Sleep loss and exercise performance
Across 227 outcome measures from 69 publications, sleeping 6 hours or less in any 24-hour period, compared with a normal-sleep control of more than 6 hours in any 24-hour period, was associated with a mean change in exercise performance of - 7.56% (95% CI - 11.9 to - 3.13), with near-total heterogeneity (I2 = 98.1%). Effects were significant for all exercise categories, but subgroup analyses indicated that the pattern of sleep loss preceding exercise is an important factor, with consistent negative effects only observed with deprivation and late-restriction (earlier-than-normal waking) protocols. In those same two protocols, a significant relationship was observed between time awake before the task and the change in performance, amounting to roughly a 0.4% decrease for every hour awake. The negative effects on tasks performed in the PM were consistent, while tasks performed in the AM were largely unaffected.
Craven J (2022) · Sports medicine (Auckland, N.Z.) · PMID 35708888
It does not say that sleep EXTENSION improves performance. This review compared normal sleep against sleep loss only, so it cannot support "sleep more than usual and you will lift more." It does not give a strength-specific effect size: the abstract reports 66 strength outcome measures and says "Effects were significant for all exercise categories," but it never states a separate number for strength, so no one may quote "7.56% off your squat." The - 7.56% is not a prediction for an individual athlete either: the abstract reports I2 = 98.1%, near-total heterogeneity, with a confidence interval running from - 11.9 to - 3.13. The ~ 0.4%-per-hour dose is reported only for deprivation and late-restriction protocols; the abstract does not extend it to early restriction (a late bedtime with a normal wake time), so it must not be quoted as a universal per-hour penalty. And the morning-vs-evening result is about the acute timing of a test after a bad night; it is not evidence that morning training is generally superior, nor that a poor night is harmless if you simply train early.
Every figure below is shown with the sentence it was taken from, so you can check it against the paper rather than trusting us.
- - 7.56%percent change in exercise performance
Pooled mean percentage change in exercise performance under sleep loss versus normal sleep, across all exercise categories and all sleep-loss protocols
Results indicated a negative impact of sleep loss on the percentage change (%Δ) in exercise performance (n = 959 [89%] male; mean %Δ = - 7.56%, 95% CI - 11.9 to - 3.13, p = 0.001, I2 = 98.1%).
- - 11.9 to - 3.1395% confidence interval, percent change
Confidence interval around the pooled mean performance change
Results indicated a negative impact of sleep loss on the percentage change (%Δ) in exercise performance (n = 959 [89%] male; mean %Δ = - 7.56%, 95% CI - 11.9 to - 3.13, p = 0.001, I2 = 98.1%).
- 98.1%I-squared
Statistical heterogeneity across the pooled outcomes - near-total, meaning the average conceals very wide variation between studies
Results indicated a negative impact of sleep loss on the percentage change (%Δ) in exercise performance (n = 959 [89%] male; mean %Δ = - 7.56%, 95% CI - 11.9 to - 3.13, p = 0.001, I2 = 98.1%).
- 959participants
Total participants contributing to the pooled percentage-change estimate
Results indicated a negative impact of sleep loss on the percentage change (%Δ) in exercise performance (n = 959 [89%] male; mean %Δ = - 7.56%, 95% CI - 11.9 to - 3.13, p = 0.001, I2 = 98.1%).
- 89%percent of participants
Share of the pooled sample that was male - the evidence base is overwhelmingly male
Results indicated a negative impact of sleep loss on the percentage change (%Δ) in exercise performance (n = 959 [89%] male; mean %Δ = - 7.56%, 95% CI - 11.9 to - 3.13, p = 0.001, I2 = 98.1%).
- ~ 0.4%percent performance decrease per hour awake
Meta-regression dose-response, reported ONLY for deprivation and late-restriction (earlier-than-normal waking) protocols: performance fell by roughly this much for every additional hour awake before the task. Not reported for early restriction (a late bedtime with a normal wake time).
A significant positive relationship was observed between time awake prior to the exercise task and %Δ in performance for both deprivation and late-restriction protocols (~ 0.4% decrease for every hour awake prior to exercise).
- 227outcome measures
Total performance outcome measures pooled
Overall, 227 outcome measures (anaerobic power: n = 58; speed/power endurance: n = 32; HIIE: n = 27; strength: n = 66; endurance: n = 22; strength-endurance: n = 9; skill: n = 13) derived from 69 publications were included.
- 66outcome measures
Number of strength-category outcome measures included - the largest single category, though no separate strength effect size is reported anywhere in the abstract
Overall, 227 outcome measures (anaerobic power: n = 58; speed/power endurance: n = 32; HIIE: n = 27; strength: n = 66; endurance: n = 22; strength-endurance: n = 9; skill: n = 13) derived from 69 publications were included.
- 69publications
Number of source publications synthesised
Overall, 227 outcome measures (anaerobic power: n = 58; speed/power endurance: n = 32; HIIE: n = 27; strength: n = 66; endurance: n = 22; strength-endurance: n = 9; skill: n = 13) derived from 69 publications were included.
- ≤ 6 hhours of sleep in any 24 h period
The review's operational definition of the sleep-loss condition; the normal-sleep control was more than 6 h in any 24 h period
Investigations measuring exercise performance under 'control' (i.e., normal sleep, > 6 h in any 24 h period) and 'intervention' (i.e., sleep loss, ≤ 6 h sleep in any 24 h period) conditions were included.
Pooled sample of n = 959, of whom 89% were male, as stated in the abstract. The abstract does not report training status, age, or any further sex breakdown beyond that 89% male figure, so this is a predominantly male evidence base and nothing in the abstract licenses a trained-vs-untrained split or a claim about female athletes.
Craven J, et al. (2022)Effects of Acute Sleep Loss on Physical Performance: A Systematic and Meta-Analytical Review.
Sports medicine (Auckland, N.Z.) · PMID 35708888
Meta-regression · 227 outcome measures derived from 69 publications · n = 959 [89%] male
Who funded itAuthors declare no conflicts of interest.
Volume, rep ranges, rest, stretching, sleep — each with its numbers and its limits.
Every figure is machine-checked against the sentence it came from. Including what that does not guarantee.
Repstra handles the programming, check-ins and admin so the conversation you have with a client is the coaching one. Free for your first 5 clients, 0% commission on your own.
See how it works