Will a ketogenic/Carnivore diet help or hurt your athletic performance?
Back in the low-carb heyday - a bizarre time when people looked to sedentary, unfit, overweight authors for advice on fat loss and sports nutrition - people would say stuff like the following:
"Your bodily stores of glycogen can be depleted within an hour! However, even lean people have enough fat stores to power them for days! Therefore, low-carb diets are superior for athletes!"
Said the people who struggled to get their XXXL butts up a brief flight of stairs.
It's true that the average person stores between 500–700 grams of glycogen in the muscles and around 100 grams in the liver (around 2,400 to 3,200 calories). And yes, even a very lean 75 kg athlete with a body fat percentage of 6% is carrying 4.5 kg of fat (around 40,500 calories).
In the terribly simple-minded world of diet ideologues, this means low-carb diets deliver at least 13 times more energy for athletic activity.
Except they don't.
The content below was originally paywalled.
In June of this year, the Journal of the International Society of Sports Nutrition published a position stand on ketogenic diets. A “position stand” is where an organization commissions a group of researchers to review the evidence on a topic and then compile a review which becomes the organization’s official stance on the matter.
Many of the position statements I’ve read over the years were commissioned by dubious heart and dietetic associations, and weren’t worth the paper they were printed on. These organizations have extensive industry ties, and the statement papers they commissioned were authored by people with significant financial conflicts of interest, or who were so thoroughly inculcated that the statement was never going to be anything other than a predictable regurgitation of mainstream dogma.
The IJSSN position statement, however, is pretty solid. Among the authors are some well known names in sports science, a research field that doesn’t harbor the same rampant criminality and audacious financial conflicts of the pharma genre.
Jeffrey Stout, Bill Campbell, Richard Kreider, Douglas Kalman & Jose Antonio are all prolific sports nutrition researchers who - unlike certain ‘scientists’ who’ve made the transition from researcher to diet book author - are not known for pushing faddist dietary approaches.
Their position statement received no funding.
Only one of the eighteen authors listed on the statement evinces a potential conflict of interest: Dominic D’Agostino “is an inventor on USF patents related to applications of therapeutic ketosis” and “co-owner of Ketone Technologies LLC.” He also acts as an advisor for a company called Levels Health, alongside anti-carb authors like Robert Lustig and David Perlmutter.
Judging by its conclusions, D’Agostino’s low-carb leanings appear to have had little impact on the position statement.
Here are some key conclusions from the statement (bold emphasis added), which I’ll expound upon in this series:
"A ketogenic diet has largely neutral or detrimental effects on athletic performance compared to a diet higher in carbohydrates and lower in fat, despite achieving significantly elevated levels of fat oxidation during exercise (~1.5 g/min)."
"All studies involving elite athletes showed a performance decrement from a ketogenic diet, all lasting six weeks or less.”
“Of the two studies lasting more than six weeks (involving recreational exercisers), only one reported a statistically significant benefit of a ketogenic diet."
"There is insufficient evidence to determine if a ketogenic diet affects males and females differently. However, there is a strong mechanistic basis for sex differences to exist in response to a ketogenic diet."
The position statement also concluded "A ketogenic diet tends to have similar effects on maximal strength or strength gains from a resistance training program compared to a diet higher in carbohydrates. However, a minority of studies show superior effects of non-ketogenic comparators."
The effects of keto diets on muscle and strength gains from resistance training are something I’ll discuss in a separate and forthcoming article.
In this article and subsequent installments, I’ll discuss what those of us who are into serious cycling, running or other modes of endurance exercise can expect from a ketogenic diet.
How to Fudge Research to Support a Keto Diet: Two Prime Examples
As noted, only one of the 16 controlled trials examined in the IJSSN statement reported a significant improvement in performance for the ketogenic diet group compared to the control group.
I’ve already thoroughly dismantled this trial, in an article available to paid subscribers here. To recap, one of the authors of that study was Jeff Volek, a staunchly pro-keto researcher who has written three low-carb books and is on the payroll of one of the world's best-known low-carb food companies (Atkins Nutritionals).
Volek’s co-authors, McSwiney et al from Ireland, have published a number of other pro-keto papers.
The study began with forty-seven recreationally-trained male endurance athletes. It suffered a huge attrition rate, with twice as many keto subjects dropping out. This meant only 20 subjects completed the final testing.
As I explained in my write-up, the trial was non-randomized; subjects were allowed to self-select which group they wanted to be in. This is problematic, because certain types of people may select a certain type of intervention, which could bias the results. In this study, it seemed the heavier and fatter subjects opted for the keto intervention.
In further violation of the “control your variables” principle, the keto subjects self-reported a higher protein intake. Despite receiving the same endurance and resistance exercise intervention, they also appear to have performed a higher volume of training throughout the study.
Three tests were performed before and after the 12-week intervention.
After an initial warm-up, subjects performed a six-second sprint on a stationary bike, for which the air resistance against pedaling was increased.
This was followed by a simulated 100 km time trial. In other words, subjects were timed while they pedaled the equivalent of 100 km.
Immediately after the 100 km TT was completed, participants completed a “critical power test.” This involved maintaining as high a power output as possible for 3 minutes, during which time the same pedal resistance used during the SS sprint was applied.
During the initial six-second sprint and subsequent critical power test, researchers measured both peak power (the maximum power measurement achieved) and also the average amount of power sustained over the six seconds and 3 minute tests, respectively.
After the twelve-week intervention, both groups similarly improved their performance on a simulated 100 km time trial performed on a stationary bike.
The researchers claim the ketogenic group experienced improvements in peak power output on the six-second sprint (6%) and critical power (17%) tests.
The high-carb group, meanwhile, experienced a slight worsening in both tests.
This, declared the researchers, showed “Keto-adaptation enhances exercise performance.”
Except it does no such thing.
To portray the final results as an “enhancement” for the keto group, the researchers had to employ some clever sleight of hand.
When you read the actual paper, you’ll see that the researchers present the power figures in watts per kilogram.
That would be fine and dandy if all the participants began and finished the study with the exact same body and lean mass weights, but that’s definitely not what occurred.
At the start of the study, the keto group weighed a mean 86.3 kg and had 17.5% body fat, which seems uncharacteristically chubby for a group claiming to have trained at least 7 hours each week for the last 2 years or more. Meanwhile, the high-carb group weighed 76.5 kg and sported 12.8% body fat at baseline.
During the study, the ketogenic subjects lost 5.9 kilograms and reduced their body fat by 5.3%.
When we look at the before and after absolute wattages produced by each group - i.e. the total wattage the subjects produced, irrespective of how much they weighed - the keto group in fact experienced declines in their power output.
In absolute terms, by the end of the study both groups experienced slight declines in the 6 second sprint peak wattage (1.8% vs 1.4% in the high-carb and keto groups, respectively).
In absolute terms, the average wattage the subjects maintained during the 6 second sprint improved by 1.8% in the high-carb group, compared to a 3% decrease in the keto group.
In absolute terms, the peak wattage during the final 3 minutes of the test decreased by 8.6% in the high-carb group, while increasing 8.9% in the keto group. However, the ultimately more important average wattage during the final 3 minutes increased by 1.2% in the high-carb group but declined by 4.4% in the keto group.
All of a sudden, the results of the LCKD group don’t look so great.
The only reason the LCKD group improved on the very selectively reported peak watts/kg measures was because they lost more weight, not because they experienced some remarkable performance-boosting metabolic shift in fat utilization, as the researchers surmised.
The researchers would probably justify their selective reporting of the results by stating that relative power is more important than absolute power in a sport like cycling. “Relative power” refers to the amount of power you can generate in proportion to your body weight, whereas absolute power is the amount of watts you can generate irrespective of your weight.
Let’s say two athletes are capable of generating the same amount of watts on a bike. If one of those athletes weighs 5 kg less than the other, then - all other factors being equal - the lighter athlete will have the advantage because he has less mass that needs to be moved through space.
Which brings us to a further problem with the results.
Remember how the keto group began with a significantly higher body fat level? Well, at the end of the study, body fat levels were similar in the high-carb and keto groups (12.1% versus 12.3%, respectively).
While a body fat level of 12% might sound like nirvana to the average overweight Westerner, it’s on the high side for elite male endurance athletes, among whom single digit body fat levels are commonplace. At 12% body fat, both groups still had considerable room to drop some weight and increase their relative power further.
The best way to drop weight, of course, is to shed body fat while retaining as much lean mass (of which muscle is a primary component) as possible. Because it is muscle that powers you along, not chub.
And this is where we discover yet another non-supportive finding in the McSwiney data. When you use the data they’ve supplied in the paper to calculate the before and after watts per lean kg of body weight, we see a similar situation to what occurred with the absolute power changes.
Expressed in terms of watts per kg of lean mass, average six second speed power marginally increased from 14.7 to 14.9 watts/kg in the high-carb group, but fell from 15.7 to 15.2 in the keto group. Average power in the 3-minute critical power test remained unchanged in the high-carb group, but dropped from 5.7 to 4.7 watts/kg in the keto group.
Which begs the question: What would have happened if both groups leaned out further? As they approached their ideal competitive weight, would the keto group suffer a further decline in power output?
That’s difficult to answer based on the results of this study, because it was a non-randomized endeavor featuring subjects with disparate baseline characteristics. Further alienating the study from the crucial “control your variables” principle is that the two groups strongly appear to have utilized disparate protein intakes and training volumes.
That’s still not the end of it. Remember how twice as many subjects dropped out of the keto group? During the study, keto participants noted a drop in energy levels and performance during the first 7–10 days, and what the researchers called a “lag” in performance for the first 4–6 weeks. Five participants found the keto diet too difficult to adhere to, while two keto participants who remained in the study were unable to complete the final testing. Another keto participant could not complete the final testing due to “technical difficulty,” but no further explanation is provided.
Not being able to complete an event because you ran out of gas on a keto diet can hardly be considered an “enhancement” in performance.
Endurance athletes need to weigh up the marked 4-6 week drop in performance that will ensue after commencing a keto diet, and decide whether it is justified when the likely result is equal or worsened performance when compared to a high-carb diet.
They also need to consider they may be so unsuited to the diet that it will cause them to terminate an event before it is completed - not a great way to achieve podium glory and keep sponsors happy (in part 2, we’ll meet a truly world class triathlete who experienced this exact fate).
The McSwiney et al study is by no means the win for ketogenic diets its authors made it out to be. They selectively presented data suggesting a performance-enhancing effect of a ketogenic diet, while ignoring other metrics that indicated a worsening of performance from that diet.
This was not really a study comparing the effect of two different diets on performance, but the effect of weight loss on selectively cited measures of performance. Weight loss is ultimately a function of calories in versus calories out, not how little carbohydrate you eat.
The Infamous Phinney Study
In the Pro-Keto Junk Science Hall of Fame, one study holds a special place. I’m talking the infamous cyclist study conducted by Steven Phinney, who also happens to be a low-carb book author (he’s even co-authored a book with Jeff Volek). The study was first reported in two 1983 papers authored by Phinney and his MIT and Harvard co-authors (see here and here).
This study involved five competitive road cyclists, who were housed in a metabolic ward for five weeks but were allowed to go about their daily activities and training. Like the McSwiney et al study, this was not a randomized trial. In fact, it didn’t even feature two groups following different diets, or the same group following the different diets for equal periods in random order.
Instead, during the first week, the cyclists ate a eucaloric (weight maintenance) mixed diet. During the remaining four weeks, they consumed a eucaloric, high-fat, zero-carbohydrate diet (15% protein, 85% fat).
The before-and-after testing consisted of a “time to exhaustion” test, which is where you pedal on a stationary bike until you either get too tired or bored to continue pedaling any longer.
To believe Phinney et al, no decrease in endurance was noted during the final time to exhaustion test. “The most striking result of this study,” they wrote, “was the ability of highly trained endurance athletes to maintain their level of training and perform the extent of endurance exercise observed at [week 4 of the ketogenic diet].”
The study, in fact, showed no such thing.
More Holes Than a Bullet-Ridden Sieve
The study design and reporting was marred by a raft of fatal flaws. Before I reveal what really happened to each subject’s performance at week 4, it’s important to understand these flaws.
Uneven periods on each diet: The cyclists were observed for only one week while following a mixed diet, but four weeks following a ketogenic diet. What would the results have been if they followed both diets for an identical period? Thanks to the poor design of Phinney's study, we'll never know.
No random assignment to treatment. Not only should the subjects have followed the mixed and keto diets for the same period of time, some should have followed the mixed diet first and then the ketogenic diet, while others should have started out on the ketogenic diet then followed the mixed diet.
Why? Because if a difference in a performance variable is noted among the two diets, we can be confident it was a result of the diet itself and not an artifact of the switch from one diet to another (it's no big secret that switching between starkly contrasting dietary regimens can result in temporary perturbations in biochemical/physiological parameters).
In the time to exhaustion test, the subjects were told to pedal at a power level equivalent to 60%-65% of their VO2 max for as long as they could. This has zero relevance to real-life endurance events. There is no officially-sanctioned cycling event in which the participants are told at the starting line, "OK guys, when the flag drops, ride at 65% of your VO2 max, and ride at that pace for as long as you can. Last man standing wins!"
Not only that, but 60%-65% of VO2 max is not an especially high level of exertion. At higher percentages of VO2 max, muscle glycogen and blood glucose become increasingly important energy sources, because triglycerides (fats) simply can’t generate the adenosine triphosphate (ATP) required for muscular contraction quickly enough (I’ll discuss this in more detail soon).
There was no familiarization test. In studies involving blood tests or x-rays, familiarization tests are not generally necessary. There is little skill or mental challenge involved in sitting still for a few minutes while you have your blood taken or chest x-rayed.
However, in studies involving physical tests, it is crucial that researchers subject all the participants to an initial familiarization test before the main study kicks off.
As noted, the five subjects in the Phinney et al study were competitive road cyclists. Real life cycling events occur in three main formats, all of which; 1) occur outdoors; 2) involve a moving bicycle, and; 3) contain a clearly defined and predetermined course.
The type of cycling event most readers will be familiar with, because they’ve seen it on TV, is a non-circuit point-to-point race, as seen in stage races like the Tour de France or one-day classics like the infamous Paris-Roubaix. These races involve a clearly defined start and finish point, and the goal is simple - get to the finish line first.
Criterion-style races involve a circuit-type course, in which the usual scenario is that the course must be completed for a certain number of laps. The first person across the line after completing the requisite number of laps wins.
The other primary event is the time trial, which can involve either an individual rider or his entire team. Like a stage race, this also involves a set course. The goal is to get from the starting gate to the finish line in the quickest possible time. The winner of a time trial event is the person or team that clocks the day's fastest time.
In the Phinney et al study, the subjects did none of these things. Instead, they were instructed to perform a time to exhaustion test - a test most road cyclists have never done before, and which bears little resemblance to their usual style of riding and competing.
So you have a bunch of guys whose normal cycling activity occurs outdoors, involves changing routes and scenery, involves the use of different gear ratios during a ride, and also involves varying tempos and levels of exertion. Both their training rides and competitive events have a predetermined start and finish point.
But now you are sticking them indoors on a piece of equipment that goes nowhere no matter how fast you pedal. You are telling them to sit on the spot and pedal that piece of equipment at a fixed pace, in a single gear, for as long as they possibly can.
In other words, you are now asking them to perform an unfamiliar activity.
Clinical trials confirm that when you take highly-trained cyclists and place them in the unfamiliar lab setting, their performance on the second test is consistently superior to that of their first test (for a sample study involving the time to exhaustion test, click here; for a sample study involving a simulated 100 km time trial test, click here).
As a result, familiarization procedures are a crucial and standard requirement in properly conducted studies examining physical performance variables.
But familiarization testing was never performed in the Phinney et al study. In the Phinney et al study, the second test, which will inevitably evince improvement over the first, was also the final test!
This fact alone basically invalidates the entire study.
So we have a study that skeptical minds could be forgiven for assuming was deliberately designed to find a benefit for the keto diet.
The second time to exhaustion test - where subjects would be expected to show improvement anyhow - was scheduled after the four-week keto intervention.
During those time to exhaustion tests, the researchers instructed the subjects to avoid higher levels of exertion that we know require glucose to generate the ATP required for muscular contraction.
Now here’s the real cracker. Despite the study being designed in a manner that favored the keto intervention, only two of the five cyclists experienced any significant improvement on their second time to exhaustion test!
Despite the research-confirmed advantage of familiarization from the first to second test, another subject experienced a negligible and non-significant 3-minute increase. The remaining two cyclists, meanwhile, experienced hefty declines in their performance (-48 and -51 minutes)!
So how were Phinney et al able to pull off the misleading claim that the keto diet caused no decline in performance?
Easy. They ignored the individual results, and instead added the combined times together and calculated an average time-to-exhaustion figure and … voila! The resultant figure made it look as if no decline in performance was observed.
But wait, there’s still more.
The results section of the second 1983 paper by Phinney et al began with the claim, “The EKD was tolerated by all five subjects without difficulty.”
Buried in the final passages of the paper were the following comments:
“There are indications in the results of this study that the price paid for such extreme conservation of carbohydrate during exercise appears to be a limitation on the intensity of exercise that can be performed.”
But Phinney et al didn’t expand upon just what they meant by this, and it took over twenty years for Phinney to finally come clean. In a 2004 paper, he casually admitted:
“The bicyclist subjects of this study noted a modest decline in their energy level while on training rides during the first week of the Inuit diet, after which subjective performance was reasonably restored except for their sprint capability, which remained constrained during the period of carbohydrate restriction.”
So not only was the submaximal performance of 3 of 5 of Phinney’s cyclist unchanged or worsened after the keto diet, but all of them experienced a sustained decrease in their performance at higher intensity levels.
For competitive athletes, and recreational athletes who like to challenge themselves at a high level, this is unacceptable.
I’ll give you a little analogy to help explain why.
Let’s say you have a car that develops an unsettling rattling noise every time you exceed 60-65 km/h. Imagine taking it to your mechanic, and instead of fixing the problem, he simply says, “well, don’t drive your car over 60-65 km/h then!”
You’d think the guy was nuts.
But let’s imagine that you take his advice. You live in an area where many of the streets have a 50 km/h speed limit, so you try to stick to those roads.
But then one day, your child has a serious accident and you need to rush him to the nearest hospital. You don’t fancy waiting around for paramedics, so you bundle him into the car and begin speeding down the freeway. Before you know it, you hit 80 km/h and then … boom!
Your engine blows up. You are now stranded by the side of the road with a very sick child who needs urgent medical care.
In performance terms, the equivalent would be trying to power up a steep hill on a bike, or trying to chase down the leaders in the final leg of a triathlon, but instead “blowing up” (cycling parlance for that awful moment when you run out of gas during an event or big ride).
Volek, Phinney and McSwiney et al will never admit it, but this is actually a common outcome among highly active people who attempt to fuel their training regimens on a ketogenic diet.
In the next installment, I’ll give some highly instructive examples.
Leave a Reply