What does one night awake cost the muscle?
In 13 young non-athletes, afternoon muscle protein synthesis was 18% lower the next day, with cortisol 21% higher and testosterone 24% lower. One night, no training, a marker not performance.
Main caveat: A marker, not an outcome. A two-hour postprandial synthesis rate is not muscle mass, strength or performance.
Thirteen healthy students (7 men, 6 women) completed two conditions at least four weeks apart in randomised order: a normal night at home (22:00 to 07:00) and one night fully awake in the laboratory under observation.
On both following days they ate standardised meals, received an infusion of labelled phenylalanine from 10:00, and had muscle taken from the thigh at 13:00 and at 15:00. Mixed-muscle fractional synthesis rate over those two hours was 18% lower after the night awake (0.072 ± 0.015 against 0.059 ± 0.014 %/h, p = .040, single-tailed).
Cortisol area under the curve from 10:00 to 16:00 was 21% higher after deprivation (186 ± 41.7 against 226 ± 44.6 AU, p = .011 in the results, p = .030 in the abstract) and testosterone area under the curve 24% lower (6.40 ± 5.28 against 4.86 ± 3.64 AU, p = .029), although the sleep-by-time interaction for testosterone did not reach significance (p = .063).
IGF-1, insulin, the IGF-1 gene isoforms, the clock genes and the degradation genes showed no clear change. Every man, but not every woman, had a lower synthesis rate after the night awake; the study was not powered to compare the sexes and the authors say so.
Our reading: the cleanest human evidence so far that a single night without sleep blunts the muscle's response to food the next day, in people who do not train, measured as a rate over two hours.
Protocol
Design: randomised crossover, control (CON) and deprivation (DEP), at least 4 weeks apart. Ethics approval 2016-028; no trial registration is stated. Sample size 13 from a power analysis on fractional synthesis rate using earlier immobilisation and exercise data (beta = 0.2, alpha = 0.05), since no study of sleep loss and protein synthesis existed when the trial was designed.
Who: 13 young healthy students, 18-35 y. Excluded: shift work in the previous 3 months, travel across time zones in the previous 4 weeks, two or more naps a week, a diagnosed sleep disorder, extreme morning or evening chronotype. Habitual bedtime 22:00-00:00, wake 06:00-08:00, at least 7 h of self-reported sleep. Women were tested in the same cycle phase in both conditions, avoiding the follicular phase.
Before each trial: habitual sleep for the preceding week, with a wrist actigraph and a diary. No hard exercise, alcohol or caffeine for 48 h. A standardised meal at 19:00 (about 20% fat, 14% protein, 66% carbohydrate, 8.4-8.9 kcal/kg).
Deprivation night: reported to the laboratory at 21:00, sedentary activities only, continuously watched and monitored by actigraphy, 30 h in a sound-attenuated room at 211 ± 14 lux and 21 ± 2°C, with low-protein snacks (fruit and vegetables) and water allowed. Control night: slept at home, woken at 07:00 by a researcher and a nurse who took blood before any activity or light exposure.
Test day, both conditions: breakfast at 07:30 (20.3 ± 1.8 g protein, 147 ± 7.2 g carbohydrate), cannulae at 08:00, primed (0.34 mg/kg) constant infusion (0.0085 mg/kg/min) of labelled phenylalanine from 10:00, lunch at 12:00 (20.6 ± 0.3 g protein, 65.7 ± 5.6 g carbohydrate), biopsies of the vastus lateralis at 13:00 and 15:00. Hourly blood from 07:00 to 17:00; IGF-1, cortisol and testosterone at 07:00, 10:00, 13:00 and 16:00; insulin at 07:00, 11:00 and 15:00.
Measures: mixed-muscle fractional synthesis rate by the precursor-product model; plasma cortisol, testosterone, IGF-1 by immunoassay and insulin by bead panel; mRNA at 13:00 for ARNTL, CRY1, PER1, IGF-1Ea, IGF-1Eb, FBXO32, TRIM63, FOXO1, FOXO3, normalised to single-strand DNA. Sleep by actigraphy, not polysomnography.
Analysis: two-way repeated-measures ANOVA for hormones with Sidak comparisons; area under the curve by the trapezoidal method; single-tailed paired t tests for synthesis rate and hormone concentrations, two-tailed for sleep and gene data; p < .05; mean ± SD.
Funding and interests: the open full text prints no funding statement; the authors declare no conflict of interest and thank two named colleagues for technical help.
Key Results
| Measure | Normal night (CON) | Night awake (DEP) | p |
|---|---|---|---|
| Mixed-muscle fractional synthesis rate (%/h) | 0.072 ± 0.015 | 0.059 ± 0.014 | .040 (single-tailed) |
| Cortisol area under the curve 10:00-16:00 (AU) | 186 ± 41.7 | 226 ± 44.6 | .011 (abstract: .030) |
| Testosterone area under the curve (AU) | 6.40 ± 5.28 | 4.86 ± 3.64 | .029 |
| Testosterone, sleep x time interaction | .063 | ||
| Cortisol, sleep x time interaction | 4.38E-5 | ||
| Cortisol at 07:00 | higher in CON | .014 | |
| Cortisol at 10:00 / 16:00 | higher in DEP at 16:00 | .940 / .048 | |
| IGF-1, insulin, IGF-1Ea and IGF-1Eb mRNA | no clear change | not significant | |
| FBXO32, TRIM63, FOXO1, FOXO3 mRNA | no clear change | not significant | |
| ARNTL, CRY1, PER1 mRNA | no clear change | not significant | |
| Sleep in the week before | 5.9 ± 0.5 h | 6.1 ± 1.4 h | .718 |
| Sleep the night before the trial | 6.8 ± 0.8 h | 7.4 ± 0.7 h | .195 |
| Sleep during the intervention | night at home | 0.0 ± 0.0 h |
The 18% is the paper's own figure for the difference in synthesis rate; the two means above are what it rests on.
No clear difference, not "the same." The degradation genes, the clock genes, IGF-1, insulin and the IGF-1 isoforms showed no significant change. No equivalence test was run, and with 13 participants a real effect of this size could be missed.
Direction in all men, not in all women. Every man had a lower rate after the night awake; some women went the other way, and the authors call this exploratory, not powered, and not tested statistically.
Participants came in mildly short of sleep. Actigraphy in the week before each arm gave 5.9 and 6.1 h, below the 7 h they reported habitually; the authors flag it and say the effect of that is unclear.
Our arithmetic: the two biopsies are two hours apart, so the rate describes 13:00 to 15:00 after a breakfast and a lunch of about 20 g protein each — one postprandial window on one day, not a daily balance.
Mechanisms
Measured: lower postprandial synthesis rate, higher cortisol exposure across the day, lower testosterone exposure; no change in the measured degradation genes or in IGF-1 and insulin.
Authors' interpretation: one night awake is enough to produce anabolic resistance — the muscle answers a protein-containing meal less strongly — together with a procatabolic hormonal environment; they propose this as a mechanistic precursor to the body-composition changes seen with chronic sleep loss. They note cortisol's own evidence points more to breakdown than to a lower synthesis rate, so the hormone data are a plausible setting rather than a demonstrated path.
Authors' interpretation: the flat cortisol profile across the deprivation day, against a falling one in the control, may indicate circadian misalignment; night-time snacks are acknowledged as a possible confounder, and the area under the curve was taken from 10:00, after intake was standardised.
Not measured (inference on what follows): protein-level signalling could not be run because there was too little tissue, the myofibrillar fraction specifically was not isolated, and nothing mitochondrial, aerobic or performance-related was measured at all. Whether this one-day difference accumulates into lost muscle, or is offset by training or extra protein, is untested here.
Practical Application
What does one night awake cost the muscle? In 13 young non-athletes, afternoon muscle protein synthesis was 18% lower the next day, with cortisol 21% higher and testosterone 24% lower. One night, no training, a marker not performance.
Good for: a reason to protect sleep when the goal is muscle, beyond how an athlete feels — by inference, since no training was done here. After a night awake the muscle answered food less strongly, in a crossover where each person was their own control.
Good for: keeping the response proportionate. One night was measured, not a week; and the base's practice after a short night is to keep the plan and let the warm-up decide the key session (SYNTH-07).
Not good for: a dose-response. Four, five or six-hour nights, the ordinary case for an athlete, were not tested; neither was recovery after the following night's sleep.
Not good for: anything about endurance, mitochondria, VO2max or an interval session. None was measured, and no exercise took place in the protocol.
Not good for: advice by sex. Every man went down and some women did not, but the study was neither designed nor powered to compare, and the authors say the pattern is not conclusive.
Sex and hormones (inference): hormone exposure moved in the catabolic direction in a group where every one of the men had a lower testosterone exposure after the night awake; whether low testosterone for part of a day changes synthesis rate at all was not tested here, and the authors call it an open question.
Limitations
A marker, not an outcome. A two-hour postprandial synthesis rate is not muscle mass, strength or performance; one day's difference cannot be projected onto a training block.
Single-tailed tests on the primary outcome, with p = .040. The hormone area-under-the-curve comparisons were also single-tailed, while the testosterone interaction term (p = .063) did not reach significance.
n = 13, non-athletes, tested once in each condition, with wide spread in the hormone data (testosterone area under the curve 6.40 ± 5.28 AU).
The conditions differ in more than sleep. The control night was at home and the deprivation night in a laboratory under observation, with light at 211 ± 14 lux, sedentary activity, and snacks allowed overnight; the authors name the night-time intake as a confounder.
Mildly sleep-restricted at entry. Actigraphy gave 5.9 and 6.1 h in the weeks before the arms, so both arms came in mildly short of sleep, although the night directly before each trial was 6.8 and 7.4 h, close to the 7 h they reported.
Sleep measured by actigraphy, not polysomnography; no overnight blood samples, so the hormones of the night itself were not measured in either condition.
Reporting. The abstract gives p = .030 for the cortisol increase and the results give p = .011 for the same comparison; the percentage changes agree.
Not blinded, and neither participants nor researchers could be. Expectation effects on a biochemical rate are unlikely but untested (our inference).
How this entry was checked and changed
Why this level
a crossover with n = 13 earns B by design (rubric: crossover under 15), lowered one step because the outcome is a two-hour molecular rate rather than muscle mass, strength or performance, the primary comparison used single-tailed paired t tests, and p = .040 sits close to the threshold. Read from the full text (CC BY)