Migraines are common headache disorders and risk factors for subsequent strokes. Acupuncture has been widely used in the treatment of migraines; however, few studies have examined whether its use reduces the risk of strokes in migraineurs. This study explored the long-term effects of acupuncture treatment on stroke risk in migraineurs using national real-world data.

A team of Taiwanese researchers collected new migraine patients from the Taiwan National Health Insurance Research Database (NHIRD) from 1 January 2000 to 31 December 2017. Using 1:1 propensity-score matching, they assigned patients to either an acupuncture or non-acupuncture cohort and followed up until the end of 2018. The incidence of stroke in the two cohorts was compared using the Cox proportional hazards regression analysis. Each cohort was composed of 1354 newly diagnosed migraineurs with similar baseline characteristics. Compared with the non-acupuncture cohort, the acupuncture cohort had a significantly reduced risk of stroke (adjusted hazard ratio, 0.4; 95% confidence interval, 0.35–0.46). The Kaplan–Meier model showed a significantly lower cumulative incidence of stroke in migraine patients who received acupuncture during the 19-year follow-up (log-rank test, p < 0.001).

The authors concluded that acupuncture confers protective benefits on migraineurs by reducing the risk of stroke. Our results provide new insights for clinicians and public health experts.

After merely 10 minutes of critical analysis, ‘real-world data’ turn out to be real-bias data, I am afraid.

The first question to ask is, were the groups at all comparable? The answer is, NO; the acupuncture group had

  • more young individuals;
  • fewer laborers;
  • fewer wealthy people;
  • fewer people with coronary heart disease;
  • fewer individuals with chronic kidney disease;
  • fewer people with mental disorders;
  • more individuals taking multiple medications.

And that are just the variables that were known to the researcher! There will be dozens that are unknown but might nevertheless impact on a stroke prognosis.

But let’s not be petty and let’s forget (for a minute) about all these inequalities that render the two groups difficult to compare. The potentially more important flaw in this study lies elsewhere.

Imagine a group of people who receive some extra medical attention – such as acupuncture – over a long period of time, administered by a kind and caring therapist; imagine you were one of them. Don’t you think that it is likely that, compared to other people who do not receive this attention, you might feel encouraged to look better after your health? Consequently, you might do more exercise, eat more healthily, smoke less, etc., etc. As a result of such behavioral changes, you would be less likely to suffer a stroke, never mind the acupuncture.


I am not saying that such studies are totally useless. What often renders them worthless or even dangerous is the fact that the authors are not more self-critical and don’t draw more cautious conclusions. In the present case, already the title of the article says it all:

Acupuncture Is Effective at Reducing the Risk of Stroke in Patients with Migraines: A Real-World, Large-Scale Cohort Study with 19-Years of Follow-Up

My advice to researchers of so-called alternative medicine (SCAM) and journal editors publishing their papers is this: get your act together, learn about the pitfalls of flawed science (most of my books might assist you in this process), and stop misleading the public. Do it sooner rather than later!

5 Responses to Does Acupuncture Reduce the Risk of Stroke in Patients with Migraines?

  • What I find highly suspicious at a first glance already is the extremely high incidence of stroke in the non-acupuncture group: 636 out of 1354, equating to roughly half (~47%) of the cohort. This is almost double the 25% lifetime risk of stroke worldwide, and still quite a bit higher than the elevated risk (~39%) in East Asia – and this even ignoring that the follow-up was 19 years, not a full lifetime.

    When a study of a particular intervention finds a relative risk in the non-intervention arm that is much higher than the known overall baseline risk, then I’d say this should be a big red flag for the researchers that they’re probably doing something wrong.

    • good point!

      • It’s even stranger when you look at the graph: in the first few years, the graph shows an extremely high stroke incidence for the non-intervention group (almost four times the normal rate in the first year) – only to fall off to normal rates after a couple more years (when both curves run parallel). In other words: when you leave out the first 4 years, there is no difference at all between the two groups.

        Whatever the cause, it appears 100% certain that acupuncture had nothing to do with it.

        • I agree it’s unlikely – but few things are 100% certain, I think.

          • Yes, you’re right of course that “100%” isn’t something one should use lightly in science.

            Then again, I think that in this case, the intervention group shows exactly what one would expect – a slow and pretty steady increase in the cumulative number of strokes towards the baseline value – while the anomaly lies fully with the non-intervention group: an inexplicable fast rise in the beginning that soon levels off to the expected behaviour. So any ‘effect’ causing the difference in outcomes appears to happen in the non-intervention arm.
            If acupuncture would have an effect, the non-intervention arm should exhibit a steady but slightly steeper rise compared to the intervention arm, not this weird fast rise followed by levelling-off.

            Anyway, I just remembered where we saw a similar anomaly: Norbert Aust’s critique of Frass’ homeopathy cancer study, where the placebo group showed a median survival way below what was normal for the type of cancer – which might indicate fudging of data.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Subscribe via email

Enter your email address to receive notifications of new blog posts by email.

Recent Comments

Note that comments can be edited for up to five minutes after they are first submitted but you must tick the box: “Save my name, email, and website in this browser for the next time I comment.”

The most recent comments from all posts can be seen here.