The Effects of Menstrual Cycle Phase on Exercise Performance in Eumenorrheic Women: A Systematic Review and Meta-Analysis
Participants
78 included studies of eumenorrheic women; 51 studies entered the initial pairwise meta-analysis and 73 the network meta-analysis. Pooled participant count is not given in the abstract.
Protocol / Analysis
Phases compared: early follicular vs all other phases (late follicular, ovulation, early luteal, mid-luteal, late luteal) in the initial pairwise analysis; all phases against each other in the network meta-analysis, ranked by SUCRA
Outcomes: endurance- and strength-based exercise performance
Study quality: modified Downs and Black checklist plus a GRADE-based strategy
Key Results
Early follicular vs all other phases pooled: trivial reduction, median ES = −0.06 (95% CrI −0.16 to 0.04) — the credible interval includes zero
Largest single contrast: early follicular vs late follicular, ES = −0.14 (95% CrI −0.26 to −0.03) — still trivial in size, but this interval does not include zero
SUCRA: the early follicular phase had the lowest value (30% — most likely to be the poorest phase); all other phases 53-55%, i.e. indistinguishable from one another
Quality of evidence for the review graded "low" (42%)
Authors' conclusion: performance "might be trivially reduced during the early follicular phase"; with trivial effects, large between-study variation and many poor-quality studies, general guidelines cannot be formed — a personalised approach is recommended
Mechanism / Context
Hypothesized: oestrogen (peaking late follicular) and progesterone (luteal) fluctuate and could influence performance
Inference, not tested: the pooled performance effects are too small and too variable to confirm any mechanistic prediction at group level
Practical Application
Across 78 studies of eumenorrheic women, performance in the early follicular phase (bleeding days) was only trivially lower than in the other phases (ES −0.06, CrI −0.16 to 0.04); the biggest gap, early vs late follicular, was −0.14 — so there is no group-level basis for phase-specific training prescriptions. The authors recommend a personalised approach based on the individual's own response.
Do not reduce training automatically during menstruation "because the meta-analysis says so" — it says the opposite: the effect is trivial and inconsistent.
Some women may show clear phase sensitivity and others none; decide from the athlete's own symptom and performance record.
Not from this paper; the base's own practice: track symptoms and performance for at least three cycles before making any phase-specific adjustment.
Limitations
Quality of evidence graded low (42%); many included studies were of poor quality
Large between-study variation; the abstract does not report the proportion of studies with hormone-verified phases
Training status of the pooled cohorts is not broken down in the abstract
Oral-contraceptive users are a separate population (see REF-338)
Phase definitions varied across studies
How this entry was checked and changed
Content correction 2026-09-27
">1,100 participants pooled" -> not in the abstract; removed (78 studies, 51/73 in the analyses).
"SMD ≈ −0.06 to −0.15, all CIs overlap 0" -> −0.06 (CrI −0.16 to 0.04) and −0.14 (CrI −0.26 to −0.03); the second interval does NOT include zero. "I² ≈ 80%" -> not in the abstract; removed.
"Only 8 of 78 studies hormone-verified", "~20-40% misclassification", "EF 1-7 vs 1-5 days" -> not in the abstract; removed. Three-phase model -> the six phases the abstract lists.
Tier "4-5 Elite" -> Tier 1-3 (the entry's own limitations said elite were under-represented); Sport "swimming, sprint" -> general. "≥3 cycles" relabelled as the base's own practice.