Research Review By Dr. Michael Haneline ©


Download MP3

Date Posted:

January 2013

Study Title:

Assessing the risk of stroke from neck manipulation: A systematic review


Haynes M, Vincent K, Fischhoff C et al.

Author Affiliations:

University of Western Australia, Australia; Société Franco-Européenne de Chiropratique (SOFEC), France; Institut Franco-Européen de Chiropratique, Paris, France

Publication Information:

International Journal of Clinical Practice 2012; 66: 940–947.

Background Information:

Strokes resulting from vertebral artery dissection (VAD) and internal carotid artery dissection (ICAD) are known to occur following cervical spine manipulative therapy (cSMT). It has been suggested that cSMT, as well as a host of other neck movements, are capable of triggering cervical artery dissections (CADs) with consequent cerebrovascular accidents in susceptible individuals. It has also been suggested that cSMT may cause expansion of a thrombus and/or produce an embolisation and consequential stroke. However, there are still doubts about the level of contribution from cSMT.

The exact incidence of CAD-related stroke is unknown, but fortunately it is generally reported to be rare. And, in persons who are afflicted, there is thought to be an inherent fragility of the arterial wall that is thought to be caused by factors such as genetic predisposition, biomechanical abnormalities, and biochemical imbalances.

Patients with developing VAD or ICAD often complain of neck pain and headache, so they may present to a manual musculoskeletal practitioner for treatment. Studies that have investigated the level of risk of stroke from cSMT are conflicting, with some suggesting a strong association and others suggesting that a strong association is absent (7). However, it has been reported that several of these studies have major limitations that render their results inconclusive.

The purpose of this review was to update a systematic review by Rubinstein et al. (1 – published in 2005, before RRS began) on cSMT and stroke, as well as to assess the quality of newer studies on the topic, and to determine whether there is conclusive evidence of a strong association between cSMT and CAD-related stroke.

Pertinent Results:

A search of PubMed yielded 159 citations. From these, two of the authors determined that the same five abstracts should be included. No additional relevant papers were found in the papers’ bibliographies, the other databases, or searches using different terms.

Synopses of the papers that were included in this study are as follows:
  • Rothwell et al. (2) conducted a retrospective population-based nested case–control study where cases were derived from hospital records and exposures were derived from health insurance records. The cases had vertebrobasilar occlusive stroke, however an unknown proportion of the strokes involved VAD. Exposure to cSMT was determined via records of cervical spine-related visits to chiropractors. The comparison group was comprised of matched non-stroke controls. A strong measured association was found within 1 week of the stroke in patients aged < 45 years, which was 3.6% of cases compared with 0.9% of controls [OR crude = 3.94 (95% CI = 0.99–15.78)].
  • Smith et al. (3) also conducted a retrospective nested case–control study, but in their study CAD patients and matched controls who had stroke from other causes were interviewed. Exposures to cSMT within 30 days of stroke were compared between cases and controls, which showed a strong association. Exposure to cSMT was 14% among cases compared with 3% in controls [OR = 6.62 (95% CI = 1.4–30)].
  • Dittrich et al. (4) conducted a case–control study wherein data were prospectively collected via interviews. Exposure to cSMT and other types of neck movement were identified. Cases were patients with CAD and controls were matched patients with other types of stroke. cSMT within 7 days of the CAD was reported by 12.8% of cases compared with 6.4% of controls. However, the odds ratios were not significant [OR crude = 2.1 (95% CI = 0.5–9.1); OR adj = 1.5 (95% CI = 0.3–6.9)].
  • Cassidy et al. (5) conducted a population-based case–control and case-crossover study that attempted to control for confounding due to the pain associated with VAD by comparing exposures of cervical related visits between chiropractors and primary care practitioners (PCPs), as well as differences in health status between chiropractic and PCP patients. Positive associations were observed for chiropractic patients aged < 45 years, especially for visits within 3 days of the VAD. However, similar associations were observed for patients who had PCP visits.
  • Thomas et al. (5), conducted a case–control study using hospital records to identify CAD cases which were compared with matched controls who had suffered other types of strokes. They asked subjects about exposures to cSMT and recent head or neck trauma within 3 weeks of the stroke. A strong association for cSMT was found, with 23% of cases reporting cSMT compared with 4% of controls [OR crude = 12.80 (95% CI = 1.58–104.3), OR adj = 12.7 (95% CI = 1.43–112.0)]. An even stronger association was found with recent head or neck trauma, with 64% of cases reporting head or neck trauma compared with 7% of controls [OR crude = 25.5 (95% CI = 5.71–96.9) and OR adj = 23.5 (95% CI = 5.71–96.9)].
Assessments of quality of the studies showed that confounders had affected each of them, with Dittrich et al., the only prospective study, attaining the highest rating. All of the studies had at least three confounders and/or biases.

Clinical Application & Conclusions:

All of the studies were subject to confounders and biases which led to their results being rather imprecise. Thus, any conclusions about the association between cSMT and CAD that are drawn from these studies would be equivocal. According to the authors… “conclusive evidence is absent for a strong association between cSMT and CAD, and is also lacking for no association.”

Given that there is uncertainty about the relationship between cSMT and CAD, the authors asserted that informed consent is warranted to advise patients who will receive cSMT that there may be an increase in the risk of a rare form of stroke, which also applies to other neck movements. However, the potential risks of cSMT should be placed in context with gastrointestinal complications (including many deaths) that are associated with NSAIDs that are commonly prescribed for neck pain.

Even though there is uncertainty about the relationship between cSMT and CAD, stroke following cSMT is rare. Therefore, when patients elect to receive care that involves cSMT, their treatment preferences should be respected.

Study Methods:

This was a systematic review in which the authors searched multiple databases for various types of studies that dealt with CAD. However, case reports, case series, abstracts and letters were excluded. In order to be included, the eligibility criteria required that the studies:
  • had a population with a confirmed or assumed diagnosis of CAD;
  • had a control group;
  • had individuals exposed to specific incidences of cSMT or mild neck trauma, noteworthy neck movements or positioning; and
  • were full reports.
Studies were excluded if the CADs were the result of surgery, arteriography or major trauma.

The reference sections in the included papers were searched and assessed for eligibility in order to reduce the effect of publication bias.

The quality of the eligible papers was judged using a rating instrument similar to the one used by Rubinstein et al. in their review.

Study Strengths / Weaknesses:

The authors discussed several weaknesses that they observed in the included studies which are listed below.
  • Information bias – Where there were potential inaccuracies in the data that were gathered or in the assumptions that were drawn from the data. For instance, both Rothwell et al. and Cassidy et al. relied on hospital records to identify vertebrobasilar occlusive stroke patients as cases, but some of these strokes were likely not due to dissection. Another example is the use of health insurance billing codes to determine chiropractic and PCP visits, because cSMT may not have been used at some of the chiropractic visits, whereas some PCPs may have administered cSMT during their visits.
  • Confounders – Such as VAD pain (patients with VAD headache and neck pain are likely to seek care involving cSMT), the use of stroke controls (because less healthy, stroke-prone individuals often have a lower socio-economic status and are less likely to visit a chiropractor), preferential extreme headache presentation to PCPs, sporadic binge drinking and acute infection patients (which are associated with CAD) are more likely to be PCP patients.
  • Bias – Especially recall bias, where patients who have suffered CAD following cSMT and told that they have a tear in their artery are more motivated to participate in a study and more likely to recall cSMT than controls who have other types of stroke.
Almost all of the biases and confounders that were mentioned tended to increase the calculated odds of association.

Additional References:

  1. Rubinstein SM, Peerdeman SM, van Tulder MW, Riphagen I, Haldeman S. A systematic review of the risk factors for cervical artery dissection. Stroke 2005; 36: 1575–8.
  2. Rothwell DM, Bondy SJ, Williams JL. Chiropractic manipulation and stroke. A population -based case-control study. Stroke 2001; 32: 1054–60.
  3. Smith WS, Johnston SC, Skalabrin EJ et al. Spinal manipulative therapy is an independent risk factor for vertebral artery dissection. Neurology 2003; 60: 1424–8.
  4. Dittrich R, Rohsbach D, Heidbreder A et al. Mild mechanical traumas are possible risk factors for cervical artery dissection. Cerebrovasc. Dis 2006; 23: 275–81.
  5. Cassidy JD, Boyle E, Cote P et al. Risk of vertebrobasilar stroke and chiropractic care: results of apopulation based case-control and case-crossover study. Spine 2008; 33(Suppl 4): 176–83.
  6. Thomas LC, Rivet DA Thomas LC et al. Risk factors and clinical features of craniocervical dissection. Man Ther 2011; 16: 351–6.
  7. Reggars J, French S, Walker B et al. Risk management for chiropractors and osteopaths: neck manipulation and vertebrobasilar stroke. Aust Chiropr Osteopathy 2003; 1: 9–15.