A Brief Response to Taubes's Food Reward Critique, and a Little Something Extra

It appears Gary Taubes has completed his series critiquing the food reward hypothesis of obesity (1).  I have to hand it to him, it takes some cojones to critique an entire field of research, particularly when you have no scientific background in it, and have evidently not read any of the scientific literature on it.  As of 2015, a Google Scholar search for the terms “food reward” and “obesity” turned up 2,790 papers.

The food reward hypothesis of obesity states that the reward and palatability value of food influence body fatness, and excess reward/palatability can promote body fat accumulation.  If we want to test the hypothesis, the most direct way is to find experiments in which 1) the nutritional qualities of the experimental diet groups are kept the same or at least very similar, 2) some aspect of diet reward/palatability differs, and 3) changes in body fat/weight are measured (for example, 2, 3, 4, 5, 6, 7, 8, 9).  In these experiments the hypothesis has both arms and one leg tied behind its back, because the most potent reward factors (energy density, sugar, fat) have nutritional value and therefore experiments that modify these cannot be tightly controlled for nutritional differences.  Yet even with this severe disadvantage, the hypothesis is consistently supported by the scientific evidence.  Taubes repeatedly stated in his series that controlled studies like these have not been conducted, apparently basing this belief on a 22-year-old review paper by Dr. Israel Ramirez and colleagues that does not contain the word 'reward' (10).

Another way to test the hypothesis is to see if people with higher food reward sensitivity (due to genetics or other factors) tend to gain more fat over time (for example, 11, 12, 13, 14, 15, 16).  In addition, studies that have examined the effect of palatability/reward on food intake in a controlled manner are relevant (17, 18, 19, 20, 21, 22), as are studies that have identified some of the mechanisms by which these effects occur (reviewed in 23).  Even if not all of the studies are perfect, at some point, one has to acknowledge that there are a lot of mutually buttressing lines of evidence here.  It is notable that virtually none of these studies appeared in Taubes's posts, and he appeared unaware of them. 
Read more »

Ancestral Health Symposium 2015

I recently returned from AHS12 and a little side trip to visit family.  The conference was hosted at Harvard University through the Harvard Food Law Society.  Many thanks to all the organizers who made it happen.  By and large, it went smoothly.

The science as expected ranged from outstanding to mediocre, but I was really encouraged by the presence and enthusiastic participation of a number of quality researchers and clinicians. The basic concept of ancestral health is something almost anyone can get behind: many of our modern health problems are due to a mismatch between the modern environment and what our bodies "expect".  The basic idea is really just common sense, but of course the devil is in the details when you start trying to figure out what exactly our bodies expect, and how best to give it to them.  I think our perspective as a community is moving in the right direction.

Read more »

The Carbohydrate Hypothesis of Obesity: a Critical Examination

Introduction

I'd like to begin by emphasizing that carbohydrate restriction has helped many people lose body fat and improve their metabolic health.  Although it doesn't work for everyone, there is no doubt that carbohydrate restriction causes fat loss in many, perhaps even most obese people.  For a subset of people, the results can be very impressive.  I consider that to be a fact at this point, but that's not what I'll be discussing here. 

What I want to discuss is a hypothesis.  It's the idea, championed by Gary Taubes, that carbohydrate (particularly refined carbohydrate) is the primary cause of common obesity due to its ability to elevate insulin, thereby causing increased fat storage in fat cells.  To demonstrate that I'm representing this hypothesis accurately, here is a quote from his book Good Calories, Bad Calories:

Read more »

Safflower Oil Study

A few people have sent me a new study claiming to demonstrate that half a tablespoon of safflower oil a day improves insulin sensitivity, increases HDL and decreases inflammation in diabetics (1). Let me explain why this study does not show what it claims.

It all comes down to a little thing called a control group, which is the basis for comparison that you use to determine if your intervention had an effect. This study didn't have one for the safflower group. What it had was two intervention groups, one given 6.4g conjugated linoleic acid (CLA; 50% c9t11 and 50% t10c12-CLA) per day, and one given 8g safflower oil. I have to guess that this study was originally designed to test the effects of the CLA, with the safflower oil group as the control group, and that the interpretation of the data changed after the results came in. Otherwise, I don't understand why they would conduct a study like this without a control group.

Anyway, they found that the safflower oil group did better than the CLA group over 16 weeks, showing a higher insulin sensitivity, higher HDL, lower HbA1c (a marker of average blood glucose levels) and lower CRP (a marker of inflammation). But they also found that the safflower group improved slightly compared to baseline, therefore they decided to attribute the difference to a beneficial effect of safflower oil. The problem is that without a control (placebo) group for comparison, there's no way to know if the improvement would have occurred regardless of treatment, due to the season changing, more regular check-ups at the doctor's office due to participating in a study, or countless other unforeseen factors. A control group is essential for the accurate interpretation of results, which is why drug studies always have placebo groups.

What we can say is that the safflower oil group fared better than the CLA group, because there was a difference between the two. However, what I think really happened is that the CLA supplement was harmful and the small dose of safflower oil had no effect. Why? Because the t10c12 isomer of CLA, which was half their pill, has already been shown by previous well-controlled studies to reduce insulin sensitivity, decrease HDL and increase inflammatory markers at a similar dose and for a similar duration (2, 3). The safflower oil group only looked good by comparison. We can add this study to the "research bloopers" file.

It's worth noting that naturally occurring CLA mixtures, similar to those found in pastured dairy and ruminant fat, have not been shown to cause metabolic problems such as those caused by isolated t10c12 CLA.