Effect of Menstrual Cycle Phase on Perceived Exertion During Aerobic Exercise in Eumenorrheic Women: A Systematic Review and Meta-analysis

Abstract

Background:

The rating of perceived exertion (RPE) is a readily available and practical tool widely used in exercise science to monitor exercise load, but a rigorous review of the effect of menstrual cycle (MC) phases on RPE within continuous aerobic exercise has not yet been completed.

Objective:

This study investigated the effects of the MC phase on RPE during aerobic exercise.

Study Design:

This was a systematic review and meta-analysis.

Methods:

The search strategy was carried out using the 5 most common scientific databases. While qualitative analyses were performed in all included studies, random effects to standard mean difference were calculated and meta-analysis was performed where possible. This study addresses comparison for RPE at the beginning, middle, and end of the exercise adopting 2 mains analysis. The first adopted early cycle (first session of the cycle) as control compared with the subsequent phases, and the second adopted days 1 to 5 (early follicular) as control compared with the subsequent phases.

Results:

A total of 17 studies (n = 160) were included in the qualitative synthesis. The meta-analysis showed that MC phases did not impact RPE (P > .05).

Conclusions:

The current meta-analysis showed that MC does not impact RPE. Although acute RPE is not impacted by MC phases, future studies and practitioners should pay attention to the impact of RPE session by session throughout the MC.

INTRODUCTION

Borg’s rating of perceived exertion (RPE) scale^1–3 has been used extensively to understand psychophysiological responses in exercise science. RPE represents the self-perception of physical work from integrative peripheral and central cues of the body (ie, with increases in physical strain, the person creates perceptual estimates generating outcomes measured using psychophysical ratio scales).^1–3 Over many years, RPE is highly related to physiological parameters, such as blood lactate, heart rate, ventilation, and oxygen uptake,^4–9 making its use in the prescription and management of exercise load more accessible.

Since the creation of RPE, it has served as a validation tool to help monitor the athlete’s psychological state,¹⁰ intensity, and internal workload in the exercise,^11–14 and they are adapting to monitor the training of strength, aerobic, and resistance.^11,12 In experimental trials, it has been demonstrated that temperate conditions,¹⁵ acute altitude,¹⁶ and sex^17,18 alter RPE. When comparing men and women with similar physical activity profiles, women have been reported to perceive exercise as being more difficult than men.¹⁷ Since the 1970s, researchers have considered the menstrual cycle (MC) as a potential factor to affect the RPE response of women performing aerobic exercise.¹⁹

The MC is characterized as a component of women’s reproductive system, with specific milestones controlled by the hypothalamic-pituitary-gonadal axis, which triggers sexual hormones, such as estrogen (E₂) and progesterone (P₄), throughout the cycle.²⁰ A typical MC ranges from from 21 to 35 days and is divided into 2 basic phases, follicular (1st to 14th days—high E₂, low P₄, and low basal body temperature [BBT]) and luteal (15th to 28th days—low E₂, high P₄, and high BBT). While this major approach generally divides the cycle in half according to menses and ovulation, the subphase classification method based on hormonal fluctuations is considered a more accurate approach to identify subtle changes throughout the cycle,^20,21 such as low E₂, high P₄ and BBT (menses); rising E₂, low P₄ and BBT (early-to-late follicular phase); peak E₂, low P₄ and rising BBT (ovulatory phase); secondary lesser peak of E₂, peak P4 and high BBT (early-to-mid-luteal phase); and falling E₂, falling P₄ and falling BBT (premenstrual period; for details see Figure 1).

Figure 1.

MC events. The illustrative hormone lines are according to the behavior of FSH, LH, E₂, and P₄ throughout 1 complete MC of 28 days. BBT indicates basal body temperature; E₂, estrogen; FSH, follicle-stimulating hormone; LH, luteinizing hormone; MC, menstrual cycle; P₄, progesterone.

Gamberale et al¹⁹ were the first to study the influence of MC’s hormonal fluctuations on RPE, where they found higher RPE in the midfollicular phase when comparing menstrual and postmenstrual periods. In later exercise studies, participants had higher RPE in the early follicular phase,²² while others showed higher RPE in the midluteal phase.²³ In contrast, some studies found no difference between phases of the MC on RPE responses to exercise.^24–26 Changes in RPE are expected mainly in the second half of the cycle (ie, low E₂, high P₄, and high BBT) due to the sensibility of hormonal receptors in brain areas (eg, prefrontal cortex, hippocampus, and amygdala), and by decreases in the excitability of the corticospinal tract (nervous system) at the level of the brain, hampering motor performance (eg, running).²⁷

A study by Marsh and Jenkins²⁸ highlighted the possibility that the heterogeneity of the designs and methods between previous studies contributed to these conflicting results around MC’s effects on RPE. However, the previous hypothesis²⁸ as well as others^29,30 was researched without a systematic strategy, which limits the overall utility of these reviews and creates a need for systematic research on this topic. Another consideration is that previous reviews focused on the impact of the MC on RPE in the overall session, and did not consider the possible impact at different moments of the session (ie, beginning, middle, and end). Therefore, we chose to conduct a systematic review and meta-analysis of studies that investigated the effects of the MC phase on RPE during aerobic exercise in eumenorrheic women. This study’s intent is to provide valuable information for exercise prescription and intensity control, exclusively for women.

METHODS

The current systematic review and meta-analysis was conducted following Cochrane and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analysis) guidelines.³¹ The eligibility of article inclusion was based on PICO (Population, Intervention, Comparison, and Outcome) criteria³² and the screening of methodological quality was based on Cochrane Handbook for Systematic Reviews of Interventions guidelines.³³ The current research was registered on the International Prospective Register of Systematic Reviews (PROSPERO) platform (Process #CRD42020172271).

Inclusion Criteria

Studies included were original publications in the English language. Considering that the aim of the present study is to compare MC impact on RPE during aerobic exercise, we only included studies with a crossover design (ie, each participant performed at least 2 sessions over the MC). Participants were eumenorrheic women of reproductive age, healthy, ranging from recreationally active to moderately trained, and not taking oral contraceptives. Prospective participants in perimenopause and menopause were excluded. Considering that aerobic exercise is a strategy that uses large muscle groups, it can be maintained continuously and is rhythmic in nature, and thus, for this study we just included constant-load aerobic exercise (eg, cycling, running, and swimming). Studies investigating resistance exercise, time trials, and intermittent and maximal graded tests were excluded. Studies comparing the major MC phases (ie, follicular and luteal phase) or between subphases (eg, menses phase, early follicular phase, late follicular phase, and midluteal phase) were included. Considering that RPE 6 to 20 points² is the most utilized scale to represent exertion during aerobic exercise with the highest reliability between perceptual and workload and physiological responses,^34,35 only studies that used this tool were included. Results derived from 1 to 20 points,³⁶ the Borg’s category ratio scale (Borg’s CR-10)¹ or other scales adapted to specific exertion (eg, legs and arms), as outcomes were not included.

Search Strategy

Literature searches were performed on PubMed, PsycINFO, Scopus, SPORTDiscus, and Web of Science, using Boolean search operations combined with MeSH terms for MC, physical exercise, and RPE (searches ended September 26, 2022). The specific search terms are reported in electronic Supplemental Digital Content Appendix A (available at: http://links.lww.com/JWHPT/A108). Gray search strategy was conducted according to field experts and citations within the selected studies.

Data Screening and Extraction

The Mendeley Desktop software (v 1.19.4, 2008–2019) was used to check and exclude any duplicate citations detected. Independently, 2 researchers (R.P. and R.S.) verified the eligibility criteria by reading the title and abstract of the articles. In cases of disagreement, a third reader (R.A.) was utilized. The data extraction was performed independently by 2 researchers (R.P. and R.S.) and reported as subject characteristics (sample size, age, sample profile, $\dot{\text{V}}{\text{o}}_{2\text{peak}/\text{max}}$, time without oral contraceptive, and hormone profile), MC characteristics (MC length, identification methods of the MC phases, and investigated phases), exercise characteristics (exercise model, exercise protocol, and environmental conditions), and RPE outcomes.

The GetData Graph Digitizer (v 2.26.0.20) was used to extract RPE in the studies that reported their data only in figure form. In cases of missing data, a researcher (R.P.) contacted the corresponding author to access the information.

Methodological Assessment

Similar to data screening procedures, the methodological quality was carried out jointly by 3 researchers (R.P., R.S., and R.A.) following standard criteria according to the Cochrane Handbook for Systematic Reviews of Interventions guidelines.³³ First, the selected articles were blinded for their title, author’s name, journal, and publication year by a person not involved in this research project. Methodological bias was considered for selection (no randomized design between MC phases), allocation sequence concealment (blinding participants), detection (blinding statistical analysis), attrition (incomplete data), and reporting (interest conflict and selective reporting). Methodological bias was classified as low risk, unclear risk, and high-risk and was expressed individually for the included studies.

Methodological accuracy of MC phase identification utilized recommendations outlined by Allen et al.²⁰ The quality of different approaches used by each study was scored as any self-report methods = 1 point, BBT, and luteinizing hormone [LH] indication/concentration = 2 points, salivary urine and serum concentration of estrogen and progesterone = 3 points, and sonography method = 4 points. For any study that used 2 or more of these approaches, a sum of scores was created and performed by the authors. Study accuracy was based on Forsyth and Reilly³⁷ and Freemas et al,³⁸ in which accuracy categories were created: low accuracy = 1 point (only self-report methods), moderate accuracy = 2 to 4 points (eg, LH concentration + BBT = 4 points), and high accuracy = 5 to 8 points (eg, BBT + LH concentration + salivary urine = 7 points). After completing the data extraction and methodological assessment, blinding of the articles was removed to include information concerning publication authors and years.

MC Phases Classification

To avoid the heterogeneous subclassification of phases adopted among studies, in which studies used a different range of days (eg, early follicular [days 1–7], [days 1–5], or [days 1–3]), our classification was based on biological fluctuations throughout the MC (see Figure 1) adopting the same approach of McNulty et al.³⁹ Therefore, we classified MC subphases of each study as early follicular (days 1–5), late follicular (days 6–12), ovulation (days 13–15), early luteal (days 16–19), mid-luteal (days 20–23), and late luteal (days 24–28).

Data Analysis

All descriptive findings regarding RPE were included in the current systematic review, while the meta-analysis included only studies that provided MC and RPE data. Due to nonstandardized measurement units between studies concerning outcomes of $\dot{\text{V}}{\text{o}}_{2\text{peak}/\text{max}}$, E₂ and P₄, each measure was standardized and adjusted to mL · kg−¹ · min−¹, pg · mL−¹, and ng · mL−¹, respectively. The meta-analysis of standard mean difference (SMD) was conducted using 2 different strategies, the first adopting the first session of the cycle (early MC) as control compared with the other phases, the second adopting early follicular as control compared with the other phases. For both methods, 3 moments of the exercise were analyzed: the beginning of exercise (first RPE measure), middle of exercise (at 50% or at the measurement point closest to 50% of exercise), and the end of exercise (last RPE measure). The χ² and I² were used to assess the heterogeneity according to Higgins et al’s⁴⁰ guidelines. The data were processed in R language (v 4.11–0) using meta package.⁴¹

RESULTS

Search Results

From 250 studies identified in 5 databases, we excluded 54 duplicates, 148 by screening the title and abstract, and 37 by the eligibility of full text. Six additional studies were included through the gray literature strategy. These decisions resulted in 17 studies^{19,23,24,38,42–54,} published between 1975 and 2021 included in the qualitative synthesis. Contact with corresponding authors regarding 6 studies with insufficient data ultimately led to their exclusion. Therefore, 11 studies^{23,24,38,43,45,46,48,49,52–54} were included for meta-analysis. These data are plotted in the PRISMA flowchart (Figure 2).⁵⁵

Figure 2.

PRISMA flowchart of study selection and eligibility criteria.

Study Characteristics

One hundred sixty physically active, trained women (range 18–33 years) were examined in studies (Table 1). Physical activity level was based on self-description within each study. Based on MC length and information provided by the authors, it is possible to assume all participants had a regular MC. All studies adopted at least one method of determination of MC phases, in which blood hormone indicators and self-reports were the most common. Eight measured estrogen and/or progesterone, with a particular study⁴⁸ accessed by urine rather than blood/saliva. Different approaches in the MC phase classification confirm a heterogeneity among studies, with some of them adopting the menstrual period (range 1–3 day), others utilizing early follicular (range 2–6 day) and midfollicular (range 6–10 day). As expected, estrogen and progesterone increased over the MC (ie, follicular-ovulatory-luteal).

Table 1.

Sample Characteristics and MC Measurement^a

Authors	Sample Characteristic			MC Measurement
	n	Age, y	$\dot{\text{V}}{\text{o}}_{2\text{max}/\text{peak}}(\text{mL}\cdot \text{k}{\text{g}}^{-1}\cdot {\text{min}}^{-1})$	MC Length, d	Methods of Determining MC Phase	MC Phases Investigated	Estrogen (pg · mL−1)		Progesterone (ng · mL−1)
							Follicular Phase	Luteal Phase	Follicular Phase	Luteal Phase
Bailey et al42	9	27.0 ± 6.8	49.6 ± 4.3	…	Serum estrogen and progesterone	FP (1–8 days after onset of menses) and LP (19–24 d after onset of menses)	37.0 ± 5.0	84.5 ± 5.6	0.6 ± 0.1	7.9 ± 1.1
Beidleman et al43	8	33.0 ± 3.0	[I] FP = 46.8 ± 4.0 and LP = 46.3 ± 5.6 [II] FP = 33.3 ± 3.7 and LP = 33.8 ± 3.7	28.0 ± 2.0	Self-report and LH urine surge	EF (3–6 d after onset of menses) and ML (6–9 d after the LH surge)	[I] 39.0 ± 22.0 [II] 53.0 ± 12.0	[I] 112.0 ± 38.0 [II] 136.0 ± 54.0	[I] 0.5 ± 0.02 [II] 0.7 ± 0.3	[I] 14.1 ± 9.3 [II] 11.7 ± 6.2
De Souza et al44	8	29.0 ± 4.2	53.4 ± 4.1	28.2 ± 2.2	LH urine surge, plasma estrogen, and progesterone	EF (2–4 d after onset of menses) and ML (6–8 d after onset of LH surge)	41.8 ± 19.4	149.6 ± 70.3	0.5 ± 0.3	12.8 ± 4.4
Eston and Burke45	21	21.7 ± 2.6	39.3 ± 4.4	29.0 ± 3.34	Self-report and BBT	M (2–3 d of menses), MF (6–9 d after onset of menses), ML (6–9 d after ovulation), and PM (72 h prior to onset of menses)	…	…	…	…
Freemas et al38	12	24.8 ± 5.6	40.1 ± 11.0	29.0 ± 3.0	Self-report, BBT, LH urine surge, salivary estrogen, and progesterone	MF (d 6–9 of the cycle) and ML (4–9 d after ovulation)	167.0 ± 55.0	206 ± 120	0.38 ± 0.1	1.0 ± 0.5
Galliven et al46	8	31.0 ± 1.0	46.2 ± 2.0	…	Plasma estrogen and progesterone	FP (3–9 d after onset of menses), midcycle (10–16 of menses onset) and LP (18–26 d after onset of menses)	54.5 ± 7.9 Midcycle = 133.2 ± 19.9	108.4 ± 15.5	0.3 ± 0.1 Midcycle = 2.1 ± 0.6	9.5 ± 0.9
Gamberale et al19	12	27	36.63	…	Self-report	M (1–2 d of menses), post-menstruation (10–18 d after onset of menses), and PM (6–2 d prior to onset of menses)	…	…	…	…
Garcia et al47	4	[I] 22.5 ± 1.7	[I] 39.6 ± 8.8	[I] 30.3 ± 1.2	Self-report, BBT, and serum progesterone	FP (5–8 d after onset of menses) and LP (22–25 d after onset of menses)	…	…	…	[I] LP = 8.3 ± 6.3
Hackney et al48	9	22.2 ± 1.7	46.0 ± 2.6	29.5 ± 1	Self-report, BBT, and retrospective self-report calendar	MF (6–9 d after onset of menses) and ML (21–24 d after onset of menses)	13.2 ± 3.2 ng · mL−1	50.4 ± 16.5 ng · mL−1	0.41 ± 0.10 μg · mL−1	5.76 ± 3.22 μg · mL−1
Hackney et al24	8	25.0 ± 4.0	57.5 ± 3.5	28 ± 3	Self-report, blood estrogen, and progesterone	MF (8 ± 2 d after onset of menses) and ML (23 ± 3 d after onset of menses)	…	…	…	…
Janse De Jonge et al49	8	23.7 ± 4.1	40.0 ± 6.9	…	Self-report, blood estrogen, and progesterone	EF (d 5–7) and ML (21–22)	[I] 33.7 ± 10.2 [Ii] 33.6 ± 9.8	[I] 107.4 ± 17.9 [II] 102.4 ± 27.1	[I] 0.4 ± 0.1 [II] 0.4 ± 0.3	[I] 11.0 ± 3.4 [II] 11.6 ± 5.7
O’Leary et al52	10	20.0 ± 2.2	50.7 ± 9.0	30.0 ± 3.0	Self-report	MF (7 ± 2 d after menses) and ML (20 ± 2 d after menses)	27.0 ± 6.1	67.6 ± 21.7	…	…
Meyer et al50	4	27.8 ± 2.5	51.2 ± 2.0	…	Self-report	M (1–2 d after onset of menses) and no menstrual (19–21 d after onset of menses)	130.4	210.4	…	…
Nicklas et al51	6	26.3 ± 2.4	44.9 ± 1.7	…	Self-report and BBT	MF (7–8 d after onset of menses) and ML (7–8 d after ovulation)	48.9 ± 13.0	188.5 ± 70.2	0.6 ± 0.2	10.5 ± 2.5
Pivarnik et al23	9	27.2 ± 3.7	42.5 ± 6.7	…	LH urine surge and blood progesterone	MF (~7 d prior to ovulation) and ML (7 d after ovulation)	…	…	0.4 ± 0.2	9.1 ± 4.1
Prado et al53	14	24.3 ± 4.2	41.6 ± 6.5	28.1 ± 3.0	Self-report and BBT	FP (1–5 d after menses) and LP (1–5 d prior to onset of menses)	…	…	…	…
Williams et al54	10	21.0 ± 1.0	53.5 ± 4.7	28.0 ± 1.0	Self-report	MF (4–10 d after onset of menses) and ML (20–27 d after onset of menses)	39.8 ± 18.3	148.1 ± 35.2	…	…

Abbreviations: BBT, basal body temperature; EF, early-follicular; FP, follicular phase; LH, luteinizing hormone; LP, luteal phase; M, menstrual; MC, menstrual cycle; MF, midfollicular; ML, midluteal; PM, premenstrual.

^a[I] and [II] show different experiments within the same study.

Table 2 summarizes the characteristics and conclusions of studies. Studies conducted exercise protocols on the treadmill or bicycle ergometer. Except for three studies that investigated fixed intensity tests until exhaustion,^42,43,51 all tests utilized fixed time protocols. Four studies^19,23,49,53 concluded a significant effect of the MC on RPE, with three^23,49,53 observing higher RPE during the luteal phase compared with the follicular phase, and one¹⁹ observing higher RPE during menstruation compared with the follicular and luteal phases. In one study,⁴⁹ which manipulated environmental conditions (temperate vs hot, humid), the effect of the MC on RPE was observed only in the hot, humid condition, with higher scores during the luteal phase compared with the follicular phase.

Table 2.

Exercise Characteristics and Outcomes of Studies^a

Authors	Ergometer	Intensity Variable	Protocol	Environment Temperature, °C	Environment Humidity, %	Study Conclusion
Bailey et al42	Bicycle	$\dot{\text{V}}{\text{o}}_{2\text{peak}}$	70% $\dot{\text{V}}{\text{o}}_{2\text{peak}}$ until exhaustion	22.7 ± 1.6	38.0 ± 19.0	↔
Beidleman et al43	Treadmill	$\dot{\text{V}}{\text{o}}_{2\text{peak}}$	70% $\dot{\text{V}}{\text{o}}_{2\text{peak}}$ until exhaustion	22.0 ± 3.0	45.0 ± 5.0	↔
De Souza et al44	Treadmill	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	40 min at 80% $\dot{\text{V}}{\text{o}}_{2\text{max}}$	…	…	↔
Eston and Burke45	Bicycle	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	[I] 3 min at 70% [II] 3 min at 90%	…	…	↔
Freemas et al38	Bicycle	AT	[I] 5 min at 10% below AT [I] 5 min at 10% above AT (interval 5 min)	…	…	↔
Galliven et al46	Treadmill	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	20 min at 70% $\dot{\text{V}}{\text{o}}_{2\text{max}}$	…	…	↔
Gamberale et al19	Bicycle	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	6 min at 40 and 70% $\dot{\text{V}}{\text{o}}_{2\text{max}}$	…	…	↑ M vs LP and FP
Garcia et al47	Bicycle	Wpeak	60 min at 60% Wpeak	32	80	[I] ↔ [II] ↔
Hackney et al48	Treadmill	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	10 min at 35% $\dot{\text{V}}{\text{o}}_{2\text{max}}$	…		↔
Hackney et al24	Treadmill	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	90 min at 70% $\dot{\text{V}}{\text{o}}_{2\text{max}}$	Environment controlled		↔
Janse De Jonge et al49	Bicycle	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	45 min at 60% $\dot{\text{V}}{\text{o}}_{2\text{max}}$	[I] 20 [II] 32	[I] 45 [II] 60	[I] ↔ [II] ↑ LP vs FP
O’Leary et al52	Treadmill	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	min at 60%−70% $\dot{\text{V}}{\text{o}}_{2\text{max}}$			↔
Meyer, et al50	Treadmill	$\dot{\text{V}}{\text{o}}_{2\text{max}}$	60 min at 65%−70% $\dot{\text{V}}{\text{o}}_{2\text{max}}$	Environment controlled		↔
Nicklas et al51	Bicycle		70% $\dot{\text{V}}{\text{o}}_{2\text{max}}$ until exhaustion			↔
Pivarnik et al23	Bicycle	$\dot{\text{V}}{\text{o}}_{2\text{peak}}$	60 min at 65% Wpeak	23.3 ± 0.9	60.0 ± 6.0	↑ LP vs FP
Prado et al53	Treadmill	AT	[I] 15 min at a speed 10% above AT [II] 15 min at a speed 20% below AT (interval >48 h)	22–25	60–65	[I] ↑ LP vs FP [II] ↑ LP vs FP
Williams et al54	Treadmill	$\dot{\text{V}}{\text{o}}_{2\text{peak}}$	60 min at 65% $\dot{\text{V}}{\text{o}}_{2\text{peak}}$	…	…	↔

Abbreviations: AT, anaerobic threshold; FP, follicular phase; LP, luteal phase; M, menstruation; $\dot{\text{V}}{\text{o}}_{2\text{max}}$, maximum oxygen consumption; Vo_2peak, peak oxygen consumption; W_peak, power peak; ↑, significant increase; ↔, no significant difference.

^a[I] and [II] show different experiments within the same study.

Early MC Versus Other Phases

Figure 3 illustrates the methodological assessment and meta-analysis findings from early MC versus other phases at the beginning, middle, and end of the exercise.

Figure 3.

Meta-analysis of the effect of early MC versus other phases on RPE and methodological assessment results. CI_95% indicates confidence interval 95%; IV, inverse of variance; MC, menstrual cycle; SMD, standard mean difference.

• Beginning of the exercise. Two outliers (SMD = ~2), one favoring the early MC and another favoring the other phases, were excluded. From the remaining four studies,^49,52–54 six SMDs were extracted at the beginning of the exercise. The meta-analysis shows a nonsignificant effect (Z = −0.66; P = .51; SMD = −0.12; IC_95%= −0.46, 0.23) of MC phases on RPE. Low heterogeneity was identified (ι²= 0.0, χ²= 0.77, I²= 0%; P = .98).

• Middle of the exercise. From 6 studies,^{23,24,49,52–54} 8 SMDs were analyzed at 50% duration of exercise. The meta-analysis shows a nonsignificant effect (Z = −1.68; P = .09; SMD = −0.27; IC_95%= −0.58, 0.04) compared with other phases. Low heterogeneity was identified (ι²= 0.0, χ²= 2.13, I²= 0%; P = .95).

• End of the exercise. From 11 studies,^{23,24,38,43,45,46,48,49,52–54} 21 SMDs were extracted at the end of the exercise. The meta-analysis shows a nonsignificant effect (Z = −0.19; P = .85; SMD = −0.02; IC_95%= −0.19, 0.15). Low heterogeneity was identified (ι²= 0.0, χ²= 15.58, I²= 0%; P = .74).

Early Follicular Phase Versus Other Phases

Figure 4 illustrates meta-analysis findings from the early follicular phase versus other phases at the beginning, middle, and end of the exercise.

Figure 4.

Meta-analysis of the effect of the early follicular phase versus other phases on RPE. CI_95% indicates confidence interval 95%; IV, inverse of variance; SMD, standard mean difference.

• Beginning of the exercise. From two studies,^49,54 3 SMDs (n = 26) were extracted at the beginning of the exercise. The meta-analysis shows a nonsignificant effect (Z = −0.42; P = .67; SMD = −0.12; IC_95%= −0.66, 0.43) of MC phases on RPE. Low heterogeneity was identified (ι²= 0.0, χ²= 0.60, I²= 0%; P = .74).

• Middle of the exercise. From two studies,^49,54 3 SMDs (n = 26) were analyzed at 50% duration of exercise. The meta-analysis shows a nonsignificant effect (Z = −0.98; P = .33; SMD = −0.27; IC_95%= −0.82, 0.28) of MC phases on RPE. Low heterogeneity was identified (ι²= 0.0, χ ²= 0.62, I²= 0%; P = .73).

• End of the exercise. From four studies,^45,46,49,54 11 SMDs (n = 168) were extracted at the end of the exercise. The meta-analysis shows a nonsignificant effect (Z = 1.11; P = .27; SMD = −0.12; IC_95%= −0.09, 0.34). Low heterogeneity was identified (ι²= 0.0, χ²= 4.15, I²= 0%; P = .94).

Methodological Quality

Figure 5 shows the summary of methodological assessment from studies. Overall, high quality of evidence from the included studies was observed, with 71% considering randomization of MC phases, 71% incomplete outcome data, and 94% involving selective report criteria. One study⁴² investigated the effect of an ergogenic supplement provided by the manufacturer, which may be classified as selective reporting bias. Furthermore, 94% showed an unclear risk of bias for blinding statistical analysis. A high accuracy in the strategies of the studies for the identification of the MC phase was also observed.

Figure 5.

Summary of the risk of bias and methodological accuracy of MC phase identification of the studies. MC indicates menstrual cycle.

DISCUSSION

The current systematic review and meta-analysis investigated the effect of MC phases on RPE responses throughout constant-load aerobic exercise. The main results of the study indicate that there is no impact of MC phases on RPE.

Similar findings described previously were also obtained by other studies^28–30; however, methodological and statistical limitations in these studies were overcome in our analysis approach. For example, while Paludo et al²⁹ pooled different sports modalities, exercise protocols, and RPE tools (eg, 6–20 and 0–10 points), we standardized these variables including only aerobic continuous exercise and use the RPE 6- to 20-point scale. Moreover, considering that our study pooled only female cyclists and runners physically active and trained, the extrapolation of present findings to other groups needs to be taken with caution.

Three studies^23,49,53 observed a higher RPE during the luteal phase compared with the follicular phase in which Janse De Jonge et al⁴⁹ investigated MC and RPE during hot and humid conditions. Curiously, Garcia et al⁴⁷ also investigated these responses under hot and humid conditions but no differences were observed. In this case, a speculative explanation is that an optimal thermoregulation profile during exercise could have been the determining factor in inhibiting the impact of the combination of the increase of central temperature (progesterone effect), heart, and ventilation rate on RPE among studies.

Higher RPE during menstrual bleeding compared with luteal and follicular phases was observed by only one study.¹⁹ Considering that previous studies suggest that feelings of discomfort may impact RPE, we speculate the results of Gamberale et al¹⁹ could be due to bodily discomfort caused by menstrual bleeding and/or using different sizes of sanitary napkins,^56,57 both factors (discomfort and sanitary napkins) not explored in this study.¹⁹

While estrogen has a protective role in perceptual responses, progesterone acts by reducing the prefrontal cortex’s ability (decrease in reactivity) to inhibit negative perceptual responses from the amygdala (increase reactivity).⁵⁸ In studies that did not evaluate sexual hormones (9 studies), we cannot confirm the normal hormonal fluctuations over MC in their participants in this review, and thus, the absence of these data is considered a limitation.

The studies included herein assessed a small sample and compared exclusively RPE within subjects during one MC. This approach is a limitation as it decreases the likelihood of reliable inferences due to an increase in the risk of expectancy effects on acute RPE responses. Therefore, future replication studies investigating cycle-by-cycle effects are recommended.

A gap was observed in the body of studies that investigate MC effects on RPE 6–20 points during continuous aerobic exercise with most studies published more than 7 years ago (~47 years), and thus, it is still necessary for new studies to update and advance knowledge in this field.

We observed some limitations in the studies included in this review. First, the absence of descriptions of familiarization/control about how and whether participants interpreted correctly the RPE concept may have biased the results. Second, blinding of MC phases for participants themselves is certainly difficult (if not impossible); however, blinding of the session (to participants and researcher), MC phases, and statistical analysis (to researcher) could be reduced by human bias effects.^{33, 59} Third, future studies should reduce the large heterogeneity in MC phase classification approaches by adopting standard recommendations for studies in sports and exercise science with female participants.⁶⁰

Finally, the present review examined physically active and healthy women (not users of hormone replacement) who were free of medical conditions related to MC dysfunction (eg, premenstrual syndrome and premenstrual dysphoric disorder). Therefore, the results found here may be unlike those that could be observed in other groups. Future studies focused on exercise training should consider investigating how small changes in RPE across multiple sessions that span many MC phases might impact behavioral responses.

CONCLUSIONS

To our knowledge, this is the first review to adopt a meta-analysis approach examining RPE and MC at different moments of constant-load aerobic exercise in healthy women. Based on the findings of this review, we conclude that regardless of the analysis approach (early vs end of MC and follicular phase versus luteal phase), MC does not impact RPE during exercise. Finally, research in this area needs greater attention and consideration. As such, we reinforce the need for further investigation on this topic.