NB: EatingAcademy is undergoing a major rebrand and redesign to bring you more high-quality longevity, self-optimization, and health-related content than ever before. Subscribe and tune in to peterattiamd.com on January 8th to learn more.

How a low carb diet affected my athletic performance (Part 4)

(« Part 3 of my personal journey: How a low carb diet reduced my risk of heart disease)

A common belief among people is that carbohydrates are “necessary” for physical exertion and any form of athletic activity. I certainly clung to that belief for the majority of my life, especially as someone focused on endurance sports. However, at least theoretically, it seemed possible that I should be able to get by utilizing far more fat than glycogen for aerobic activity, if I could only figure out how.

To test this hypothesis I did a VO2 max test before and after my final dietary intervention – ketosis. There is no shortage of places to read about this test, but I’ve included a figure, below, showing you what it looks like, and what it measures. Parenthetically, this test is only slightly more pleasant than a root canal without anesthetic (I’ve had this done once), as you are forced to exert yourself to the point of complete failure. Typically, it’s either done with you running on a treadmill or riding your bike on a stationary trainer, as I’m doing below. You wear a mask that excludes any inflow or outflow of air from your nose and a pair of sensors measure the exact concentration of oxygen and carbon dioxide leaving and entering your mouth. From this, a computer calculates VO2, the minute volume of oxygen – that is, the actual volume of oxygen your body is processing. This is typically recorded in milliliters per minute (mL/min). As a general rule, every 1,000 mL of oxygen processed requires about 5 calories of energy expenditure. Hence, a VO2 of 2,500 mL/min of oxygen, requires about 12.5 calories per minute, or 750 calories per hour.

During this test, a number of other physiologic parameters are measured: heart rate, power output, and lactic acid levels in the blood (hence, the blood stained towel in the picture below hanging over the handlebars from the constant finger sticks to get blood).

Peter Attia VO2 max test

All of this data is important to get an overall picture of performance. Let me start with the one folks are probably least familiar with – the respiratory quotient.

The respiratory quotient, or RQ, is simply the ratio of carbon dioxide you produce to oxygen you consume.  This ratio typically varies from about 0.70 to just over 1.00.  Why it’s important and helpful to know this ratio is that you can infer, based on the ratio, what you are utilizing for energy (i.e., how much fat versus glycogen) at any point in time during the test.  The figure below shows this relationship:

  • When RQ is 0.70 (i.e., when you breathing out 70% as much carbon dioxide as you are consuming oxygen), you are effectively getting all of your energy from fat.
  • When RQ is 1.00 (i.e., when you are breathing out an equal amount of carbon dioxide to the amount of oxygen you consume), you are effectively getting all of your energy from glycogen.

The goal of any endurance athlete is to derive as much energy as possible come from fat, rather than glycogen, for a given level of exertion.  Why?  I’ll be doing a series on this in great detail in the next few months, but for now consider this: we can store about 1,200 to 1,600 calories worth of glycogen versus 100,000 calories of fat.  Furthermore, replacing glycogen during training/competing is full of problems.  Hence, you want to “spare” glycogen for only those times when it is essential (i.e., when you are anaerobically active) and use fat as much as possible when you can afford to (i.e., when you are aerobically active).

Respiratory quotient explained

So how did my performance change over the final 12 weeks of my nutritional experiment, when I was in a state of ketosis?  The table below summarizes my findings, but let me add a bit of commentary.

summary of low carb diet on athletic performance

Aerobic base– defined as the point at which you transition to more than 50% of your energy being derived from glycogen instead of fat.  The higher this number (i.e., the higher the level of exertion) the better, because it means you can “spare” more glycogen for when you need it, and use as much fat as possible energy.

  1. Prior to ketosis – I hit this point at a heart rate of 104 and a VO2 of 1,630 mL/min (about 500 calories per hour of energy consumption).
  2. Post ketosis – I hit this point at a heart rate of 162 and a VO2 of 3,690 mL/min (about 1,100 calories per hour of energy consumption).

Implication: My aerobic efficiency improved dramatically (in the words of the person who tested me, “Like nothing I’ve ever seen before.”)


60% peak VO2– this is roughly the highest level of energy output a well-conditioned person can sustain for several hours.  I call this “all-day speed.”  When I’m doing a very long swim or bike ride (say, north of 4 hours), this is the maximum average power output I can sustain.  In me this corresponds to about 2,500 mL/min of oxygen consumption, which requires about 750 calories per hour

  1. Prior to ketosis – At this level of energy output, I required 95% of my energy to come from glycogen, and was only able to get 5% of my energy from fat.
  2. Post ketosis – At this same level of energy output, I was now only 22% dependent on glycogen, and therefore able to get nearly 80% of my energy from fat.

Implication: Another example of improved aerobic efficiency, illustrating that I can now rely on much more fat, rather than glycogen, during prolonged exertion.  This frees me up from needing to be constantly eating on long swims and bike rides.


Anaerobic threshold (AT)– as we measure it, this is the point at which your body starts to accumulate lactic acid faster than it can metabolize, or clear, it.  We use this as a pretty good (but not perfect) approximation for when your body transitions from being aerobic (able to process fat or glycogen in an oxygen-rich cellular environment) to being anaerobic (only able to process glycogen in an oxygen-poor cellular environment).  Aerobic metabolism is much more efficient than anaerobic metabolism, hence you want this threshold to be as high as possible, and ideally you want this point to be determined by lactate generation, and not substrate cross-over (i.e., inability to burn fat).

  1. Prior to ketosis – Occurred at VO2 of 3,100 mL/min (about 930 calories per hour) in a state of complete glycogen dependence.
  2. Post ketosis – Occurred at VO2 of 3,800 mL/min (about 1,140 calories per hour) in a state of 70% glycogen dependence.

Implication: Most people at the point of AT are entirely dependent on glycogen (as I was).  It’s not entirely clear to me why AT increased (a good thing), let alone why I still had some ability to use fat (30%), though some have suggested that the presence of ketones in the blood may blunt the effects of acidosis.


Max VO2– this is where you fall off the bike or treadmill.  It’s the last bit of what we refer to as “anaerobic cap” performance.  You can only sustain it for fraction of time, but it’s a 100% glycogen-dependent state of maximum output.

  1. Prior to ketosis – Hit this level at VO2 of 4,960 mL/min (or about 63 mL/min/kg, which is how folks typically describe max VO2.
  2. Post ketosis – Hit this level at VO2 of 4,350 mL/min (or about 56 mL/min/kg)

Implication: The one drawback, it seems, to completely eliminating carbohydrates from my diet was a loss of all-out top end power.   For someone like me, this doesn’t seem to hinder performance too much, but if I was trying to win an Olympic gold medal in the 400 meter run or the 100 meter freestyle, it seems I’d be better off with some carbohydrate in my diet.

So what did I learn?  Keto-adaptation made me far more metabolically flexible and efficient in the aerobic environment.  This seems particularly important for folks who compete in events longer than a few minutes (e.g., 10K, marathon, triathlon), but less so for folks doing short-burst activity.

The real question is how can you get the best of both worlds?  That is, is there a way to reap the benefits of keto-adaptation of on the aerobic side, without any of the anaerobic cap costs?

In short, I believe the answer is yes, and I look forward to writing about this in great detail in the near future.