) and peak annular tissue strains (%) during physiological flexion limits and simulated frontal impact at anterior (Ant), middle (Mid) and posterior (Post) regions of the disc, at intervertebral levels C2–C3 through C6–C7 (see Fig. 2). Significant increases (p<0.05) in the frontal impact strains over the corresponding physiological limits are indicated by *
Full size table
Table 2 Peak disc shear strains. Average (SD) neutral disc heights (mm) and peak disc shear strains (radians) during physiological flexion limits and simulated frontal impact at the anterior (Ant), middle (Mid) and posterior (Post) regions of the disc (see Fig. 2), at intervertebral levels C2–C3 through C6–C7. Significant increases (p<0.05) in the frontal impact strains over the corresponding physiological limits are indicated by *
Full size table
Fig. 4
figure 4

Average peak anterior (Ant), middle (Mid) and posterior (Post) disc annular tissue strains at C2–C3 through C6–C7 during simulated frontal impact (4 to 10 g) and physiological limits. Significant increases (p<0.05) in frontal impact strains over the corresponding physiological limits are indicated by *

Full size image
Fig. 5
figure 5

Average peak disc shear strains at the anterior (Ant), middle (Mid), and posterior (Post) disc regions of C2–C3 through C6–C7 during simulated frontal impact and physiological limits. Significant increases (p<0.05) in frontal impact strains over the corresponding physiological limits are indicated by *

Full size image

The C2–C3 middle disc annular tissue strain first exceeded physiological limits at 6 g and subsequently spread to all intervertebral levels at 10 g, excluding C4–C5 (Table 1, Fig. 4). At 10 g, anterior annular tissue strain exceeded the physiological limits throughout the cervical spine, with the exception of C3–C4 and C4–C5. The highest peaks for the anterior and middle annular tissue strains occurred at 10 g. These were 43.1% and 39.9% at C2–C3 and 29.6% and 28.6% at C5–C6, respectively. The posterior annular tissue strains beyond the corresponding physiological limits were observed at C3–C4 (116.5%) and C5–C6 (105.6%), both during the 10 g impact.

Disc shear strain at C2–C3 exceeded physiological limits beginning at 4 g (Table 2 and Fig. 5). At 10 g, increases in disc shear strain were observed at all intervertebral levels, excluding C4–C5. Of these increases, the greatest disc shear strain was most often noted in the posterior disc region, reaching 0.9 radians and 1.0 radians (125% and 150% increases from physiological limits) at C2–C3, during the 6 g and 8 g impacts, respectively.

Discussion

Injuries to the cervical disc sustained during automobile collisions may result in chronic pain [23, 24, 29], yet there are currently no biomechanical studies investigating disc injury mechanisms during simulated frontal impact. The present study determined the dynamic disc annular tissue strain and disc shear strain during simulated frontal impact, using a biofidelic whole cervical spine model with muscle force replication. At 4 g and 6 g, significant increases (p<0.05) over physiological limits were observed in the C2–C3 disc shear strains and middle disc annular tissue strains, respectively. The largest disc shear strain was observed in the posterior disc region, reaching a maximum of 1.0 radian (150% increase over the physiological limit) at C2–C3 during the 8 g impact. At 10 g, anterior disc annular tissue strains exceeded physiological limits at all intervertebral levels, excluding C3–C4 and C4–C5, while the posterior annular tissue strains exceeded physiological limits at C3–C4 and C5–C6. The disc annular tissue strain and disc shear strain results suggest that during frontal impact, intervertebral flexion combined with disc shearing may be the injury mechanism producing subfailure disc injuries. In support of these findings, previous clinical hypotheses have suggested that frontal impact would likely produce horizontal shear between cervical vertebrae, and result in excessive strain of the anterior disc annular tissues with compression of zygapophysial joints [2, 58].

There are several limitations of the present model that must be considered. Firstly, the T1 vertebra was fixed to the sled, which may have affected the cervical spine kinematics, especially at C7–T1. While no data exist that quantify the in vivo intervertebral rotations during simulated frontal impact, the peak head-T1 flexion of young volunteers with pre-tensed muscles ranged between 49.4° and 74.0° during simulated frontal collisions up to 15.6 g [57]. The average (SD) peak head-T1 flexion of the present study ranged from 43.1° (5.9°) to 84.6° (6.0°), during the 4 g and 10 g impacts, respectively, and exceeded the in vivo corridor, since the current model simulated the response of an unwarned subject. Secondly, the incremental trauma approach was used to determine the relationship between disc strain and impact acceleration. The many advantages of the incremental trauma protocol have been well documented [26, 48]. The incremental and single trauma protocols have been shown to produce equivalent soft-tissue injury severity due to frontal impacts of a porcine cervical spine model [21]. In that study, the incremental trauma group underwent 2 g, 3.5 g, 5 g, 6.5 g and 8 g impacts, while the single trauma group underwent only the 8 g impact. Flexibility testing was performed on each specimen while intact, and following each impact. There were no statistically significant differences in the flexibility parameters between the incremental and single trauma groups following the 8 g impacts. This well-controlled experiment validates the use of the incremental trauma protocol for determining disc strains due to frontal impact. Lastly, the current model did not simulate an active neuromuscular response: the MFR system provided postural stability and passive resistance to intervertebral motion following impact, simulating the response of an unwarned occupant. While individuals may be able to anticipate frontal collisions, they may not be able to generate adequate neck muscle force to restrict head flexion. The average peak extension moment of 18.7 Nm at C7–T1 produced by the present MFR exceeds the peak extension moment developed by weaker subjects (in vivo female range: 9 to 33 Nm) [49]. In addition to not being able to generate enough force to resist motion, the muscles may not be able to achieve peak tension quickly enough to protect the cervical spine from injury. The peak disc annular tissue strain and disc shear strain of the current study occurred between 148.5 ms and 151.3 ms, respectively, following impact. A recent study measured the electromyographic activity of the cervical muscles during simulated frontal impacts up to 1.4 g and found that the trapezius muscle was the first to achieve peak activity at 189 ms following impact [36]. Thus, the posterior cervical muscles may neither be able to develop sufficient strength, nor respond quickly enough to restrict head flexion or protect the neck following frontal impact.

Comparison of the current frontal impact disc shear strain data to previously reported rear impact data [50] revealed differing intervertebral levels at risk for injury (Fig. 6). The C2–C3 intervertebral level displayed greater increases during frontal impact, while in rear impact, C5–C6 was at the greatest risk. Although studies have demonstrated injuries to the C2–C3 disc during frontal impacts [9, 28, 56], a larger body of evidence notes that the C5–C6 intervertebral level is most commonly injured in both frontal and rear impact collisions [9, 12, 29, 32, 52, 62]. The injuries at C5–C6 during frontal impact occurred at 10 g, while injuries due to rear impact began at 3.5 g. This observation concurs with epidemiological evidence, which suggests that injuries may occur at lower impact accelerations during rear impact, as compared with frontal impact [34, 53, 55]. The data of the current study suggest that injuries observed at C2–C3 and C5–C6 might be due to the increases in disc annular tissue strain and disc shear strain beyond the physiological limits.

Fig. 6
figure 6

Comparison of peak disc shear strain during frontal and rear impact [50] at 8 g (C2–C3 through C6–C7)

Full size image

Understanding how subfailure injuries may cause pain is perhaps the most important issue when investigating disc injuries sustained during frontal impact. Apart from neck pain, headache is the most commonly reported symptom following frontal and rear impacts [2, 53]. Occipital or cervicogenic headache is known to result from referred pain via injury to the C2–C3 zygapophysial joint, which is innervated by the third occipital nerve [2, 3, 30]. Injury to the C2–C3 disc, suggested by the present study, may lead to accelerated degenerative changes over time, and abnormal loading of the adjacent zygapophysial joints [8]. Frontal impact patients may be at higher risk for occipital headaches and thus obtain relief from therapeutics targeting the occipital nerve [14, 15]. Other sources of chronic pain from subfailure injuries have been ascribed to mechanisms directly involving the disc, such as the presence of injured nociceptive fibers on the annulus fibrosus [4, 10], and the subsequent production of irritative inflammatory mediators [7, 33]. A final, commonly proposed pain mechanism is the accelerated degeneration of the injured disc, leading to facet degeneration and painful facet inflammation [8, 42, 43].

Conclusions

The present study demonstrated that disc annular tissue strain and disc shear strain exceeded physiological limits during simulated frontal impact. These data suggest that subfailure disc injuries that occur during frontal impact might be caused by combined intervertebral flexion and disc shearing. A comparison to previously reported rear impact data revealed that the C5–C6 intervertebral disc was at the greatest injury risk during rear impact, while during frontal impact, in addition to C5–C6, subfailure injuries are likely at superior intervertebral levels, including C2–C3. The disc injuries occurred at lower impact accelerations during rear impact, as compared with frontal impact. Disc injuries may be a source of chronic pain following both frontal and rear impacts.