REF-218 · Gejl & Nybo, 2021 · Journal of the International Society of Sports Nutrition, Vol. · Paper
Performance effects of periodized carbohydrate restriction in endurance trained athletes
Systematic review and meta-analysis
Design
Rationale in the abstract: training or recovering with low muscle glycogen stimulates factors of importance for mitochondrial biogenesis and favourable metabolic adaptations, but chronic CHO restriction compromises training quality, especially the intensity of peak intervals — hence "periodized" restriction of selected sessions only.
Inclusion criteria: females and males with VO2max ≥55 and ≥60 ml/kg/min respectively; total intervention and training period ≥1 week; training and/or recovery with periodized CHO restriction at least three times per week; endurance performance measured before and after.
Search: 407 papers -> 9 studies met the criteria.
Key Results
| Metric | Effect |
|---|---|
| Standardized mean difference (periodized CHO restriction vs normal/high CHO control training) | 0.17 [−0.15, 0.49] |
| p-value | 0.29 (not significant) |
| Conclusion | Periodized CHO restriction does not per se enhance endurance performance in endurance-trained athletes |
Findings
Based on the available literature, periodized CHO restriction (training and/or recovery with low CHO availability at least three times per week) does not per se enhance performance in endurance-trained athletes compared with control training on normal (high) CHO availability.
The review discusses the different periodization approaches across studies with a focus on identifying potential physiological benefits — i.e. the metabolic rationale stands, the performance effect does not.
Critical Context for Practice
Not from this paper: the single trial most often quoted for a performance gain, Marquet 2016 REF-003 (Sleep-Low), is one of the included studies; pooled across the 9 studies the effect is not significant.
The confidence interval [−0.15, 0.49] includes zero on one side and a small positive effect on the other: the data neither show harm nor a reliable benefit.
Inference, not tested here: the physiological signals the abstract cites (mitochondrial biogenesis, metabolic adaptation) are a reason to consider selected low-CHO sessions, not a reason to expect faster racing from them.
Application
Nine training studies in endurance-trained athletes (VO2max ≥55 women / ≥60 men), each ≥1 week with periodized CHO restriction at least three times per week: pooled standardized mean difference on endurance performance 0.17 [−0.15, 0.49], p = 0.29 — no performance effect versus normal high-CHO training. Train-low is not a performance intervention; if used, it is for the metabolic rationale, with no expectation of a faster time trial.
Inference from the abstract's rationale, not tested: keep the sessions where intensity matters on normal (high) CHO availability, because compromised quality of peak intervals is the drawback the authors name for chronic restriction.
Not from this paper: how much carbohydrate to take in during training and racing is in Jeukendrup 2017 REF-008; race-day fuelling is not a train-low context.
Limitations
Nine studies with heterogeneous train-low approaches; the abstract does not give participant totals.
Performance was the only pooled outcome; the physiological benefits are discussed, not meta-analysed.
Nine small training studies with heterogeneous train-low protocols and performance tests.
A null pooled effect whose confidence interval spans zero does not exclude a small benefit — it rules out a large one.
How this entry was checked and changed
Why this level
regraded 2026-09-26 from A: meta-analysis of nine small RCTs
Content correction 2026-09-27
"9 RCT" -> abstract: nine training studies meeting the criteria (design not stated as RCT).
"Marquet 2016 +3% TT" moved to a "Not from this paper" line pointing to REF-003; "train-low 1-2×/week" and "race day 90-120 g/h" -> house advice not in the abstract; removed (dose pointed to REF-008).
Inclusion criteria restated exactly (VO2max ≥55 women / ≥60 men; ≥1 week; restriction at least three times per week) and the abstract's rationale (mitochondrial biogenesis vs compromised peak-interval intensity) added.