A Prevention of the Vertebrobasilar Accidents Following Cervical Thrust Manipulations

Recommandations of the SOFMMOO

Jean-Yves Maigne, MD, Hôtel-Dieu Hospital, Paris, France

Home page English Section / Home page sofmmoo.com

Page Content : Recommendations of the SOFMMOO / References


It is now a well established fact that cervical thrust manipulations can harm the vertebral artery. This accident was formerly regarded as very rare, although severe, and related to atherosclerosis. Clinical tests were proposed to detect patients at risk. This problem is now better known. It is no longer attributed to atherosclerosis (and ageeing process) but to a dissection of a vertebral artery, a clinical entity observed in younger patients (20-45 years). It remains very rare, but mild symptoms appear to be not so unfrequent. Finally, the predicting tests seem to be deprived of any value.

What are the risks of a cervical manipulation?

The largest published series was that of Assenfeldt et al. They reported 165 cases of vertebrobasilar accidents, and 13 cases of vascular accidents in other territories. The mean age was 38 years, and there was a female predominance. Hurwitz et al published a smaller series of 118 patients. In 55 cases, they were able to specify the manipulation technique involved, which was, in 82% of the cases, a thrust technique with rotation. The most recent study is that by Haldeman et al, reporting on cases that had given rise to claims of malpractice. Their conclusion was that this complication appeared to be unpredictable, a concept which will be discussed below.

These series allowed some authors to determine the prevalence of vertebrobasilar accidents. For Dvorak (1985) et al, the figure was one serious accident per 400,000; for Henderson (1988), none out of 500,000; for Patijn (1991): one in 519,000; for Carey (1993): one in 3 million; and for Klougart (1996): one in 414,000 (rotation techniques). A study by Lecoq and Vautravers, in France, showed an estimate of one reported accident for 5 millions of manipulations. But in a latter study by Dupeyron et al, it appears that this first figure was largely underestimated, because a vast majority of accidents are not recorded by the statistics of the insurance companies, and only the most impressive ones have been published. For these authors, a more realist figure would be one severe accident for 150,000 manipulations.

Also, there are minor cases of dissection that do not necessitate hospitalization and there must be other cases that remain asymptomatic because of the adequacy of collateral blood supply. In a comprehensive assessment of the complication rate, these cases should, strictly speaking, be considered.

What are the clinical symptoms of a vertebro-basilar accident?

The mean age is 37 years and the sex ratio is 2/3 of females. A very large majority of the patients are under 50 (in the series from Haldeman et al, the most aged is 51).
The latency to onset range from immediate to 48 hours after the manipulations. Cases with an interval of time of 7 days have been described, but there is no absolute proof.
The symptoms range from a mere headache with nausea and neck pain to a more severe clinical syndrome, associating vertigo, vomiting, and sometimes neurological deficits. They consist in visual signs (diplopia, blurred vision, gaze paresis, Horner’s syndrome…), motor symptoms (hemiparesis, tetraparesis, facial palsy, dysarthria…) and in some very rare cases, in vital signs (coma, locked-in syndrome...) Of note, a lateral neck pain and some headache can be the only symptoms of a spontaneous dissection.

Clinical course of cerebro-vascular accidents

From the 126 cases reported by Terrett (1992), 35% resolved without residual and, oppositely, 23% resulted in death. In the smaller series from Hufnagel (1999), on 10 cases, 5 resolved, 3 remained with marked deficits and 2 with locked-in syndrome or vegetative state. These are ancient series, and one can assume that the prognosis is now better, but these figures point out the potential severity of the vertebrobasilar accidents. For information only, we (JYM) recall of four patients having had a definite vertebrobasilar accident following a cervical manipulation (and consulting us a long time later for another reason) in the last decade. None of them had actual significant sequellae (and, naturally, none were reported in the literature). We also recall of two others patients, one with an accident following a standard clinical examination of the neck by a physiotherapist, and the other with a spontaneous dissection (of the carotid artery), without any known risk factor.

What are the causes of cerebro-vascular accidents?

There is a vulnerable portion of the vertebral artery at the atlantoaxial joint. A rotation of the neck can stretch the artery to a certain extent, depending on the range of rotation, leading to a shearing force at the vessel wall. Thus, in some specific cases where the arterial wall is very thin and fragile, a tear of the intima may appear. A cascade of events may follows: intramural hematoma, thrombus formation and stroke. This condition is known as a vertebral artery dissection. Unfortunately, these cases (of fragile artery) are not identifiable before the occurence of the dissection. The injuring rotation may be provoked by a manipulation of the neck or even by a mere clinical assesment of the range of motion or done by the patient himself/herself (turning around...) But, anyway, the thrust carries an extra risk.

There is a simple way to explore the effects of different neck movements on the artery. When the vertebral artery is stretched to a certain extent, the blood flow may slow down or even stop (naturally, the brainstem remains supplied by the anastomosis of the polygone of Willis). The blood flow can be recorded by ultra sounds (Doppler). A study by Haynes has compared the respective effects of rotation and lateral flexion. In 148 patients, the blood flow was stopped in 5% of the cases with rotation, and always persisted in lateral flexion. This means that rotation is the real harmful movement.


Prevention has always been a major concern. Based on the fact that these accidents were formerly attributed to atherosclerosis clogging the artery, many tests were developped which were supposed to compromise the blood flow and to stop it if the artery was severly stenosed. In such a case, clinical symptoms would appear, the presence of which would make the manipulation contra indicated. All these tests combined extension and rotation from 30 seconds to 3 minutes. They were the Dix & Hall-Pike, Hautant, Kleyn, Maigne, Stejskal, and Unterberger tests. Rancurel test consisted in a manual compression of the artery in the suboccipital area. None of them was validated.

Other solutions have been proposed. Imaging techniques (Doppler, MRI) have no real predicitive value. Obtaining an informed consent of the patient is certainely the best commendable way. But how to inform a lay person about the possibility of a dissection, the risk of an ischemia of the brainstem the clinical significance and the rarity of such a complication? This choice has been made by the Australian Physiotherapy Association, which proposes this set expression: « I wish to manipulate your joint using a quick movement in the position in which I am holding your neck. You may hear a click and this is normal. Neck manipulation can be dangerous but this is extremely uncommon. I have carried out the recommended precautionnary tests and in my opinion, there is little risk in your case. Are you agreeable for me to go ahead? ». Objections can be raised: what are the recommended tests and their reliability? What mean "dangerous"? What is a "little" risk? This could be an illustration of the difficulties encountered when informing the patient.

Other proposals include forbidding cervical manipulations, which should also lead to ban NSAIDs, aspirin and analgesics (and surgery) because of their complications, or, oppositely, to consider cervical manipulations as a mere risk factor, among others, of a vertebral artery dissection, without any firm conclusion (see Haldeman et al).

Recommendations of the SOFMMOO

The French Society of Orthopaedic and Osteopathic Manual Medicine brought at a round table the prominent personalities in the field of French Manual Medicine in 1997. After a thorough presentation of the problem by various speakers (including anatomists, neurologists and radiologists), they agreed on the proposals made by the author, now known as the recommendations of the SOFMMOO. They were published:

Acknowledging the fact that prevention is out of reach, the aim of these recommendations is to reduce the number of (not to say to suppress) rotational cervical thrust manipulations in a targeted population. This population consists mainly in females of less that 50 years old.
Five recommendations were developped, in addition to classic contraindications of spinal manipulative therapy.

Recommendation #1

Seeking any undesirable effect following previous manipulative neck treatment such as nausea, headache, dizziness or vertigo. They could testify of a previous dissection with a favourable spontaneous outcome. This is an absolute contraindication to further cervical thrust manipulation.

Recommendation #2

No thrust manipulation for recent (i.e. acute) neck pain (less than 3-4 days), because it may be a symptom of a spontaneous dissection of the vertebral artery.

Recommendation #3

Neurologic exam mandatory before any cervical thrust (same reason as #2: the risk of a current dissection).

Recommendation #4

No cervical thrust in rotation in females less than 50 years. No cervical thrust in rotation in males less than 50 years at the first visit (but allowed at the 2nd visit if the first treatment was not efficient). Instead of rotational thrust, it is highly recommended to use mobilisations, MET (muscle energy techniques), soft tissue cervical techniques and upper thoracic spine thrust manipulations (which certainly act on the cervico thoracic muscles).

Recommendation #5

Only physicians with a Universitary Diploma passed at least one year before should be allowed to perform cervical manipulations. This latter recommendation should be adapted to the context of foreign countries. The idea is that a physician should not be allowed to thrust a neck without at least one year of full practice. Such an interval of time may allow him or her to feel comfortable and confident with other techniques (thoraci and lumbar spine).


Top of the page