MICROSTRUCTURAL DAMAGE IN ARTERIAL TISSUE EXPOSED

TO REPEATED TENSILE STRAINS

Neal Austin, BSc,a Lisa M. DiFrancesco, MD,b,c and Walter Herzog, PhDa

a Graduate StudCalgary, Calgary,

b Doctor of MedThe University of

c Professor, CaCanada.

Submit requestHuman PerformaFaculty of KinesioCanada T2N 1N4

Paper submittedaccepted Septembe

0161-4754/$36Copyright © 20doi:10.1016/j.jm

14

ABSTRACT

Objectives: Vertebral artery (VA) damage has been anecdotally linked to high-speed, low-amplitude spinal manipulativetreatments (SMTs) of the neck. Apart from a single study quantifying the maximum stresses and strains imposed on the VAduring cervical SMT, there are no data on the mechanics of the VA for this treatment modality, and there is no informationon the possible long-term effects of repeat exposure to cervical SMT. The purpose of this study was to quantifymicrostructural damage in arterial tissue exposed to repeat strain loading of a magnitude similar to the maximum strainsmeasured in the VA during high-speed, low-amplitude cervical SMT.Methods: Twenty-four test specimens from cadaveric rabbit ascending aorta were divided into 2 control groups (n = 12)and 2 experimental groups (n = 6 each). Specimens were exposed to 1000 strain cycles of 0.06 and 0.30 of their in situlength. A pathologist, blinded to the experimental groups, assessed microstructural changes in the arteries usingquantitative histology. Pearson χ2 analysis (α = .05) was used to assess differences in tissue microstructure between groups.Results: Control and 0.06 strain tissues were statistically the same (P = .406), whereas the 0.30 strain group showedmicrostructural damage beyond that seen in the control group (P = .024).Conclusions: We conclude that cadaveric rabbit arterial tissue similar in size and mechanical properties of that of thehuman VA can withstand repeat strains of magnitudes and rates similar to those measured in the cadaveric VA duringcervical SMT without incurring microstructural damage beyond control levels. (J Manipulative Physiol Ther2010;33:14-19)

Key Indexing Terms: Manipulation, Spinal; Arteries; Chiropractic; Neck Injuries

Vertebral artery (VA) damage is commonly thought tooccur as a result of everyday tasks such as turning thehead while backing up a vehicle, lifting objects,

coughing, sport-related injuries, and undergoing spinalmanipulative therapy.1 It is reported that the incidence ofVA damage associated with cervical manipulation is between1 and 1.5 cases per 100 000 people annually, and the popu-lation in which it is primarily identified in is middle-aged menand women.2,3 Although the occurrence of VA damage is low,the consequences of injury are invariably serious.4

The anatomical orientation of the VA is thought to leave itprone to injury during cervical high-speed, low-amplitude

ent, Faculty of Kinesiology, The University ofAlberta, Canada.icine, Departments of Pathology and Oncology,Calgary, Calgary, Alberta, Canada.lgary Laboratory Services, Calgary, Alberta,

s for reprints to: Walter Herzog, PhD, KNB 402nce Laboratory, The University of Calgary,logy, 2500 University Drive NW, Calgary, AB,(e-mail: walter@kin.ucalgary.ca).April 7, 2009; in revised form August 28, 2009;r 25, 2009..0010 by National University of Health Sciences.pt.2009.11.006

manipulation because of neck rotation.5 The VA runs fromthe subclavian artery through the transverse foramen of thecervical vertebrae before forming the vertebral loop betweenC1 and the foramen magnum.6 The vertebral loop is thoughtto accommodate stretch of the VA during neck rotation andcould theoretically be prone to injury during spinalmanipulation. In an experiment involving 5 un-embalmedgeriatric cadavers and 1 chiropractic clinician, the internalforces and associated strains acting on the VA duringcervical spinal manipulation were quantified.7 In thisexperiment, it was found that the VA is “slack” when thehead is in a neutral position and is only unraveled or stretchedtoward the end of passive neck motion.7 The head positionduring passive neck motion, which would cause a stretch inthe VA, is not reached during normal spinal manipulativeprocedures; therefore, it has been suggested that a single necktreatment has no ill effect on the mechanical integrity of theVA.7 However, a chiropractic patient may receive a dozen ormore cervical spinal manipulations per year,8 and thepossible mechanical effects of repeat exposure to arterialstrain remain unknown, because the possible microstructuralchanges due to such mechanical loading have never beenquantified to our knowledge.

The relationship between the number of spinal manip-ulations a patient receives and the side effects resultingfrom repeat treatments has been studied and has been found

Fig 1. The ascending aorta of a New Zealand white rabbit. The testsection is indicated by arrows.

15Austin et alJournal of Manipulative and Physiological TherapeuticsArterial Tissue DamageVolume 33, Number 1

to not exist.9 Furthermore, investigation into the effects ofcervical spinal manipulation on preexisting arterial lesionshas been made, but no significant differences in arteriallesion dimensions (length, cross-sectional area, and vol-ume) could be detected after 20 treatments.10 Biomechan-ical research on cervical spinal manipulations has focusedon the external forces applied by chiropractors on patientsduring cervical spinal manipulative treatment (SMT).7,11,12

Only a single study focused on the stresses and strains onthe VA for a variety of high-speed, low-amplitude cervicalSMTs.7 The average peak forces applied externally to theneck region of subjects during spinal manipulation rangefrom approximately 100 to 150 N and are achieved withinapproximately 100 milliseconds. For these external forcesand neck manipulations, it was found that cervical spinalmanipulation resulted in average peak strains of cadavericVAs of 0.06 from their lengths with the head and neck inthe neutral position.7

Experiments designed to investigate VA microstructuraldamage due to straining and/or repeat exposure to manip-ulative treatments have been unsuccessful in humans.Furthermore, questions pertaining to the relationship be-tween repeat cervical spinal manipulation and its effect onVA microstructural damage have arisen in court, specificallythe inquest into the death of Lana Dale Lewis.13 The purposeof this study was to investigate the effects of repeat strainapplication on the microstructural integrity of arterial tissue.

METHODS

Ethical approval for this study was obtained fromthe University of Calgary's Conjoint Health ResearchEthics Board.

Animal ModelThe ascending aortas from 12 healthy New Zealand white

rabbits (8.2 ± 2.0 months, 4.6 ± 1.0 kg) were harvestedimmediately after killing. The choice for this animal modelwas based on the similarity of size and structure between therabbit aorta and the human VA. Some of the similarities areoutlined below.

The average diameters of the rabbit aorta and the humanVA are 5 to 5.5 mm14 and 3.75 to 4.5 mm,15-18 respectively.The human VA greatly varies in size from the average value;it differs from person to person and within a person betweenleft and right.15 The wall thickness of the ascending aorta hasbeen reported to range between 0.18 and 0.20 mm,14 whereasthe human vertebral wall thickness is approximately 0.5mm.19 Both arteries contain the 3 typical structural layers: thetunica intima, the tunica media, and the tunica adventitia. Thetunica intima is thinner in the VA than the rabbit aorta,20 but itprovides little to no resistance to tissue strain,21 and therefore,differences in tunica intima thickness would likely not affectthe stress-strain relationship of these arteries. The tunica

media in the rabbit ascending aorta consists of circumferen-tially orientated elastin with interspersed smooth muscle andcollagen,20 whereas the tunica media of the human VA ismainly composed of smooth muscle, with little collagen andelastin.20,22 The tunica adventitia of the human VA is thickerthan that in the rabbit aorta and is considered its stiffest layer.The adventitia of the human VA is primarily composed oflongitudinally orientated fibers on the inner portion, similarto rabbit aorta, but also has circumferentially orientated fiberson the outer portion.20,23

The elastic modulus for the rabbit aorta is approximately5 MPa24 and between 3.2 and 5 MPa for the human VA.25 Inaddition, the ascending aorta of New Zealand white rabbit isa highly homogeneous tissue ensuring consistent resultsbetween specimens.15,26

Before arterial tissue extraction, excessive fat andconnective tissue were removed from the arterial tissuemechanically with the careful use of forceps and a scalpel.The removal of such material has little effect on themechanical properties of the cadaveric rabbit arterial tissueas the elastic modulus of white adipose tissue has beenreported to be much lower (approximately 28.9 kPa)27 thanrabbit aorta (5 MPa).24 The in situ length of the specimenwas carefully marked with Verhoeff stain (Fig 1) and thenmeasured with electric calipers. The in situ length of eacharterial specimen was defined as zero strain and served as thestarting position for all mechanical testing. All measurementand testing was made according to engineering strain definedas follows:

e =change in lengthoriginal length

=L − L0L0

=ΔLL

whereL = current lengthL0 = original (in situ or zero strain) lengthΔL = change in length

ig 2. Tissue fixation at in situ length after mechanical testing andefore embedding.

16 Journal of Manipulative and Physiological TherapeuticsAustin et alJanuary 2010Arterial Tissue Damage

The ascending aorta was then carefully removed andplaced in a 9% saline solution until testing.

Before testing, each sample was divided longitudinallyinto 2 test specimens and cleansed of further fat. The removalof excess fat and connective tissue from the aortic sampleswas necessary to ensure an uncompromised clamping of thetissue sample into the mechanical testing system. A total of24 rectangular cadaveric specimens (length, 13.0 ± 2.9 mm;width, 6.9 ± 1.2 mm) were used for testing. Histologicanalysis was performed on the midsection of the specimens(length = 7.6 ± 1.2 mm) approximately 3 mm away from theends so that any arterial damage as a result of tissue clampingand extraction were minimized. Arterial strain was definedas the average length change of the tissue from its initial insitu length and was measured between the tissue clamps.

Mechanical TestingSpecimens were divided into 4 groups (n = 6 per group).

Group 1 specimens (zero control group) were kept in 9%saline solution, and no testing or fixation of the tissue into thematerial testing machine was performed. Group 2 specimens(zero strain group) were fixed into tissue clamps and attachedto the mechanical testing system that was used to apply therequired strains to the experimental tissue specimens (858Mini Bionix II, MTS, 1999; Eden Prairie, MN). Afterattachment into the mechanical testing system, the specimensin group 2, along with groups 3 and 4, were all remeasuredusing electric calipers and the length adjusted until they wereset at their initial in situ length. The group 2 specimens werekept at a constant zero strain (initial) length for the durationof a normal test (20 minutes). The specimens of groups 3 and4 were attached to the tissue clamps of the 858 Mini Bionix IItesting machine and brought to their zero strain length.Group 3 specimens were exposed to 1000 strain cycles of0.06 strain (6% elongation) at 0.6 strain/s speed. Theseconditions matched the maximal strain and strain rateobserved for cadaveric VAs during cervical spinal manip-ulations as performed by a chiropractor.7

Group 4 specimens underwent 1000 cycles of 0.3 strain(30% elongation) at 0.6 strain/s speed. The 0.3 strainmagnitude corresponded to approximately 50% of theultimate failure strain observed in human cadaveric VAsfor a single cycle of stretch exposure.7

The number of strain cycles that the cadaveric arterialtissue was exposed to was 1000. This number represents amaximal value for the number of cervical SMTs a patientmay receive within a lifetime. Literature suggests that theaverage number of visits a patient makes to a chiropractorper year ranges from 4 to 18, and it has been suggested thatthe number of cervical manipulations per visit is greater than1.8,28,29 Therefore, if one considers that the maximal numberof visits per year is approximately 18 and assumes anaverage number of 2 manipulations per visit, then thenumber of manipulations a patient would receive in 1 year is

Fb

approximately 36. If we consider a patient who has beenusing chiropractic services for 30 years, then the number ofmanipulations would be approximately 1000. Thus, thenumber 1000 is roughly estimated to be the maximalexpected number of manipulations a patient may experienceover a lifetime. Although we would not expect a patient toreceive 1000 manipulations in a short period (eg, 1 day), wechose this number for the present testing procedures.

After the mechanical testing, samples were carefullyremoved from the testing system and fixed in a 10% neutralbuffered formalin solution at their in situ (zero strain)length using hypodermic needles (Fig 2). The hypodermicneedles were placed at the edges of the sample, ensuringthat the mid part of the tissue was not compromised for thehistologic testing.

Histologic AnalysisSamples from the midportion of the test specimen, to

ensure that any damage to the tissue was caused by the repeatmechanical exposure rather than by the tissue clamping,were processed using a Leica TP 1020 (Wetzlar, Germany)for dehydration and paraffin embedding. The tissue was thensectioned (8 μm) using a Leica RM 2165 microtome.Samples were stained with a Musto Movat and ahematoxylin and eosin stain. The Musto Movat stain targetselastin fibers blue/black, collagen yellow/orange, and fibrinred and, therefore, was used to highlight damage to the fibersin the arterial wall. The hematoxylin and eosin stain targetsnuclei blue/purple, cytoplasm varying shades of pink,collagen pale pink, and muscle deep pink-red, and wasused to identify damage to the arterial wall. A pathologistblinded to the experimental groups devised a scale forhistologic analysis to evaluate microstructural tissue damageassociated with the repeat straining of the arterial tissues. Thepathologist evaluated the stained slides, and arterial damagewas graded as none (0), mild (1), or severe (2). Mild damageconsisted of focal or superficial tearing, splitting, ordiscontinuities within the artery. Severe damage wasassigned when extensive or full-thickness tears or defects

Fig 3. Percentage of occurrence of damage within each experi-mental group: percentage of occurrence of tissue damage (normal,mild, severe) to the arterial specimens after exposure to 1000repeat strain cycles. Significant tissue damage was seen betweengroup 1/2 (control) and group 4 (30% strain) specimens.

17Austin et alJournal of Manipulative and Physiological TherapeuticsArterial Tissue DamageVolume 33, Number 1

were seen. Not only was the pathologist blinded to thegroupings of the samples, she was unaware of the number ofgroups, the mechanical testing performed, and the purpose ofthe study.

StatisticsResults were analyzed using a nonparametric Pearson χ2

test, with a level of significance of .05. Specimens fromgroups 1 and 2 (zero control and zero strain group) werefound to be the same and thus were grouped together (n = 12).The control groups were then compared to each of theexperimental groups (0.06 and 0.30 strain).

RESULTS

Histologic results of group 3 specimens (0.06 strain) werethe same as those found in the control group (zero controland zero strain) specimens (P = .406). Group 4 specimens(0.3 strain) showed significantly more microstructuraldamage than groups 1 and 2 control samples (P = .024;Fig 3). Examples of normal tissue and tissue with severedamage are shown in Fig 4.

DISCUSSION

Groups 1 and 2 control specimens showed somemicrostructural tissue damage. Specifically, both controlgroups contained 4 normal samples and 2 samples with milddamage. The mild tissue damage found in the control tissuesis likely associated with the dissection and handling of thearterial tissues.

Group 3 specimens (0.06 strain) were statistically thesame as the control samples (Table 1 and Fig 3), suggestingthat the repeat straining of cadaveric rabbit arterial tissue

similar to that experienced by the cadaveric VA duringchiropractic neck manipulations does not cause micro-structural damage.

Microstructural damage in group 4 specimens (0.3 strain)was significantly greater than those observed in the controlgroup providing a positive control for our test procedures.This result illustrates that repeat strain exposure of arterialtissues might produce microdamage that is not observed inmechanical assessments of tissue integrity in a singleloading cycle.7

Cerebrovascular accidents associated with neck manip-ulation likely have amechanical origin. It has been speculatedthat chiropractic SMT might cause VA dissection possiblyleading to stroke. Our previous work on the mechanics of theVA in human cadavers during high-speed, low-amplitudecervical spinal manipulative therapy suggests that this is notpossible in a normal VA.7 However, it is well established thatbiological tissues can fail because of microstructural damagewhen exposed to repeat mechanical loading that in itself is notdamaging for a single loading bout.30,31 Thus, it has beenargued that repeat exposure to chiropractic neck manipula-tions might predispose the VA tomicrostructural damage thatmight lead to stroke.13 This study cannot refute this argumentas we used an animal model rather than human vertebralartery and performed instrumented testing rather thanperforming repeated chiropractic manipulations. Neverthe-less, we can conclude from this study that cadaveric rabbitarterial tissue of similar size, structure, and mechanicalproperties to the human VA does not incur microstructuraldamagewhen exposed to 1000 strain cycles of magnitude andspeed corresponding to the maximal values observed in thehuman cadaveric VA during chiropractic SMT to the neck.

LIMITATIONS

There are a number of limitations that need to be kept inmind when interpreting the results of this study. First, thestrain magnitude (0.06) and the strain rate (0.6 strain/s) of thetest specimens were based on a study using cadavericspecimens. This study had a limited number of independentobservations (n = 6) that measured the surface strains in anonperfused artery.7 However, in the absence of any otherdata on the mechanical stresses and strains of VAs duringcervical spinal manipulations, these remain the only data ofrelevance to this study and thus were used here as ourstarting point. Second, the rabbit ascending aorta is not thehuman VA, independent of the similarities in structure andmechanical behavior. Therefore, the results should not betranslated to the human VA without due consideration.Third, the strain measurements of our experimental speci-mens were measured as the average change in lengthbetween the tissue clamps during the mechanical testing andwere not done directly at the site of histologic analysis.Although not ideal, the assumption that the tissue strain washomogeneous across the experimental specimen was made

ACKNOWLEDGMENT

Fig 4. Stained aorta specimens. Aorta specimens with normal histologic appearance (left). Aorta specimens with severe tissue damage(right). Musto Movat stain (upper): elastic fibers are blue-black; collagen, yellow-orange; and fibrin, red. Hematoxylin and eosin staining(lower): nuclei are blue-purple, cytoplasm is varying shades of pink, and muscle fibers are deep pink-red.

Table 1. Number of tissue specimens in each testing groupaccording to their histologic grade (ɛ = strain)

No clamping ɛ = 0.00 ɛ = 0.06 ɛ = 0.30

Normal 4 4 3 1Mild 2 2 2 2Severe 0 0 1 3

Practical Applications

• An animal model is necessary to study arterialtissue microdamage.

• Strains similar to those occurring in the VA duringSMT of the neck were reproduced mechanically.

• One thousand repeat strain cycles mimicking SMTdid not cause microdamage in arterial tissue.

• One thousand repeat strain cycles of a magnitudecorresponding to approximately 50% of ultimatefailure strain (0.30) causes significant microdamagein arterial tissue.

18 Journal of Manipulative and Physiological TherapeuticsAustin et alJanuary 2010Arterial Tissue Damage

and further ensured by preparing the arterial specimens intorectangular pieces. This method was chosen to achieve thenecessary strain while minimizing the damage to the arterialspecimen from any other factor than the mechanical strainapplied. Fourth, the 1000 repeat loading cycles wereperformed in a single session, whereas a patient wouldreceive maybe 2 or 3 repeat manipulations per session beforereceiving further neck manipulations at a later date. In themeantime, the arterial tissue in a patient would have thepossibility to adapt to any imposed loading; thus, theprotocol used here must be considered a worst-case scenarioin producing microstructural damage to the VA that wouldlikely not occur in real life. Finally, the histologic scale usedto grade the experimental tissue was developed specificallyfor this study and has not previously been validated in theliterature. Although the scale has not been validated, we areconfident in its sensitivity. The pathologist, who was blindedto the purpose of the project, the mechanical testing, theexperimental groups, and the number of experimentalgroups, identified significant differences in microstructuraldamage between the 0.30 strain group and the control and0.06 strain groups.

CONCLUSION

Cadaveric arterial tissues of New Zealand white rabbitwith similar size, structure, and mechanical properties ofhuman vertebral artery did not exhibit histologicallyidentifiable microdamage when exposed to repeatedmechanical loading equivalent to the strains observed inhuman vertebral artery during chiropractic cervical spinemanipulative therapy.

The authors thank Tim Leonard and Ruth Seerattan for

technical help.

19Austin et alJournal of Manipulative and Physiological TherapeuticsArterial Tissue DamageVolume 33, Number 1

FUNDING SOURCES AND POTENTIAL CONFLICTS OF INTEREST

No funding sources or conflicts of interest were reportedfor this study. The Canadian Chiropractic Association, theCollege of Chiropractors of Alberta, and the CanadianChiropractic Protective Agency provided financial support.

REFERENCES

1. Haldeman S, Kohlbeck FJ, McGregor M. Risk factors andprecipitating neck movements causing vertebrobasilar arterydissection after cervical trauma and spinal manipulation. Spine1999;24:785-94.

2. Schievink WI. Spontaneous dissection of the carotid andvertebral arteries. N Engl J Med 2001;344:898-906.

3. Saeed A, Shuaib A, Al-Sulaiti G, Emery D. Vertebral arterydissection: warning symptoms, clinical features and prognosisin 26 patients. Can J Neurol Sci 2000;27:292-6.

4. Assendelft WJ, Bouter LM, Knipschild PG. Complications ofspinal manipulation: a comprehensive review of the literature.J Fam Pract 1996;42:475-80.

5. Stevinson C, Ernst E. Risks associated with spinal manipula-tion. Am J Med 2002;112:566-71.

6. Moore KL, Dalley AF. Clinically oriented anatomy. 5th ed.Baltimore: Lippincott Williams and Wilkins; 2006.

7. Symons BP, Leonard T, Herzog W. Internal forces sustained bythe vertebral artery during spinal manipulative therapy.J Manipulative Physiol Ther 2002;25:504-10.

8. Von Kuster T. Chiropractic health care: a national study of costof education, service, utilization, number of practicing doctorsof chiropractic and other key policy issues. Washington, DC:The Foundation for the Advancement of Chiropractic Tenetsand Science; 1980. p. 91-6.

9. Cagnie B, Vinck E, Beernaert A, Cambier D. How common areside effects of spinal manipulation and can these side effects bepredicted? Man Ther 2004;9:151-6.

10. Wynd S, Anderson T, Kawchuk G. Effect of cervicalmanipulation on pre-existing vascular lesion within the caninevertebral artery. Cerebrovasc Dis 2008;26:304-9.

11. Herzog W, Conway PJ, Kawchuck GN, Zhang Y, Hasler EM.Forces exerted during spinal manipulative therapy. Spine 1993;18:1206-12.

12. Triano J, Schultz A. Loads transmitted during lumbosacralspinal manipulative therapy. Spine 1997;22:1955-64.

13. Ontario Hospital Association. Inquest into the death of: LanaDale Lewis. Coroner's jury verdict and recommendations.Ontario, Toronto Canada; 2004. Available at http://www.oha.com/KnowledgeCentreLibrary/CoronersReports/Documents/Lana%20Dale%20Lewis%20Inquest%20-%20Clinical%20Practice%20Guidelines.pdf.

14. Wolinsky H, Glagov S. A lamellar unit of aortic medialstructure and function in mammals. Circ Res 1967;20:99-111.

15. Mitchell J. Differences between left and right suboccipitaland intracranial vertebral artery dimension: an influence onblood flow to the hindbrain? Physiother Res Int 2004;9:85-95.

16. Abd El-Bary TH, Dujovny M, Ausman JI. Microsurgicalanatomy of the atlantal part of the vertebral artery. Surg Neurol1995;44:392-401.

17. Thiel, HW (1991). Gross morphology and pathoanatomy of thevertebral arteries. J Manipulative Physiol Ther 14:133:141.

18. George B, Laurian C. The vertebral artery pathology andsurgery. Austria: Springer-Verlag/Wein; 1987.

19. Sato T, Sasaki T, Suzuki K, Matsumoto M, Kodama N, HiraiwaK. Histological study of the normal vertebral artery. Etiology ofdissecting aneurysms. Neurol Med Chir (Tokyo) 2004;44:629-36.

20. Fawcett DW, BloomW. A textbook of histology. Philadelphia:W.B. Saunders Company; 1986.

21. Nagal A. Die mechanischen eigenschaften der kapillerwandund ihre bezeihungen zum bindegewebslager. Z ZellforschMikrosk Anat 1934;21:376-87.

22. Mann T, Refshauge KM. Causes of complications fromcervical spine manipulation. Aust J Physiother 2001;47:255-66.

23. Piffer CR, Zorzetto NL. Microscopy anatomy of the vertebralartery in the suboccipital and intracranial segments. Anat Anz1980;147:382-8.

24. Sokolis DP, Kefaloyannis EM, Kouloukoussa M, Marinos E,Boudoulas H, Karayannacos PE. A structural basis for theaortic stress-strain relation in uniaxial tension. J Biomech 2006;39:1651-62.

25. Johnson CP, How T, Scraggs M, West CR, Burns J. Abiomechanical study of the human vertebral artery withimplications for fatal arterial injury. Forensic Sci Int 2000;109:169-82.

26. Vernhet H, Demaria R, Juan JM, Oliva-Lauraire MC, Senac JP,Dauzat M. Changes in wall mechanics after endovascularstenting in the rabbit aorta: comparison of three stent designs.AJR Am J Roentgenol 2001;176:803-7.

27. Gefen A, Haberman E. Viscoelastic properties of ovine adiposetissue covering the gluteus muscles. J Biomech Eng 2007;129:924-30.

28. Nyiendo J, Haldeman S. A prospective study of 2000 patientsattending a chiropractic college teaching clinic. Med Care1987;25:516-27.

29. Thiel H, Bolton J. Estimate of the number of treatment visitsinvolving cervical spine manipulation carried out by membersof the British and Scottish Chiropractic Associations over aone-year period. Clin Chiropr 2004;7:163-7.

30. Wang XT, Ker RF. Creep rupture of wallaby tail tendons. J ExpBiol 1995;198:831-45.

31. Wang XT, Ker RF, Alexander RMcN. Fatigue rupture ofwallaby tail tendons. J Exp Biol 1995;198:847-52.