Cervical manipulation: How risky is it?

My coauthors and I regret that Brian Lecker and Tim Pethrick found our commentary¹ so disturbing. We are primarily interested in collecting cases of probable dissection of the cervical arteries following sudden neck movement. The incidence of stroke following cervical manipulation, estimated by Lecker and Pethrick to be 1 per 1–3 million manipulations, is surely conjectural and in any case is irrelevant. We need careful studies that include rigorous follow-up and investigation of all cases of stroke following therapeutic neck manipulations. Only then we will be able to estimate the true extent of this problem.

We agree that there are other risk factors for stroke and that strokes may occur, by coincidence, following any activity.

In our study, however, we are only interested in documenting cases of cervical artery dissection. We have found that there are about 50 cases per year of stroke associated with cervical artery dissection and that manipulation of the cervical spine is associated with 27% of these.

Risk factors for stroke must be distinguished from risk factors for cervical artery dissection. There is no evidence that migraine, diabetes or smoking are risk factors for dissection. Our most recent, as yet unpublished, analyses confirm these findings; the only certain risk factors are neck trauma and in some cases genetic abnormalities of the vessel wall. Neck pain is the hallmark of the arterial tear in most cases, both in our study and in all other published studies. It is sudden and severe, and easily distinguishable from the chronic pain seen in patients attending for neck manipulation.

It might be helpful to attempt a case–control study, as Lecker and Pethrick suggest. However, the research question is not whether neck manipulation can result in dissection of a cervical artery, for it surely can, but rather whether some types of manipulation have a lower risk of dissection than others. We also need careful research to document the efficacy of neck manipulation as a therapy. Only then will we be able to weigh its cost, in terms of risk of cervical artery dissection and stroke, against its benefits.

We agree with Robert Sydenham that absence of proof is not proof of absence. Members of the medical profession and those manipulating necks for whatever reason must try to find the reason for these occasional tragic accidents and not try to pretend they never happened.

Alan O'Connor raises some interesting points. First, we must reassure him that all data were collected prospectively, but, of course, after the injuries occurred. Patients were questioned in hospital regarding details of their clinical picture and laboratory tests were performed accordingly. As he questions, what delay can one accept between neck movement and later stroke? There are several autopsy reports of causal dissection with thrombus months after neck injury. We believe that the answers to O'Connor's other questions may be found by studying the materials on our Web site (www.strokeconsortium.ca/PG.08.spontads.html). Many of our findings are posted as they emerge and we are striving to keep the site up to date.

We thank our colleagues at the Canadian Memorial Chiropractic College, with whom we have had numerous helpful discussions. We initially attempted to share all data with them but came across the barrier of patient confidentiality, which we have not yet managed to overcome.

Signature

John W. Norris
Professor of Neurology Stroke Research Unit Sunnybrook & Women's College Health Sciences Centre Toronto, Ont.