Cervical Spine Manual Therapy & Stroke

After considering the discussion that emerged from this blog post, and these cartoons, I thought it would useful to write a follow up piece, to provide readers with information to consider if / when they find themselves in a situation where they offer or receive manual therapy to the neck. The debate that preceded this piece, and others like it typically have boiled down to a difference in interpretation of data that describes associations between manipulation of the neck, and the occurrence of a vertebral artery stroke. Typically, one side of the debate takes the stance that the risk of stroke is not identifiable, and when weighed against probable benefits, is not worth applying. The other side of the debate tends to interpret the data in a manner that provides scope for the mitigation of any risk of stroke associated with application of thrust manipulations to the neck. The stance essentially, is that there are clear benefits to cervical manipulation, that outweigh any risk of stroke. The risk of stroke being caused or catalyzed by manual therapy to the neck is, in practical terms considered negligible, so long as a thorough history and examination are completed.

In this recent debate, the contribution of Roger Kerry, a researcher in the field of cervical artery pathology as it relates to manual therapy, was particularly interesting. Roger took an admirably neutral position as the debate unfolded, and raised the point that debates surrounding the issue of stroke and manipulation of the neck, should shift away from the question “Does neck manipulation cause stroke?”, and towards a discussion of how best to go about identifying patients who present to a manual therapist with signs and symptoms of an early stage stroke event. He suggested that his research team’s assessment of the literature points to the idea that the association between manipulation of the neck and stroke is best explained (epidemiologically) by the failure of a manual therapist to recognise early signs of stroke in a patient presenting with musculoskeletal-like neck pain and / or headache. He also indicated that his research team’s assessment of the literature suggested that the association between thrust manipulation and stroke has been over-reported, relative to stroke events associated with non-thrust manual therapy interventions.

I am unaware that there exists a published association between non-thrust manual therapy (such as mobilization techniques) and stroke. I am still yet to uncover any of the evidence that Roger Kerry has pointed to. It’s unclear to me if the claim is that there exists evidence that demonstrates an association between non-thrust manipulation and stroke, or if it has been surmised that the association between stroke and thrust manipulation and the association between stroke and non thrust manipulation would probably be similar if equal efforts were made to report both scenarios. If it is the latter case, I don’t think that constitutes a reasonable base upon which to make a claim that the type of manual therapy intervention that is applied is irrelevant to any association that exists between manual therapy and vertebral artery stroke.

However, I do acknowledge that the question of whether or not the association between stroke and thrust manipulation, is greater, equal or lesser than the association between stroke and non thrust manipulation becomes a moot point if we proceed from the position that the data available is insufficient to explain either association in a cause and effect manner. I appreciate that it is from this position that Roger Kerry’s research is proceeding. The development of an updated framework to better assist the piecing together of disparate pieces of subjective history and objective findings to help form a clinical picture of the patient whose symptoms may be indicative of a stroke, is obviously likely to be very valuable for patients and their manual therapists. So I wish Roger Kerry all the best with his ongoing research.

Having said that, I still find myself stuck on the idea that it is unreasonable to shift completely away from the “Does a thrust manipulation specifically, or manual therapy generally, cause stroke?” Obviously what is required is more informative data. But in the absence of data, I think it is incumbent upon health professionals to carefully weigh the potential risks of an intervention against its likely benefits. It is this risk versus benefit assessment that I think many manual therapists err.

Although it is very very very unlikely, it is plausible that the next patient I see on Monday could have a thrombus that is just waiting to be dislodged. Or, he or she may have a congenitally weak artery wall that is susceptible to dissection. Of course, it is possible that such a tenuously poised thrombus could also be dislodged during a simple head turn while driving. Likewise, a spontaneous dissection of a congenitally weak artery wall could occur from, say, forceful cervical extension during a sneeze. Again, I am at pains to point out that the likelihood of these events are very very very very small, perhaps 1 in 500000. What i do know however, is that such an event won’t result from me performing a thrust manipulation or end range extension and rotation mobilisation to the upper cervical spine… because I just don’t do it. I play it safe because I have no way to know for sure, even with the most thorough clinical assessment, if my next patient might be the patient on the verge of a stroke. If a sneeze could be enough to form an embolus, I don’t see why I should start adding additional forces to the neck unless I am really sure that the outcome is going to be worth it. Given that I tend to see patients whose neck pain is most often self limiting, some over a longer time course than others, I don’t see a need to push for the quick fix. I think that in the absence of clear data demonstrating risk size, the plausibility of risk should be considered against the probability of benefit.

The issue is that the risk I am guarding against is too small to express numerically. For some manual therapists that might mean the risk is negligible. If those same manual therapists happen to adhere to clinical processes that operate from an assumption that any specific manual therapy intervention is ever “necessary”, well, we conceivably have scenarios emerge where a clinical presentation (according to the the therapist) demands a specific intervention, for which the risk of stroke (according to the therapist) is “negligible”. That is dangerous. We end up with is a risk/reward analysis that lacks critical judgment, in that it overstates the actual benefit (no specific manual therapy intervention is ever necessary), while dismissing potential risk as negligible. I think that for professions whose basic tenet is “do no harm”, the burden of proof for risk is very low, particularly when the actual benefit is unclear (good short term, equivocal long term for the individual patient… and potentially harmful for the broader culture in the long term). I’m not calling for a ban on any specific techniques. I think we need to be able to regulate ourselves within our professions, and think critically and conservatively, when there is a paucity of data to guide us.

What I am suggesting, is that those who utilise thrust manipulation or end range mobilisation techniques on the cervical spine, attempt to step outside of their clinical habits, and take a broader view of what it means to claim that a given manual therapy technique is ever “indicated” or “necessary” during the clinical encounter. When we do this, we give ourselves an opportunity to try “doing less” to patients. And we give ourselves the opportunity to observe how patients progress without applying the intervention that we might otherwise have described as “necessary”. From that position, it becomes a whole lot easier to simply choose interventions that are probably lower risk. The biggest challenge, i think, is convincing one’s self that one’s clinical processes, which have probably always produced good outcomes, with no serious adverse events, can benefit from some modification.

I look forward to writing an additional post, focusing on outlining Roger Kerry’s ideas, and providing a bit of a clue as to ow to proceed with assessing patients for the purpose of differentiating neck pain and headache derived from the CNS response to nociception from the cervical artery versus other more typical, and less consequential sources of nociception.



Leave a Reply

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