Claim high-fat diets can prevent diabetes 'unproven'

"Diets laden with butter, cream and cheese 'can help combat surge in type 2 diabetes'," the Mail Online reports.

But the study it reports on only followed a small group of men for 12 weeks – not long enough to determine whether the diet would prevent diabetes or other chronic diseases.

The study involved 38 overweight to obese men randomised to one of two strictly controlled diets containing the same amount of calories, which the body "burns" to create energy.

In the first group, the energy mainly came from carbohydrates (53% of total calorie intake), while the energy mainly came from fats (73% of total calorie intake) in the second group.

Men in both groups lost weight and body fat after 12 weeks on the diets. There were only minor differences in certain blood sugar and cholesterol markers – nothing you could draw any conclusions from.

The main problem with this trial is it is so small, and it looked only at short-term effects.

You can't conclude anything from these results about the longer-term effects the diet would have on diabetes and cardiovascular disease.

An important fact not mentioned in the reporting of the study is both of the diets involved eating fewer calories than the men had previously consumed.

This reinforces the fact there is no magic bullet to weight loss – simply eat less and move more.

The study certainly doesn't give the green light to eating as much fat as you like. But healthy unsaturated fats should be included as part of a balanced diet.

Where did the story come from?

The study was carried out by researchers from the University of Bergen in Norway, and was published in the peer-reviewed American Journal of Clinical Nutrition.

It was funded by the Western Norway Regional Health Authority, Meltzerfondet, Bergen Medical Research Foundation, and the University of Bergen. Several companies provided products used in the study.

The Mail Online's reporting of the study was poor. In the final paragraph of the study, the researchers warn about trying to extrapolate their short-term results to longer-term effects on metabolic and cardiovascular disease risk. Yet this is precisely what the Mail has done.

It's very difficult to see how the Mail concluded a high-fat diet "can combat [the] surge in type 2 diabetes". The reduction in fasting blood sugar was in fact seen in the low-fat group, not the high-fat one.

And even then there was little difference seen between the groups for other markers – the study categorically did not find one of these diets was better than the other.

What kind of research was this?

This randomised controlled trial (RCT) aimed to investigate the theory that consuming fats or carbohydrates would have different effects on the amount of fat around body organs and on markers of metabolic syndrome. 

Metabolic syndrome is a collection of signs and symptoms – high blood pressure, high cholesterol, poor blood sugar control and obesity – that put a person at increased risk of cardiovascular disease and diabetes.

It's thought different dietary components could make some people more likely to develop the syndrome.

A randomised controlled trial is the best way of investigating the effects of an intervention.

But the difficulty with a randomised controlled trial assessing diet is that because of the practicalities of running the trials, they often include small numbers of people assessed on a short-term basis.

This means the results of such studies can't lead to sweeping conclusions on a population level.

What did the researchers do?

The trial recruited 46 overweight to obese men aged 30 to 50 with a body mass index (BMI) over 29 via a newspaper advertisement.

Researchers excluded men with serious illness and those taking regular medication or who had recently been trying to lose weight.

The men were randomised to follow 12 weeks of one of two diets:

  • very high-fat, low-carbohydrate (VHFLC) diet with 73% of energy from fat and 10% from carbohydrate
  • low-fat, high-carbohydrate (LFHC) diet with 30% of energy from fat and 53% from carbohydrate

The two diets provided identical daily energy intake (8,750 kJ/day), with 17% from protein. Both diets followed a low glycaemic index (GI) pattern, meaning they wouldn't cause a rapid rise in blood glucose levels after eating.

Both groups were told to consume more than 500g of fruit and vegetables a day and eat fish twice a week, and were given standardised supplies of butter, coconut oil and sugar substitutes.

Each group was also given recipe booklets for their specific diet and attended a course prior to the trial to ensure they understood the diet.

Each month the men were asked to keep five-day food records and weigh out their food on a daily basis.

The men were asked to keep their physical activity the same, were questioned about their ability to follow a diet strictly, and told about the importance of accuracy and honesty during the trial.

The researchers analysed blood samples to look at fat levels and blood sugar, examined the men's respiratory function, and used CT scanners to assess body composition.

What did they find?

After various drop-outs, only 38 of the original 46 men were available for analysis – just 18 in the LFHC group and 20 in the VHFLC group.

Body weight dropped by about 11-12kg, or 3.6 BMI points in both groups over the 12-week period.

Total abdominal fat and fat around the organs decreased by roughly 20-30% in both groups. Waist circumference decreased by 11-13cm. There were no significant differences between the groups.

Fasting blood sugar only reduced in the LFHC group, but there were no other between-group differences for other measures of blood sugar control, such as insulin.

Levels of one type of fat (triglycerides) decreased in both groups. Low-density ("bad") cholesterol decreased only in the LFHC group, but high-density ("good") cholesterol increased only in the VHFLC group.

Improvements were noted to occur within the first eight weeks in the VHFLC group, but were more gradual in the LFHC group.

What did the researchers conclude?

The researchers concluded that, "Consuming energy primarily as carbohydrate or fat for three months did not differentially influence [organ] fat and metabolic syndrome in a low-processed, lower-glycaemic dietary context.

"Our data do not support the idea that dietary fat per se promotes [body fat] and cardiometabolic syndrome in humans." 


This small trial aimed to see whether there is a difference between strictly controlled low-GI diets that contain the same amount of energy, but are either predominantly fat or carbohydrate based.

Overall, researchers found the diets caused both weight and fat reduction, with little difference between the two – with the exception of minor differences in certain blood sugar and cholesterol markers, the significance of which is difficult to interpret. These could just be down to chance.

The researchers were careful to control the diets and other lifestyle aspects to try to ensure any observed effects were only coming from the diets.

However, the trial had a couple of important limitations. For one, it was very small to start with, even before losing an extra eight to follow-up.

As the researchers acknowledged, the study may not have had sufficient numbers to reliably detect differences in outcomes between the groups.

The groups also included a specific group of overweight to obese men, so the effects in these people may not be comparable with other populations.

Most importantly, short-term measures of weight, blood sugar and body fat at three months tell you nothing about the possible longer-term effects.

This means you can't conclude anything about a person's risk of diabetes or cardiovascular disease in the longer term.

Little can be concluded from this relatively brief, small study. It certainly doesn't change our current understanding about diet and health.

The best way to lose weight or maintain a normal weight and reduce your risk of disease is to follow current healthy eating and exercise guidelines. You should aim to eat balanced amounts of carbohydrate, protein and fat.

NHS Attribution