Glycogen-Depletion Workouts (Training Low): When They Help, When They Backfire

Educational journalism, not medical advice. Every claim here is checked against its cited sources by editor Tim Bunce — a health writer, not a physician. It isn’t specific to your situation: for health decisions, talk to your own clinician. How we work →

What the published evidence shows

Hansen 2005 was the first published controlled trial in trained endurance athletes. One leg was trained with full muscle glycogen (twice-daily protocol that didn’t deplete stores); the other leg performed the second daily session with depleted glycogen. After 10 weeks, the trained-low leg showed:

• Higher resting muscle glycogen
• Greater citrate-synthase activity (mitochondrial enzyme)
• Improved time-to-exhaustion in tests Hansen 2005

Subsequent trials (Yeo, Burke, Bartlett) have broadly replicated the cellular signalling findings. The molecular response to low-glycogen exercise is roughly 2-3 times the AMPK/PGC-1α signal of the same workout with normal glycogen Bartlett 2015.

How to actually train low

• Fasted morning session. Wake up, drink water, do a 60-90 minute zone-2 session before breakfast. Overnight fast plus exercise gradually depletes glycogen.
• Sleep low. Hard evening session, then skip carbs at dinner. Wake up for next-morning easy session. Refuel after.
• Twice-daily protocol. Morning hard session (full glycogen), afternoon recovery without carb intake between. The second session is the “low” one.
• 1-2 sessions per week is the maximum. More than that produces accumulating immune-system strain and degraded training quality.
• Always recover with carbs after. Train low, recover with carbs. The catabolic state matters during training; not during recovery.

Cautions

• Don’t train high-intensity with low glycogen. Performance drops 10-15% in glycogen-depleted state. The adaptation benefit doesn’t outweigh the lost intensity.
• Watch for immune-system signs. Frequent colds, slow recovery, declining HRV. Train-low protocols stress the immune system; cap weekly volume at 1-2 sessions.
• Race day: train high. The trained adaptations show up most clearly when racing with full glycogen. Carb-load normally before competition.
• Female athletes: some published data suggests women may be more sensitive to low-glycogen training stress, with menstrual-cycle disruption a risk if overdone. Conservative dosing.

Practical takeaways

• 1-2 weekly zone-2 sessions in glycogen-depleted state produces 2-3x the molecular signal for mitochondrial biogenesis vs. fed sessions.
• Methods: fasted morning, sleep-low, twice-daily protocol.
• Train low at low intensity only. Train high (full glycogen) for hard sessions and races.
• Refuel with carbs after train-low sessions; the depleted state matters during training, not during recovery.
• Cap at 1-2 sessions weekly; more produces immune-system strain.

The gap between cell signalling and the finish line

The most important caveat in this whole topic is one the early enthusiasm glossed over: a bigger molecular signal is not the same thing as a faster race. The cell-signalling story is genuinely strong — the “fuel for the work required” framework collated by Impey and colleagues found that commencing a session within a specific low range of muscle glycogen amplified the adaptive signal in roughly three-quarters of the studies they reviewed (augmented cell signalling in 73% of 11 studies, gene expression in 75% of 12, and oxidative-enzyme increases in 78% of 9) Impey 2018. But signalling is an intermediate marker. What an athlete actually cares about is the clock, and there the picture is far messier.

On the optimistic side sits the trial that put “sleep low” on the map. Marquet and colleagues randomised 21 trained triathletes to either a sleep-low protocol or a control group; both consumed an identical 6 g of carbohydrate per kilogram per day, differing only in when it was eaten. After three weeks, every athlete in the sleep-low group improved their 10-km run time (by an average of 73 ± 20 seconds) and their supramaximal cycling capacity, while the control group did not change Marquet 2016. That is an unusually clean result, and it is the single strongest piece of human performance evidence in favour of periodised carbohydrate.

The problem is that it has not reliably replicated in better-trained athletes. Gejl and colleagues ran a 4-week randomised trial in 26 elite male endurance athletes (mean V̇O₂max 65 mL·kg·min) who layered carbohydrate restriction onto their normal training three days a week. Both groups improved — V̇O₂max rose 5%, muscle glycogen rose 18%, citrate-synthase activity rose 11% — but the changes were essentially identical whether or not carbohydrate was periodised. The authors concluded bluntly that periodic carbohydrate restriction had “no superior effects on performance and muscle adaptations in elite endurance athletes” Gejl 2017. When the same group later pooled the field in a systematic review and meta-analysis of nine trials (about 140 athletes), the combined performance effect was small and non-significant (standardised mean difference 0.17; 95% CI −0.15 to 0.49; p = 0.29). Their summary judgement: the evidence that carbohydrate periodisation enhances endurance is “weak,” and “train-low is not per se associated with enhanced endurance” Gejl & Nybo 2021.

Reading those side by side, the honest position is: train-low reliably turns up the adaptive signalling dial, it may translate to faster racing in sub-elite athletes who are not already maximally adapted (as in Marquet), and it has so far failed to beat ordinary fuelled training in already-elite athletes. It is a plausible marginal-gains tool, not an established performance multiplier — and anyone selling it as guaranteed free speed is reading the mechanism studies and ignoring the outcome studies.

What the immune-system risk actually is — and isn't

The frequently repeated warning that train-low “suppresses immunity” deserves a more careful reading than it usually gets, because the best-controlled study of the sleep-low protocol found the opposite of a dramatic effect. Louis and colleagues tracked 21 trained triathletes through three weeks of nine twice-daily sessions a week, half of them done with reduced glycogen. White-blood-cell counts, plasma cortisol and salivary interleukin-6 did not change meaningfully; the incidence of upper-respiratory-tract infections was unchanged and low in both groups; and sleep was barely affected. Salivary IgA (a mucosal-immunity marker) dipped slightly in the sleep-low group, but the overall conclusion was that sleeping and training with reduced glycogen had “minimal effects” on immunity and sleep — crucially, provided total daily carbohydrate and energy intake stayed adequate Louis 2016.

That conditional clause is the whole point. The genuine danger of train-low is not the individual depleted session; it is using “train low” as cover for chronically under-eating. When low carbohydrate availability is paired with low overall energy availability and sustained over weeks, it stops being a training stimulus and becomes a driver of Relative Energy Deficiency in Sport (REDs). The 2023 International Olympic Committee consensus statement — built on more than 170 new studies since 2018 — explicitly flags low carbohydrate availability as an emerging contributor to REDs, a syndrome that impairs bone health, endocrine and reproductive function, metabolic rate, mood and, eventually, performance itself in both female and male athletes Mountjoy 2023. In other words, the immune-suppression headline is a symptom of a larger problem: train-low becomes hazardous precisely when it slides into under-fuelling. Keep the day's total carbohydrate and energy targets intact, restrict only the timing around one or two sessions, and the immune signal in the controlled data is small. Let total intake fall, and you are no longer periodising carbohydrate — you are starving your training.

A cleaner decision rule, and who should skip it

The most useful reframing to come out of this literature is Impey and colleagues’ “fuel for the work required” principle: instead of eating a fixed high-carbohydrate diet every day, match carbohydrate to the specific demand of the next session, and reserve the low-glycogen state for easy work where glycogen is not the limiting fuel Impey 2018. The practical translation is a simple if-then: if the session is long and easy (zone 2, technique, recovery), it is a candidate to perform low; if the session has any quality — intervals, threshold, race-pace, a hard group ride — it should be fully fuelled, because the meta-analytic evidence is clear that low carbohydrate availability blunts the intensity you can hold in peak intervals, and losing that intensity can cost you the very adaptations the hard session exists to drive Gejl & Nybo 2021. This is the evidence-based version of the gym-floor maxim “train low, race high”: protect quality and competition with carbohydrate, and confine depletion to sessions that genuinely don’t need it.

Some people should not experiment with train-low at all, or should only do so under professional supervision. Anyone with a current or past eating disorder, disordered-eating tendencies, or a history of REDs should avoid deliberately training in an under-fuelled state — the protocol normalises exactly the behaviour they need to move away from, and the IOC framework treats prior low-energy-availability exposure as a key moderating risk factor Mountjoy 2023. People with diabetes or on glucose-lowering medication should not train fasted or glycogen-depleted without medical guidance, given the hypoglycaemia risk. Adolescents and masters athletes, pregnant or breastfeeding athletes, and anyone with low bone-mineral density have more to lose from any drift toward energy deficiency than they stand to gain from a marginal mitochondrial signal. And recreational exercisers chasing general health or weight management simply don’t need it — train-low is a periodisation tool aimed at squeezing oxidative adaptations out of athletes already doing high training volumes, not a fat-loss hack. If any of these apply to you, the sensible move is to discuss it with a sports physician or accredited sports dietitian before changing how you fuel, rather than copying a protocol designed for healthy, well-trained, well-monitored competitors.

References