Neurophysiological and biomechanical characterization of goat cervical facet joint capsules

ELSEVIER

Journal of Orthopaedic Research

Journal of Orthopaedic Research 23 (2005) 779-787

www.elsevier.com/locate/orthres

Neurophysiological and biomechanical characterization of goat cervical facet joint capsules Ying Lu *, Chaoyang Chen, Srinivasu Kallakuri, Ajit Patwardhan, John M. Cavanaugh Bioengineering Center, Wayne Stafe University, 818 W. Hancock, Detroit, M I 48202, USA

Accepted 5 January 2005

Abstract

Cervical facet joints have been implicated as a major source of pain after whiplash injury. We sought to identify facet joint capsule receptors in the cervical spine and quantify their responses to capsular deformation. The response of mechanosensitive afferents in C5-C6 facet joint capsules to craniocaudal stretch (0.5 d s ) was examined in anaesthetized adult goats. Capsular afferents were characterized into Group I11 and IV based on their conduction velocity. Two-dimensional strains across the capsules during stretch were obtained by a stereoimaging technique and finite element modeling. 17 (53%) Group I11 and 14 (56%) Group IV afferents were identified with low strain thresholds of 0.107 f 0.033 and 0.100 f 0.046. A subpopulation of low-strain-threshold afferents had discharge rate saturation at the strains of 0.388 f 0.121 (n = 9, Group 111) and 0.341 k 0.159 (n = 9, Group IV). Two (8%) Group IV units responded only to high strains (0.460 k 0.170). 15 (47%) Group I11 and 9 (36%) Group IV units could not be excited even by noxious capsular stretch. Simple linear regressions were conducted with capsular load and principal strain as independent variables and neural response of low-strain-threshold afferents as the dependent variable. Correlation coefficients (R2)were 0.73 f 0.1 1 with load, and 0.82 f 0.12 with principal strain. The stiffness of the C W 6 capsules was 16.8 f 11.4 N/mm. Our results indicate that sensory receptors in cervical facet joint capsules are not only capable of signaling a graded physiological mechanical stimulus, but may also elicit pain sensation under excessive deformation. 0 2005 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. Keywords: Mechanoreceptor;Nociceptor; Strain; Whiplash; Cervical facet joint capsule

Introduction Cervical facet joints (FJ) have been implicated as a major source of pain after whiplash injury [2,7,16]. Local anaesthetic blocks of cervical FJ indicate that 40-73% of whiplash patients have pain arising from FJ [2]. In cadaveric studies, high-speed cineradiography revealed that C5-C6 FJ capsules (FJC) experienced maximum strains of 5140% under low-speed rear impacts [6], while simulated whiplash loading on whole cervical spines showed maximum strains of around 40% in lower FJC [19,20]. In a behavioral study, Lee at al. reported Correspondingauthor. Tel.: +1313 577 3440; fax: +1 313 577 8333. E-mail address: [email protected](Y. Lu).

that CGC7 FJC strains of 1 1 4 2 % may be sufficient t o generate neck pain in an in vivo rat model [ 151. These studies provided compelling evidence indicating that excessive FJC strain may be responsible for the ensuing clinical symptoms. Despite the strong evidence for FJC involvement in whiplash, sensory neurophysiologic studies of cervical FJC, let alone FJC neural responses t o capsular deformation, are non-existent. These issues have gone unanswered, primarily due to the challenges of recording nerve activity from rather short cervical nerve roots, as compared to much longer lumbar nerve roots. Only recently, several studies examined the neurophysiologic properties of cervical paravertebral tissues [4,5,25]. Chen et al. first identified mechanoreceptors in goat cervical

0736-0266/$ - see front matter 0 2005 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.orthres.2005.01.002

780

Y. Lu et ul. I Journal of Orthopaedic Research 23 (2005) 779-787

FJCs and studied their responses to stretch [5]. Similarly, in lumbar FJCs, mechanoreceptors and nociceptors were identified and characterized in rabbit and cat for their responses to stretch in various directions [21,28]. However, none of these studies quantified capsular deformation in investigating FJC neural response. Thus, the aims of this study were to identify FJC receptors in the cervical spine and quantify their responses to capsular deformation. The current study characterized FJC neural response under a low-strain rate and along with further dynamic studies aids in the understanding of FJC injury mechanism in whiplash.

roots. A C 4 C 7 laminectomy was then performed to expose C6 roots (Fig. 1 inset). The C6 roots were immersed in mineral oil during the experiments. Because cervical roots were short and partly covered by dura, resulting in an accessible length of 2 4 mm, neural activities were recorded with a custom-designed miniature bipolar electrode as described by Chen et al. [5]. The left C5 superior articular process was removed to the cranial boundary of a tuberosity on the dorsolateral aspect of C5 articular pillar (Fig. l inset). The C5 inferior articular process was then freed from the pedicle underneath without disturbing the vertebral artery or the spinal nerves exiting the foramen. Two holes were drilled 5 mm apart in the freed process, and two stainless steel hooks were inserted to connect the freed process to the testing system. The location of the freed C5 process was adjusted by the actuator using the bony landmarks including the tuberosity as references so that the original length of the capsule could be maintained. Mechunicul tesi setup

Materials and methods .Surgical procedures All surgical procedures have been approved by the Institutional *\nimal Care and Use Committee (IACUC). Nine adult Lamancha o r Alpine female goats (33-55 kg) were used. Anesthesia was induced with diazepam (0.5 mg/kg, im), pentothal (15 mg/kg, iv), butorphenol (0.22 mg/kg, im), and atropine (0.066 mg/kg, im) and maintained with i\oflurane inhalation (2.5-3%). A C2 to T2 midline incision was made, .ind splenius capitus, semispinalis capitus, and multifidus muscles were \eparated from spinous processes and laminae. The ventral rami of C4 C6 nerves were cut to minimize their neural activity in C6 dorsal

The experiments were conducted in a testing frame (Fig. I ) that accommodated a spine fixator, a stereoimaging system, and a computer-controlled Gemini GV6 digital servo actuator (Parker Hannifin Corp., Roherk Park, CA) with a miniature 50 Ib load cell (Entran, Fairfield, NJ). The spine fixator consisted of a hinged strut and an assembly to fix the TI spinous process. The hinged strut was adjustable in length so that the strut angle could be maintained at 30" f 5". The load cell was attached via an adaptor to the two stainless steel hooks. The shaft of the actuator was aligned with the midpoint between the two holes on the C5 articular process. Two Motion Corder Analyzers with SR-500 cameras (8-bit monochrome, 512 x 480 pixels; Kodak. San Diego. CA) were used, each equipped with a Nikkor macro lens (Nikon, Tokyo, Japan).

Fig. 1. Schematic drawing of the test setup. The inset shows the surgical preparation and the positioning of capsular markers. Two holes are shown drilled in the rostra1 end of the freed C5 process to attach to the actuator via two stainless steel hooks.

Y. Lu et al. I Journal of Ortliopaedic Research 23 (2005) 779-787 E.uperiniental procedure

78 1

(A) Loadiug Pattern Actuator

Left C6 dorsal nerve rootlets (&lo) were first screened to find the responsive rootlet to probing on receptive fields on the left C 5 4 6 FJC as described by Lu et al. [17]. Electrical stimulation (0.1 ms duration, 8 or 15 V) was then applied on receptive fields with a bipolar stimulating electrode to obtain conduction velocities (CV) [17]. Once a receptive field was identified, all the subsequent tests were performed with this rootlet. An array of tantalum spheres (0.5 mm diam) was then applied on the FJC using a minimal amount of acrylic paint as adhesive. One axis of the array was aligned with the craniocaudal direction of the goat. The markers were positioned to cover the maximum FJC surface area, with cranial and rostra1 boundary markers less than 1 mm away from the capsule insertion on the articular processes. The arrays were 4 x 4, 5 x 5, or 6 x 6, with intervals of 1-3 mm between markers. Additional markers were positioned on the C5 and C6 articular processes for tracking FJ displacement (Fig. 2). Before testing, the C5-C6 FJC was loaded with an applied displacement of 2 mm at a rate of 0.5 r n d s and unloaded at the same rate. The FJC then underwent a series of stretch tests. Each test had a trapezoidal loading pattern consisting of a displacement ramp, a 10 second hold, and a release ramp (Fig. 3A). A displacement rate of 0.5 m d s was used in both ramps. The first test was 2 mm actuator displacement, with 2 mm increments for subsequent tests until the capsule sustained catastrophic failure (capsular rupture). Nerve discharge was recorded during the test and for 30 s after unloading. An interval of at least 4 min was allowed to download large quantities of images after each test. Similar intervals of 4-5 min have been adopted in knee and hip joint capsule studies to minimize the effect of repetitive capsule loading on afferent response [I 1,221. All neural activity and load data were amplified and recorded on analog tape and were synchronized with image data as described in Lu et al. [17]. Capsule images were captured at a rate of 3 frames per sec. The two cameras were triggered synchronously by a 5 V square puke. Images of a calibration frame (30x20 x 16mm, 15 markers) were captured at the conclusion of the experiment with the frame positioned to include the FJC spatial volume. Data reduction

Nerve discharge data were analyzed using a window discrimination system-Enhanced Graphics Acquisition and Analysis (RC Electronics, Santa Barbara, CA). Single units were identified by their response

Fig. 2. C S C 6 FJC at 6 mm stretch. Also shown are the C5 anchoring hooks, the 5 x 5 array of markers, and two vertebral markers. The joint line is indicated by a dotted line.

Displacement 8 m m P 4 I

t

l ~ n a l

Time

Fig. 3. Loading patterns (A) and force vs. displacement curves (B) for a series of tests. Paths of loading and unloading are illustrated with arrowheads. Stress relaxation in the 10 s hold is shown by a downward arrow. The values of peak load and load reduction are shown for each test.

to electrical stimulation of receptive fields on the capsular surface [17]. These units were categorized by CV into Groups I1 (20-70ds), 111 (2.5-20 d s ) , and IV (0.5-2.5 d s ) [23]. For strain analysis, ImageExpress 3D Wizard (SAI, Utica, NY) was used to track the tantalum spheres to obtain their 3D coordinates using the calibration frame as the reference coordinate system. Linear quadrilateral membrane elements were developed in Hypermesh (Altair, Troy, MI) using the original positions of the markers as nodes. Digitized marker displacement history was imposed on the nodes to reconstruct capsular deformation. LS-DYNA (LSTC, Livermore. CA) was then used to perform strain analysis. Similar techniques have been adopted in various kinematics studies [13,27]. The error involved in marker digitization was evaluated by performing the same procedures on the calibration frame as described by Lu et al. [16]. A finite element model (FEM) was constructed using the known coordinates of the frame as nodes and digitized coordinates as displacement history. The model resulted in no more than 1% principal strain, with 0% strain signaling no error. Displacements across the FJ were obtained by digitizing the C5 and C6 markers on the articular processes and calculating the distance between markers for plotting force vs. displacement curves. These values were less than actuator displacement. The capsular stiffness values were obtained by finding the slope of the linear region of the curve after the toe region. Since toe regions were observed up to displacements of 24 mm, force vs. displacement curves in 8 mm tests were used to calculate stiffness. The stiffness was calculated as the change in force (from peak force to 50% peak force) divided by the change in displacement (from peak displacement to the displacement at 50%" of peak force).

782

Y. Lu et al. /Journal of Orthopaedic Research 23 (2005) 779-787

(6) Capsule s h i n distribution (nine dehed elements) under stretch M

g

0.05 0.02 4mm( 0.12 0.05 6mml 0.17 0.09 8mm 0.19 0.f3 1Omm 0.27 0.27 2mm

I

I

l2mm 14mm 16mm

0.33 0.29 0.41 0.29 0.43 0.32

I I I I

c2

m 0.02 0.10 0.03 0.14 0.06

OM 0.21 027 0.f9 031 0.20 0.36 0.21 0.40 0.25

I 1 I I

0.08

0.06 0.11 0.10 0.15 0.11 0.19 0.12 023

0.14 028 0.17 032 0.19 034 0.18

7 0.12 0.18 0.17 023 0.16 0.26 0.15 0.30 0.19 0.33 0.20 0.42 0.25

0.09

0.24 0.14 0.29 0.17 0.32 0.10 0.35 0.12 0.13 0.19 0.52 0.19

0.14

OdO

0.15

0.07

Od3 0.07

~

Od5

0.10

0.10

0.02 021 0.07 029 0.07

0.09 041

I

0.19

0.05 053

c3 om 0.03 0.16 0.05 022 0.08 0.32 0.10 0 3 0.1f 0.30 0.12 0.48 0.23 0 .!a 0.21

Fig. 4. (A) FEM representation of a C 5 C 6 FJC at 8 mm stretch with the vector plot of principal strains. Outward arrows indicate positive strains, most of which were along the rostral~audaldirection; inward arrows indicate negative values. most of which were along the medial-lateral direction signifying compression. (B) Means and standard deviations of the principal strains from 9 goats are shown for nine defined elements representing capsular strain distribution.

The strain history of a mesh element, where the receptive field was located, was obtained for principal strain. Direction deviations (deg) of principal strains relative to the loading direction were obtained for each element at peak displacement in each test. Deviations were positive if the principal strain was oriented clockwise from the loading direction in the top view (Fig. 4A). Because of varied number and size of FEM elements in FJCs, the mean of the direction deviations in 9 goats was calculated by weighting individual deviations by the element area at peak stretch. In addition, 9 elements were picked to quantitatively represent strain distribution (Fig. 4A) including rostral-medial (Rl), rostralcenter (R2), rostral-lateral (R3), middle-medial (MI), middle-center (M2), middle-lateral (M3), caudal-medial (CI), caudal+enter (C2), and caudal-lateral (C3) elements. The corner and boundary elements on the mesh were picked for this purpose. The middle elements were always along the joint line, while the center elements were centered approximately between the medial and lateral boundaries. The strain threshold of an afferent was defined as the principal strain at its receptive field at a time point when the afferent discharge rate (DR) increased during stretch (Fig. 5). The increase was considered only when the instantaneous DR was at least 1 impulse/s greater

than average baseline DR if baseline DR was continuous or at least 0.2 impulse/s greater if baseline DR was sporadic or non-existent. Average baseline DR was calculated over 30 s before the onset of stretch; the strain sensitivity of an afferent was defined as the slope of the linear regression of its DR and principal strain (above threshold and below saturation); it was calculated as DR increase per 1% strain (impulse/ s/%).Aflerent saturation was defined as no increase in DR with increasing stretch. The suturution strain of an afferent was defined as the local principal strain at which afferent saturation occurred. Stat istical analysis

Strain distribution was statistically analyzed using the specified 9 elements. x2 goodness of fit was used to assess whether the maximum strain occurred more frequently in one area than another. Two-way repeated-measure analysis of variance was applied with stretch levels and capsular areas as independent variables and principal strains as dependent variables. Linear regressions were conducted with capsular load and principal strain as separate independent variables and DR (above threshold and below saturation) as the dependent variable. Indepen-

0

0.-

s


=

15.80 S /Div)

75.88

s

Fig. 5. A C afferent (CV = 1.6 m/s) response to 10 mm stretch. Discharge rate (DR) increased after threshold and reached saturation before maximum stretch.

Y. Lu et al. I Journal of Orthopaedic Research 23 (2005) 779-787

dent t-tests were used to compare strain threshold, saturation strain, and strain sensitivity of afferents between afferent groups. p c 0.05 was considered significant. All statistical analyses were performed in SPSS (SPSS, Chicago, IL). All experimental values are presented as mean f standard deviation.

Results Biomechanical characterization

The capsule stiffness averaged 16.8 k 11.4 N/mm ( n = 9 , Fig. 3B). The load decrease at maximum displacement (10 s hold) increased as greater capsule displacement was applied, with a mean linear regression coefficient of 0.822 _+ 0.060 from 9 goats (all p C 0.05). FJC strain distribution is shown in Fig. 4A at an 8 mm stretch in a female Lamancha goat. Principal strains in most elements were oriented in approximately the same direction as the loading, with the mean direction deviation of 1.8" k 18.2" for 9 goats. Principal shear strains were primarily directed perpendicular to the loading direction. Mean principal strains from 9 goats are shown in Fig. 4B for the 9 elements. No significant difference was found among nine elements, but greater stretch caused significantly higher strains (p < 0.05), which was independent of capsular areas. In addition, no significant pattern for the location of maximum strain was identified using x2 goodness of fit test. FJC tears occurred in the test range of 18-3Omm. Because this study was focused on physiologic and subcatastrophic ranges of stretch, FJC neural response and strain are only presented for tests from 2 to 16 mm. Neurophysiological characterization

Fifty-seven single afferents in 9 goats were analyzed and categorized by CV into 32 Group I11 and 25 Group IV units. In characterizing CV by electrical stimulation, with the short innervation pathway (20-30 mm), Group I1 units tended to fall within the stimulus artifact and were difficult to identify; they are not presented in this study. Seventeen (53%) Group 111 and 14 (56%) Group IV units responded to stretch with low strain thresholds (Table 1). The strain thresholds were 0.107 k 0.033 (n = 17) and 0.100 k 0.046 (n = 14) for Group I11 and IV units, respectively. No significant difference was observed in threshold between groups. Nine of 17 low-strain-threshold Group I11 and 9 of 14 low-strain-threshold Group IV units displayed discharge saturation at various strains (Table 1). The saturation strains were 0.388 f 0.121 for the 9 Group I11 and 0.341 +_ 0.159 for the 9 Group IV units, with no significant difference between the two groups. Two (8%) Group IV units responded only to high strains (0.460 k 0.170),

783

Table 1 Characterizationof Group 111 and IV afferents Group 111 (n = 17)

Group IV ( n = 16)

Strain threshold Mean SD n

Low 0.107 0.033 11

Low 0.100# 0.046 14

High 0.460# 0.17 2

Saturation strain Mean SD n

0.388 0.121

0.341 0.159

NA

9

9

Strain sensitivity (impulselsPA) Mean 0.336 SD 0.406 n 17

0.403 0.475 14

Linear regression Load

Loud

Prin. strain

NA

Prin. strain

(2) Mean SD n

0.737 0.831 0.115 0.121 17 17

0.716 0.817 0.113 0.12 14 14

NA

Note: Pairs of # indicate significant difference (p 0.05). NA-not examined for high-strain-thresholdGroup IV afferents.

which were significantlyhigher than the strain thresholds of the 14 low-threshold Group IV units. Fifteen (47%) Group I11 and 9 (36%) Group IV units could not be excited even by noxious capsular stretch. Seven of these units displayed resting discharges while the other 17 (30% of Group III/IV) were silent or sporadic (DR c 1 impulse per 10 s). The 3 1 low-strain-threshold afferents displayed a wide range of strain sensitivity (Table 1). Group I11 afferents showed strain sensitivity from 0.02 to 1.36 impulse/s/% and Group IV from 0.04 to 1.47 impulse/ s/%. No significant difference was observed between the two groups. The relationships between DR increase, load, and principal strain are shown in Fig. 6 for Group I11 and in Fig. 7 for Group IV units. No significance difference in correlation coefficients was found between groups. The neural correlation to load was less than to principal strain, but with no significant difference in either Group I11 (p = 0.06) or IV (p = 0.07).

Discussion The goat animal model has been used in many cervical spine studies including disk healing and spine fusion [l]. Despite its quadruped nature, the goat holds its head upright, axially loading the neck. The size of accessible regions in the cervical FJ is important to allow quantification of neural discharge and two-dimensionalcapsular strains across the joint, thus discouraging the use of widely-adopted but smaller animal models for neurophysiology, such as rat and cat.

Y. Lu et al. I Journal of Orthopaedic Research 23 (2005) 779-787

784

Group III Afferent Responses (A)

1

50 Discharge Rate Increase

1 401

50 Lapnlsedsec

Impnlsedsec

40

-+-0-

30

- a---A- -

+ 20 --t

--x-- 10

(n=8) 0 0

10

7

20

30 Capsular Load (N)

1

40

1

6 hpalredsec

6 Impulredrec

5

5

Fig. 6. Neural responses of 17 Group I11 mechanoreceptors to stretch. Afferent DR increase was plotted against capsular load and principal strain. Each symbol represents the same afferent in the row of two graphs. Afferents displayed a wide range of DR increase. For the purpose of better illustration, 8 afferents with greater stretch sensitivity are shown in the top row (A) while 9 with less sensitivity in the bottom row (B).

Although a few studies have evaluated the hypothesized role of FJC deformation in whiplash injury, this is the first study to directly demonstrate a quantitative relationship between nerve discharge in cervical FJC and applied stretch. In addition, our study supports the previous findings in cat of the presence of mechanoreceptors and nociceptors in and around cervical FJ [4,251. Around half of Group III/IV afferents that we observed had low strain thresholds and were sensitive to stretch. Schaible and Schmidt reported 31% of Group I11 and 13% of Group IV units in knee joint were activated by non-noxious joint movement [23]. Millar observed that the majority of articular fibers responded to small elbow extensions [18]. We hypothesize that as in these synovial joints, a population of FJC receptors may act as stretch receptors and signal the presence and magnitude of a mechanical stimulus when FJC is deformed or loaded. As in limb joints [22,26], full-range proprioceptive function in spine FJ are probably provided by the coordination of muscle spindles and Golgi

tendon organs in paravertebral muscles along with FJC afferents. Among low-threshold afferents, 53% of Group I11 and 64% of Group IV units showed saturated responses at high strains. Comparatively, 73% of hip joint afferents showed saturation at the extreme internal rotation up to 40" [22]. Tensile saturation (>300 kPa) was observed in Group III/IV gracilis muscle afferents [ 141. Similarly, Garell et al. observed that more than half of Group III/IV afferents in feline skin demonstrated compressive saturation at forces above the human sharpness threshold but below or close to the pain threshold [9]. This population of Group III/IV units in skin was thought to be associated with non-painful sharpness perception [9,lo]. Correspondingly, saturated afferents in our study do not appear to encode stimulus intensity above pain threshold; they may provide signals relevant to nonpainful mechanical "over-stimulation.'' Two (8%) of the Group IV units were observed to have high strain thresholds and were characterized to be nociceptive, compared to 14 (56%) of the Group IV

785

Y. Lu ei al. I Journal of Orthopaedic Research 23 (2005) 779-787

otoup IV AfferentResponses (A)

Discharge Rate Increase

-A-

*

20

* 10 * 0

20 30 capsular Load (N)

10 10

40

0

0.0

0.2

plincipal L

0.4

0.6

m strain

5 4

3 2 1

0 0

-

10 20 30 40 Capsular Load ( I ? )

Fig. 7. Neural responses of 14 Group IV mechanoreceptors to stretch. Seven afferents with greater sensitivity are shown in the top row (A) while 7 with less sensitivity in the bottom row (B).

low threshold units. Several other joint studies also reported that non-nociceptive fine afferents outnumbered nociceptive ones, but at lower rates than the current study. Schaible and Schmidt reported 4'7% of Group 111 and 13% of Group IV were strongly activated by forceful movement against the resistance of knee joints [23]. Rossi and Grigg recorded 12% of hip joint afferents responded to an activation angle of 30" in internal rotation as opposed to 88Yo below 20" [22]. Seventeen (30%) of Group III/IV units in our study were silent or sporadic and did not respond to any tensile stimuli. Schaible and Schmidt reported 22% of Group I11 and 43.5% of Group IV units in medial articular nerve could not be excited by any knee joint movement [23]. Grigg et al. reported that no Group IV units in posterior articular nerve were responsive to even noxious knee joint movement in a normal joint, but 55% of Group IV were responsive under inflammation [12]. These unresponsive afferents are likely nociceptors, which either have very high mechanical threshold or are only activated in inflamed conditions. The mechanosensitivity increase of these afferents in inflamed conditions has been observed in other joint studies [29]. In our study two high-threshold units were activated at 3458% strain while 18 low-threshold units showed saturated responses at 3439%. These strain values fall

within the reported range of subcatastrophic injury (35.0-64.6%), but were well below reported catastrophic failure ranges (94.0-103.6%) in human FJCs [24,27]. They also do not distinctly differ from the strains that lower FJCs experienced during whiplash loading (3560%) [6,19,20]. In addition, from a physiologic point of view, Lee et al. reported that the FJC strain range of 1142% may produce pain symptoms in the rat as evidenced by behavioral sensitivity and astrocytic activation in the spinal cord [15]. Taken together, these findings support the FJC strain mechanism of whiplash injury, which postulates that capsular strain beyond the physiological limit during whiplash results in capsular injury and initial pain. FJC nociceptors that are unresponsive in normal conditions may become activated after capsular injury and contribute to sustained FJ pain. However, further studies are needed at strain rates comparable to those seen in whiplash events. The lack of consistent maximum strain locations on FJCs and the high variability of strains between specimens were also observed in other cervical and lumbar FJ studies [13,27]. This phenomenon may result from individual differences in capsular insertion locations and heterogeneities in capsular thickness. Goat C5-C6 FJCs displayed a mean stiffness of 16.8 k 11.4N/mm at a loading rate of 0.5 m d s , which

786

Y. Lu et al. I Journal of Orthopaedic Research 23 (2005) 779-787

is comparable to the value of 19.6 k 11.9 N/mm measured for human cervical FJCs at 0.0083 mm/s [27]. In contrast, other dynamic studies showed much higher stiffness, with a mean of 35 N/mm at 10 mm/s [30] and 35.4 N/mm at 100 mmls [27] for human cervical FJCs. In addition, the load relaxation displayed a significant linear correlation with capsular displacement, consistent with the Kelvin model of a viscoelastic solid during stress relaxation [8]. This study examined the neural response of the cervical FJC to low-rate loading. Compared to the FJC strain rate of around 3OOn/n/~during simulated whiplash [20] and the quasistatic rates of 0.0083 or 0.01 mm/s used in cadaveric studies [24,27], our strain rate of approximately 5 % / ~probably better represents the physiological strain rate in cervical spine movement in daily life. Our study demonstrated the possible physiological and nociceptive functions of FJC afferents in nonimpact conditions. The potential roles of these afferents in whiplash injury remain to be examined under dynamic loading and under conditions of injury and inflammation. Comparisons can provide critical insight into afferent behaviors during physiologic and higher acceleration movements. In summary, we observed that fine afferents in the goat cervical FJC were excited by various strain levels. Low-strain-threshold afferents displayed graded sensitivity to stretch with saturated response at high strains while high-strain-threshold and unresponsive receptors could act as nociceptors. These findings may shed light on cervical spine function in physiologic and noxious environments. Further studies addressing muscle response and dynamic loading are warranted.

Acknowledgement This work was supported by CDC Grants # R49CCR519751 and # R49-CEO00455 and a Ford Biomedical Engineering Graduate Fellowship provided by Ford Motor Company to the first author. The technical assistance of Amrita Vempati and Anita Singh is gratefully acknowledged.

References [I] Baisden J, Voo LM, Cusick JF, et al. Evaluation of cervical laminectomy and laminoplasty. A longitudinal study in the goat model. Spine 1999;24:1283-8. [2J Barnsley L, Lord SM, Wallis BJ, Bogduk N. The prevalence of chronic cervical zygapophysial joint pain after whiplash. Spine 1995;20:20-5. [4] Bolton PS, Holland CT. An in vivo method for studying afferent fibre activity from cervical paravertebral tissue during vertebral motion in anaesthetised cats. J Neurosci Methods 1998;85: 21 1-8.

[5] Chen C, Lu Y,Cavanaugh JM, et al. Recording of neural activity from goat cervical facet joint capsule using custom-designed miniature electrodes. Spine, in press. [6] Deng B, Begeman PC, Yang KH, et al. Kinematics of human cadaver cervical spine during low speed rear-end impacts. In: Proc of 44th Stapp Car Crash conference 2000. p. 171-88. [7] Dwyer A, April1 C, Bogduk N. Cervical zygapophyseal joint pain patterns. I: A study in normal volunteers. Spine 1990;15:453-7. [8] Fung YC. Biomechanics-mechanical properties of living tissues. Springer-Verlag; 1981. [9] Garell PC, McGillis SL, Greenspan JD. Mechanical response properties of nociceptors innervating feline hairy skin. J Neurophysiol 1996;75:1177-89. [lo] Greenspan JD, McGillis SL. Thresholds for the perception of pressure, sharpness, and mechanically evoked cutaneous pain: effects of laterality and repeated testing. Somatosens Mot Res 1 994; 11:31 1-7. [I I] Grigg P, Hoffman AH. Properties of Ruffini afferents revealed by stress analysis of isolated sections of cat knee capsule. J Neurophysiol 1982;47:41-54. [I21 Grigg P, Schaible HG, Schmidt RF. Mechanical sensitivity of Group I11 and IV afferents from posterior articular nerve in normal and inflamed cat knee. J Neurophysiol 1986;55:63543. [13] Ianuzzi A, Little JS, et al. Human lumbar facet joint capsule strains: I. During physiological motions. Spine J 2004,4 141-52. [I41 Khalsa PS, Ge W. Encoding of tensile stress and strain during stretch by muscle mechano-nociceptors. Muscle Nerve 2004;30 21G24. [I 51 Lee KE, Davis MB, Mejilla RM, Winkelstein BA. In vivo cervical facet capsule distraction: mechanical implications for whiplash and neck pain. In: Proc of 48th Stapp Car Crash conference, 2004. [I61 Lord SM, Barnsley L, Wallis BJ, et al. Percutaneous radiofrequency neurotomy for chronic cervical zygapophyseal-joint pain. N Engl J Med 1996;335:17214. [17] Lu Y, Chen C, Kallakuri S, et al. Development of an in vivo method to investigate biomechanical and neurophysiological properties of spine facet joint capsules. Eur Spine J, doi: 10.10071 SOO586-004-0835-9. [I81 Milbar J. Flexion-extension sensitivity of elbow joint afferents in cat. Exp Brain Res 1975;24:209-14. [I91 Panjabi MM, Cholewicki J, Nibu K, et al. Capsular ligament stretches during in vitro whiplash simulations. J Spinal Disord 1998;11:227-32. [20] Pearson AM, Ivancic PC, Ito S, Panjabi MM. Facet joint kinematics and injury mechanisms during simulated whiplash. Spine 20O4;29:390-7. [21] Pickar JG, McLain RF. Responses of mechanosensitive afferents to manipulation of the lumbar facet in the cat. Spine 1995;20: 2379-85. [22] Rossi A, Grigg P. Characteristics of hip joint mechanoreceptors in the cat. J Neurophysiol 1982;47:1029-42. [23] Schaible HG, Schmidt RF. Responses of fine medial articular nerve afferents to passive movements of knee joints. J Neurophysiol 1983;49:1118-26. [24] Siegmund GP, Myers BS, Davis MB, et al. Human cervical motion segment flexibility and facet capsular ligament strain under combined posterior shear, extension and axial compression. In: Proc of 44th Stapp Car Crash conference, 2000. p. 15970. [25] Thunberg J, Hellstrom F, Sjolander P, et al. Influences on the fusimotor-muscle spindle system from chemosensitive nerve endings in cervical facet joints in the cat: possible implications for whiplash induced disorders. Pain 2001;91:15-22. [26] Williams WJ. A systems-oriented evaluation of the role of joint receptors and other afferents in position and motion sense. Crit Rev Biomed Eng 1981;7:23-77.

Y. Lu et ul. I Journul of Orthopaedir Reseurch 23 (2005) 779-787

[27] Winkelstein BA, Nightingale RW, Richardson WJ, Myers BS. Cervical facet joint mechanics: its application to whiplash injury. In: Proc of 43rd Stapp Car Crash conference, 1999. p. 243-52. [28] Yamashita T, Cavanaugh JM, El Bohy AA, et al. Mechanosensitive afferent units in the lumbar facet joint. J Bone Joint Surg Am 1990;72:865-70.

787

[29] Yamashita T, Cavanaugh JM, Ozaktay AC, et al. Effect of substance P on mechano-sensitive units of tissues around and in the lumbar facet joint. J Orthop Res 1993;11:205-14. [30] Yoganandan N, Pintar FA, Kumaresan S, Elhagediab A. Biomechanical assessment of human cervical spine ligaments. In: Proc of 42nd Stapp Car Crash conference, 1998.