In January 2016, I wrote a blog post about the potential benefits of doing occasional training sessions in a glycogen depleted state, in other words with low carbohydrate stores. Various research studies have shown that this has benefits for endurance training, specifically in upregulating the process that creates mitochondria, the “batteries” of the muscle cell which burn fuel for energy. The more mitochondria you have in each muscle cell, the better for performance in endurance events. Training with low glycogen has also been shown to increase fat oxidation, your body’s ability to burn fat as fuel. This is thought to have benefits for endurance performance by sparing the limited glycogen stores available in the muscle. The longer you can use fat as your primary energy source, the longer you can make your carbohydrate stores last. You can find my earlier blog post here for more detailed information on how training with low glycogen stores at certain times may be beneficial.

Here are some ways in which you can ‘train low’:

  • Train before breakfast on an empty stomach, ie fasted.
  • Do your first workout of the day after consuming some carbohydrate and then do not eat any carbs afterwards. Perform your second daily session with low carb stores.
  • Train for more than 90 minutes without taking on any carbohydrate during the session.
  • Don’t have any carbohydrate for 2-4 hours after training instead of your usual recovery meal.
  • Train after an evening meal, then again the next morning without consuming any carbohydrate until after the second workout (“sleep low”).

What have perhaps been missing from the research up to now, however, are specific studies that tested out these protocols in endurance athletes and showed measureable improvements in performance. But since I last wrote on this topic, there have been two interesting studies done on triathletes and cyclists using a carefully controlled “sleep low” protocol (the last on the list above) and comparing the results to a similar group who undertook the same training after consuming carbohydrate. The results were very encouraging.

In the first study 1, 21 triathletes were divided into two groups – “sleep low” and “control” – and undertook a three-week training and diet intervention. During this time, both groups performed three blocks of 6 supervised training sessions over four days: high intensity cycling or running intervals in the afternoon after consuming a high amount of carbohydrate and a low intensity session the next morning. Importantly, both groups consumed the same amount of carbohydrate across each 24 hour period (6g/kg/day, so a moderate intake). This is not a ‘low carb diet’ study. But while the “control” group performed all their workouts in a fully fuelled state, with carbohydrate consumed before and after both workouts, the “sleep low” group did not consume any carbohydrate after their afternoon interval session, only protein and fat, and then performed the morning session in a fasted state, refuelling with carbohydrate only after finishing it.

The performance of both groups was measured via several tests both before and after the three week intervention. The results were clear: the “sleep low” group showed improvements in sub-maximal cycling efficiency (including reduced heart rate and increased fat oxidation), supra-maximal cycling to exhaustion (cycled 12.5% longer) and 10km run performance (3% faster on average). They also saw a decrease in fat mass. No performance improvements were seen in the “control” group.

What’s particularly significant about this study is its practicality. While following a low carb diet for a specific period has been shown to increase fat oxidation in various studies, it also appears to downregulate the ability to use carbohydrate as fuel, making it harder to perform at higher intensities. But the protocol used in this study only lasted three weeks and did not involve restricting carbohydrate overall, just around certain training sessions. The participants also undertook high intensity training sessions with fully stocked glycogen stores, doing only low intensity sessions with a glycogen depleted state. The results speak for themselves.

It would be possible to replicate the protocol used in this study in your own training, eg by inserting a three-week training block using the “sleep low” approach at an appropriate point. But the researchers wanted to see if similar results could be obtained from just a one week diet and training intervention, so they set up a second study 2. In this case, 11 trained cyclists undertook just a single four-day high intensity/low intensity training block, with half the group completing all six sessions with full glycogen stores and the other half performing the morning low-intensity session with depleted glycogen stores, as before. The tests this time, before and after the intervention, were a two-hour sub maximal ride followed by a 20km time trial. The result was similar: the “sleep low” group saw a 3.2% improvement in their time trial, associated with a higher power output during the second half of the test. The “control” group saw no improvement. Interestingly, there was no difference in fat vs carbohydrate fuel utilisation during the sub-maximal ride, so the improvement in the time trial isn’t explained by “sparing” muscle glycogen.

Implementing a one week block of the “sleep low” training and diet protocol into your race preparation plan is a lot easier than a three-week block, and this latter study suggests that you will still see benefits in performance from a shorter intervention.

If you want to try it, here’s an example of how your diet and training might look over a 24 hour period during the “sleep low” block:

1.00pm: Lunch (including carbs): baked sweet potato, chicken, salad, apple, low fat Greek yogurt

4.00pm Snack (including carbs): banana and honey on toast or oatcakes

5.00pm Interval training session

7.00pm Dinner (no carbs): salmon, eggs, avocado

10.00pm Snack (no carbs): drink made with casein protein


7.00am Steady training session (done fasted)

8.00am Recovery drink: carbs and whey protein

9.00am Breakfast (including carbs): porridge oats, banana, honey, Greek yogurt, almond butter.

Use an app like to calculate the portion sizes you need to give you sufficient carbohydrate over the course of each day, based on your bodyweight (6g/kg per day).

Repeat this protocol four times in a single week, ideally on four consecutive days. The amount of carbohydrate depends on your body weight (aim for 6g/kg, around 55-60% of overall calorie intake). Ideally, do the same performance test before and after the four day block to assess your progress.

It should be noted that both these studies were undertaken with male subjects only, so it isn’t known if female athletes would see similar positive results.

Jo Scott-Dalgleish BSc (Hons) is a BANT Registered Nutritionist, writing and giving talks about nutrition for endurance sportBased in London, she also works as a Registered Nutritional Therapist, conducting onetoone consultations with triathletes, distance runners and cyclists to help them eat well, be healthy and perform better through the creation of an individual nutritional plan. To learn more about these consultations, please visit



1 Marquet et al. Enhanced endurance performance by periodisation of carbohydrate intake: “Sleep Low” strategy. Med Sci Sports Exerc. 2016. Apr; 48(4):663-72

2 Marquet et al. Periodisation of carbohydrate intake: short term effect on performance. Nutrients. 2016. Nov 25; 8(12) pii E755 (full text available)