17 February 2006

On Fats

ST Feb 15, 2006
Is this back on the menu?

Supporters of low-carb, high-fat diets gloated last week as doctors scrambled to contain the damage of a huge and expensive study which set out to show the benefits of a low-fat diet - and couldn't.

The study involved nearly 49,000 women aged 50 to 79 who were followed for eight years. Those on a low-fat diet were found to have the same rates of breast cancer, colon cancer, heart attacks and strokes as those who ate whatever they pleased.

Hurray! said everyone else, while doctors hurried to come out in support of low-fat diets anyway. Even if the largest study ever done on the subject showed that it didn't cut the risk of heart disease and cancer, they were sticking by the mantra that low fat was good for the heart.


The thing about these studies is how contradictory they can sound. One day, you're told to eat margarine; the next day, you find out that butter is better. One day, you're loading up on fibre; the next, some surgeon says the way to cure constipation is to go slow on the greens.

What's one to make of these mixed signals?

Mr Wang is neither a doctor nor a nutritionist. So he may not be very qualified to comment on this topic. However, from a layman's commonsensical point of view, Mr Wang feels that this study on low-fat diets and their effect on the risk of major diseases may be missing a rather important point.

In Mr Wang's mind, the reason for being on a low-fat diet is so as not to be fat. The reason for not being fat is that you become less likely to get cancer, heart disease etc. However, being on a low-fat diet may or may not succeed in making you thin.

For example, if you're on a low-fat diet but you consume huge quantities of low-fat food (for example, lots of rice and bread every day), you could still be very fat. Conversely, you might not be on a low-fat diet but if you exercise a lot (run, swim, cycle etc), you may still be quite slim.

The study did not seem to have focused on whether the women on low-fat diets were indeed thinner than the women who ate whatever they pleased. The scientific conclusion was merely:

Women on a low-fat diet were found to have the same rates of breast cancer, colon cancer, heart attacks and strokes as those who ate whatever they pleased.
This is NOT the same as saying:
Women who were thin were found to have the same rates of breast cancer, colon cancer, heart attacks and strokes as those who were fat.
Who knows? Perhaps the truth is simply that if you're thin, you can eat whatever you like. And if you're fat, you ought to lose some fat. Oh, and that just going on a low-fat diet may not be an effective way to do that.


Gilbert Koh aka Mr Wang said...

An interesting point. :) Perhaps vanity is a key factor in determining which women are likelier to get cancer, heart disease etc. Those who are more vain make more effort to keep slim thereby reducing their health risks.

Anonymous said...

evolution is the survival of the prettiest - from popular the sitcom

Mr Jherek said...

Curious, have some of the illustrious Mr Wang's posts disappeared.

Elia Diodati said...

Disclaimer: IANADoctor, but I am a chemist and I know enough biochemistry to be dangerous.

The thing is, a low fat diet is not necessarily one that will make you lose weight. A diet that is one tablespoon peanut butter and two dozen boxes of Willy Wonka Nerds is just as likely to make you too heavy to walk as a diet of two jars of peanut butter every day. Good nutrition is not as simple as low fat, which does not automatically imply weight loss. There are simply far too many other reasons (genetics and sedentary lifestyle being two of the most important factors) for people to be fat.

Scientific studies in themselves have no intrinsic "point" per se. Scientists make a hypothesis, they raise funding for experiments, they collect data, they test whether or not the data support the hypothesis. That's how science works. Making a pragmatic statement such as "this study misses the point" itself misses the point that science is an empirical process of testing claims with observations, not one that prescribes value judgments.

However, the article you have posted does very little justice to the point of the study. The justification for the study is still of tremendous importance because it overturns decades of "conventional wisdom". Doctors who continue to espouse it do so at the peril of established notion.

Besides, doctors may consider themselves well-justified for other reasons. For example, it is now quite a well-established fact that not all fats are created equal. Olive oil actually lowers your cholesterol levels, salmon oil keeps your skin smooth, while generic vegetable oil is more likely to be laden with trans fats which have been shown to promote risks of cancer. The problem is that this study did not distinguish between the kinds of fats now known to exhibit such varied and different bodily reactions. This point that not all fats are created equal was not as well appreciated 5 years ago as it is today. Therefore this study, which was planned and approved about 10 years ago, did not take this into account and it's not really anyone's fault.

As to why the journalist appears to be so confused about the whole point (and appears to have succeeded very well in spreading the ignorance), I postulate that people not trained in the technical sciences are not used to kind of hair-splitting reductionist intellectual debate that characterizes science. Frustrating? Undoubtedly, especially when one has to plow through the vocabulary of a technical field. But pointless? Not if the scientists know what they are doing!

One has to bear in mind that when one does science, one has to pick something very specific to study. Science writers face the challenge of putting the big picture together from thousands of such studies and applying some well-established principles of biochemistry to draw a picture that makes any kind of sense.

It is hard enough for experts such as doctors and trained medical scientists to keep abreast of all the data now being published. Perhaps I shouldn't be so surprised to see an article like this that basically says, "Eh, last time they say this, now they say that. How ah?" It's still depressing, though.

People who read health articles in the newspaper: go learn some biochemistry and test if these studies make sense to you! This is the 21st century, where an afternoon of due diligence in front of Google can make you more expert on a rare illness than 99.99% of doctors out there who don't have the time to do so. You have to believe that all this can indeed make sense to you!

Gilbert Koh aka Mr Wang said...

Well, here's an article that talks about the flaws of the study. Naturally it is written by someone who has a vested interest in promoting low-fat diets, heheh. Not saying that he is necessarily wrong - just saying that he has a vested interest.

Which leads me to my next point - that science isn't necessarily as objective and free of value judgments as you have suggested it to be. I can think of different examples - how about, for example, the intelligent design vs evolution debate; the purported "effectiveness" of reparative therapy for gays - where parties have motivations for taking particular stances, and those motivations aren't related to the burning desire to scientifically investigate the universe.

It's pertinent to point out as well that science is often driven by commercial interests - you don't really think that the S'pore govt is promoting life sciences research just out of scientific curiosity, do you - but the fact that science is driven by commercial interests also means that it necessarily loses some objectivity - for example, if Teflon is really carcinogenic, you wouldn't expect DuPont's scientists to be the first to point it out, would you?

More later .... the Little Kids want Mr Wang's attention now.

Heavenly Sword said...

Related to Elia Diodati's comments: The SSK, 'Edinburgh School'

expat@large said...

Thin people never die!

Mr Wang Say So!

"Rolling down a mountain - I'm sure the fat man wins." Jethro Tull.

expat@large said...

Diodata, a lot of 'clinical' (i.e. in the patient) biochemistry is presumption and guesswork, so I doubt that spending your entire life seeking The Truth on Google (as some do) would enlighten more than confuse.

Explain for example why Prozac "works" where there is no change in measured serotonin levels. Why suldefanil works on the penile arteries and not the coronary (actually that has been explained, but long after the drug was initally tested for angina).

Another instance: olive oil, being a mono-unsaturated fat, may NOT be as good for you as poly-unsaturated oils, like sunflower or safflower. There is a alternative school of thought extant which says that the supposed benefits of olive oil are more related to the rest of the so-called mediterranean diet (fish mainly), and that the olive oil component is in fact the worst part of it. Olive oil is just along for the ride. And mediterranean people die eventually too.

This whole diet issue is all to do with Big-Pharma's plan to take over the world by making every aspect health and living some potential disease requiring a patented pill or potion.

expat@large said...
This comment has been removed by a blog administrator.
expat@large said...

Actually Mr Wang, just thinking further about what you said:

I think it may be YOUNGER people who BELIEVE they are overwieght who diet to be thin. Health is of no concern as they assume they will live for ever.

Older people, who know differently, the genuinely overweight, such as um, someone around here, diet in order to live long enough to buy next year's calendar.

This report is of very great concern to me, I mean them.

Most rational nutritionists accept that any reduced-calorie diet will enable you to loose weight (and that's what most of the crash-diet fads - atkins, zone, south-beach, chicken-soup - actually do, lower your calorie intact) but the argument is about which particular type of diet is HEALTHIER, as in less likely to "cause" those diseases they studied.

Gilbert Koh aka Mr Wang said...

Interesting comments, all. Mr Wang will digest and possibly post further thoughts on health information and what we can all do with it.

singaporean said...

Aiyoh, you guys are all so passe.

Infectious Viruses Linked to Obesity

And so is cancer!

So...dont waste time exercising, eat whatever you want and live in a bubble!

Elia Diodati said...

Mr Wang, I had read the articles in question and they had admitted that they did not test the hypothesis that lowering the dosages of certain kinds of fats vs. null (i.e. regular diet) as testing the hypothesis of reducing fat intake overall. Yes there are limitations to the study, but that is my point, that the study is carried out according to the initial hypotheses present, namely that a low-fat diet (defined as being ca. 25% lower than the national average) reduces risk factors in three kinds of leading diseases. The study states that the data collected in the study were statistically inconclusive about disproving the null hypothesis.

As a scientist, I do not believe that the evolution v. intelligent design debate is scientific. The evo-devo debate is not about the science of the origin of life, it is simply trying to get the public to be aware of the scientific method. To date, ID proponents have provided no testable hypotheses, which by definition is the only things that science is concerned with. If you are interested in refutation of the specific debates that crop up, I can refer you to excellent blogs such as The Panda's Thumb that counter ID claims with the latest in scientific thinking. To call it a scientific debate is really a non sequitur.

A little bit of context might be order for the issue of medical treatment for homosexuality: until the 1960s-70s, the medical community had classified homosexuality as a psychiatric condition. It was only with increasing sociological evidence mounting in the 1970s that identical twin studies showed high correlations of homosexuality (tipping the scales in favor of nature over nurture) and the litany of genetic evidence in the 1990s for genetic correlations with sexual preference that this viewpoint is now discredited among the medical profession. Note again the fluxional quality of medical "truth".

As for the motivations of the people on the rostrum, it is painfully obvious that on one side of these debates are the vast majority scientists doing their thing, and on the other are religio-polical extremists advocating some kind of agenda. You may criticize this statement for not being value-neutral, but I would say the available evidence in my opinion validates my choice of words.

The issue of sponsorship is an important point, and ultimately it boils down to trust in the sponsor's reputation and the perception of conflict of interest (or lack thereof). The relevant case to look at were the Phillip Morris-funded studies indicating that smoking did not lower the incidence of lung cancer. P.M. later admitted to biasing the data by selective shredding to negative data, and I believe they settled out of court because some federal agency sued them for misrepresentation.

In the case of the low-fat diet study, the sponsor was the US National Institute of Health, which I would believe is a platform that is as neutral as it gets in that it espouses no commercial or political agenda. If anything, it would be predisposed to uphold "conventional wisdom" since government agencies are notorious for wanting to uphold the status quo. Yet here they are freely accepting the outcomes of the study. No public histrionics from the NIH over how stupid they were in spending $400m in taxpayers' money to discover nothing. Consider that, if you will.

expat@large: I freely admit my ignorance of the topics you have chosen to challenge me with. However, I disagree with your dismissal of clinical biochemistry as mere "presumption and guesswork". It is the same quality of presumption and guesswork that is in all of science. Nothing can be proven in science, but it is easy to refute a theory by presenting unambiguous data to the contrary. Scientific truth is defined as whatever is left after striking out all the statements that have been proven to be false to date. Nothing more, nothing less. This gives scientific truth a fluxional quality (unlike mathematical theorems) that can give rise to confusion. Despite this, the empirical worldview of science is not diminished by this axiomatic flaw, given that all our modern technology has allowed us to do (including having this discussion on this nondescript server at an undisclosed physical location).

As for the issues of clinical trials, the human body is simply too complex. Questions like "Is drug X going to work in patient Y without side effect Z?" are undoubtedly important, but it is impossible to answer them with certainty short of carrying out of the experiment (and bearing the responsibility of all the ethical baggage therein). X, Y and Z are all incredibly complex phenomena which are at best poorly understood. This is why many issues in social and medical sciences are answered statistically. This is why all (good) doctors talk about are risk factors for cancer, not the certainty of you dropping dead from (say) cardiac arrest on your 54th birthday when you are blowing out the candles on your cake. Statistical methods, when properly applied, are about making the best of limited data, and are not merely guesswork. If it is, at least it's educated guesswork.

On the other hand, basic principles of biochemistry, for example that the metabolism of fat is more time-consuming than the metabolism of simple sugars, are inviolate in that in the entire recorded scientific literature no one has shown convincing evidence to the contrary.

I am not advocating the search for Nirvana on Google, but I am a proponent of people doing their own research on Google and collect information to answer specific questions, and then make informed choices based on the available information as compared to just following the herd in whatever the current fashionable health advice is.

Science is not value-neutral, considering that the practice of science is carried out by humans. However, the method of experimentation is value-neutral, in the sense that once a hypothesis is framed, and data is analyzed, the hypothesis can be tested. Yes, one can deliberately construct a skewed hypothesis, and yes, methods of analysis can be biased, but in science, the facts and methods are laid out for all to see (modulo a subscription to the actual publication in question). Given sufficient time and effort, scientific claims can be validated by repetition and verification of the raw data. This is what happened behind-the-scenes vide the South Korean stem cell scandal a few months ago, and is so far effective at rooting out fraud once suspicion has been presented.

On a side note, this issue of value-neutrality is much more prevalent in the social and medical sciences than in the physical sciences. Just because you are a conservatist American doesn't mean that the laws of gravity don't apply to you because they discovered the law in England.

Maybe I should have advocated instead that people learn biochemistry and statistics and then go and look up stuff on Google. My point is still that many people are not used to scientists presenting data to debate very narrow points and have a very large propensity to overgeneralize.

Elia Diodati said...

As a subsidiary point, serendipity does occur in science, but that still does not detract from its value. Your Viagra example is a good case in point. I heard a story from a clinician claiming to have been involved in the initial Pfizer trial that the wives of male patients were bitterly adamant about keeping their spouses on the regimen, which supposedly was how Pfizer figured out what kind of potency their drug had.

expat@large said...

ED, I suppose the points I was trying to make are

a) there is a lot of crap information on the internet

b) the human body is more complex than we know at present.

This does not mean we should abandon the quest for sceinetific truth. Far from it - merely that we should be aware that even experts disagree -

I have worked with professors at the highest levels of medicine, and they argue with each other over points that readers of their own textbooks would have thought were settled aeons ago...

As you rightly say, truth in science is essentially the scratching out what theories are wrong, but don't discount that EVERYTHING is open for reinterpretation and in my experience, personality clashes tend to rule most things in life, from science to politics to what's for dinner..

"Science validates bias precisely by appearing to remove it." Stephen Jay Gould.

Of course my opinion is relative and open to reinterpretation over a few beers...

Anonymous said...

I think everyone also forgot one very relevant point regarding the study - the sample class are women above 50.

IF they have been having nice, fried chicken wing and coconut milk rich curries for the last 40 years, 8 years of low fat diet isn't going to help very much.

Elia Diodati said...

expat@large and klimmer: Your points are well taken.

One final comment: science is somewhat open source in its approach. With a subscription and a sound familiarity with the fundamentals (and also an afternoon with nothing better to do, I suppose), people can view the claim and the evidence and judge for themselves whether they believe the work or not.

On a formal level this is all well and good, but what if you don't want to (or can't) read the primary literature? I believe this is where most of the complications come in, when one has to resort to second-hand (or worse) reporting. How much can you trust indirect accounts? And on a related note, should I trust a quote from a perceived expert, even though it is technically a logical fallacy to appear to higher authority?

Good luck on your diet, by the way. It's a real shame that virtually all foods fall into three categories (in my non-expert and non-scientific classification, which BTW is especially open for reinterpretation):
1. They increase your risk for heart disease.
2. They increase your risk for cancer.
3. They taste bad.

Gilbert Koh aka Mr Wang said...

4. Fruits.

I like fruits. They taste good. :)

Personally, I think that in health, as in many other areas, the amount of conventional, extremely-likely-to-be-correct-&-reliable advice already grossly exceeds the ability of most individuals to apply it. Never mind the latest this or the latest that, in scientific studies.

For example, I think that the following would be very good, albeit basic advice:

1. Exercise regularly;
2. Don't smoke;
3. Don't be overweight;
4. Eat your fruits and veggies;
5. Get enough rest;

... but I think that very few people would be able to implement all these into their lifestyles. Point 1 alone trips up at least 50% of the general population, I imagine.