I thought I was prepared for the notorious taste of ketone supplements, which are claimed to enhance endurance sports performance and improve recovery. I'd read about the countless negative experiences with it and seen the grimaces of athletes drinking it on camera. Plus, being a fan of grappa (a potent Italian digestif), I thought ketones were nothing to be scared of. I was wrong.
After a month of testing ketones, I can confidently say they're among the worst things I've ever drunk. However, did they enhance my cycling performance like they’re allegedly doing so with World Tour teams at the Tour de France? Let's take a look.
In January, HVMN, a company producing ketone supplements, asked if we wanted to try their new product, Ketone-IQ. The supplement, they said, tasted better and was cheaper than its predecessor (among the main downsides of these supplements). Of course, we gave it a go. But we didn’t want just to sip a cuppa of ketones and give it a pretentious review. We wanted to do something more fun.That's why we approached Adam Isherwood, a PhD student at the University of Oxford studying the role of ketones on endurance performance. Isherwood's scholarship is funded by another company producing the ketone supplements, called DeltaG and based in Oxford. He says that by the criteria of the University, his research is independent and doesn’t have any involvement in the commercial side of the company. And, viceversa, the company doesn’t have any direct involvement in his research (you’ll notice also from the results of the tests).
However, we decided that in the following months we’ll run another test with HVMN providing the support Isherwood did with this test. So stay tuned for more in the future.
This time around, though, we sent Isherwood our HVMN's Ketone-IQ, and he sent us back six different small bottles, labelled: A1, A2, B1, B2, C1, and C2. For each number, two contained ketone supplements (HVMN and DeltaG), while one of them contained a placebo. But, of course, we didn't know which was which.
Isherwood designed a blinded, placebo-controlled, crossover design test, where the order of drinks was determined randomly. The experiment was aimed at determining which supplement (or placebo) was enhancing (or inhibiting) my performances over a 30min all-out time trial. I also had to perform the TT after an hour ride at a power 15% lower than my FTP (functional threshold of power).Before the end of the three-week trial, I wasn’t allowed to look at the power numbers – or at the blood-test results my wife took along the way (thank you!).
Finally, because I undertook the test at home, I had to keep the setting as consistent as possible: using the same turbo set-up, the same power meter (charged and calibrated), same air temperature (where possible), same fatigue in the legs (same training programme for the days prior) and same food eaten in preparation for the effort.
To consume enough carbohydrates for the time trial, Isherwood recommended eating 3.3g of carbs per kg of bodyweight in the evening before the test and 2g/kg for breakfast. That equated to 240g of rice plus sides for dinner and 100g of porridge (with apples, bananas and maple syrup) for breakfast. Hydration was also essential and I made sure I drank at least 1,000 ml of water before starting.
During the first hour of the test, I drank a water bottle with 30g of carbs but, during the TT, I was asked to drink plain water.
I drank the first bottle of ketones (a double dose one) 60mins before the start. This was a bigger dose than the second one – planned for an hour later – and aimed at raising my ketone levels.Thirty minutes after the first drink, my wife took my first blood-test sample via finger pricks. The blood samples targeted the ketone and lactate levels. At the same time, I monitored my glucose concentrations through Supersapiens continuous glucose monitor. My wife then wrote the results on a notepad and placed it out of my reach (and no, I didn't look, despite my curiosity).
Fifteen minutes later, I started the 10min warm-up at 180W, stopped for 5mins and received a second finger prick.
At this point, I reported my Rate of Perceived Exertion (RPE) for the first time, drank the second bottle and started the first hour on ERG mode (a setting tat fixes the power output on indoor trainers) at 200W. Even in this case, it was important not to look at the power numbers on my bike computer.
In that long hour where I only stared at the time passing by, I reported my RPE after 30 and 60mins, while my wife took a third blood sample at the end. Finally, after all the meticulous preparation beforehand, it was showtime.
The time trial
The goal was to go as hard as possible for those 30mins. And, once again, I wasn’t allowed to look at my power numbers. It was a brutal and alienating experience, but I think I had learned how to tackle it week after week. And that's something that may (or may not) have influenced the test itself.
In those 30 long minutes, I also reported my RPE after 15mins (I took a mental note and wrote it down at the end) and then at the very end of the TT. The latter task was easy to report, as my RPE was always around nine to 10 on a scale from zero to 10 (10 being the absolute maximum). Job done.
And with the great relief of my wife, who was working in the kitchen and had alarms set on her phone, that was the last time she had to step into the pain cave and take my blood samples.
The guessing game
Before catching up with Isherwood after the test, I tried to guess which supplement was in the bottle and then looked at the performance results.After the first taste, I had the impression that samples C and B were somewhat different to sample A in both their taste and texture. However, one sample had almost made me sick and gave me the feeling of not being able to push enough: it was sample B on week two. On that occasion, I even had a hard time finishing the two bottles, which were hands down the most horrible thing I have drunk. Grappa, in comparison, tastes like a soft drink.
On week three, on the other hand, when I drank sample A, I felt the best and the strongest, also mentally. I felt prepared for the task and more ready to suffer.
So when Isherwood asked me to guess the contents, I responded that C and B were ketones and A the placebo. But I couldn't tell which of the two ketones was HVMN or DeltaG. I just tried my luck and said HVMN was sample C. If they had improved the taste of their first iteration as they claimed, it must have tasted better than the horrible taste of sample B.
The moment of truth
The results were astonishing. Not only had I guessed all three of the bottles (making the experiment a little less rigorous), but I discovered I had performed the best when I drank the placebo. And by a big margin of 40W.
The average power output for the TT using the C sample (HVMN) was 237W (154bpm) and very similar to the power output of when I drank the B sample (DeltaG): 240W and 154bpm. In comparison, my max HR is around 170bpm and my Anaerobic Threshold around 250W. When I drank the placebo, on the other hand, I averaged 285W (heart rate monitor failure, though), and I did so feeling great and physically strong.
I attribute this result to different factors. First, over the weeks, I learned how to suffer. I was mentally fresher and ready to dig deeper before the third trial. I embraced the pain more and I felt like a different person.
When I drank the two supplements, though, I struggled to chug them down because of their taste and I almost felt sick when I used DeltaG. I’m sure that the poor performance of my second time trial was 100% down to that stomach upset.
I’d felt better when I used HVMN and the taste was more bearable. However, when we looked at the ketone levels in my blood, we noticed that the level of ketones using HVMN was very small and probably not even significant for the study (between 0.4 to 0.8mmol/L). On the contrary, DeltaG spiked my ketones a lot more (4.2 to 4.8 mmol), but the higher level of ketones didn't go down well in my body.
Why such a big difference in ketone levels?
The different percentages were also due to the fact that in the first drink Isherwood had applied two commercial doses of HVMN (70ml containing 20g of butanediol) and two doses of DeltaG, which contained 50g of ketone monoester or exogenous ketone. In the second drink, on the other hand, he’d added one dose of HVMN (10g of butanediol) and one dose of DeltaG (25g of ketone monoester).Why the difference? Did he want to limit the amount of ketones in my blood from the competitors? Not really but, even if he did, the results were still in favour of HVMN.
“I used multiples of the commercial doses rather than using a personalised dose for bodyweight because I know how to scale DeltaG on bodyweight but not HVMN,” he explained. “If anything, more DeltaG made it less likely to be beneficial, as more drink meant more GI upset and this was definitely too much for you.”
Different content between brands
There’s also an important difference between the supplement used in HVMN (1,3 butanediol) and the DeltaG’s monoester. From a metabolic point of view, DeltaG’s monoester breaks down into two molecules: one of Beta-hydroxybutyrate (BHB, one of the three ketone bodies produced in the liver), which flows into the bloodstream and increases blood ketone levels immediately; and one molecule of butanediol, which is then converted into BHB in the liver.
The butanediol contained in the HVMN drinks, on the other hand, covers just half of that process and that’s why it takes longer to raise ketone levels in the blood. Butanediol is an alcohol (strictly speaking a diol) rather than a ketone, but it’s converted into BHB in the liver. And because in the liver it’s competing with other substrates for this conversion to take place, especially when you're fed and there's lots of stuff in your system, it can take a while for the BHB to build up in your bloodstream.
Limitations of short tests
Isherwood also says they usually plan a longer preload (2hrs and then the TT). Still, we had to work around that because of my time restrictions. A longer preload would have probably shown different results."But the obvious questions of this test were: does ingesting these drinks elevate your blood-ketone levels? And is there a benefit to having elevated levels? I think that from the results these answers are clear," said Isherwood in our debrief.
Despite the limitations of a home test, it was still clear that the stomach distress given by DeltaG had impeded me from performing better. And the level in my blood following HVMN was also too reduced to impact my performance.
Can you grow fond of ketones?
As athletes train their gut to absorb a significant amount of carbohydrates during an event, can you also acclimatise to ketones to find that miraculous edge that ketones promise?
"No one's looked at it yet – whether you can kind of train the gut to tolerate ketones," says Isherwood. "But I suspect that's because it would be horrendously expensive."
Through this test, we've been reminded that the performance system is a complex one and made of many different elements. And some are more important than others. Training, resting and nutrition are still at the base of the pyramid (the baking of the cake of endurance sports). They’re the most critical pieces of the puzzle.Then comes the marginal gains, those smaller pieces that sit at the top of the pyramid and make good ‘perfect’, if the basics were nailed. Among these, we can include small aero gains, high-tech testing through performance software and, at the time of writing, ketone supplements.
Pro teams use them successfully because they treat the basics savagely well. Plus, they need to find those marginal gains if they want to beat their competitors.
For amateurs, on the other hand, marginal gains don't have the same ROI as they have for pros. I would never risk compromising an important race because of stomach upset, nor am I willing to invest a lot of money to get used to the sharp taste that may, or may not, give me that edge.