Keto made easy

The DIETFITS Study: Low Fat vs Low Carb for a Year

Quick answer: DIETFITS was a year-long trial of 609 adults who followed either a healthy low-fat or a healthy low-carb diet. The low-fat group lost about 5.3 kg on average and the low-carb group about 6.0 kg, a difference the researchers did not find significant (Gardner et al., 2018). Genotype and insulin secretion didn't predict who did better on which diet. The useful lesson isn't "keto doesn't work". It's that the diet you can actually keep up, built on real food, decides the result.

On this page

What DIETFITS tested

The researchers weren't only asking "which diet wins?" They were also asking whether you could predict, from someone's genes or how much insulin they secrete, which diet would suit them.

So 609 adults were split into two groups for 12 months. One group ate a healthy low-fat diet, the other a healthy low-carb diet. "Healthy" is doing real work in both names: these were the whole-food versions of each diet, not low-fat cookies or low-carb snack bars.

What it found

The headline numbers, from the paper's abstract (Gardner et al., 2018):

Do the back-of-envelope arithmetic and the gap is about 0.7 kg, or roughly a pound and a half, over a whole year. That's well inside the normal swing of water weight on any given morning. You couldn't tell the two groups apart by standing them on a bathroom scale.

What it did not test

Here's where people over-read it in both directions.

It wasn't a keto trial. The low-carb arm was a moderate low-carb diet, not a ketogenic one. If you're eating under 20 to 50 g of carbs a day, you weren't the diet being tested.

It wasn't a metabolic ward study. Those are the studies where people live in a research unit and every bite is weighed, like the Kevin Hall studies with 17 and 19 people (Hall et al., 2016; Hall et al., 2015). DIETFITS ran 609 people through a year of ordinary life. That's its strength (it looks like real life) and its weakness (it can't tell you exactly what anyone ate).

An average isn't you. "No difference on average" can hide a lot of people doing very well on one diet and poorly on the other. The trial's own question was whether you could predict that in advance from genes or insulin, and the answer was no. That doesn't mean everyone responds the same. It means those particular tests didn't sort people.

Why "no winner" is still useful news

If two quite different diets produce nearly the same result over a year, the thing they had in common is probably doing most of the work. Both were the "healthy" version of their diet, built on real food. Both groups had to stick with it for a year.

So what you do is pick the way of eating that makes the deficit easiest for you to live with, and then make it boring enough to repeat. If cutting carbs turns your hunger down and you stop thinking about food all afternoon, that's your edge. If you'd be miserable without bread, a healthy low-fat plan may get you just as far.

And if you're working on blood sugar rather than the scale, that's a different question from the one DIETFITS answered. Talk it through with your doctor, especially if you take diabetes medication, because cutting carbs can change how much of it you need.

What the research says

DIETFITS fits a pattern that shows up across the bigger, longer trials:

Put together: a modest early edge for cutting carbs, a small or vanishing edge at a year, and a lot of individual variation underneath every average.

KetoDick's take

DIETFITS came up in his 2019 response to a Men's Health article that used it as evidence against keto (video 1JbK1_ClDmY). He questioned the study, guessing from the article's description that it was self-reported rather than a ward feeding study, and pointing to the wide spread of individual results the article described. His bigger point, from the video, was that a calorie deficit gets you there either way, and that appetite control is what makes the deficit comfortable.

The research partly backs him and partly doesn't. He's right that it wasn't a ward study. He didn't raise the bigger limit, that the low-carb arm wasn't ketogenic. And the trial was randomized and large, so "it was self-reported" isn't a reason to throw it out. It's a reason to read it for what it is: a real-life test of two healthy diets.

His long-standing position lines up with the study's lesson more than the article's spin did. He's said that adherence is the whole game, and that every extra rule makes a diet harder to keep (videos UygKGoP_BtA, LbPnm9T0oxc). He's also said he doesn't think keto works for everybody, because genetics and gut differ (video cTGYIFOTPl8). That's close to what DIETFITS found.

Key takeaways