Monday, September 5, 2011

Nonlinearity and the industrial seed oils paradox

Most relationships among variables in nature are nonlinear, frequently taking the form of a J curve. The figure below illustrates this type of curve. In this illustration, the horizontal axis measures the amount of time an individual spends consuming a given dose (high) of a substance daily. The vertical axis measures a certain disease marker – e.g., a marker of systemic inflammation, such as levels of circulating tumor necrosis factor (TNF). This is just one of many measurement schemes that may lead to a J curve.


J-curve relationships and variants such as U-curve and inverted J-curve relationships are ubiquitous, and may occur due to many reasons. For example, a J curve like the one above may be due to the substance being consumed having at least one health-promoting attribute, and at least one health-impairing attribute. The latter has a delayed effect, and ends up overcoming the benefits of the former over time. In this sense, there is no “sweet spot”. People are better off not consuming the substance at all. They should look for other sources of the health-promoting factors.

So what does this have to do with industrial seed oils, like safflower and corn oil?

If you take a look at the research literature on the effects of industrial seed oils, you’ll find something interesting and rather paradoxical. Several studies show benefits, whereas several others hint at serious problems. The problems seem to be generally related to long-term consumption, and to be associated with a significant increase in the ratio of dietary omega-6 to omega-3 fats; this increase appears to lead to systemic inflammation. The benefits seem to be generally related to short-term consumption.

But what leads to the left side of the J curve, the health-promoting effects of industrial seed oils, usually seen in short-term studies?

It is very likely vitamin E, which is considered, apparently correctly, to be one of the most powerful antioxidants in nature. Oxidative stress is strongly associated with systemic inflammation. Seed oils are by far the richest sources of vitamin E around, in the form of both γ-Tocopherol and α-Tocopherol. Other good sources, with much less gram-adjusted omega-6 content, are what we generally refer to as “nuts”. And, there are many, many substances other than vitamin E that have powerful antioxidant properties.

Chris Masterjohn has talked about seed oils and vitamin E before, making a similar point (see here, and here). I acknowledged this contribution by Chris before; for example, in my June 2011 interview with Jimmy Moore. In fact, Chris has gone further and also argued that the vitamin E requirement goes up as body fat omega-6 content increases over time (see comments under this post, in addition to the links provided above).

If this is correct, I would speculate that it may create a vicious feedback-loop cycle, as the increased vitamin E requirement may lead to increased hunger for foods rich in vitamin E. For someone already consuming a diet rich in seed oils, this may drive a subconscious compulsion to add more seed oils to dishes. Not good!

21 comments:

Sam Knox said...

Hey Ned,

You may have already seen this, but here's an article that offers an alternative explanation for the "seed oils paradox":

http://www.newyorker.com/reporting/2010/12/13/101213fa_fact_lehrer

It suggests that initial results showing a benefit from seed oils may have been nothing other than dumb luck.

Sam Knox

Sam Knox said...

If that link doesn't work, search for "The Truth Wears Off".

Stabby said...

That is certainly one plausible mechanism. Another is that in the short-term dietary fat is in fact anti-inflammatory in a few ways. http://www.ketotic.org/2011/08/dietary-fat-cck-and-cholinergic.html?showComment=1314891170337&m=1#c7613550842783782797

Of course you could get the same benefit with any fat, without overloading your tissue phospholipids with linoleic acid.

Ned Kock said...

Hi Sam, thanks. Interesting article.

Ned Kock said...

Hi Stabby, thanks. Right, the all-important vagus nerve. That is a very short-term effect, one that indeed I’ve seen in connection with saturated fat and blogged about here. It is much shorter-term than the short-term effects of seed oils I’ve referred to on this post.

Ron Lavine said...

Thanks for this contribution to our thinking. It seems challenging to draw conclusions about the health effects of various dietary components for a number of reasons. One of them is that you can't study one particular dietary component out of context with someone's entire diet.

For instance, coffee is "good" for you because it's a high source of anti-oxidants. But is that only significant for someone who eats almost zero fresh vegetables? Then coffee would comprise their largest source of "vegetables".

Ned Kock said...

Hi Ron. To compound that problem, there is little incentive (funding) to conduct research on things like the effect of coffee consumption on health, including covariates like vegetable consumption. The research that ends up being done is of low quality, often failing to account for multivariate relationships and nonlinearity.

Ned Kock said...

I should also say that nutrition research in general tends to be self-funded (i.e., the researcher does without any external funding), at least from what I hear from colleagues in the field.
On the other hand, research on the effects of drugs to treat degenerative diseases, such as cholesterol-lowering drugs, is very well funded.

David Isaak said...

@Ron--

At the risk of seeming to miss your real point, I'd like to quibble with "coffee is good for you because of anti-oxidants."

Whenever researchers find that some plant food seems to have a health benefit and can't put immediately put their finger on why, they trot out the fact that the food contains anti-oxidants. Closer looks--if anyone bothers--usually reveal the benefits to be from something else. Apparently researchers are afraid to investigate benefits without first proposing a feasible mechanism for those benefits.

Still, your fundamental point is a sound one. For instance, almost all experiments showing problems from saturated fat include huge quantities of carbs or industrial oils along with the sat fat. Eades summarizes one study quite nicely:

http://www.proteinpower.com/drmike/lipid-hypothesis/saturated-fat-study-sucks/

Chris Masterjohn said...

Hi Ned,

Thanks for acknowledging me! I caught that in your interview with Jimmy and I appreciate that you have been so persistent in crediting me despite the lack of concrete source documentation. :)

Here is the letter I published:

http://www.ncbi.nlm.nih.gov/pubmed/17466237

That letter was a spinoff of this much longer article I put on my web site:

http://www.cholesterol-and-health.com/One-High-Saturated-Fat-Meal-Can-Be-Bad-Carrot-Cake-Coconut-Oil.html

There, I wrote, "Safflower oil may raise the amount of vitamin E in lipoproteins immediately after a vitamin E-rich meal, but what is the long-term effect on vitamin E status of a high intake of polyunsaturated fats?"

I may have written or spoken about this elsewhere, perhaps on a podcast interview or two. I will eventually write a post specifically addressing this, but there is good data that as adipose tissue becomes saturated in PUFA over the course of say four years on a high-PUFA diet, the requirement for vitamin E steadily increases such that the benefit of vitamin E in the oil is lost. Likewise, switching to a low-PUFA diet in such a state could lead to vitamin E deficiency for a few years while adipose PUFA normalizes.

Chris

Chris Masterjohn said...

By the way Ned, I'm not sure how it is at other universities but in our department I don't know of anyone who is self-funded. I think generally in published papers you will see outside sources of funding listed.

Chris

Ned Kock said...

Great Chris, thanks. I knew that you had written about seeds oils and vitamin E from a comment you made on a post, but I wasn’t aware that you had written so much and so cogently about it.

Ned Kock said...

What I mean by self-funded is something like a faculty member doing research as part of his/her normal duties, usually with the help of students, but with no significant external funding as a principal investigator (PI) from places like the NIH, NSF etc.

Indeed, I may be overestimating the percentage of nutrition researchers that do truly self-funded research. Many of the researchers that are not PIs themselves may be getting funding from grants received by someone else in their institutions, often to pay for their salaries.

Ned Kock said...

I revised the last paragraph of the post, adding the links … and some speculation.

Andrew said...

I know the community doesn't like seed oils but how come so many studies show reduced heart disease end-points when saturated fat is replaced with unsaturates?

Andrew said...
This comment has been removed by the author.
Ned Kock said...

Hi Andrew. The benefits seem to be generally related to short-term consumption. That is the left side of the J curve.

In the long term is where the problem lies …

Chris Masterjohn said...

Hi Ned,

I can't speak for others except to say that all of the research that goes on in my lab is from outside sources and I've been under the impression that is true for most other labs here. I'm sure it depends on the nature of the university too.

I'm not sure when I'll get to it, but I'll try to write a blog about this topic soon. If and when I do that, could I use your graphic from this post, with the appropriate link and credit to you?

Thanks,
Chris

Ned Kock said...

> I'm not sure when I'll get to it, but I'll try to write a blog about this topic soon. If and when I do that, could I use your graphic from this post, with the appropriate link and credit to you?

Of course. I look forward to reading the post.

Ned Kock said...

Btw, Chris. I’ve read your recent posts on the “animal protein causes cancer” controversy now erupting again with the release of Forks Over Knives, and they are fantastic. So was Denise’s post.

online bookie said...

great theme thanks for writing