Here’s a nice post about dietary puzzles in which a group of people who should have a high or low rate of heart disease don’t. For example,
Spanish paradox. Those naughty Spaniards are eating more fat and less carbs and getting LESS heart disease, now there’s a surprise. Good thing their medical system is so marvelous.
Sri Lanka paradox. In Sri Lanka they eat <25% calories from fat and still get lots of heart disease. Tut tut.
I have blogged about the Israeli Paradox. These paradoxes go away, the author notes, “when you realize saturated fat is not the cause of heart disease.”
Elsewhere on his blog he discusses studies that found that eating less fruits and vegetables improved health. Thanks to Dave Lull.
Not unrelated to all this, Ken Carpenter, a co-author of mine and an excellent speaker, is giving a talk next week in the Nutrition Department at UC Berkeley titled “When Nutritionists Go Wrong”. It is on Wednesday, December 19, 2007, at 4:10 pm Room 114 Morgan Hall. Ken has written wonderful books and articles about the history of nutrition and I am eagerly awaiting this talk.
Do we really know the degree to which nutrition is important in health, that is, assuming we get basic amounts of protein, carbs and fat? Maybe the fundamental assumptions about the relation between nutrition and health outcomes are just plain wrong. And no matter how many times correlations are repeatedly found in epidemiological studies, it can never show causality. There are so many correlational studies controlling for confounds that are themselves based on correlations. It’s all a big guessing game and it’s obvious now that nutritionists have been guessing wrong.
The lesson is: when it’s nearly impossible to do controlled, randomized studies, be humble and admit that you are just guessing…
I think Tim understates the utility of correlations. They are relatively cheap and can be guides about whether theories are reasonable and then what to do a randomized study on. Correlation suggests causation. Then we have to go further. We also have to be careful not to base policy on correlations, especially when we have good reason to believe confounding factors will be important.
The first evidence that folate prevents birth defects came from correlational studies. See for example
https://www.kofinasperinatal.org/download.cfm?DownloadFile=244A9AD1-BB1E-A39E-81BE645F04C090B2
Later this idea was tested with experiments. It turned out to be correct.
Seth: So, what *does* cause heart disease? Seriously, I’d like your opinion.
The current state of knowledge would have it that it’s multifactorial, for instance related to diet, exercise, smoking, hypertension, etc. Obviously lots of people get heart disease in this country, but they don’t all have the same risk factors. How does one explain, e.g., the runner Jim Fixx, who died of a heart attack and who seemingly had only one risk factor, high cholesterol? Then there’s data like that in which vegetarian Adventists outlive non-vegetarian ones.
If the ideas that you’ve been writing about here turn out to be even half-correct, well, I just don’t get it.
They’re only paradoxes if you believe Ancel Keys and Senator McGovern were right. It’s getting pretty clear they were disastrously wrong.
Distance running IS a risk factor. It’s a bizarre behavior with no evolutionary raison d’etre, and it’s incredibly inflammatory. We didn’t evolve to do anything of the kind. We evolved to be largely stationary, with short bursts of extreme exertion — sprinting. And in fact, sprinting exercise is anti-inflammatory and anabolic. (The idea that the Marathon is the pinnacle of human physical health is kind of insane, even before you acknowledge that it’s the commemoration of a runner’s death.)
We also evolved to eat — also in short bursts — large amounts of (can’t help myself, dude) protein and fat. Carbs just weren’t around in non-trivial quantities. And even then, we evolved to go quite a while between feeds. And indeed, intermittent fasting is anti-inflammatory, compared to constant noshing. And, also, anabolic. In other words, you get big and strong if you eat one meal a day of protein and fat. Six small meals, less so. Add carbs, even less. Why? We’re probably tuned for our ancient environment.
David: Most correlations only suggest hypotheses to be tested; if they are enormously strong they may begin to suggest causation. I agree that when something jumps out at us in a correlation, it should be tested to see if there is causality. The problem is that it’s enormously difficult, expensive, and ethically limited to do controlled nutritional experiments. It’s not clear to me that replication of correlational studies should be taken very seriously, because you may just be replicating correlations that have nothing to do with cause.
The controlled nutritional studies like the Women’s Health Initiative seem to suggest that diet matters less to health than was previously believed. I don’t know how we go about sorting out all of the correlational data and controlled studies to try to determine if nutrition, within a basic range of getting reasonable protein, carbs and fats is really that important.
The Japanese remain the world’s longevity champions. CHD mortality rates have declined even as meat consumption has doubled. When I first lived in Japan 25 years ago, the beef bowl and katsudon had already become staples of the Japanese fast food diet. (As Homer Simpson would say, “Mmmm . . . katsudon.”)
Most paradoxical is the persistently high smoking rate. And it would seem that the stressful Japanese lifestyle would only make things worse. But this is where I think alcohol enters the equation in an interesting way. It’s still de rigueur for Japanese businessmen to hit the bars after work on a daily basis. Getting soused and blowing off steam is culturally acceptable, even smiled upon.
Add to this a cohesive, community/religion-centered culture (though most Japanese claim not to be “religious”). This is a “cradle-to-grave” society that manages to avoid being actually socialistic. Free-market capitalist that I am, there’s a lot to the argument that the stress alleviation provided by a paternalistic society and “good enough” socialized medicine may well offset the deficiencies.
Another beneficial aspect of the Japanese lifestyle is the huge amount of walking and biking they do. Not for “exercise,” but merely to get around. Seriously, the only people I ever saw jogging in Japan on a regular basis were other Americans. (Plus, this is a culture that puts a high premium on hygiene.)
Tim: I agree with you. I think the main problem with studying nutrition is that it is so multidimensional. How do you begin to do a reasonable mapping of the eating behaviors of a population onto a set of outcomes that you care about? If the goal is to set some recommended nutrition policy this is very hard. Even if you could do studies that plausibly estimate causality for some outcome over some range of the nutrition variable, you would still need to worry about its interactions with everything else, and with the multiplicity of outcomes that people are concerned about. Its even worse that the outcomes we are really interested in are long run mortality and morbidity.
Dennis, I don’t know what causes heart disease. However, if I had to guess, I think the evidence is strongest for:
1. Stress. The social gradient is large and everywhere and seems to be due to lass control of one’s job at lower positions. Kendrick’s examples of displaced peoples provides a different kind of evidence for this factor. Perhaps stress acts by disturbing sleep, which causes one’s immune system to malfunction.
2. Inflammation. The strong correlation between gum disease and heart disease. I think gum disease is due to too much inflammation.
Tom, this is very interesting:
> sprinting exercise is anti-inflammatory and anabolic.
> you get big and strong if you eat one meal a day of protein and fat.
Can you please pont me to some further reading?.
Thanks
I’ve heard a theory of the evolution of human beings in the ecological niche of East Africa that suggests our species spent a lot of time following herds of large migrating animals and scavenging those that died or were killed. This is said to explain bipedalism since that configuration of the body is the most efficient at shedding heat. In this setting it made sense to eat the most energy dense portions of the animal first–brains and so forth–that are mainly fat. I haven’t read the science, just a lay account, but I found it very suggestive.