THANK YOU FOR VISITING SWEATSCIENCE.COM!
As of September 2017, new Sweat Science columns are being published at www.outsideonline.com/sweatscience. Check out my bestselling new book on the science of endurance, ENDURE: Mind, Body, and the Curiously Elastic Limits of Human Performance, published in February 2018 with a foreword by Malcolm Gladwell.
- Alex Hutchinson (@sweatscience)
***
There’s an article by “paleonutrition” advocate Kurt Harris, posted on the Psychology Today website, that is currently making the rounds, in which he argues that aerobic exercise causes heart disease. The study he’s talking about was published in Radiology back in 2009 by a German group led by Frank Breukmann and Stefan Möhlenkamp. They compared 102 marathon runners between the ages of 50 and 72 with 102 age-matched controls, and used cardiac MRI with late gadolinium enhancement to look for evidence of permanent heart damage. Five of the runners had evidence of past heart attack, and another seven had heart damage not associated with a heart attack, for a total prevalence of 12%. Among the controls, two had evidence of past heart attack, and another two had other damage, for a total prevalence of 4%.
The difference between the two groups wasn’t statistically significant. This doesn’t matter to Harris:
I am obliged to point out that by the conventional arbitrary criteria used in biomedical publishing, the difference was “not statistically significant”… Stop doing what you are told and read the statistics without letting the authors or editors tell you what is “significant”.
That’s certainly a convenient approach: if you already think you know what the right answer is, then you don’t need to worry about little details like statistical significance.
Still, the results are very much worth considering. Harris wonders why the study was ignored by the “nutrition and fitness” blogosphere, and takes a shot at the New York Times:
[I]t’s published in Radiology, which is not exactly Gina Kolata territory.
Actually, the Times ran a Bloomberg article on the early results from this study cohort way back in 2006. And Runner’s World interviewed cardiologist Paul Thompson, a co-author of the most recent results from the study, in January. Thompson points out some important caveats about the study group:
[H]is marathon group includes a number of former smokers and others who might have been quite unhealthy before they began running…
The key issue with Möhlenkamp’s runners is that their cardiac risk scores are compared using their present cholesterol, blood pressure, and other health numbers. They might have had terrible numbers before they started running, so when their coronary calcium is compared with folks who are not athletes, but had good risk numbers all their lives, it looks like the runners had more calcium, ie, more atherosclerosis than predicted by their risk factors… Many of the runners “got religion” when they turned 40 or so.
To his credit, Harris acknowledges that 50% of the runners in the study had a history of smoking, compared to just 42% of the controls — but immediately dismisses the possibility that this (and any other underlying factors that differ between the two groups) could play a role. Again, why worry about statistics and the possibility of confounders when you already know the right answer?
Let me perfectly clear: it’s entirely possible that repeated marathon running (the runners in the study had completed an average of 20 marathons) produces damage to the heart. By all means we should continue to study this, and find out. Despite Harris’s snide comments about aerobic exercise not producing immortality, we’ve known that aerobic exercise doesn’t grant immunity from heart disease for decades. In fact, very first time noted iconoclast Tim Noakes made a splash in scientific circles was way back in 1976, with a paper about heart attacks in veteran marathon runners.
Of course, there’s a difference between running marathons and “aerobic exercise” as practiced (or not) by the vast majority of the population. I’m a lifelong runner, but if someone asked me: “What’s the optimal exercise routine for cardiac health?” I wouldn’t necessarily recommend running marathons. I’d probably suggest something like an hour a day, with two to three days a week devoted to shorter, faster interval workouts. And of course, I would also recommend doing some resistance training — and I wouldn’t take, for example, a study showing that competitors in the World’s Strongest Man competition die early as evidence that “resistance training is bad for you.”
That’s basically what Harris is doing here: concluding, on the basis of this one study of a fairly extreme group of outliers, that aerobic exercise in general is as bad for you as boxing or football. The only sensible exercise, he argues, is resistance training:
I still find no grounds at all to believe that high levels of “cardio” protects your heart or makes you live longer. Certainly not “the more the better” which is what we’ve been led to think since the 1970s running craze.
The reason he finds no grounds, of course, is that he’s never looked. In his article, he waxes philosophic about the fact that “the picture or the test result is not the patient or even the disease.” Then, on the basis of a test result, he concludes that running will kill you. But what if we look at “the thing itself,” and find out whether runners die less from heart attacks — and whether “the more the better” holds up.
Paul Williams did this study and published it back in 2009. He followed 35,402 runners for 7.7 years, during which time 467 men reported heart attacks, angina or had to have bypass surgery, while another 54 died of heart disease. The risks of all these symptoms, including death, decline with every additional kilometre run. Those running more than 9 km/day “produced risks 65% lower for angina, 29% lower for nonfatal CHD, and 26% lower for fatal and nonfatal CHD” compared to those running fewer than 3 km/day (the level corresponding to national physical activity guidelines for adults).
I’ve rambled on a bit here, so let me finish by repeating one simple point. Even if you take the results of Möhlenkamp’s study at face value (which I don’t, as explained above), the conclusion you can draw is that running a large number of marathons may damage your heart. To go from this to arguing that 20 minutes on the elliptical is bad for you requires a Beamon-esque leap of logic, along with a cheerful disregard of literally hundreds of epidemiological studies.
A minor point, but it looks like Harris uses a ‘prosecutor’s fallacy’ in dismissing the 5% significance level.
“To meet the standard definition, there would have to be a 5% chance the difference is due to chance. Instead, the significance level was 8% by McNemar’s test, so there is only a 5% [I assume 8% is meant] probability that the difference is due to chance.”
The chance of finding a difference, assuming there is no connection between running and heart disease, is not equal to the chance that the difference is due to chance. That depends on the prior odds. This would be a mere footnote if there were only two competing claims here (that, without any prior knowledge, can be assumed to have an equal chance of being true). But in fact there are three claims here, namely
1 marathoning is healthy
2 marathoning has no effect
3 marathoning is unhealthy
Harris rejects 1 and 2 and says that there is an 8% chance that 1 or 2 is true, so there is a 92% chance that 3 is true, while the statistic tells us that (assuming 50-50 prior odds) the there is an 8% chance that that 2 is true and thus the chance that either 1 or 3 is true is 92%.
So, Harris conflates the theory that either 1 or 3 is true with the theory that 3 is true.
Excellent point, RH — thanks very much for pointing it out!
@RH
Seriously – my head got dizzy for a second there trying to process the logic. Nice interpretation.
@David Csonka
Mine was too writing it down :-). As a non native speaker of English, my way of putting things may be a bit cumbersome. In this particiular case I plead in my defense that statistical significance simply is a complicated subject. It is to me anyway.
The Psychology Today piece is a reposting of one of Harris’ blog entries. That “cardio is unhealthy” is an obsession of his, and he has several posts on the subject. Here is another example: http://www.paleonu.com/panu-weblog/2010/3/21/still-not-born-to-run.html.
“That’s certainly a convenient approach: if you already think you know what the right answer is, then you don’t need to worry about little details like statistical significance.”
That’s a convenient way to dismiss my analysis. Unfortunately you are completely wrong. I run 15 k a week and have for over ten years. I never considered the idea that excessive running could actually cause myocardial fibrosis until I read this study. Believe what you want, I call them how I see them. My opinion is as a physician not a runner or exercise advocate or critic.
“Actually, the Times ran a Bloomberg article on the early results from this study cohort way back in 2006. ”
I’m surprised you did not notice that it was the same cohort but a completely different paper that did not use cardiac MRI. NYT did not report on the Radiology study.
As far as the statistical signifcance, the difference between a 5 and 8% significance level is arbitrary. That is a fact and it is also a fact that runners had three times the fibrosis. Even at a level of 10%, we can’t exclude the possibiltiy that the effect is real. I have a more recent study that compares elite marathoners to controls with no smokers among the runners. There the difference was 50% fibrosis in the elite marathoners and 0% in the controls. Easily statistically significant. Explain that one. Perhaps the runners just thought about smoking or were exposed to second hand smoke? If you want to think running a lot is magic, you can come up with something, I am sure. I have a hard time seeing how 20% more former (usually years ago) smokers results in 200% more fibrosis. Maybe you can explain such a non-linearity. At the very least, the running is certainly not reversing any heart disease, is it?
“arguing that 20 minutes on the elliptical is bad for you”
I didn’t say its bad for you. I implied it is a waste of time. I was quite explicit that high levels of high intensity cardio is what I consider dangerous. And I’ve advocated reasonable levels of runnning as having benefits as well as saying that I run myself. You need to read more carefully and with less of your non-defensive bias perhaps. I’m also not disregarding anything. A self selected sample of runners where running more miles correlates with morbidity is an observational study with all the usual defects that implies. Healthy people run and people that are health conscious may run more if they are more health conscious. Duh!
“Misinterpreting” indeed.
@Koop
3 posts out of over 100 is an “obsession”? Um, OK.
I should add that you have misquoted me. I said:
“We should keep calling marathons, centuries on the bicycle and hours on those ridiculous stairmasters and treadmills “cardio” to remind us which organ we may be putting at risk.”
Hours on the stairmaster is not 20 minutes. I did not say 20 minutes anywhere. If someone is snowbound and needs a little exercise, 20 minutes on a stairmaster is better than nothing. Doing hours of it is a waste of time. Strive for accuracy. It’s not your only misquoute.
I also did not say that ANY amount of aerobic exercise causes heart disease. Only excesssive amounts. Another misquote on your part.
@Other readers
Please read the original PT article so can see accurately what it says.
@RH
Your analysis is confusing, mostly because it is completely incorrect.
Thanks for taking the time to comment, Kurt. There’s not a lot in your responses that I can argue with, mainly because you just repeated what you said in your original article, except more loudly. Plus you added a nice appeal to authority by emphasizing that this is your opinion “as a physician.”
You repeat several times that I’ve “misquoted” you, without giving any specific examples. Are you now suggesting that, contrary to your title, cardio exercise DOESN’T cause heart disease? You did, after all, write:
“I think that not only does sustained “cardio” not protect you from atherosclerosis, I think it is quite likely that through repetitive shear stress with endothelial damage and promotion of an inflammatory state, that it may promote atherosclerosis and/or direct cardiac muscle damage.”
Perhaps you could clarify: are you now saying that “sustained” actually means “excessive,” and “excessive” means running multiple marathons? If so, then I (and I suspect everyone else who read your article) misunderstood.
You also wrote, re. the NYT article: “I’m surprised you did not notice that it was the same cohort but a completely different paper that did not use cardiac MRI.”
Believe it or not, the reason I wrote “this study cohort” was because I meant “this study cohort,” not “the same study.” My point was that people have been studying, talking about, and writing about the relationship between prolonged running and cardiovascular health (including the ongoing study of this precise group of runners) for years, even though you seem to think there’s a conspiracy not to talk about it. The Times reported on it. Runner’s World wrote about it. We’re all interested in these results, notwithstanding your claim that “you may have to be an academically-oriented cardiologist or radiologist to really understand the significance of the findings.”
As for your rehash of the statistics, it’s once again a question of interpretation. Of course p=0.05 is an arbitrary threshold. The reason we’ve come up with rules of thumb like that is to guard against the very human tendency to treat study results as a Rorschach blot in which we see the outcomes we already believe. (i.e. “They” want you to believe that these results aren’t statistically significant, but it’s CLEAR to anyone with a brain what’s going on…)
Lastly, I quite like how you dismissed the results of the Williams study: “Healthy people run and people that are health conscious may run more if they are more health conscious. Duh!”
Meanwhile, in your Psych Today article, you write: “Could “cardio” promote atherosclerosis and myocardial damage by being confounded by diet? That is, could the wheat, excess sugar and linoleic acid found in low fat “healthy” diets be more prevalent in marathon runners by virtue of their greater caloric intake of this noxious garbage?”
So the kinds of “healthy” behaviours that are prevalent in runners are “noxious” in one context. But when a study shows that running more is associated with living longer, it’s because running is associated with other “healthy” behaviours? As Gary Taubes himself argued, when you have to keep patching up a theory to explain results that don’t agree with it, it’s probably time to take a hard look at the theory.
@ Kurt Harris
I am happy that you were able to understand my argument, despite my confusing formulation, since you were able to spot that my analysis is ‘completely incorrect’.
I also noticed that you changed the passage I attacked, and substituted it with a correct analysis in terms of rejecting the H0 hypothesis. Perhaps my analysis was not completely incorrect. Otherwise I appreciate your preparedness to learn from someone who is completely wrong.
As to the second part of my analysis, perhaps I am wrong. In that case I’d be interested to hear what is wrong with it.
As to you general message: I do share your suspicion that, above a certain milage, adding more miles will make you healthier. I simply do not think that there is enough evidence to say anything definitive about that.
Sorry: As to you general message: I do share your suspicion that, above a certain milage, adding more miles will NOT make you healthier. I simply do not think that there is enough evidence to say anything definitive about that.
@RH: “I do share your suspicion that, above a certain milage, adding more miles will NOT make you healthier.”
And I’d add that I agree with this point — in fact, I’d be hard-pressed to think of ANY activity, habit, food, etc., no matter how wonderful, that doesn’t become counterproductive (at least within the narrow definition of “optimal health”) beyond a certain threshold.
The “J-curve” hypothesis is generally used to discuss the interaction between exercise and immune system function: a little bit of exercise strengthens immunity, more exercise makes it even better, but too much hurts it. While it remains a topic of active research, several studies suggest that immunity starts to be compromised (on average) at about the level of moderately serious marathon training — say 100K a week.
I think this sort of curve is likely pretty generic: certainly, the latest heart damage studies hint that the effects of running on heart health may follow a similar curve. But for 99.99% of people in North America and Europe, they’re on the part of slope where more exercise improves immunity and heart function.
As for the people who are on the far side of the slope, training at a very high level… Well, let me put it this way: none of the marathoners I know run marathoners in order to be healthy. They may run to be healthy, but they run marathons to conquer a personal challenge. Conflating those two motivations is like arguing that going for long hikes is dangerous because people sometimes fall off Mount Everest and die.
Yes, I cautiously stated the blinding obvious.
All of this hand-wringing over a single study mangled by dubious analysis? Puh-leeze.
What’s astounding is that a supposedly trained MD would bang his drum so loudly over a single study in a field where, as the author here notes, hundreds of other applicable studies exist and are readily available for comparison.
If I didn’t find it amusing, I probably wouldn’t bother to think much about it at all.
Except that…there are a lot of people who do extreme exercise & get cancer, heart disease, neurological diseases etc…
When we start to research the countries that have the longest history of healthy centenarians, these people don’t have diabetes, cancer, heart disease, neurological diseases, etc…at rates even close to ours & they’re living longer.
They don’t do extreme forms of exercise…they are active throughout their days…participating in community, walking, biking, tai chi with breathing, gentle yoga with breathing, cooking…every day. They don’t participate in extreme forms of exercise.
Isn’t it possible that just like we’ve been wrong about so many other things, we could possibly be going too far? Perhaps our addiction to extreme forms of exercise comes from our extreme stress levels?
After all, the sensations that come from cardio are addictive sensations just like an alcoholic or drug addict experiences…yet our culture likes to make it okay because it fits into our work hard, play hard ideas. Meditation which is shown to balance our attachments & aversions works with this idea of us being addicted to sensations where some are good & some are bad, yet the healthiest people are those that create balance & do not strive to feel any sensation, pleasant or unpleasant, for a prolonged period of time.
Our ideas that we’ve been taught since we were little may not be what’s healthiest for us…let’s open our minds.
@Stef:
“Isn’t it possible that just like we’ve been wrong about so many other things, we could possibly be going too far?”
Sure it’s possible. Lots of things are possible — so to distinguish between what’s possible and what’s actually true, we have to rely on evidence. Is there any evidence that people who do lots of cardio exercise die of heart disease?
@Stef,
I will agree that extreme exercise is probably counter-productive. However, you point on that other countries don’t have the rates of degenerative diseases that we do, but fail to realize that may be due to other factors besides lack of extreme exercise. Don’t forget that other countries tend to eat less processed foods, get more omega 3’s from fish and other sources and probably have less stress. Not only that, but you seem to allude to the notion that most Americans are exercising excessively. Given our obesity rates and sedentary lifestyles, I certainly do not believe that is true and don’t need a study to reasonably come to that conclusion.