Train low cycling means starting or finishing selected easy sessions with little carbohydrate, so the muscle adapts to working with less fuel. The UCI's review of nutritional periodisation, by James Morton and colleagues and published in 2026, says to fuel every session for the work it asks of you, and finds the performance evidence for train low limited.
What does fuel for the work required mean?
The review is one of the 14 topic reviews behind the UCI's nutrition position statement. Morton, of Liverpool John Moores University, and five co-authors define nutritional periodisation as the deliberate manipulation of energy, macronutrient and micronutrient intake across the days, weeks and season of training, to help adaptation, recovery, body composition and racing. In practice you set carbohydrate meal by meal according to the length, intensity and purpose of the next session, instead of eating the same amount every day.
The reason is muscle glycogen, the carbohydrate stored in your legs. The authors class it as low below 250 mmol per kg of dry muscle, medium from 250 to 600 and high above 600, and below about 250 the muscle fatigues sooner. Work from Morton's laboratory shows how much the starting level matters. Recreationally active participants who ate 7.6 g of carbohydrate per kg of body weight overnight after a depleting ride managed 44 minutes of efforts at 80% of peak power the next morning. On 3.6 g per kg they managed 36 minutes, and on none they managed 18, with six of the eight unable to finish the planned 8 x 3 minute session. In a separate study by Impey and colleagues, riders who started a long session with high glycogen kept going for 158 minutes against 100 when they started with medium stores.
What is train low in cycling?
The review describes four ways researchers have trained riders with less carbohydrate. A fasted ride is a morning session before breakfast with no carbohydrate on the bike. Twice a day means a second session after a first one, with no carbohydrate in between. Sleep low is a hard evening session, no carbohydrate overnight and an easy ride the next morning. Recover low withholds carbohydrate after a depleting ride.
Inside the muscle the idea holds up. Starting exercise with low glycogen switches on signals linked to building mitochondria and burning fat, and Morton and colleagues report that the effect shows most when a session ends with glycogen at about 200 to 300 mmol per kg or lower.
Do fasted rides and sleep low make you faster?
That is where the evidence thins. The review's authors write that definitive evidence of better performance from train low "is currently limiting", and the trials disagree.
| Study | Who | What they did | Performance result |
|---|---|---|---|
| Marquet and colleagues, 2016 | 21 trained triathletes, 11 on sleep low | 3 weeks of sleep low, both groups on 6 g of carbohydrate per kg a day | 10 km run 2.9% faster against 0.1%, cycling efficiency up 11% against 1.4%, ride to exhaustion not significant (P = 0.06) |
| Gejl and colleagues, 2017 | 26 elite male endurance competitors | 4 weeks, 3 days a week of about 80 g of carbohydrate between a hard 1 hour morning ride and a 2 hour afternoon ride, against 414 g | VO2max and 30 minute time trial power up about 5% in both groups, no difference |
| Gejl and Nybo, 2021 | Meta-analysis of 9 studies in endurance-trained people | Periodised carbohydrate restriction against normal to high carbohydrate | Effect size 0.17, 95% CI minus 0.15 to 0.49, P = 0.29 |
| Hearris and colleagues, 2019 | 8 recreationally active participants | 7.6, 3.6 or 0 g per kg overnight before a hard interval session | 44, 36 and 18 minutes of efforts completed |
None of these diets could be blinded, so riders knew which group they were in, and Marquet's gains came with wide spreads (2.9% plus or minus 2.15%). Gejl and Nybo pooled the trials and concluded that cutting carbohydrate around training does not by itself improve endurance performance. The weight of evidence sits with them, and the sleep-low result stays a lead.
Which sessions should you fuel fully?
The review gives a worked example for a 70 kg male professional over four days. A five hour session of hard intervals gets 3 g per kg of carbohydrate before, 90 to 120 g an hour on the bike and 3 g per kg after, 13.4 to 16.1 g per kg across the day. A five hour steady ride below the first lactate threshold starts on 0 to 0.5 g per kg, takes nothing or a carbohydrate mouth rinse for the first two hours and 30 to 60 g an hour after that, a cell the journal corrected in an erratum on 26 November 2025. That day still totals 8.3 to 10.1 g per kg, because dinner and recovery meals refill the legs for the next hard day. A one hour recovery ride gets no carbohydrate during it.
The authors keep train low for easy, aerobic sessions, and only when the planned intensity and volume won't suffer. Hard intervals, tests and races are fuelled in full, before, during and after. When two sessions fall within 8 hours of each other, the review gives 1.2 g per kg an hour for the first 4 hours after the first one.
The authors also warn that holding the lower intakes through key training blocks would drastically reduce energy availability, with consequences for health and performance.
What does this mean for an amateur rider?
Our view at WattPlan is that most amateurs on 6 to 10 hours a week should take the first half of the idea and be wary of the second. Fuelling the hard days properly has a clear payoff in the session itself, while the gain from train low is unproven and a flat interval session is a certain cost. If you want to try it, keep it to an easy ride you would do anyway, never the day's key session.
A WattPlan plan already separates the hard days. When your week changes, its week repair turns one of two demanding sessions that land back to back into an easy endurance ride, unless both are fixed, such as a race or a locked club ride. The feed reminders in WattPlan's workout player run on structured workouts of 60 minutes or more and on free rides, asking for 60 g of carbohydrate at each feed, every 45 minutes, a pace of about 80 g an hour. That rate doesn't change with the session, so on an interval session follow it, and on an easy ride you have chosen to keep low, decide the amount yourself.
Our guide to endurance ride nutrition covers what to carry on the long days.
What to do this week
- Mark the hard sessions in the week ahead, intervals, tempo blocks and anything over 2.5 hours, and eat a carbohydrate meal of 1 to 4 g per kg in the 1 to 4 hours before each one.
- After a hard session with another within 8 hours, start eating carbohydrate straight away rather than waiting for the next meal.
- If you try a fasted ride, make it an easy one below your first threshold, take water and carry food, and eat as soon as the power or your legs fade.
- On low-carbohydrate days keep protein up. The review cites work suggesting train low adds about 0.12 g per kg a day to protein needs, and protein before or during the ride doesn't appear to blunt the signal.
- Don't combine train low with cutting weight. If you think you are under-eating for your training, talk to a doctor or a registered dietitian.
How does this fit with our earlier nutrition posts?
A few figures here differ slightly from our older posts, and this one is based on the newest evidence, the 2026 UCI reviews. Our recovery nutrition guide gives 1 to 1.2 g of carbohydrate per kg an hour for 4 hours whenever the next hard ride is within a day, while the UCI review keeps its fastest refuelling for sessions less than 8 hours apart. The same guide gives 1.4 to 2.0 g of protein per kg a day from a 2017 position stand, where the 2026 UCI statement gives 1.6 to 2.1 g. Where they disagree, go with the newer figures in this post.
What we don't know yet
Most train-low trials are small, last 2 to 6 weeks, recruit men and test performance in efforts of an hour or less. Morton and colleagues call for studies in elite men and women with tests of 4 to 6 hours, where the fat-burning adaptations would most likely show. They also note one cost, a 1.1% drop in sleep efficiency during a 3 week sleep-low block in male triathletes. Nobody has yet tested a train-low plan in amateur riders fitting 8 hours around work.
Sources
- UCI Sports Nutrition Project: Nutritional Periodization: Strategies to Enhance Training Adaptation and Recovery, International Journal of Sport Nutrition and Exercise Metabolism
- Erratum. UCI Sports Nutrition Project: Nutritional Periodization, International Journal of Sport Nutrition and Exercise Metabolism
- Enhanced Endurance Performance by Periodization of Carbohydrate Intake: "Sleep Low" Strategy, Medicine and Science in Sports and Exercise
- No Superior Adaptations to Carbohydrate Periodization in Elite Endurance Athletes, Medicine and Science in Sports and Exercise
- Performance effects of periodized carbohydrate restriction in endurance trained athletes: a systematic review and meta-analysis, Journal of the International Society of Sports Nutrition
- UCI Sports Nutrition Project: Position Statement on Nutrition for Cycling, International Journal of Sport Nutrition and Exercise Metabolism