Gary Taubes is probably the most single-minded person I have ever met. In 2002, when he was a little-known science journalist and author of two books on scientific controversies, an article of his was published in the New York Times, headlined: What If It’s All Been a Big Fat Lie? In it, he argued that the low-fat dietary advice of the previous couple of decades wasn’t only incorrect, but actively dangerous and the reason for, as he put it, the “rampaging epidemic of obesity in America”. For Taubes, dietary fat wasn’t a problem at all. Instead, the real danger was carbohydrate, he asserted, sparking a backlash, and fuelling the ongoing conversation about what constitutes a “healthy diet”. He wasn’t the first to assert that carbs were bad (Robert Atkins got there before him), but perhaps because of his serious and scientific background – he has a physics degree from Harvard and studied aerospace engineering at Stanford – he has been a polarising figure, with as many ardent followers as detractors.
Since 2007, Taubes has published five books on sugar, fat and carbohydrate, including his latest, Rethinking Diabetes, in which he posits that low-carb diets have been under-used as a way to manage blood glucose in type 1 and type 2 diabetes, in favour of drug-heavy treatment regimes which, he suspects, may do more harm than good.
His writing on nutrition has won several awards, notably from the US National Association of Science Writers, but it has also been sharply criticised, mainly for his almost evangelical attachment to the keto diet, in which you eat so little carbohydrate (50g or less per day) that the body goes into a state called ketosis, meaning you stop burning stored glucose and start burning fat instead. In 2021, he published The Case for Keto, a self-help book, after which, as he says, he went “from being a respected source of information to somebody who may indeed be a crank after all”. Taubes, who follows the diet himself, is now proposing it should be offered to people living with diabetes. (Keto diets were originally developed as a way to treat certain types of childhood epilepsy.) For Taubes, keto means he doesn’t “eat starches, grains or sweets, and I don’t eat breakfast because I think better in the morning without it. When I snack, it’s nuts or good cheese. If we were to go to dinner together, I’d order a piece of fish or half a roast chicken and ask the waiter to hold the rice or potatoes and give me a green salad or green vegetables instead.”
When we talk, it’s a bright morning in Berkeley, California, where Taubes, 67, lives with his wife and sons. With his open-neck shirt, tan, salt-and-pepper short hair and slightly drawling delivery, he seems more like a professorial Owen Wilson than someone seeking to radically alter how diabetes is understood and treated.
The majority of his latest book is an exhaustive retelling of the history of diabetes research and how, in the first half of the 20th century, it went – as Taubes sees it – wrong, with the emergence of a treatment doctrine that mistakenly allowed people living with diabetes to eat whatever they wanted, all the while using insulin and drugs, such as metformin, to manage the blood-glucose consequences. Taubes has always been fascinated by bad science and for him this was bad science of the highest order, because the regime was based on dietary hypotheses which he says had not – and still have not – been rigorously tested.
Before the discovery of insulin in the 1920s, diet was the only way to manage diabetes and although various options were tried by early practitioners, low-carb was, says Taubes, among the most popular (with medics, at least). Insulin was a gamechanger. Not only did it almost magically save the lives of children with type 1 diabetes, who would often arrive at hospital comatose and die swiftly afterwards, but it also meant that people with diabetes of both types could eat a more or less normal diet.
Another example, for Taubes, of how early researchers were mistaken, concerns the differences between type 1 and type 2 diabetes, which are so different they almost shouldn’t share a name. Type 1 diabetes is an autoimmune disease in which the body’s immune system attacks and destroys the cells in the pancreas that produce insulin, the hormone which regulates the level of glucose in our blood; people living with type 1 need insulin injections or an insulin pump, to survive. Like type 2, type 1 can cause complications such as heart, kidney or eye disease, and nerve damage.
Type 2 diabetes accounts for about 90% of cases, and is a metabolic disorder in which the body either can’t make or can’t use insulin (AKA insulin resistance) to metabolise glucose, leading to persistently high levels in the blood. Ultimately, people living with type 2 diabetes may need insulin and other diabetes medications, too, but for a lot of people, diet and lifestyle modifications can defer that need. Many, but not all, specialists think there is often a causal relationship between weight and type 2 diabetes, which has led to a high level of stigma around the diagnosis. Diagnoses of both are rising globally – five million people live with diabetes in the UK.
What Taubes would like to see is low-carb diets being offered alongside or instead of diabetes medications. “When insulin therapy started in the 1920s, they had no idea what the long-term side-effects were or what the long-term consequences of living with diabetes were [because most people with type 1 died],” he says. “Then doctors find out that it’s just easier to let patients eat whatever they want and give them drugs to cover them. Then it’s another five, 10 or 20 years before they start seeing the long-term complications, which they think of as long-term complications of the disease.” What he wishes scientists at the time had concluded was: “The reason we’re keeping them alive is insulin therapy. So what we’re seeing is the long-term complications of the disease as controlled by insulin therapy, and the insulin therapy might be causing the complications as much as the disease is.
“By the late 1930s, you have this tidal wave of diabetic complications: the heart disease, the atherosclerosis, the neuropathy, the kidney failure, the blindness, amputations. And nobody ties it back.” By then, the low-carb diet had fallen far from favour. “Nobody wants to eat a diet. So nobody’s being told: ‘Look, if I give you insulin, I’m going to keep you alive until you’re 30, especially if I give you a lot of insulin and you do eat your carbs. But if I tell you not to eat the carbs and we minimise the insulin use – which for type 2 could be no insulin – I might keep you alive as long as anyone else in your family.’”
In the book, which is laden with references, studies and dense historical detail, Taubes mentions case records from the 1700s in which patients on low-sugar diets beg for a medical solution, suggesting that the preference for medication over a highly prescriptive diet has been with us for a long time. “If you’re told, a pill or a diet, we all want the pill. But if you’re told a pill or a diet and the diet will keep you healthy and the pill will give you a chronic degenerative disorder where you’re still going to have these horrible complications, they are just going to be 20 to 30 years later… the pill is going to be easier, because it always is. But if you change the diet, it’s not a hypothetical change: you can put your diabetes into remission, you can stop taking these medications.”
Convincing as this sounds, there are some apparent flaws in Taubes’s arguments, which are by no means widely accepted in the academic or medical communities. Professor Roy Taylor is a leading British diabetes researcher. “When a subgroup of the UK Government Scientific Advisory Committee on Nutrition was convened in 2021 to look at low-carbohydrate as an approach to diabetes in general, the literature was very thoroughly assessed and I was part of that panel. Very low-carbohydrate diets had no better results than the modest reduction of carbohydrates,” he says. Other studies, such as one in 2022 at Stanford that compared low-carb diets and the Mediterranean diet, have shown that while they both work when it comes to controlling blood glucose, the Mediterranean diet is easier to stick to.
Second, low-carb diets are now offered as one way of managing diabetes of both types, the pendulum swinging back in their direction after almost a century, possibly more so here in the UK than in the US (Taubes doesn’t include contemporary case histories in the book). Two members of my own family have been put on a very low-carbohydrate diet in recent years when they were deemed at risk of developing type 2 diabetes in their 70s (a risk both of them reversed).
Jack Leeson, 55, was diagnosed with type 2 diabetes six years ago and on the advice of his NHS doctor, radically altered his diet. “She put me on the diabetes drug metformin and told me about people who lose limbs with it. So I was very motivated.” She didn’t suggest keto, “but she made clear the volume of sugars in bread and pasta, and supposedly healthy things like fruit juice, which is just sugar. I gave them up. I didn’t replace them with fat, just more protein, lots of vegetables, berries, soya milk and yoghurt, beans and lentils and pasta made from Japanese konjac root.” Leeson also does an hour of aerobic exercise every day. “I binned off the diabetes, cholesterol and blood pressure issues in 18 months and lost about 5st.”
Diabetes UK, the biggest diabetes charity in the UK and diabetes.co.uk, an online support community, have information on their sites about low-carb diets, particularly for people with type 2 diabetes. Diabetes UK states that “there is no consistent evidence that a low-carb diet is any more effective than other approaches in the long term. So it shouldn’t be seen as the diet for everyone… At the moment, there is no strong evidence to say that a low-carb diet is safe or effective for people with type 1 diabetes. Because of this, we do not recommend low-carb diets to people with type 1 diabetes.” What Taubes would probably add is that there isn’t much evidence that they’re unsafe either, because low-carb diets haven’t been studied intensively either within or beyond the diabetes research community. But diet is incredibly difficult to study – especially in a context like diabetes where subjects are also often medicated. One of the tropes of nutritional science is that drawing long-term health conclusions from what people eat is nigh on impossible, because diet interacts with lifestyle, because people lie, intentionally or not, about what they eat, and because longitudinal studies of diet are so expensive.
Munjeeta Sohal, 39, was diagnosed with type 1 diabetes as a teenager and has mixed feelings about the idea it could or should be managed through carb restriction. “A low-carb diet might be a good way to control blood sugars,” she says, “but I am now on an insulin pump system that allows me complete freedom over what I eat. If I eat less carbs, my insulin stays more in range. I see the impact it has on my blood sugars, but that isn’t enough to make me want to do it full-time. My blood sugars are finally, thanks to the technology that’s available, in range between 70% and 90% of the time on an average day, and that is brilliant. I don’t need to eat low-carb for this to be the case.” Having diabetes is also a risk factor for eating disorders and Sohal’s relationship with food veered towards unhealthy when she was first diagnosed as a teen. Like many others she worries that “asking people to restrict may, for some, lead to secret bingeing, or guilt and shame around enjoying food.”
“They’re right, of course, to worry,” says Taubes. “But if diabetes, like obesity, is triggered in susceptible individuals by the carbohydrate content of the diet and can be put into remission by avoiding carbohydrate-rich foods, what would you tell patients?”
Another complication is that where Taubes is able to look at diabetes – indeed at diet as a whole – through a single, high-fat-low-carb lens, few others can. “I eat the same thing every day,” he says. “As long as I like it, I will continue to like it and be happy to eat it.” I suspect this makes it hard for him to understand why many of us have such a complicated relationship with cake.
While Taubes himself has stuck to keto for the last two decades, the rest of us might find it tough to follow. As Professor Taylor says: “The fall-off, in keto, is quite high. People have families and friends, and eating is part of social interaction.” Even in a highly motivated group, like people trying to control diabetes, adherence to low-carb is pretty patchy: a 2022 paper tracked this low adherence and pointed to cultural, religious and – perhaps most important – economic barriers. Keto can be expensive and labour-intensive, as well as socially awkward, at least in the beginning. (There are also some rare but potentially very serious health risks, says Taylor.)
Gregory Dodell, a New York-based endocrinologist who takes a weight-inclusive approach to managing diabetes, says: “You have to look at the social determinants of health. We’re not treating a population as a research experiment, we’re treating a population with a lot of different complicated variables and issues and a very complex, multifactorial chronic condition. One size does not fit all.”
In conversation, Taubes isn’t quite as dogmatic about diet as his writing makes him seem. He regularly says things like, “assuming what I’m arguing is correct”, and at one point casually notes that he could “have a heart attack tomorrow, which is possible the way I eat, and which, God knows, I keep expecting”. He does use a lot of caveats in his books (which apparently drives his publisher slightly crazy), but on the page, he nonetheless comes across as unwaveringly committed to the high-fat, low-carb way of life, so I find his concern surprising. “Well, my world is full of people pointing out the age other people died who believe what I believe. Which, of course, is selection bias, because you don’t see the people who are still alive, you only see the people who die. If I die tomorrow, maybe I would have died 10 years ago, had I not eaten the way I did. It’s always an experiment.”
Given that there is no control version of any of us against which to measure success, none of us will ever know if we chose the “right” diet. The paradox, of course, is that this kind of diet-by-hypothesis is exactly what Rethinking Diabetes rails against. But Taubes sees no alternative: “The science has been pretty awful. So many of the conceptions that have evolved around eating behaviour and nutrition are based on assumptions that may be wrong. The problem is that people don’t change their [dietary or health] advice because the longer they give it, the more invested they are that it had better have been right. I write from this perspective – of the history – so folks can see the damage that is done by allowing assumptions to be embraced as facts without definitive evidence.”
• This article was amended on 15 January 2024 because an earlier version described Diabetes UK and diabetes.co.uk as “the two biggest diabetes charities in the UK”. To clarify: Diabetes UK is the biggest UK diabetes charity whereas diabetes.co.uk, is an online support community, not a charity.
Rethinking Diabetes – What Science Reveals About Diet, Insulin and Successful Treatments by Gary Taubes is published by Granta at £16.99 on 18 January. Buy a copy for £14.44 at guardianbookshop.com
