J Manipulative Physiol Ther 2006 (Mar); 29 (3): 236—144
Donald R. Murphy, DC, DACAN, Eric L. Hurwitz, DC, PhD, Amy A. Gregory, DC
Rhode Island Spine Center, Pawtucket, RI, USA. firstname.lastname@example.org
OBJECTIVE: The purpose of this study is to present information from a series of patients with imaging findings of encroachment on the cervical spinal cord who were treated with chiropractic cervical manipulation.
CASE SERIES: There were 27 patients (18 females, 9 males; age range, 23-65, mean age, 44.3 years) with neck and/or arm pain with findings of cervical spinal cord encroachment on magnetic resonance imaging. None of these patients had severe or acute myelopathy or advanced signal changes in the spinal cord indicative of myelomalacia. These patients were treated with a variety of approaches that included some form of cervical manipulation. The mean number of treatments that included manipulation was 12 (range, 2-32). Nineteen patients were treated with high-velocity, low-amplitude "thrust" manipulation, 9 patients were treated with low-velocity muscle energy technique, and 1 patient was treated with both methods. The mean patient-rated subjective improvement at the last follow-up reexamination was 70.0% (range, 10%-100%). From baseline to the last follow-up examination, the mean improvements in outcome measures were as follows: Bournemouth Neck Disability Questionnaire, 23.7 points (31%); Neck Disability Index, 6.4 points; and Numerical Pain Rating Scale, 3.9 points. In 3 patients, there was increased pain after manipulation that lasted from 1 to 4 days. There were no major complications, and in no patient did any increased pain after treatment last more than 4 days. No new neurologic symptoms or signs were seen in any of these patients.
CONCLUSION: The finding of cervical spinal cord encroachment on magnetic resonance imaging, in and of itself, should not necessarily be considered an absolute contraindication to manipulation. However, because radicular and myelopathic complications to cervical manipulation have been reported in the literature, great care should be taken in all cases, particularly those in which anatomic conditions such as cord encroachment are present.
From the Full-Text Article:
Cervical manipulation is a treatment that has received increased attention in recent years, both for its alleged therapeutic benefits and its alleged potential for harm. Experimental evidence suggests that cervical manipulation can be an effective tool as part of an overall approach to the management of patients with neck pain [1-4] and headache. [5-10] Less is known about the role that cervical manipulation can play in the management of patients with radiculopathy, though some evidence exists that it may be helpful. [11, 12] However, the rate and types of complications of this treatment are largely unknown. There are some who feel that the risks of cervical manipulation outweigh the benefits,  and that cervical manipulation in the presence of a disk protrusion is dangerous or contraindicated. [14, 15] Saal et al,  in a study of nonsurgical management of patients with cervical radiculopathy secondary to disk protrusion, made a point of stating, “Forceful joint manipulation was not used.” Haas et al,  in an article reporting on a randomized controlled trial of cervical manipulation, listed disk protrusion and cervical radiculopathy as, “… contraindications to manipulation.” Others, particularly practitioners who use manipulation, feel that this type of treatment is quite appropriate, even in the presence of a disk protrusion. 
It is important for practitioners of manipulation to be aware of those conditions in which manipulation is indicated or contraindicated and those that should cause the practitioner to use particular caution or use an altered technique. It is also important for physicians who refer patients to practitioners who use manipulation to be aware of those conditions in which manipulation, in skilled hands, is relatively safe. Specifically, with regard to patients with disk protrusion and/or spinal stenosis, there is little in the literature that guides the referring or the treating practitioner as to whether, and when, it is safe to use manipulation.
The purpose of this article is to present information from a series of patient records that had imaging findings suggestive of spinal cord compression, with or without clinical signs of myelopathy, who were treated with an approach that included manipulation. Because imaging signs of cord compression may be considered by some to be an absolute contraindication to cervical manipulation, the clinical outcome and complications and adverse reactions in this patient population are reported.
It would appear from these data that the finding of cervical spinal cord encroachment on MRI should not be considered an absolute contraindication to cervical manipulation. Even in patients with mild myelopathic signs, this treatment may be provided in some cases. However, because the finding of cord encroachment does not necessarily indicate cord pathology, it cannot be said that all cases of cord encroachment can be treated safely. None of the patients treated here had severe or acute myelopathy. None had advanced signal changes in the spinal cord indicative of myelomalacia. Because spinal cord and nerve root related complications have been reported after cervical manipulation, appropriate clinical reasoning must be used regarding the presence of pathologic changes in the spinal cord before deciding to use cervical manipulation.
Several reports can be found in the literature of patients who have developed signs and symptoms of cervical radiculopathy and/or myelopathy after cervical manipulation. Rinsky et al  reported a 44-year-old patient with “severe ankylosing spondylitis” since the age of 18 years who fell in his kitchen and developed neck pain and “tingling” in his left arm. He consulted a chiropractor and received cervical manipulation. He developed immediate left-sided weakness that progressed to tetraplegia. At surgery, fracture through the spondylotic fusion at the C3-4 interspace was found, with cord compression at that level. A similar case was reported by Schmidley and Koch.  A patient with ankylosing spondylitis fell to the floor while getting out of bed, developing weakness in his legs and stiff neck. He consulted a chiropractor and received cervical manipulation after which his symptoms worsened and involved the upper extremities. He was found to have cervical spondylotic myelopathy, which was treated with surgery.
Davis  reported 2 cases of patients who developed myelopathy after cervical manipulation by osteopaths, although in 1 of these cases, there was evidence of extradural mass in the cervical cord that apparently represented metastasis from a lung carcinoma; however, this article is short on details. Kewalramani et al  reported 3 patients who developed signs and symptoms of myelopathy after cervical manipulation. In 1 case, radiographs showed fracture of the C6 vertebral body and unilateral facet luxation at C5 through C6. In another, surgery revealed subarachnoid hemorrhage. In the third patient, chronic osteomyelitis was found. In none of these patients was it clear what role the manipulation played in the progress of the disease, though the manipulation may have hastened this progression. Destee et al  reported a patient who developed myelopathic signs and symptoms, including bladder dysfunction, 3 days after cervical manipulation. He was found to have large protrusions at C5-C6 and C6-C7, the latter of which was located partially intradural. He was operated on, apparently successfully. Tolge et al  reported a patient with phrenic nerve dysfunction after cervical manipulation. No anatomic explanation was determined in this case. Tseng et al  reported 2 cases, 1 in which a patient developed quadriparesis and difficulty with voiding after the first manipulative treatment and the second in which a patient had been manipulated 3 times before and developed severe neck, shoulder, and arm pain after the fourth manipulative treatment. The type and level of training of the manipulative practitioner was not identified in either case.
Misra et al  reported a case in which a patient developed spinal cord and brainstem dysfunction immediately after manipulation by a barber. Magnetic resonance imaging showed a spinal cord neurofibroma at the C1 through C2 level. Interestingly, despite the fact that the manipulation in this case was delivered by a barber, the keywords for this article included “chiropractic manipulation.” This is consistent with other reports that have appeared to implicate “chiropractic manipulation” in cases of postmanipulative complications that do not involve chiropractors.  Equally noteworthy is the fact that these authors state  that “neck manipulation can be dangerous and should be avoided.” Chung et al  reported a patient treated by a bonesetter who, “suddenly and forcefully rotated his head to one side and then the other side.” The patient immediately developed “numbness of his whole body and difficulty of breathing.” This was attributed to contusion to the upper cervical cord and lower brainstem brought about by the manipulation. Segal et al  reported a case of a young and otherwise healthy woman who developed sensory disturbances and bowel and bladder difficulty starting 15 minutes after cervical manipulation by a chiropractor. No speculation was made as to what could possibly have been the mechanism in this case, and the authors acknowledged that this was an “extraordinarily rare” complication to cervical manipulation. Lipper et al  reported a case of a woman who developed Brown-Sequard syndrome after her third session of cervical manipulation by a chiropractor, which resolved completely with steroids and immobilization. Padua et al  reported on 4 patients with radicular and/or spinal cord lesions after treatment with manipulation. Two of these had central disk protrusions, and the other 2 had cervical spondylotic myelopathy.
Malone et al  published the largest series, a retrospective chart review from a 6-doctor neurosurgical practice. In it, they report 22 patients whom they describe as having had a significant complication to cervical manipulation. Of these, 21 had radiculopathy, 11 myelopathy, 2 Brown-Sequard syndrome, and 1 vertebral artery occlusion secondary to far lateral disk protrusion.  Because of this study's design, it is impossible to tell if any of these cases represented true complications to manipulation or were just part of the natural history of the condition.
Firm conclusions cannot be drawn from these studies. Gradual progression of radiculopathic and myelopathic signs and symptoms is not uncommon. [40, 41] Thus, it is impossible to determine in these reported cases whether the progression resulted from the manipulation or was incidental to it. Nonetheless, it is certainly reasonable to state that spinal cord and/or nerve root injury is possible after cervical manipulation. However, the original data presented in the current study would suggest that the finding of spinal cord encroachment should not be considered an absolute contraindication to this form of treatment. All patients in this series were treated with cervical manipulation, among other treatments. There were no major complications, and the only adverse reactions were short-lived, transient increase in pain in 3 patients, which resolved in all 3 cases. Not only was manipulation not harmful in this patient population, but clinical improvement was seen.
Three cases of transient increase in pain of 27 is a rate of 11.1%. This is substantially lower than the 34% rate of transient symptoms reported by Senstad et al  in a study of 85 patients. It is not clear as to why there was such a great difference in rates of transient pain increase after treatment between this study and that of Senstad et al,  although the different designs may preclude comparing the studies. Nonetheless, it should be emphasized that great care should be taken by the practitioner in applying cervical manipulation in all cases, particularly in the presence of anatomic findings that have the potential to lead to injury. A meticulous approach to positioning, premanipulative assessment, and manipulative technique would seem prudent in minimizing complication in cases such as those reported here.
All of the patients in this case series were placed in the “premanipulative position”  before instituting manipulation. With the use of the premanipulative position, a contact is taken in preparation for manipulation as close as possible to the segment for which manipulation is intended, and the cervical spine is moved into the intended direction of manipulation. Movement in this direction continues slowly until the practitioner feels the onset of the “restrictive barrier,”  that is, the point at which resistance to further movement is perceived by the practitioner. The movement is then stopped, and the patient is asked if this position causes pain or any other symptoms. Particularly, in patients with radiculopathy, the patient is asked if this position causes “peripheralization”  of symptoms, that is, increase or extension of symptoms into the arm as opposed to the cervical area only. If peripheralization of symptoms occurs, that movement is not chosen as the desired direction of manipulation, and a different direction is taken.
The first direction of manipulation that was typically chosen in these patients was A-P rotation, because this is the direction that, in the experience of the first author of this study, is least likely to cause peripheralization of symptoms. This movement involves attempting to make a contact on the anterior portion of the vertebra on the side of involvement, then moving the cervical spine into slight flexion, lateral flexion away from the side of contact, and slight rotation toward the side of contact, which, in most of the patients reported here who were treated for radiculopathy, was also the side of radiculopathy. A possible explanation for decreased likelihood of symptom peripheralization with this direction of manipulation is that flexion combined with lateral flexion away from the involved side has both been shown to increase the size of the intervertebral foramen,  and while rotation toward that side tends to decrease the size of the foramen, rotation combined with flexion increases foramen size (Nuckley DJ, personal communication, July 2, 2002). In addition, the slight flexion of the cervical spine as a whole may allow the practitioner to avoid extension at the targeted level, thus, avoiding any spinal cord compression that may result from extension.47 It should be noted, however, that the flexion movement should be slight to avoid increased axial stress on the cord.  Nonetheless, it is felt that if peripheralization of symptoms with this direction of movement is not seen, it is unlikely that careful and skilled manipulation in this direction will compromise the nerve root.
The 2 types of manipulation used on the patients in this case series were HVLA manipulation and ME manipulation. High-velocity, low-amplitude manipulation involves applying a short quick thrust to a targeted intervertebral segment. This is designed to increase what is perceived to be restricted joint play  in the joints of that segment. Muscle energy manipulation involves the use of isometric contraction, eye movements, and breathing in an attempt to first facilitate, then inhibit, the muscles around the vertebral segment being manipulated so that movement can be introduced to the segment without the need for a HVLA thrust. The treating practitioners in this study (DRM and AAG) use ME procedures primarily in those patients who are uncomfortable with the “cracking” sound that is typically associated with HVLA procedures, and because clinical experience has suggested that ME procedures are less likely to cause increased pain or other complications. Interestingly, the perception that ME procedures are less likely to produce increased pain is not supported by the data presented here, in that in those 3 patients who reported transient increased pain after manipulation, 1 was treated with HVLA, 1 with ME, and the third with both methods. However, the small sample size does not allow firm conclusions to be drawn about this.
Scott-Dawkins  compared HVLA with ME procedures in 30 patients with neck pain. It was found that the patients treated with HVLA manipulation experienced a greater immediate relief of pain, but that at the end of the treatment period, pain intensity decreased in both groups, with no difference between the groups. In another randomized controlled trial of 336 subjects, Hurwitz et a [l2] compared HVLA manipulation and low-velocity mobilization that involved non-ME,2 “…low velocity, variable amplitude movements applied within the patient's passive range of motion …” They found that both treatments produced reduction in pain severity and in disability (measured by the NDI) at 2, 6, and 13 weeks and 6 months of follow-up. No difference in outcome was seen between the groups. Patients treated with HVLA manipulation were more likely to report transient minor discomfort than were patients treated with mobilization. Importantly, this had a negative impact on satisfaction with treatment.  Thus, it would seem prudent to use ME or mobilization methods in those patients in whom potential for increased pain or other complication is highest.
The patients in this case series were treated by practitioners with a minimum of 4 years of training and between 2 and 16 years of experience in the daily use of cervical manipulation. The findings cannot necessarily be applied to practitioners with a lesser degree of training and experience with the use of this tool. Also, practitioners with this level of training and experience should be confident in their manipulative skill (and realistic with this confidence) before attempting to treat patients with the findings presented here.
A weakness of the study is that the data on complications and adverse reactions on 10 of the 27 patients were gathered retrospectively by abstracting the patient's chart, rather than being gathered prospectively, as was the case with the remaining 17 patients. Although this could potentially lead to bias (eg, an underestimate in the frequency of complications), it is doubtful that a significant complication would have occurred in any of the patients that would not have been recorded in the chart. There were no appreciable differences in the clinical presentation and outcomes between those patients followed prospectively and retrospectively; thus, we believe that differences in data collection methodology did not materially affect the results.
This study was done using a case series design, in which a series of patients is evaluated in the absence of comparison to a control or comparison group. The primary purpose of this study was to assess complications to treatment, for which case series designs are considered to be valuable.51 Absence of a control or comparison group limits interpretation with regard to treatment outcomes; therefore, firm conclusions cannot be drawn with regard to the outcomes data reported here.
The findings from this case series suggest that cervical spinal cord encroachment, as seen on MRI or CT, should not be considered an absolute contraindication to chiropractic cervical spine manipulation, provided the manipulation is applied by an appropriately trained and experienced practitioner, and is performed with the utmost care and skill.