April 2016 update: The blog is not dead, I’ve just been very busy!

In today's blog, this and more: 

Moving countries (from Australia to Norway),

Low CHO, High-Fat diets, 

Recent Scientific publication update, 

Future posts: Weight loss for cyclists & prolonged cycling metabolism

It’s been about a year since my last blog-post and so there is a long overdue update for the audience of the blog. 

This post will be dived into the following sections, jump to whichever one you find relevant/interesting:

1)  What happened in 2015? why such big gap between posts? à Moving countries 

2)   Low Carbohydrate, High Fat (LCHF) diets post in Cycling-tips in 2015.

3)   Carbohydrate dependence during prolonged time-trial scientific publication.

4)   Prospective posts for 2016à Weight loss for cyclists and metabolism during very prolonged (>12 h) cycling.

1) What happened in 2015? why such big gap between posts? à Moving countries 

I was appointed post-doctoral research fellow at the Norwegian School of Sport Sciences and we moved with Kat here to Oslo from Melbourne in June 2015, and it's been very hard to keep up with the blog.

This position is a fantastic opportunity to develop in my academic career and we are very excited of starting this new chapter in our life. As with all major changes in one’s life, it is full of challenges, but I see it as opportunities for growth.

As you can imagine (or maybe you can’t) changing countries and starting a new research position can be quite demanding as there are many new things to adjust to. Additionally, and as part of the venture, I’ve been working on several research projects in parallel, which depleted my ‘functional reserve’ (i.e. that extra energy needed to write a blog). These projects incorporated, among other things, travelling to the August Krogh institute in Denmark to analyse human muscle samples for a study on post-exercise protein intake for recovery after cycling, plan and set-up experimental studies and work on a meta-analysis and many other things.

Also throw on top of that coaching athletes and training when I’m not working at the school, and even some attempt at learning some Norwegian (I’m still very bad at it!).

Currently, despite I’m far from being ‘not busy’ things are slightly more settled and I expect I will be able to write more often.

Figure 1, A, B and C. Aerial photo of the Norwegian School of Sport Science and surrounding areas: beautiful, isn’t it? Photo credits: www.nih.no.

2) Low Carbohydrate, High Fat (LCHF) diets post in Cycling-tips in 2015.

The last post on this blog was in March 8^th 2015, but I did write one –quite long- post for Cycling tips on LCHF diets and cycling which got published in early June 2015.

I should give credit where it is due and thank Ken (now in HPTek) in Melbourne, for suggesting me to write the article and for putting me in touch with Matt de Neef (one of the editor of cycling tips), who embraced my criticism to a former article on their page (if you're reading this, thank you guys!).

The article is about a year old, but you might still find it interesting. If you like the topic I would recommend you read the review on this topic published by Louise Burke that came out in November 2015. The review article is excellent, it adds more depth to what I described in that blog article written earlier that same year. Here is a link to Louise's article, which is free.

3) Carbohydrate dependence during laboratory-based time-trials scientific publication.

In early 2016 our publication titled “Carbohydrate dependence during prolonged simulated cycling time trials” was published on European Journal of Applied Physiology.

Here is a link where you can find the article.

I hope it is understandable to the general audience (sadly most likely not considering all the scientific formatting and jargon within), but it is still part of my plans to explain it in more lay terms in this same blog.

4) Prospective posts for 2016. A) Weights loss for cyclists and B) Power-output and metabolism of very prolonged exercise.

They say one picture is worth a thousand words. So I’ll let this figure do the job:

Figure 2 – Weekly average body mass vs weekly net calorie balance over the course of 3 months.

This is a sneak-peek of what the post will be about, which is basically about weight loss for cyclists. Provided there is not a LOT of information on the scientific literature about this, let alone with tight exercise and diet control/record, and I can't run a full-blown scientific study on this at the minute I decided to run an experiment using my experimental subject by default (myself). It aims at showing the power exercise nutrition and to document changes in body mass and power-output (absolute and W/kg) over the course of a preparatory phase of training when coupled with an planned and optimised nutritional intervention (and it is also a great excuse to get in shape for the races coming up!).

The experiment (or I should call it intervention) is happening as I write this (body mass is currently down to ~75.5 kg), and I’ve just killed a full bowl of banana split with vanilla ice-cream  and Nutella:

Figure 3 – My Sunday dessert during weight loss: It’s not about starving. It’s not about low carbs either.

As it is happening while I write, it is going to take me another while until it is finished and I can publish it here. Additionally there is going to be a LOT of data and numbers that will need to be crunched (today is 90 days of uninterrupted daily food and exercise data logging), so it is likely going to be published as a series of posts rather in a one long post.

I’m aiming at shedding some light on this topic by including unpublished data on very prolonged exercise in the lab together with other insights on the topic. This will be happening most likely after the post on weight loss.

That is all for today, stay tuned for the blog-posts coming up!

Using a power-meter to estimate energy expenditure. Are you getting the right information?

Overview

This Blog-post is about how you can use a power-meter to estimate energy expenditure (a.k.a. calories burnt) during a bike ride. The main aim is to debunk some misconceptions and to shed light on the link between mechanical work and metabolic energy expenditure.

The core of this post will be broken down in three sections.

1.   The link between mechanical work and metabolic energy expenditure. (THEORY)

2.  The use of power-meter data to estimate energy expenditure, differentiating between kilojoules of work and kilocalories of energy expenditure. (PRACTICE)

3.   How your Garmin cycling-computer might be calculating your energy expenditure. (ANALYSIS)

Introduction

One of the several uses of a power-meter is to estimate energy expenditure. While the information derived from power meters makes easy to measure the amount of mechanical work, the relationship between mechanical work and the energy expenditure is not obvious.

This information is particularly important when integrating power-meter derived information onto different nutritional (refuelling or weight loss/control) practices.  

The aim of this post is to debunk some misconceptions and define some terms and concepts necessary to better understand energy expenditure. This post will be followed by a post on weight loss, performance and health.

1.       The link between mechanical work and metabolic energy expenditure

Turning over the pedals of the bike (mechanical work) is possible thanks to the biochemical events that happen inside the skeletal muscle to generate energy (metabolic energy). Consequently, metabolic energy is ultimately turned into mechanical work, but a part of the metabolic energy is turned into heat (‘wasted energy’).

Efficiency is an important concept that is used to indicate what percentage of the metabolic energy is turned into actual work and what percentage is turn into heat.

A process that is 100% efficient generates no heat. To understand this better I like to think about lamps and cyclists (Figure 1).

Figure 1. A schematic representation of 3 different systems with different efficiencies. A tungsten lamp (2.5 % efficiency), a LED lamp (15% efficiency) and a cyclist (21% efficiency).

The input of energy for lamps is electricity, tungsten lamps are probably the least efficient lamps in the market at the moment and they generate a lot of heat (I dare you to hold a tungsten light-bulb with your bare hand to test this). LED lamps instead are more efficient (this varies hugely depending on the type of LED lamp) and generate less heat (you can touch them without getting burnt).

In a way, a cyclist can be interpreted as the light-bulbs. To some extent we can say that the metabolic energy generated is analogous to electricity, turning over the pedals is analogous to the light generated by the light-bulbs and heat is heat.*

Power-meters measure mechanical work and not metabolic energy expenditure, therefore to calculate energy expenditure during a ride, it is necessary to know how efficient a cyclist is.

2.       Using a power-meter to estimate metabolic energy expenditure

Now that you understand the relationship between metabolic energy, heat and mechanical work, and the concept of efficiency it is time to jump into the practical application of this.

As you well know power meters measure power, mechanical power, but what we want to know is the metabolic energy to generate mechanical work. Power is the amount of work done (or energy used) in a unit of time. By convention it is measured in Watts (W) which is Joules per second (1 W = 1 Joule/s). So, by knowing the amount of power held for a given amount of time you can know the amount of mechanical work completed.

Example: Frank Lampard rides his bike for 1 h (or 3600 seconds) at an average power of 150 W (or 150 joules/s). Frank has done 540000 Joules (or 540 kilojoules) of mechanical work.

However, as explained above, mechanical work is not equivalent to metabolic energy. To know how much metabolic energy (derived from fat and carbohydrates metabolism) has been used to do this amount of mechanical work it is necessary to know the efficiency of the process. Efficiency varies between people. Therefore it needs to be measured for each cyclist, and it requires specialized equipment (an ergometer and indirect calorimetry metabolic cart). Thankfully efficiency is fairly constant in standard conditions and it has been shown to be in average 20.7% (range 18.3-22.6%) for well-trained males [1] and 19 % for very well-trained females [2]. Therefore, despite measuring is the best way to accurately make any estimations, it can be assumed that you fall somewhere close to these average values. This makes things easy for estimating the actual metabolic energy spent on a ride.

Example: Frank Lampard did 540 kJ of mechanical work during his 1 h ride. How much metabolic energy did Frank have to generate for his ride? Assuming Frank’s efficiency is 20.7%, he used around 2609 kJ of metabolic energy (540 x [100/20.7]).

In short, assuming an efficiency of 20.7% you can estimate metabolic energy expenditure by multiplying mechanical work by 4.8 (100/20.7).

The seemingly simple outcome of this simple calculation can become very confusing when taken into consideration that different units, calories and joules (or kilocalories and kilojoules), are used to refer to energy expenditure.

A problem with terminology and units, Joules and calories:

The calculations above have been done using Joules. Now what is the role of the calories here? Calories are another unit to measure energy. So joules (J) or calories are interchangeable, 1 calorie is 4.184 J. It is analogous to comparing kilometres to miles. Joules (like meters and Watts), belong to the International System of Units (SI), and calories don’t.

On the example above Frank’s metabolic energy expenditure is 2609 kJ. Therefore, this amount is equivalent to 623 kcal. However this can be confusing as this amount is remarkably close to the 540 kJ of mechanical work that Frank did (Figure 2, top panel). This can so confusing that some manufacturers of cycling computers seem to not have quite understood it (see below).

Many people think or assume that the value obtained in kJ from the power-meter computer is the amount of ‘calories burnt’. It is easy to say that the value of metabolic energy expenditure can be calculated by changing the unit of the value of mechanical work from kJ to kcal, but this leads to incorrect results and it is not correct conceptually either (Figure 2).

Figure 2. Two ways of estimating metabolic energy expenditure, the right way and the wrong way.

3.       How your Garmin cycling-computer might be calculating your energy expenditure

Most cycling computers provide an estimate of energy expenditure based on different measures. As discussed above, using information coming from the power-meter meter is one way to do it and a fairly straightforward one. However this is not the only way in which cycling computers estimate energy expenditure. Moreover, cycling computers can be wrong and provide misleading information.

Before I proceed I must make clear that the information of this section applies to Garmin Edge 800 units (the one I own and used to get the information for this post). You should check how your brand and unit model works before you generalize what I describe here. I do not hold anything against Garmin or firstbeat companies and it is not the point to denigrate them but to share my findings about how their products work according to my findings. Any comments are expected to be taken as constructive criticism.

On my garmin unit I can set the screen to display Calories and/or kilojoules. While ‘kilojoules’ indicates the amount of mechanical work, ‘Calories’ is supposed to indicate the metabolic energy expenditure.  Let’s jump straight into the problem and see what the computer shows after two different rides (Figure 3).

Figure 3. Mechanical work (Kilojoules) –green boxes- and Energy expenditure (Calories) –red boxes- displayed on my Garmin Edge 800 unit when used after a short ride with (left) and a long ride without (right) a heart rate (HR) strap.

So we can see that there are two problems here:

1) When used with a HR strap it underestimates energy expenditure (my estimation says it should be somewhere around 1732 kcal, 44% higher).

2) When used without a HR strap it also underestimates energy expenditure (my estimation says it should be somewhere around 4209 kcal, 13% higher).

While it is clear what it is doing wrong in scenario #2 (it simply converts the value of mechanical work (in kilojoules) into energy expenditure (Calories) by just changing the unit –as I what I say in Figure 2 is the wrong way of doing it-), it is not so clear what it is doing in scenario #1.

To unveil this it is necessary to look under the hood and see how the computer operates (Figure 4).  

Figure 4. Schematic figure of how my Garmin Edge 800 seems to estimate energy expenditure.

To estimate energy expenditure the computer prioritizes the information coming from the heart rate strap, when available, over the information coming from the power-meter. The computer uses an estimation method developed by a company from Finland called firstbeat (http://www.firstbeat.com/). This method integrates different variables that are entered in the computer (age, gender, activity level, etc) and using some sort of algorithm that integrates heart rate and heart rate variability, it estimates what is your actual energy expenditure. In my case the value that comes from the heart-rate based estimation is about 44% off from my power-meter and efficiency based estimation. This algorithm has not seem to be properly scientifically validated through the peer review process. There is some information about how this algorithm works on firstbeat’s webpage ‘white papers’, but these are not scientific articles, therefore it is not science.

In my opinion these head units would provide a much more accurate estimation of energy expenditure if they

a)    Prioritized the information coming from the power meter over that of the heart rate strap.

b)    Assumed a standard value of efficiency or provided the chance of entering the rider’s efficiency into the computer.

c)     Properly reported the energy expenditure based on that efficiency value.

Conclusion

Estimating metabolic energy expenditure using a power meter requires understanding the concept of efficiency (Figure 1).

Provided the common use of joules and calories to measure energy, the output of some calculations can be confusing and it is important to follow the right logical steps to make this calculation (Figure 2).

Some cycling computers on the market provide misleading information regarding metabolic energy expenditure (Figures 3 and 4).

I hope this post has made this topic clearer to you and now when you look at your cycling computer in search for the kilojoules or the Calories display you ask yourself:

•    What is this value on the screen, mechanical work or metabolic energy expenditure?
•     Is it derived from my power-meter or my heart-rate?
•    If derived from my power-meter, is it assuming any specific efficiency value for me or is it just mistakenly changing the units from joules to calories?
•    If derived from my heart-rate, is it accurate? Has it been validated?

References:

1.            Coyle, E.F., L.S. Sidossis, J.F. Horowitz, and J.D. Beltz, Cycling efficiency is related to the percentage of Type I muscle fibers. Medicine & Science in Sports & Exercise, 1992. 24(7):782.

2.            Haakonssen, E.C., D.T. Martin, L.M. Burke, and D.G. Jenkins, Energy expenditure of constant-and variable-intensity cycling: power meter estimates. Medicine and science in sports and exercise, 2013. 45(9):1833.

Further reading:

• Ettema G, Loras HW. Efficiency in cycling: a review. Eur J Appl Physiol. 2009;106(1):1–14  

---

* This is a simplification because for the lamp what it generates the energy (the power station) is not part of what uses that energy. In case of a human the ‘power station’ (mainly the mitochondria) is on the same place that uses that energy, the skeletal muscle.