Estimating glycogen usage from power meter data

In trying to manage my diet and training plan, I've found the "kj = calories" rule of thumb to be pretty helpful. The problem is knowing how much of that energy needs to be replaced in the form of carbs.

Is there another estimate for the grams of muscle glycogen consumed on a given ride? My gut feeling is that this would be a factor of kJ, intensity factor, time spent in given zones, etc. but I have yet to find any good research on this. One interesting theory is that TSS ~= grams of glycogen but i'm not sure if there are any numbers to back that up.

Obviously this would vary between individuals but it would be helpful to scale your post-ride recovery meals by having a reasonable estimate of your glycogen deficit.

Any ideas?

I went down this path a few years ago and started looking at typical substrate utilization as a function of intensity curves. If you dig a bit you can come up with published data for glycogen vs. percentage of VO2 Max for instance. But it's very hard to tie that to percentage of power at a particular reference point (VO2 Max, MAP, etc.) and it varies between individuals and even for the same individual depending on training.

If you take the whole concept further through ideal refueling, glycogen resynthesis rates and RMR/BMI concepts it turns out not to matter very much.

Here's the problem, say you could accurately determine based on AP/NP or IF that you burned 1000 Calories of which 600 were from your stored muscle glycogen and blood/liver glucose. Naturally you'd want to replenish that glycogen and not much more if you're interested in weight loss but want to train on subsequent days. Trouble is you can't ingest 600 Calories of glucose and immediately resynthesize and store it as muscle glycogen. Even during the peak glycogen resynthesis phase immediately following exercise (the critical half hour) you can only resynthesize about half that amount of glycogen.

So you make a giant smoothie with the ideal (debabtable exactly what constitutes ideal btw) ratio of carbs to protein and try to get exactly 600 Calories of carbohydrate on board. Well half or more of that will be stored as fat since you've exceeded peak glycogen resynthesis rates.

Well it turns out that's not a big deal either. In the end your body requires energy even at rest and it still comes down to calories in vs calories out. So if your net caloric balance is negative you'll burn through those carbs you ingested that converted to fats. But you won't have replenished all the glycogen you burned on your ride.

Sure I guess you could come up with detailed schedules that attempted to restore just enough glycogen over your complete day and multiple meals and somehow estimate the glycogen burn rate during rest and daily activities to add to that burned during exercise. But then what, measure every meal, time it just right and hope your estimates of resting glycogen/glucose requirements are accurate?

In the end it makes sense to refuel with 200 to 300 Calories of healthy carbs combined with some protein (3:1, 4:1, 2:1 take your pick, the research isn't conclusive) immediately after exercise since that's about all you can process at once (and why the standard serving sizes for Endurox, Metabolol, Recoverite, etc. are all about the same). Follow it up with a healthy meal or another recovery drink 45 minutes to an hour later and eat a balanced diet including plenty of healthy carbs throughout the day.

And most of all, maintain an appropriate daily caloric balance to gain, lose or maintain weight. You don't have to exactly dial in the macronutrients to make sure you store just the right amount of glycogen. You do have to make sure you include plenty of healthy carbs across the day and the critical half hour refueling is key but it gets real hard to dial it in a lot more closely than that outside of a controlled lab environment.

-Dave

PS, There could be a relationship between TSS and glycogen consumed but it's certainly not a 1:1 relationship.

Take a rider with 300 watt FTP who rides a full hour FTP test to generate 100 TSS in an hour. That's 1080 kJ of energy consumed in an hour at relatively high intensity. Even taking a lowball swag at glycogen utilization being 75% at FTP (I'd guess much higher for most athletes) the rider would burn through more than 800 Calories of glycogen not 100 predicted by "glycogen = TSS"....

-Dave

Glycogen doesn't belong to the "world" of feeding as much as it belongs to the "world" of training. Like Dave said, you can't eat Glycogen. By training, we increase our ability to store more. Key word here, is training. Good news though, is that you made your way in pinpointing such an important component to performance/fatigue. We do train in order to increase our potential to store it. And by training we deplete it.

I am no expert in training with Power (although I do own a pm). But as far as I can understand the concept (TSS/CTL/ATL etc), to me, CTL/ATL impersonates (among other things):
1- your potential for storing Glycogen (CTL which translate into greater work capacity) as well as
2- your current glycogen level (ATL, in a reverse relationship of course). That is, the higher ATL, the lower the Glycogen levels (or so I would expect).

My goal here isn't to oversimplify this wonderful concept by narrowing it down to just the Glycogen factor (there's more to it, Central and peripheral nervous system fatigue, muscle tissue damage rebuild etc). But I believe that this concept encapsulate the glycogen matter so well that there's no need to follow (in parallel) glycogen levels, even if it was indeed possible. Same thing with Lactate levels during workouts or any other factor that contributes to fatigue. This, I find, is yet an other occurrence of "Best predictor for performance, is performance itself" (A.Coggan)

So keep going with your quest to try to understand relations between nutrition and performance, follow Dave's advices, and trust PMC to monitor your Glyco levels.
- - - - -
here's little food for your thoughts: "Athletes can become nearly depleted of glycogen after one or two training sessions or by long, intense training sessions of 1hr or more (Houston 1978; Beltz et al. 1988). "Conclusive proof exists that successive days of intense training can almost completely deplete glycogen (Costill et al. 1988). Finally, Coggan wrote a huge paper on sugar (if I am not mistaken). I'd give this paper (which is fairly recent) a priority of you want to dig deeper into the topic.

There is (with an R^2 of 0.75).

Pretty good fit considering the range of things involved.

So what does the best fit function look like? I'm thinking it has to also be a function of absolute FTP since that impacts kJ per hour at a given AP or NP.

-Dave

I posted graphs of glycogen use vs. TSS, rate of glycogenolysis vs. IF, and and rate of glycogenolysis vs. blood lactate concentration to the wattage list earlier this year in threads titled "For your viewing pleasure" and "Two more plots":

Wattage | Google Groups

Wattage | Google Groups

Rather than attempt to reproduce them and the discussion that followed here, I'll just point you "that way!" (since I know you're on the wattage list).

Ahhh yeah, I remember at least one of those charts.

Thanks...

Guys, when you finished speaking Chinese, if you could dum your findings down for us, I'm sure I wouldn't be the only one to be interested in understanding 'em...

Well those are Andy's charts to describe and a fair amount has been hashed out on the Google Wattage Group: http://groups.google.com/group/wattage but you have to apply for the free membership to view the discussions.

But the one that intrigues me the most is the chart he's posted showing rate of glycogenolysis vs IF

Basically it shows a surprisingly good linear fit between rate of glycogenolysis in units of mmol/min/kg to (Intensity Factor)^2 which is the foundation of TSS.

The fitted line is described by:
rate of glycogenolysis = 2.6754(IF^2) - 0.9965
and its R^2 is 0.7862 showing a pretty good but not a perfect fit to the data (ideally, R^2 = 1)

But in english it means you can estimate rate of glycogen burn fairly well based on IF^2 and if you play around with units a bit can estimate glycogen burned from TSS. It also implies resting rate of glycogenolysis of negative 0.9965 mmol/min/kg if you take IF=0 which doesn't really make sense so I'd guess this model only applies for moderate to hard exercise.

But it's not as simple as Glycogen burned = TSS as the OP suggested and there's still the question of whether knowing the amount of glycogen burned on a ride with great accuracy is all that useful in terms of daily fueling strategies for the reasons listed above.

-Dave

Dave, thanks.

Was I right then in simply suggesting to OP to rely on performance management chart for this matter?

I mean, reason for this relatively good "fit" is probably because without it, the PMC wouldn't be doing that much of a good job would it?

(BTW, whenever I issue doubtful or wrong recommendations, please do advise. I don't bite ;-)

Well, the OP was asking how to estimate glycogen utilization for the purpose of refueling for weight management purposes. From that standpoint I wouldn't suggest basing refueling strategies on the PMC.

Trouble is CTL represents a weighted rolling average of roughly your last 3 to 4 months of daily training load and ATL represents roughly your last 3 to 4 weeks with the default time constants. So I wouldn't plan day to day caloric intake based on a 3 week or 3 month rolling average.

The PMC does a great job of modeling long term training load with CTL or in effect the long term average daily training load you've sucessfully adapted to. The sucessful adaptation is implied because it's hard to overtrain continuously for nearly 4 months so if you achieve a certain CTL then by definition you've been working at that average daily load for quite some time. It also does a good job of modeling short term or recent training load via ATL and predicting immediate fatigue vs. freshness by taking the difference between the two. But none of that is immediate enough to guide daily refueling strategies.

Actual energy in kj burned on the ride gets you much closer in terms of estimating immediate refueling needs. The graphs above imply that you could use TSS or IF to estimate the sugars that you burned but again in the end what really matters is total calories from a broad based and healthy diet and as long as that diet doesn't overly restrict carbs or protein there's not much value in trying to dial it in more accurately.

-Dave

Negative glycogenolysis at rest is not clearly wrong: at rest you are always resynthesizing glycogen at some non-zero rate, unless stores are 100% topped off.

I can understand zero glycogenolisis at rest, but negative? Would glycogen resynthesis actually represent itself as negative glycogenolisis?

'Guess it could but it hits me as strange that this data fit could predict both glycogen utilization and glycogen resynthesis based on the model's intercept as they're very different processes. Looks to me like a modeling artifact rather than part of a greater glycogen utilization-resynthesis model. But as I said above, I'm just eyeballing a couple of curves and that's what hit me, I don't claim to know the limits of those graphs or even the major points demonstrated by those curves.

-Dave

"Negative" glycolysis, glycogen resynthesis: they're the same process, putting glycogen back in the muscles. And yes, if you're at rest, and have a glycogen deficit, then the glycogen is getting put back at some non-zero rate. Or am I missing something?