Clinical Biomechanics

Transcription

1 Clinical Biomechanics 2 (29) Contents lists available at ScienceDirect Clinical Biomechanics journal homepage: Whiplash injury prevention with active head restraint Paul C. Ivancic a, *, Daohang Sha b, Manohar M. Panjabi a a Biomechanics Research Laboratory, Department of Orthopaedics and Rehabilitation, Yale University, School of Medicine, New aven, CT, USA b Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, School of Medicine, Philadelphia, PA, USA article info abstract Article history: Received 1 April 29 Accepted 3 June 29 Keywords: Whiplash Injury prevention Active head restraint Biomechanics Cervical spine Background: Previous epidemiological studies have observed that an initial head restraint backset greater than 1 cm is associated with a higher risk of neck injury and persistent symptoms. The objective of this study was to investigate the relation between the active head restraint position and peak neck motion using a new human model of the neck. Methods: The model consisted of an osteoligamentous neck specimen mounted to the torso of a rear impact dummy and carrying an anthropometric head stabilized with muscle force replication. Rear impacts (7.1 and 11.1 g) were simulated with and without the active head restraint. Physiologic rotation was determined from intact flexibility tests. Significant reductions (P <.5) in the spinal motion peaks with the active head restraint, as compared to without, were identified. Linear regression analyses identified correlation between head restraint backset and peak spinal rotations (R 2 >.3 and P <.1). Findings: The active head restraint significantly reduced the average peak spinal rotations, however, these peaks exceeded the physiologic range in flexion at head/c1 and in extension at C/5 through C7/. Correlation was observed between the head restraint backset and the extension peaks at C/5 and C5/6. Interpretation: Correlation between head restraint backset and spinal rotation peaks indicated that a head restraint backset in excess of. cm may cause hyperextension injuries at the middle and lower cervical spine. The active head restraint may not be fully activated at the time of peak spinal motions, thus reducing its potential protective effects. Ó 29 Elsevier Ltd. All rights reserved. 1. Introduction * Corresponding author. Address: Biomechanics Research Laboratory, Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, 333 Cedar St., P.O. Box 271, New aven, CT , USA. address: paul.ivancic@yale.edu (P.C. Ivancic). Rear automobile collisions commonly cause neck injuries resulting in acute and persistent symptoms (Quinlan et al., 2). A recent epidemiological study found that, among European countries, Switzerland had the highest cost per whiplash claim of 35, Euro (Chappuis and Soltermann, 2). The origin of the persistent neck pain may be an acutely injured neck tissue (Lord et al., 1996) or psychosocial processes, possibly initiated by an acute organic lesion (Ferrari and Schrader, 21). Persistent symptoms continue to be reported despite the 1969 implementation of Federal Motor Vehicle Safety Standards, FMVSS 22 (NTSA, 1969), which required manufacturers to incorporate head restraints in new automobiles. This standard did not establish requirements for head restraint position. Previous epidemiological studies have found that FMVSS 22 has been ineffective in preventing neck injuries (Minton et al., 2; Morris and Thomas, 1996; Sturzenegger et al., 1995), while others have reported reduction in neck injuries of up to only 2% (Chapline et al., 2; Kahane, 192; Olney and Marsden, 196; O Neill et al., 1972; States et al., 1972). In these studies, neck injuries were assessed by evaluating patient symptomatology for up to 1 year. Updated standards in FMVSS 22a (NTSA, 2) attempt to reduce the risk of whiplash injuries by requiring improvements in head restraint geometry and position. These standards rely mainly upon epidemiological studies of whiplash patients, with little supporting biomechanical data. These retrospective epidemiological data indicate that a gap of greater than 1 cm between the head restraint and the back of the head prior to rear impact was associated with persistent neck symptoms lasting longer than a year (Jakobsson, 2; Olsson et al., 199). Active neck injury prevention systems, such as the active head restraint (Wiklund and Larsson, 199) and energy absorbing seat (Jakobsson et al., 2), have been developed for some automobiles. Previous studies have indicated the potential benefits of active injury prevention systems in reducing the incidence of whiplash injuries by as much as 75% (Farmer et al., 23; Jakobsson et al., 2; Viano and Olsen, 21). owever, their implementation is without thorough understanding of 26-33/$ - see front matter Ó 29 Elsevier Ltd. All rights reserved. doi:1.116/j.clinbiomech

2 7 P.C. Ivancic et al. / Clinical Biomechanics 2 (29) their injury prevention mechanisms. To reduce injurious head and neck loads and motions, the active head restraint rotates forward based on occupant momentum pressing into the seatback during whiplash (Viano and Olsen, 21). Previous biomechanical studies have evaluated the effects of head restraint position on head and neck loads and motions during simulated whiplash of volunteers (Siegmund et al., 1999), whole cadavers (Geigl et al., 199), anthropometric test dummies (Svensson et al., 1996; Szabo et al., 23), and computer models (Eriksson, 2; Szabo and Welcher, 1996; Tencer et al., 22; Tencer et al., 21). Volunteer and cadaver experiments have shown that standard, non-active automobile seats and head restraints were insufficient to limit excessive head motions (Geigl et al., 199). Simulated rear impacts of anthropometric test dummies have demonstrated the ability to virtually eliminate neck extension by altering the head restraint and seatback properties (Svensson et al., 1996; Szabo et al., 23). Computational models have found that an initial head restraint gap of at most 5 6 cm was needed in order to reduce neck loads and motions to within the physiologic range (Sendur et al., 25; Stemper et al., 26). Although these studies imply relationships between the initial head restraint position and neck loads and motions, such relationships have not been evaluated or proven. The goals of the present study were to develop a new human model of the neck (UMoN) for whiplash simulation, consisting of a neck specimen mounted to the torso of a rear impact dummy and carrying an anthropometric head, and to use the model to investigate the relation between the active head restraint position and peak neck motion. 2. Methods In the present study, we define the potential for whiplash injury due to average high-speed neck motion during simulated rear impact beyond the physiologic limit at any spinal level. The physiologic rotation limit was defined at each spinal level as the average peak rotation obtained from the flexibility tests while the physiologic rotation range was defined as the average peak rotation ±1 SD Specimen preparation and flexibility testing Six fresh-frozen human osteoligamentous whole cervical spine specimens (occiput-) were mounted in resin at the occiput and in normal neutral posture. When mounted to the torso of the rear impact dummy, the average vertebral flexion angles relative to the horizontal measured by radiography were:.6 C1, 19. C2, 25.9 C3, 25.7 C, 29. C5, 33.5 C6, 35.1 C7, and 27., consistent with the in vivo neutral posture (Descarreaux et al., 23). The average age of the specimens was 3. years (range: 79 9 years) and there were four male and two female donors. Apart from typical age related changes, the specimens did not suffer from any disease that could have affected the osteoligamentous structures. Lightweight motion-tracking flags with 9.5 mm diameter markers were rigidly attached to each vertebra (C1 C7) and to the occipital and mounts. An additional flag marker was fixed at the center of mass (CoM). Sagittal flexibility testing was performed to determine physiologic spinal motions of the intact specimens. Pure moments up to 1.5 Nm were applied to the occipital mount via a loading jig in four equal steps while the mount remained fixed. The weights of the loading jig and occipital mount were counterbalanced throughout the tests. To allow for viscoelastic creep, 3 s break periods were given following each load application. After two preconditioning cycles, spinal motion data were recorded at each load increment of the third loading cycle uman model of the neck (UMoN) The UMoN consisted of the neck specimen with its mount rigidly connected to the torso of a rear impact dummy, BioRID II (Denton ATD Inc, Milan, O, USA), and an anthropometric surrogate head rigidly connected atop the occipital mount (Fig. 1a). The head (.2 kg mass;.2 kg m 2 sagittal plane moment of inertia) had a motion-tracking marker fixed at its CoM. The head and neck were stabilized using the compressive muscle force replication (MFR) system (Ivancic et al., 25). The MFR was symmetric about the mid-sagittal plane and consisted of two anterior, two posterior, and eight lateral cables. The anterior cables originated at the head and ran through separate guide posts anterior to the C vertebra. To apply the posterior MFR, small holes were drilled bilaterally into each lamina (C3 through C7) in which wire loops were inserted into the spinal canal and tightly secured above each y Global +Rx z Gap (cm) AR 3 AR 2 AR 5 AR AR 1 a) UMON b) AR Position eight (cm) Above CoM Below CoM Fig. 1. (a) Photograph of the human model of the neck (UMoN) and rear impact apparatus. Motion-tracking flags were fixed to the head, vertebrae, sled, seatback, and active head restraint. Bi-axial accelerometers were fixed to the head CoM, CoM, and sled. A contact switch was mounted to the back of the head. The global coordinate system had its z-axis horizontal and positive forward, y-axis vertical and positive upward, and x-axis horizontal and positive to the left. Rotation was defined as positive for flexion (+Rx) and negative for extension ( Rx). (b) Schematic of the average active head restraint (AR) position immediately prior to impact for each of the five positions studied.

3 P.C. Ivancic et al. / Clinical Biomechanics 2 (29) vertebral spinous process. The two posterior MFR cables originated from the head and ran through the wire loops. To apply the lateral MFR, lateral guide rods were inserted in the frontal plane into the vertebral bodies of C2 through C7. The lateral rods at C3 through C7 were positioned at the approximate centers of rotation of C2/ 3 through C6/7, respectively. The bilateral MFR cables originated from the head, C2, C and C6 vertebrae and passed alternately along the lateral guide rods. All cables passed through low-friction housing below the cervical spine. The housing was fixed at the mount and the pelvis, while the cables glided freely within the housing. After exiting bilaterally from the pelvis, each cable was connected to a separate preloaded spring. The stiffness coefficients of the anterior, lateral, and posterior springs were.,. and. N/mm, respectively. The preload was 15 N in each anterior and posterior spring and 3 N in each lateral spring. With this MFR arrangement the compressive neutral posture pre-loads at each spinal level were: 12 N (head/c1, C1/2); 1 N (C2/3, C3/ ); 2 N (C/5, C5/6) and 3 N (C6/7, C7/). This muscle force replication system, which provides postural neck stability and passive resistance to intervertebral motions during rear impact, produces a high-speed kinematic response similar to in vivo data (Ivancic et al., 25). No preload was applied to the anthropometric thoracic or lumbar spine regions Rear impact apparatus and monitoring The apparatus consisted of an automobile seat mounted on a custom sled in which the UMoN was seated and secured with a seatbelt (Fig. 1a). The front seat of a 26 Kia Sedona minivan with active head restraint (AR) was used (Kia Motors America, Inc, Irvine, CA, USA). The AR was activated by UMoN s momentum pressing into the seatback during whiplash and rotated forward via a pivoting mechanism between it and the seatback. By raising and moving forward relative to the seatback during rear impact, the AR was designed to reduce neck loads and motions even if the head restraint is not properly positioned at the time of impact (Viano and Olsen, 21). The seat base angle was 1 from the horizontal while the initial seatback angle was 2 from the vertical. Motion-tracking flags were rigidly attached to the sled, seatback, and AR. A custom floor pan was mounted to the anterior of the sled to support UMoN s legs and feet. The sled was mounted on low-friction linear bearings, which translated along two precision ground, stainless steel shafts. The sled was accelerated using an acceleration generation system consisting of a piston, high-energy compression springs, and a computer-controlled electromagnetic release. At the time of the electromagnet release, a trigger signal initiated high-speed camera recording. The high-speed digital camera (MotionPRO, Redlake MSAD, San Diego, CA, USA) recorded the sagittal motions of all flag markers (head, vertebral, sled, seatback, and AR) at 5 frames/s. A contact switch was mounted to the back of the head to determine the time of head/ar contact. A bi-axial accelerometer (5 g capacity; Part No. ADXL25JQC, Analog Devices, Norwood, MA, USA) was fixed to the sled and continuously sampled at 1 kz using an analog-to-digital converter and a personal computer. A custom pneumatic braking system was designed to gradually decelerate UMoN following impact. 2.. Simulated whiplash with and without AR Whiplash was simulated with the AR in the each of the five positions immediately prior to impact (Fig. 1b): minimum gap and height (AR 1), midrange gap and maximum height (AR 2), maximum gap and height (AR 3), midrange gap and minimum height (AR ), and midrange gap and height (AR 5). The average (SD) head/ar gap ranged from 3.2 (2.1) cm for AR1 to 7.5 (2.) cm for AR3 while AR height ranged from 1.2 (2.) cm below to.5 (2.) cm above the head CoM. Subsequently, whiplash was simulated without the AR. The impacts were first performed at a maximum measured horizontal sled acceleration of 7.1 g and subsequently at 11.1 g Data analyses Custom Matlab programs (Matlab, The Mathworks Inc., Natick, MA, USA) were used to obtain the coordinates of the flag markers with sub-pixel accuracy and to compute the spinal motions throughout the flexibility tests and during whiplash. The horizontal velocities of the head CoM, CoM, and sled were obtained by numerical differentiation of the translation data. Sled acceleration was determined using the accelerometer while the accelerations of the head and CoMs were obtained by numerical double differentiation of the motion data. Acceleration and velocity data were expressed in the global coordinate system (Fig. 1a) which was fixed to the ground and had its z-axis horizontal and positive forward, y- axis vertical and positive upward, and x-axis horizontal and positive to the left. Rotation was defined as positive for flexion (+Rx) and negative for extension ( Rx). The head/ translations were expressed in the coordinate system which was fixed to and moved with the vertebra. The /sled translations were expressed in the sled coordinate system which was fixed to and moved with the sled. The and sled coordinate systems were aligned with the global coordinate system immediately prior to impact. All data were digitally filtered using a 3rd order, dual pass, Butterworth low-pass filter at a cutoff frequency of 3 z. The motion, velocity, and acceleration peaks and their occurrence times relative to the onset of the sled acceleration were determined. Single factor, repeated measures ANOVA and Bonferonni post hoc tests (P <.5) were performed to determine significant reduction in the spinal motion peaks with the AR, as compared to without. Linear regression analyses were performed to identify correlation (R 2 >.3; P <.1) between the AR position, peak spinal motions, and head/ar contact time for all impacts combined Error analyses A custom jig was designed to determine the overall error in the calculation of the high-speed spinal rotations during whiplash. The jig consisted of a high-speed rotator, with constant velocity, which carried a motion-tracking flag. The jig was rigidly fixed to the torso to the BioRID II, which was rear impacted at 7.1 g. The motiontracking software was used to track the flag marker positions and to compute the angle change between frames. The average rotation error was. (SD.3 ). The average flag marker translation error, as determined in a separate study, was.2 mm (SD.1 mm) (Ivancic et al., 26). 3. Results Sample time histories of acceleration, translation, and rotation data during simulated whiplash with and without the AR were illustrative of the series (Fig. 2). The head/ and /sled motions included extension rotation and posterior shear translation (Fig. 2e, g, i, j). The AR caused axial separation at head/, as compared to compression with no AR (Fig. 2f). Up to the time of head/ar contact, little difference was observed in the spinal rotations with or without the AR (Fig. 2m t). Peak AR/seatback flexion and head/ar contact occurred simultaneously at 66 ms (Fig. 2l). All spinal motion peaks occurred later, with C/5 extension occurring at 156 ms (Fig. 2q). The cervical spine formed an S-shaped curvature, defined by flexion rotation of the upper spinal levels, head/

4 72 P.C. Ivancic et al. / Clinical Biomechanics 2 (29) Translation (cm) Acceleration 15 a) Sled (g) 15 b) CoM(g) 15 c) ead CoM(g) 6 d) α ead (rad/s 2 ) 1 Az Az 5 Ay Az Ay e) Tz ead/ 12 f) Ty ead/ 12 g) Tz 12 h) Ty i) ead/ 5 j) 5 k) Seatback/Sled 5 l) AR/Seatback m) ead/c q) C/ n) C1/ r) C5/ o) C2/ s) C6/ p) C3/ t) C7/ Time (ms) Time (ms) Time (ms) Time (ms) Without AR With AR Fig. 2. Sample time-history data for acceleration, translation, and rotation during simulated whiplash without the AR (thin line) and with AR 5, midrange gap and height (thick line), for specimen #6. Data include: (a c) horizontal accelerations of the sled, CoM, and head CoM; (d) head angular acceleration; (e h) translations of head/ and /sled; (i l) rotations of head/, /sled, seatback/sled, and AR/seatback; (m t) intervertebral rotations. The time of head/ar contact for the impact with AR 5 is shown by a vertical line in panel l. C1 through C3/, and simultaneous extension of the lower levels. The spinal rotation peaks and rotation durations were generally less with the AR, as compared to without (13. vs head/ C1; 1.5 vs C/5;.7 vs. 7. head/). The average peak accelerations and velocities (sled,, and head), rotations (head/, /sled, seatback/sled, and AR/seatback), and occurrence times of the peaks and head/ar contact are presented in Table 1. Peak horizontal acceleration of the sled

5 P.C. Ivancic et al. / Clinical Biomechanics 2 (29) Table 1 Peak accelerations (g), velocities (kph), rotations (deg) and occurrence times (ms). Average (SD) peak data with and without the active head restraint (AR) during rear impacts of 7.1 and 11 g: (a) horizontal accelerations and pulse durations (sled,, and head); (b) horizontal velocities (sled,, and head); (c) vertical accelerations ( and head); rotations of (d) extension (head/, /sled and seatback/sled); (e) flexion (head/ and AR/seatback); (f) head/ar contact times. 7.1 g 11.1 g AR 1 AR 2 AR 3 AR AR 5 No AR AR 1 AR 2 AR 3 AR AR 5 No AR (a) orizontal acceleration (+Az) Magnitude (g) Sled 7.1 (.7) 7.1 (.5) 7.1 (.5) 7.2 (.3) 7.2 (.3) 7.2 (.2) 1.6 (.) 1.9 (.5) 11.1 (.) 11.2 (.2) 11.2 (.3) 11.5 (.3) 5. (.7) 5.6 (.7) 5.6 (.6) 5.3 (.5) 5.5 (.6) 5.1 (.5) 7. (.6) 9.9 (.7) 9. (1.) 9. (.6) 9.7 (.) 9.7 (.5) ead 3. (.5) 2.9 (1.1) 2.7 (1.1) 3.1 (.6) 3.3 (.) 2.1 (.7) 5.5 (.) 7. (.7) 6.7 (.5) 6.6 (.) 6. (.3).6 (.9) Sled 2. (1.) 27.7 (.) 27 (1.1) 26.3 (.) 26.7 (1.) 27.7 (.) 2.7 (1.) 27.3 (1.6) 2. (1.3) 27.3 (1.6) 27. (1.7) 27. (1.3) 75.7 (7.1). (6.3) 7.7 (5.) 76. (7.2) 77. (.3) 7.7 (6.2) 7.3 (6.6) 76. (3.6) 76.7 (2.) 73. (3.3) 73.7 (2.3) 77. (.5) ead (1.) (.5) 15 (31.5) 15.3 (27.1) 16. (19.2) Pulse duration (ms) Sled.3 (3.2) 7. (3.) 6.3 (2.9) 5.7 (2.) 5.3 (1.6) 7. (2.) 5.3 (3.) 7. (2.1) 6.7 (2.1) 6.7 (2.) 5.7 (2.9) 5.7 (2.) (3.7) (17.3) (19.3) (2.1) (15.3) (16.7) (19.7) (1.5) (17.3) (2.9) (23.3) (2.2) ead (.) (69.5) 21.7 (65) 197. (5.) 171. (3.5) 21.3 (.7) 13.3 (17.) (.1) 19. (13.1) 12.3 (9.) (9.2) 1. (17.) (b) orizontal velocity (+Vz) (kph) Sled 7. (.) 6. (.3) 6.9 (.3) 6.7 (.) 6. (.) 6.9 (.) 9. (.3) 1.3 (.2) 1. (.) 1. (.3) 1.3 (.) 1.7 (.) 7. (.1) 7.1 (.) 7.1 (.) 7. (.) 7.2 (.7) 6.9 (.1) 11.1 (.5) 11.7 (.5) 11.1 (.5) 11.2 (.7) 11.5 (.6) 1. (.5) ead.9 (.7). (1.).3 (.7). (.7).9 (.). (.) 13. (.7) 13.3 (1.6) 11. (1.6) 13.5 (.5) 1. (.6).5 (2.) (c) Vertical acceleration +Ay Magnitude (g) 1.2 (.2) 1.2 (.2) 1.3 (.2) 1.2 (.1) 1.2 (.1) 1.3 (.2) 2. (.2) 2.2 (.2) 2. (.3) 2. (.3) 1. (.3) 2.5 (.2) ead 2.3 (.) 2.1 (.) 2.1 (.) 2.1 (.5) 2.2 (.5) 2. (.) 2.3 (.3) 2. (.3) 2.6 (.2) 2.5 (.5) 2.9 (.) 3.1 (.3) 75. (11.9) ead 69.3 (1.6) 72.7 (.) 12.3 (7.5) 1.3 (1.) 11.3 (11.3) 1. (.3) 3. (5.). (6.6) 76. (6.6) 5. (6.5) 1.3 (3.) 79.3 (1.) 1.3 (2.1) 7.7 (3.) 7.7 (.3).7 (1.) 7. (2.5) 69.3 (2.) 12.7 (2.9) 5.3 (29.1) (1.6) (2.3) 11. (3.3) 73.3 (2.1) 75.7 (7.1) 1. (32.1) 11.7 (2.7) (7.9) 6.3 (2.3) 176. (15.5) Ay magnitude (g) 1. (.2) 1.1 (.2) 1.2 (.2) 1.2 (.2) 1.1 (.1) 1.2 (.2) 1. (.2) 1.9 (.2) 2. (.1) 1.9 (.2) 1.9 (.1) 2. (.2) ead 1.6 (.3) 2. (.5) 2.5 (.) 1. (.5) 1.7 (.6) 2.6 (.5) 2.3 (.6) 2. (.9) 3.7 (1.) 3. (1.) 2.5 (.5) 5.9 (.9) (2.7) ead (3.) (5.3) (6.1) (13.) (3.9) 16.3 (16.) (5.) (6.) (9.) (d) Extension ( Rx) Magnitude (deg) ead/ 9. (.9) 2. (11.2) 26. (7.3) 1.3 (.2) 1. (.9) 35.7 (1.9).5 (3.5) 1.5 (.5) 22.3 (7) 11.1 (3.6) 9. (3.6) 53.5 (13.3) /sled 12.2 (3.2) 11.9 (1.9) 12.5 (1.) 13. (1.) 12.5 (.5) 1. (2.1) 17.3 (.9) 15.5 (1.5) 15.9 (1.) 16. (.9) 1.7 (1.) 17.7 (.) Seatback/sled 2. (.2) 2. (.2) 2.1 (.1) 1.9 (.1) 2. (.1) 2. (.2) 2. (.1) 3.2 (.3) 3.3 (.1) 3.1 (.2) 3.1 (.2) 3.2 (.1) ead/ 163 (17.) (23.1) (19.9) 179 (1) 16 (22.2) /sled (22.2) (9.) (9.2) (15.) (11.) Seatback/sled 56.7 (3.9) 59.3 (1.6) 59. (2.1) 56.3 (2.3) 61. (13.3) (16) 12. (.9) (1.7) (37.5) (6.5) 11.7 (11.) (5.9) 11. (6.) 121. (2.) 12.7 (3.) 2.7 (2) (23.7) (1.2) 17 (15.6) (25.) (1.5) (7.) (5.5) (.5) (3.3) 65.3 (5.3) 7. (2.) 5.7 (3.) 59. (1.7) 63.7 (1.3) (2.1) 126. (39.5) 172 (15.6) (3.9) 63. (11.6) 12.7 (7.7) 112. (3.6) 15.7 (21.3) 17.3 (.5) 66.7 (3.) (e) Flexion (+Rx) Magnitude (deg) ead/ 5.5 (1.3) 5. (2.1) 6. (1.) 5.7 (1.5) 6.2 (1.7) 5.6 (1.3) 7.1 (1.).7 (.9) 7.5 (1.2) 6. (.) 7.5 (.) 7.3 (.9) AR/seatback. (.7) 3.5 (.7) 3.5 (.5).5 (.3).3 (.) n/a 5. (.7).6 (.6) 5.2 (.6) 5.6 (.6) 5.3 (.3) n/a ead/ 79.7 (6.3) (7.6) 7.3 (.3) 7.3 (7.) 7.7 (6.) 7.7 (3.3) 1.3 (3.5) 77. (1.1) 75.7 (2.) 7. (1.3) 75. (1.1) (1.1) AR/seatback 66.3 (2.3) 5.7 (.3) 1.7 (2.7) 69. (3.5) 71. (3.3) n/a 61.7 (3.9) 7.3 (1.5) 5.3 (.3) 61. (6.) 65.3 (3.5) n/a (f) ead/ar Contact time (ms) 72. (13.) 9.5 (1.6) 13. (6.9) 1.7 (1.1) 6. (19.3) n/a 61. (2.9) 73.7 (2.7) 92.3 (22.) 65.3 (.6) 65.2 (2.5) n/a occurred first, followed by, and the head (Table 1a). The duration of the horizontal acceleration pulse ranged from 5.3 to.3 ms for the sled, 6.7 to 9.7 ms for, and 1 to 21 ms for the head. The highest peak horizontal velocities were 1.7,

6 7 P.C. Ivancic et al. / Clinical Biomechanics 2 (29) , and 1. kph for the sled,, and head, respectively (Table 1b). Lower vertical acceleration peaks were generally observed with the AR, as compared to without, with the exception of the positive peaks during the 7.1 g impacts (Table 1c). Dramatic reductions in the head/ extension peaks were observed with the AR, as compared to without (Table 1d). Peak /sled extension, reaching 17.7 with no AR, was generally reduced with the AR. Peak AR/seatback flexion ranged between 3.5 and 5.6 (Table 1e) while the head/ar contact time ranged between 61. and 1.7 ms (Table 1f). The average spinal rotation peaks are presented graphically together with the physiologic limits (Fig. 3). The AR significantly reduced the extension peaks throughout the middle and lower cervical spine, however, motion beyond the physiologic limit was observed. At C5/6 and C6/7, peak extension during the 11.1 g impacts was significantly reduced by all AR positions with the exception of AR 3 (maximum gap and height). AR height did not correlate with any biomechanical parameters. In contrast, correlation was observed between head/ar gap and the extension peaks at C/5 and C5/6 and head/ar contact time (Fig. ). The average times at which the peak accelerations and motions were attained demonstrate temporal differences due to the AR (Fig. 5). The AR altered the pattern of peak head/ translations. Intervertebral extension peaks occurred earlier with the AR, as compared to without, with the exception of head/c1 and C3/. Peak AR/seatback flexion occurred prior to all spinal motion peaks. 3 ead/c1 C1/ C2/3 C3/ C/ * * * * * 3 C5/6 C6/7 C7/ * * * * * * * * No No No With AR AR With AR AR With AR AR 7.1 g 11.1 g Fig. 3. Average peak intervertebral rotations with and without the active head restraint (AR). The physiologic rotation range (average ± 1 SD) is indicated in grey shading. The 7.1 g data are indicated by white squares (h) while the 11.1 g data are indicated by black squares (j). Significant reduction (P <.5) in average peak extension with the AR, as compared to without, is indicated by a single asterisk for the 11 g data and by a double asterisk for both the 7.1 and 11.1 g data.

7 P.C. Ivancic et al. / Clinical Biomechanics 2 (29) or Time (ms) a) C/5 Extension vs. Gap Rx (C/5) =.9 Gap + 2. R2 = ead/ar Gap (cm) b) C5/6 Extension vs. Gap Rx (C5/6) =. Gap R2 = c) Time of ead Contact vs. Gap Time = 6.7 Gap +. R2 = ead/ar Gap (cm) ead/ar Gap (cm) Fig.. Correlation between the head/ar gap and: (a) C/5 extension; (b) C5/6 extension; (c) head/ar contact time. The in vivo physiologic rotation ranges at C/5 and C5/ 6, indicated in grey shading, are based on the overall averages (±1 SD) reported by previous studies (Lind et al., 199; Ordway et al., 1999; Penning, 197).. Discussion In attempt to reduce whiplash injuries, updated Federal Motor Vehicle Safety Standards, FMVSS 22a (NTSA, 2), require manufacturers to improve head restraint geometry and position in new automobiles. These standards rely primarily upon epidemiological studies, with little supporting biomechanical data. Relationships between head restraint position and neck loads and motion have not been validated. The present study utilized a new human model of the neck (UMoN, Fig. 1a) to investigate the relation between the active head restraint (AR) position (Fig. 1b) and peak neck motion. UMoN consisted of a neck specimen mounted to the torso of BioRID II and carrying an anthropometric surrogate head. While the AR significantly reduced the average peak spinal motions throughout the middle and lower cervical spine, as compared to no AR, these peaks exceeded the physiologic limits in flexion at head/c1 and in extension at C/5 through C7/ (Fig. 3). These data indicated potential soft tissue injuries at these spinal levels during whiplash even in the presence of the AR. Among AR positions, AR 3 (maximum gap and height) generally allowed the largest motions above the physiologic range, with the greatest potential for extension injuries at C/5 through C7/. Neither the C5/6 nor the C6/7 extension peaks allowed by AR 3 could be differentiated from those attained with no AR, while all other AR positions significantly reduced these peaks. The AR positions with the smallest gap (AR 1 and 5) generally allowed the least motion in excess of physiologic. The present model has limitations that should be considered before our results are further interpreted. The average duration of the sled acceleration pulse, ranging between 5.3 ms and.3 ms, was less than the 1 ms pulse used for standardized evaluation of head restraint effectiveness (RCAR-IIWPG, 2). Real-life whiplash injuries can occur at impact severities less than that prescribed in standardized tests. UMoN utilized an anthropometric head with mass.2 kg, which may have been heavier than the actual head mass of some specimens. A single AR and seat from a 26 Kia Sedona minivan were used, thus the effects of different AR or seat designs were not investigated. To obtain sufficient data needed to understand the role of the AR in neck injury prevention, multiple impacts were performed using the same specimen. In order to minimize the effect of injury on data obtained in subsequent impacts, testing was conducted first with the AR, followed by without, at an impact severity of 7.1 g followed by 11.1 g. These limitations are contrasted by UMoN s numerous advantages. UMoN is a logical evolution of our previous head neck model for the study of whiplash injury prevention systems (Ivancic et al., 25). Muscle force replication, applied to the head and neck of UMoN, provided postural stability and produced a high-speed kinematic response similar to in vivo data. In our previous model, whiplash was simulated by applying horizontal acceleration to a mini-sled on which the vertebra was rigidly fixed. This previous model did not replicate the in vivo /sled motions of extension, posterior shear, or axial separation due to straightening of the thoracic spine during whiplash (Fig. 2g, h, j). Our present highest average peak /sled extension of 17.7 (Table 1d) is in good agreement with the 19.3 of extension that was reported for kph impacts of volunteers (Siegmund et al., 21). The present data may be used to understand the relationship between AR position and whiplash injury risk, specific to each spinal level. AR height did not correlate with any biomechanical parameters, indicating that AR height may not be able to predict peak spinal motions during whiplash. These results are consistent those of a previous study which observed correlation only between head restraint height and peak head angular acceleration, upper neck loads, and Nij (Siegmund et al., 25). We observed correlation between the head/ar gap and extension peaks at C/5 and C5/6 (Fig. a and b). Based upon these correlations, motion beyond the average in vivo physiologic limit may occur at these spinal levels due to head/ar gaps in excess of. and. cm, respectively, causing extension injuries. These results are consistent with previous epidemiological studies which observed higher neck injury risk and greater incidence of chronic symptoms for a head restraint gap larger than 1 cm (Jakobsson, 2; Olsson et al., 199). Federal Motor Vehicle Safety Standards and international rating systems support our relationships between AR position and peak spinal motion. The updated Federal Motor Vehicle Safety Standards, FMVSS 22a, requires a head restraint gap of at most 5.5 cm in newly manufactured automobiles (NTSA, 2). The current rating system of the International Insurance Whiplash Prevention Group classifies a head restraint gap of less than 7 cm as good, 7 9 cm acceptable, 9 11 cm marginal, and greater than 11 cm poor (RCAR-IIWPG, 2). In order to achieve the greatest protective effect for reducing whiplash injuries, the AR must be fully activated at the time of peak intervertebral rotations. The AR rotated forward based upon the dummy momentum pressing into the seatback and was, on average, fully activated at 65.3 ms when head/ar contact and peak seatback/sled extension occurred (11.1 g impact; AR 5; Fig. 5). The peak rotations at the spinal levels most susceptible to injury occurred later, after the AR had deactivated (153. ms head/c1; 16.7 ms C/5; ms C6/7; and ms C7/). This is demonstrated in the sample time-history data (Fig. 2l), in which AR/seatback flexion is zero at and beyond 12 ms. Deactivation of the AR may be due to several factors: inertial head loading may be

8 76 P.C. Ivancic et al. / Clinical Biomechanics 2 (29) Time (ms) Acceleration +Az +Ay -Ay Rotation +Rx -Rx Translation -Tz +Ty -Ty Without AR With AR A R T A R T of sufficient magnitude to deactivate the AR upon contact while torso loads applied to the seatback may be of insufficient magnitude to maintain prolonged AR activation. Peak seatback/sled extension occurred at 63 ms, on average, indicating that torso loads applied to the seatback may subsequently decrease. Inability of the AR to remain fully activated at the time of peak spinal rotations may reduce its protective effect for reducing whiplash injuries. 5. Conclusions Sled /C1 Seatback/Sled C3/ C1/2 C1/2 C7/ /C1 C6/7 C5/6 C/5 C2/3 Sled ead/ar Contact Seatback/Sled AR/Seatback C1/2 C7/ C2/3 C6/7 C1/2 C5/6 /C1 C/5 C3/ /C1 Fig. 5. Temporal analyses based upon the average occurrence times of key events during the 11.1 g rear impacts without the AR and with AR 5 (midrange AR gap and height). Key events include: peak accelerations of the sled,, and head (+Az horizontal, +Ay superior, Ay inferior); rotations of each spinal level (head/c1 through C7/), head/, /sled, seatback/sled, and AR/seatback (+Rx flexion, Rx extension); and translations of head/ and /sled (+Ty separation, Ty compression, Tz posterior shear). The relationship between the AR position and peak spinal motion was determined in the present study. Our results indicated that while the AR significantly reduced spinal rotation peaks throughout the middle and lower cervical spine, motion beyond the physiologic limit was observed. Injuries at the middle and lower cervical spine may occur with a head/ar gap of. cm or greater. The AR may not be fully activated at the time of peak spinal motions, thus reducing its potential protective effect. Continued epidemiological, clinical, and biomechanical studies of whiplash injury prevention systems will lead to improvements in head restraint and seat deign, thus reducing the frequency of whiplash injuries. Conflict of interest statement We declare no conflicts of interest. Acknowledgement This research was supported by Grant 5R1CE1257 from the Centers for Disease Control and Prevention (CDC). References Chapline, J.F., Ferguson, S.A., Lillis, R.P., Lund, A.K., Williams, A.F., 2. Neck pain and head restraint position relative to the driver s head in rear-end collisions. Accident Anal. Prev. 32, Chappuis, G., Soltermann, B., 2. Number and cost of claims linked to minor cervical trauma in Europe: results from the comparative study by CEA, AREDOC and CEREDOC. Eur. Spine J. 17, Descarreaux, M., Blouin, J.S., Teasdale, N., 23. A non-invasive technique for measurement of cervical vertebral angle: report of a preliminary study. Eur. Spine J. 12, Eriksson, L., 2. Neck injury risk in rear-end impacts. Risk factors and neck injury criterion evaluation with Madymo modelling and real-life data. Ph.D. thesis, Chalmers University of Technology, Gotenborg, Sweden. Farmer, C.M., Wells, J.K., Lund, A.K., 23. Effects of head restraint and seat redesign on neck injury risk in rear-end crashes. Traffic Injury Prev., 3 9. Ferrari, R., Schrader,., 21. The late whiplash syndrome: a biopsychosocial approach. J. Neurol. Neurosurg. Psychiat. 7, Geigl, B.C., Steffan,., Leinzinger, P., Roll, Muhlbauer, M., Bauer, G., 199. The movement of the head and cervical spine during rearend impact. In: International Research Conference on the Biomechanics of Impact, Lyon, France, pp Ivancic, P.C., Panjabi, M.M., Ito, S., Cripton, P.A., Wang, J.L., 25. Biofidelic whole cervical spine model with muscle force replication for whiplash simulation. Eur. Spine J. 1, Ivancic, P.C., Wang, J.L., Panjabi, M.M., 26. Calculation of dynamic spinal ligament deformation. Traffic Injury Prev. 7, 1 7. Jakobsson, L., 2. AIS 1 neck injuries in rear-end car impacts. Biomechanical guidelines and evaluation criteria based on accident data and parameter studies. Chalmers University of Technology. Jakobsson, L., Isaksson-ellman, I., Lindman, M., 2. WIPS (Volvo cars Whiplash Protection System)-the development and real-world performance. Traffic Injury Prev. 9, Jakobsson, L., Lundell, B., Norin,., Isaksson-ellman, I., 2. WIPS volvo s Whiplash Protection Study. Accident Anal. Prev. 32, Kahane, C.J., 192. An evaluation of head restraints, Federal Motor Vehicle Safety Standard 22. Report DOT-S-6-1 US Department of Transportation, Washington, DC. Lind, B., Sihlbom,., Nordwall, A., Malchau,., 199. Normal range of motion of the cervical spine. Arch. Phys. Med. Rehabil. 7,

9 P.C. Ivancic et al. / Clinical Biomechanics 2 (29) Lord, S.M., Barnsley, L., Wallis, B.J., Bogduk, N., Chronic cervical zygapophysial joint pain after whiplash. A placebo-controlled prevalence study. Spine 21, Minton, R., Murray, P., Stephenson, W., Galasko, C.S., 2. Whiplash injury are current head restraints doing their job? Accident Anal. Prev. 32, Morris, A.P., Thomas, P., Neck injuries in the UK co-operative crash injury study. Society of Automotive Engineers Paper No NTSA, 2. Federal motor vehicle safety standards; ead restraints; Final Rule. 9 CFR Part (239), pp NTSA, Federal motor vehicle standard No. 22, Code of Federal Regulations, Section Olney, D.B., Marsden, A.K., 196. The effect of head restraints and seat belts on the incidence of neck injury in car accidents. Injury 17, Olsson, I., Bunketorp, O., Carlsson, G., Gustafsson, C., Planath, I., Norin,., Ysander, L., 199. An in-depth study of neck injuries in rear end collisions. In: International Research Council on the Biomechanics of Impact, Bron, France. O Neill, B., addon Jr., W., Kelley, A.B., Sorenson, W.W., Automobile head restraints frequency of neck injury claims in relation to the presence of head restraints. Am. J. Public ealth 62, Ordway, N.R., Seymour, R.J., Donelson, R.G., ojnowski, L.S., Edwards, W.T., Cervical flexion, extension, protrusion, and retraction. A radiographic segmental analysis. Spine 2, Penning, L., 197. Normal movements of the cervical spine. AJR Am. J. Roentgenol. 13, Quinlan, K.P., Annest, J.L., Myers, B., Ryan, G., ill,., 2. Neck strains and sprains among motor vehicle occupants-united States, 2. Accident Anal. Prev. 36, RCAR-IIWPG, 2. RCAR-IIWPG Seat/ead Restraint Evaluation Protocol. < Sendur, P., Thibodeau, R., Burge, J., Tencer, A., 25. Parametric analysis of vehicle design influence on the four phases of whiplash motion. Traffic Injury Prev. 6, Siegmund, G.P., einrichs, B.E., Chimich, D.D., Lawrence, J., 25. Variability in vehicle and dummy responses in rear-end collisions. Traffic Injury Prev. 6, Siegmund, G.P., einrichs, B.E., Lawrence, J.M., Philippens, M., 21. Kinetic and kinematic responses of the RID2a, ybrid III and human volunteers in lowspeed rear-end collisions. Society of Automotive Engineers Paper No Siegmund, G.P., einrichs, B.E., Wheeler, J.B., The influence of head restraint and occupant factors on peak head/neck kinematics in low-speed rear-end collisions. Accident Anal. Prev. 31, States, J.D., Balcerak, J.C., WIlliams, J.S., Morris, A.T., Babcock, W., Polvino, R., Riger, P., Dawley, R.E., Injury frequency and head restraint effectiveness in rearend impact accidents. Society of Automotive Engineers Paper No Stemper, B.D., Yoganandan, N., Pintar, F.A., 26. Effect of head restraint backset on head-neck kinematics in whiplash. Accident Anal. Prev. 3, Sturzenegger, M., Radanov, B.P., Di Stefano, G., The effect of accident mechanisms and initial findings on the long-term course of whiplash injury. J. Neurol. 22, 3 9. Svensson, M.Y., Lovsund, P., aland, Y., Larsson, S., The influence of seat-back and head-restraint properties on the head-neck motion during rear-impact. Accident Anal. Prev. 2, Szabo, T.J., Voss, D.P., Welcher, J.B., 23. Influence of seat foam and geometrical properties on BioRID P3 kinematic response to rear impacts. Traffic Injury Prev., Szabo, T.J., Welcher, J.B., uman subject kinematics and electromyographic activity during low speed rear impacts. Society of Automotive Engineers Paper No Tencer, A.F., Mirza, S., Bensel, K., 22. Internal loads in the cervical spine during motor vehicle rear-end impacts: the effect of acceleration and head-to-head restraint proximity. Spine 27, 3 2. Tencer, A.F., Mirza, S., Bensel, K., 21. The response of human volunteers to rearend impacts: the effect of head restraint properties. Spine 26, (Discussion ). Viano, D.C., Olsen, S., 21. The effectiveness of active head restraint in preventing whiplash. J. Trauma 51, Wiklund, K., Larsson,., 199. Saab active head restraint (SAR) seat design to reduce the risk of neck injuries in rear impacts. Society of Automotive Engineers Paper No