Modelling Carbohydrate Requirements During Exercise: An Overview of The EatMyRide Approach

Exercising requires energy. The two main sources of energy are fat and carbohydrates. Understanding and modelling the fat and carbohydrate requirements during exercise is important for a proper nutrition plan.

The basis

Two physiological concepts form the basis of nutrition intake during exercise:

1. Carbohydrate burn: When exercise intensity increases, the share of energy that comes from carbohydrates increases and hence the relative amount of fat burn decreases. Additionally, since an increase in exercise intensity also means an increase in energy burn, the absolute carbohydrate burn increases as well. Carbohydrate burn usually increases more than linear, i.e. when the absolute intensity doubles, the carbohydrate burn will be more than doubled.
2. Carbohydrate availability: The limiting factor in energy supply is the amount of carbohydrates available, rather than the amount of fats available. Usually the body contains enough fat to fuel days of exercise, while carbohydrate stores can be depleted within 1,5 – 2 hours.

To get to a personalized nutrition plan, these two concepts need to be quantified. When both the carbohydrate burn and the amount of available carbohydrates are estimated, the minimum required carbohydrate intake is simply the difference between these two.

Modelling carbohydrate burn

The amount of fat and carbohydrate burn – also known as substrate utilization- has been studied by a number of researchers in the past decades. Romeijn et al. (1993) provided a very useful framework. They studied the role of four energy systems at three different exercise intensities: 25, 65 and 85% of VO2Max. When modelling the carbohydrate burn as function of the relative exercise intensity (%VO2Max) based on the numbers provided by the authors, the relationship is approximately quadratic.

In the years after 1993 there has been a lot more research into exercise and substrate utilization. For example, different authors studied the concept of FatMax and the substrate utilization at anaerobic threshold (e.g. Asker and Jeukendrup, 2003; Purdom et al., 2018; and Peric et al., 2016). However, in general those articles don’t provide a holistic framework with the substrate utilization across different exercise intensities. That makes those results harder to use in a model that estimates the substrate utilization based on the relative exercise intensity and does so across all intensities.

Several authors sticked to the results from Romeijn et al. (1993) to model the carbohydrate burn of athletes (Rapoport, 2010; Van Dijk et al., 2017). EatMyRide does it similarly, but with a slight adaptation. Based on the more recent insights that FatMin (point at which fat burn becomes negligible) and anaerobic threshold usually correlate (Peric et al., 2016), we estimate the relative fat burn at 85% VO2Max to be lower than Romeijn et al. (1993) indicate.

To use the model in practice there is once step left. The power data of the exercise should be translated to relative exercise intensities in terms of percentage of VO2Max. When ones’ VO2Max is known this translation can be made. Rapoport (2010) provides the mathematical relationship between both, under a few assumptions.

In order to use the results for workouts based on heart rate rather than power, the substrate utilization per heart rate zone has to be estimated. This is done by first calculating the power zones based on the athlete’s Functional Threshold Power and calculating the average substrate utilization per power zone. Assuming that each power zone on average corresponds to a heart rate zone, these average carbohydrate and fat burn values per power zone constitute the burn per heart rate zone as well.

Modelling carbohydrate availability

Carbohydrates are stored as glycogen in muscles and the liver, with the muscles containing the largest part. The amount of glycogen per kilogram muscles is approximately 150 mmol (equaling roughly 2% of the muscles’ weight) after an overnight rest and proper nutrition with enough carbohydrates (Sherman et al., 1981). With a carbohydrate loading regime glycogen levels can increase to approximately 200 mmol. Research has also indicated that below 70 mmol / kg muscle mass there is a risk of impaired exercise performance (Schweitzer et al., 2017).

To estimate the amount of available carbohydrates on a personalized level, these benchmark values are taken as starting point. For each athlete we estimate the muscle mass using the formulas provided by Heymsfield et al. (2020). Muscle glycogen is only available to be used locally. So, the total muscle mass is multiplied by a factor to account for the fact that not all muscles are used during sports like cycling. The total available glycogen stores are then simply the amount of active muscle mass multiplied by the amount of glycogen per kg muscle mass. For the first sports activity on a day we assume that the athlete has properly fueled before, so the glycogen level is estimated to be 150 mmol per kg muscle mass. If the athlete indicates he or she has followed a carb loading regime, the glycogen level is assumed to be 200 mmol.

When an athlete has already done a sports activity on the day and starts a second activity, the starting glycogen level is lower: during the exercise it decreases approximately with an amount equal to the carbohydrate burn minus the carbohydrate intake. After exercise glycogen is resynthesized however. Assuming that one follows the carbohydrate intake advice after exercise, glycogen resynthesizes at a rate of 10 mmol / kg wet muscle weight per hour during the first four hours and 4-6 mmol afterwards (Burke et al., 2017; Ivy, 2013). So for any activity after the first activity on the day the glycogen level can be estimated based on the glycogen level at the end of the previous activity and the time between the activities.

From carbohydrate burn and availability to required intake

Using the two models described above the minimum required carbohydrate intake can be estimated. First, the total carbohydrate burn of the exercise is calculated, by applying the carbohydrate burn model to the power or heart rate data of the sports activity. Next, the amount of available glycogen is the difference between the glycogen starting level (calculated using the model described above) and a threshold value of 70 mmol glycogen / kg muscle mass. In the Energy Level graph in the app the red zone indicates the zone below 70 mmol.

The difference between the total carbohydrate burn and the available glycogen is the minimum amount of carbohydrates that one should take during exercise.

The importance of proper nutrition and timing

The carbohydrate burn usually is not constant over the duration of exercise, due to varying exercise intensity. Also the rate at which carbohydrates in nutrition products are absorbed vary between products. Since glucose and fructose are oxidated by different pathways in the body, the amount and ratio of different carbohydrates are important. EatMyRide takes these factors into account and notifies the user if a product is not optimal for a given sports activity.

Using your own metabolic data

The approach outlined in the previous sections forms a useful method to create personalized nutrition plans and get insights into burn and intake. However, it is widely acknowledged that there is a lot of interindividual variance in substrate utilization that cannot be solely explained by differences in fitness. One of the most used solutions to know your own metabolic profile is the INSCYD test. The INSCYD algorithm provides an estimation of your metabolic profile based on a few exercise tests you perform. The results of this test can also be entered into the EatMyRide app, to tailor the calculations even more to your personal profile.

References

• Achten, J., & Jeukendrup, A. E. (2003). Maximal fat oxidation during exercise in trained men. International journal of sports medicine, 24(08), 603-608.
• Burke, L. M., van Loon, L. J., & Hawley, J. A. (2017). Postexercise muscle glycogen resynthesis in humans. Journal of applied physiology.
• Heymsfield, S. B., Stanley, A., Pietrobelli, A., & Heo, M. (2020). Simple skeletal muscle mass estimation formulas: what we can learn from them. Frontiers in Endocrinology, 11, 31.
• Ivy, J. L. (2013). The regulation and synthesis of muscle glycogen by means of nutrient intervention. The Encyclopaedia of Sports Medicine: An IOC Medical Commission Publication, 19, 113-125.
• Peric, R., Meucci, M., & Nikolovski, Z. (2016). Fat utilization during high-intensity exercise: When does it end?. Sports Medicine-Open, 2, 1-6.
• Purdom, T., Kravitz, L., Dokladny, K., & Mermier, C. (2018). Understanding the factors that effect maximal fat oxidation. Journal of the International Society of Sports Nutrition, 15(1), 3.
• Rapoport, B. I. (2010). Metabolic factors limiting performance in marathon runners. PLoS computational biology, 6(10), e1000960.
• Romijn, J. A., Coyle, E. F., Sidossis, L. S., Gastaldelli, A., Horowitz, J. F., Endert, E., & Wolfe, R. R. (1993). Regulation of endogenous fat and carbohydrate metabolism in relation to exercise intensity and duration. American Journal of Physiology-Endocrinology And Metabolism, 265(3), E380-E391.
• Schweitzer, G. G., Kearney, M. L., & Mittendorfer, B. (2017). Muscle glycogen: where did you come from, where did you go?. The Journal of Physiology, 595(9), 2771.
• Sherman, W. M., Costill, D. L., Fink, W. J., & Miller, J. M. (1981). Effect of exercise-diet manipulation on muscle glycogen and its subsequent utilization during performance. International journal of sports medicine, 2(02), 114-118.
• Van Dijk, H., Van Megen, R., & Vroemen, G. (2017). The secret of cycling: maximum performance gains through effective power metering and training analysis. Meyer & Meyer Sport.