EXPERT REACTION: Bread or butter? High carb diets may be worse than high fat

Embargoed until: Publicly released:

People with high carb diets are more likely to have poor health than people with high fat diets, according to an international study of over 135,000 people from 18 countries. The study authors say that diet advice to reduce fat is based on evidence from North America and Europe that may not apply to other countries. They found that on average, global diets consisted of over 60 per cent energy from carbs and 24 per cent from fats, but the best diets should have 50-55 per cent carbs and around 35 per cent fats.

Journal/conference: The Lancet

DOI: 10.1016/S0140-6736(17)32252-3

Organisation/s: McMaster University, Canada

Media Release

From: The Lancet

The Lancet: Replacing fat with high carbohydrate intake may be linked to worse health outcomes, according to study

Reducing total fat intake, and replacing it with a high intake of carbohydrates may be linked to worse health outcomes, according to an international study of diets, published in The Lancet.

The study involved more than 135000 people from 18 countries, and found that high fat diets (about 35% energy – including both saturated and unsaturated fats) were associated with a lower risk of mortality, whereas a high intake of carbohydrates (above 60% energy) was associated with a higher risk.

The study found that on average, globally, people’s diets consisted of over 60% energy from carbohydrates and 24% energy from fats, suggesting that rather than focusing on reducing fat intake in diets, guidelines should instead focus on reducing carbohydrate intake, particularly in low- and middle-income countries where carbohydrate intake was highest. Additionally, while current guidelines recommend reducing saturated fat intake to below 10%, the study found that very low intake of saturated fats (below 3%) was associated with a higher risk of mortality, compared to diets with a higher intake of saturated fats of up to 13%.

The study is being presented at the European Society of Cardiology Congress 2017.

“For the first time, our study provides a global look at the realities of people’s diets in many countries and gives a clearer picture of people’s fat and carbohydrate intake,” says lead author Dr Mahshid Dehghan, McMaster University, Canada. “The current focus on promoting low-fat diets ignores the fact that most people’s diets in low and middle income countries are very high in carbohydrates, which seem to be linked to worse health outcomes. In low- and middle-income countries, where diets sometimes consist of more than 65% of energy from carbohydrates, guidelines should refocus their attention towards reducing carbohydrate intake, instead of focusing on reducing fats. The best diets will include a balance of carbohydrates and fats – approximately 50-55% carbohydrates and around 35% total fat, including both saturated and unsaturated fats. Our study did not look at trans fats, typically from processed foods, and the evidence is clear that these are unhealthy.” [1]

Existing global guidelines recommend that 50-65% of a person’s daily calories come from carbohydrates, and less than 10% from saturated fats [2], but this is mostly based on evidence from North America and Europe, and it is unclear how applicable it is to other countries.

The study included 135335 people aged 35-70 years old from 18 countries across various regions including the Middle East, South America, Africa, China, North America and Europe, and South Asia [3] to look at the links between diet, cardiovascular disease and death.

Cardiovascular disease is a global epidemic, with 80% of the burden of disease in low-income and middle-income countries. Diet is one of the most important modifiable risk factors for cardiovascular disease and other non-communicable diseases.

At the start of the study, each participant provided information on their socioeconomic status, lifestyle behaviours, medical history, family history of cardiovascular disease, weight, height, waist and hip measurements, and blood pressure. They also completed a questionnaire on the types of foods and beverages they consumed, what size portion and how often, which the researchers used to calculate the amount of calories they derived from carbohydrate, fats and protein every day.

Participants’ completed follow-up visits with the research team at least every three years to record information on cardiovascular disease (ie, fatal and non-fatal heart attacks, stroke and heart failure) and death for 7.4 years. The team analysed rates of cardiovascular events (for example, heart attack, stroke, heart failure), non-cardiovascular deaths and deaths.

Globally, the average diet consisted of 61.2% carbohydrates, 23.5% fat (including 8% saturated fats, 8.1% monounsaturated fats, and 5.3% polyunsaturated fats), and 15.2% protein. However, this balance varied by region: on average, carbohydrate intake was highest in China (67%), South Asia (65.4%) and Africa (63.3%); total fat intake was highest in North America and Europe (30.5%), Middle East (30.3%), and Southeast Asia (29.2%).

High carbohydrate diets were common, with more than half of the people in the study eating a high carbohydrate diet (at least 60% of energy from carbohydrates), and about a quarter deriving 70% of their daily calories from carbohydrates. Half of the people involved derived less than 7% of their energy from saturated fats, and three-quarters (75%) ate less than 10% from saturated fats.

During the study, 5796 people died (including 1649 from cardiovascular disease and 3809 from non-cardiovascular disease) and 4784 had a major cardiovascular disease event (2143 heart attacks and 2234 strokes)

Diets high in carbohydrates (average of 77% energy from carbohydrates) were associated with a 28% higher risk of death, compared with low carbohydrate diets (46% energy from carbohydrates) [7.2 deaths per 1000 people years, compared with 4.1]. Rates of major cardiovascular events remained similar for low and high carbohydrate diets.

Comparatively, diets with high total fat intake (35.3% energy from fat) were associated with a 23% lower risk of death, compared with low fat diets (11.0% energy from fat) [4.1 deaths per 1000 people years in people, compared with 6.7]. Rates of major cardiovascular events remained similar for low and high fat diets.

The trends were similar when looking at saturated fats, with very low intake (below 3%) associated with a higher risk of mortality, compared to diets with a higher intake of saturated fats of up to 13%. Similarly, diets with very low intake of mono and polyunsaturated fats (3.6% and 2.2%) were associated with a higher risk of mortality, compared to diets with higher intakes (13% and 9.1%).

Saturated fats are typically found in animal fat products, like milk and meat, while monounsaturated and polyunsaturated fats are typically found in vegetable oils, olive oils, high fat fruits like olives and avocados, nuts and fish.

“Despite there being no association between low carbohydrate intake and health outcomes, this does not provide support for very low carbohydrate diets – less than 50% energy. A certain amount of carbohydrate is necessary to meet energy demands during physical activity and so moderate intakes, of around 50-55% of energy, are likely to be more appropriate than either very high or very low carbohydrate intakes.” explains Dr Dehghan. [1]

The authors note some limitations, including that the diet measures used were based on when the study began. Additionally, the study did not look at the specific types of foods the nutrients were derived from, but the will assess this in future analyses.

Writing in a linked Comment, Drs Christopher Ramsden and Anthony Domenichiello, National Institute on Aging, USA, says: “The relationships between diet, cardiovascular disease, and death are topics of major public health importance, and subjects of great controversy… The PURE study is an impressive undertaking that will contribute to public health for years to come. Initial PURE findings challenge conventional diet-disease tenets that are largely based on observational associations in European and North American populations, adding to the uncertainty about what constitutes a healthy diet. This uncertainty is likely to prevail until well-designed randomised controlled trials are done.”


The study was funded by Population Health Research Institute, the Canadian Institutes of Health Research (CIHR) , Heart and Stroke Foundation of Ontario, support from CIHR’s Strategy for Patient Oriented Research (SPOR) , through the Ontario SPOR Support Unit, as well as the Ontario Ministry of Health and Long- Term Care and through unrestricted grants from several pharmaceutical companies (with major contributions from Astra Zeneca [Canada], Sanofi-Aventis [France and Canada], Boehringer Ingelheim [Germany and Canada], Servier, and GlaxoSmithKline), and additional contributions from Novartis and King Pharma and from various national or local organisations in participating countries. It was conducted by researchers from McMaster University, St John’s National Academy of Health Sciences, Fuwai Hospital, Peking Union Medical College & Chinese Academy of Medical Sciences, Madras Diabetes Research Foundation, Aga Khan University, North-West University, Sahlgrenska Academy, University of Gothenburg, Health Action by People, Dante Pazzanese Institute of Cardiology, University of Zimbabwe, Estudios Clínicos Latinoamérica, Birzeit University, Simon Fraser University, Fundacion Oftalmologica de Santander- FOSCAL, Eternal Heart Care Centre and Research Institute, Isfahan University of Medical Sciences, Istanbul Medeniyet University, Universiti Teknologi MARA, Universidad de La Frontera, Wroclaw Medical University, University of the Western Cape, University of Ottawa, Independent University, Bangladesh, Dubai Medical University, Université Laval.

[1] Quote direct from author and cannot be found in the text of the Article.


[3] Development levels are based on gross national income per capita from the World Bank classification for 2006, when the study started. The countries included in the study are three high-income countries: Canada, Sweden, and United Arab Emirates, seven upper-mid-income: Argentina, Brazil, Chile, Malaysia, Poland, South Africa, Turkey, four low- and middle-income countries: China, Colombia, Iran, the occupied Palestinian territory, and four low-income countries: Bangladesh, India, Pakistan, and Zimbabwe.


  • The Lancet
    Web page
    Please link to the article in online versions of your report (the URL will go live after the embargo ends).
  • The Lancet
    Web page
    Please link to the article in online versions of your report (the URL will go live after the embargo ends).

Expert Reaction

These comments have been collated by the Science Media Centre to provide a variety of expert perspectives on this issue. Feel free to use these quotes in your stories. Views expressed are the personal opinions of the experts named. They do not represent the views of the SMC or any other organisation unless specifically stated.

Dr Alan Barclay is an accredited practicing dietitian and nutritionist, author and academic and a Research Associate at the University of Sydney.

This is an observational study which only shows associations and cannot prove causation. It is of relatively short duration for a study of this design, with only a median follow-up of 7.5 years.

Food intake was only assessed at base-line using a Food Frequency Questionnaire (FFQ) that was validated using unstated methods and no results of the validation studies are presented in the paper. We therefore do not know how well the FFQ assessed people's carbohydrate, fat (saturated, mono and polyunsaturated) and protein intakes in each country. Finally, the Hazard Ratios are modest for the extreme quintiles  (Q1 and Q5).

From Figure 1, we can see that total mortality was lowest for carbohydrate intakes between approximately 45 per cent of energy and 60 per cent of energy. With respect to saturated fats, risk of mortality was highest at very low intakes and started to increase when more than 6 per cent of energy was consumed from this nutrient.

Overall, the conclusions of the paper are overstated – a major overhaul of existing dietary guidelines is not warranted based on this additional evidence.

It is important to put the results of the Dehghan, et al. paper in to an Australian context: based on our most recent national health survey, Australian’s consumed on average just over 43 per cent of energy from carbohydrate and 11.5 per cent of energy from saturated fat. On face value, this means that Australians should be consuming more carbohydrate and less saturated fat – consistent with our current dietary guidelines.

Last updated: 16 Aug 2018 11:56am
Declared conflicts of interest:
None declared.
Professor Jennie Brand-Miller is Professor of Human Nutrition in the School of Life and Environmental Sciences and Charles Perkins Centre at the University of Sydney

Australians are ahead of the curve. We recognised that carbohydrates were not created equal over 3 decades ago – some were harmful because they increased fluctuations in blood glucose (ie they had a high GI). Since then we have reduced our intake of added sugars and as well as high GI starches.  Australians and the Australian food industry should be congratulated.

Last updated: 29 Aug 2017 2:58pm
Professor Amanda Lee is a professor within the School of Public Health at Queensland University

The Prospective Urban Rural Epidemiology (PURE) study is a very large prospective observational study that assesses association of intake of dietary components (estimated by food frequency questionnaire) with health outcomes, in more than 135,000 people from 18 countries. This paper reports initial associations, mainly in low and middle-income countries, between macronutrients (with a focus on carbohydrate, total and different types of fat) and cardiovascular disease and mortality.
The findings that higher intakes of fats including saturated fatty acids, monounsaturated fatty acids, and total polyunsaturated fatty acids, and also animal protein, were associated with lower mortality, whereas carbohydrate intake was associated with increased mortality make an important contribution to nutrition.
However, the types and food sources of carbohydrate are not reported in this paper. Yet the PURE study has shown previously an association between higher intakes of fruit, legumes and vegetables (that provide carbohydrates) and increased mortality, suggesting that it is mainly carbohydrate from added sugars and refined grains that may be problematic. Further, the food sources of fats have also not been reported or controlled for in this paper.

An explanation of the findings could be that, in the early stages of the nutrition-transition to more western diets, animal products may help increase life expectancy in low and middle income countries, as they are a rich source of micronutrients that can be lacking in many of the countries in the PURE study.
In addition, the results may not be generalisable to Australia, as the upper levels of intakes of carbohydrate reported in the study are much higher and the lower intakes of fats are very much lower than consumed here.
It would be useful to see greater control of cultural, social, economic and other confounding variables, and more detailed analysis, including adoption of a food-based approach, in future reports of the PURE study.

Last updated: 29 Aug 2017 2:56pm
Professor John Funder is a Distinguished Scientist at the Hudson Institute of Medical Research and a Professor in the Department of Medicine at Monash University

This is a good study, across a range of countries with high, mid and low average income populations. What it shows is that fats - saturated, mono-unsaturated, polyunsaturated - are not the no-no we have all been brought up to believe in the context of high carbohydrate diets when the source of the latter is refined sources such as sugar, rice etc.

Highly significantly, lower levels of various cardiovascular morbidity and mortality were seen between diets with up to 35 per cent of calories from fat over those low in fat and high in carbohydrate. Complex carbohydrates - as in fruit and vegetables - are probably another thing, and not similarly a problem.

So go for dairy, olive oil and even the occasional wagyu beef burger, have lots of grains, fruit and vegetables, and lay off the sweet stuff - especially the empty calories in the 16 teaspoonfuls of trouble in sugar-sweetened soft drinks. Sounds a bit like the Mediterranean Diet, with wagyu rather then meat sauce.

Last updated: 29 Aug 2017 2:53pm
Mr Bill Shrapnel is a nutritionist and Director of Shrapnel Nutrition Consulting Pty Ltd

This study provides further evidence that high carbohydrate diets are not the preferred model for healthy diets. Although this evidence has been accruing for years it was largely ignored by the National Health and Medical Research Council during the development of the latest Australian Dietary Guidelines. 

The NHMRC chose not to review the Nutrient Reference Value for total carbohydrate intake, or the evidence for glycaemic load and chronic disease risk, or the evidence on the associations between dietary saturated fat, carbohydrate and coronary heart disease, all of which would have shed light on the issue.

During the development of the Guidelines several health agencies wrote to the NHMRC arguing that advice to replace saturated fat with carbohydrate to lower coronary disease risk was no longer evidence-based, but this was ignored. 

Consequently, the latest Australian Dietary Guidelines were 10 years out-of-date when they were published.

Last updated: 29 Aug 2017 2:51pm
Emeritus Professor Mark Wahlqvist is former Professor and Head of Medicine at Prince Henry’s Hospital and Monash Medical Centre, Associate Dean (International Health and Development) and Director of the Asia Pacific Health and Nutrition Centre

The controversy that this study inevitably accentuates will not be resolved by such macronutrient-based studies.

Food pattern, food habit and socio-ecological studies already provide us with the advice we need.That is to support biodiversity and access to public open space so that we can increase our prospects of plant-based minimally processed, biodiverse, affordable diets and be physically active so that we can eat enough without being overfat.

Last updated: 29 Aug 2017 2:48pm
Professor Grant Schofield, director, Human Potential Centre, Auckland University of Technology (AUT)

There is a sea-change in establishment thinking around the diet-heart hypothesis, and what constitutes a healthy diet. The long awaited PURE study results, follow a favourable review of best-selling book “The Big Fat Surprise” also in the The Lancet(last week, and 2 years after publication) summarising the evidence for where we got it wrong in demonising fat.

PURE found that lower intakes of fat and saturated fat (SFA) were associated with higher mortality, and had no benefit in terms of cardiovascular mortality. Importantly, this study refutes the claim often made that SFA variations in epidemiology just aren’t low enough to see the benefit of lower SFA.

In the PURE study, there are low intakes of SFA, and plainly we not only see no evidence of benefit, we see harm in increased all-cause mortality.

If there were such harmful effects of saturated fat, you’d expect to see some evidence in studies of these sorts.

PURE tells us, along with mounting other evidence, that it is finally time to move on from banishing dietary fat, including saturated fat.

I’d concentrate on getting people to eat less highly processed and refined foods, especially sugar and highly refined carbohydrates. The totality of the evidence says to 'eat foods low in human interference'. If our food was clearly recently alive in nature, that’s a good start to a healthy diet.

Last updated: 29 Aug 2017 10:12am
Professor Jim Mann, director, Edgar Diabetes and Obesity Research, University of Otago

This study presented at the European Society of Cardiology Congress and published in The Lancet suggests that a high carbohydrate intake is associated with a higher risk of total mortality, whereas total fat and individual types of fat were related to lower total mortality.  

This observation together with the findings that total fat and types of fat were not associated with cardiovascular disease and saturated fat was inversely associated with stroke lead the authors to suggest that 'removing current restrictions on fat intake but limiting carbohydrate intake (when high) might improve health'. They conclude that global dietary guidelines should be reconsidered in the light of their findings.  

Given that their findings are based on observations made on over 135,000 people in 18 countries one might assume that such conclusions are definitive. There are however major limitations to these conclusions some of which are acknowledged by the authors. The limitations apply to countries like New Zealand as well as to countries which traditionally have a high carbohydrate intake such as China, which the authors suggest may particularly benefit from their recommendations.

It is important to consider this study in the context of a large body of evidence regarding nutrition and health, and not consider the results of this single study in isolation. Importantly the strengths and limitations of each study must be considered. Important strengths of this study are the large sample size, and inclusion of populations from a wide variety of countries and regions throughout the world. The pooling together of such diverse populations with diverse patterns of lifestyle and dietary patterns poses some challenges however with respect to interpretation of the results.

A key limitation is that there is no distinction between carbohydrates which have been repeatedly shown to be detrimental to health (e.g. free sugars such as table sugar, refined grains) and those which have been clearly shown to have health benefits (e.g. fibre-rich wholegrains, legumes, vegetables and fruits).  

Other very large cohort studies have shown that there are health benefits when saturated fat is replaced either by polyunsaturated fat or wholegrains but not when replaced by sugars or refined grains.

A major difficulty in interpretation results from the fact that the most striking effects were seen when comparing extreme levels of intake. For example, the risk of death was 28% higher among those with diets high in carbohydrate than in those with the lowest intakes. However, those with the highest intake were obtaining 77% of energy from carbohydrate and those with lowest intake 46% energy from carbohydrate. 

In New Zealand, current intakes are not appreciably different from those in the low intake category. The 'benefits' of fat are similarly mainly apparent when comparing extreme levels of intake.

Carbohydrate intakes are highest amongst the predominantly rice-eating countries including China and countries in South Asia (carbohydrates providing 65 to 68% total calories) and it is these countries which the authors suggest might be particularly at risk from their high carbohydrate intakes.  

However, in the largest cities in China fat intake has increased appreciably at the expense of carbohydrate consumption and rates of obesity, diabetes and cardiovascular disease have increased. In China, this new dietary trend which appears to be compatible with the recommendations of the authors is also associated with increasing cholesterol levels. There would seem to be considerable risk associated with offering recommendations which are in conflict with traditional dietary patterns.  

Japan was not represented in this study but it is noteworthy that in that country where rice is a staple food life expectancy is the greatest in the world.

National and international dietary guidelines are increasingly emphasising diet quality and that a wide range of macronutrient intakes can contribute to a diet associated with positive health benefits. Recommending the optimal sources of carbohydrate and fat is more important than precise amounts.  

Current guidelines which we endorse recommend that people continue to eat a diet that is rich in vegetables and fruit, legumes, pulses, nuts, wholegrains, and vegetable oils. Importantly, people should limit the amount of free sugars, salt and highly processed food.  A range of dietary patterns, including Mediterranean, Asian style and other traditional dietary patterns can be consistent with this approach.

Last updated: 29 Aug 2017 10:09am

News for:


Media contact details for this story are only visible to registered journalists.