Why high fat diet studies on rats and mice are not to be trusted

By Jerome Burne

Over the past year I have been wondering whether there is something deeply flawed about research into the effects of high fat diets on rats and mice, done presumably to clarify the effects on humans. The rodent work consistently tells us that high fat diets make you fat and diabetic, while research on humans finds they do the opposite. What is going on?

The forty-year-old standard advice warning us off saturated fat is looking increasingly dated, even though experts in some quarters – looking at you Diabetes UK– still cling to it. But the world of the lab rat seems to exist in a parallel universe or in a time warp when low fat high carbohydrate (LFHC) diet was widely believed to be the way to lose weight and stay healthy.

So are the results telling us that the increasingly popular low carb high fat approach is wrong? That after all there’s no need for official bodies to perform a major U-turn? Not as far as I can tell. In fact it seems the rodent work is highly misleading. Not only are the so called ‘high fat diets’ they are fed nothing like the low carbohydrate diets any informed human would follow, but the animals have been selectively bred to ensure they become fat and diabetic on a high fat diet. This is not research, it is a rigged game.

Low carbohydrate diet helps humans lose weight

The confusion this fake research is (intentionally?) generating can be seen by comparing a couple of recent results from human studies with a couple from the labs. Here is a randomised controlled study published in 2014 in the Annals of Internal Medicine and involved 148 healthy men and women, who had their weight and various health biomarkers measured over a year.

Half were on a standard low fat high carbs (LFHC) diet; the others were on a HFLC diet. The conclusion was that: ‘The low carbohydrate (HFLC) diet was more effective for weight loss and cardiovascular risk factor reduction.’

Around the same time Yale University put out a press release on a rodent study with the headline: ‘Mother’s high fat (HFLC) diet alters metabolism in offspring leading to higher obesity risk’.

There was no attempt to suggest that the reaction of rodents might be different to that of humans – a basic caveat with any animal research. The assumption is that what’s bad for mouse mums is bad for humans as well.

The study found the diet damaged a part of the mouse brain (hypothalamus) that controls basic functions like sex drive, appetite and the way energy is handled. As a result: ‘the offspring remained overweight and had abnormalities in glucose metabolism throughout life. ‘This is another way of saying they were fat and diabetic.

High fat diet damages human foetuses

The release then goes on unhesitatingly to declare that the human foetus would be harmed by high fat in the same way. ‘The findings suggest that the third trimester of pregnancy in humans is the most critical period,’ the release continues. What’s more, say the authors, by avoiding high fat meals and ‘altering their food intake’ (presumably returning to a low fat diet), ‘mothers can control or even reverse their offspring’s disposition to obesity.’

There are a number of high quality studies just in the last year that strongly suggest that high levels of fats are not a problem for humans. One was also published in the Annals of Internal Medicine – a review of 44 trials involving over 500,000 people, looking at how much fat they ate and how much fat they had in their blood.

The conclusion: ‘Current evidence does not clearly support cardiovascular guidelines that encourage high consumption of polyunsaturated fatty acids and low consumption of total saturated fats.’ A good and detailed summary of the saturated fat debate can be found in an article entitled ‘Don’t Blame Fat ‘in Time magazine, June 23 2014.

Meanwhile studies on the risks of high fat are still being pumped out from the parallel rodent world. Here’s one from October 2014 from the University of California, Berkeley, which found that drinking grapefruit juice could cut the amount of weight gain caused by feeding mice a high fat diet.

The mice were fed a diet that contained either 60% fat or 10% for 100 days. This pushed up their weight and their glucose levels; grapefruit juice reduced both. Grapefruit juice, the researchers concluded, might help to cut diabetes risk.

Rodent high fat studies deeply misleading

Grapefruit juice may or may not help obese and diabetic humans; a good way to find out would be to give it to humans in a trial. So why does fat has this effect on rodents and does it tell us anything useful about humans? It seems unlikely since it flies in the face of dozens of other studies (see example from HIUK post.)

In fact they are deeply misleading and not simply because humans are different to mice. Cancer, for instance, is easy to cure in rats and mice but, as researchers are well aware, the results usually fail to translate to humans. However nutrition researchers such as the ones at Yale don’t seem to think that is even worth mentioning.

I have no idea if a responsibly designed high fat diet eaten during a human pregnancy is more likely to predispose a child to diabetes and obesity than a low fat one with lots of glucose generating carbohydrates, but it seem unlikely.

It’s now clear that rodent research which is being used as the basis for advice about how best to feed our children and stay healthy has been done on lab animals fed a high fat diet that is virtually guaranteed to make them fat. What’s more there is a widely used strain that have been bred over generation from animals that naturally put on weight in response to fat.

Rodent high fat diet studies not to be used to advise humans

The first witness to support this serious charge is an article on the site of the University of California Davis Health System uploaded in 2008. This says that although rodent studies provide: ‘the foundation for much of the belief that high fat diets are detrimental to health,’ most aren’t properly designed: ‘so we can’t use the results to give people recommendations on diet’.

The article goes on to identify what’s wrong. It reveals that high fat diet often consists of 60% lard, 20% sugar and 20% milk protein, which it described as ‘the mouse equivalent of pork rinds, ribs and coke’. In the trials this combination is compared with a natural low fat vegetable based high fibre diet known as chow. The problem is that the trials ignore other elements in each diet that have direct health effects.

The UC Davis article is referring to a paper published in Cell Metabolism the previous year, which goes on to explain that rodent chow also contains lots of soy which has an oestrogen-like effect on how much food and water the animals want, how active they are and how much fat they store. Meanwhile the sugar in the high fat diet is linked to weight gain and insulin resistance. Combining high fat with sugar is what makes human junk food moreish and damaging.

This dubious mixture is then fed to animals that are unlike any found in nature as they have been selected for fat sensitivity; one widely used variant puts on weight and develops raised insulin even when being fed zero carbohydrates!

The rat that is specially designed to get fat on fat

Evidence for this comes from the second witness Richard Feinman Professor of Biochemistry at State University of New York Downstate Medical Center who is a long-term advocate of the benefits of the HFLC diet and has written about it for HealthInsightUK. He naturally wanted to find an explanation for the dramatically different effects the diet had on rodents and humans.

In a paper for Nutrition Metabolism in 2012, he goes into detail about this ‘zero carbs’ mouse, known as the C57BL/6 (also as B/6 or Black six ) strain. It is used in the vast majority of mouse experiments and has been selected specifically to put on weight and raise glucose in response to a high fat diet. The only way a regular rodent high fat study could produce a surprise result using B/6 would be for weight and glucose levels to drop.

In fact it is possible to feed them in such a way that doesn’t result in damaging diabetic symptoms but it is tricky and involves a delicate balance. In one study putting B/6’s on a diet made up of 95% fat, 5% protein (very low) and no carbohydrates at all, got the same beneficial results found in humans on very low carbohydrate diets. These include weight loss with no cut in calories, increased energy, improved glucose tolerance and reduced insulin.

However when Feinman ran a study keeping B/6s on very high fat, zero carbohydrates but raising the protein proportion to a more normal 20%, all the benefits were lost and the animals went back to weight gain plus a rise in glucose and insulin. So the idea that these animals can tell us anything about what happens to humans eating a nutritionally informed high fat diet is a gross distortion of the research.

Make high fat diets on rodents transparent

Remarkably, these seriously misleading high fat rodent studies are relied on as evidence for the dangers of fat by experts who should know better, such as Jeremy Pearson, vascular biologist and associate medical director at the British Heart Foundation. In an article in the magazine New Scientist in July last year entitled ‘The truth about saturated fat’ he defended the benefits of the low fat diet, commenting that the idea that a diet high in saturated fat raises the risk of heart disease: ‘remains persuasive’.

In support of this he pointed to the ‘hard evidence from studies in animals where the dietary control is possible to a degree it is not in people. They repeatedly show that high saturated fat leads to high LDL cholesterol and hardened arteries.’

I appreciate that I have combined a number of different things in this post – studies showing that high fat causes diabetes related problems; research showing that high fat doesn’t lead to heart disease. And that I’ve used fat and saturated fat interchangeably. But this is not a scientific paper. Its intention is to point up a serious and avoidable source of confusion in an area already rife with confusion and misinformation.

A useful start to correct it would be for all studies to make clear what the high fat diet actually consisted of and what strain of specially selected rodent was being used and what it had been selected for. Until then, before swallowing any rodent high fat result, season it with a large pinch of salt.

With special thanks to Peter Dobromylskyj a vet with physiology degree, who runs a very sophisticated academic blog on diet and health at http://high-fat-nutrition.blogspot.co.uk/ His comments were invaluable in sorting out various questions I had about animal research but any errors are mine.

Jerome Burne

Jerome Burne

Jerome Burne is the editor of HealthInsightUK. He is an award-winning journalist who has been specialising in medicine and health for the last 10 years and now works mainly for the Daily Mail. His most recent book “The Hybrid Diet” was written with nutritionist Patrick Holford, published 2018. Award: 2015: Finalist for 'Blogger of the Year' Medical Journalists' Association.


  • How fat are the sewer rats that inhabit our cities, eating our waste (much of it fatty)? An investigation into their health might be revealing!

  • I once received a distressed note from a local scientist whose PhD had involved experiments to develop arterial damage in a rat model. She indicated that the high fat diet they gave the rats did not cause obesity unless the rats also received a high sugar intake in their drinking water supplemented I seem to think with condensed milk! But no one ever mentioned that inconvenient point when they published their papers. Perhaps there are other strains of rats that need carbohydrates to become fat but no one is drawing attention to those strains.

  • This is excellent. Whenever I read a rodent paper I always try to find the data on the diet; and I usually have to copy the code # into the supplier’s website to find it. This is not proper disclosure. The protein is always casein, the fat usually a mix of lard and soy oil (more unsaturated than saturated), and the carbohydrate a mix of sugar and a short-chain starch. Like a cheese cake recipe you’d buy in a box.
    Have you seen the guinea pig studies of low carb and CVD? Low carb protects guinea pigs from cholesterol overfeeding.

  • The systems analyst approach makes for a clearer understanding of obesity and its drivers. Some simple physics and the simplest endocrinology then complete the picture. And it is best to think first about conditions in the wild. As warm-blooded creatures mammals supply species that provide the best specific examples.

    Every animal has its energy budget. If finds enough food to match that energy budget all is well. But every animal faces competition, generally, either from others of its own kind or from other species. Every ecology must rest in a kind of balance pivoting about balance. Its a numbers game and demand for food cannot exceed supply.

    But is not unusual for supply to fluctuate on a seasonal basis. What happens then is that a creature may not find enough food to match its budget. A creature in this budget will begin to look to its own tissues to supplement the limited energy it is getting from food. Creatures with fatty reserves are better placed. Converting fat to energy is more efficient and less consequential than is converting protein. Food stress is the cue that permits fats to be released into the blood from fat cells. Think about this, if fats in the blood alone automatically became fatty tissue we’d be stuck with body-fat for ever.

    Seasonal food stress is worth preparing for. The grisly bear makes the best example. Water freezes where he lives for about four months of the year. Water and food is harder to find. So this chap finds a den and doesn’t come out for 120 days or so, not even to urinate or defecate. He’s hibernating. I cannot quote the energy budget for a hibernating grisly but he still has one. The annual budget is the order of two to three million calories at least, but there are only 240 days to hunt for food. In those 240 days our grisly has to average finding and eating at least one half as much again as he needs. The bears physiology drives hyperphagia (overfeeding). When it’s time to hibernate the bear has adequate condition (body fat in reserve) to last.

    The grisly has no calendar, and he hasn’t given that bright source of light in the sky a name. How does he know its time to bed down for winter? Hormones take care of it all, basically, and provide the cues that guide his actions. Light leverages cortisol levels on a diurnal and seasonal basis. Cortisol leverages insulin, and insulin leverages hyperphagia. Insulin is the co-factor that presses blood fats in to fat cells, and insulin leverages lipogenessis too. Between cortisol and insulin gluconeogenesis and lipogenesis ensure we can synthesise fats for storage if we never ate fats at all. Our physiology is much the same as the grislys, or any other mammals, his just operates to extreme to cope with extreme conditions. Without insulin fats in the blood will not be pressed into fat cells, but with higher concentrations of insulin in the blood they will, and they do not have to come from fats in food.

    Physiology encourages hyperphagia because when there’s competition for food and variation in supply it pays to have some energy set aside. Having some body-fat is evolutionary advantage. Environment and habitat tend to periodically enforce hypophagia.

    The modern human has been encouraged to eat less fat, and it lives under conditions that can disturb cortisol. The human can soon find itself with levels of cortisol and insulin that drive hyperphagia, and it cannot escape the vicious circle. The only way out is to get insulin levels down, The only macro-nutrient that will not leverage insulin levels is fat, because if it did we would never be able to make use of fatty reserves if we had to.

    Anybody that says the high fat diet does not work hasn’t managed to de-leverage insulin levels enough to become adequately hypophagic. It is not necessary to work off body-fat. The grisly does it sleeping and not raiding the fridge during the big freeze. We have less conscious control over how much we eat than we willingly think. Nature has tasked insulin with a significant level of control. Most other creatures need robust autonomic responses granted by nature, and only the human is daft enough to think it knows best. The researcher whose work is grant aided by a sponsor whose enterprises is invested in the dogma has no incentive to challenge the dogma.

    Makes sense?

    Researchers that experiment on lab rats are dumber than their subjects. Dogma drives their delusions. If the post-grads feeding the rats learned the truth it would be at odds with what they were taught as undergraduate students.

    • Bravo, Christopher!! What a brilliant and scholarly argument. Absolutely amazing. I learned something very valuable from your analysis. Congratulations on a job well done and well thought out!! No rodent brains at your house!! Thanks for the post!!!

    • Hi Chris,

      Great stuff, do you have a blog?

      Chris Adams

      • Thanks Jerome,

        This rodent vs human conflict has nagged away at the back of my mind, I had picked up on the dodgy chow but hadn’t realised quite how significant the “breeding programme” was/is.

  • “Bravo, Christopher!! What a brilliant and scholarly argument.”

    Well thank you, Mary, I am suitably fluffed and buffed by your praise. Sometimes the simplest arguments can be the most difficult to put across, so finding a route that makes it far easier for folks to follow is the elusive holy grail people trying to promote progressive arguments must pursue. Inspiration and perspiration, eh? And after the work-related week I’ve had it was refreshing to break into a smile and laugh in delight. I regret I had not proofed it to the extent I should.

    But let’s not forget that it was Jeromes excellent analysis in his post that inspired my response. Clearly I have been possessed of this ‘answer’ for a while, but simply having possession of answer doesn’t automatically mean a person could explain it well. Jeromes analysis was more questioning about the use and abuse of experimental evidence and that is an excellent line to take to undermine the claims, in contrast mine was a more contextual approach that cast aspersions. providing readers could follow the logic within its context.

    If you cast your mind back there was a convention given to Alzheimers and research. UK Prime Minister David Cameron pledged additional money (funded by taxpayers) for research into the degenerative condition. The trouble is that research funding tends to bring on bias, in that persons involved tend to seek treatment based solutions over cause oriented preventative measures. What was curious though was that in the run before that meeting there was a flurry of articles in newspapers and magazines discussing Alzheimers, though not necessarily in a truly balanced way.

    New Scientist ran a feature that fell far short of gold standards of science journalism and that was not a fair representation of the facts. The reporter had been spoon fed an unbalanced analysis, and the research and results that supposedly backed up the articles had neither been written up nor tested by peer review and publication. I had short email exchanges with the academic behind the research, and there were aspects that I did not find encouraging. The methods and analysis seemed primitive to me with scope for confounding factors.

    Research (the greater bulk of it) is firmly under the influence of the corporate agenda (which is to make money) and cause oriented prevention does not permit the marketing of treatment based protocols. Hence regard for the real roots of the cause of many a disease have slipped down the agenda of concern. The public are being conned, so too are politicians, and it isn’t easy to spot when it is arising, but in the days before the convention given to Alzheimers the propaganda was rank. Once sanity is in decline truly sane solutions will become far harder to deliver to a confused and bewildered public.

    The reasons for the way we humans carry on in this are explicable. ‘Stupidity’ plays a part as does ‘plain fat-headedness’, but there are other reasons. To discern them its worth thinking about human origins and the energetics of how we trended form primate to human, along with the drivers for sociological trends too. Both work and money can be scarce. Work is scarce because the labours of the few can meet the genuine needs of the many. Money is scarce as a little known feature of design attributes and the process of supply. Then asymmetric distribution of wealth and indebtedness seals the deal. Some people have to get busy and work at supplying other people with things they don’t really need and could well do without. Statins make for a most excellent example. A wonder drug to prevent onset of Alzheimers will next in line, unless we get wise and begin campaigning for sane solutions to problems we humans have brought upon ourselves. And that is one big big-up for the ethos behind the site Jeromne selflessly hosts.

    Although discussions hosted by Jerome may strike some followers as coming from left field, they are not so left field when given the requisite attention they deserve.

    We can take inspiration from Yoshimi, we haven’t got to let those pink robots defeat us. The battle will never be won, but just so long as enough of us keep fighting back the battle will not be lost either. Inspire as many people as you can, for numbers are our most potent weapon within the battle that is democracy in this half-assed state of affairs.

  • I am a proofreader for one of the world’s leading cancer research centres. I’ve been preparing documents prior to submission for publication for 18 years. There is very little interest in something as mundane as diet. HFD is given in conjunction with, for example, a carcinogen, to see the change in effect of carcinogens, tumour development, genetic mutations, treatment, etc. In other words, the diet is not what is being considered. And it won’t be. Standard rat diet (chow) has been around 70(?) years. You start questioning that, you start questioning 70 years of rodent model-based cancer and health research.

    • Editorial

      Hi HT what a fascinating comment. Thanks very much for posting. Would be very interesting to learn more about the obscure world of rat chow. It is indeed mundane but since animal work is a key part of the “evidence based” underpinning for mainstream dietary advice, details of its vagaries should be more widely known. Be delighted to discover more if you had the time to contact me via the web site.

  • The Medical Journal of Australia recently trumpeted headlines, ‘The Paleo Diets Kill’ neglecting to add ‘rodents.’ Professor Simpson, author of this nonsense gleefully garbled the syntax of this sentence in the press: “The Paleo people will be screaming and howling and they have a personal and professional interest in doing it,” says Professor Simpson the academic director of Sydney University’s Charles Perkins Centre.

    “One of the best ways to combat the lure of fad diets is to provide hard, experimental evidence,” he says.

    He should take his own advice. Not only was his comment above, less than academically articulate but it is dead wrong as others have noted here, but the climate in Australia on statins, cholesterol, CVD, obesity and diabetes is stuck in the tired old 30 year paradigm that has given us all of the above disorders. We just don’t seem to be able to embrace the truth: new idea resistance has become the latest course offered to medical trainees in Australia and they are excelling at it with Olympic virtuosity.

    Pete Evans, a celebrity chef and champion of a Paleo dietary pattern in Australia has been subjected to ridicule and rudeness that is almost unparalleled by ‘medical experts,’ especially for his new book which contains Paleo-style recipes and advice for prospective parents and children.

    The name calling, as we know, is the usual response used in an argument when people are losing it, and when their facts are, well, non-factual: but with perfect timing, the American Academy of Nutrition and Dietetics issued their new guidelines, vindicating Pete and throwing egg on the faces of his critics. Have the critics apologized to him? No, of course not. That would take integrity.

  • I disagree with the comments posted by “HT”. How in the world can diet be considered as mundane and not important in the arena of disease research? What about the vast body of literature (in both humans and animals) showing that diet clearly affects health and disease?

    “HT” is also incorrect in that diet won’t be considered by the scientific community. The literature proves otherwise. No doubt there is a subset of scientists who don’t consider diet important but that number is shrinking, as it should be. The sheer number of studies showing the effects of diet on disease argues against diet being ‘unimportant’.

    “HT” makes an argument that since chow diets have been around for 70 years, we shouldn’t question them. This has no basis in science. Science IS about questioning. This is how we refine and improve our understanding of the world around us. To be sure, no diet is perfect and this certainly includes chow. And yes, we should question past research. It either holds up, or it doesn’t. This is how we move forward.

    Science is also about understanding the tools used in experiments. While I’m sure spending 18 years as a proofreader gives “HT” a valuable experience, it does not make him/her an expert in diet research. To dismiss diet out of hand as unimportant flies in the face of thousands of published papers done by thousands of capable scientists. Furthermore, HT’s attitude slows our understanding of disease models, since it enables yet more studies to be done with poor diet choices or poorly matched control diets. Such studies (despite being published in high impact journals) are largely uninterpretable and unfortunately waste valuable scientific resources.

    We need to get past thinking that diet doesn’t matter and focus on doing good science in diet-induced disease research.

  • The article and nearly all the comments were well worth reading. I would point out that one matter not discussed here is the linoleic acid content of mouse chow. Rather than go on about this I suggest that curious readers Google the following:
    Gorilla heart disease obesity
    Mazuri Primate Growth & Repo Biscuit
    Joseph Hibbeln linoleic acid
    Frances Sladek linoleic acid
    Stephen Guyenet linoleic acid
    Bill Roberts linoleic acid
    Tom Brenna linoleic acid vegetarian diet
    William T. Neville linoleic acid
    David Gillespie linoleic acid
    David Brown linoleic acid

    • Editorial

      Yes interesting. Liked the gorilla biscuits link- presumably their nutritional needs are a bit closer to us. I’ve heard accounts of how veterinary approach to keeping animals healthy is much more likely to add in minerals and supplements but don’t have anything useful. Advice for us all to fill up with linoleic oils (polyunsaturated vegetable oils) showed up as not so good in recent paper on two trials that had “disappeared” for years and found that replacing saturated fat with polyunsaturated pushed up heart disease rates. http://www.bmj.com/content/346/bmj.e8707. Is that what the animal research found??

      • Here in the United states, besides gorilla obesity, we also have problems with feline, canine, equine, elephantine, and murine obesity, to mention a few. Elena Less of the Cleveland Zoo seems to be at the forefront of Gorilla Obesity research. If you Google – ADIPOSITY IN ZOO GORILLAS you can access her doctoral dissertation.

        Interestingly, on page 79 she wrote, ” It is well-known that consuming more calories than are being burned in a given day leads to weight gain. Diets that are high in saturated fat, simple sugars, and starch and lower in fiber are typically considered calorically dense and can cause a rapid accumulation of fat tissue (Astrup, 2008).”

        Less did not realize that the crude fat in primate biscuits is soybean oil which is 50 to 60 percent linoleic acid.

        Quote from page 92: “It may be that the fiber provided by the biscuits improved insulin sensitivity, but the paradoxical increase in cholesterol may be explained by saturated fat present in the biscuits (Medallion Labs analysis of Marion Leafeater® biscuits).”

        Note: the crude fat in the Marion Leafeater biscuits is soybean oil (stabilized). It’s possible that “stabilized meant partially hydrogenated in which case some of the linoleic acid and oleic acid would have been hydrogenated to stearic acid (which has no effect on cholesterol levels) while the remainder of the altered fat molecules would have become trans oleic and trans linoleic acids which do affect cholesterol levels.

        Quote from page 94: “Biscuits may be a healthy diet item as the only source of fiber, but the high levels of sugar, starch and saturated fat they contain could have deleterious health effects.”

        So you see, like most researchers, Dr. Less was programmed to blame saturated fats for the effects of excessive linoleic acid and added sugars intake. Hope this adequately answers your question.

Leave a Reply

WP-Backgrounds by InoPlugs Web Design and Juwelier Schönmann