Too much fat, sugar, flavour… or processing?

By Alistair Brown, Guest Blogger.  Alistair is currently studying for his MPH at University of Edinburgh and is on placement with Obesity Action Scotland as part of work on his dissertation.

Debate continues about what properties of certain foods raise our risk of overweight and obesity: is it fat, or carbs, or sugar, or a lack of fibre or protein, or hyper-palatability, or something else? Some people talk about junk food, though it’s not always clear what that means; or about processed food, but isn’t all the food we eat processed anyway?

After a rigorous new study by Kevin Hall’s team [1], it’s clearly important to consider processing. The study found that participants who ate a diet mostly comprising ‘ultra-processed foods’ ate around 500 calories more every day than participants who ate a diet mostly comprising unprocessed foods.

As we would expect from the increased calories, participants on the ultra-processed diet gained almost a kilo, while participants on the unprocessed diet lost the same. This weight fluctuation happened over only two weeks.

(An unprocessed breakfast served in the study: spinach omelette and sweet potato hash)

What happened in the study?

Twenty participants were randomised into two groups to stay as inpatients in a hospital ward for 4 weeks. For the first two weeks, half were fed with at least 80% of their foods classified as ‘ultra-processed’, and the other half with at least 80% of their foods ‘unprocessed’, according to the NOVA definition [2]. Diets in both groups were matched for calories, calorie-density, carbs, fat, protein, fibre, sugar and sodium – so any differences in the results could not be due to those factors. After the study, participants rated both diets equally for pleasantness and familiarity, so neither diet was more palatable than the other either.

Participants were given three meals per day and were allowed to eat as much or as little as they wanted, plus snacks. Participants also took part in equal amounts of low-level daily exercise.

(An ultra-processed breakfast: Honey Nut Cheerios, a muffin, and margarine)

After two weeks, the groups swapped around, and spent two weeks eating the other diet.

What did they find?

The headline result is that participants ate 500 calories per day extra on the ultra-processed diet – that’s 25% of an average women’s daily requirement, and 20% of an average man’s requirement, extra, every day.

Such overeating would be expected to cause weight-gain and in just two weeks it did exactly that, of almost a kilo on average:

What does this mean?

These results are astonishing. It may not seem much like a surprise to those who have always favoured “real food”. Others may have noticed that a unifying feature of successful diet trials, be they low fat, low carb, or something else, is their tendency to use unprocessed foods.

But here an attempt has been made to rigorously test that idea, and the result is enormous. The additional 500 calorie per day intake cannot be attributed to fat, sugar, carbs, calorie-density, fibre, protein, or palatability. So why did the participants eat so much more on the ultra-processed diet?

The graph above shows that participants on the ultra-processed diet ate much faster than those on the unprocessed diet. This is because the foods in the ultra-processed diet were around 85% more calorie-dense (calories per gram) than in the unprocessed diet. This suggests that people can consume more calories with every bite when eating ultra-processed foods.

It may also be that ultra-processed foods are quicker to chew and swallow, as some of the mechanical action of chewing has been done for us during processing. 

Another interesting finding was an increase in an appetite-suppressing hormone, and a decrease in an appetite-increasing hormone, while the participants were eating the unprocessed diet.

Could the study be flawed?

Every trial has limitations. Arguments could be made that this trial only involved twenty participants, all in their early thirties, so the results may not be generalizable to entire populations, or other age groups.

As this trial was conducted on a hospital metabolic ward, everything could be counted and measured exactly. So its limitations are less important than most trials that involve people living normal lives, who can’t be monitored and measured so precisely. Furthermore, the effect size (+500 kcal/day) is massive.

What does all of this imply for public health?

It is thought that overweight and obese adults in the UK over-consume an average of 200-300 calories per day [4]. This study suggests that that over-consumption could be eliminated by people favouring unprocessed foods rather than ultra-processed foods, no matter their fat, carb or sugar content.

If the effect of ultra-processed foods in the general population is similar to what was observed in this study, then it should be questioned whether ultra-processed foods should be marketed as healthy. It would also mean we need to work out how to help people choose unprocessed foods when they tend to be more expensive and less convenient and how could we lower those barriers.

[1] Hall et al. 2019 available from: https://www.cell.com/cell-metabolism/pdf/S1550-4131(19)30248-7.pdf

[2] Monteiro et al. 2018 available from: https://www.cambridge.org/core/journals/public-health-nutrition/article/un-decade-of-nutrition-the-nova-food-classification-and-the-trouble-with-ultraprocessing/2A9776922A28F8F757BDA32C3266AC2A/core-reader

[3] BBC News Ultra-processed foods 'make you eat more' May 2019: https://www.bbc.co.uk/news/health-48280772

[4] Public Health England ‘One You’ campaign: https://www.nhs.uk/oneyou/for-your-body/eat-better/keep-track-of-calories-400-600-600/