The evidence that losing sleep impairs athletic performance is strong and consistent; the evidence that extending sleep beyond normal further improves it is thin, mostly uncontrolled, and explicitly described as tentative by its own reviewers
Interpretation
Silva et al.'s systematic review found only 2 studies met its eligibility criteria for sleep-extension research in athletes, with the authors' own conclusions described as tentative given the limited and mostly high-risk-of-bias evidence. The one study actually conducted in cyclists/triathletes — Roberts et al., 9 athletes — found restricted sleep clearly hurt performance, while extended sleep was only modestly faster than normal sleep, framed as 'maintaining' rather than boosting performance. By contrast, Craven et al.'s much larger meta-analysis (69 studies, 959 participants) found acute sleep loss reduced performance by 7.56% on average, significant across every category including endurance — a far more robust and consistent body of evidence than exists for the extension side of the question.
Practical implication
Lead sleep advice to athletes with the strong, well-evidenced message — protect against sleep loss, which reliably hurts performance — rather than the much weaker claim that extending sleep beyond a normal amount will measurably boost it. The two claims don't have equal evidentiary weight and shouldn't be presented as if they do.
Sources
3 sources, independently verified.
Out of an extensive search, only 2 studies met the review's eligibility criteria for sleep-extension research in athletes (a closer look at the accepted manuscript's own tables shows 7 trials considered, with inconsistent eligibility counts flagged as a limitation). Extension of time in bed ranged 26-106 minutes across studies; 6 of 15 performance measures showed large effects, the rest trivial-to-medium, but risk of bias was high in the one RCT and low in the non-randomised studies. The review's own conclusion is explicitly tentative — this is a thin, mostly male, few-sport evidence base.
The one study in this batch's sleep research actually conducted in cyclists/triathletes: across 4 days per sleep condition (restricted ~4.8h, normal ~6.8h, extended ~8.4h), restricted sleep produced a slower time trial by day 3, while extended sleep was slightly faster than normal sleep by day 4 (56.8 vs 58.7 minutes, p<0.02) — framed by the review above as extended sleep 'maintaining' performance better than it 'boosting' it outright. Very small sample (9 athletes); directional, not definitive.
Sleep loss of 6 hours or less reduced physical performance by 7.56% on average (95% CI -11.9 to -3.13), a significant effect across every performance category tested including endurance. The asymmetry worth noting against the sleep-extension evidence above: the case that losing sleep hurts performance is considerably stronger and more consistent than the case that extending sleep beyond normal helps it.