Tip · Tracking

Aligning training with the menstrual cycle (women)

Q: Am I really stronger in the follicular phase?

Not measurably. The meta-analysis of 78 studies found, as the largest single effect, an effect size of −0.14 between the early and late follicular phase, which is considered trivial. The ranking of the phases shows the same: only the early follicular phase, that is, menstruation, stands out with a SUCRA value of 30 percent; all other phases lie between 53 and 55 percent and cannot be distinguished from one another. There is no strength phase in the data.

Q: Should I train more lightly in the luteal phase?

There is no basis for that. In the meta-analyses, the weakest phase is not the luteal phase but the early follicular phase. And at the molecular level, muscle protein synthesis after strength training did not differ between the late follicular and mid-luteal phases, measured with stable isotopes and muscle biopsies.

Q: Why does the scheme of days 1 to 14 and 15 to 28 not fit me?

Because it does not fit a great many women. In the analysis of more than 600,000 ovulatory cycles, the follicular phase lasted 16.9 days on average, with a range of 10 to 30 days, and the luteal phase 12.4 days on average, with a range of 7 to 17 days. Cycle length also shortens with age, by about 0.18 days per year of age between 25 and 45.

Q: Which app can I use?

It depends on what it evaluates. In a structured assessment of twelve apps, purely calendar-based apps scored 0 out of 30 points, because they only calculate from past cycles. Symptothermal apps, which evaluate basal body temperature and cervical mucus in the current cycle, reached 11 to 20 out of 30 points. For detecting ovulation, the physiological measurement of a ring also performed well: 96.4 percent of ovulations detected, mean error 1.26 days versus 3.44 days with the calendar method.

Q: Does the pill make a difference for muscle building?

According to the available evidence, no. A meta-analysis of eight studies with 325 participants found no effect of oral contraceptives on muscle building, power or strength; all confidence intervals included zero. So there is no evidence-based reason to argue for or against the pill when it comes to strength training.

Q: If the effect is so small, why do the recommendations persist?

Because the hormone patterns are plausible and because the studies were of poor quality for a long time. The meta-analysis rates the quality of evidence at 42 percent and names the small samples, the lack of hormone measurements and the inconsistent definition of phases as the reason why no general recommendations are possible. In the review that admitted only studies with hormone measurements, the mean sample size was around 14 people.

Follicular phase for strength, luteal phase for endurance – so the theory goes. Reviews find only small effects with large individual differences.

Cycle-based training means: strength in the follicular phase, endurance in the luteal phase. The tip is included here because both have been properly tested — performance in meta-analyses, muscle building with stable isotopes, and the calendar scheme on 612,613 real cycles.

In short

The phase effect exists, but it is trivially small. Across 78 studies, performance in the early follicular phase was minimally reduced compared with all other phases combined (ES = −0.06; 95% credible interval −0.16 to 0.04); the largest single effect, between the early and late follicular phase, was ES = −0.14. At the molecular level, no difference at all is measurable. And the scheme of days 1–14 versus days 15–28 does not apply to many women: the follicular phase lasted 16.9 days on average, with a range of 10 to 30 days. The only phase with a measured disadvantage is menstruation itself.

What is behind it

Estradiol rises in the follicular phase, reaches a peak shortly before ovulation and then falls; after ovulation, progesterone rises and, to a lesser extent, estradiol again. Training theory derives its hypotheses from this: estrogen is credited with an anabolic and neuromuscularly beneficial role, progesterone with a catabolic role that places a burden on thermoregulation.

When tested, this derivation does not hold up. Measured with stable isotopes and muscle biopsies, muscle protein synthesis and protein breakdown after strength training did not differ between the late follicular and mid-luteal phases. The same research group also found no difference for the phases of the pill.

What the meta-analyses show on performance

The key paper included 78 experimental studies, 51 in the pairwise meta-analysis and 73 in a network meta-analysis, calculated with multilevel models on a Bayesian basis. The result was the same for endurance and strength endpoints: ES = −0.06 (95% CrI −0.16 to 0.04) for the early follicular phase compared with all other phases, largest single effect ES = −0.14 (−0.26 to −0.03).

The ranking of the phases shows it most clearly. The SUCRA value was 30% for the early follicular phase and between 53 and 55% for all other phases. There is no strength phase and no endurance phase in the data, but rather one slightly less favorable phase and an undifferentiated remainder.

Why the calendar scheme does not hold up

Analyzed across 612,613 ovulatory cycles from 124,648 users of a cycle app, recorded via menstruation, basal body temperature and LH tests, reality looks different from the scheme: mean cycle length 29.3 days, mean follicular phase 16.9 days with a 95% range of 10 to 30 days, mean luteal phase 12.4 days with a range of 7 to 17 days.

The variability therefore lies almost entirely in the follicular phase. Age adds to this: between 25 and 45 years, cycle length decreased by 0.18 days per year of age and the follicular phase by 0.19 days. Anyone who plans by calendar days is planning for the wrong phase in a considerable share of cycles.

What is well supported

The test figures are the real yield here. Performance: ES = −0.06 (95% CrI −0.16 to 0.04) across 78 studies, largest single effect ES = −0.14 (−0.26 to −0.03), SUCRA 30% for the early follicular phase versus 53 to 55% for all others. Strength: in a second meta-analysis of 21 studies, all comparisons were non-significant and small to trivial (p ≥ 0.26; Hedges’ g ≤ 0.35; I² = 0%). Muscle building: myofibrillar protein synthesis rate in the trained leg 1.52 ± 0.27% per day in the follicular phase and 1.46 ± 0.25% in the luteal phase — significant training effect (P < 0.001), no phase effect. Phase length: 16.9 days (10–30) versus 12.4 days (7–17) in 612,613 cycles.

What the studies show

McNulty 2020, Sports Medicine

The key paper on the topic. 78 studies included, 51 in the pairwise and 73 in the network meta-analysis. Compared with all other phases combined, performance in the early follicular phase was minimally reduced (ES = −0.06; 95% CrI −0.16 to 0.04), for endurance and strength endpoints alike. SUCRA 30% for the early follicular phase, 53 to 55% for all others. Quality of evidence low (42%); the authors explicitly do not make any general recommendations.

Blagrove 2020, J Sci Med Sport

Systematic review with meta-analysis specifically on strength measures. 21 studies; maximal voluntary contraction, isokinetic peak torque and explosive strength were compared between the early follicular phase, ovulation and the mid-luteal phase. All comparisons remained non-significant and small to trivial (p ≥ 0.26; Hedges’ g ≤ 0.35); the confidence intervals ranged from g = −0.42 to 0.48 (I² = 0%).

Colenso-Semple 2025, Journal of Physiology

The methodologically most elaborate single study. 12 young women completed, in randomized order, two six-day phases — late follicular phase and mid-luteal phase — with unilateral strength training. Myofibrillar protein synthesis rate 1.33 ± 0.27% and 1.52 ± 0.27% per day in the follicular phase, 1.28 ± 0.27% and 1.46 ± 0.25% in the luteal phase. Significant training effect (P < 0.001), no phase effect, no difference in protein breakdown (P = 0.24).

Bull 2019, npj Digital Medicine

The most important paper for practice, because it tests the basis of the training scheme. 612,613 ovulatory cycles from 124,648 users, recorded via menstruation, basal body temperature and LH tests. Mean cycle length 29.3 days, mean follicular phase 16.9 days (95% CI 10–30), mean luteal phase 12.4 days (95% CI 7–17).

What the studies do not show

The main problem is study quality, not the direction of the effect. The meta-analysis rates the quality of evidence at 42% and explicitly names the number of low-quality studies as the reason why no general recommendations are possible. A review that admitted only studies with hormone measurements found just 19 papers with 279 participants and a mean sample size of 13.9 ± 7 people; 58% reported a significant phase effect, but direction and size varied.

How to do it

Do not determine the phase by calendar. Given a mean follicular phase of 16.9 days with a range of 10 to 30 days, the fixed division into days 1–14 and days 15–28 does not apply to a considerable share of cycles. If you track, do so with physiological measurements: the authors of the large analysis write explicitly that parameters such as basal body temperature must be recorded. The same criterion separates the useful from the useless apps in an assessment of twelve apps: symptothermal apps 11 to 20 out of 30 points, calculothermal ones 2 and 3 out of 30, purely calendar-based ones 0 out of 30. A ring can detect ovulation physiologically: in 1,155 cycles, the physiological method detected 1,113 (96.4%) of ovulations with a mean error of 1.26 days, while the calendar method was off by 3.44 days. Anyone who wants to adjust something should adjust in the early follicular phase.

Safety

Cycle tracking and training planning are physically harmless. Three points nonetheless belong here. First: anyone who notices over months that menstruation is absent or that no ovulation occurs has a finding that should be investigated — especially with high training volume and low energy availability. An absent cycle is not a side effect of good training but a warning sign. Second: cycle apps are not suitable for contraception if they only evaluate the calendar — in the assessment of twelve apps, the six calendar-based ones scored 0 out of 30 points. Third: with severe pain, very heavy or very long bleeding, intermenstrual bleeding or an absent period, the cause should be examined by a gynecologist.

BK-Score Long used, barely studied

Human evidence		3
Mechanism		5
Safety data		9
Hype gap		3
Track record of use		7

Reviews of cycle-based training reach a sobering conclusion: the effects are small, the studies methodologically weak, and the individual differences greater than the phase effect.

The score rates the state of knowledge, not the effect. “Safety data 9” means well studied – not harmless.
Subjective assessment by Biohacking Kompakt based on published scoring rules – not a scientific rating and not a medical recommendation. Rules and all ratings (German)

Frequently asked questions about aligning training with the menstrual cycle (women)

Am I really stronger in the follicular phase?