New study: high intensity exercise on a low carb diet.

Switch an athlete from their standard carbohydrate-rich diet to a low carb ketogenic one and suddenly performance tanks.  It is known.  Give them a few weeks to adapt, however, and it recovers.  This much was established for mainly endurance-related performance parameters by Steve Phinney and colleagues in the 1980’s (eg, Phinney et al., 1983).  Then, along came Antonio Paoli, Dominic D’Agostino, and others who showed a similar phenomenon in gymnasts, a population that routinely exercises at higher levels of intensity (Paoli et al., 2012).  Notably, in these studies the athletes were allowed adequate time to adapt to the new metabolic milieu – sometimes referred to as ketoadaptation.  Three weeks appears to be the minimum amount of time required for ketoadaptation; ie, studies of shorter duration generally show: low carb = poor physical performance.

…which is why I was surprised to see this one:

Effects of a short-term carbohydrate-restricted diet on strength and power performance (Sawyer et al., 2014)

These researchers subjected ~30 strength-trained individuals to a battery of performance assessments before and after 7 days of a low carb [ketogenic] diet.  Usually I would’ve stopped reading at this point because 7 days is too short.  But there were some nuances in the way this particular study was designed which piqued my interest.

After a week of measurements on their habitual diet, the participants switched over to the experimental diet.  And as expected, the ketogenic diet induced a spontaneous reduction in appetite.  Food intake over the control and experimental weeks was as follows:

Diet grams

diet kcals

Since their normal training routines apparently didn’t change, we can safely say the [modest] changes in body composition were diet-induced.  The control, or habitual diet didn’t have any impact because these participants [theoretically] weren’t doing anything different.

Low carb dieters lost a little bit of weight.  Given the nature of carbohydrate restriction, some of the weight lost was water weight.  But given the caloric deficit & relative hypoinsulinemia, some body fat was lost as well (statistically significant in women only):

Body comp

Lean mass was somewhat preserved, likely due at least in part to the increased protein intake.  Note that lean mass didn’t increase, which could’ve theoretically masked a negative impact of carbohydrate restriction on physical performance; it was merely preserved.

Instructions for the low carb group were simply to limit carbs to less than 50 grams.  They tried, but failed, to adequately compensate for the lost carbohydrate calories by eating more protein and fat.  Some might argue that this, the increased protein intake, may have confounded the results.  I won’t argue against that, but I don’t think it invalidates this study because: 1) it’s what would likely happen in real life; and 2) glycogen depletion without adequate ketoadaptation seems like a greater immediate detriment to physical performance than increased protein intake would be a performance “booster.”

The performance assessments: handgrip dynamometry, vertical jump, max bench press (1RM), 1RM back squat, maximum-repetition bench press power output, and Wingate anaerobic test.  Needless to say, these tests gave the researchers deep insight into the participants’ maximal strength and power output.  To put this in context, other studies have assessed time to exhaustion during a long slow bike ride or jog – the opposite of “high intensity.”

For the reasons mentioned above, we might have expected strength and power output to plummet.  This didn’t happen.  In fact, some of these measurements actually improved (modestly, albeit statistically significantly):


Nuance: the participants refrained from exercise for 48 hours prior to testing.  Not all studies do this.  Worded another way: a low carb [high protein] diet might improve strength and power output given adequate rest…  Or, eating less carbs and adding a few more rest days = increased training-induced gains in strength and power… any takers?!  Of course, this wasn’t assessed directly, and it’s difficult to say how much, if any, the added rest days contributed to the results.

Given the theories underlying the relationship between ketoadaptation and power output, I wouldn’t have been surprised to see these results had the study lasted longer than 3 weeks.  One possibility as to why it worked is that the process of ketoadaptation was expedited in these individuals because they were: 1) exercising; 2) eating significantly fewer carbs than in most other short-term studies; and 3) in an energy deficit… three well-known, potent ketogenic stimuli.  The added protein might have helped, also, and certainly wouldn’t have hampered ketoadaptation given their energy deficit.  But still, that’s putting a lot of stock into this process as being a major reason why the diet worked in such a short time.

What we know: weight loss is generally accompanied by a loss in strength and power output.  A counter-argument could be that some of the weight lost was water weight… although when has dehydration ever been shown to improve performance?  If anything, this would’ve had a net negative (or neutral) effect…

The authors speculated that the increased handgrip strength could’ve been due to the novice effect, because, despite the stereotype (at least in my experience), gym buffs don’t actually walk around squeezing a grip crusher…

grip crusher

…but they should’ve been roughly familiar with most of the other tests… well, maybe not the Wingate test.  NO ONE SHOULD BE FAMILIAR WITH THIS TEST.

At this point, I don’t think this experiment was bollixed, and I’m not entirely sure how much the nuances and subtleties impacted the results.  And I’d still recommend any athletes who want to see if their performance improves on a low carb diet to stick it out for at least 3 weeks.  But this study certainly piqued my interest.

An abbreviated history of low carb diets and exercise – take a quick glance at the macronutrients, exercise routines, participants’ level of experience, rest periods, etc., etc. and you’ll see what I meant when I said these studies are very nuanced.  Or just bookmark ’em for future reference.

The human metabolic response to chronic ketosis without caloric restriction: preservation of submaximal exercise capability with reduced carbohydrate oxidation. (Phinney et al., 1983)

Enhanced endurance in trained cyclists during moderate intensity exercise following 2 weeks adaptation to a high fat diet.
(Lambert et al., 1994)

Effects of dietary fat on muscle substrates, metabolism, and performance in athletes. (Vogt et al., 2003)

Endurance capacity and high-intensity exercise performance responses to a high fat diet. (Fleming et al., 2003)

An isoenergetic high-protein, moderate-fat diet does not compromise strength and fatigue during resistance exercise in women. (Dipla et al., 2008)

Ketogenic diet does not affect strength performance in elite artistic gymnasts. (Paoli et al., 2012)


calories proper


Become a Patron!



Be Sociable, Share!
  • Jack Kruse

    Great stuff Bill. You keep this up and you’ll have all the paleo meat heads trying to come after you! 😉 The longer you go ketotic the more performance you can get. It maintains your redox potential and improves EEG, EMG, and fMRI data when you look for it. The key for humans it improves our flow of electrons over cell membranes and in ETC. This generates large voltages everywhere in a cell. When you do this to electrons, sleep improves and the DC current in interfascial water below myelin layers and outside the axons improve. As this improves we see massive changes in EEG’s. Magneto-encephalograms also show increase magnetic fluxes in white mater. This is tied to better energy transduction of water confined in microtubules of the brain leading to an eventual increase in performance across all organ systems. It takes 24-36 months to be measured in a non athlete. Robert O. Becker established that the DC current in interfascial water is the source of the regenerative current in all life forms. The DC current decreases as life becomes more complex. In salamanders they can regenerate totally to re-establish their performance. Becker found humans only true ability to regenerate is found in bone and finger tips. The reason was the DC current was centered in the brain’s neocortex to drive complex neocortical function. Fritz Popp took his work deeper and found some rather shocking results tied to this. When electrons in CSF interact with DHA a photon is given off to the brain tissue. Popp found the number of photons emitted seemed to be linked to the organism’s position on the evolutionary scale of life. This finding was similar to Becker’s findings in electrons in the tree of life at the neuroepithelial junction where injury occurs. Popp the more complex the organism, the “fewer photons” were emitted. Rudimentary animals, like salamander and plants tended to emit 100 photons/cm2/sec at a wavelength of 200-800 nm, corresponding to a very-high-frequency electromagnetic wave well within the visible range, whereas humans emit only 10 photons/cm2/sec at the same frequency. Remember that water absorbs best in the 270-80 nm range according to Pollack’s work. What increases water inside cells and microtubules? Ketosis. A ketogenic diet provides mitochondria more electrons than any other template. The more electrons provided the more photons liberated to the brain to be distributed all over the body to increase performance. This is how it all happens.

    • Dr. Volek might be publishing some research on athletic performance after *very* long-term ketosis (in humans). I’m not expecting to be surprised, but still looking forward to some more clinical studies on this population… those data are few & far between.

      • Dan Ordoins

        I would like to read that research… Let us know when it comes out please…. Also nice posting. Thanks!

        • Hi Dan, thanks!

          …regarding Volek’s research… it might be a while (ie, years not months); last I heard, they were still recruiting.

  • Marc Perry

    Hey Bill, interesting and thoughtful analysis. Thanks for sharing.

  • I think that the 48hr rest prior to testing allowed muscle glycogen stores to be “trickle-charged” from blood glucose. More muscle glycogen = more high-intensity exercise possible.

    • Penelope

      Is muscle glycogen the limiting factor for high intensity exercise? I assumed it was oxygen. Has this been tested?

      • When people “hit the wall” during high-intensity exercise (e.g. marathons), it’s because they’ve run out of muscle glycogen.

        • CynicalEng

          A marathon is endurance, not high intensity. Glycogen can be exhausted but not on account of the intensity – more the integral of the intensity over a long time period.

          • O.K. I was just hypothesising that having a 48hr rest before doing the tests allowed muscle glycogen to accumulate, which is why strength and power output didn’t plummet.

          • CynicalEng

            For sure. I liked your “trickle charge” analogy, I think the total quantity of glycogen present is lower in keto due to hypoinsulinaemia.

          • I used to be an Electronic Engineer.

          • Wenchypoo

            Well, Jaminet used to be a rocket scientist.

        • Danny J Albers

          Which is precisely why ketoadaptation is so appealing to ultra runners like Mike Morton argueably a legend of the sport who consistenly fllows a keto diet and swears by it.

          “The Bottom line is the diet works for me in training, life and racing and so far the “side effects” are all positive…other than not being able to eat a jar of nutella. Again I feel compelled to state I know that no one diet works for everyone. I just want to share my experience with a ketogenic diet.”

        • johnnyv

          Marathons are the opposite of high intensity exercise.
          1rm squats and dead lifts are high intensity and use phosphocreatine which can be replenished via beta oxidation with adequate rest between repetitions.

      • I think “force,” in this case, would be more closely related to the speed of energy production. Glycogen to lactate generates ATP very rapidly.

        • Jack Kruse

          only when the redox potential is high Bill. during intense exercise it is not high. This has been Peter’s (hyperlipid) main point with delta psi for a while now.

          • that said, I suspect a similar (or greater) rate of energy production can be had given adequate ketoadaptation, resulting in the ability to generate equal or greater power output (which could explain the findings of this study)…

          • Jack Kruse

            The answer is in Gilbert Ling’s work and Gerald Pollack’s work. It is all tied to the relationship of K+ to water charge separation, and ATP molecules stochastically joined. When you capture the energy in a powered up electron you are catching the light power of the photon fro free. This is what drives biochemistry.

    • at 30 grams of carbs per day and very little insulin to help, the keto group likely had substantially lower glycogen than controls (265 grams carbs/d).

  • rs711

    I Haven’t read the study yet so I apologize if this has already been addressed. Was there any mention/speculation concerning their GNG capacity ‘upregulating’ in order to meet the glucose demands of glycolytic activity?

    • that wasn’t addressed; the discussion section was mostly about changes in body composition.

  • John Lushefski

    I haven’t seen anything about adaptability regarding ATP-CP system. Has there been research there? I have never noticed any issues with low carb, but some friends have. I do weightlifting 4 days, bouldering 1-2 days, ballet 4-7. I walk a lot too.

    • Much of that research was done by Veech and colleagues in the 90s.

      • John Lushefski

        My problem is usually that I feel the information may not translate well to me. This blogpost is cool, since most studies use things like leg extension exercises or bike sprints. I would be more interested in vertical jump, 50m sprint, etc, especially in advanced/elite athletes.
        I think there was a mid-80s study that showed some significant muscle and strength gains in elite weightlifters with high protein (3.5g/kg maybe). I’ve never read it however, so I don’t know if they even controlled for calories.
        We can design ketogenic diets that look very different, so I wonder how that comes into play. I eat a lot of fermented dairy and coconut for instance.

        • Uh, you mean like this one from a couple years ago:

          “Ketogenic diet does not affect strength performance in elite artistic gymnasts

          We analyzed body composition and various performance aspects (hanging straight leg raise, ground push up, parallel bar dips, pull up, squat jump, countermovement jump, 30?sec continuous jumps) before and after 30?days of a modified ketogenic diet. The diet was based on green vegetables, olive oil, fish and meat plus dishes composed of high quality protein and virtually zero carbohydrates

          No significant differences were detected .. in all strength tests. ..After VLCKD there was a decrease in body weight and fat mass”

          • John Lushefski

            Yea, I just mentioned that on Hyperlipid. Protein intake was very high–I wonder about the long term health effects.

          • 200g – certainly more than you need, but I don’t consider it excessive, especially for “elite gymnasts”.

          • disclaimer: I’m not really in the “high protein = always bad” camp.

        • They actually did test vertical jump, it improved slightly (

          And yeah, increased protein intake *could’ve* been somewhat of a confounder…

          “We can design ketogenic diets that look very different, so I wonder how that comes into play.”

          Check out some of the studies linked at the end of the post – no two are alike!

          • John Lushefski

            Yea, the vertical jump test is why I said this one is cool. Something interesting would be maybe 100m sprint, since ATP-CP system should handle most of it?

          • John, check out the Wingate test. It measures the same thing but is done on a cycle ergometer.

          • John Lushefski

            An interesting observation: when I have tried bouts of high carb (1000g/day), it seems my nervous system “tires” more easily. When I do many heavy squat/deadlift singles or doubles, I lose motivation and strength more quickly than when low carb (<=100).

          • I agree, that is pretty interesting.

            in acute overfeeding studies, similar (I think) to your “bouts of high carb (1000g/d),” I’ve seen something like this… participants report what could be interpreted as “nervous system ‘tires’ more easily.” iirc, i’s not entirely consistent across studies, but interesting nonetheless.

  • Tuck

    “…although when has dehydration ever been shown to improve performance?”

    “Inverse relationship between percentage body weight change and finishing time in 643 forty-two-kilometre marathon runners”

    Noakes covers more of his research on this in his book Waterlogged. He finds that the fellow who’s the most dehydrated in a marathon is typically standing at the top of the podium.

    • interesting. Thanks! Any mention of high intensity or anaerobic exercise in that book?

      • Danny J Albers

        Waterlogged by Noakes is a stellar read and as you know he is a big fan of your book

        • word

          Noakes is great, no doubt… but a whole book about drinking water? or is that grossly oversimplifying..

  • Pingback: THE VIKING MANIFESTO Part 4: LCHF and Energy Metabolism Within the Muscle()