- Email: [email protected]
Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine
Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine Kevin M. Bell a,n, Yiguo Yan a, Richard E. Debski b, Gwendolyn A. Sowa c, James D. Kang a, Scott Tashman a a
Ferguson Laboratory for Spine Research, Department of Orthopaedic Surgery, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA Department of Bioengineering, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA c Department of Physical Medicine and Rehabilitation, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA b
art ic l e i nf o
a b s t r a c t
Article history: Accepted 21 November 2015
The human cervical spine supports substantial compressive load in-vivo arising from muscle forces and the weight of the head. However, the traditional in-vitro testing methods rarely include compressive loads, especially in investigations of multi-segment cervical spine constructs. Various methods of modeling physiologic loading have been reported in the literature including axial forces produced with inclined loading plates, eccentric axial force application, follower load, as well as attempts to individually apply/model muscle forces in-vitro. The importance of proper compressive loading to recreate the segmental motion patterns exhibited in-vivo has been highlighted in previous studies. However, appropriate methods of representing the weight of head and muscle loading are currently unknown. Therefore, a systematic comparison of standard pure moment with no compressive loading versus published and novel compressive loading techniques (follower load – FL, axial load – AL, and combined load – CL) was performed. The present study is unique in that a direct comparison to continuous cervical kinematics over the entire extension to flexion motion path was possible through an ongoing intrainstitutional collaboration. The pure moment testing protocol without compression or with the application of follower load was not able to replicate the typical in-vivo segmental motion patterns throughout the entire motion path. Axial load or a combination of axial and follower load was necessary to mimic the in-vivo segmental contributions at the extremes of the extension-flexion motion path. It is hypothesized that dynamically altering the compressive loading throughout the motion path is necessary to mimic the segmental contribution patterns exhibited in-vivo. Published by Elsevier Ltd.
Keywords: Cervical spine Kinematics Range of motion ROM Follower load Compressive preload Hybrid control Robotics Bi-plane radiography In-vivo
1. Background In-vitro biomechanical testing has been critical in the design and evaluation of surgical instrumentation. Determination of realistic physiologic loading levels for the cervical spine has, however, proven difficult outside of the in-vivo setting. Unconstrained pure moment testing combined with the hybrid testing method is currently the gold standard test protocol for evaluation of motion preservation technology and adjacent level effects. Pure moment testing was specifically designed to apply uniform loading at each cross section throughout the length of a spinal construct, permitting irregularities to be identified (Panjabi, 1988). Pure moment testing is well suited for making relative comparisons between treatments, but is currently not based on or representative of in-vivo motion (Panjabi, 2007). n
Corresponding author. Tel.: þ 1 412 657 7857; fax: þ1 412 648 8548. E-mail address: [email protected] (K.M. Bell).
Additionally, the human cervical spine supports substantial compressive load in-vivo arising from muscle forces and the weight of the head. However, the traditional in-vitro testing methods rarely include compressive loads; especially in investigations of multi-segment cervical spine constructs. Various methods of modeling physiologic loading have been reported in the literature including axial forces produced with inclined loading plates, eccentric axial forces application, follower load, as well as attempts to individually apply/model muscle forces in-vitro (Adams and Dolan, 2005; Cook, 2009; Cripton et al., 2000; DiAngelo and Foley, 2004; Goel et al., 2006; Miura et al., 2002; Panjabi, 1988, 2007; Panjabi et al., 2001; Patwardhan et al., 2000; Wilke et al., 1994, 2001, 1998). Miura et al. (2002) and DiAngelo and Foley (2004) published articles directly aimed at determining the most appropriate loading mechanism to produce physiologic motion patterns. Miura et al. (2002) presented pure moment testing combined with follower load and, through adjusting moment targets, was able to
http://dx.doi.org/10.1016/j.jbiomech.2015.11.045 0021-9290/Published by Elsevier Ltd.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
achieve 20% agreement with segmental range of motion reported in literature, however typical segmental motion patterns were not observed with this technique. DiAngelo and Foley utilized an eccentric axial compressive method in attempt to mimic the weight of the head. DiAngelo and Foley (2004) was able to show reasonable agreement with the segmental motion patterns, but the magnitudes dramatically underestimated the average in-vivo segmental kinematics. The objective of this project is to identify and verify the appropriate in-vitro loading conditions that would replicate the in-vivo kinematics of the cervical spine, with the overall goal of improving the biofidelity of the experimental platform. A systematic comparison of standard pure moment with no compressive loading versus published and novel compressive loading techniques (follower load – FL, axial load – AL, and combined load – CL) was performed. It is hypothesized that an optimized follower load, passing through the segmental centers of rotation, will add stability to the system but will not dramatically affect the segmental motion patterns observed throughout the extension to flexion motion path. In contrast, axial load applied perpendicular to superior most vertebral body, will not maintain the pure moment assumption, whereby enabling the segmental motion patterns to be altered. 2. Methods 2.1. Protocol N ¼12 fresh-frozen human (C3-7) cervical cadaveric specimens (51.8 years7 7.3) were pre-screened with CT and dissected, preserving osteoligamentous structures. Specimens were mounted in a robot-based spine testing system, consisting of a serial linkage robotic manipulator (Staubli RX90, Staubli Inc., Duncan, SC) with an on-board six-axis load cell (UFS Model 90M38A-150, JR3 Inc., Woodland, CA) and custom specimen-mounting fixtures (Bell et al., 2007; Gilbertson et al., 1999; Hartman et al., 2009). Four clinical lateral mass screws were used to secure the specimens to the mounting fixtures (one in each pedicle and two in the anterior portion of the vertebral body). After mounting, specimens were wrapped in 0.9% saline soaked gauze and periodically sprayed with saline in order to prevent dehydration. The robot was controlled via MATLAB (Mathworks, Inc.) and operates under adaptive displacement control to a pure moment target of 2.0 N m for flexion and extension (FE) for each state in a randomized order (no compression (Fig. 1A), follower load, axial rotation, combined loading). Due to the quasi-static nature of the adaptive displacement control algorithm the system operated at a rate of 0.0677 0.0014°/s. Two consecutive full extension to flexion loops were performed with the data from the second cycle being presented to account for preconditioning (Cripton et al., 2000). Segmental motion was recorded using a five camera VICON
system tracking passive reflective markers rigidly attached as a marker group to each vertebral body. A hand held VICON digitizer was utilized to digitize the anatomical coordinate system for each vertebral body relative to the marker group and the Euler angle rotations of C34, C45, C56, and C67 were determined and reported. 2.2. Follower load Follower load application was accomplished by loading the specimen with bilateral cables passing through cable guides inserted into the vertebral bodies and over pulleys attached to the base (Fig. 1B). A novel active system was implemented in our laboratory using linear actuators coupled with load cells. Control of the system was integrated with the custom-built PC-based control program written in MATLAB that is currently used to control the robot testing system, and enabled active control of the loading throughout the motion path. The follower load system (Fig. 1B) consists of two independently controlled 24 V servo motor linear actuators (Ultramotion – 3-B.125-DC426_24-4-/4) and compression/tension load cells (Transducer Techniques – MLP-100). A Galil Motion Controller (DMC-4183-BOX8 (-16BIT)-D3040-D4040) controlled this system using on-board closed loop Proportional-Integral-Derivative (PID) load control. The 3/64‶ diameter stainless steel wire rope lanyard was threaded through a custom designed adjustable cable guide system attached with clinical pedicle screws to enable the follower load cable to interface with the specimen in a manner consistent with the design criterion: (a) tangent to the curvature of the spine and (b) pass through the specimen’s center of rotation (COR). Optimization of the follower load path to align with the specimen’s COR was accomplished through an offline iterative feedback process using the moment output of the testing system’s on-board six-axis load cell. With the specimen in the neutral position, 100 N of follower load was applied to the specimen and resulting change in moments was recorded. Follower load magnitude of 100 N was chosen as it is representative of the most common follower load magnitude presented in literature (Cho et al., 2010; Finn et al., 2009, 2011; Lee et al., 2011; Martin et al., 2011; Paxinos et al., 2009; Snyder et al., 2007). The position of the cable guide was then adjusted to counteract the moment change and the process was repeated until less than 0.1 N m change in moment was observed. Preliminary testing of the described optimization process was performed ensuring that the previously described maintenance of segmental curvature angle criteria was upheld (Patwardhan et al., 2003). 2.3. Axial load Although less popular than follower load as a method to apply compressive load due to published instability issues with this testing method, some authors believe axial loading to be the most physiologic loading scheme—mimicking head weight (DiAngelo and Foley, 2004). Axial loading can be applied along an axis locally fixed to the specimen or globally fixed to the world coordinate system. Previous reports have shown that the cervical spine buckles at very low loads when an axial load is applied globally, therefore for this study the axial load was applied along an axis locally fixed to the specimen (perpendicular to the robot end effector – Fig. 1C). The axial load was applied using the robotic arm to a load target of 50 N (DiAngelo and Foley, 2004), representative of the approximate weight of the head, using the adaptive displacement control algorithm enabling the load to be applied purely in the axial direction and be maintained throughout the flexion–extension rotation path.
Fig. 1. Schematic of the four loading states implemented in this study, (A) no compression, (B) follower load, (C) axial load rotated state showing perpendicular line of action depicted for clarity and (D) combined load.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 2.4. Combined load Simulation of muscle loading and the weight of the head with the follower load and axial compressive load have been independently shown to result in modest improvements in making the in-vitro motion more physiologic (DiAngelo and Foley, 2004; Miura et al., 2002). In this study, the two loading schemes were combined. It was hypothesized that the combination of axial and follower loading will have a synergistic effect, producing more physiologic kinematics (Fig. 1D). 2.5. In-vivo data The previously reported in-vivo data set (Anderst et al., 2013a, 2013b) utilized in the study was reanalyzed for direct comparison with the current in-vitro data. In-vivo data consisted of N ¼20 asymptomatic control patients (45.5 years7 5.8) consented to participate in an Institutional Review Board (IRB) approved protocol. Subjects performed continuous, full ROM flexion–extension at a rate of one complete cycle every 3 s. Subject-specific bone models of C3-C7 were created from CT scans. A previously validated tracking process determined three-dimensional vertebral position with sub-millimeter accuracy by matching bone models from the CT scan to the biplane X-rays (Anderst et al., 2011). 2.6. Data analysis Results of the in-vitro testing at full range of motion (ROM) were summarized and expressed in bar graphs representing mean 7 95% confidence intervals. A Shapiro–Wilk test was performed, which indicated that not all groups were normally distributed, therefore a non-parametric Wilcoxon signed-rank test for paired design was used to identify significant (*p o 0.05) differences between loading modes (no compression versus follower load, axial load and combined load). The in-vivo and in-vitro ROM data was normalizing to percent ROM. The quality of fit between in-vivo and in-vitro motion was assessed by calculating root mean square error (RMSE) at 20% increments of percent ROM.
3. Results Fig. 2 displays the segmental contributions at 100% of the overall extension to flexion motion path. At 100% ROM the segmental distribution for no compressive load state exhibited approximately equal segmental contribution from each level. Adding follower load only had minimal effects on the segmental contribution from each level. However, adding the axial load significantly reduced the C67 contribution to the motion path (p ¼0.004), shifting the majority of the motion toward the middle (C4-5 and C5-6) segments. This significant reduction was also observed in the combined loading state (p ¼0.019). The overall extension to flexion ROM was not affected by the application of a compressive load, with all four compressive loading states exhibiting virtually identical ROM (Fig. 3A). However, although not statistically significant (p ¼0.14), the in-vitro ROM was on average 12.3% smaller than the in-vivo data set. Therefore, in order to directly compare the segmental contributions between the in-vivo and in-vitro data the in-vivo segmental rotation was uniformly scaled by a factor of 87.7% (Fig. 3B). The scatter plot of the in-vitro segmental rotation plotted versus percent of total ROM (Fig. 4) demonstrates clearly that the adding compressive load effects the segmental motion distribution. A small change is observed with application of follower load, wherein C34 and C45 appear to contribute more than C56 and C67 throughout middle portion of the extension to flexion curve. This same change is further magnified with the application of an axial compressive load. The axial compressive loading state and the combined loading curves show a very similar pattern with the upper segments appearing to be recruited from superior to inferior as the motion path progresses. The average RMSE between the in-vivo data and in-vitro data increased with the application of compressive loading. However, the RMSE was dependent on percentage ROM (Fig. 5A) and segmental level (Fig. 5B). No compression had the lowest RMSE at
3
20%, 40% and 60% ROM, which corresponds to the extension and neutral region of the overall path of motion. However, the addition of axial and combined loading resulted in the lowest RMSE at 80% and 100% ROM, which corresponds to the flexion region of the motion path (Fig. 5, Table 1). No compression and follower load had the lowest RMSE at C34 and C56, however axial load and combined load had the lowest RMSE at C45 and C67.
4. Discussion This study systematically evaluated standard pure moment with no compressive loading versus published and novel compressive loading techniques (follower load, axial load, and combined load). Consistent with previous in-vitro reports, the pure moment testing protocol with no compression was not able to replicate the typical in-vivo segmental motion pattern reported in literature (DiAngelo and Foley, 2004; Miura et al., 2002). Although variability existed from specimen to specimen due to local degenerative differences between specimens, overall applying a pure moment at each segment resulted in approximately equal segmental contribution to the overall extension to flexion motion path. Adding follower load allowed the specimen to be tested at a physiologic loading magnitude, while maintaining the pure moment assumption. However, this loading scheme was still not able to replicate the in-vivo segmental motion patterns near the end range of motion. The follower load state in the present study reproduced flexion–extension parameters used in the Miura et al. (2002) study. Although, similar trends are observed between the two studies in terms of percent contribution, the ROM magnitude observed at 2.0 N m in the Miura study was approximately 20% more than the ROM observed in this study. Differences in specimen grade and testing methods could account for some of this disagreement; however in both cases a larger moment target would have been necessary to replicate the magnitude of the reported in-vivo kinematics. Additionally, the Miura et al. study only reported end-range kinematics, which is the predominate trend in literature, however the end-range kinematics are not necessarily representative of the mid-range cervical kinematics and are highly variable (Dvorak et al., 1988; Frobin et al., 2002; Reitman et al., 2004; Wu et al., 2007). High variability was also a limitation of the present study, although attempts were made to limit the variability by controlling recruitment age and pre-screening subjects and specimens. The present study is unique in that a direct comparison to continuous cervical kinematics over the entire extension to flexion motion path was possible through an ongoing intra-institutional collaboration. As hypothesized, application of an axial load perpendicular to the superior most vertebral body altered the segmental motion patterns. The most dramatic differences were
Fig. 2. Average segmental kinematics at 100% of the full extension to full flexion motion path for each loading mode (*p o0.05).
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
4
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
observed in the middle portion of the curve (the neutral zone). Very few authors have reported segmental distribution data at sub-maximal points on the motion curves, making this data difficult to interpret (Wu et al., 2010). Further investigation is necessary to understand this fully, however it is theorized that the
Fig. 3. (A) Overall extension-to-flexion ROM for in vivo and each in vitro loading scheme. P-values are for differences between the in-vivo and in-vitro data. (B) Scaled in-vivo segmental motion path plotted versus the percentage of the overall range of motion (full extension to full flexion for C3-C7).
inherent laxity of the neutral zone combined with the lordotic curvature of a healthy cervical spine account for these dramatic shifts in the segmental motion pattern when axial load is applied. Although the changes in segmental motion pattern were less pronounced at full range of motion, axial load and combined loading were able to produce a motion pattern similar to in-vivo at
Fig. 5. RMS errors between in-vivo and in-vitro motion for each loading condition. Errors varied with both: (A) Percentage of the overall range of motion; and (B) Segmental level.
Fig. 4. In-vitro segmental motion paths plotted versus the percentage of the overall range of motion (full extension to full flexion for C3-C7), showing the effects of altered compressive loading.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Table 1 Scaled in-vivo (deg.) % ROM 20% 40% 60% 80% 100%
C34 1.7 7 0.3 4.1 70.6 6.8 7 0.9 9.5 7 1.1 11.9 7 1.2
C45 2.4 7 0.3 5.2 7 0.5 8.17 0.7 10.9 70.9 13.4 71.1
C56 2.7 70.6 5.17 1.1 7.3 7 1.5 9.5 71.9 11.9 7 2.2
C67 2.4 7 0.8 4.2 7 1.3 5.7 7 1.7 7.4 7 2.0 9.5 7 2.3
No compression (deg.) % ROM C34 20% 2.7 7 0.8 40% 5.5 7 1.6 60% 7.6 7 2.1 80% 9.1 72.2 100% 11.2 7 2.4 RMSE 1.1
C45 2.3 7 0.3 5.3 7 0.8 7.7 7 1.3 9.6 7 1.4 11.7 7 1.6 1.7
C56 1.8 7 0.7 3.8 71.0 6.4 71.6 9.17 2.0 11.5 7 2.4 0.9
C67 2.4 7 0.8 4.2 7 1.0 6.5 7 1.3 9.7 7 1.8 12.3 7 2.2 3.0
RMSE 0.7 1.0 0.7 1.3 1.7 1.3
Follower load (deg.) % ROM C34 20% 3.0 7 0.9 40% 6.2 7 1.7 60% 8.3 7 2.0 80% 9.6 7 2.0 100% 11.2 7 2.1 RMSE 1.5
C45 2.3 7 0.4 5.6 7 0.6 8.17 0.9 9.7 7 0.9 11.4 7 1.0 2.0
C56 1.9 7 0.6 3.5 71.1 6.2 71.9 9.17 2.4 12.0 7 2.9 1.0
C67 2.0 7 0.6 3.3 7 0.9 5.4 7 1.2 9.0 7 2.2 12.17 2.8 2.8
RMSE 0.8 1.4 1.0 1.0 1.7 1.4
Axial load % ROM 20% 40% 60% 80% 100% RMSE
(deg.) C34 3.4 7 0.9 7.8 7 1.9 9.7 7 2.2 10.6 7 2.3 11.7 7 2.4 2.3
C45 1.9 7 0.5 5.4 7 0.8 9.2 7 1.5 10.8 71.6 12.1 71.6 1.4
C56 1.6 7 0.7 2.5 70.9 5.5 71.7 9.4 72.6 11.9 7 2.9 1.5
C67 2.3 7 1.1 2.9 7 1.1 3.6 7 0.8 6.5 7 1.2 10.9 72.1 1.9
RMSE 1.0 2.4 2.1 0.7 1.0 1.6
Combined % ROM 20% 40% 60% 80% 100% RMSE
load (deg.) C34 3.7 7 1.0 7.4 7 1.6 9.5 7 1.7 10.6 7 1.8 11.5 7 2.0 2.2
C45 2.0 7 0.4 5.8 7 0.9 9.4 7 1.2 11.17 1.2 12.1 71.2 1.4
C56 1.5 7 0.7 2.4 70.9 5.3 71.8 8.9 72.8 11.6 7 3.0 1.7
C67 1.8 7 0.6 2.4 7 0.8 3.17 1.0 5.8 7 1.5 10.3 72.5 1.8
RMSE 1.2 2.3 2.2 1.0 0.8 1.6
the extremes of the extension-to-flexion motion path. This observation is consistent with the study published by DiAngelo and Foley (2004), which utilized an eccentric axial compressive method in attempt to mimic the weight of the head. DiAngelo and Foley (2004) were able to show reasonable agreement with the segmental motion patterns, but the reported magnitudes dramatically underestimated the average in-vivo segmental kinematics. Although, the two methods share similarities, an eccentric axial loading method has limited stability, an issue that is corrected in the present study with the combined loading methodology. The fact that axial compressive and/or combined loading resulted in the lowest RMSE at 80% and 100% ROM is interesting in light of the clinical scenario. The need for axial compressive load may be reflective of the in-vivo influence of increased muscular contribution with increasing rotation. This is consistent with muscular models of the cervical spine which have shown that the resultant load experienced on the head increases with increasing rotation (Cheng et al., 2008; Johnston et al., 2008; Schuldt, 1988). In the present study the AL force was applied perpendicular to the superior most vertebral body (C3) which mimics the resultant muscular force on the head as it is transmitted to the vertebral column through the occiput. This is consistent with the inverse dynamic model proposed by Anderst et al. that was based on muscle driven in-vivo kinematics (Anderst et al., 2013c). The muscle loading discussion resonates with Patwardhan et al. (2000) conclusion that the stability provided by application of a follower load “may suggest a mechanism of action by which
5
muscles stabilize the cervical spine under physiologic compression”. In the present study, follower load was shown to be beneficial in the neutral zone for added stability, whereas an axial compressive load was required in the elastic zone for proper segmental distribution. Taking this beyond the mechanism of muscle loading, this distinction is interesting when considering the cervical spine osteoligamentous structures. Throughout the neutral zone, the stiffness of the construct is provided primarily from the intervertebral disc. However, with increasing flexion the ligamentous components are recruited and the instantaneous axis of rotation has been shown to shift anteriorly (DiAngelo and Foley, 2004). It is theorized that the physiologic loading may linked to the anatomy and therefore may be dependent upon the flexion– extension moment-rotation curve. The present study had several limitations. The rate of the invitro testing system was at least an order of magnitude slower than the rate at with the in-vivo data was recorded, due to the quasi-static testing algorithm. This discrepancy will affect the kinematics, but the influence should be minimized due to the invitro preconditioning protocol. Additionally, this study focused on flexion and extension due to the limitations of the follower load methodology, therefore future work is necessary to determine the applicability of the present methodologies to axial rotation and lateral bending of the cervical spine. Finally, the current study applied the compressive load across a multi-level cervical spine construct. Although the results indicate that this methodology enabled the segmental motion patterns to be altered, it is possible that direct manipulation of the individual motion segments is necessary to fully simulate physiologic loading. Despite the limitations of this study, future investigations focused on optimizing the axial load and combined loading methods for replication of in-vivo cervical spine kinematics are recommended. It is hypothesized that dynamically altering the compressive loading throughout the motion path is necessary in order to mimic the segmental contribution patterns exhibited invivo. However, it should be noted this study is only the first step in the process to develop a new testing protocol to mimic in-vivo kinematics in-vitro and will also serve as the foundation for future in-vitro experimentation where physiologic kinematics/kinetics are critical to the clinical relevance of the data such as determination of native biomechanical properties, establishment/validation of computational models, and as a loading scheme for mechanobiology experiments.
Confict of interest We wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome.
Acknowledgments In-vitro testing was performed with the support of the Albert B. Ferguson, Jr. MD Orthopedic Fund of The Pittsburgh Foundation. In vivo data was acquired with support from the National Institutes of Health (Grant # R03 AR056265) and the Cervical Spine Research Society's Society 21st Century Development Grant. I would also like to acknowledge Bill Anderst for valuable contributions to the content of this manuscript. Funding from NIH/NCCAM K08AT00471802 award is also acknowledged.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
References Adams, M.A., Dolan, P., 2005. Spine biomechanics. J. Biomech. 38, 1972–1983. Anderst, W.J., Baillargeon, E., Donaldson 3rd, W.F., Lee, J.Y., Kang, J.D., 2011. Validation of a noninvasive technique to precisely measure in vivo threedimensional cervical spine movement. Spine 36, E393–E400. Anderst, W.J., Donaldson 3rd, W.F., Lee, J.Y., Kang, J.D., 2013a. Cervical motion segment percent contributions to flexion–extension during continuous functional movement in control subjects and arthrodesis patients. Spine 38 (9), E533-9. Anderst, W.J., Donaldson, W.F., Lee, J.Y., Kang, J.D., 2013b. Cervical spine intervertebral kinematics with respect to the head are different during flexion and extension motions. J. Biomech. 46, 1471–1475. Anderst, W.J., Donaldson, W.F., Lee, J.Y., Kang, J.D., 2013c. Subject-specific inverse dynamics of the head and cervical spine during in vivo dynamic flexion– extension. J. Biomech. Eng. 135, 61007–61008. Bell, K.M., Gilbertson, L., Cook, H., Arthur, A., Doehring, T., Muriuki, M., Smolinski, P., Kang, J.D., 2007. Investigation of the integrated passive, active, and control subsystems of the spine utilizing a robotics-based spine testing system. In: Proceedings of the 53th Annual Meeting of the Orthopaedic Research Society (ORS), San Diego, CA. Cheng, C.H., Lin, K.H., Wang, J.L., 2008. Co-contraction of cervical muscles during sagittal and coronal neck motions at different movement speeds. Eur. J. Appl. Physiol. 103, 647–654. Cho, B.Y., Lim, J., Sim, H.B., Park, J., 2010. Biomechanical analysis of the range of motion after placement of a two-level cervical ProDisc-C versus hybrid construct. Spine 35, 1769–1776. Cook, D.J., 2009. Characterization of the Response of the Cadaveric Human Spine to Loading in a Six-Degree-of-Freedom Spine Testing Apparatus. University of Pittsburgh. School of Engineering, Pittsburgh, PA. Cripton, P.A., Bruehlmann, S.B., Orr, T.E., Oxland, T.R., Nolte, L.P., 2000. In vitro axial preload application during spine flexibility testing: towards reduced apparatusrelated artefacts. J. Biomech. 33, 1559–1568. DiAngelo, D.J., Foley, K.T., 2004. An improved biomechanical testing protocol for evaluating spinal arthroplasty and motion preservation devices in a multilevel human cadaveric cervical model. Neurosurg. Focus 17, E4. Dvorak, J., Froehlich, D., Penning, L., Baumgartner, H., Panjabi, M.M., 1988. Functional radiographic diagnosis of the cervical spine: flexion/extension. Spine 13, 748–755. Finn, M.A., Brodke, D.S., Daubs, M., Patel, A., Bachus, K.N., 2009. Local and global subaxial cervical spine biomechanics after single-level fusion or cervical arthroplasty. Eur. Spine J. 18, 1520–1527. Finn, M.A., Samuelson, M.M., Bishop, F., Bachus, K.N., Brodke, D.S., 2011. Two-level noncontiguous versus three-level anterior cervical discectomy and fusion: a biomechanical comparison. Spine 36, 448–453. Frobin, W., Leivseth, G., Biggemann, M., Brinckmann, P., 2002. Sagittal plane segmental motion of the cervical spine. A new precision measurement protocol and normal motion data of healthy adults. Clin. Biomech. 17, 21–31. Gilbertson, L.G., Doehring, T.C., Livesay, G.A., Rudy, T.W., Kang, J.D., Woo, S.L.Y., 1999. Improvement of accuracy in a high-capacity, six degree-of-freedom load cell: application to robotic testing of musculoskeletal joints. Annu. Biomed. Eng. 27, 839–843. Goel, V.K., Panjabi, M.M., Patwardhan, A.G., Dooris, A.P., Serhan, H., 2006. Test protocols for evaluation of spinal implants. J. Bone Jt. Surg. Am. 88 (Suppl 2), S103–S109. Hartman, R.A., Bell, K.M., Kang, J.D., 2009. Analyses of the Components of the Posterior Column in a Distractive-Flexion Injury Model. In: Proceedings of the 55th Annual Meeting of the Orthopaedic Research Society (ORS), Las Vegas, NV.
Johnston, V., Jull, G., Souvlis, T., Jimmieson, N.L., 2008. Neck movement and muscle activity characteristics in female office workers with neck pain. Spine 33, 555–563. Lee, S.H., Im, Y.J., Kim, K.T., Kim, Y.H., Park, W.M., Kim, K., 2011. Comparison of cervical spine biomechanics after fixed- and mobile-core artificial disc replacement: a finite element analysis. Spine 36, 700–708. Martin, S., Ghanayem, A.J., Tzermiadianos, M.N., Voronov, L.I., Havey, R.M., Renner, S.M., Carandang, G., Abjornson, C., Patwardhan, A.G., 2011. Kinematics of cervical total disc replacement adjacent to a two-level, straight versus lordotic fusion. Spine 36, 1359–1366. Miura, T., Panjabi, M.M., Cripton, P.A., 2002. A method to simulate in vivo cervical spine kinematics using in vitro compressive preload. Spine 27, 43–48. Panjabi, M.M., 1988. Biomechanical evaluation of spinal fixation devices: I. A conceptual framework. Spine 13, 1129–1134. Panjabi, M.M., 2007. Hybrid multidirectional test method to evaluate spinal adjacent-level effects. Clin. Biomech. 22, 257–265. Panjabi, M.M., Miura, T., Cripton, P.A., Wang, J.L., Nain, A.S., DuBois, C., 2001. Development of a system for in vitro neck muscle force replication in whole cervical spine experiments. Spine 26, 2214–2219. Patwardhan, A.G., Havey, R.M., Carandang, G., Simonds, J., Voronov, L.I., Ghanayem, A.J., Meade, K.P., Gavin, T.M., Paxinos, O., 2003. Effect of compressive follower preload on the flexion–extension response of the human lumbar spine. J. Orthop. Res.: Off. Publ. Orthop. Res. Soc. 21, 540–546. Patwardhan, A.G., Havey, R.M., Ghanayem, A.J., Diener, H., Meade, K.P., Dunlap, B., Hodges, S.D., 2000. Load-carrying capacity of the human cervical spine in compression is increased under a follower load. Spine 25, 1548–1554. Paxinos, O., Ghanayem, A.J., Zindrick, M.R., Voronov, L.I., Havey, R.M., Carandang, G., Hadjipavlou, A., Patwardhan, A.G., 2009. Anterior cervical discectomy and fusion with a locked plate and wedged graft effectively stabilizes flexiondistraction stage-3 injury in the lower cervical spine: a biomechanical study. Spine 34, E9–E15. Reitman, C.A., Mauro, K.M., Nguyen, L., Ziegler, J.M., Hipp, J.A., 2004. Intervertebral motion between flexion and extension in asymptomatic individuals. Spine 29, 2832–2843. Schuldt, K., 1988. On neck muscle activity and load reduction in sitting postures. An electromyographic and biomechanical study with applications in ergonomics and rehabilitation. Scand. J. Rehabil. Med. 19, 1–49. Snyder, J.T., Tzermiadianos, M.N., Ghanayem, A.J., Voronov, L.I., Rinella, A., Dooris, A., Carandang, G., Renner, S.M., Havey, R.M., Patwardhan, A.G., 2007. Effect of uncovertebral joint excision on the motion response of the cervical spine after total disc replacement. Spine 32, 2965–2969. Wilke, H.J., Claes, L., Schmitt, H., Wolf, S., 1994. A universal spine tester for in vitro experiments with muscle force simulation. Eur. Spine J. 3, 91–97. Wilke, H.J., Rohlmann, A., Neller, S., Schultheiss, M., Bergmann, G., Graichen, F., Claes, L.E., 2001. Is it possible to simulate physiologic loading conditions by applying pure moments? A comparison of in vivo and in vitro load components in an internal fixator. Spine 26, 636–642. Wilke, H.J., Wenger, K., Claes, L., 1998. Testing criteria for spinal implants: recommendations for the standardization of in vitro stability testing of spinal implants. Eur. Spine J. 7, 148–154. Wu, S.K., Kuo, L.C., Lan, H.C., Tsai, S.W., Chen, C.L., Su, F.C., 2007. The quantitative measurements of the intervertebral angulation and translation during cervical flexion and extension. Eur. Spine J. 16, 1435–1444. Wu, S.K., Kuo, L.C., Lan, H.C., Tsai, S.W., Su, F.C., 2010. Segmental percentage contributions of cervical spine during different motion ranges of flexion and extension. J. Spin. Disord. Tech. 23, 278–284.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine Kevin M. Bell a,n, Yiguo Yan a, Richard E. Debski b, Gwendolyn A. Sowa c, James D. Kang a, Scott Tashman a a
Ferguson Laboratory for Spine Research, Department of Orthopaedic Surgery, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA Department of Bioengineering, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA c Department of Physical Medicine and Rehabilitation, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA b
art ic l e i nf o
a b s t r a c t
Article history: Accepted 21 November 2015
The human cervical spine supports substantial compressive load in-vivo arising from muscle forces and the weight of the head. However, the traditional in-vitro testing methods rarely include compressive loads, especially in investigations of multi-segment cervical spine constructs. Various methods of modeling physiologic loading have been reported in the literature including axial forces produced with inclined loading plates, eccentric axial force application, follower load, as well as attempts to individually apply/model muscle forces in-vitro. The importance of proper compressive loading to recreate the segmental motion patterns exhibited in-vivo has been highlighted in previous studies. However, appropriate methods of representing the weight of head and muscle loading are currently unknown. Therefore, a systematic comparison of standard pure moment with no compressive loading versus published and novel compressive loading techniques (follower load – FL, axial load – AL, and combined load – CL) was performed. The present study is unique in that a direct comparison to continuous cervical kinematics over the entire extension to flexion motion path was possible through an ongoing intrainstitutional collaboration. The pure moment testing protocol without compression or with the application of follower load was not able to replicate the typical in-vivo segmental motion patterns throughout the entire motion path. Axial load or a combination of axial and follower load was necessary to mimic the in-vivo segmental contributions at the extremes of the extension-flexion motion path. It is hypothesized that dynamically altering the compressive loading throughout the motion path is necessary to mimic the segmental contribution patterns exhibited in-vivo. Published by Elsevier Ltd.
Keywords: Cervical spine Kinematics Range of motion ROM Follower load Compressive preload Hybrid control Robotics Bi-plane radiography In-vivo
1. Background In-vitro biomechanical testing has been critical in the design and evaluation of surgical instrumentation. Determination of realistic physiologic loading levels for the cervical spine has, however, proven difficult outside of the in-vivo setting. Unconstrained pure moment testing combined with the hybrid testing method is currently the gold standard test protocol for evaluation of motion preservation technology and adjacent level effects. Pure moment testing was specifically designed to apply uniform loading at each cross section throughout the length of a spinal construct, permitting irregularities to be identified (Panjabi, 1988). Pure moment testing is well suited for making relative comparisons between treatments, but is currently not based on or representative of in-vivo motion (Panjabi, 2007). n
Corresponding author. Tel.: þ 1 412 657 7857; fax: þ1 412 648 8548. E-mail address: [email protected] (K.M. Bell).
Additionally, the human cervical spine supports substantial compressive load in-vivo arising from muscle forces and the weight of the head. However, the traditional in-vitro testing methods rarely include compressive loads; especially in investigations of multi-segment cervical spine constructs. Various methods of modeling physiologic loading have been reported in the literature including axial forces produced with inclined loading plates, eccentric axial forces application, follower load, as well as attempts to individually apply/model muscle forces in-vitro (Adams and Dolan, 2005; Cook, 2009; Cripton et al., 2000; DiAngelo and Foley, 2004; Goel et al., 2006; Miura et al., 2002; Panjabi, 1988, 2007; Panjabi et al., 2001; Patwardhan et al., 2000; Wilke et al., 1994, 2001, 1998). Miura et al. (2002) and DiAngelo and Foley (2004) published articles directly aimed at determining the most appropriate loading mechanism to produce physiologic motion patterns. Miura et al. (2002) presented pure moment testing combined with follower load and, through adjusting moment targets, was able to
http://dx.doi.org/10.1016/j.jbiomech.2015.11.045 0021-9290/Published by Elsevier Ltd.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
achieve 20% agreement with segmental range of motion reported in literature, however typical segmental motion patterns were not observed with this technique. DiAngelo and Foley utilized an eccentric axial compressive method in attempt to mimic the weight of the head. DiAngelo and Foley (2004) was able to show reasonable agreement with the segmental motion patterns, but the magnitudes dramatically underestimated the average in-vivo segmental kinematics. The objective of this project is to identify and verify the appropriate in-vitro loading conditions that would replicate the in-vivo kinematics of the cervical spine, with the overall goal of improving the biofidelity of the experimental platform. A systematic comparison of standard pure moment with no compressive loading versus published and novel compressive loading techniques (follower load – FL, axial load – AL, and combined load – CL) was performed. It is hypothesized that an optimized follower load, passing through the segmental centers of rotation, will add stability to the system but will not dramatically affect the segmental motion patterns observed throughout the extension to flexion motion path. In contrast, axial load applied perpendicular to superior most vertebral body, will not maintain the pure moment assumption, whereby enabling the segmental motion patterns to be altered. 2. Methods 2.1. Protocol N ¼12 fresh-frozen human (C3-7) cervical cadaveric specimens (51.8 years7 7.3) were pre-screened with CT and dissected, preserving osteoligamentous structures. Specimens were mounted in a robot-based spine testing system, consisting of a serial linkage robotic manipulator (Staubli RX90, Staubli Inc., Duncan, SC) with an on-board six-axis load cell (UFS Model 90M38A-150, JR3 Inc., Woodland, CA) and custom specimen-mounting fixtures (Bell et al., 2007; Gilbertson et al., 1999; Hartman et al., 2009). Four clinical lateral mass screws were used to secure the specimens to the mounting fixtures (one in each pedicle and two in the anterior portion of the vertebral body). After mounting, specimens were wrapped in 0.9% saline soaked gauze and periodically sprayed with saline in order to prevent dehydration. The robot was controlled via MATLAB (Mathworks, Inc.) and operates under adaptive displacement control to a pure moment target of 2.0 N m for flexion and extension (FE) for each state in a randomized order (no compression (Fig. 1A), follower load, axial rotation, combined loading). Due to the quasi-static nature of the adaptive displacement control algorithm the system operated at a rate of 0.0677 0.0014°/s. Two consecutive full extension to flexion loops were performed with the data from the second cycle being presented to account for preconditioning (Cripton et al., 2000). Segmental motion was recorded using a five camera VICON
system tracking passive reflective markers rigidly attached as a marker group to each vertebral body. A hand held VICON digitizer was utilized to digitize the anatomical coordinate system for each vertebral body relative to the marker group and the Euler angle rotations of C34, C45, C56, and C67 were determined and reported. 2.2. Follower load Follower load application was accomplished by loading the specimen with bilateral cables passing through cable guides inserted into the vertebral bodies and over pulleys attached to the base (Fig. 1B). A novel active system was implemented in our laboratory using linear actuators coupled with load cells. Control of the system was integrated with the custom-built PC-based control program written in MATLAB that is currently used to control the robot testing system, and enabled active control of the loading throughout the motion path. The follower load system (Fig. 1B) consists of two independently controlled 24 V servo motor linear actuators (Ultramotion – 3-B.125-DC426_24-4-/4) and compression/tension load cells (Transducer Techniques – MLP-100). A Galil Motion Controller (DMC-4183-BOX8 (-16BIT)-D3040-D4040) controlled this system using on-board closed loop Proportional-Integral-Derivative (PID) load control. The 3/64‶ diameter stainless steel wire rope lanyard was threaded through a custom designed adjustable cable guide system attached with clinical pedicle screws to enable the follower load cable to interface with the specimen in a manner consistent with the design criterion: (a) tangent to the curvature of the spine and (b) pass through the specimen’s center of rotation (COR). Optimization of the follower load path to align with the specimen’s COR was accomplished through an offline iterative feedback process using the moment output of the testing system’s on-board six-axis load cell. With the specimen in the neutral position, 100 N of follower load was applied to the specimen and resulting change in moments was recorded. Follower load magnitude of 100 N was chosen as it is representative of the most common follower load magnitude presented in literature (Cho et al., 2010; Finn et al., 2009, 2011; Lee et al., 2011; Martin et al., 2011; Paxinos et al., 2009; Snyder et al., 2007). The position of the cable guide was then adjusted to counteract the moment change and the process was repeated until less than 0.1 N m change in moment was observed. Preliminary testing of the described optimization process was performed ensuring that the previously described maintenance of segmental curvature angle criteria was upheld (Patwardhan et al., 2003). 2.3. Axial load Although less popular than follower load as a method to apply compressive load due to published instability issues with this testing method, some authors believe axial loading to be the most physiologic loading scheme—mimicking head weight (DiAngelo and Foley, 2004). Axial loading can be applied along an axis locally fixed to the specimen or globally fixed to the world coordinate system. Previous reports have shown that the cervical spine buckles at very low loads when an axial load is applied globally, therefore for this study the axial load was applied along an axis locally fixed to the specimen (perpendicular to the robot end effector – Fig. 1C). The axial load was applied using the robotic arm to a load target of 50 N (DiAngelo and Foley, 2004), representative of the approximate weight of the head, using the adaptive displacement control algorithm enabling the load to be applied purely in the axial direction and be maintained throughout the flexion–extension rotation path.
Fig. 1. Schematic of the four loading states implemented in this study, (A) no compression, (B) follower load, (C) axial load rotated state showing perpendicular line of action depicted for clarity and (D) combined load.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 2.4. Combined load Simulation of muscle loading and the weight of the head with the follower load and axial compressive load have been independently shown to result in modest improvements in making the in-vitro motion more physiologic (DiAngelo and Foley, 2004; Miura et al., 2002). In this study, the two loading schemes were combined. It was hypothesized that the combination of axial and follower loading will have a synergistic effect, producing more physiologic kinematics (Fig. 1D). 2.5. In-vivo data The previously reported in-vivo data set (Anderst et al., 2013a, 2013b) utilized in the study was reanalyzed for direct comparison with the current in-vitro data. In-vivo data consisted of N ¼20 asymptomatic control patients (45.5 years7 5.8) consented to participate in an Institutional Review Board (IRB) approved protocol. Subjects performed continuous, full ROM flexion–extension at a rate of one complete cycle every 3 s. Subject-specific bone models of C3-C7 were created from CT scans. A previously validated tracking process determined three-dimensional vertebral position with sub-millimeter accuracy by matching bone models from the CT scan to the biplane X-rays (Anderst et al., 2011). 2.6. Data analysis Results of the in-vitro testing at full range of motion (ROM) were summarized and expressed in bar graphs representing mean 7 95% confidence intervals. A Shapiro–Wilk test was performed, which indicated that not all groups were normally distributed, therefore a non-parametric Wilcoxon signed-rank test for paired design was used to identify significant (*p o 0.05) differences between loading modes (no compression versus follower load, axial load and combined load). The in-vivo and in-vitro ROM data was normalizing to percent ROM. The quality of fit between in-vivo and in-vitro motion was assessed by calculating root mean square error (RMSE) at 20% increments of percent ROM.
3. Results Fig. 2 displays the segmental contributions at 100% of the overall extension to flexion motion path. At 100% ROM the segmental distribution for no compressive load state exhibited approximately equal segmental contribution from each level. Adding follower load only had minimal effects on the segmental contribution from each level. However, adding the axial load significantly reduced the C67 contribution to the motion path (p ¼0.004), shifting the majority of the motion toward the middle (C4-5 and C5-6) segments. This significant reduction was also observed in the combined loading state (p ¼0.019). The overall extension to flexion ROM was not affected by the application of a compressive load, with all four compressive loading states exhibiting virtually identical ROM (Fig. 3A). However, although not statistically significant (p ¼0.14), the in-vitro ROM was on average 12.3% smaller than the in-vivo data set. Therefore, in order to directly compare the segmental contributions between the in-vivo and in-vitro data the in-vivo segmental rotation was uniformly scaled by a factor of 87.7% (Fig. 3B). The scatter plot of the in-vitro segmental rotation plotted versus percent of total ROM (Fig. 4) demonstrates clearly that the adding compressive load effects the segmental motion distribution. A small change is observed with application of follower load, wherein C34 and C45 appear to contribute more than C56 and C67 throughout middle portion of the extension to flexion curve. This same change is further magnified with the application of an axial compressive load. The axial compressive loading state and the combined loading curves show a very similar pattern with the upper segments appearing to be recruited from superior to inferior as the motion path progresses. The average RMSE between the in-vivo data and in-vitro data increased with the application of compressive loading. However, the RMSE was dependent on percentage ROM (Fig. 5A) and segmental level (Fig. 5B). No compression had the lowest RMSE at
3
20%, 40% and 60% ROM, which corresponds to the extension and neutral region of the overall path of motion. However, the addition of axial and combined loading resulted in the lowest RMSE at 80% and 100% ROM, which corresponds to the flexion region of the motion path (Fig. 5, Table 1). No compression and follower load had the lowest RMSE at C34 and C56, however axial load and combined load had the lowest RMSE at C45 and C67.
4. Discussion This study systematically evaluated standard pure moment with no compressive loading versus published and novel compressive loading techniques (follower load, axial load, and combined load). Consistent with previous in-vitro reports, the pure moment testing protocol with no compression was not able to replicate the typical in-vivo segmental motion pattern reported in literature (DiAngelo and Foley, 2004; Miura et al., 2002). Although variability existed from specimen to specimen due to local degenerative differences between specimens, overall applying a pure moment at each segment resulted in approximately equal segmental contribution to the overall extension to flexion motion path. Adding follower load allowed the specimen to be tested at a physiologic loading magnitude, while maintaining the pure moment assumption. However, this loading scheme was still not able to replicate the in-vivo segmental motion patterns near the end range of motion. The follower load state in the present study reproduced flexion–extension parameters used in the Miura et al. (2002) study. Although, similar trends are observed between the two studies in terms of percent contribution, the ROM magnitude observed at 2.0 N m in the Miura study was approximately 20% more than the ROM observed in this study. Differences in specimen grade and testing methods could account for some of this disagreement; however in both cases a larger moment target would have been necessary to replicate the magnitude of the reported in-vivo kinematics. Additionally, the Miura et al. study only reported end-range kinematics, which is the predominate trend in literature, however the end-range kinematics are not necessarily representative of the mid-range cervical kinematics and are highly variable (Dvorak et al., 1988; Frobin et al., 2002; Reitman et al., 2004; Wu et al., 2007). High variability was also a limitation of the present study, although attempts were made to limit the variability by controlling recruitment age and pre-screening subjects and specimens. The present study is unique in that a direct comparison to continuous cervical kinematics over the entire extension to flexion motion path was possible through an ongoing intra-institutional collaboration. As hypothesized, application of an axial load perpendicular to the superior most vertebral body altered the segmental motion patterns. The most dramatic differences were
Fig. 2. Average segmental kinematics at 100% of the full extension to full flexion motion path for each loading mode (*p o0.05).
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
4
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
observed in the middle portion of the curve (the neutral zone). Very few authors have reported segmental distribution data at sub-maximal points on the motion curves, making this data difficult to interpret (Wu et al., 2010). Further investigation is necessary to understand this fully, however it is theorized that the
Fig. 3. (A) Overall extension-to-flexion ROM for in vivo and each in vitro loading scheme. P-values are for differences between the in-vivo and in-vitro data. (B) Scaled in-vivo segmental motion path plotted versus the percentage of the overall range of motion (full extension to full flexion for C3-C7).
inherent laxity of the neutral zone combined with the lordotic curvature of a healthy cervical spine account for these dramatic shifts in the segmental motion pattern when axial load is applied. Although the changes in segmental motion pattern were less pronounced at full range of motion, axial load and combined loading were able to produce a motion pattern similar to in-vivo at
Fig. 5. RMS errors between in-vivo and in-vitro motion for each loading condition. Errors varied with both: (A) Percentage of the overall range of motion; and (B) Segmental level.
Fig. 4. In-vitro segmental motion paths plotted versus the percentage of the overall range of motion (full extension to full flexion for C3-C7), showing the effects of altered compressive loading.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Table 1 Scaled in-vivo (deg.) % ROM 20% 40% 60% 80% 100%
C34 1.7 7 0.3 4.1 70.6 6.8 7 0.9 9.5 7 1.1 11.9 7 1.2
C45 2.4 7 0.3 5.2 7 0.5 8.17 0.7 10.9 70.9 13.4 71.1
C56 2.7 70.6 5.17 1.1 7.3 7 1.5 9.5 71.9 11.9 7 2.2
C67 2.4 7 0.8 4.2 7 1.3 5.7 7 1.7 7.4 7 2.0 9.5 7 2.3
No compression (deg.) % ROM C34 20% 2.7 7 0.8 40% 5.5 7 1.6 60% 7.6 7 2.1 80% 9.1 72.2 100% 11.2 7 2.4 RMSE 1.1
C45 2.3 7 0.3 5.3 7 0.8 7.7 7 1.3 9.6 7 1.4 11.7 7 1.6 1.7
C56 1.8 7 0.7 3.8 71.0 6.4 71.6 9.17 2.0 11.5 7 2.4 0.9
C67 2.4 7 0.8 4.2 7 1.0 6.5 7 1.3 9.7 7 1.8 12.3 7 2.2 3.0
RMSE 0.7 1.0 0.7 1.3 1.7 1.3
Follower load (deg.) % ROM C34 20% 3.0 7 0.9 40% 6.2 7 1.7 60% 8.3 7 2.0 80% 9.6 7 2.0 100% 11.2 7 2.1 RMSE 1.5
C45 2.3 7 0.4 5.6 7 0.6 8.17 0.9 9.7 7 0.9 11.4 7 1.0 2.0
C56 1.9 7 0.6 3.5 71.1 6.2 71.9 9.17 2.4 12.0 7 2.9 1.0
C67 2.0 7 0.6 3.3 7 0.9 5.4 7 1.2 9.0 7 2.2 12.17 2.8 2.8
RMSE 0.8 1.4 1.0 1.0 1.7 1.4
Axial load % ROM 20% 40% 60% 80% 100% RMSE
(deg.) C34 3.4 7 0.9 7.8 7 1.9 9.7 7 2.2 10.6 7 2.3 11.7 7 2.4 2.3
C45 1.9 7 0.5 5.4 7 0.8 9.2 7 1.5 10.8 71.6 12.1 71.6 1.4
C56 1.6 7 0.7 2.5 70.9 5.5 71.7 9.4 72.6 11.9 7 2.9 1.5
C67 2.3 7 1.1 2.9 7 1.1 3.6 7 0.8 6.5 7 1.2 10.9 72.1 1.9
RMSE 1.0 2.4 2.1 0.7 1.0 1.6
Combined % ROM 20% 40% 60% 80% 100% RMSE
load (deg.) C34 3.7 7 1.0 7.4 7 1.6 9.5 7 1.7 10.6 7 1.8 11.5 7 2.0 2.2
C45 2.0 7 0.4 5.8 7 0.9 9.4 7 1.2 11.17 1.2 12.1 71.2 1.4
C56 1.5 7 0.7 2.4 70.9 5.3 71.8 8.9 72.8 11.6 7 3.0 1.7
C67 1.8 7 0.6 2.4 7 0.8 3.17 1.0 5.8 7 1.5 10.3 72.5 1.8
RMSE 1.2 2.3 2.2 1.0 0.8 1.6
the extremes of the extension-to-flexion motion path. This observation is consistent with the study published by DiAngelo and Foley (2004), which utilized an eccentric axial compressive method in attempt to mimic the weight of the head. DiAngelo and Foley (2004) were able to show reasonable agreement with the segmental motion patterns, but the reported magnitudes dramatically underestimated the average in-vivo segmental kinematics. Although, the two methods share similarities, an eccentric axial loading method has limited stability, an issue that is corrected in the present study with the combined loading methodology. The fact that axial compressive and/or combined loading resulted in the lowest RMSE at 80% and 100% ROM is interesting in light of the clinical scenario. The need for axial compressive load may be reflective of the in-vivo influence of increased muscular contribution with increasing rotation. This is consistent with muscular models of the cervical spine which have shown that the resultant load experienced on the head increases with increasing rotation (Cheng et al., 2008; Johnston et al., 2008; Schuldt, 1988). In the present study the AL force was applied perpendicular to the superior most vertebral body (C3) which mimics the resultant muscular force on the head as it is transmitted to the vertebral column through the occiput. This is consistent with the inverse dynamic model proposed by Anderst et al. that was based on muscle driven in-vivo kinematics (Anderst et al., 2013c). The muscle loading discussion resonates with Patwardhan et al. (2000) conclusion that the stability provided by application of a follower load “may suggest a mechanism of action by which
5
muscles stabilize the cervical spine under physiologic compression”. In the present study, follower load was shown to be beneficial in the neutral zone for added stability, whereas an axial compressive load was required in the elastic zone for proper segmental distribution. Taking this beyond the mechanism of muscle loading, this distinction is interesting when considering the cervical spine osteoligamentous structures. Throughout the neutral zone, the stiffness of the construct is provided primarily from the intervertebral disc. However, with increasing flexion the ligamentous components are recruited and the instantaneous axis of rotation has been shown to shift anteriorly (DiAngelo and Foley, 2004). It is theorized that the physiologic loading may linked to the anatomy and therefore may be dependent upon the flexion– extension moment-rotation curve. The present study had several limitations. The rate of the invitro testing system was at least an order of magnitude slower than the rate at with the in-vivo data was recorded, due to the quasi-static testing algorithm. This discrepancy will affect the kinematics, but the influence should be minimized due to the invitro preconditioning protocol. Additionally, this study focused on flexion and extension due to the limitations of the follower load methodology, therefore future work is necessary to determine the applicability of the present methodologies to axial rotation and lateral bending of the cervical spine. Finally, the current study applied the compressive load across a multi-level cervical spine construct. Although the results indicate that this methodology enabled the segmental motion patterns to be altered, it is possible that direct manipulation of the individual motion segments is necessary to fully simulate physiologic loading. Despite the limitations of this study, future investigations focused on optimizing the axial load and combined loading methods for replication of in-vivo cervical spine kinematics are recommended. It is hypothesized that dynamically altering the compressive loading throughout the motion path is necessary in order to mimic the segmental contribution patterns exhibited invivo. However, it should be noted this study is only the first step in the process to develop a new testing protocol to mimic in-vivo kinematics in-vitro and will also serve as the foundation for future in-vitro experimentation where physiologic kinematics/kinetics are critical to the clinical relevance of the data such as determination of native biomechanical properties, establishment/validation of computational models, and as a loading scheme for mechanobiology experiments.
Confict of interest We wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome.
Acknowledgments In-vitro testing was performed with the support of the Albert B. Ferguson, Jr. MD Orthopedic Fund of The Pittsburgh Foundation. In vivo data was acquired with support from the National Institutes of Health (Grant # R03 AR056265) and the Cervical Spine Research Society's Society 21st Century Development Grant. I would also like to acknowledge Bill Anderst for valuable contributions to the content of this manuscript. Funding from NIH/NCCAM K08AT00471802 award is also acknowledged.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i
K.M. Bell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
References Adams, M.A., Dolan, P., 2005. Spine biomechanics. J. Biomech. 38, 1972–1983. Anderst, W.J., Baillargeon, E., Donaldson 3rd, W.F., Lee, J.Y., Kang, J.D., 2011. Validation of a noninvasive technique to precisely measure in vivo threedimensional cervical spine movement. Spine 36, E393–E400. Anderst, W.J., Donaldson 3rd, W.F., Lee, J.Y., Kang, J.D., 2013a. Cervical motion segment percent contributions to flexion–extension during continuous functional movement in control subjects and arthrodesis patients. Spine 38 (9), E533-9. Anderst, W.J., Donaldson, W.F., Lee, J.Y., Kang, J.D., 2013b. Cervical spine intervertebral kinematics with respect to the head are different during flexion and extension motions. J. Biomech. 46, 1471–1475. Anderst, W.J., Donaldson, W.F., Lee, J.Y., Kang, J.D., 2013c. Subject-specific inverse dynamics of the head and cervical spine during in vivo dynamic flexion– extension. J. Biomech. Eng. 135, 61007–61008. Bell, K.M., Gilbertson, L., Cook, H., Arthur, A., Doehring, T., Muriuki, M., Smolinski, P., Kang, J.D., 2007. Investigation of the integrated passive, active, and control subsystems of the spine utilizing a robotics-based spine testing system. In: Proceedings of the 53th Annual Meeting of the Orthopaedic Research Society (ORS), San Diego, CA. Cheng, C.H., Lin, K.H., Wang, J.L., 2008. Co-contraction of cervical muscles during sagittal and coronal neck motions at different movement speeds. Eur. J. Appl. Physiol. 103, 647–654. Cho, B.Y., Lim, J., Sim, H.B., Park, J., 2010. Biomechanical analysis of the range of motion after placement of a two-level cervical ProDisc-C versus hybrid construct. Spine 35, 1769–1776. Cook, D.J., 2009. Characterization of the Response of the Cadaveric Human Spine to Loading in a Six-Degree-of-Freedom Spine Testing Apparatus. University of Pittsburgh. School of Engineering, Pittsburgh, PA. Cripton, P.A., Bruehlmann, S.B., Orr, T.E., Oxland, T.R., Nolte, L.P., 2000. In vitro axial preload application during spine flexibility testing: towards reduced apparatusrelated artefacts. J. Biomech. 33, 1559–1568. DiAngelo, D.J., Foley, K.T., 2004. An improved biomechanical testing protocol for evaluating spinal arthroplasty and motion preservation devices in a multilevel human cadaveric cervical model. Neurosurg. Focus 17, E4. Dvorak, J., Froehlich, D., Penning, L., Baumgartner, H., Panjabi, M.M., 1988. Functional radiographic diagnosis of the cervical spine: flexion/extension. Spine 13, 748–755. Finn, M.A., Brodke, D.S., Daubs, M., Patel, A., Bachus, K.N., 2009. Local and global subaxial cervical spine biomechanics after single-level fusion or cervical arthroplasty. Eur. Spine J. 18, 1520–1527. Finn, M.A., Samuelson, M.M., Bishop, F., Bachus, K.N., Brodke, D.S., 2011. Two-level noncontiguous versus three-level anterior cervical discectomy and fusion: a biomechanical comparison. Spine 36, 448–453. Frobin, W., Leivseth, G., Biggemann, M., Brinckmann, P., 2002. Sagittal plane segmental motion of the cervical spine. A new precision measurement protocol and normal motion data of healthy adults. Clin. Biomech. 17, 21–31. Gilbertson, L.G., Doehring, T.C., Livesay, G.A., Rudy, T.W., Kang, J.D., Woo, S.L.Y., 1999. Improvement of accuracy in a high-capacity, six degree-of-freedom load cell: application to robotic testing of musculoskeletal joints. Annu. Biomed. Eng. 27, 839–843. Goel, V.K., Panjabi, M.M., Patwardhan, A.G., Dooris, A.P., Serhan, H., 2006. Test protocols for evaluation of spinal implants. J. Bone Jt. Surg. Am. 88 (Suppl 2), S103–S109. Hartman, R.A., Bell, K.M., Kang, J.D., 2009. Analyses of the Components of the Posterior Column in a Distractive-Flexion Injury Model. In: Proceedings of the 55th Annual Meeting of the Orthopaedic Research Society (ORS), Las Vegas, NV.
Johnston, V., Jull, G., Souvlis, T., Jimmieson, N.L., 2008. Neck movement and muscle activity characteristics in female office workers with neck pain. Spine 33, 555–563. Lee, S.H., Im, Y.J., Kim, K.T., Kim, Y.H., Park, W.M., Kim, K., 2011. Comparison of cervical spine biomechanics after fixed- and mobile-core artificial disc replacement: a finite element analysis. Spine 36, 700–708. Martin, S., Ghanayem, A.J., Tzermiadianos, M.N., Voronov, L.I., Havey, R.M., Renner, S.M., Carandang, G., Abjornson, C., Patwardhan, A.G., 2011. Kinematics of cervical total disc replacement adjacent to a two-level, straight versus lordotic fusion. Spine 36, 1359–1366. Miura, T., Panjabi, M.M., Cripton, P.A., 2002. A method to simulate in vivo cervical spine kinematics using in vitro compressive preload. Spine 27, 43–48. Panjabi, M.M., 1988. Biomechanical evaluation of spinal fixation devices: I. A conceptual framework. Spine 13, 1129–1134. Panjabi, M.M., 2007. Hybrid multidirectional test method to evaluate spinal adjacent-level effects. Clin. Biomech. 22, 257–265. Panjabi, M.M., Miura, T., Cripton, P.A., Wang, J.L., Nain, A.S., DuBois, C., 2001. Development of a system for in vitro neck muscle force replication in whole cervical spine experiments. Spine 26, 2214–2219. Patwardhan, A.G., Havey, R.M., Carandang, G., Simonds, J., Voronov, L.I., Ghanayem, A.J., Meade, K.P., Gavin, T.M., Paxinos, O., 2003. Effect of compressive follower preload on the flexion–extension response of the human lumbar spine. J. Orthop. Res.: Off. Publ. Orthop. Res. Soc. 21, 540–546. Patwardhan, A.G., Havey, R.M., Ghanayem, A.J., Diener, H., Meade, K.P., Dunlap, B., Hodges, S.D., 2000. Load-carrying capacity of the human cervical spine in compression is increased under a follower load. Spine 25, 1548–1554. Paxinos, O., Ghanayem, A.J., Zindrick, M.R., Voronov, L.I., Havey, R.M., Carandang, G., Hadjipavlou, A., Patwardhan, A.G., 2009. Anterior cervical discectomy and fusion with a locked plate and wedged graft effectively stabilizes flexiondistraction stage-3 injury in the lower cervical spine: a biomechanical study. Spine 34, E9–E15. Reitman, C.A., Mauro, K.M., Nguyen, L., Ziegler, J.M., Hipp, J.A., 2004. Intervertebral motion between flexion and extension in asymptomatic individuals. Spine 29, 2832–2843. Schuldt, K., 1988. On neck muscle activity and load reduction in sitting postures. An electromyographic and biomechanical study with applications in ergonomics and rehabilitation. Scand. J. Rehabil. Med. 19, 1–49. Snyder, J.T., Tzermiadianos, M.N., Ghanayem, A.J., Voronov, L.I., Rinella, A., Dooris, A., Carandang, G., Renner, S.M., Havey, R.M., Patwardhan, A.G., 2007. Effect of uncovertebral joint excision on the motion response of the cervical spine after total disc replacement. Spine 32, 2965–2969. Wilke, H.J., Claes, L., Schmitt, H., Wolf, S., 1994. A universal spine tester for in vitro experiments with muscle force simulation. Eur. Spine J. 3, 91–97. Wilke, H.J., Rohlmann, A., Neller, S., Schultheiss, M., Bergmann, G., Graichen, F., Claes, L.E., 2001. Is it possible to simulate physiologic loading conditions by applying pure moments? A comparison of in vivo and in vitro load components in an internal fixator. Spine 26, 636–642. Wilke, H.J., Wenger, K., Claes, L., 1998. Testing criteria for spinal implants: recommendations for the standardization of in vitro stability testing of spinal implants. Eur. Spine J. 7, 148–154. Wu, S.K., Kuo, L.C., Lan, H.C., Tsai, S.W., Chen, C.L., Su, F.C., 2007. The quantitative measurements of the intervertebral angulation and translation during cervical flexion and extension. Eur. Spine J. 16, 1435–1444. Wu, S.K., Kuo, L.C., Lan, H.C., Tsai, S.W., Su, F.C., 2010. Segmental percentage contributions of cervical spine during different motion ranges of flexion and extension. J. Spin. Disord. Tech. 23, 278–284.
Please cite this article as: Bell, K.M., et al., Influence of varying compressive loading methods on physiologic motion patterns in the cervical spine. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.11.045i