noigroup logo

Convoluted manipulations

By Timothy Cocks Science and the world 22 Aug 2014

In his post, Different Cracks, David Butler asked whether anyone was still “cracking” joints. It generated a lot of interest and remains one of the most commented-on posts on noijam. Comments came from a range of different perspectives and the question of risk and cervical manipulation was inevitably raised.

A paper published earlier this month in Stroke and free to access online has raised the question again and sought to review the evidence and provide a scientific statement from the American Heart and Stroke Associations:

Cervical Arterial Dissections and Association With Cervical Manipulative Therapy 

Conclusions—CD is an important cause of ischemic stroke in young and middle-aged patients. CD is most prevalent in
the upper cervical spine and can involve the internal carotid artery or vertebral artery. Although current biomechanical
evidence is insufficient to establish the claim that CMT causes CD, clinical reports suggest that mechanical forces play
a role in a considerable number of CDs and most population controlled studies have found an association between CMT
and VAD stroke in young patients. Although the incidence of CMT-associated CD in patients who have previously
received CMT is not well established, and probably low, practitioners should strongly consider the possibility of CD
as a presenting symptom, and patients should be informed of the statistical association between CD and CMT prior to
undergoing manipulation of the cervical spine

But scratching the surface of the cervical-manipulation-safety debate reveals a convoluted story. The paper in Stroke was picked up and published in a post by Steven Novella on the Science-Based Medicine blog. The author of the post on SBM questions the Stroke paper’s treatment of a key study, Cassidy et al (2008) published in Spine, that is often used to support the notion that the correlation between cervical manipulation and cervical arterial dissection is NOT causal – suggesting that cervical manipulation is safe.

Novella links to another SBM post by a Mark Crislip in 2008 that took a very close look at the Cassidy et al (2008) article. Crislip makes a number of very strong points against the Cassidy article questioning their conclusions, methodology and, in his opinion, very selective writing of the abstract.

Both of the SBM posts generated hundreds of comments that became very heated, personal, and at times degenerated into ad hominem attacks and flights of logical fancy. If you have time, a read through the comments is quite interesting.

The Science-Based Medicine blog has been very forthright in its approach to the topic of cervical manipulation, particularly in relationship to it’s use by chiropractors. In this post, Jann Bellamy reports on a hearing of the Connecticut Board of Chiropractic Examiners held in January 2010. The hearing was convened to decide whether chiropractors should warn patients about the risk of stroke following neck manipulation and provide a discharge summary listing the symptoms of stroke.

The hearing occurred because two Connecticut women, Janet Levy and Britt Harwe, suffered strokes resulting from chiropractic cervical manipulation. Bellamy’s post recounts the harassment that these two women faced from some chiropractors, which ended up with the FBI getting involvement and the eventual arrest and conviction of one harasser. There’s a transcript of the Superior Court hearing here if you’re really interested and a timeline of the whole story here.

During the hearing J David Cassidy, (of Cassidy et al 2008) gave evidence as a representative of the International Chiropractors Association, and according to Bellamy, the Board relied heavily on Cassidy’s testimony and article in deciding that chiropractors were not required to inform patients of the risk of stroke prior to performing a cervical manipulation.

Fast forward a few years and the British Medical Journal published a piece by Wand, Heine and O’Connell (2012) (unfortunately behind a paywall here) entitled “Should we abandon cervical spine manipulation for mechanical neck pain? Yes” and a press release. The Wand et al paper was criticised by at least one blogger as manipulative cherry-picking of articles by authors that had no track record on the topic (no link, you’ll have to look that one up yourselves if you want to), and there was also an opposing view published in there BMJ by (no surprise) Cassidy, Bronfort and Hartvigsen (2012). A copy of their response is available here and includes a disclosure of competing interests that makes for interesting reading.

The response from Cassidy et al (2012) cite patient preference as a key point in the discussion, reasoning that cervical manipulation must be preferred by many as 6-12% of the population receives it annually. The idea of patient preference might lead one to speculate whether patient’s would still undergo spinal manipulation in the same numbers if the Connecticut Board had decided in 2010 that chiropractors were required to inform their patients of a potential link between cervical manipulation and stroke.

Coming back full circle, what is to be made of the American Stroke and Heart Associations’ scientific statement? They make four key points that I think every clinician should be aware of:

– Although current biomechanical evidence is insufficient to establish the claim that CMT causes CD, clinical reports suggest that mechanical forces play a role in a considerable number of CDs

– Most population controlled studies have found an association between CMT and VAD stroke in young patients.

– Although the incidence of CMT-associated CD in patients who have previously received CMT is not well established, and probably low, practitioners should strongly consider the possibility of CD as a presenting symptom

– Patients should be informed of the statistical association between CD and CMT prior to undergoing manipulation of the cervical spine

What does this mean for a clinician deciding whether to manipulate a patient’s neck or not? Here are my* thoughts

– There IS a risk. Regardless of the statistical likelihood of cervical arterial dissection occurring after cervical manipulation, the catastrophic potential impact puts this risk in the red in my mind.

– Patients may present as a result of symptoms caused by an already existing cervical arterial dissection – a careful assessment and examination and awareness of red-flags is always essential.

– Informing patients of the risks associated with cervical manipulation (while legally mandated in Australia for physiotherapists anyway) is morally and ethically a must, no question at all.

I’ll admit to bias** here (I’ve never manipulated a neck in a clinic- I never saw the need) but, given the magnitude of the risk, with the potential for stroke or death, and the dubious (at best) benefits, I would find it hard to defend this procedure.

Dissenting thoughts and arguments welcomed!

-Tim Cocks

Get your think on and get up to date at a noigroup course or immerse yourself in some brainy books with Explain Pain 2nd Ed and The Graded Motor Imagery Handbook.

*Just want to point out here that I have not undertaken a systematic review nor meta-analysis of the literature findings. I have read the relevant literature and information linked to in the post(s) and with my own admitted bias** have formed my own understanding on this issue – this is not clinical advice or in any way suggestive of clinical guidelines proposed by the author and/or endorsed by noigroup; just sayin’- people have taken posts on this site the wrong way before.


  1. I don’t crack Cx at all (not need to take that risk). Lx and Tx I’ll do occasionally, on request mainly.

    To test whether it works, you can manipulate yourself when you next feel tight or sore (this way you are free of therapeutic alliance effects). My experience is that it definitely works, but that the effect wears off very quickly. Longest effect I’ve had is maybe 10 minutes; the shortest effect, seconds. I have also made an effort to ask people who crack themselves and found a similar pattern of responses – it definitely works, but the tightness and pain returns very quickly.

    Considering this, I’d say I have no problem with people using it in the Tx and Lx, so long as it’s reasonably gentle and infrequent. As always, the critical element in any therapy is the quality of the interaction, not the physical technique.

    Regards, EG

    1. Hey EG, that’s an interesting experiment. While the therapeutic alliance bias may be reduced, I reckon the bias from beliefs and prior knowledge may be greater?

      I wonder if the relief from pain and/or stiffness is purely because the human (who has been primed to believe that manipulation is “good” either through training to be a therapist or being told by a therapist), having heard a pop has enough credible evidence of safety for the constructions of pain and stiffness to be deconstructed for a short time? I’m thinking about Lorimer’s ‘formula’ for pain here which suggests that pain will occur when there is more credible evidence of danger [to the body/body tissues/self] than credible evidence of safety. I think this then goes beyond the notion of placebo/beliefs/expectation.

      The dilemma then is the perennial one of whether an intervention that will only have short-term benefit at best, is delivered in exchange for money – certainly not a trivial matter.


      1. Hi Tim,

        Yes, beliefs could have made a difference in the self-treatment experiment. But I figure if everyone I ask has the same experience with self-manipulation it adds a little weight to the idea that it might be a real effect (real but short-lived). But you’re right there’s no way to know for sure.

        I’ve done the same self experiments with massage and hot packs. Probably everyone reading this has too. The treatments were all pretty much equivalent in effect – they would reduce stiffness/discomfort but only for a very short time. Manipulation was definitely quicker however. Luckily, back soreness usually fixes itself so I don’t need to go seeking treatment anywhere.

        Fear (as a reaction to real or perceived threat) is the only issue I am aware of that needs assistance, and I can do that under the guise of a physical treatment. For acute pain, there’s no other way to do it without getting the patient off-side. To mention stress, threat and fear to someone with acute or chronic back pain is *a threat in itself* for most clients, and this really must be understood.

        Milton Ercikson was very careful never to tweak a client’s structured ego defenses. Take Tiger Woods. Any psychologist (or self-trained radical like myself) can see a mile off what personal stress is causing his pain. But to broach such a subject in session is probably the worst thing you could do. His ego defenses must be respected if any ground is to be made. So long as the therapist knows how to create rapport, Tiger himself will offer up the problem soon enough.

        Physical treatments don’t seem to have any lasting benefit. But psychological ones do. Phsyiotherapy might be better termed Psychosomatics. In the future it probably will be.


  2. Tim,

    It seems perhaps your perceptions of the literature is out of step with the scientific consensus?

    And a deconstruction of the AHA study itself which shows how it was manipulated to state the opposite of the biomechanical data of stress and strain patterns of the VA, ICA and subtly threw in a CMT (i.e. chiropractic manipulative therapy) for good measure.

    If NOI claims to be evidence-informed, then why does it promote an view that is contradicts the literature regarding safety and effectiveness of c-SMT?


    1. Hi Marc

      Thanks for taking the time to comment.

      I’ve written a detailed and long response- please accept this not as an attempt to shut down debate but rather, as a sign of respect for you, your comment, and the topic, and the desire to encourage a robust discussion about ideas.

      I’ve reviewed the links that you provided and have some thoughts on both, but there are a couple of things I’d like to say first. This post is not a position statement by the NOI group on cervical manipulation – I tried to make that really clear in my footnote and at the same time acknowledge my own bias on the subject. This is my post and my views on the subject.

      I will say that when I asked for other views and comments, my use of “dissenting” was sloppy as by definition dissenting suggests that views are different to the “commonly or officially held” view, which I did not wish to suggest my views are – mea culpa.

      Secondly, who gets to decide what the “scientific consensus” is? Is it the AAOMPT to whose statement you linked? Or the Palmer Centre for Chiropractic Research in your youtube link who have promoted a “leading scientist” (using an appeal to authority right there in the video title does not engender great confidence in me) as taking the AHA/ASA to task? Is it you? Is it me? If the debate on the safety of cervical manipulation demonstrates anything, it demonstrates that there is far from a consensus on the topic.

      The AAOMPT make three broad points on the safety of cervical manipulation that I can see in their responses.

      1. The AAOMPT suggest that cervical manipulation is less risky than anti-inflammatory medication, injections or spinal surgery. I think this is a very poor argument – demonstrating that intervention A is less risky than intervention B does not suggest that intervention A is safe. The same argument is made by the leading scientist in the Palmer Centre for Chiropractic research video.

      2. The AAOMPT cite Halderman et al (2002) and state, “Incidents of stroke associated with cervical manipulation of the spine are rare. In a 2002 review of 64 cases of cerebrovascular ischemia, or lack of blood flow to the brain, associated with cervical spine manipulation, researchers concluded that strokes after manipulation appear to be unpredictable and should be considered a rare complication of this treatment approach”. Far from making any comment on whether there are risks associated with cervical manipulation, Halderman et al (2002) start with the a priori assumption that the 64 cerebrovascular ischaemia events were associated with cervical manipulation.

      Suggesting that something is “rare” does not mean it doesn’t happen and does not make it safe. Lawrence Krauss has repeatedly made this point in relation to people making arguments from religion or faith- in a very large universe, rare events happen all the time. How many 10’s of millions of manipulations are performed each year in the US alone? I’ve done the math and using the lowest possible range from the figures used in the Cassidy et al (2012) response it’s upwards of 18 million. A probability of adverse events of 6 in 10 million manipulations (cited in the American Physical Therapist Association position paper on Thrust Joint Manipulation) doesn’t look so remote in comparison.

      The Halderman (2002) article (I’ve retrieved and read it) suggests that cerebrovascular ischaemia is a rare complication of cervical spine manipulative therapy, with onset of symptoms within 2 days of 94%, and 30 minutes within 75% of the cases in the 64 cases they reviewed.

      Additionally, Halderman et al (2002) state that
      “At this writing, a physician wishing to advise a patient considering cervical manipulation is limited in the information that can be offered. Screening the patient for arteriosclerotic vascular disease or risk factors would not exclude a patient from receiving cervical manipulation or guarantee that the patient will not have a stroke after manipulation. It is, however, possible to inform the patient of this risk, to discuss the relative risk of other common treatments for neck pain such as non steroidal anti-inflammatory drugs, surgery, and other common medical procedures, and to explain that according to the current understanding of this problem, cerebrovascular symptoms are unpredictable, inherent, and rare complications of cervical manipulation.”

      The argument that cervical manipulation is less risky than drugs and surgery gets another run, but the statement that cerebrovascular symptoms are unpredictable, inherent and rare complications of cervical manipulation is, frankly, terrifying. Here is an article cited by the AAOMPT in their statement purporting to support the safety of cervical manipulation suggesting that even with the correct screening measures being undertaken prior to manipulation there is no “guarantee that the patient will not have a stroke after manipulation”

      Granted, the authors point out that most vertebrobasilar artery dissections occur in the absence of cervical manipulation, either spontaneously or after trivial trauma or common daily activities (sneezing, tennis, backing out the driveway, yoga), but why take the (obvious) risk?

      3. The AAOMPT statement cites an article by Carlesso et al (2010) suggesting “In addition, a 2010 systematic review in the scientific journal Manual Therapy found no strong evidence linking the occurrence of adverse events to cervical manipulation and/or mobilisation.” The Carlesso et al (2010) paper systematically reviewed the occurrence of adverse events after cervical manipulation/mobilisation during *other scientific studies* (not within the broader, general population) and undertook a meta-analysis of 14 RCTs and 3 Observational Studies.

      The authors state in their discussion:
      “While it is acknowledged that an RCT is not primarily used to detect adverse events, specially those that are rare, the strength of meta-analysis with sufficient data cannot be disregarded.”

      But then later state:
      “We are not suggesting that RCTs are the study of choice to detect harms. In general, large scale observational studies are the most appropriate to detect adverse events and are likely the only way that serious adverse events will be captured”

      The value of citing Carlesso et al (2010) is questionable in the AAOMPT’s statement and does not convince me that the AAOMPT’s position is the “scientific consensus” on this topic.

      While I’ve been involved in online discussions that can end up playing out like a card game as participants try to trump one another with this systematic review, or that RCT supporting their side (it’s exhausting and generally ends with no useful outcome), I’ll add that there are at least 2 recent reviews that I’m aware of that suggest that my position does not “contradict the literature” on the safety of cervical manipulation.

      Puentedura et al (2012) (!po=3.84615) and Ernst (2007) are both open access at those links.

      My view, as presented in the post can be summed up as – there is a risk and patients should be informed of this risk. Additionally, given my understanding of the risks vs benefits of cervical manipulation, I could not defend using this technique. I find no evidence in the AAOMPT’s statement or the video from the leading scientist that suggests this view is out of step or contradictory to the body of literature on the topic.

      I would like to her your view on this topic (I don’t know what it is and I think it would be unfair of me to infer your position from the few links posted), whether you manipulate cervical spines, whether you see any bias in your position, whether you think there is any risk associated with cervical spine manipulation and your clinical and scientific rationale for performing (or not) the technique.

      Thanks again


  3. Hi Tim
    A question for you. I have not read your citations yet but I do plan too. Do you know if there is a difference in risk between end of range cervical manipulation, or mid range manipulation?

    1. Hi bigd7876
      The criteria for inclusion as a manipulative technique is a high velocity, low amplitude thrust technique. Whether this thrust is performed at mid- or end- range as a starting position has not, as far as I can tell, been teased out in the literature in terms of differing risks.


  4. The arguments for and against Cervical manipulation have a tendency to wander off into opposing camps of held opinion, usually dependent upon validity justification for either approach. It might seem that justification is more pertinent than appropriateness. Here are some reasons why I think Cervical manipulations might be inappropriate.

    The neck is basically a self-exercising organ. Normally, its needs for exercise are satisfactorily met with unattended everyday activities….seldom would anyone think they have to apply external methods to restore or improve the functionalities of the neck. If anything, the neck’s most common demand is one of requiring rest from being over-taxed physically. Excluding weight-lifting, discus throwing etc, the neck usually copes quite comfortably with everything we throw at it… must do because it’s the conduit for several vital physical functionalities : breathing, swallowing, blood flow to brain, flexibility to assist the protective senses (seeing, hearing, smelling), and perhaps most importantly of all, protection for the nerve roots before and during their distribution to various localised areas. As such, it’s probably safe to assume that the neck comes with it’s own corrective abilities when any of those functions are threatened. It is probably the most over-worked organ in the body, and it probably has its own means of prioritising the needs of the functions which use it. Quite often, those needs are in conflict with each other, and the neck must somehow provide better functionality for one at the expense of another. There would seem to be a hierarchy of priorities determining which functions take precedence, and I’d be inclined to see that as good reason for the introduction of radiated, referred and mimicked symptoms whenever a nerve is threatened in the cervical area. Any stiffening or pain in the neck itself can negatively influence other functions which must be maintained at some survival level. It’s a trade-off whose formula is , as yet, unknown to us.

    When it comes to manipulations to restore pain free flexibility, we really have to be careful that we are not disturbing this balance of inherent priorities, because we run the risk of creating more complex conflicts, which in turn, the neck must respond to. The ‘stroke’ issue would seem to point to the fact that doing so has the potential to disturb the balance, thus allowing a vital function, blood flow to brain, to become compromised…..something that is unlikely to occur without external interference. That, in itself, should be enough of an argument against manipulations, and particularly against manipulations when the neck is already displaying stiffness due to a threatened or trapped nerve. Do we really know what protective conflicts we might be generating with manipulations, and if we don’t, then the only way forward is to rely on the neck’s own adaptive processes. Until we understand those processes fully, any manipulations are risky and unproven. No manufactured technique should ever be tested without knowing the actual detail of possible negative outcomes, and that seems to be exactly what’s happening with justifications for cervical manipulations for nerve threatened conditions.

    In conclusion, as I see it, the neck has its own nerve response rule-book when it comes to protecting the functions which use it… rise to variable referred responses which don’t over-disturb the balance of prioritised functionality. Unless we fully understand processes, any interference is a ‘shot in the dark’ , only supported by guesswork…..not by science.
    Why anyone would assume that a stiff neck requires applied force to enable de-stiffening, is simply beyond my reasoning capabilities. The only process that I know for easing neck stiffness is the sleep process, and if more effort was put into exploring that simple remedy, we might be able to drag ourselves out of the applied manipulations circular arguments, which only serve to hinder progress in treatments. If ‘Wait and see’ v ‘Applied Manipulations’ is the only argument on the block, then we might be missing out on investigating that ‘Wait and see’ option in the detail it deserves, and thus willingly overlooking the merits of adaptive processes.

Your email address will not be published. Required fields are marked *


Product was added to cart.