Extreme rotation of the atlas on the axis (at the atlantoaxial joint) stretches the vertebral artery.

Extreme rotation of the atlas on the axis (at the atlantoaxial joint) stretches the vertebral artery.

The American Heart Association and the American Stroke Association recently published in the journal Stroke a thorough analysis of the evidence for an association between cervical manipulative therapy (CMT) and both vertebral artery dissection (VAD) and internal carotid artery dissection (ICAD). The full article is online: “Cervical Arterial Dissections and Association With Cervical Manipulative Therapy: A Statement for Healthcare Professionals From the American Heart Association/American Stroke Association.” For background, an arterial dissection is essentially a tear in the inner lining of the artery. This tear disrupts the normal flow of blood, and also causes platelets to gather at the site of injury. This can result in a blood clot at the site of the dissection. This blood clot can block flow through the artery, or it can break off and lodge downstream, blocking flow at that point. Dissections, therefore, can result in a stroke (a lack of blood flow to a portion of the brain causing damage). There are four arteries in the neck that bring blood from the heart to the brain, two carotid arteries in the front, and two vertebral arteries in the back. A dissection in one or more of these arteries is associated with 2% of all strokes, but with 8-25% of strokes in patients <45 years old. This is mostly because strokes associated with processes like atherosclerosis are much less common in the younger population. Arterial dissections are classified as either spontaneous or traumatic. Trauma can be either severe, such as whiplash injury from a car accident, or subtle, such as from yoga or simply turning one’s neck to look past the shoulder.

There is an ongoing controversy as to the relationship between CMT and cervical dissection, either ICAD or VAD. Chiropractors, who are one of the main practitioners of CMT, have generally denied a causal connection, while neurologists have been more cautious. As the reviewers point out, it is now clear, based on several studies, that there is an association between CMT and cervical dissection. What is not clear is if the CMT is causing the cervical dissection. They write:

In summary, a few case-control studies suggest that CMT is associated with CD. These studies did not specifically distinguish whether the CMT included a thrust maneuver or not; the former is typically used with chiropractic manipulation. In the absence of prospective cohort or randomized studies, the current best available evidence suggests that CD, especially VAD, may be of a low incidence but could be a serious complication of CMT. Although these studies suggest an association, it is very difficult to determine causation because patients with VAD commonly present with neck pain, which may not be diagnosed prior to any CMT.

There is one study in particular, the Cassidy study, which is often cited as the evidence that the correlation between dissection and CMT is not causal. Specifically, they found that patients are just as likely to visit a medical practitioner as a chiropractor in the weeks prior to a stroke. Therefore, the authors argue, patients go to the doctor for the neck pain, rather than CMT causing the neck pain through dissection. While the review was otherwise thorough, I was disappointed in their treatment of the Cassidy study. Our own Mark Crislip did a much better job, pointing out the many flaws in the Cassidy study. Specifically, they included older strokes and a longer time frame, both resulting in a statistical dilution of cases where CMT might have caused VAD. When you look only at stroke patients <45 and within 24 hours of having CMT, there is a clear association – about a 5-fold increased risk of VAD. Other information discussed in the review is also illuminating. For example, the ratio of ICAD to VAD in spontaneous dissection is 2:1, but following CMT the ratio is 3:1 in favor of VAD. This can have two interpretations – either CMT causes some VAD, or that patients with VAD are more likely to seek treatment for neck pain than patients with ICAD. Both conditions result in neck pain, but VAD is more posterior so it is not entirely implausible that this difference in symptoms is partly responsible for the increased association between VAD and CMT. They also discuss a possible mechanism. There is no question that CMT stretches the vertebral artery, although in most cases probably not enough to cause dissection. This makes sense, as VAD is a rare complication of CMT (even if we assume that 100% of the association results from CMT causing VAD). In the rare cases of VAD associated with CMT, this may result from aggressive CMT, or from the particular of the patient’s anatomy.

Conclusion: Cervical manipulation probably causes strokes

The authors of the new review conclude:

Clinical reports suggest that mechanical forces play a role in a considerable number of CDs, and population controlled studies have found an association of unclear etiology between CMT and VAD stroke in young patients. Although the incidence of CD in CMT patients is probably low, and causality difficult to prove, practitioners should both strongly consider the possibility of CD and inform patients of the statistical association between CD and CMT, prior to performing manipulation of the cervical spine.

This is a typically conservative conclusion, and it’s difficult to fault the authors for being so cautious. I do think they are understating the probability of a causal relationship, however, probably because they were overly influenced by the Cassidy study, which they should not have been given its fatal flaws. That aside, we can state clearly that, although rare, VAD is associated with CMT. There is a very plausible mechanism for injury, and it is clearly known that even mild trauma can cause VAD. Further, there are numerous case reports of young healthy patients developing symptoms of VAD, including stroke, shortly after their CMT, sometimes immediately. It is true that none of this proves a causal connection, because we can’t do randomized trials (although prospective trials would be helpful). But this is true in the same way that we can’t prove smoking causes lung cancer, we only have an association. When we have a clear association with multiples lines of evidence suggesting the most likely causal explanation, and that explanation points to a medical risk, then I think it is reasonable to act on that risk. We recommend that patients do not smoke. We should also recommend that they do not have CMT, especially high cervical manipulation and forceful manipulation. Medicine is a game of risk vs benefit, and so considering the risk is not enough. What is the benefit of CMT for the specific conditions it is used to treat? The evidence for the benefit of CMT is less than the evidence that it causes VAD. A Cochrane review of CMT for neck pain and headache concluded:

Done alone, manipulation and/or mobilization were not beneficial; when compared to one another, neither was superior.

So CMT does not work, but perhaps may have some benefit when added to medical management, and even then it is no better than the gentler mobilization. Therefore, since CMT (if it works at all) is not superior to mobilization, and may involve a rare but serious risk of VAD and stroke, it seems to me it is unethical to perform CMT for neck pain or headache rather than mobilization. It can further be noted that, even if chiropractors are correct in saying that most people with VAD and CMT presented with an existing VAD causing neck pain, that still does not justify CMT. In such a case (someone with a VAD and neck pain) the neck should not be manipulated at all because such manipulation could very plausibly provoke a stroke. Cases of suspected VAD need emergent medical evaluation and treatment. The current review is further evidence that CMT should be eliminated as a medical procedure.


Posted by Steven Novella

Founder and currently Executive Editor of Science-Based Medicine Steven Novella, MD is an academic clinical neurologist at the Yale University School of Medicine. He is also the host and producer of the popular weekly science podcast, The Skeptics’ Guide to the Universe, and the author of the NeuroLogicaBlog, a daily blog that covers news and issues in neuroscience, but also general science, scientific skepticism, philosophy of science, critical thinking, and the intersection of science with the media and society. Dr. Novella also has produced two courses with The Great Courses, and published a book on critical thinking - also called The Skeptics Guide to the Universe.