Nobody ever changes a strongly held opinion because of a single study. It鈥檚 too easy to find potential flaws in any given set of results: they tested the wrong people, or used the wrong protocol, or simply got a fluke result. This isn鈥檛 necessarily a bad thing. You 蝉丑辞耻濒诲苍鈥檛 discard all your previous opinions every time you read a new study. But it raises a question that鈥檚 worth asking yourself now and then with regard to controversial topics: What level of evidence would it take to convince you to change your mind?
That鈥檚 the implicit question underlying a 聽on the long-standing controversy about whether low-carb, high-fat (LCHF)聽diets enhance endurance performance. Back in 2017, a team led by Louise Burke at the Australian Institute of Sport published 聽suggesting聽that three and a half聽weeks of an LCHF diet consisting of at least 75 percent fat and less than 50 grams a day of carbohydrates (the equivalent of two bananas) turned elite racewalkers into fat-burning machines聽but so that their overall race performance suffered. The results put Burke at the epicenter of a maelstrom of controversy鈥攕o she and her colleagues doubled down and repeated the whole study, calling their effort Supernova 2.
The basic theory behind LCHF (also known as ketogenic) diets for athletes is fairly straightforward. We all carry around plenty of energy in the form of fat stores, but we generally can鈥檛 burn it quickly enough to fully fuel anything more intense than a mild jog. If you adapt to an LCHF diet, you can double or even triple your maximum rate of fat burning. In theory聽that might allow you to fuel long-distance events mostly with fat, reducing the need to cram in carbohydrates during the event聽and eliminating the risk of bonking.
One thing everyone agrees on is that going LCHF will indeed ramp up your fat-burning ability. Beyond that, there are lots of different bones of contention. For example, some studies suggest that the improved fat burning comes at the direct expense of carbohydrate burning, which means you lose some of your ability to summon quick bursts of energy for climbing hills and covering surges. But the objection raised by Burke鈥檚 study is more fundamental: in exchange for access to the bottomless pit of fat stores, you become less efficient. That means you require more oxygen to produce a given level of muscular output. If you鈥檙e exercising so intensely聽that聽you鈥檙e out of breath, increasing your oxygen consumption is bad news.聽
But that only matters if you actually believe Burke鈥檚 results. In the new paper (which is free to read online), she devotes an entire page and a half to a table summarizing criticisms of the original Supernova study that appeared聽in peer-reviewed literature and even on social media, along with the adjustments they made in response. For example, the original study was split into two separate training camps at different times, raising the possibility that conditions weren鈥檛 the same for everyone. This time聽they brought in 28 athletes鈥攁 larger sample size鈥攖o a single camp, where everyone聽went through the same protocols at once. Also, they added a final race two and a half聽weeks after the end of the test diet to check whether there was a delayed benefit to training on an LCHF diet before returning to a high-carb diet.
The design of the new study is fairly intricate, but the results are simple to summarize: they鈥檙e pretty much identical to last time. The racewalkers on the LCHF diet got way better at getting energy from fat, more than doubling their fat-burning rate at typical race speeds. But they also got less efficient鈥攖hey consumed an additional 7.1 percent of oxygen at their approximate 20K race speed聽and an additional 6.2 percent at their 50K race speed. Is that a big deal? Well, the infamous Nike Vaporfly 4% shoes were so named because they reduced oxygen consumption by 4 percent on average. And in both a 10K race immediately following the experimental diet and a 20K race two and a half聽weeks later, the high-carb group got faster while the LCHF group got slower.
As before, there are still some loopholes and caveats. The efficiency penalty may be less pronounced at lower intensities, making it more suitable for ultradistance races. Some people may not even care about small changes in efficiency if they鈥檙e able to become less dependent on constantly stuffing in carbs throughout a long race. And some people don鈥檛 care about race performance at all聽and stick with聽LCHF diets for other reasons. But for the population studied here, who race all out for four hours or less and care about every second, it鈥檚 getting harder and harder to escape the conclusion that an LCHF diet is (at best) not an improvement on conventional mixed diets.
From a methodological point of view, of course, a scientist replicating her own study isn鈥檛 as convincing as a totally different group reaching the same conclusions. But here鈥檚 where we come back to the question at the top: How much evidence does Burke need to provide to change minds? And at what point does the onus shift to those who disagree with her? After the original Supernova study, the Journal of Physiology published enumerating all the possible ways Burke鈥檚聽conclusion might have been incorrect. What struck me was how many coulds and mays and it is possibles it contained. It may be, for example, that it takes longer than three weeks to reap the full 鈥渋mmune, neurological, microbiome and hormonal effects鈥 of LCHF diets. Burke鈥檚 results can鈥檛 disprove this idea,聽but to my knowledge, no one has presented any evidence that proves it, either.
It鈥檚 been nearly a decade since I started reading (and writing) articles about the theoretical endurance benefits of LCHF diets. During that time, I鈥檝e certainly spoken to lots of people who swear by this approach. I think it鈥檚 a cool idea聽and theoretically plausible. And I understand that studying these things properly is a huge endeavor. But given the now replicated Supernova findings, I have to let data override hypotheses. The most convincing way to refute Burke鈥檚 results isn鈥檛 to list all the things she might have done wrong; it鈥檚 to do it right, demonstrate the performance benefits, and publish the results. Better yet, do it twice.
For more Sweat Science, join me on and , sign up for the , and check out my book .