Sunday, December 7, 2014

Bad Science

Hot Yoga? Really?

I started reading the Breaking Muscle blog and it's been a mixed bag. But then I saw a post on "3 Ways Hot Yoga May Benefit You". I have a strong aversion to "alternative" practices. By definition alternative means that it's not supported by the science. But this post is about some published scientific studies on the potential benefits of hot yoga.

I have no problem with yoga - when it's benefits are accurately described. Yoga can be a perfectly good way to meet the flexibility aspects of a complete physical activity program.

I just get all uppity when people say it's great for everything. (And I get that way when anything is called great for everything.)

Some of you may be wondering why I'm going to go on this rant about one study of hot yoga, when this is supposed to be a HEMA training blog. The point is to demonstrate the skills necessary to be a good consumer of science. To learn appropriate caution when encountering science-y sounding fitness claims on the internet.

So, is Hot Yoga good for Fitness?

Well, this blog post went over one study that looked at strength and cardiovascular measures and another for cardiometabolic risk factors. And both studies are flawed. And the results of both studies are overstated. This is a common problem with reporting on science - the results are overstated, and generalized beyond the scope of the study. Furthermore making strong claims about anything based on one study is bad science.

The basic prescriptions for how to improve strength and cardiovascular ability are based on hundreds of studies. So before it can be claimed that yoga improves other fitness components we should have a good evidence base. Not. Just. One. Study.

Bikram Yoga and Physical Fitness in Healthy Young Adults

This study appeared in the Journal of Strength and Conditioning Research. It's not terrible. But it's results are weak and then exaggerated. The study looked at just 21 people. Well, it started with 32 but 11 dropped out of the hot yoga group. That's a big red flag right there! 50% dropout from only hte experimental group severely effects the outcome of a study. The subjects were divided into two groups: hot yoga and control. The hot yoga group was allocated twice as many people, because they expected a high number of dropouts. So, they knew in advance that there would be a problem and went ahead anyways. Not encouraging.

Why did people dropout? We are told that some had scheduling issues and others were "dissatisfied". Quite a number of studies are published in this field without such high dropout rates from scheduling issues. So I think it's much more likely that half of their experimental group just couldn't take the hot yoga program - which is done with the thermostat set extra high and plenty of humidity to boot.

If only the strong survive your study then that will skew the results towards a positive outcome. We don't end up measuring the people who are struggling.

And since they knew in advance that the program was so hellish that half of the subjects (who are volunteers getting paid to do yoga) would dropout you have to wonder whether or not any benefit is going to be worth it.

The control group was doing no exercise and the subjects were all sedentary to start with. In such a group of people any exercise will show improvement. So comparing four and a half hours of exercise a week to no exercise isn't a test of hot yoga versus other methods. It's a test of whether or not it's better than literally nothing.

So what were the results? Unimpressive. Two strength measures were used: grip strength and isometric deadlift. The grip strength showed no improvement, but there aren't any gripping activities in Bikram so that's not surprising. The other test showed some modest improvement*. But it's an isometric test following an isometric program. It's not going to translate well into real world performance and health measures. Since in the real world you don't just pull on a chain attached to the floor, you move.

* Or it didn't because they used shady statistics. I'm not certain about this. But they used a statistical measure that is not common in this field. And they also reported the common measure, and that one came up negative. However, this other test (group x time point comparison) did show statistical significance. Anytime I see a brand new test I get curious - I asked my stats professor about this. So it seems like the test is legit for this purpose. But did the researchers pick this because the regular test didn't support their hypothesis? I don't know.

Flexibility: They measured two flexibility components, low back/hamstring, with a sit-and-reach; and shoulder flexibility. Unsurprisingly, the yoga training produced substantial increases in the sit-and-reach test. But that measure of flexibility is not well correlated with either performance measures or health/fitness measures (like low back pain). So, who cares if that improved?

They also measured active shoulder range of motion. The test involves laying face down and holding a stick overhead. Then you lift the stick as high off the mat as you can. This measure improved as well. But not by a huge amount. Given some of the postures in Bikram this result is not surprising either.

Body Composition: The changes here were slight to none. And we get to see the researchers use the phrase, "trended towards significance". Ha! Individual p-values do not have a trend. When most of the p-values you get in a category are not significant, then you definitely don't have a trend. This phrase is researcher bunk for, "um, we didn't get a positive result, but, we . . uh . . . almost got one."

So, 4.5 hours of light exercise a week barely beat sitting around for improvements in body composition (weight, waist-to-hip ratio, body fat % and lead body mass) - not exactly surprising. Nor noteworthy. Only one actually significant result out of four tested also smells a bit like a fishing expedition.

Cardiorespiratory: I don't know why the researchers thought that standing still in various poses for 90 minutes would have improved cardiovascular health. Maybe because it was so hot that the heart had to work extra hard to keep the body cooled? Suffice it to say there was no improvement in blood pressure or aerobic capacity.

Why is it so Damned HOT in Here?

Seriously? No attempt is made by the authors to explain why they thought that yoga in a super hot, humid room would be better than regular, comfortable room yoga. And none of the things they measured seemed to be related to the heat. None of the minor benefits that were seen can be attributed to the heat since they were all similar to what you would expect from a regular yoga program.

Why do I care, though? Because hot yoga is an unnecessary risk for heat injury. It's hot enough in that room to mess you up. There's a reason that half people in the program decided not to finish it.

A nice description of the risks is given here, and from the same website to boot.

And, no, hot yoga does not sweat toxins out of the body! That's just not a thing. Human physiology doesn't work that way. Sweat glands only selectively excrete specific substances - they don't sieve your bloodstream and leak out whatever they find.

In summary:
Higher risk than regular yoga or exercise.
Minimal benefits that can just as easily be achieved cheaper and comfier.

Conclusion: not worth it.

But why did I rip so hard into this one study? Because proper science means that the more outlandish the claim the better the evidence required. Saying that a single exercise mode will help everything while in an excessively hot room is an extraordinary claim. So it requires extraordinary evidence. This study fails to meet that bar.

The bibliography:
Tracy, B. L., & Hart, C. E. F. (2013). Bikram Yoga Training and Physical Fitness in Healthy Young Adults: Journal of Strength and Conditioning Research, 27(3), 822–830. doi:10.1519/JSC.0b013e31825c340f

Yoga for Cardiovascular Risk

The same blog entry also talks about a study looking at cardiometabolic risk factors. That is risk for heart attacks, strokes and diabetes.

You've got to be kidding me. Light intensity exercise like this has no plausible mechanism for improving any of the things they measured.

And this can't be taken seriously. We can see that with information contained within the citation alone. It's from the Journal of  Alternative and Complementary Medicine, which makes the Quackwatch list of Non-recommended Periodicals. This journal has an impact factor of 1.5 which means that most scientists won't cite stuff from it. Neither of these pieces of information is perfect in determining quality, but they show a trend.

The actual study continues the trend.
  • No control group. Seriously, this journal published a study without a control group. Why?
  • A large number of disparate measures are used. I feel like they just used everything they had in their lab.
  • Paradoxical results. One group improved in one measure the other group did not. But the other group improved in something else, unrelated, which the first group also did not improve.
This is not a good study. It should be ignored. Instead of cherry-picking weak studies that show what you want if you squint, the blogger should have looked for good research and clear patterns in the literature.

Yoga for Mental Health

There was one other study in the blog post that I'm ripping on. It was on mental health improvements. However, that exercise and a meditative practice improved some mental measures is not noteworthy. Of course it did. And that's a good thing. If you want to do yoga for the meditative aspect then more power to you.


We have three weak studies of hot yoga. None of them even attempts to address the core issue: does hot yoga have any benefit over regular yoga? Read any text on exercise, personal training, sports medicine etc. and you will have a section on how high heat and humidity is dangerous. How training in those conditions should be avoided, and when it can't be avoided you have to take additional precautions to keep people safe. 90 minutes straight of activity in that heat would not be considered safe in any other context.

Only in alternative practices would this even be considered. Take something known to be dangerous and then declare it superior. Then don't bother testing it. Remember, none of these studies compared regular temperature yoga to hot yoga.

This is the kind of research that's considered good enough for alternative medicine. And it's why these kinds of claims must be scrutinized more closely.



  1. My favorite response was an experienced yogi emailing me this, " I can't see how [listening to your body] is possible if you're too focused on falling in your own sweat ad possibly dropping from dehydration and heat exhaustion."

  2. What would you say about Tai Chi, an alternative practice?

    1. Tai Chi is not alternative per se. It depends on the claims being made.

      Like any of a wide variety of practices, Tai Chi can be used for meditative purposes which have well demonstrated benefits to mental health and some minor physical health components (like blood pressure).

      Tai Chi can also be used as a form of balance training, in particular in older populations that are a fall risk. It's certainly better than inactivity, but it's poor specificity means that there are other more effective modes to achieve this objective.

      Tai Chi can be used as part of a complete approach to physical activity, specifically for the neuromotor component and, to a lesser extent, flexibility and range of motion. It will not however create significant cardiovascular or strength benefits.

      I'm not aware of any other claims about Tai Chi that are supported by the literature.