Carbohydrate periodisation (train-low)
Evidence: contested
Training with low glycogen reliably amplifies the molecular signals of mitochondrial adaptation, but the translation to real performance gains is inconsistent and the downsides are real. A niche, periodised tool, not a default.
Not medical advice
This is a general knowledge base, not medical or dietary advice. If you are injured, unwell or weighing up a supplement or a change to your diet, speak to a doctor, physiotherapist or registered dietitian who knows your situation.
Carbohydrate periodisation means deliberately varying carbohydrate availability across the training week rather than eating high all the time. The organising idea, set out by Impey and colleagues as “fuel for the work required”, is to match carbohydrate to the demands of the upcoming session: fuel hard for the sessions that need quality, and deliberately go in with low glycogen for some easier sessions to harvest a stronger adaptive signal (Impey et al. 2018). This is distinct from a chronic low-carbohydrate or ketogenic diet: periodisation keeps total carbohydrate adequate over the week and simply moves it around, where keto restricts it permanently.
The train-low methods
Several protocols all aim to put a session in a low-glycogen state (Impey et al. 2018):
- Fasted or overnight-low sessions. An easy aerobic run done before breakfast, after an overnight fast, with glycogen partly drawn down.
- Twice-a-day train-low. A first session depletes glycogen; carbohydrate is withheld between sessions so the second session is performed glycogen-depleted.
- Sleep-low. A hard session in the evening with adequate carbohydrate beforehand, then a low-carbohydrate overnight period so glycogen stays low, then a fasted easy session the next morning. This concentrates the low-glycogen window over a long stretch (Marquet et al. 2016).
Across the literature, train-low typically covers only 30 to 50% of sessions, not all of them (Impey et al. 2018).
The mechanism
Low muscle glycogen amplifies the molecular signalling that drives mitochondrial adaptation. When glycogen is low, exercise produces a larger activation of the energy-sensing and stress kinases AMPK and p38 MAPK, and greater downstream expression of PGC-1alpha and p53, the master regulators of mitochondrial biogenesis (Hawley et al. 2018; Impey et al. 2018). This is the same theme as the antioxidant story in reverse: there, blunting a stress signal blunts adaptation; here, sharpening a stress signal is meant to enhance it. Impey and colleagues frame this as a glycogen threshold, a window of glycogen low enough to maximise signalling yet high enough to still complete the session’s work (Impey et al. 2018).
The honest verdict
The molecular case is strong and the performance case is not. Reviewing the controlled trials, Impey and colleagues found train-low augmented cell signalling in 73% of studies and gene expression in 75%, yet only 37% showed a performance improvement while 63% showed no change (Impey et al. 2018). The adaptive signal reliably strengthens; the performance gain does not reliably follow. The one well-known positive trial, Marquet’s sleep-low study, did show better cycling efficiency and 10-km running time over three weeks, but in a small sample of trained triathletes and without consistent replication (Marquet et al. 2016).
The downsides are concrete. Training with low glycogen reduces the quality of sessions performed at intensity, so it is ill-suited to the hard work where most fitness is built (Impey et al. 2018). Chronic under-fuelling around training also raises illness risk and is a slippery slope toward low energy availability and RED-S, where the costs to bone, hormones and health dwarf any marginal adaptive gain (Hawley et al. 2018).
How to use it, if at all
Keep it for occasional easy aerobic sessions, never for quality work, and only against a backdrop of adequate weekly energy and carbohydrate. Fuel the hard sessions and races properly. As ever, the basics of consistent volume, sleep and adequate fuelling matter far more than this manipulation.