[swombat]

Saturated fat has been demonised since the 1970s when a landmark study concluded that there was a correlation between incidence of coronary heart disease and total cholesterol, which then correlated with the percentage of calories provided by saturated fat, explains Malhotra. “But correlation is not causation,” he says. Nevertheless, we were advised to “reduce fat intake to 30% of total energy and a fall in saturated fat intake to 10%.”

It is frightening to think that such a fundamental piece of public policy is based on the "correlation is not causation" fallacious thinking.

Is there no one in power who is in the habit of actually thinking before approving actions?

[achileas]

Most things in the biological sciences are correlations, to be as precise as we are in the sciences, you have to point out that some relationship is correlational unless you specifically test for causation (which is unfeasible in most cases)...if you assume causation, even when there seems to be very little doubt as to the causative agent, your paper is going to get sent back to you by the reviewers (in most cases).

Add to that the fact that recommendations are based on erring on the side of caution, and that tight correlations do become predictors of an outcome, then you have a recipe for these recommendations. Nothing bad is going to happen by reducing saturated fat intake, for the most part, but there is a chance of it if intake is excessive, and in fact saturated fat intake is a fairly strong predictor of future heart incidents.

[wwweston]

> Nothing bad is going to happen by reducing saturated fat intake

From what I've read, it depends on what you eat instead. For example, as the article says:

"a JAMA study revealed that a 'low fat' diet showed the greatest decrease in energy expenditure, an unhealthy lipid pattern, and increased insulin resistance (a precursor to diabetes) compared with a low carbohydrate and low glycaemic index (GI) diet."

[Volpe]

Insulin resistance is caused by sugar intake (completely unrelated to cholesterol)... Arguing that in order to lower cholesterol you have to consume more sugar seems rather strange.

[tolmasky]

Although individuals did not consciously decide to replace fat with sugar in their diets, the reality is that this policy advisement lead food producers to reduce fat in their products and replace that fat with sugar -- so the practical result was indeed this. Most products that advertise low-fat have accomplished this with sugar. Sure, it didn't have to happen this way, but thats irrelevant because it did happen this way: their was public policy that labeled fat as the bad guy, fat was reduced (sugar increased), and these actions were applauded.

[Volpe]

But I've seen no evidence that, that is what happened. There doesn't seem to be a logical reason as to why you'd need to increase sugar in order to lower cholesterol, so why would companies do that?

Can you point to an example food(s) that had high fat, and low sugar, and then after this policy, low fat, and high sugar... ?

[malkia]

Maybe to make the food still tasty. Go to your local store - check all low-fat, 2%-enters, etc. and then full milk - check the sugar content percentage. That's all... At least here in US, Los Angeles - usually it's the normal yogurt/milk (e.g. hihg-fat) that has lower sugary content (percentage-wise) than the rest. Sometimes it's quite significantly lower.

And I don't know how the whole craze about non-plain yogurt came. I'm bulgarian native, and it's "foreign" to me that yogurt is mixed with any sugar at all.

It's still yummy - you can mix it equal portions with water, shake it well - and you get a very good drink. Or make this cold soup - http://en.wikipedia.org/wiki/Tarator

Or just eat it like this. No need for sugar...

[rosser]

...so why would companies do that?

Because flavor. It turns out that when you remove fat from a foodstuff, it tends not to taste nearly as good.

[nitrogen]

Read the ingredients list for a simple salad dressing: basically oil, vinegar, seasonings. Now look a the "lite" version of the same dressing: oil is replaced with high fructose corn syrup, thickeners (probably starches), etc.

[tghw]

Straight from the article:

Malhotra also points to the United States, where percentage calorie consumption from fat has declined from 40% to 30% in the past 30 years (although absolute fat consumption has remained the same), yet obesity has rocketed. One reason, he says, is that the food industry “compensated by replacing saturated fat with added sugar.”

[mistermann]

Yogurt is one. I don't know about the previous sugar levels, but it is extremely difficult to find anything other than low or no fat yogurt, and they typically have what I consider to be very high sugar levels, even many brands of plain yogurt.

[elrzn]

How is that? My understanding is that lipid oxidation promotes insulin resistance itself.

If anything, and under the right dosage, sugar is protective.

[marze]

And as the quote states, not just one correlation:

heart disease correlated with total cholesterol

then

total cholesterol correlated with saturated fat consumption

[yetanotherphd]

I think it's more a matter of different beliefs arising from complicated, hard to interpret evidence. It's very comforting to believe in good guys and bad guys, but the real world of research is more complex than that.

[patrickg_zill]

It benefited the wheat farmers and "Big Ag" so there was possibly political pressure to focus on this.

[yetanotherphd]

What about "Big Dairy" or "Big Meat", they got the short end of the stick from the push to high carb low fat diets.

Industry lobbying as a theory lacks predictive power.

[mistermann]

Perhaps their lobbyists aren't as convincing.

This is the second time I've encountered the phrase "lacks predictive power on HN tonight" by the way. Is a theory without predictive power now considered wrong?

[voidlogic]

IMHO if it doesn't have predictive power it cannot be a theory:

the·o·ry noun \ˈthē-ə-rē, ˈthir-ē\ : an idea or set of ideas that is intended to explain facts or events

[swombat]

Explanation and prediction are two different things.

Your definition doesn't even require explanation - it just requires intent to explain.

So for example, "the moon is made of cheese" is a theory according to the definition you just quoted. It's an idea that is intended to explain the composition of the moon.