DIETFITS – looking at the data

Executive Summary

* A Monday note in January reviewed the latest paper from the DIETFITS study. This study randomised 609 men and women to either a healthy low-carb (HLC) or healthy low-fat (HLF) diet and examined weight loss and other measures over a 12-month period.

* The diet was genuinely low carb or low fat for 8 weeks – 20g of carbohydrate or fat daily on each diet – but then participants were advised to move to the level of carb or fat intake that they believed they could sustain for the year-long trial.

* The study researchers kindly provided access to all data from the study. I used it to ask a different research question – what can we learn from the people who did best (which I defined as a weight loss of 10% or more, of their starting body weight, at 12 months)?

* In this note, I describe my methodology and then present the results.

* A first interesting finding was that, of the 436 participants for whom 12-month data were available, 76 gained weight over the year, 4 stayed at the same weight and 356 lost weight. The weight gains thus impacted average weight loss results when all 609 people were averaged.

* I tested the hypothesis that people who stayed closest to their initial low carb or low fat levels would do best. It turned out to be a valid hypothesis.

* The hypothesis held across all groups: women who lost more than 10% of their body weight doing low carb maintained the lowest carb intake (in grams and as a % of calorie intake). Women who lost less than 10% of their body weight (or stayed the same weight) maintained the next lowest carb intake (in grams and as a % of calorie intake). Women who gained weight had the highest carb intake (in grams and as a % of calorie intake). This pattern held for men doing low carb and it held for men and women doing low fat (but with fat grams and fat as a % of calorie intake measured as opposed to carb intake).

* This analysis summarises the grams of carb and fat and the percentage of carb and fat as a proportion of total calories that the biggest losers consumed compared to those who gained weight.

Introduction

On January 25th, 2021, we looked at the latest publication from the DIETFITS trial (Ref 1). This acronym comes from “Diet Intervention Examining The Factors Interacting with Treatment Success.” This is the study that randomised 609 men and women to either a healthy low-carb (HLC) or healthy low-fat (HLF) diet and examined weight loss and other measures over a 12-month period.

I reported in that recent note that I approached the authors with a query and they very kindly shared all the original data from the trial with me in reply (Ref 2). I have had a chance to look at the data and it has been fascinating.

The publications from the trial so far have looked at the four groupings – men doing HLF, men doing HLC, women doing HLF and women doing HLC. This has given us insights into which diet might perform better (low carb or low fat) and which sex might perform better (men or women) and which combinations might perform better (men doing low carb vs women doing low fat, for example). Having recruited 609 men and women and having assigned them to HLC or HLF, this was the right approach. However, looking at the entirety of the raw data, there is another interesting approach – what can we learn from the people who did best? Instead of averaging everyone, is there a way of looking at the best (and perhaps worst) performers and learning from what they did? That’s what I set out to do for this week’s note.