Do not trust the cheering, for those persons would shout as much if you or I were going to be hanged.”
~ Oliver Cromwell

In the blogosphere, the proponents of chiropractic often quote the following paper, with the abstract:

Risk of Vertebrobasilar Stroke and Chiropractic Care
Results of a Population-Based Case-Control and Case-Crossover Study
Spine. 2008 Feb 15;33(4 Suppl):S176-83.
by Cassidy JD, Boyle E, Côté P, He Y, Hogg-Johnson S, Silver FL, Bondy SJ.

Why be different? Here is the abstract.

STUDY DESIGN: Population-based, case-control and case-crossover study. OBJECTIVE: To investigate associations between chiropractic visits and vertebrobasilar artery (VBA) stroke and to contrast this with primary care physician (PCP) visits and VBA stroke.
SUMMARY OF BACKGROUND DATA: Chiropractic care is popular for neck pain and headache, but may increase the risk for VBA dissection and stroke. Neck pain and headache are common symptoms of VBA dissection, which commonly precedes VBA stroke.
METHODS: Cases included eligible incident VBA strokes admitted to Ontario hospitals from April 1, 1993 to March 31, 2002. Four controls were age and gender matched to each case. Case and control exposures to chiropractors and PCPs were determined from health billing records in the year before the stroke date. In the case-crossover analysis, cases acted as their own controls.
RESULTS: There were 818 VBA strokes hospitalized in a population of more than 100 million person-years. In those aged <45 years, cases were about three times more likely to see a chiropractor or a PCP before their stroke than controls. Results were similar in the case control and case crossover analyses. There was no increased association between chiropractic visits and VBA stroke in those older than 45 years. Positive associations were found between PCP visits and VBA stroke in all age groups. Practitioner visits billed for headache and neck complaints were highly associated with subsequent VBA stroke.
CONCLUSION: VBA stroke is a very rare event in the population. The increased risks of VBA stroke associated with chiropractic and PCP visits is likely due to patients with headache and neck pain from VBA dissection seeking care before their stroke. We found no evidence of excess risk of VBA stroke associated chiropractic care compared to primary care.

Seems impressive, especially the conclusion. Chiropractic is not to blame for VBA strokes. But abstracts are like movie trailers. They give a flavor of the movie, but often leave out many important plot devices and characters. The Dark Knight evidently has a plot line concerning Two Face, but you would not know it from the trailers (9). Writing an abstract is an art, and you have to choose those feature you wish to emphasize in a limited space. But if you want to know the rest of the story (12), you have to actually pull and read the reference. If you were to read this article in its entirety, you would not be so sanguine about the safety of chiropractic.

As best as I can tell, no one bothers to read the whole reference to see what it says. So I did (6). Dr. Hall has covered the issue of stroke and chiropractic nicely in other posts, but I figure if I am going to go to the trouble of reading this stuff, why should I be the only one to suffer? Consider this an extended footnote to Dr. Hall’s posts, if she will be so kind as to humor me.

Does the meat (7) of the article support the abstract? Lets wander through the content instead of the abstract and see.

First they found 818 vertebral artery strokes on the basis of discharge codes.
First problem: discharge codes are not a reliable way to know the real diagnosis (13).

The discharge diagnosis is sometimes a best guess, often not based on the strictest of criteria, and when people compare the real diagnosis based on a chart review to the discharge coded diagnosis, they often have a poor correlation. And, at least in the US, coding is an art to maximize reimbursement, not a method of determining what the patient really had. You have to do chart reviews to know what the patients ‘real’ diagnosis is, and even then, given the vagaries of testing and disease presentation, you often have to go with best bet.

Also, note the patients have the diagnosis of stroke, not the reason for the stroke. For the elderly, where stroke is common, the most common cause of stroke is thrombus, a blood clot forming in an artery that feeds the brain. The second most common cause is emboli, flipping a blood clot to the brain. In the young, where stroke is rare, the cause is more commonly a tear in the vertebral artery, a dissection, which can be precipitated by trauma.

The worry in chiropractic is a stroke CAUSED BY A DISSECTION. Sorry for yelling. Already the study is flawed as we do not know if there were dissections. Stroke may be a surrogate for dissection, but not a good one, especially in the elderly.

I would bet that the diagnosis of stroke is more accurate in younger people. When a young person presents with a stroke or other neurologic symptoms, they get the million dollar work up as they are not supposed to have a stroke, whereas if an obese, elderly, hypertensive, smoking, diabetic presents with transient dizziness or double visions from medications or viral illness, well, they might be labeled as a vertebral artery stroke without undergoing a precise diagnostic work up. So while the number of VBA strokes is likely to be accurate in the young, in the elderly it is possible to over represent the number of strokes.

Also, that was 818 vertebral strokes in 100 million person years, so a VBA stroke is a rare thing. This is important in that if you have a high background of vertebral artery strokes from thrombus or emboli, which is a common cause in the elderly, you will lose the rare dissection. If you want to find a rare event like a dissection as cause of stroke, the elderly may not be the best place to look, as the noise of the thrombo-embolic stokes will overwhelm the effects of the dissection. In the elderly, looking for a dissection based on stroke codes would be like locating your cell phone by listening in a concert hall by calling it while The Who is playing (11).

They compared these 818 strokes to a control group in a case control study.

Case control studies are good for a rare disease and a common exposure and can only establish an association, which is not, I will emphasize, establishing causation.

They compared those that visited a chiropractor with those that visited a family practitioner, over a month’s time to see if it was associated with a vertebral artery stroke. You know their conclusion.

So does the body of the paper support the conclusions of the abstracts? Well, yes and no.

For people over age 45, there was the same association of vertebral artery stroke after a visit to a chiropractor or a primary care practitioner. A stroke mind you. The cause of the stroke was not determined. It would be reasonable to assume that most strokes in the elderly would be due to thrombi or emboli and that the rare event of a dissection, the worry of a chiropractic manipulation of the neck where you tear the artery, might go undiagnosed.

In the over 45 age group, strokes due to chiropractic could be lost in the sea of strokes due to other reasons, and since we do not know if there were a dissection or blood clot as a cause of strokes, it is hard to conclude that there were no extra strokes from chiropractic. The study was not powered to determine the rare event of a chiropractic event against the background to usual stokes in the elderly. The mean age in this study was 63 (11).

Young people should not have any strokes. In the young, vertebral artery dissection is a common cause of a rare event. It is also the worry from chiropractic neck manipulation. If you could find an effect of chiropractic, it would be in the young. And they do. The people who have an increase in stroke are those under age 45. And it is a big association: odds ratios from 3 to 12.

The association is most noticeable in the first 24 hours after seeing a chiropractor. Usually if you rip an artery it is symptomatic right away. Again, we do not know if these people had dissection or not. We only know they had stroke of some sort, within a day after seeing a chiropractor. One would predict that if there were an association between chiropractic and stroke you would most easily find it in the young and the effect would be most noticeable in the first day or so after the chiropractic visit.

And this article confirms this association.

Note the word association. Association is not causation. You would need a prospective study comparing stoke rates from chiropractic patients vs non chiropractic patients to find causality.

But if you are less than 45 and visit a chiropractor, there is a strong association between that visit and a stroke in the next 24 hours.

Table 3 has the smoking gun. The highest odds ratio (14) for a stroke is 12, 4 times any other, in the age less than 45 in the first 24 hours after a visit to a chiropractor. And, to give credit where credit is due, they mention this as a key point in the discussion and, in the introduction, note two other studies that show an association between visiting a chiropractor and having a stroke. The risk of stroke is 5 to 6 times in these studies.

Part of what they do is smear the data by taking it out for a month after a visit, diluting the effect of the first 24 hours after visit.

How do the authors deal with this data?

They emphasize in the abstract, which is all anyone will read, that there were equal strokes after visits to chiropractic and primary care physicians. They do not bother to mention the increased association in the young until you reach the discussion, which most will not read.

They account for this by suggesting that patients with headache and neck pain, symptoms of an incipient vertebral artery stoke, lead people in equal proportion, to seek care from DC and MD

In other words, they had a stroke in progress at the time they sought care.

The codes used by the chiropractic and primary care physicians for the visit, admittedly inaccurate, were by and large NOT vertebral artery stroke symptoms.

“Neck-related chiropractic visits were identified using diagnostic codes: C01–C06, cervical and cervicothoracic subluxation; C13–C15, multiple site subluxation; C30, cervical sprain/strain; C40, cervical neuritis/ neuralgia; C44, arm neuritis/neuralgia; C50, brachial radiculitis; C51, cervical radiculitis; and C60, headache.

For PCP visits, we included community medicine physicians if they submitted ambulatory fee codes to OHIP. Fee codes for group therapy and signing forms were excluded. Headache or neck pain-related PCP visits were identified using the diagnostic codes: ICD-9307, tension headaches; 346, migraine headaches; 722, intervertebral disc disorders; 780, headache, except tension headache and migraine; 729, fibrositis, myositis and muscular rheumatism; and 847, whiplash, sprain/strain and other traumas associated with neck (These codes include other diagnoses, and we list only those relevant to neck pain or headache)”

Compare these to the symptoms for vertebral artery stroke (from E medicine)

“Ipsilateral facial dysesthesia (pain and numbness) – Most common symptom, Dysarthria or hoarseness (cranial nerves [CN] IX and X), Contralateral loss of pain and temperature sensation in the trunk and limbs, Ipsilateral loss of taste (nucleus and tractus solitarius), Hiccups, Vertigo, Nausea and vomiting, Diplopia or oscillopsia (image movement experienced with head motion), Dysphagia (CN IX and X), Disequilibrium, Unilateral hearing loss, headache”

Beside headache, there is no overlap in symptoms of a vertebral artery stroke and reasons in this study for visiting an MD or DC.

To suggest, based on the codes, that they had a stroke in progress is disingenuous at best (17).

That is a hypothesis they came up with to account for the data. Spin, as it were. That hypothesis may be true, but it is not proven by the data. The study was not designed to demonstrate that result. People who do not read the article carefully take this hypothesis, and they call it a plausible hypothesis in the paper, as the conclusion of the data. This unproved hypothesis is emphasized in the abstract as if it were a conclusion from the data. And those that do not read past the abstract repeat it as if it were a valid conclusion from the data.

To repeat: “The increased risks of VBA stroke associated with chiropractic and PCP visits is likely due to patients with headache and neck pain from VBA dissection seeking care before their stroke” is a hypothesis to account for the data, not the conclusion of the data.

It will take another study to prove or disprove that hypothesis.

My read of the paper is different.

My conclusion, from reading the paper in its entirety, rather than the abstract, is that a population that should not have a stroke, the young, has a marked increase association with stroke 24 hours after visiting a chiropractor and that given the rarity of a vertebral artery dissection as a cause of stroke in the elderly, the elderly is not a group that one could easily find an increase in stroke after chiropractor visit.

To quote the paper:

“We have not ruled out neck manipulation as a potential cause of some VBA strokes. On the other hand, it is unlikely to be a major cause of these rare events. Our results suggest that the association between chiropractic care and VBA stroke found in previous studies is likely explained by presenting symptoms attributable to vertebral artery dissection.”

Note the word major. I agree it is not a major cause of stroke, at least in the elderly. In the young, the highest odds ratio, 12, for a stroke is in the first 24 hours after visiting a chiropractor is twice what has been found in prior studies. As mentioned above, a case control study cannot determine causality, just association.

Given the number of chiropractor visits and the rarity of vertebral artery tear, chiropractic is probably a rare cause of a rare event.

However, given that chiropractic neck manipulation is worthless magical thinking, ANY stroke is one too many.

Less than a movie trailer, this abstract is more like the Superman covers of my youth, where the cover had only glancing resemblance to the story (16).

Go to the website and peruse the chiropractic section. It is striking how many young people had a stroke just after having chiropractic neck manipulation. Of course, these are just anecdotes, and the plural of anecdote is anecdotes not data. But wait. Anecdotes are the prime proof for alt med proponents, since evidence usually proves their pet quackery is a crock. Dr Weil calls anecdotes, “uncontrolled clinical observations.” That’s what whatstheharm has: uncontrolled clinical observations that Weil says should have the same weight as clinical trials (15).

Given three studies that now show an increase in the stroke rates in the young after neck manipulation, I would not let a chiropractor come close to my neck.

In real medicine, it takes less data than this to bring a drug under scrutiny to decide if the benefits are worth the potential risks of a therapy. In my own world of infectious diseases, there have been millions of doses of the antibiotic telithromycin given and only a handful of liver failure and death associated with the antibiotic. I would not prescribe this antibiotic unless I had no other options given the potential risks, albeit very small, and this is for a therapy that actually works.

As a slight tangent with some biologic correlates, it should be pointed out that in a good hanging, the victim should not strangle to death (1).

A good hanging should be set up such that there is a fall just far enough so that the first and second vertebral bodies are separated, breaking the neck and quickly killing the victim. You do not want them to fall too far, as the head may come clean off and that is aesthetically unpleasant. Most people who die these days from hanging do not get a ‘good’ hanging; they suffocate at the end of a rope, a particularly gruesome way to die.

The vertebral artery is often damaged in suicidal hanging (2); “The vertebral artery was shown to be injured quite frequently (rupture, intimal tear, sub-intimal hemorrhage), namely in one quarter of all cases, and indeed in more than half taking into account the perivascular bleeding.” This easy injury is in part due to mechanical reasons “The vertebral arteries appear to be particularly susceptible to injury in trauma of the cervical spine because of their close anatomical relationship to the spine” (3).

A passive hanging (no drop) gives about 686 Newton’s of force around the neck for a 70 kg human. In chiropractic, “the mean force of all manual applications (is) 264 Newton’s and the mean force duration (is) 145 milliseconds (8)”. So a chiropractic neck manipulation, for a short period of time, can provide 38% the force of a hanging. And a bad hanging at that.

Neck injuries are not that frequent because the muscles of the neck prevent injury by preventing sudden, disastrous, movement. If you want to increase the chance of injury from relatively minor trauma, have the person relax. If the muscles are relaxed because the person is not expecting the trauma, the chance of injury goes up. It is why whiplash can occur after minor injuries (4). Chiropractors often have their patients relax just before the coup de grace, I mean manipulation, helping to maximize the chance of injury despite having less force applied to the neck than a noose and gravity.

Given the above, to claim that the VBA occurred before the patient had chiropractic neck manipulation is like saying the hanging victim had a broken neck, but it occurred on the steps up to the scaffold.

References and Snide Asides

“Hanging is too good for a man who makes puns; he should be drawn and quoted.”
Fred Allen

(1) evidently in the 1600’s the English, when the hung (hanged?) their prisoners had them suffocate and did not tie their legs to that the ghastly death dance would serve as a deterrent.

(2) Forensic Sci Int. 1984 Aug;25(4):265-75.
Injury of the vertebral artery in suicidal hanging.

(3) Neurosurgery. 1991 Dec;29(6):912-5.
Subintimal dissection of the vertebral artery in subluxation of the cervical spine

(4) And faking an injury to get money.

(5) I will relegate the issues of the author to a footnote, since I want to focus on the content of the article, but if one is judged by the company they keep.….

(6) Caveat: I am not an epidemiologist nor a statistician. I took and dropped statistics each year I was in college. Not that I was bad at math, my undergrad degree was in physics (this is not bragging, well, yes it is)
So we have, at Good Samaritan Hospital, a bunch of real smart residents, a few of them, such as the resident on my service, are trained in epidemiology and such studies.
Since 3 out of 2 Americans do not understand statistics, and I am one of those people, so I asked one of my residents to help explain the article to me.

(7) Or the soy protein for you vegans.


(9) Written two weeks before the movie released.

(10) The Who. I know they are old, but they did get the record as the worlds loudest band. My son says Motorhead is louder. I don’t know. Let me know what the worlds loudest band is, it you have a moment.

(11) Elderly. Crap. I’m 51 and I’m now calling those in their 50’s elderly. But then I do have my ARUP card, so I can get cranky any time I want for any reason I want. Damn kids and their tattoos/visible underwear/goatee/boom box. In my day, well, my day was the day of the hippie. I suppose I have nothing, really, to complain about. Pop culture of my youth was worse.

(12) Copyright Paul Harvey, all rights reserved. Do not use without express written approval of Major League Baseball

(13) To give the DC their due, they have a complete discussion of the potential drawbacks of the study in the discussion that includes most of the issues I raise here. But, as mentioned, you have to get past the abstract.

(14) The odds ratio compares whether the probability of an event is the same for two groups. An odds ratio of 1 implies that the event is equally likely in both groups. An odds ratio greater than one implies that the event is more likely in a group. An odds ratio less than one implies that the event is less likely in a group.

(15) This is sarcasm.

(16) Superman never married Lois Lane, at least in the 1970’s.

(17) Or talking out your …. Steve wants us to be upscale. Never mind.


7/19/08 Addendum

Since the initial post, thanks to the comments, I have come up with an alternative hypothesis to account for the data. Like the authors, it is neither proven nor disproved by the data, but I think has more plausibility than their explanation.

There is a baseline number of VBA strokes in a population. When patients have serious symptoms, they tend to preferentially seek care with MD’s rather than alternative providers. The group patients who sought care from their primary providers were indeed having incipient VBA strokes. It is why they went to the doctor.

The other group, who were not having an incipient VBA stroke (suggested by the codes for chiropractor visits) had neck manipulation and VBA strokes were induced in this population to a the rate equal to or exceeding the rate of the PMD group.

The study could be interperted that visits to a chiropractor for neck manipulation induce strokes at a rate equal to or greater than the control population ie those presenting with stroke.

Again: a hypothesis. It explains the data as well as the authors, and with more plausibility when viewed in the light of anatomy and physical forces to the neck, other uncontrolled clinical observations :), and prior epidemiologic studies of the association for manipulation and stroke.

Posted by Mark Crislip

Mark Crislip, MD has been a practicing Infectious Disease specialist in Portland, Oregon, since 1990. He is a founder and  the President of the Society for Science-Based Medicine where he blogs under the name sbmsdictator. He has been voted a US News and World Report best US doctor, best ID doctor in Portland Magazine multiple times, has multiple teaching awards and, most importantly,  the ‘Attending Most Likely To Tell It Like It Is’ by the medical residents at his hospital. His growing multi-media empire can be found at