Stephanie Choi, B.Math, Eleanor Boyle, PhD, Pierre Côté, DC, PhD,
J. David Cassidy, DC, PhD, DrMedSc
Toronto Western Research Institute,
University Health Network,
Toronto, Ontario, Canada.
PURPOSE: The current evidence suggests that association between chiropractic care and vertebrobasilar artery (VBA) stroke is not causal. Rather, recent epidemiological studies suggest that it is coincidental and reflects the natural history of the disorder. Because neck pain and headaches are symptoms that commonly precede the onset of a VBA stroke, these patients might seek chiropractic care while their stroke is in evolution. However, very little is known about the characteristics of these patients. In fact, only small clinical case series and physician surveys have described the characteristics of chiropractic patients who later develop a VBA stroke. To date, no population-based study has described this group of patients. Therefore, the objective of our study is to describe the characteristics of Ontario VBA stroke patients who consulted a chiropractor within the year before their stroke.
METHODS: We conducted a population-based case series using administrative health care records of all Ontario residents hospitalized with VBA stroke between April 1, 1993, and March 31, 2002. Three databases were deterministically linked to extract the relevant information. We describe the demographic, health care utilization, and comorbidities of VBA patients.
RESULTS: Ninety-three VBA stroke cases consulted a chiropractor during the year before their stroke. The mean age was 57.6 years (SD, 16.1), and 50% were female. Most cases had consulted a medical doctor during the year before their stroke, and 75.3% of patients had at least one cerebrovascular comorbidity. The 3 most common comorbidities were neck pain and headache (prevalence, 66.7%; 95% confidence interval [CI], 57.0%-76.3%), diseases of the circulatory system (prevalence, 63.4%; 95% CI, 54.8%-74.2%), and diseases of the nervous system and sense organs (prevalence, 47.3%; 95% CI, 38.7%-58.1%).
CONCLUSIONS: Our population-based analysis suggests that VBA stroke patients who consulted a chiropractor the year before their stroke are older than previously documented in clinical case series. We did not find that women were more commonly affected than men. Moreover, we found that most patients had at least one cardio- or cerebrovascular comorbidity. Our analysis suggests that relying on case series or surveys of health care professionals may provide a biased view of who develops a VBA stroke.
From the Full-Text Article:
Our study advances knowledge about the characteristics of the VBA stroke patient population who consulted a chiropractor before their stroke. Overall, we found that our sample was in the fifth and sixth decade of life and that men and women were equally represented. Most cases had consulted a medical doctor during the year before their stroke, and 75.3% of patients had at least one cerebrovascular comorbidity. The 3 most common comorbidities were neck pain and headache, diseases of circulatory system, and diseases of nervous system and sense organs.
Only clinical case series and physician surveys have previously been used to describe the characteristics of VBA stroke patients who had consulted a chiropractor before their stroke. These study designs are particularly liable to selection bias, information bias, and poor precision. These biases may lead to erroneous conclusions about the true characteristics of VBA stroke patients. We find several important differences when we compare our results to those of the studies included in Table 1 . First, previous studies found that VBA stroke patients are likely to be younger than 40 years and to be female. In our sample, the mean age was 57 years; and men and women were equally represented. Second, we report differences with regard to prestroke comorbidities. Previous studies found that VBA stroke patients who saw a chiropractor before their stroke are more likely to be healthy individuals and are less likely to have risk factors for cardio- and cerebrovascular disease. These findings are not supported by our study. As shown in our results, more than 75% of our sample had at least one cerebrovascular risk factor.
Recent evidence suggests that chiropractic care is not associated with an increased risk of developing a VBA stroke. [2, 7] Nevertheless, chiropractors who treat patients with neck pain and headaches need to be aware of the possibility that these patients may be suffering from a stroke in evolution. Although no screening test is currently available to determine who may suffer from a VBA stroke, chiropractors need to be alert to this possibility. Our study suggests that these patients are likely in their sixth decade of life and have a history of cerebrovascular comorbidities. Therefore, we recommend that clinicians inquire about the presence of hypertension, heart disease, diabetes, high cholesterol, cerebrovascular disease, upper respiratory tract infections, addiction, and obesity  Similarly, a history of migraine headache has been shown to be associated with ischemic stroke. [20, 21] Although these comorbidities are not contraindications for cervical spine manipulation, they should alert clinicians that these patients may be at a higher risk of ischemic stroke. In very rare instances, patients who present to chiropractors with “unusually” severe and acute neck pain and/or headache could be experiencing the prodrome of dissection-related VBA stroke. All clinicians should screen these patients for neurological signs and symptoms of stroke and follow them carefully. Until we have better epidemiological studies that address risk factors for VBA stoke, we cannot quantify these risks.
Our study design improves the knowledge in this field for 2 reasons. First, we conducted a population-based study and included all cases recorded over a 9-year period in a comprehensive data set. Second, our use of administrative data limited the impact of diagnostic bias for the ascertainment of prestroke comorbidities. All comorbidities were reported in the database before the stroke diagnosis. Therefore, the measurement of comorbidities was independent of the stroke.
Our study also has limitations. First, misclassification of ICD-9 codes for VBA stroke is possible.  A major limitation of using health administrative data is misclassification bias and the possibility of bias in assignment of VBA-related diagnoses, which has previously been raised in this context. Liu et al  have shown that ICD-9-CM hospital discharge codes for stroke have a low positive predictive value when compared with chart review. Furthermore, not all VBA strokes are secondary to vertebral artery dissection. Our analysis likely includes both dissecting and nondissecting strokes, and our data did not allow us to address what percentage of strokes were sequelae to arterial dissections. Our administrative data include ICD-9-CM codes up to the fourth digit. The identification of vertebral artery dissection would require a 5-digit ICD-9-CM code (ICD-9-CM: 443.24). Therefore, it is likely that our sample includes nondissecting and dissecting strokes. Second, we may have missed ambulatory visits to physicians and emergency departments because patients may have consulted a physician who bills under an alternative payment plan. However, this would be a very small minority of cases. We explored this issue and found that VBA cases without ambulatory physician visits were slightly younger than those who had ambulatory physician records.
Patients who visit a chiropractor before a stroke may or may not be different from patients who visit a medical doctor before a stroke. Although this comparison would be relevant, it could not be done with our data because nearly all cases consulted a medical physician in the year before their stroke. Our analysis suggests that the greatest number of stroke cases occurred in patients with 6 or more primary care visits or 3 or more specialist visits during the year preceding their stroke. This would suggest that their health is not optimal.
A study such as this one is the first step in determining who may be at risk of developing a VBA stroke. We want to emphasize that the information provided in this study is purely descriptive and cannot be used to make prediction or make causal inferences about the role of chiropractic in the etiology of VBA stroke. Our previous research suggests that there is no causal relationship between chiropractic care and VBA stroke. Nevertheless, these 2 events can coincide in time. A prediction tool allowing chiropractors to identify who is at risk of this very rare event would be extremely valuable. However, such a study would be difficult to conduct because it would require a large population of cases assembled from an even larger source population of millions of people.
Our population-based study suggests that VBA stroke patients who consulted a chiropractor during the year before their stroke are older than previously documented. Moreover, we found no sex difference and that three quarters of patients had at least one cardio- or cerebrovascular comorbidity. Our analysis highlights that relying on clinical case series or physician surveys may provide a biased view of who develops a VBA stroke.
Only clinical case-series and physician surveys have previously been used to describe the characteristics of VBA stroke patients who had consulted a chiropractor prior to their stroke.
Using population-based data, we found that VBA stroke patients who consulted a chiropractor during a year prior to their stoke are older than previously documented. Males and females were equally represented among the cases.
Most VBA cases had consulted a medical doctor during the year prior to their stroke and 67.7% of the patients had at least one risk factor for cardio- or cerebrovascular disease.
The three most common comorbidities diagnosed among the VBA cases during the year prior to their stroke were: neck pain and headache (prevalence: 66.7%; 95% CI: 57.0% – 76.3%), diseases of circulatory system (prevalence: 63.4%; 95% CI: 54.8% – 74.2%) and diseases of nervous system and sense organs (prevalence: 47.3%; 95% CI: 38.7% – 58.1%).
Relying on case-series or surveys of health care professionals may provide biased description of who develops a VBA stroke.
Funding Sources and Potential Conflicts of Interest
Dr J. David Cassidy: Canadian Chiropractic Protection Association (CCPA), Government of Nova Scotia, and lecture at CCPA meeting.
Dr Pierre Côté: grants received from Ministry of Health and Long-Term Care, research grants from Canadian Chiropractic Protective Association, expert testimony from Canadian Chiropractic Protective Association, and lecture payment received from (1) Societé des Experts an Evaluation Medico-Légale du Québec, (2) School of Health and Rehabilitation Sciences, University of Pittsburgh, (3) Trillium Health Centre Spine Institute, and (4) Canadian Chiropractic Protective Association.