Research Review:
Nutrition and exercise recovery - Part 1

By Helen Kollias


Checklist for winning the Tour de France: Elite genetics. A sweet ride. Looking good in yellow. Carbohydrates alone… or carbohydrates plus protein? Dr. Berardi’s research provides the answer.

This is a special research review article, because I’m reviewing our own John Berardi’s new article that just came out on Christmas Eve. In addition to the usual review, I’ll be asking John questions throughout the article to get a better idea of his thought process as he developed the study. It’s like being in the lab with Dr. Berardi!

Before we get John involved, I’ll briefly go over why carbohydrates, proteins or any combination thereof would matter to you and for your aerobic recovery.


Have you ever heard of “hitting the wall”? “Bonking”? (No not boinking — this is a family friendly blog!) Hitting the wall and bonking are what runners and cyclists, respectively, describe what happens to your body when you run out of glycogen.

Hitting the wall feels like you have completely run out of energy. If you have ever suffered from hypoglycemia, then you have some idea of what this is like. Some people even have hallucinations, dizziness and weakness. It doesn’t feel good, and you’ll definitely have a hard time finishing whatever you are doing. The average weekend warrior running a marathon normally hits the wall around 20 miles (32 km) into the 26 mile (42km) run. Hitting the wall is so much of an issue in rookie marathon runners that running experts suggest that newbies have at least one training run until they hit the wall, in order to be able to handle the feeling during a real race.

Physiologically, “hitting the wall” is similar to what happens when your car runs out of gas. The faster you drive the faster you use the gasoline, which is what happens with glycogen -– the more intense the exercise, the faster you go through glycogen.

Big deal, you say. So, you hit the wall, run out of glycogen… don’t you have fat to use? Yes and no. It turns out that if your exercise intensity is high enough (above about 50% of your max aerobic capacity) fat just can’t be converted fast enough to provide enough energy (ATP) for your muscles to use. If you’re better trained you can use more fat at a given intensity, but you still need glycogen to go as fast as you can.

Glycogen is a polysaccharide (literally, “many sugars”) made up of glucose. Think of a glucose as a Lego block and glycogen the wall you make with the Legos. The average glycogen molecule is made up of around 60,000 Legos, or glucose molecules. Your body stores glycogen primarily in your liver and your muscles.

What happens you run low on gas? You go to the gas station and fill up your car. It’s pretty much the same thing when you run out of glycogen: you need to fill up with carbohydrates. Can you make (re-synthesize) glycogen faster if you take in protein at the same time? Could you pump gas into your car — or glycogen into working muscles — faster?


Research question

This is the kind of question that JB wanted to answer in his study.

Does supplementing with a liquid carbohydrate-protein supplement improve cycling performance compared to carbohydrate alone?

Study citation:

Berardi JM, Noreen EE, Lemon PW. Recovery from a cycling time trial is enhanced with carbohydrate-protein supplementation vs. isoenergetic carbohydrate supplementation. J Int Soc Sports Nutr. 2008 Dec 24;5(1): 24.

My first question for JB is why he asked this question. It may seem odd, but when you read or hear of studies you should ask yourself: why did they decide to do this? Is this a follow-up of another study? Is this in response to someone else’s study? Is this the first study to address this question?

Question for JB: What drew you to looking at cycling performance and supplementation? You’re known more for your involvement in weight training than endurance exercise.

JB: It’s no secret I believe that recreational exercisers should emphasize a mixture of weight training and high intensity interval exercise to get into the best shape of their lives. However, my work with elite athletes does run the gamut from strength and power athletes to endurance athletes and everything in between.

In fact, right now, two of our sponsored athletes are Dede Griesbauer, America’s top triathlete, and Chandra Crawford, gold medalist in the 2006 Olympics in the sport of cross country skiing. I’ve also chased through France with the Tour de France on more than one occasion. I’m definitely a cycling fan and that helped with my decision to use a cycling model for this study.

There are also many scientific reasons for choosing cycling performance. Indeed, with the right equipment, the cycling model is one of the best for looking at nutrition and performance.

For starters, it’s easy to take blood and respiratory samples when a subject is sitting down cycling. Imagine trying to take the same samples while someone is running. Random needle stabs, respiratory head gear bouncing all over the place. Not good.

Next, with cycling we can have cyclists race against pre-recorded performances in a “virtual reality” environment. So we have the comfort and accessibility of the lab with the “feel” of real road cycling.

Finally, recovery nutrition is partly aimed at carbohydrate replenishment. And in cycling, there’s a simple relationship between glycogen content of the quadriceps and performance. We can measure carb content of the quads with and without different recovery interventions to see just how they impact glycogen and performance. With other activities, especially weight training, the relationship between glycogen in a specific muscle and performance isn’t so clear.


15 male cyclists came to the lab 3 times, twice for familiarization and once for day-long testing.

Familiarization involved the cyclists biking on a Serotta Size Cycle at either a 5% or a 7% grade (depending on cycling strength) for 60 minutes with the goal of going as far as possible. Since this was a stationary bike,  “as far as possible” was calculated based on the cyclist’s mass, percent grade and how fast the cyclist was going (estimated by the power [in watts] they produced). The cyclists warmed up and cooled down for 5 minutes before the 60-minute ride.

The importance of a “testing” session is pretty obvious, but “familiarization” is probably not at as obvious. Why bother with a familiarization session? Don’t these guys already know how to ride a bike?

In any new situation you have a learning curve. In this case, even though all the participants were cyclists and knew how to cycle, they may have never cycled on this particular type of bike nor in this particular manner. The experimenter can also tweak things at this stage to optimize the future test session. In this case, JB probably did minor adjustments to saddle height and figured out which gradient (5% or 7%) was best for the given cyclist. By having two familiarization sessions JB was able to minimize any effect that the learning curve might have had on the final testing.

Just for reference, the famous Alpe d’Huez, a mountain regularly featured in the Tour de France, has an average gradient of 7.9% and is 13.8 km long. According to CyclingNews, the current record for doing this is held by Marco Pantani, who managed 36 minutes and 50 seconds in 1995.

Research subject cycling in the lab. Notice the headgear recording ventilation.
Research subject cycling in the lab. Notice the headgear recording ventilation.

Another reason in this case for a familiarization session is to get a good base for the cyclists’ abilities. JB set a computer pacer during the first familiarization that was based on the cyclists’ speed 2 seconds earlier. But the cyclists didn’t know this. They were told that the pacer was based on their cycling ability and to try to beat it as convincingly as possible. (Scientists are sneaky. Not because we like fooling around with people and messing with their heads, but because we have to make sure the participants are giving it their all in order to make sure that effort doesn’t play a factor anywhere during the experiments.)

In the next familiarization session, the pacer was set to be exactly the same as the first session. The best performance of the two-familiarization sessions was used as the pacer for the morning test session.

Question to JB: Were the familiarization sessions helpful? Would you consider these guys to be in pretty good cardiovascular shape? Why not go with “untrained” individuals?

JB: No performance study is any good without familiarization sessions. The learning effects are just too great. The best example is working out a different gym. Usually, during the first session at a brand new gym, I don’t give it my all because of the uncertainty. The bars might be different, the visuals different, the benches different, etc. But after a few sessions, when I’m familiar, I can go for it.

In our study, we used only two familiarization sessions because the average performance between session 1 and 2 wasn’t statistically different. If there been a significant difference, we’d have kept bringing the cyclists back until their performance normalized. Then we’d have known they were truly familiarized and ready for some sort of evaluation.

Actually, the fact that it only took two familiarization sessions was a testament to the fitness of these participants. Most of them were high level cyclists and triathletes. They already knew cycling well. Many of them have indoor training bikes at home. So they were pretty comfortable coming in and giving ‘er.

With untrained individuals, we’d have had a different study on our hands. More familiarization would have been required, motivation levels would have been lower, and our study might not have been at all applicable to the population we were most interested in studying here – competitive cyclists.

Breakfast of champions?

On testing day the cyclists came to the lab in the morning. Two hours before testing, they were given a standardized breakfast: Vector Meal Replacement cereal, Vector cereal bars and skim milk.

Table 1 – Meal nutritional breakdown

Kcal/kg g/kg Carbs g/kg Protein g/kg Fat
Vector Meal Replacement cereal, Vector cereal bars & skim milk 7 1.2 0.3 0.1

Question for JB: Why Vector?

The breakfast meals were designed to be easy, to provide adequate kcals, and to deliver a “typical” intake of protein, carbs, and fats. We looked at what these individuals normally ate for breakfast and tried to approximate it as best we could.

Vector calls their product “meal replacement” instead of cereal or granola bars because they contain whole grains, less sugar than normal, and are spiked with protein. Not whole foods nor organic by any means, but way better than regular sugary cereal. In all, the breakfast was a suitable choice and close enough to their regular intake not to affect the results.

In addition, Kelloggs was generous enough to donate all the meal replacement cereal and bars so it was a no-brainer.

Round 1 – One hour of cycling

After a 10 minute warm up, the athletes cycled as far as possible in 60 minutes at either a 5% or 7% grade. This ride was the morning or AMex session. During the ride, they were allowed to drink as much water as they wanted, but nothing else. After the ride, they cooled down for 5 minutes.

Intermission: Supplementation and feeding

5 minutes after the cool down, cyclists were given 1 L (4 cups) of post-exercise liquid supplement that contained carbohydrate-only or carbohydrate and protein.

Liquid supplement nutritional breakdown

Carbohydrate + protein

  • 33% maltodextrin
  • 33% glucose
  • 33% whey protein hydrolysates
  • 4.8 kcal/kg bodyweight
  • 0.8g/kg carbs
  • 0.4g/kg protein


  • 100% maltodextrin
  • 0% glucose
  • 0% whey protein hydrolysates
  • 4.8 kcal/kg bodyweight
  • 1.2g/kg carbs
Question to JB: Why have them cycle as far as possible versus, say, having them cycle to exhaustion?

Cycling to exhaustion doesn’t really mimic sport that well. You see, there aren’t many sports that measure how long someone can run or ride a bike until they drop from fatigue. Instead, most endurance sports measure how long it takes to go a certain distance. Now, we could have chosen to do that in our study.

For example, we could have had folks ride a certain number of kilometres and see how fast they did that distance. However, in the sport of cycling, typically many time trials are ridden in 1 hour or less. So we wanted to standardize the time (1 hour) instead of the distance. Either would have worked. And both are better than cycling to exhaustion in terms of real world applicability.

An hour and two hours after the ride the cyclist were given the same 1 L of post-exercise liquid supplement that they received right after the ride. Then four hours after the ride they were given another “breakfast” meal. See figure 1 below for a diagram of the test session.

Figure 1 – A diagram of the test session

Round 2 – Another hour of cycling

The PMex, or the afternoon exercise session, was the real test. Until this final exhausting ride, there was no data to show which supplementation was better. All the previous sessions and morning ride had one purpose: to set up and provide comparison for this final ride.

The last ride, PMex, was paced from the AMex performance. The cyclists were given exactly the same amount of water as they drank in the AMex to make sure that hydration levels didn’t affect the results.

Measuring results

Along with measuring distance and power output (in watts), JB collected expired gas (air the cyclists breathed out) and blood during the two test rides. By collecting expired gas –- specifically, a volume of oxygen and a volume of carbon dioxide -– JB could calculate how much carbohydrate and fat the athletes used. He used the blood samples to figure out how much glucose and lactate was in the blood serum.

Expired gas was sampled from 5-15 minutes and 35-45 minutes, but only the last 5 minutes of each sampling was analyzed. Blood was taken at 15 and 45 minutes into the rides. In case you’re wondering how a researcher can take blood from somebody while they’re cycling, let’s just say it takes some forethought. It’s a bad idea to try to jab someone with a needle while they’re riding as fast as they can, so before they started cycling JB put a venous catheter into the crook of the cyclists’ elbow (aka the antecubital region) for easier sampling during the ride.

Question for JB: Why sample expired gas for 10 minutes, but only analyze the last 5 minutes?

It has to do with the equipment we used. Air gets trapped in the collecting tubes and the measurement device itself when sitting idle. So the first few minutes of collection actually push that “stale” air out. This ensures that during the actual 5 minute analysis period, you’re analyzing only the expired air of the participant.

Figure 2 – Times of gas collection and blood collection during the two 60-min test rides (AMex and PMex)

Question to JB: It must have been a really long day. You had each cyclist in the lab for at least 10 hours. How long were you in the lab on your average testing day?

On days I’d run only one subject through the protocol, it was about a 13-14 hour day with set-up and wrap-up. After the subjects left, I’d have to treat their blood, freeze part of it, and then take the rest down to another lab for analysis. Some days, however, I’d run 2 subjects and those days would last even longer. No complaints, though. I’d sleep in on days I wasn’t collecting data!

What did JB find out after all those long days in the lab? Did carbohydrates alone do the job? Find out in Part 2!

Eat, move, and live… better.

The health and fitness world can sometimes be a confusing place. But it doesn’t have to be.

Let us help you make sense of it all with this free special report.

In it you’ll learn the best eating, exercise, and lifestyle strategies — unique and personal — for you.

Click here to download the special report, for free.