This story is so important that I’m going to forgo the usual 3-story format of the newsletter and dedicate this entire issue of “In Step with Jonny” to a discussion of the new study on low-carb diets and mortality.
You may have seen some alarming headlines recently about this new study on low-carb diets. (They seemed alarming to me too — until I read the actual study.)
A sample of the headlines I’ve seen include “Low-carb diet has high risks”, “New study shows low-carb diets increase mortality” and — the most odious example — Atkins Diet Increases All-Cause Mortality” by none other than Dr. Dean Ornish. (More on that particularly vile distortion in just a moment.)
The lead author of the study — Dr. Teresa Fung — has just graciously agreed to be interviewed by me, and I will talk at length about the interview next week, in part two of this article. As far as I know, I am the first blogger to get this interview, and I’m very excited to share what I learn next week.
Now let’s get to the research — and to the disgraceful reporting on it by the Mainstream Media.
The study in question — “Low-Carbohydrate Diets and All-Cause and Cause — Specific Mortality” — was published recently in the Annals of Internal Medicine. It was not an actual clinical experiment , but rather a statistical analysis of data that’s been generated by the Nurses Health Study and the Health Professionals Follow-Up Study, two very long-term studies of diet that have been going on for thirty years and have involved close to 200,000 participants.
*All-Cause” mortality means death from anything at all, from cancer to a car accident; “Cause-Specific” mortality refers to death that can be traced to a specific illness or event, like cancer or a heart attack.
For this particular investigation, the researchers looked at about 45,000 men and 85,000 women whose histories were followed from the 1980’s through 2006. Their diets were assessed through food questionnaires and then rated by the researchers as being either “high” or “low” carbohydrate.
Then the researchers broke down the “low-carbohydrate” eaters into two sub-groups: those who got most of their protein and fat from vegetable sources, and those who got most of their protein and fat from animal sources.
With me so far? Good.
Next, the researchers looked at how many people had died over the course of the 20 some years that data was collected, a number which turned out to be roughly 21,000 of the 125,000 people in the study. They then looked to see if there was any relationship between low-carb eating and an increased risk of dying.
The “animal” low-carb group had a slightly higher risk of dying, the “vegetable” low-carb group had a slightly smaller risk of dying.
Now one thing you need to know before we go on is that assignment to the “low-carb” or “high-carb” group depended entirely on food questionnaires. I don’t know if you’ve ever filled out a food questionnaire before, but they are a standard way of doing dietary research. But just about everyone in the field knows they are notoriously unreliable. I don’t know about you but I can barely remember what I had for breakfast yesterday. A food questionnaire asks things like “how many times last month did you have meat”? “how many times a week do you eat vegetables?”. Good luck with that.
Now I don’t want to completely trash food questionnaires because they are a very useful tool for dietary research. But if you think about it even for a minute, you’ll see exactly what some of the problems are, even aside from faulty memory.
For example: In one persons mind, “two vegetable servings” consists of the sliver of lettuce and tomato he put on his Big Mac. Meanwhile, another person counts large portions of steamed broccoli and cauliflower as his “two vegetable servings”. The questionnaire- and the researchers — can make no distinctions between these two people, just as they make no distinctions between people eating high-nitrate, high-sodium baloney and people eating a grass-fed antibiotic free burger from Novy Ranches! As far as the food questionnaire goes, each of those is a “serving” of “meat”; similarly, a vegetable “serving” of ketchup is equal to a vegetable serving of Brussels sprouts.
And if you think this is a small point, think again. A recent study from Harvard actually divided “meat eaters” into two groups (for the first time ever in a serious, peer-reviewed research study). They found that people eating processed meat did indeed have a higher risk of heart disease, but people eating non-processed meat had zero increased risk. The type of meat you eat — as well as the type of vegetables — makes a profound difference.
This study did not make any distinctions among the type of meat that people ate. It’s also interesting to note that the “animal products low-carb” group were also more overweight, were more likely to be smokers, and had lower intakes of fruits and vegetables. Of course, these things are “corrected” for in the statistical analysis, but it gives you a good sense of the kind of lifestyle people in the “animal-based diet” group were likely to have.
So what this study really shows is a very slight correlation between all-cause mortality and reporting that you eat a lot of meat (of dubious quality) and very few fruits and vegetables.
Now consider the most absolutely intellectually dishonest example of spinning this study for the sake of advancing an agenda. The article by Dean Ornish in the Huffington Post.
Let’s start with the headline: “Atkins Diet Increases All-Cause Mortality”.
The first thing you should know is that nowhere in the entire study– and I read every word- is the name “Atkins” mentioned (except when the researchers reference a study done earlier this year that was dubbed the “Eco-Atkins” study). No one was put on an Atkins diet. (In fact, as I explained above, no one was put on any diet.) So the headline is a complete and utter lie.
The researchers themselves specifically caution against making a statement like Ornish did, but apparently Dr. Ornish missed that part of the article. Here’s what the researchers themselves said.
“The low-carbohydrate diet scores were not designed to mimic any particular versions of low-carbohydrate diets available in the popular literature.
Therefore, the risk estimates do not directly translate to the assessment of benefit or risk associated with the popular versions of the diet.”
I’ll have a lot more to say about this study after I interview Dr. Fung, and even more to say about how the media has reported the findings. For now, let’s just say that this is a perfect “teaching moment” where we can really explore what research says and doesn’t say, and how individual agendas (such as Dr. Ornish’s) and misinformed reporters can completely slant the facts so that what you actually hear is very different from what actually happened.
Most people are aware that this happens in politics all the time.
What most people are not aware of is that it also happens in science. And is happening right now, again, as it frequently has over the past several decades, in the “Diet Wars”.