Cervical Manipulation and The AHA

The recent statement published by the American Heart Association on the risk of cervical artery dissection (CD) with cervical manipulative therapy (CMT) and the responses to this statement by the APTA and AAOMPT have gotten me thinking about this issue again. One of my biggest concerns is that our profession (namely the manual therapists) are too biased in favor of CMT to objectively assess the risk benefit ratio laid out in the literature. It is very easy to be influenced by anecdotes of successful treatment with CMT and to lose sight of the overall benefits in proportion with the potential risks. It is also easy to cherry pick articles which support our bias by showing positive outcomes with CMT while down playing or calling into question the association with CD risk. In light of this inherent bias within our profession (and within me personally), I find statements from an external stake holder such as AHA very compelling. Just like surgeons should heed our concerns about excessive and unwarranted surgeries, we should take very seriously the opinions of other intelligent, well meaning experts on theAHA issue of CMT and CD. It is important that we fight the reflexive urge to be defensive and look at this statement by the AHA as an opportunity to challenge our biases and take a fresh look at the issue. For even more outside perspective here are some other opinions on the topic.

I was not surprised when I read the response of the APTA and AAOMPT to this article. I think they both highlight the bias and protectionism that our profession has towards cervical thrust manipulation. In this post I am going to examine the arguments put forth by the APTA and AAOMPT in defense of CMT. This is a complex issue and I look forward to hearing the opinions of others on the topic.


When I started to read the statements by the APTA and AAOMPT (which are very similar) I was encouraged that both letters acknowledged the likely connection between CD and CMT.

AAOMPT-“The orthopaedic manual physical therapist community recognizes the rare but significant risks associated with cervical manipulation”

                APTA-“The physical therapy community recognizes the rare but significant risks associated with cervical manipulation

Although arguments can be made against the correlation of CD and CMT, it does seem highly plausible that they are related. Case reports (such as this, this, this, this, and this) also suggest that CMT can cause CD (smoking gun), however this cannot be determined with certainty from these case reports and the case controlled studies we have on the topic.

After acknowledging the correlation between CMT and CD both statements go on to make some pretty weak arguments in defense of CMT. For example, both statements make comments about mitigating the risk of CD.

                Quote from Tim Flynn: APTA“In fact, physical therapists in the United States, along with our international colleagues, have developed a clear framework designed to mitigate this risk”

                AAOMPT:” In fact, the AAOMPT, in collaboration with the International Federation of  Orthopaedic Manual Physical Therapists (IFOMPT) has studied this issue in great detail and   previously provided an international framework for mitigating risk associated with these      techniques. This framework is reflective of best practice, intending to place risk in an appropriate context that is informed by the available evidence.”

They have developed a framework to mitigate risk, however I have seen no convincing evidence that this framework is effective at reducing risk. The current evidence does not provide us with a way of predicting who will suffer a CD from CMT. Many of the reported cases of CD after CMT occur in young healthy subjects with no major risk factors for arterial disease. In fact, later on in the APTA statement they say

                “In a 2002 review12 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.”

So they appear to be unpredictable yet we have a framework which can mitigate the risk? Are we fooling ourselves? I guess if we are going to continue with CMT we need to show that we are putting forth an effort; however, we are far from proving that any measures other than abstinence can reduce the risk of CD with CMT. At best, the framework developed by the IFOMPT will help to identify patients who are already having an ischemic event prior to CMT.

Both statements then make a Tu quoque argument, under the guise of context. They bring up the fact that commonly used interventions such as medications, injections and surgery also have significant risks. I agree that these treatments have risk (even a greater risk than CMT) however this argument is one sided and may be seen as an erroneous argument because it is irrelevant to the safety/efficacy of CMT. I would suggest however, that context is not totally irrelevant, although the APTA and AAOMPT articles did not present a balanced view of the context. Both statements mentioned interventions with greater risk but they did not address the effectiveness of those interventions and they did not address the effectiveness of interventions with a lower risk than CMT. Bringing up this issue, in this way, simply points out that we need to perform a well thought-out risk/benefit analysis of all the interventions available for a given condition. If an intervention’s potential benefit is not worth the risk of the procedure then its use should be questioned, irrespective of context.

My next concern with these statements pertains to their framing of the potential benefits of CMT. It took me a few reads before I noticed that all three articles cite the same systematic review but make very different points. All three articles spoke about a 2007 systematic review in the Journal of Rheumatology from Gross et al which looked at conservative management of mechanical neck pain. The AHA stated this about the article:

Data on the effectiveness of CMT for neck pain are sparse and questionable. The most recent Cochrane review consisting of some low-quality trials found that the effect of cervical manipulation was comparable to that of mobilization, which does not include a thrust”

Both the statements from the APTA and AAOMPT paint a different picture.

AAOMPT-“ To date, many scientific studies support cervical manipulation for the treatment of   head and neck pain of mechanical origin. One systematic review in the Journal of Rheumatology3 from 2007 reviewed 88 randomized controlled trials and concluded that exercise combined with manipulation or mobilization demonstrated both pain relief and functional improvement in adults with acute, subacute, or chronic mechanical neck disorders.”


So despite their critique of the AHA article as being “limited in its scope and context” the AHA paper was the only one of three which pointed out the important difference between thrust and non-thrust CMT techniques. Given CDs relationship to trauma and force, the most defensible position is to assume that thrust techniques are more likely to result in CD than non-thrust. There are also several case reports documenting CD after thrust and a lack of case reports documenting CD after mobilization. I do realize that I am making an assumption by saying thrust is more dangerous than non-thrust CMT. There is no direct evidence which proves this, however it seems to be a reasonable position given the indirect evidence which is currently available.

Ignoring this difference between thrust and non-thrust CMT is a big flaw in the APTA and AAOMPT’s statements. I am sure that most stakeholders looking at this issue are thinking about thrust techniques, not non-thrust. As the AHA statement rightfully points out, there is no convincing evidence that thrust techniques to the cervical spine are more effective than the, likely much safer, mobilization techniques.

As stated above I am not surprised by the statements from the APTA and AAOMPT. Both of these organizations represent practitioners not patients. Most of the time their mission to represent PTs is in line with the best interest of the patient. However, in this case I question whether they are sacrificing (probably unintentionally due to bias) the best interest of patients for the benefit of the clinicians they represent. I see no reason why the AHA has any interest in the topic other than to promote safety and efficacy for patients. I feel that the AHA’s statement was well constructed, well supported, and I strongly disagree with the statement that the AHA’s position “may place unnecessary fear in the public of an effective low risk treatment intervention.”


I feel very strongly that patients should be educated about the risks involved in any procedure they may undertake and the AHA’s statement provides a well balanced resource for this education. When offering thrust manipulation to the cervical spine as a treatment option it is important that we discuss the risks and benefits with our patient. How we frame that conversation can have a large impact on the “informed” part of the consent. For those of you who perform cervical thrusts, what do you tell patients about the risks and potential benefits of the technique?






21 replies »

  1. I understand what you are saying, and have a lot of sympathy for the stance you are taking. I just have some concern that what may end up happening is that because spinal manipulative therapy is generally seen as no more effective than mobilisation, that manipulation will ceased to be advocated and taught.
    However, it is effective, and is a useful tool in the toolbox of manual therapists in treating pain of joint origin. Back pain and neck pain is difficult to treat, hence why despite much research spinal pain is still a significant cause of time lost from work and a significant cost to healthcare systems worldwide. Why would we want to remove a tool that we can use to help reduce this cost?

    When treating patients, and manual therapy is indicated, I offer manipulation, mobilisation and activator as options for treatment (along with other modalities), and allow the patient to decide. I inform them that no one treatment is better than the other. I inform them that there is a risk of stroke with manipulation of the neck and that present research suggests this is around a 1 in 100,000 chance, possibly slightly more in people under 45. I then allow the patient to decide. If the patient is unsure, I suggest trying mobilisation first, with the option of changing to manipulation should mobilisation not be getting the results the patient desires

    I understand that allowing the patient more active involvement in the choices of their treatment improves outcomes. I have also had patients change as they are not responding to one form of manual therapy (going both ways)

    Surely this happens in all healthcare professions. Offering one form of treatment which may be safer or more effective, but if the patient reacts or doesn’t respond, changing tack and trying something else which maybe less effective or have more risks in the research, but may end up being more effective for that individual.

    I hope this is a reasonable and sensible approach to incorporating evidence when deadling with individual patients.

  2. D. Back,

    Thanks for your comment. I think your approach sounds very reasonable and I have no major problems with it. The only thing that I may do different is not wait for them to be unsure before I biased/influenced the patient to try mobilization first. Unless the patient was adamant that thrust be tried first, I would feel much more comfortable (in the cervical spine) trying mobilization/exercise first.

    It may be a slippery slope to keep an intervention because it may appear (in the clinic) to produce outcomes in a small number of patients. If that is our standard then we would be keeping around just about every intervention. I am sure that rubbing peanut butter on some patients will reduce their neck pain, especially if both the patient and the clinician really believe in it. However, I am not going to keep a jar of jiffy in my toolbox. I know that is an extreme example but where we draw the line is a complicated question.

    • But surely the rubbing of peanut butter (smooth or chunky? I prefer Nutella myself ;-)) would be shown to be no more effective than placebo, whereas manipulation has been shown to be an effective intervention. I don’t feel using manipulation (with present research) keeps me in the dark ages and that i am using it in blind faith, but i incorporate the information in informing my patients. At the end of the day, the risk is a small one.
      I feel the line would be if manipulation were to be proved ineffective, then we would have to consider it. But it is a complicated question.

  3. D Back,
    Nutella is a good idea :).

    “would be shown to be no more effective than placebo, whereas manipulation has been shown to be an effective intervention.” This is exactly the issue, manipulation has not been shown to be more effective than placebo, or many other commonly used interventions. The AHA statement points this out. Even when a study finds that thrust manipulation is superior to whatever they have compared it to, the effects sizes are often pretty small. And we have seen what happens when we combine studies of manual therapy with small effect sizes (Menke 2014). It gets dangerous when we have the attitude that a treatment needs to be proven ineffective before we stop using it. It should be the other way around. It should to be proven effective before we use it and I would argue that thrust manipulation has yet to proven as effective.

    I do agree with you that it is a complicated issue.

  4. Thank you for a thoughtful post and thoughtful comments below. Really all I can add is that we need to improve 1) our diagnosis of neck pain – we are in the dark as to what is causing most peoples’ neck pain, so it is no wonder that no treatments are shown to be particularly efficaceous for a heterogenous group. We need to identify subgroups most likely to respond to manipulation so that it is not offered to those not likely to benefit and, as such, improve clinical outcomes in both groups (responders and non-responders). Clinical prediction rules so far have not proven fruitful but that doesn’t mean we should not continue to try and identify useful rules (with tighter methodology than that employed in the existing literature). Equally, or perhaps more usefully, the identification of objective parameters (e.g. reduced inter-vertebral motion) could help guide the appropriate targeting of treatment. 2) related to the previous point we need to better understand the mechanism of manual therapies. If they work in different ways then presumably that is a case for targeting treatment. If they work through a common mechanism (as has been suggested) then they need not be targeted. We need to know which it is. We need more research, and we need more money to fund it.

    My personal opinion is that the risks associated with CMT will turn out to be exaggerated (once we have the requisite data which is lacking) however I also believe the therapy is over utillised in the treatment of spinal pain and alternatives to manipulation should be used/offered more often. In fact, slightly parallel to this discussion, I think patients need to be offered much less in the way of any passive treatment (risking the creation of dependency and reinforcing unhelpful health beliefs) and more active treatment (giving patients ownership over their problem).

  5. Read the IFOMPT vertebral artery guidelines and stop this foolish back and forth. The odds are so low and if you take proper precautions you have a higher chance of injuring someone with traction or a hot pack than a c-sp hvlat.

    • Stevenak66,
      Sorry you feel that this conversation is foolish. I am sure any patients who were unlucky enough to have a stroke after manipulation would disagree with you. As I addressed in my post there is no evidence to support that following the IFOMPT guidelines will reduce the risk, especially when considering the evidence reporting that we have no way to predict these events. Also, as I addressed in my post tu quoque arguments are irrelevant to this discussion.

  6. The thing that I would like to share with all of you,is either we want it or not there are no methods or global techniques which are suitable to all patients!a good healthcare professional must be first of all well trained in the specific method by global education institutions,second to have the ability to evaluate the patient with an holistic approach with the rest of healthcare staff and family and no with only statistically parameters and finally when he reaches at the end to apply a technique he has to calculate the effect of his approach and if it’s the ideal for his patient.in conclusion I believe after 10 years as physiotherapist that there are not bad or good techniques.there are bad professionals,unsuitable patients for techniques and finally non well trained professionals to be able to establish which intervation is the appropriate to every condition!greetings to all of you.

    • Giorgos,

      I disagree that all treatments have their place. There are good and bad techniques and there are even more techniques. It is very easy to say we just need to be “holistic” and we need to identify “unsuitable patients” however how do you do that? What is your evidence to support your ability to differentiate between unsuitable and suitable patients. Jonathan asked a very good question above “What is the clinical indication for thrust or high-velocity manipulation?” There is not a clear and agreed upon answer to this question. I am curious how you would answer that question.

    • Thanks for sharing harrypersson. As mentioned in the AHA article and my post this video also points out some remaining questions about this issue. She brings up the Cassidy study which is an interesting study but it has some significant problems which have been addressed by others:

      She also makes the same Tu quoque fallacy by trying to create context by bringing up the risks of another treatment. If you are going to create context then you need to look at treatments which are both potentially less and more risky and also compare their effectiveness. As stated above, with likely rare but significant risks with cervical thrust and the lack of evidence for its effectiveness (especially long term and compared to other interventions) it seems reasonable that it is at least not a first line treatment in most cases.

  7. First of all as about thrusts and high velocity manipulations I gave to say that in clinical practice I haven’t apply them.because of the legal status of my country,because is something that is not for everyday use so I didn’t have the ability to improve this technique so I can not speak if it’s something so important and in what situations.the only thing that I know is that there are studies with unclear results and some that say that there are and side effects.to be honest we have to see these studies by whom are written and what training he has past.for sure we will find different thoughts from physio’s,chiropractors and doctors!now for the subject of holistic approach I want to explain that i mean when we conclude in every technique we have to be sure that she is the preferable and the most appropriate for our patient.i said that because during my ten year working experience we see healthcare professionals how trained in one method afterwards many of them is like something blocking their mind!they see to all patients the necessity of what they learn.that make them losing their target.i won’t say that they might harm the patient but to apply unuseful treatment and modalities.many times by tacking an extended history of our patient in many musculoskeletal conditions we will see that what we say and we call to apply a technique is only to deal with the symptom.the truly cause of the problem maybe something else.for example a bad daily habit which with the necessery exercise program or ergonomical advise plus the communication with our patient,maybe more effective with duration rather than making the magic technique which indeed it will help the patient,but the results will be short-term.i believe that writing this opinion I am answering to you about the suitable and no suitable patient for one technique.i consider as rule that we have take in our mind the general precautions and red flags for every technique that we are planning to use.

  8. Two interesting articles:



    “SBM” is linked to ACSH( That is an “industyfriendly”frontgroup). See this link(and the links in that article):


    So you comment on post you do not accept. I guesss you are a member of the “sceptic” movement.

  9. harrypersson,

    The reputation of SBM is not relevant to this discussion. I welcome thoughtful critiques of the SBM link that I posted above. Here I will help you with some pointed questions. Why do you disagree with the points made in the article? Do you feel the criticism of the Cassidy article is unfair and incorrect and why?

    I could have easily dismissed the Cassidy article and the video you posted because they are both produced by chiropractors who are very likely to have a bias in favor of manipulation. However, I felt that would be a poor tactic and choose to point out specific flaws (or provide links that point out flaws) in the reasoning portrayed in the video and in the article.

    “So you comment on post you do not accept. I guesss you are a member of the “sceptic” movement.” I have no idea what this means. You and I are both commenting on a blog post that I wrote. If you define sceptic movement as people who do their best to make conclusions based on evidence and critical thinking then yes, I do my best to be a member of that group.

  10. It is the data presented, that determines whether if an article is good or not. I searched Mark Crislip (MC) on PubMed, but did not find anything. But the article can of course even though contain valuable information.

    In the beginning of MCs blog-article MC says: . “In the blogosphere, the proponents of chiropractic often quote the following paper, with the abstract:” Why do MC only give a link to the abstract, when the full story is available for free online? It is an important study read it!


    MC said: “It is striking how many young people had a stroke just after having chiropractic neck manipulation.”

    Has MC some facts? Facts proving his claim that: “many young people had a stroke just after having chiropractic neck manipulation.” Some facts would be nice facts proving assertion that: “many young people had a stroke” after “neck manipulation”.

    Not least because of what the statistics say: Vertebral artery dissection (VAD) is about 1 in 5.85 million adjustments ( It is more likely to die from lightning ).

    It is not wrong to compare risks with other forms of treatment for head and neck pain. On the contrary it is important when choosing the best method of treatment for patients.

    MC is trying to make it likely that HVLA techniques can cause stroke by providing a link to a study that involves Vertebral Artery Dissection after a FRACTURE of the neck, which is something other than a spinal dysfunction, which can be treated with HVLA techniques.

    MC makes an incorrect comparison of the diagnostic criteria that patients had prior
    stroke with diagnostic criteria for stroke ???

    The study involves visits to the MD or chiropractor before they got the stroke, and the
    patients had symptoms before the stroke, which are those that can also be caused by a vertebral artery dissection.

    Sometimes it looks like MC does not know what he’s writing, or do not care about it. Why( I gave you a possible explanation in this link http://buggesblogg.blogspot.se/ )?

    Bias does not determine whether an article is good or bad, but it can provide a possible explanation for
    why a lousy article is lousy.

    MCs article suffers from a serious lack of facts, but consists largely of propaganda and suspicions against HVLA- manipulations, and they have no probative value in a scientific discussion about the advantages and disadvantages of neck treatments.

    There are many other great scientific works of HVLA and neck problems why not discuss them instead of discussing vacuous propaganda articles?

    Here are three articles:




    The stroke topic is important, and there is a lot of research, and it does not point to a causal relationship between HVLA and stroke, but research must of course continue.

    PS I do not define “sceptic movement” as people who do their best to make conclusions based on evidence and critical thinking
    , and MCs article describe what they are. See: http://buggesblogg.blogspot.se/

  11. HarryP,

    I will give you one more chance at this. I linked to the article by Mark Crislip to highlight one point. Flaws in the Cassidy study which may limit the conclusions we can make based on the results of that study. You have done everything but address the specific criticisms made by Crislip against the Cassidy study. Maybe you are correct and his criticisms of that study are in error. However nothing you have written so far provides evidence for that.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s