An Etiologic Model to Help Explain the Pathogenesis of Cervical Artery Dissection: Implications for Cervical Manipulation

A Conceptual Etiologic Model of CAD 

Figure 1 represents a conceptual etiologic model of CAD. The 4 major categories of risk factors associated with CAD, which have been presented in an earlier publication,2 are depicted as ovals in this diagram: (1) genetic predisposition/underlying familial disorder (eg, connective tissue disease, hyperhomocysteinemia, migraine, vessel abnormalities), (2) environmental exposure (eg, infection, oral contraceptive use), (3) trivial trauma (eg, common neck movements, sporting activities, manipulative therapy to the neck), and (4) common risk factors associated with atherosclerosis (eg, hypertension, diabetes mellitus, smoking). The size of the ovals is meant to suggest point prevalence, and not, for example, other measures of risk, such as an odds ratio or relative risk ratio. Although we have no hard data for the determination of the size of these ovals, it is thought that the risk factors associated with environmental exposure, trivial trauma, and atherosclerosis are quite common in the general population. On the other hand, the prevalence of genetic predisposition remains to be examined, although presumably present in only a small percentage of the entire population.

View Large Image Download to PowerPoint

Fig 1. Theoretical model on the etiology of CAD. Depicted are the 4 major categories of risk factors for CAD. Dissection is represented by the smallest gray shaded oval (areas 1 and 2) and should be viewed as a subset of those individuals with a genetic predisposition. Area 1 (light gray area) represents dissection occurring in a young individual (younger than 45 years) with a genetic predisposition triggered by environmental exposure(s) in the absence of significant atherosclerosis. Area 2 (darker gray area) represents dissection occurring in the presence of both environmental factors and trivial trauma.

We propose that a genetic predisposition is necessary for CAD to occur, and that the process of dissection is triggered by environmental factors such as infection, estrogen, or other, as yet, unidentified risk factors, with or without trivial trauma. If trivial trauma, in the absence of other triggers, was sufficient to precipitate dissection, one might expect a dose-response curve. That is, increased exposure would lead to increased risk. To our knowledge, only 1 study has examined this relationship (a study of manipulative therapy to the neck), which found that stroke occurred at random during the course of treatment.3 The presence of atherosclerotic risk factors and older age, on the other hand, might have a protective effect because dissection tends to occur in younger patients with fewer of these risk factors, as compared with age and sex-matched cerebral ischemic controls.2

In Figure 1, genetic predisposition is depicted as the smaller circle overlapping the 3 larger ovals, which represent the other 3 risk factors mentioned above. Dissection is designated as the smaller gray shaded area within the circle representing those individuals with a predisposition. The light gray shaded area (area 1) should be viewed as dissection occurring in a young individual (younger than 45 years) with a genetic predisposition triggered by environmental exposure(s) in the absence of significant atherosclerosis, whereas the darker gray shaded area (area 2) represents dissection occurring in the presence of both environmental factors and trivial trauma.

The model also suggests that dissection is highly unlikely in an otherwise healthy individual free of an underlying genetic predisposition, as well as unlikely in the absence of the necessary triggers in someone with a predisposition. This may explain why CAD is so rare and why it typically occurs in young people. It is also possible that multiple short-lived triggers simultaneously present may exponentially increase risk, which might be supported by the following observations. Dissection was found to be more common among those with pathologic skin biopsies who developed an infection than among those with an infection but had no pathologic tissue findings.4 In subanalysis of 4 cases drawn from a larger case-control study, dissection following manipulative therapy to the neck was more likely in those with hyperhomocysteinemia.5 Furthermore, it is highly unlikely that trivial trauma, in the absence of other triggers, or infection, in the absence of an underlying arteriopathy, are sufficient to cause dissection, given the widespread daily occurrence of neck movements and sporting activities, the millions of cervical manipulations given yearly, and the almost universal exposure to infection in the population. Major trauma, due to, for example, a motor vehicle accident, is not depicted in the diagram, although it is thought that such severe trauma may be the only catalyst necessary to cause disruption of the cervical arteries leading to dissection.

Discussion 

There are a number of reasons why an etiologic model of CAD is valuable. Most simply, it provides a visual coalescence of our present understanding of the pathogenesis of CAD. The model also attempts to reflect the multifactorial nature of CAD. Although seemingly a simple well-received message, it is clear from a recent systematic review of case-control studies, which examined risk factors for dissection, that only a minority of studies take this into account.2 In that comprehensive review, only 7 studies were identified that performed multiple regression analyses.4, 6, 7, 8, 9, 10, 11 Because of the methodological problems that were evident in many of those studies, the real risk to the population remains unclear. Even in a recent study,12 which examined the relationship of C-reactive protein to spontaneous CAD, such confounders as mechanical stress associated with infection (ie, coughing, sneezing, and vomiting) were not reported.

Another element of this model is the relationship of atherosclerosis to dissection. At present, common risk factors for atherosclerosis might serve a protective role for dissection, perhaps because of adherence of the arterial layers due to atheroma. This has important ramifications because subjects with ischemic stroke are often used as controls in dissection studies. Therefore, these studies might be subject to selection bias because controls must be a representative sample of the study base and must have an equal chance of developing the target disease as the cases. This relationship, however, must still be confirmed in future studies using a robust sample of cases.

Conclusion 

Although the notion of dissection caused by a trigger in those predisposed to CAD is not a new concept,13 no study has adequately examined the impact of multiple triggers on CAD (or their interaction) in a methodologically sound manner as suggested by this model. High-quality multifactorial studies, including sufficient number of diseased subjects and appropriate (regression) analyses, are necessary to accept or refute this proposed model as a tenable hypothesis.

Practical Applications