I put together a YouTube playlist of Athletic Performance on the Ketogenic Diet.
An interesting study which is said to show good results for the keto diet and athletic performance (Hyun-seung Rhyu1 and Su-Youn Cho. The effect of weight loss by ketogenic diet on the body composition, performance-related physical fitness factors and cytokines of Taekwondo athletes . J Exerc Rehabil. 2014 Oct; 10(5): 326–331.).
The participants were randomly assigned to 2 groups, 10 participants to each group: the ketogenic diet (KD) group, and the non-ketogenic diet (NKD) group.
The diet/training period was only 3 weeks. The performances were compared:
Aerobic capacity was evaluated by measuring the time taken to finish a 2,000 m sprint. Whereas anaerobic capacity was evaluated by the Wingate test (Bar-Or, 1987), by measuring peak power, mean power and fatigue index using a Monark cycle ergometer (Monark 894-E, Sweden). Muscle strength was evaluated based on the measurement of: (1) grip force (TKK 5401, Takei, Japan) and back muscle strength (TKK 5402, Takei, Japan) using a digital measuring instrument, (2) muscle endurance by measuring the number of sit-ups performed in 60 sec, (3) instantaneous reactionary force by measuring time and distance on 100 m sprint and standing broad jump, respectively, and (4) balance by measuring duration on single leg standing with eyes closed.
The diet was only three weeks long. The body composition results were not great for Low Carb.
|Variables||KD (n= 10)||NKD (n= 10)||F-value|
|Weight (kg)||64.11± 7.19||60.34± 6.59||63.69± 7.64||61.16 ± 7.84||G||0.004|
|%Body fat (%)||12.59± 3.96||12.21± 3.59||11.31± 2.77||10.23 ± 2.63||G||1.283|
|Lean body mass (kg)||54.65± 3.93||52.47± 4.67||54.94± 6.50||53.55 ± 8.16||G||0.067|
|BMI (kg/m2)||21.44± 2.10||20.18± 1.79||21.08± 1.94||20.23 ± 1.97||G||0.032|
The KD kids lost twice as much weight which is good. The %Body fat wasn’t changed nearly as much in the KD group as in the NKD group. Worse than that, the KD group lost much more Lean Body Mass. BMI tracked the weight loss again demonstrating the weakness of the BMI measurement for tracking body composition. I think it may be the case that the LBM mass was mostly from lost water weight.
But how about performance?
- KD did better on the 2,000 meter sprint after training than before. The NKD didn’t show much of an improvement. However, the timeframe listed in the results was 500 minutes. I don’t know how to make sense of that timeframe. It took the participants over 8 hours to sprint for a mile and a quarter? Could the units be in seconds? If it was seconds then the time would be 8 minutes to run a little over a mile. That is possible in high school athletes.
- What is striking is how much worse both of the groups did after training. Both groups lost peak power and mean power and the KD group lost significantly more power than the NKD group.
- Another striking parameter was how much worse the NKD group got in anaerobic fatigue after the training. That is surprising.
- The KD group did worse on grip strength gains.
- The KD group did not improve as much on back muscle strength gains.
- Both groups took longer on the 100m sprint.
- Both groups could not jump as far in the broadjumps after training.
- Equally disturbing was the lack of improvement in the sit-ups on the NKD group compared with the KD group.
All in all this points to a fatigued group for the final tests. It is possible to speculate that the KD group was less fatigued because the performed at a lower rate during the week or two of keto adaptation and stored up strength during that period. They show less signs of fatigue.
Issues with the study
- One of the issues is that the performance tests were performed after fasting for 12-hours. That seems like a unlikely scenario which greatly benefits the Lower Carb cohort. In a more likely scenario both groups would have followed their diet.
- There seems to have been no test such as RER or even measurements of urine ketones to verify ketosis in the Low Carb cohort.
- The High School students were given lists of foods to eat. There were no food logs and no verification of compliance. There was no dietary analysis at all.
- Muscle endurance was done by seeing how many situps could be done in 60 seconds.
- There was no control group in diet.
- The standard deviation bars on the data were much bigger than the change effects.
Taekwondo is characterized as:
a comprehensive physical exercise involving high intensity movement of the muscles and joints of the whole body at the mean 85–95% HR-max
Yet, I can’t find an explanation of the time duration of Taekwondo.
A video from 2015 where Dr. Volek talks about the FASTER Study (Jeff S. Volek, Daniel J. Freidenreich, Catherine Saenz, Laura J. Kunces, Brent C. Creighton, Jenna M. Bartley, Patrick M. Davitt, Colleen X. Munoz, Jeffrey M. Anderson, Carl M. Maresh, Elaine C. Lee, Mark D. Schuenke, Giselle Aerni, William J. Kraemer, Stephen D. Phinney. Metabolic characteristics of keto-adapted ultra-endurance runners. Metabolism, Volume 65, Issue 3, March 2016, Pages 100-110.).
- Fat adapted athletes become “bonk proof” (see my post about that).
- Group of ultra-runners.
- More athletes volunteered than could be tested.
- Matched groups.
- LCD group was 70-20-10 F-P-C.
- HCD group was 25-15-60 P-F-C.
- Day 2 was three hours at 65% of VO2max (see my post about that). He later stated it ended up being at 64% of VO2max.
- They thought peak fat oxidation would be lower due to other studies documenting lower rates. They could have told from any low carb VO2max test that the peak rates were higher in Low Carb dieters.
- It looks as if they picked the 65% number based on this study (Achten J1, Gleeson M, Jeukendrup AE. Determination of the exercise intensity that elicits maximal fat oxidation. Med Sci Sports Exerc. 2002 Jan;34(1):92-7.).
- Volek showed the same graph from the VESPA article with the shift up and to the right of the fat oxidation curve (see my post about that).
- The statistically identical glycogen levels before, after and at the end of recovery were a surprise to Volek (as they are to me). Does fat allow the glycogen stores to refill? He thinks there is a chronic adaptation in LC athletes. It isn’t likely to be peripheral insulin resistance since the athlete’s muscles were biopsied to measure the glycogen levels, right? Alaskan sled dogs may provide a clue?
- Athletes were on LC for an average of 19 months.
- Gene expression differences between the two groups still being analyzed. Glycogen metabolism gene differences.
- LDL Cholesterol levels were much higher in LC athletes. HDL was also much higher in LC athletes.
- LDL Particle distributions were better in LC athletes (fewer smaller and more large LDL).
- Insulin Resistance scores were much better in LC athletes (top 1% of population).
- Half the high carb athletes have switched to low carb diet after the study.
I’ve got another friend, call him Van Wilder, who got his VO2max tested. He’s a 35 year old triathlete who has done an Ironman and marathons. He has about five years of running training. He did the Ironman fat adapted with very few carbs during the event. He’s recently gotten off the Low Carb bandwagon, at least partly. He is still lower carb.
Van Wilder’s VO2max came in at a respectable 59.8. That puts him within 5% of the elite athletes like Zach Bitter and Ben Greenfield. Van Wilder’s max fat oxidation rate is very close to the top levels measured in FASTER (Jeff S. Volek, Daniel J. Freidenreich, Catherine Saenz, Laura J. Kunces, Brent C. Creighton, Jenna M. Bartley, Patrick M. Davitt, Colleen X. Munoz, Jeffrey M. Anderson, Carl M. Maresh, Elaine C. Lee, Mark D. Schuenke, Giselle Aerni, William J. Kraemer, Stephen D. Phinney. Metabolic characteristics of keto-adapted ultra-endurance runners. Metabolism, Volume 65, Issue 3, March 2016, Pages 100-110.).
Here’s Van Wilder’s %VO2max vs Fat and Carbs Oxidation rates (in kcal/min).
At his peak he’s burning somewhere around 12 kCals a minute of fat. At the end Van Wilder is burning more than 22 kCals a minute of carbohydrates. His point where he’s burning 50% fat-50% carbs is at 80% of his VO2max. His peak fat oxidation is around 58% of VO2max. However, the point where he start burning carbs is relatively low at 45% of VO2max.
Compare his data with mine. I am fat adapted and only eat keto/low carb. I’m also 23 years older and not as trained by a long stretch.
Here’s the differences:
|Parameter||Van Wilder||Doug LCS|
|Max Fat Oxid.
|Max CHO Oxid.
Max Fat Oxid
Max CHO Oxid
50% Fat Oxid
50% CHO Oxid
Max Fat Oxid
Zero CHO Oxid
I would like to suggest that the main difference is found in the last row. My rate at which is expend no carbs and burn the most fat is at about 59% of my VO2max. Van Wilder’s point is about 43% of his VO2max.
Now my VO2max at 34.1 is significantly lower than Van Wilder’s at 59.8. And that probably explains a lot of the difference above. Van Wilder still looks to me like an efficient fat burner. Especially when compared to Damian Stoy.
Competition Fueling Strategies for Van Wilder?
What should the fueling strategy be for Van Wilder? Seems like he currently has the advantage of metabolic flexibility. He can certainly use fat for fuel at lower intensities.
Vespa has a graph on their site that shows %VO2max vs Fat oxidation in Low Carb and High Carb athletes from the FASTER study (Fat Adaptation: The Emerging Science from FASTER). Here’s the chart as it appears on the Vespa site:
I can’t find this graph in the FASTER Study paper (Jeff S. Volek, Daniel J. Freidenreich, Catherine Saenz, Laura J. Kunces, Brent C. Creighton, Jenna M. Bartley, Patrick M. Davitt, Colleen X. Munoz, Jeffrey M. Anderson, Carl M. Maresh, Elaine C. Lee, Mark D. Schuenke, Giselle Aerni, William J. Kraemer, Stephen D. Phinney. Metabolic characteristics of keto-adapted ultra-endurance runners. Metabolism, Volume 65, Issue 3, March 2016, Pages 100-110.).
But I do have some of the VO2max data tests from two of the athletes; Ben Greenfield and Damian Stoy. And Ben was LCD and Damian was HCD. So we should be able to check the graph using their data.
Here is Ben’s curve:
Here is Damian’s curve:
Damian’s peak rate of fat oxidation at around 0.35 g/min was about one-third of Ben’s top rate of around 1.1 g/min. So in this regard the curves do match the relative magnitudes in the Vespa graph.
The VESPA graph for the LCD vs the HCD shows a shift to the right for the peak fat oxidation for LCD as compared with HCD. In fact, the VESPA graph shows the peak of the LCD at 70% of VO2max and shows the peak of the HCD at 50%.
This doesn’t match Ben’s data at all. Ben’s fat oxidation peak is clearly around 55% of VO2max.
There is a small shift to the left for vegan Damian Stoy. His peak is somewhere around 45%.
I want to see the other data to see if Ben is at one end of the LC data but he pretty clearly doesn’t match the %VO2max vs maximum fat oxidation rate that the VESPA graph implies.
Why Should I Care?
I care because my own data matches Ben Greenfield’s data.
There is another slow steady heart rate training called Niko Niko. Here’s a good page describing it (NIKO NIKO PACE – THE GENTLE PATH TO SUCCESS).
The formula for heart rate is different than the Maffetone MAF heart rate. Niki Niko uses:
138 minus half your age.
So for me at 58 that would be 138 – (58/2) = 109 bpm. That’s not too far below my MAF heart rate of 112-122. It it at an interesting point on myVO2max test results as well. A heart rate of 109 actually has a higher fat oxidation rate than the extrapolated curve (where the real world data veers from the idealized curve). It is certainly a good point for fat burning.
The MAF number has shifted my performance upwards in just the past three weeks. Here’s my splits from my MAF test three weeks ago compared with my MAF 5 mile this AM. I’ve had to go from walking to a little jogging and walking mixed together.
That’s a pretty good improvement – around a minute and a half.
Ben Greenfield was a participant in the FASTER study (Jeff S. Volek, Daniel J. Freidenreich, Catherine Saenz, Laura J. Kunces, Brent C. Creighton, Jenna M. Bartley, Patrick M. Davitt, Colleen X. Munoz, Jeffrey M. Anderson, Carl M. Maresh, Elaine C. Lee, Mark D. Schuenke, Giselle Aerni, William J. Kraemer, Stephen D. Phinney. Metabolic characteristics of keto-adapted ultra-endurance runners. Metabolism, Volume 65, Issue 3, March 2016, Pages 100-110.).
Ben’s VO2max number was 61.1. The FASTER study was supposed to be performed at 64% of VO2max. For Ben that should have been 61.1 * 0.64 = 39.1. At that rate Ben would have been using 67% fat and 37% fat as his fuel for the three hour treadmill test.
But that’s not what Ben’s actual data shows.
Ben’s VO2 numbers were 35.8-37.0. Now that’s not that much different, but in this case it’s a significant difference. At 36.5 Ben is at a much different fuel mixture (80% fat and 20% carbs).
I am not suggesting there was any cheating here but the numbers really didn’t match the words of the study.
And repeated here:
There may be a clue here:
Was Ben adjusted downward? Even if he was the adjustment should have been done based on the VO2max testing. The number should have been 39.1 not 36. Here’s Ben’s VO2max testing result.
I am concerned about this difference. I’d like to know what’s up. There’s a huge difference between 2/3 Fat to 1/3 Carb and 4/5 Fat to 1/5 Carb. Especially over three hours. Especially when you end the race with low glycogen stores.
The details matter.
An interesting study put a group of endurance athletes on a Ketogenic diet and measured their performance as well as body composition changes (Zinn C, Wood M, Williden M, Chatterton S, Maunder E. Ketogenic diet benefits body composition and well-being but not performance in a pilot case study of New Zealand endurance athletes. J Int Soc Sports Nutr. 2017 Jul 12;14:22.). The study concluded:
All athletes increased their ability to utilise fat as a fuel source, including at higher exercise intensities.
Mean body weight was reduced by 4 kg ± SD 3.1 (p = 0.046; effect size (ES):0.62), and sum of 8 skinfolds by 25.9 mm ± SD 6.9; ES: 1.27; p = 0.001).
But how was their performance?
Mean time to exhaustion dropped by ~2 min (±SD 0.7; p = 0.004; ES: 0.53). Other performance outcomes showed mean reductions, with some increases or unchanged results in two individuals (VO2 Max: -1.69 ml.kg.min ± SD 3.4 (p = 0.63); peak power: -18 W ± SD 16.4 (p = 0.07), and VT2: -6 W ± SD 44.5 (p = 0.77).
Was this an adaptation problem?
Athletes reported experiencing reduced energy levels initially, followed by a return of high levels thereafter, especially during exercise, but an inability to easily undertake high intense bouts. Each athlete reported experiencing enhanced well-being, included improved recovery, improvements in skin conditions and reduced inflammation.
In the end the athletes likes the health benefits even with the performance losses.
A man with two clocks never knows what time it [really] is…
I’ve got two heart rate monitors now. One is a Polar H7 chest strap. The other is a Samsung Gear Sport Watch. They are fairly close to each other.
Here is the data from the Polar H7 chest strap as recorded by Strava.
Here is the data from the Samsung Watch as sent from Samsung Health to Strava.
The average is within one and the general trends are close.
I’ve spent a little bit of time thinking about the compatibility of MAF Heart Rate Training and weightlifting – generically termed resistance training (RT). Since the activity is relatively short duration and the heart rate isn’t past the MAF Heart Rate it seems on the surface like it would be compatible to do both.
One thing to consider is that VO2max testing is done on a treadmill which increases the speed and angle every couple of minutes. Resistance training lasts for seconds. The Rate of Perceived Exertion (RPE) of the VO2max testing isn’t all that hard until it gets towards the end of the test. The RPE of weightlifting is substantial under significant loads so using RPE as a test this would indicate that there is an issue.
My measurement for whether an activity is aerobic or anaerobic is the Respiratory Exchange Ratio (RER). RER is correlated to heart rate in the VO2max test but rarely considered in RT. There is a study which looked at RER in RT (Scott. Quantifying the Immediate Recovery Energy Expenditure of Resistance Training. The Journal of Strength and Conditioning Research · April 2011) in terms of Excess Postexercise Oxygen Consumption (EPOC). To review:
The respiratory exchange ratio (RER) is calculated as steady-state CO2 produced divided by steady-state O2 consumed and is typically defined from values of 0.70 representing total fat oxidation to 1.00 representing total glucose oxidation.
Here’s the RER data from the study for RT. Note the RER values are all well over 1.0 which indicates anaerobic exercise range.
Another interesting comment helps explain the RER values above 1.0:
During and after exercise, RER values above 1.00 are generally thought to be the result of nonrespiratory CO2 production: The bicarbonate buffering system, for example, involves the removal of hydrogen ions with concomitant CO2 production and hyperventilation blows off ‘‘extra’’ CO2. Yet a true measure of the RER is best found only when the system is in a steady state of gas exchange.
To the subject at hand:
Rapid glycolysis (as part of anaerobic metabolism) ceases when muscle contraction stops so that recovery is considered to be aerobic in nature. If this is true, both fatty acid and lactate oxidation may play a significant role in fueling the immediate energy expenditure needs of recovery. Unfortunately, substrate oxidation immediately postexercise and particularly after anaerobic-type exercise has not been studied well enough to draw specific conclusions. Because of this, it must be assumed here that when muscle contraction immediately stops, glycolysis is limited to the point where fat and lactate are the predominantly oxidized fuels.