Oh no, Carbs

(Updated for clarity based on Facebook feedback)

There is a lot of hullabaloo about carbs in nutrition circles, especially in clean eating types of diets. Especially if those carbs are from sugar. And extra especially if that sugar is high-fructose corn syrup. But the reality is that for weight loss and general health goals* carbs are not, in and of themselves, a big deal.

The science is clear that low carb and low fat diets produce the same long term outcomes when measuring weight loss. These two links do a good job of addressing this:

Low Fat vs Low Carb – No Difference

Sorry, but Low Carb and Low Fat Diets Get Pretty Much the Same Results

This is the study being discussed: Effect of Low-Fat vs Low-Carbohydrate Diet on 12-Month Weight Loss in Overweight Adults and the Association With Genotype Pattern or Insulin Secretion

So, rather than repeat that content, I want to address why this is the more reliable research. Because, of course, there will be replies about research showing that low carb works better.

To analyze diet science we need to look at a couple of factors:


How long was the study? Any study of just a few months can show weight loss. The reality is that if you can’t show weight loss in a 12-16 week study then you are just bad at your job. Any possible diet strategy can show weight loss in this time frame. Such results are meaningless to this conversation.

The health outcomes that are important for this kind of research are ones that evolve over the course of decades. Some of the landmark studies in this field have been running longer than most of my readers have been alive.

So let’s be clear. 6 months or less is short. Too short for me to care about. Too short to inform human interventions. 1 year is the bare minimum to matter.

But why are studies like this even done if they are not meaningful? Well, they are preliminary research. Preliminary research plays an important role in the process of science but they are clearly not the end point. News stories about them are usually counterproductive and they shouldn’t be used to guide interventions in humans.

A preliminary study might just be used to determine in an approach is feasible, or to determine dosing, safety and similar characteristics. But they don’t tell us if an intervention works.

There is, of course, another more cynical reason for short-term diet studies: using them to sell stuff. Since positive results are just about guaranteed in a short-term study they create an easy way to apply the “clinically proven” sticker to your diet and resist FTC objections.

Petri Dishes and Animals

A lot of the objection to sugar and carbs is the result of a belief that calories from those sources are particularly bad for human health compared to protein or (good) fats. However, this comes pretty much entirely from research in petri dishes or animal studies.

Again this is another example of over-interpreting preliminary research. The results of applying a substance directly to a cell in a petri dish cannot be generalized to the whole human body. There are so many other factors at play in understanding how a substance will affect clinical outcomes. How is it digested? How does it interact with other substances in the body? How do homeostatic mechanisms operate to limit any changes? Etc.

And while animal research is a useful step in health research for humans we are, forgive me for stating the obvious, not mice. Or rats. Or guinea pigs. Or regular pigs. Or even monkeys. The question of what animal are we most like, a mouse or a rat, suffers from the fact that it depends on what part of the human system you are studying. And even then, we cannot just blithely assume that any animal results will apply to humans.


An extension of the last point is research into biomarkers. In short biomarkers are substances found with chemistry tests that are thought to indicate a particular (disease) state. Some are well validated, like troponin for a heart attack, others are not, like telomere length for aging. It is important to understand that a biomarker is correlated to the state of the body and is not a direct measure of the state of interest. They can be suggestive, but they don’t always predict actual health outcomes.

One of the most commonly misused biomarkers are any markers for inflammation. The assumption being that such markers indicate that something Bad is happening. So if eating sugar boosts the levels of some inflammatory biomarkers it must be terrible for you! But the research into actual health outcomes – death, disability etc. – doesn’t bear this out. Sure, eating too many calories leads to problems but the source of those calories doesn’t actually predict outcomes.


Yes it’s calories. Clearly though, understanding that is not the hard part. Behavior change is the hard part. Establishing new eating patterns and habits is the hard part. Fad diets that focus on one, or a few, components like sugar or carbs attempt to create the impression that it’s what you eat, not how much, that really matters. And it’s just not true.

* General health goals = is the phrase I’m using to mean premature death and disability/loss of function due to lifestyle modifiable diseases and risk.