select search filters
roundups & rapid reactions
factsheets & briefing notes
before the headlines
Fiona fox's blog

expert reaction to study looking at low-carb and low-fat diets and risk of death in a US cohort

Research, published in JAMA Internal Medicine, reports on the connection between low-fat and low-carb diets and risk of death.


Prof Nita Forouhi, MRC Epidemiology Unit, University of Cambridge, said:

“Despite the challenges and limitations of nutritional research that apply to this study and more broadly, it is a positive step forward that this research assessed, in a single analysis, both the quantity and quality of each of the nutrients – carbohydrate, fat and protein.  This is useful in contrast to many research studies that report on one nutrient at a time, and focus on quantity alone or quality alone.

“This study did not directly evaluate food sources of each of the nutrients, but could indirectly help to clarify that it is not enough to focus on the nutrient alone in messages of either low-carbohydrate or low-fat diets, but that the type, quality and source of the nutrient is important.

“The reductions or increases in mortality with unhealthy or healthy low-carbohydrate and low-fat diets were of a modest degree which some may think is too small to focus efforts on.  But, given that dietary habits are readily modifiable and together with other lifestyle behaviours such as physical activity, smoking and alcohol intake, there is much to play for in implementing these findings for people who may choose to follow these dietary regimes.  This should be placed in context of the added consideration that dietary advice should move away from a focus on nutrients to a focus on foods and dietary patterns.

“Critics may argue that this is not clinical trial evidence but observational, yet doing clinical trials of diets for long-term outcomes is not realistic, and this study made efforts to account for some of the limitations, and was open in discussing other limitations.  The study was large in size, included men and women of different ethnic groups, and applied modern advanced statistical approaches including accounting for confounding factors, doing sensitivity analyses and checking for potential reverse causality.  The study findings can help to inform the conduct of future clinical trials, at least for intermediate health points such as weight, blood fats or glucose and other metabolic factors.”


Prof Gunter Kuhnle, Professor of Nutrition and Food Science, University of Reading, said:

“This is a very well conducted observational study.  The authors used data from a representative survey of the US population, and the results therefore apply to the general public – at least in the US.  However, their definition of ‘low carb’ is quite unusual, in that it isn’t what we would think of as low.

“The results make very interesting reading: if the associations shown here are causal, it would suggest that when following a healthy dietary pattern, it does not really matter whether the diet is lower in carbs or lower in fat.

“There is currently some ongoing public debate about whether current recommendations are causing harm and should be replaced with recommendations to follow a low-carb diet – it should be noted that current dietary guidelines recommend having 35% of energy intake from fat; that is not low fat, it is moderate fat.  This discussion is often vociferous and often uses inappropriate generalisation, such as likening all carbohydrates with sugar.  However, this study is unlikely to resolve this argument, because the amount of carbohydrate consumed in the low-carb group was still quite high: indeed, the cut-off point of 47% is very close to current recommendations, and is much higher than what would be recommended for a low carb diet.  Conversely, the fat intake of those classed as following a low carbohydrate diet was on average in line with current UK recommendations (35% of total energy) – that is not a high-fat diet.

“In summary, the study supports current advice to follow a healthy, balanced diet rich in high quality carbohydrates (in this study: whole grains, whole fruit, legumes, and non-starchy vegetables) and unsaturated fatty acids.  Interestingly however, while consuming a healthy diet reduced overall mortality, following an ‘unhealthy’ diet with low quality carbohydrates (in this study: refined grains, added sugar, fruit juice, potato, and other starchy vegetables) showed no significant association with premature mortality.”


Dr Ian Johnson, Nutrition researcher and Emeritus Fellow, Quadram Institute Bioscience, said:

“This is a large and well conducted observational investigation but it is difficult to see a very clear or novel message for consumers in this paper, other than that moderate variations in the proportions of fat and carbohydrate in our diets, within the range selected freely by an American population, were not associated with big differences in long-term health or risk of premature death.  This is perhaps not very surprising, given the remarkable adaptability of humans to different types of diet.

“What does seem to matter is the quality of those carbohydrates and fats, where quality has been defined according to some pretty conventional criteria.  High-quality carbohydrates for example are lightly refined whole grain cereals, fruits and non-starchy vegetables, and high-quality fats would contain a higher proportion of polyunsaturated lipids from plant sources, and less fat from animal sources.

“Overall, the results of this study seem consistent with the current dietary recommendations given by public health authorities in the UK and elsewhere.”


Dr Duane Mellor, Registered Dietitian and Senior Teaching Fellow, Aston Medical School, Aston University, said:

“This is another population study that looks at the ongoing question of whether low fat or low carbohydrate diet is best.  However, as the people in the study were not asked if they were trying to eat ‘low carb’ or ‘low fat’, there is no evidence of intent, which is important when looking at the effect of any diet or way of eating.   This is because how a person plans and attempts to eat can ultimately influence how healthy and varied their diet is.

“One limitations of this study is that it only looks at food intake over a limited time – although collected in a robust way, a single day’s food intake may not represent their average food intake.  However, probably the biggest problem is that the definitions used for low fat and especially low carbohydrate are very different to what many followers of this type of diet would describe as ‘low carb’.  This is because the population was simply divided into five groups and the lowest intake group although called ‘low-carb’ in fact seemed to consume more than twice the amount of carbohydrate most ‘low carbers’ would be trying to eat.

“The simplistic nature of labels such as ‘low carbohydrate’ or ‘low fat’ is highlighted well in this study, as it showed what really seemed to matter when it comes to reducing risk of early death is the overall quality of a person’s diet.  This is the key message from this study – whatever type of diet you choose, from Mediterranean to vegan to ‘keto’ to low-fat to Paleo, it is not the label your diet ‘wears’ that counts, it is the fact it needs to include a variety of mostly healthy foods.

“What we know in general from the wealth of evidence on diet is that a healthy dietary pattern tends to be minimally processed and typically based on plenty of vegetables, fruits, nuts, seeds, pulses with some healthy fats, dairy and meat, fish and poultry according to your tastes and preferences.

“So as consumers, as these types of diets such as low carbohydrate become more popular we need to be aware there may be more healthy and less healthy foods which are technically ‘low carb’, as may be the case with a number of low fat foods.”


Dr Rosie Cornish, Senior Research Associate, Population Health Sciences, University of Bristol, said:

“The researchers found that what might be regarded as ‘healthier’ diets (higher proportions of whole grains, fruit, legumes, non-starchy vegetables, and nuts, and lower proportions of added sugar, refined grains, fruit juice, animal protein, and saturated fats) were associated with lower death rates.

“There are certain to be many differences between people with healthier and less healthy diets (some of these differences are indeed illustrated in the paper) and these differences could be the actual reason for the differences in death rates.  The researchers did take account of many of the key differences: smoking, physical activity, body mass index (BMI), alcohol intake, and others.  However, in an observational study like this, you cannot ever be sure that you have taken account of all the differences.  In this research this could be because things like smoking, BMI, physical activity and so on might have changed over time but were measured only once, or because there might be things that were not measured at all.

“So, the differences in death rates could be due to differences in diets or they could be due to other factors.”


‘Association of low-carbohydrate and low-fat diets with mortality among US adults’ by Zhilei Shan et al. was published in JAMA Internal Medicine at 16:00 UK time on Tuesday 21 January 2020.

DOI: 10.1001/jamainternmed.2019.6980


Declared interests

Prof Nita Forouhi: “I am a member of the Joint SACN/NHS-England/Diabetes-UK Working Group on ‘lower carbohydrate diets compared to current government advice for adults with type 2 diabetes’.  Views expressed are my own, not the Group’s.”

Prof Gunter Kuhnle: “I don’t have any interest to declare.”

Dr Ian Johnson: “No conflicts of interest.”

Dr Duane Mellor: “I am a BDA spokesperson.”

Dr Rosie Cornish: “No conflicts.”

in this section

filter RoundUps by year

search by tag