FE-MALE Study: Female Exercisers – Menstrual (follicular And Luteal) Effects A Pilot Study PDF Free Download
1 / 177/177
100%
FE-MALE Study: Female Exercisers – Menstrual (follicular And Luteal) Effects
A Pilot Study
Michaela Mary Rogan
A thesis submitted for the degree of
Master of Science
Department of Human Nutrition
University of Otago, Dunedin, New Zealand
December 2022
ii
Abstract
Background: Fluctuations in endogenous sex hormones over the menstrual cycle
appear to influence dietary intake. Research suggests that energy intake may be greater
in the luteal phase compared with the follicular phase of the menstrual cycle, while
patterns of macronutrient intake are less clear. However, high-quality evidence is
lacking due to poor adherence to recent best-practice guidelines for menstrual cycle
research. Furthermore, existing research has almost exclusively been conducted in
sedentary populations: the dietary intake and energy availability of female exercisers
across the cycle are unclear.
Aims: This pilot study aimed to provide data on: 1) the variability of habitual energy
availability, energy intake, and macronutrient intake across the menstrual cycle in
female exercisers; 2) phase-related dietary intake in the post-exercise period; and 3)
retention rates to aid future research.
Methods: Six regularly menstruating female exercisers (>2.5hrs·week-1 exercise)
completed three-day weighed image-assisted diet records and training diaries in the
early-follicular phase (days 2-4 of the menstrual cycle) and mid-luteal phase (7-9 days
after a positive urinary ovulation test) for two menstrual cycles. On one day during each
three-day recording period, participants had their body composition measured via
bioelectrical impedance analysis (BIA) and completed a fasted 60-minute cycling time
trial, followed by an ad-libitum meal.
Results: Mean energy availability was slightly greater in the mid-luteal phase
(35.7 ± 7.1 kcal·kgFFM-1·day-1) compared with the early-follicular phase
(32.6 ± 11.5 kcal·kgFFM-1·day-1) of the menstrual cycle. When one participant with a
dramatic ~600 kcal·day-1 decrease in EEE in the mid-luteal phase of both cycles was
excluded, mean energy availability in the early-follicular and mid-luteal phases still
showed the same pattern (31.5 ± 12.5 vs. 34.7 ± 7.5 kcal·kgFFM-1·day-1). Daily
macronutrient intakes were similar between the two phases in both the full, and
reduced samples, but many participants failed to meet sports nutrition guidelines for
iii
carbohydrate and protein in either phase. Energy and macronutrient intakes from the
post-exercise ad-libitum meal were similar between phases, apart from fat, which
increased from the early-follicular phase (20.7 ± 12.0 g) to the mid-luteal phase
(25.5 ± 12.5 g). The rate of menstrual irregularities was very high, causing 43.8% (n=7)
of participants to be excluded.
Conclusion: The trends shown in this pilot study are worthy of further exploration, and
future interventional and observational studies investigating energy availability may
need to consider the effects of the menstrual cycle. Ideally, studies with large sample
sizes and more diverse populations, which adhere to best-practice guidelines for
menstrual cycle research, are required. However, without greater collaboration
between research centres and increased funding for female-specific research, the
feasibility of this is challenging.
iv
Co-Authorship Form
This form is to accompany the submission of any thesis that contains research reported in co-authored work that
has been published, accepted for publication, or submitted for publication. A copy of this form should be at the
front (after the thesis abstract) of the thesis submitted for examination and library deposit.
Expectations of theses including publications:
The thesis must be an integrated and coherent body of work. The body of the thesis may include entire
publications verbatim, but these need to be reformatted to a consistent chapter style, and commentary may need
to be added to link the publications together. If the thesis contains multiple publications, candidates need to
ensure that they have: included an introductory chapter; demonstrated their ability to critically engage with the
literature; carefully considered their research design and justified choices of methodology and methods where
necessary; and synthesised and discussed the findings from the publications. It is the candidate’s responsibility
to ensure that any published work (or parts thereof) included in the thesis comply with the copyright provisions
of the publisher and that any guidelines with regard to self-citation are followed.
Details of publications included in and/or appended to this thesis (please add rows as needed).
Chapter/
Append.
Paper title
Authors
Contribution of candidate
and co-authors – please
detail the nature and
extent (%)
Journal
Status (e.g.
under review,
forthcoming,
published)
Section 2.4.1
Figure 1
Figure 2
Appendix A
(Table 9)
Appendix B
(Table 10)
Dietary energy
intake across the
menstrual cycle: a
narrative review
Michaela M
Rogan
Katherine E
Black
The candidate (MMR) and
supervisor (KEB) both
contributed to the
conception, design,
interpretation, and analysis
of the publication. MMR
drafted the article and KEB
provided critical revision.
Nutrition
Reviews
Published
(advance
online)
Certification by Primary Supervisor:
The undersigned certifies that the above table correctly reflects the nature and extent of the candidate’s
contribution to this co-authored work
Name: Katherine Black Signature: Date: 5/12/2022
v
Preface
This Master of Science thesis was conducted in the Department of Human Nutrition,
University of Otago, New Zealand, under the supervision of Dr Katherine Black. At the
time of writing, the FE-MALE study is ongoing.
Aspects of this thesis formed the basis for the following publication:
Rogan MM, Black KE. Dietary energy intake across the menstrual cycle: a narrative
review. Nutr Rev 2022 Nov 11: (Epub ahead of print; doi: 10.1093/nutrit/nuac094).
From December 2nd 2021 until September 12th 2022 New Zealand was under the
COVID-19 Protection Framework. On January 23rd New Zealand moved to the ‘Red’
traffic light setting, which involved vaccine and mask mandates on University of Otago
campus, capacity limits for indoor settings, physical distancing, contact tracing, and
14-day self-isolation periods for infected individuals or close contacts. Over the
following months, these regulations were gradually amended, with self-isolation
periods reducing to 10 days, and then 7 days; self-isolation requirements shifting from
close contacts to household contacts; removal of contact tracing and vaccine passes; and
increased implementation of rapid antigen tests. On April 13th New Zealand moved to
the ‘Orange’ traffic light setting, as requirements for masks and physical distancing were
reduced, and indoor capacity limits were removed before the COVID-19 Protection
Framework was lifted in September.
COVID-19 impacted this thesis in several ways, including participants and researchers
contracting the virus or being close contacts, and recruitment, the majority of which
occurred when New Zealand was under the ‘Red’ traffic light setting. There were likely
fewer people at places where recruitment posters were advertised, such as gyms, and
on University of Otago campus, and perhaps less incentive to engage in face-to-face
research (especially of the nature of the present study) relating to any number of the
widespread effects of COVID-19. The original study protocol was also amended to
minimise the time participants spent at the clinic.
vi
Six participants tested positive for COVID-19 during the study period, and particular
caution had to be taken given the risks of returning to exercise too quickly following
infection. Because data collection was restricted to very specific time points throughout
the menstrual cycle, each time data collection was disrupted, it had to be postponed
until at least the next menstrual cycle. As many participants were students, this often
ran into exam periods or university holidays, causing further postponement. One
participant had ongoing complications from COVID-19, which delayed her data
collection by five months. Several other factors also disrupted data collection: two
further participants contracted other illnesses during the study period that delayed data
collection until at least the following cycle, whilst data collection for one participant had
to be postponed for the month of Ramadan. The last clinic session for the final
participant was unable to take place due to snow, preventing the inclusion of her data in
this thesis.
The candidate was responsible for:
• Developing the study protocol, in conjunction with the candidate’s supervisor Dr
Katherine Black
• Obtaining ethical approval, in conjunction with the candidate’s supervisor Dr
Katherine Black
• Designing the participant information sheet, in conjunction with the candidate’s
supervisor Dr Katherine Black
• Developing and testing the screening questionnaire (Appendix D), well-being
questionnaire (Appendix I), COVID-19 screening questionnaire (Appendix H),
and exit questionnaire (Appendix K) in REDCap, in conjunction with the
candidate’s supervisor Dr Katherine Black
• Designing the food list for the ab-libitum post-exercise clinic meal (Appendix J)
• Designing the instruction sheets for MealLogger® (Appendix) and the diet
record, training diary, and accelerometer wear time diary (Appendix G)
• Recruiting study participants via social media, local women’s cycling groups, and
advertising posters at gyms, the swimming pool, and around University of Otago
campus
vii
• All communication with participants, including distributing information sheets
to interested participants, obtaining written informed consent, dropping off and
picking up accelerometers and other study equipment, scheduling the clinic
sessions, sending txt/email reminders throughout the study (for example, to
begin the ovulation tests, wear the accelerometer and begin diet records and
training diaries), sending out the online COVID-19 screening questionnaire
before the clinic sessions, and the exit questionnaire at completion of the study,
and monitoring diet records and training diaries in MealLogger®
• Purchasing food for the clinic session and grocery vouchers for participants upon
completion of the study
• Setting up and running all the clinic sessions, in conjunction with the candidate’s
supervisor Dr Katherine Black and APPS499 postgraduate Human Nutrition
students (for health and safety purposes at least two researchers were required
to be present throughout each session, however, the candidate was responsible
for conducting all the sessions)
• Entering 57 of 92 daily diet records into FoodWorks
• Analysing training diaries, converting exercise sessions to MET values, and
calculating energy availability for all participants
viii
Acknowledgements
Firstly, a massive thank you to all our amazing participants, who were always so
enthusiastic and cooperative, throughout what was a pretty time-consuming protocol. I
know you all lead very busy lives, so thank you for fitting in all the early morning
sessions (without breakfast!), the many days of diet records, and putting up with the
tedious process of collecting those 1mL fingerprick blood samples. We expected
recruitment to be challenging but were blown away by the number of people who
expressed interest in participating.
I cannot express how grateful I am to my incredible supervisor, Katherine, who amongst
everything else going on in her busy life, not least of all a pregnancy and a new-born,
very willingly took me on as a Master’s student, and created what was really a dream
project. Thank you for spending your precious weekends reading over my work,
tolerating my many questions, and always being so prompt with your feedback. For all
the cold, early, mornings spent bundled up in the clinic (right up until going on
maternity leave), the opportunity to write and publish a review alongside my thesis and
all the work you put into that, and your unwavering support and encouragement, you
deserve more thanks that I can give in this acknowledgements section.
Thank you to everyone in the Human Nutrition department who made this project
possible: Jen Gale, who took care of all the accelerometers and made sure they were
always ready to go, often at short notice; Kirsten Webster, for answering all my
FoodWorks questions; Chaya, Em, Tianna, and Rylie, for your help with the diet records
and clinic sessions; Sara and Sharon, for always checking in on me, and being willing to
step in whenever needed; and to everyone else who contributed in one way or another.
To the rest of our small but mighty MSc cohort: Dorothy, Bailey, and Steph, you have
been great company and sounding boards for many a statistical quandary these past
two years, and I wish you all the best in your future endeavours. Dorothy, your joy and
passion for knowledge is contagious, and you have challenged me to be better
academically like no one else has before: I will forever admire the way your mind
ix
works. Your friendship, persistent encouragement, and hilariously marvellous insights
truly enriched my whole postgraduate experience in a way I never could have imagined.
And of course, thank you to all my other wonderful friends, flatmates, and family for all
your support and love over the years, and for always so patiently providing an ear for
me to off-load stress and complain about the joys of academia. You’ve made life outside
of university an absolute delight and a true refuge, and I’m very lucky to be surrounded
by such exceptional people. To my Grandma, who passed away a few weeks before
submitting this thesis, I wish you could have been around to see this through till the
end. I know you would have loved telling your friends about it, and I hope I’ve done you
proud.
x
Table of Contents
Abstract ii
Preface v
Acknowledgements viii
Table of Contents x
Tables xii
Figures xiii
Abbreviations xv
1. Introduction 1
2. Literature Review 3
2.1 Defining the Menstrual Cycle ......................................................................................................... 4
2.1.1 Fluctuations in Sex Hormones Across the Menstrual Cycle ....................................... 4
2.1.2 Menstrual Status and Menstrual Irregularities ............................................................... 6
2.1.3 Measuring Menstrual Cycle Phase ........................................................................................ 7
2.1.4 Defining Menstrual Cycle Phase ............................................................................................ 8
2.1.5 Hormonal Contraception........................................................................................................ 11
2.2 The Role of Nutrition in Exercise Performance .................................................................... 12
2.2.1 Carbohydrate .............................................................................................................................. 13
2.2.2 Protein ........................................................................................................................................... 23
2.2.3 Fat .................................................................................................................................................... 26
2.2.4 Energy Availability ................................................................................................................... 27
2.3 Measuring Energy Availability .................................................................................................... 30
2.3.1 Dietary Assessment .................................................................................................................. 30
2.3.2 Measuring Energy Expenditure of Exercise .................................................................... 35
2.3.3 Measuring Fat-Free Mass ....................................................................................................... 37
2.4 Dietary Intake Across the Menstrual Cycle ............................................................................. 40
2.4.1 Energy Intake .............................................................................................................................. 40
2.4.2 Macronutrient Intake .............................................................................................................. 50
xi
2.5 Conclusion ........................................................................................................................................... 57
3. Methods 58
3.1 Ethical Approval and Informed Consent ................................................................................. 58
3.2 Participants ......................................................................................................................................... 58
3.3 Study Design ....................................................................................................................................... 59
3.3.1 Cycle One ...................................................................................................................................... 62
3.3.2 Cycle Two ..................................................................................................................................... 62
3.3.3 Cycles Three and Four ............................................................................................................. 63
3.3.4 Dietary Assessment .................................................................................................................. 69
3.3.6 Body Composition ..................................................................................................................... 70
3.3.7 Energy Availability Calculation ............................................................................................ 70
3.3.8 Study Size ..................................................................................................................................... 71
3.3.9 Remuneration ............................................................................................................................. 72
3.3.10 Statistical Methods ................................................................................................................. 72
4. Results 74
4.1 Participants ......................................................................................................................................... 74
4.2 Daily Dietary Intake, Exercise Energy Expenditure, Energy Availability, and Body
Composition ............................................................................................................................................... 76
4.3 Post-Exercise Ad-Libitum Meal Intake ...................................................................................... 84
5. Discussion 88
5.1 Daily Energy Intake, Exercise Energy Expenditure, and Energy Availability ........... 88
5.2 Daily Macronutrient Intakes......................................................................................................... 93
5.3 Post-Exercise Ad-Libitum Meal .................................................................................................... 95
5.4 Participant Retention ...................................................................................................................... 96
5.5 Strengths and Limitations ............................................................................................................. 96
6. Conclusions and Future Research 102
7. References 104
8. Appendices 128
xii
Tables
Table 1 Menstrual cycle phase definitions based on hormonal profiles proposed by
Elliott-Sale et al. 2021 ................................................................................................................................ 10
Table 2 Daily carbohydrate targets for athletes to provide high carbohydrate
availability ...................................................................................................................................................... 14
Table 3 Acute carbohydrate fuelling strategies to promote high carbohydrate
availability for optimal performance during competition or key training sessions .......... 15
Table 4 Carbohydrate and protein intakes of adult female athletes from studies
published between 2017-2022............................................................................................................... 20
Table 5 Participant characteristics ...................................................................................................... 76
Table 6 Daily energy and macronutrient intake, exercise energy expenditure and
energy availability from three-day weighed image-assisted diet records and training
diaries in the early-follicular and the mid-luteal phases across two menstrual cycles .... 78
Table 7 Body composition measurements obtained via bioelectrical impedance analysis
during the early-follicular and mid-luteal phases across two menstrual cycles ................. 82
Table 8 Energy and macronutrient intake from an ad-libitum post-exercise meal in the
early-follicular and the mid-luteal phases across two menstrual cycles ............................... 84
Table 9 Studies assessing dietary energy intake of naturally menstruating females
across different phases of the menstrual cycle with phases verified by serum oestrogen
and progesterone concentrations....................................................................................................... 128
Table 10 Studies assessing dietary energy intake of naturally menstruating females
across different phases of the menstrual cycle without verification of phases with serum
estrogen and progesterone concentrations .................................................................................... 133
xiii
Figures
Figure 1 Schematic diagram of the relative rise and fall in oestrogen, progesterone and
luteinising hormone concentrations across an idealised 28-day menstrual cycle with
ovulation occurring on day 14. Hormonal fluctuations are superimposed over the
menstrual cycle phases representing four distinct hormonal environments: the early-
follicular phase, the late-follicular phase, the ovulatory phase, and the mid-luteal phase ..
………………………………………………………………………………………………………………………………….6
Figure 2 Schematic diagram of the hypothesised changes in dietary energy intake
across an idealised 28-day menstrual cycle with ovulation occurring on day 14 (a), and
the corresponding relative oestrogen, progesterone and luteinising hormone
fluctuations (b). Both (a) and (b) are superimposed over the menstrual cycle phases
representing four distinct hormonal environments: the early-follicular phase, the late-
follicular phase, the ovulatory phase, and the mid-luteal phase ............................................... 50
Figure 3 Schematic diagram of the general study protocol ........................................................ 61
Figure 4 Schematic diagram of the exercise session protocol................................................... 68
Figure 5 STROBE flow diagram of participant recruitment and follow-up.......................... 75
Figure 6 Scatterplots showing energy intake, exercise energy expenditure, and energy
availability of participants in the early-follicular and mid-luteal phases of a) cycle three,
b) cycle four, and c) the mean of cycles three and four ................................................................. 79
Figure 7 Scatterplots showing macronutrient intake of participants in the early-
follicular and mid-luteal phases of a) cycle three, b) cycle four, and c) the mean of cycles
three and four ................................................................................................................................................ 81
Figure 8 Scatterplots showing participant weight in the early-follicular and mid-luteal
phases of a) cycle three, b) cycle four, and c) the mean of cycles three and four ............... 82
Figure 9 Scatterplots showing participant body composition in the early-follicular and
mid-luteal phases of a) cycle three, b) cycle four, and c) the mean of cycles three and
four .................................................................................................................................................................... 83
Figure 10 Scatterplots showing energy intake of participants from a post-exercise ad-
libitum meal in the early-follicular and mid-luteal phases of a) cycle three, b) cycle four,
and c) the mean of cycles three and four ............................................................................................ 85
xiv
Figure 11 Scatterplots showing macronutrient intake of participants from a post-
exercise ad-libitum meal in the early-follicular and mid-luteal phases of a) cycle three, b)
cycle four, and c) the mean of cycles three and four ...................................................................... 87
Figure 12 Timeline of participant movement throughout the study .................................. 162
xv
Abbreviations
ASA24
BBT
BMI
CI
EAR
EDTA
EEE
EI
ES
FFM
FM
FSH
GnRH
LH
MET
MPS
OCP
PMDD
PMS
RDI
REDCap
RED-S
RER
RPE
SD
STROBE
TEE
TEI
VCO2
VO2
Automated self-administered 24-hour
Basal body temperature
Body mass index
Confidence interval
Estimated average requirement
Ethylenediaminetetraacetic acid
Exercise energy expenditure
Energy intake
Effect size
Fat-free mass
Fat mass
Follicle-stimulating hormone
Gonadotrophin-releasing hormone
Luteinising hormone
Metabolic equivalents of task
Muscle protein synthesis
Oral contraceptive pill
Premenstrual dysphoric disorder
Premenstrual syndrome
Recommended daily intake
Research electronic data capture
Relative energy deficiency in sport
Respiratory exchange ratio
Rating of perceived exertion
Standard deviation
Strengthening the reporting of observational studies in
epidemiology
Total energy expenditure
Total energy intake
Volume of carbon dioxide
Volume of oxygen
1
1. Introduction
Females have long been under-represented as research participants in a range of fields,
hindering progress in better understanding female physiology (1-5). For example,
across more than 5200 papers published in six of the leading sports science journals
between 2014 and 2020, only one-third of all participants were female, with only six
percent of studies conducted exclusively in females (1). Yet, important physiological
differences exist between the sexes, such that research conducted in males may not
always be directly applicable to females (6). One of these differences being the
menstrual cycle, involving dramatic fluctuations in endogenous sex hormones over a
21–35-day period, from menarche until menopause (7, 8).
Awareness of this sex-data gap and the need to better understand female-specific
nuances of exercise physiology and sports nutrition is growing, especially with the
increasing female participation in both recreational and elite-level sport (9, 10). There
is currently a swell of social media commentary targeted at athletes to train and eat in
alignment with their menstrual cycle (11-14). However, high-quality published
literature on the effects of endogenous sex hormones on athletic performance, training
responses, and dietary considerations in exercising populations is sparse (15-18). As
such, insufficient evidence currently exists to warrant female-specific guidelines for
sports nutrition or training practices as the relationships between the menstrual cycle,
exercise, and dietary intake are not yet fully understood (8, 15, 19).
In the general population, energy intake appears to be greater in the luteal phase of the
menstrual cycle compared with the follicular phase in regularly menstruating females
(18). However, the magnitude of difference varies widely between studies (18), and the
consensus on macronutrient intakes across the menstrual cycle is even less clear (20-
26). This may in part be due to the large amount of methodological heterogeneity and
poor-quality studies present in this research field (18). For example, studies that do not
confirm ovulation or measure serum sex hormones risk collecting data outside of the
targeted phases or including participants with abnormal hormonal profiles (8).
Recently, Elliott-Sale et al. proposed a working guide for standards of practice for
2
research on females, including strict guidelines for defining and quantifying each
menstrual cycle phase (8). Highlighted is the importance of ensuring consecutive
regular periods before any data collection, confirming ovulation and a normal hormonal
profile via urinary ovulation tests and serum oestrogen and progesterone
concentrations, and repeating outcome measures in a second cycle (8). To date, no
published research regarding dietary intake across the menstrual cycle has followed all
these guidelines (18).
Furthermore, the majority of research on dietary intake across the menstrual cycle has
been conducted amongst inactive populations (18). But for free-living physically active
females, dietary patterns to meet the needs of exercise may not always correspond to
hunger or desirability of food, and physical activity can have variable influences on
appetite (27, 28). Moreover, the prevalence of those at risk of low energy availability
(LEA) and menstrual disturbances is high in exercising populations (29-31). Energy
availability has emerged as a critical concept in sports nutrition and is a more
appropriate index for quantifying nutritional status than energy intake alone in active
populations (29). Low Energy Availability (LEA) has numerous negative health and
performance consequences, but understanding of its complexities and optimal methods
of quantification are still developing (29, 32, 33). Fluctuations in energy availability
across the menstrual cycle could have implications for how it is measured in research
and the field, and identifying phases where additional nutritional support may be
required is important.
The immediate post-exercise period presents an important nutritional opportunity to
maximise training adaptations and recovery (34), yet little is known about the influence
of menstrual cycle phase on habitual dietary intake following exercise, and how this
may relate to appetite and gastrointestinal function (19). Therefore, this thesis will
investigate energy availability, daily energy intake, and daily macronutrient intake, as
well as post-exercise energy and macronutrient intakes, across the menstrual cycle in
female exercisers, to aid in the design of future studies and the interpretation of their
findings.
3
2. Literature Review
This literature review is presented in four main parts:
• Section 2.1 describes the physiology of the menstrual cycle, defines menstrual
status, and discusses how to define and measure the menstrual cycle phases.
• Section 2.2 discusses the role of nutrition in exercise performance, current
sports nutrition guidelines for macronutrient intakes and energy availability, and
female-specific considerations.
• Section 2.3 discusses the methods for measuring dietary intake and energy
availability in free-living populations.
• Section 2.4 discusses the observational research on energy intake, energy
availability, and macronutrient intake across the menstrual cycle.
Literature searches were conducted using Scopus and PubMed between April 2021 and
November 2022 using a combination of the following keywords: menstrual cycle,
energy intake, energy availability, diet, appetite, female, athlete, exercise, carbohydrate,
fat, protein, amino acid, oxidation, catabolism, muscle protein synthesis, requirements,
image-assisted, image-based, dietary assessment, hormonal contraception, New
Zealand, resting metabolic rate, prevalence, body composition, bioelectrical impedance,
BIA, dual x-ray absorptiometry, DEXA, DXA, and skinfold. Searches were limited to
articles published in the English language, and reference lists were additionally
screened to locate further relevant literature. The World Wide Web was used to locate
online articles about menstrual cycle-related exercise and nutrition strategies, to
describe the recent media commentary on the topic. Physical and online textbooks were
used to describe basic physiology, hormonal contraception methods, and the
fundamentals of nutrition assessment. Position stands and reviews by respectable
sports nutrition bodies such as the American College of Sports Nutrition, Dieticians of
Canada, the Academy of Nutrition and Dietetics, the International Society of Sports
Nutrition, and the International Olympic Committee (IOC), were used to describe the
current sports nutrition guidelines.
4
Throughout this thesis, the term ‘female’ refers to the biological concept of sex, as
opposed to the socially constructed and individualised experience of gender (35, 36).
That is, those possessing the reproductive organs and physiological functions derived
from the XX chromosomes (35).
2.1 Defining the Menstrual Cycle
2.1.1 Fluctuations in Sex Hormones Across the Menstrual Cycle
The ‘menstrual cycle’ jointly describes the ovarian cycle, which prepares and releases
oocytes, and comprises the follicular and luteal phases; and the concurrent uterine
cycle, which prepares and maintains the uterine lining, and comprises the menstrual,
proliferative, and secretory phases (37). A regular menstrual cycle lasts 21-35 days,
from menarche until menopause (8): a period typically spanning 35-40 years (38).
The hypothalamus in the brain ultimately controls the menstrual cycle by secreting
gonadotrophin-releasing hormone (GnRH), which stimulates the anterior pituitary
gland to release follicle-stimulating hormone (FSH) and luteinising hormone (LH) (37).
The first day of menstrual bleeding is considered day one of the menstrual cycle and
denotes the beginning of the follicular phase (37). As the uterine lining is shed, FSH
rises gradually, stimulating several ovarian follicles to mature, which compete for
dominance (37). As one follicle matures, the granulosa cells surrounding the ovum
multiply and secrete oestrogen (37). Oestrogen rises throughout the follicular phase,
causing the endometrial lining to become thicker and more vascular (37). At low
concentrations, oestrogen exhibits a negative feedback relationship with LH, but at
higher concentrations, stimulates the release of LH: only when oestrogen has remained
above 200 pg·mL-1 for approximately 50 hours will the transition from LH suppression
to stimulation occur (37). Oestrogen concentration must remain elevated until the LH
positive feedback loop has been established, for the LH surge to prove successful (37).
The rapid increase in both LH and FSH suspends granulosa cell proliferation, allowing
the oocyte to mature (37).
5
The LH surge begins approximately 34-36 hours before ovulation, reaching peak levels
approximately 8-24 hours before the follicle ruptures and releases the oocyte into the
fallopian tube (although LH surge patterns vary between individuals (39)) (40-42).
Ovulation is often quoted to occur on day 14 of the menstrual cycle, however, the
duration of the follicular phase is highly variable, both between and within individuals
(43-45). As the luteal phase of the ovarian cycle and the secretory phase of the uterine
cycle begin, the ruptured follicle develops into the corpus luteum, which produces both
progesterone and oestrogen (37). If the oocyte is not fertilised within approximately
12-24 hours, it disintegrates in the fallopian tube (37). As progesterone and oestrogen
rise across the luteal phase, they suppress FSH and LH production, eventually causing
the corpus luteum to atrophy (37). The fall in progesterone and oestrogen levels
triggers the onset of menstruation, and the cycle repeats (37).
In research, the menstrual cycle is commonly divided into just the follicular and luteal
phases, however, within one cycle there are four distinct hormonal environments: 1)
the early-follicular phase, characterized by low oestrogen and low progesterone; 2) the
late-follicular phase, characterized by high oestrogen and low progesterone; 3) the
ovulatory phase, characterized by medium oestrogen and low progesterone; and 4) the
mid-luteal phase, characterized by medium oestrogen and high progesterone (8), as
Figure 1 depicts.
6
Figure 1 Schematic diagram of the relative rise and fall in oestrogen, progesterone
and luteinising hormone concentrations across an idealised 28-day menstrual cycle
with ovulation occurring on day 14. Hormonal fluctuations are superimposed over
the menstrual cycle phases representing four distinct hormonal environments: the
early-follicular phase, the late-follicular phase, the ovulatory phase, and the mid-
luteal phase1,2
1The solid orange line represents oestrogen, the dashed blue line represents
progesterone, and the dotted green line represents luteinising hormone
2Figure was created with Adobe Photoshop 2020 (Adobe Inc., San Jose, United States)
2.1.2 Menstrual Status and Menstrual Irregularities
Definitions of menstrual status and cycle phases are inconsistent throughout the
literature (8). With growing interest in the field of menstrual cycle research, there has
been a call to standardise terminology and methodology, to improve research quality
(8). The following definitions of menstrual status are based on the recent standards of
practice for research in females proposed by Elliott-Sale et al. (8).
‘Naturally menstruating’ describes those with a menstrual cycle length of 21-35 days,
with nine or more consecutive periods per year (8). To be classed as ‘eumenorrheic’,
participants must meet the criteria for ‘naturally menstruating’, but also exhibit a
7
urinary LH surge and the correct serum hormonal profile (see section 2.2.4.1), as well as
having refrained from using hormonal contraception in the three months prior to
recruitment (8). ‘Menstrual irregularities’ refers to any disruptions in this
eumenorrheic state, including oligomenorrhoea, amenorrhoea, anovulation, and luteal
phase deficiency (8). ‘Oligomenorrhoea’ is a cycle length greater than 35 days, while
‘amenorrhoea’ is the absence of menstruation, and is further classified into primary
amenorrhoea: failing to reach menarche by age 15 when the development of secondary
sexual characteristics is evident, or by age 14 when no secondary sexual characteristics
are evident; and secondary amenorrhoea: the absence of at least three consecutive
periods in non-pregnant females with past menses (8). ‘Luteal phase deficiency’ refers
to cycles with a luteal phase progesterone concentration below 16.00 nmol·L-1, while
‘anovulation’ is the presence of menstruation without ovulation (defined by the absence
of a urinary LH surge, or confirmed by serum hormone concentrations: see section
2.2.4.1) (8).
2.1.3 Measuring Menstrual Cycle Phase
The methods for identifying menstrual cycle phase vary in their sensitivity and
specificity (46). Calendar-based counting is the most basic method and involves
indirectly estimating cycle phases based on an arbitrary number of days relative to
menstruation onset (46). However, given the large inter-individual variation in
follicular phase duration (44), accurately determining the date of ovulation from
menstruation alone is incredibly difficult. Furthermore, without verifying serum
hormones, one cannot identify luteal phase deficiency, or distinguish between ovulatory
and anovulatory cycles (46). Basal-body temperature (BBT) is another indirect method
used to identify cycle phase and is based on the assumption that body temperature rises
by approximately 0.3℃ during the luteal phase (46). However, BBT correlates poorly
with serum progesterone concentration (47, 48), can be influenced by many other
factors (49), and ultimately, doesn’t measure actual hormone levels (46). Furthermore,
luteal increases in BBT may not occur in all females with ovulatory cycles (47, 50).
Therefore, used on their own, these two methods are considered unreliable for
identifying menstrual cycle phases (46).
8
Urinary ovulation tests detect the urinary LH surge, which proceeds ovulation by
approximately 24 hours (42, 51). These tests can identify ovulation and luteal phase
timing, but can’t eliminate the possibility of a luteal phase deficient cycle - which may
follow as many as 30% of positive tests (30) - so are not recommended for use in
isolation when trying to assess the menstrual cycle (46). Salivary hormone analysis may
be used to verify cycle phase, although given the low concentrations of oestrogen and
progesterone present in saliva, requires highly sensitive tests (52). Furthermore,
salivary oestrogen occurs in pulsatile cycles of 60-90 minutes (53), whilst salivary
progesterone varies considerably over 24 hours (54); reliable information requires
multiple daily saliva samples (52). While Gandara et al. could reportedly discern
between regular and abnormal cycles using salivary oestrogen and progesterone
profiles, they did not verify these against serum hormone concentrations (55). Other
infrequently-used methods exist, such as salivary ferning, and serial follicular scanning,
however, serum hormone analysis definitively emerges as the gold standard method for
verifying menstrual cycle phase as it directly measures circulating levels of oestrogen
and progesterone, to most accurately reflect hormonal status (8, 46).
2.1.4 Defining Menstrual Cycle Phase
While the methods of measuring menstrual cycle phase vary throughout the literature,
definitions of phases are inconsistent as well (see Appendices A and B). Earlier research
often compared the 10 days post-menstruation onset, with the 10 days pre-
menstruation onset; although nomenclature soon shifted to the follicular and luteal
phases, the definitions of these vary widely. The days of menstrual bleeding often
represent the follicular phase (or “menses” as it may also be called), but occasionally
studies may use a later timepoint, such as 7-9 days after the onset of menstruation (see
Appendices A and B) (what one might call the “mid-follicular” phase, although there is
no standardised definition of this (8)). Researchers often define the luteal phase by a
variable number of days relative to either menstruation or ovulation (see Appendices A
and B). Only recently have the late-follicular and ovulatory phases been recognised and
defined as two distinct phases (8); historically researchers may have used the “peri-
ovulatory” phase to refer to the several days leading up to and including ovulation, but
this period is often neglected in research (see Appendices A and B). To improve
9
uniformity, new guidelines propose stricter menstrual cycle phase definitions that
encompass the four distinct hormonal environments: 1) the early-follicular phase, 2)
the late-follicular phase, 3) the ovulatory phase, and 4) the mid-luteal phase (8) (Figure
1).
2.1.4.1 Gold Standard Protocol for Measuring and Defining Menstrual Cycle Phase
Described below are the gold standard protocols for measuring and defining the
menstrual cycle phases in research on females, proposed by Elliott-Sale et al. (8).
Firstly, researchers should track menstrual cycles for at least two months before any
testing, and participants should not have taken any form of hormonal contraception
three months before recruitment. In each phase, blood samples are to be analysed for
serum oestrogen and progesterone to both confirm the target phase has been captured,
and to identify menstrual irregularities.
Table 1 shows the definitions of menstrual cycle phases based on hormonal profiles
proposed by Elliot-Sale et al. (8). The onset of menstrual bleeding denotes day one of
the cycle, with the early-follicular phase defined as days 1-5, when oestrogen and
progesterone concentrations are at their lowest. From day eight, participants begin
taking a daily urinary ovulation test in the morning until they receive a positive result,
indicating a surge in LH. Identifying the late-follicular phase is perhaps the most
difficult: testing should be done as close to the LH surge as possible, with oestrogen
higher than the other phases, and progesterone higher than the early-follicular phase,
but lower than 6.36 nmol·L-1. Elliott-Sale et al. define the late-follicular phase as 14-26
hours prior to ovulation (8), however, physiologically it may make more sense to define
it as 14-26 hours before the onset of the LH surge, (as oestrogen reaches peak levels
before inducing the rise in LH) (37), especially as Elliott-Sale et al. also define the
ovulatory phase as the 24-36 hours following a positive urinary LH test (potentially
resulting in overlap for some individuals (39-41)). Using urinary LH tests to identify
cycle phases also depends on the sensitivity of the test, and which part of the surge
(from the initial rise to peak LH concentration) triggers a positive result. This likely
varies between individuals as LH surges resulting in ovulation display considerable
10
variability in profile, amplitude, and length (39); therefore, serum oestrogen and
progesterone concentrations should ultimately confirm cycle phase.
Table 1 Menstrual cycle phase definitions based on hormonal profiles proposed by
Elliott-Sale et al. 2021 (8)
Menstrual Cycle Phase
Days
Oestrogen
Progesterone
Early-follicular
1-5
Lower than the late-
follicular, ovulatory,
and mid-luteal
phases
Lower than the late-
follicular, ovulatory and
mid-luteal phases
Late-follicular
14-26 hours
prior to LH
surge
Higher than the
early-follicular,
ovulatory and mid-
luteal phases
Higher than the early-
follicular phase, but
lower than 6.36 nmol·L-1
Ovulatory
24-36 hours
following a
positive LH
test
Higher than the
early-follicular
phase, but lower
than the late-
follicular and mid-
luteal phases
Higher than the early-
follicular phase, but
lower than 6.40 nmol·L-1
Mid-luteal
7 days
following a
positive LH
test
Higher than the
early-follicular and
ovulatory phases,
but lower than the
late-follicular phase
Higher than
16.00 nmol·L-1
For the ovulatory phase, oestrogen should be higher than the early-follicular phase, but
lower than the other phases, and progesterone higher than the early-follicular phase,
but lower than 6.40 nmol·L-1. If no positive result occurs after 21 days of urinary LH
tests, testing is delayed until the following cycle, and any data collected during the
preceding phases, excluded. While Elliot-Sale et al. encourage measuring the late-
follicular and ovulatory phases in addition to the follicular and luteal phases, the
challenges of doing so (such as the short notice, additional costs, and participant
burden) must be recognised (8, 46). Testing in the mid-luteal phase should begin seven
days after the positive urinary LH test, with oestrogen higher than the early-follicular
11
and ovulatory phases, but lower than the late-follicular phase, and progesterone greater
than 16.00 nmol·L-1. Outcome measures should then be repeated in a second cycle to
reduce variability.
Cycles failing to meet these hormonal profiles should be excluded from analysis a-
posteriori, as should data from time points failing to capture the intended cycle phase
(with reasons for exclusion reported). It is not uncommon for those with apparently
regular menstrual cycles to occasionally experience anovulation or luteal phase
deficiency (56, 57), so confirming luteal progesterone has reached 16.00 nmol. L-1 is
important for quantifying the true effect of the menstrual cycle on outcome measures
(46). This becomes particularly important when studying active populations: the
prevalence of luteal phase deficiency in recreationally active females (>2.5 hours
exercise·week-1) may be as high as 30% (30), while over half of heavy exercisers
(>7.5 hours exercise·week-1), may have abnormal cycles (27% luteal phase deficient,
and 25% anovulatory) (31). Without measuring serum oestrogen and progesterone,
anovulatory or luteal phase deficient cycles are likely to be included, potentially
diminishing or masking any effects of the menstrual cycle on outcome measures.
Inconsistencies in the results from menstrual cycle research are apparent, but much of
these may be explained by methodological issues and poor adherence to recent
guidelines (8). While the methods described above are considered the gold standard,
that is not to completely disregard studies using other methods or discourage research
in this area (the participant, time, resource, and financial challenges of adhering to such
guidelines are certainly appreciable) but rather to exercise more caution when
interpreting results (46). For example, using urinary ovulation tests alone to discern
phases is still valuable, and provides more confidence in estimates than approximating
phases based solely on menstruation onset (46).
2.1.5 Hormonal Contraception
‘Hormonal contraception users’ refers to those taking any type of contraception capable
of altering the endogenous hormonal milieu (8). Hormonal contraception provides
either: a) progestins only (the synthetic form of progesterone), or b) both oestrogen and
12
progestins, to downregulate endogenous sex hormone production via a negative
feedback loop (37). The oral contraceptive pill (OCP) is the most common form of
hormonal contraception in New Zealand (58, 59), but other delivery methods include
implants, injections, patches, vaginal rings, and inter-uterine devices (37). Each type
and specific brand will differ in the amount of oestrogen and/or progestin it delivers, as
well as in progestin potency and androgenicity (37).
The exogenous hormones that contraception delivers are not equivalent to the
endogenous fluctuations and associated physiology of a natural menstrual cycle (8, 37).
As it may take several months for a regular cycle to return following the cessation of
hormonal contraception, research on eumenorrheic females requires participants to
have refrained from using hormonal contraception at least three months before
recruitment (8).
However, hormonal contraception users (typically OCP users) may be used as a control
group as they produce a consistent and predictable hormonal profile, so can potentially
eliminate the effects of changing endogenous sex hormones and provide consistent
conditions for testing (60). But to do so appropriately, only one brand of OCP should be
used: grouping different brands of OCPs together results in large variations in hormone
concentrations and may contribute to the conflicting results observed by studies with
OCP users as either a control or experimental group (8, 60). Furthermore, all
participants recruited as OCP users should have used the specific hormonal
contraception for at least three months (8).
2.2 The Role of Nutrition in Exercise Performance
The type, quantity, and timing of nutrients during and around training and competition
play a critical role in athletic performance, exercise recovery, preventing injury and
illness, and overall health (61). Current sports nutrition guidelines are based on
research conducted in predominantly male populations, which may not always directly
translate to female exercisers (19). For example, sexual dimorphisms in substrate
metabolism (62), invite the question of whether sex-specific guidelines for fuelling
before, during, and after training may be beneficial (19). Further, fluctuations in
13
endogenous sex hormones across the menstrual cycle exert widespread effects on
physiological and metabolic systems, such that nutritional requirements and habitual
intakes could potentially differ between cycle phases (19).
2.2.1 Carbohydrate
2.2.1.1. Carbohydrate Intake Guidelines for Female Athletes
Manipulating carbohydrate availability to match the fuel demands and goals of exercise
is critical to optimise training quality and achieve peak performance (63). Carbohydrate
is a substrate for both anaerobic and oxidative pathways, providing an important fuel
source for intermittent high-intensity exercise and sustained submaximal-intensity
exercise, as well as playing a discretionary role in brief high-intensity work (64). It is
the obligate fuel of the central nervous system (65), plays a role in bone health (66, 67)
and immunity (68), and food sources of carbohydrate provide essential and health-
promoting nutrients (69). The limited body stores of carbohydrate (liver and muscle
glycogen) are often insufficient to meet the daily fuel requirements of exercising
individuals: depleting these stores results in reduced work rates, impaired skill and
concentration, and an increased rating of perceived exertion (RPE) (63). Therefore,
exogenous carbohydrate before, during, and after exercise is required to promote high-
quality training (63). Daily carbohydrate recommendations are given according to
exercise volume and intensity and relative to body mass (61). Table 2 outlines the daily
targets for athletes to meet the carbohydrate needs of the muscle and central nervous
system to promote high carbohydrate availability.
14
Table 2 Daily carbohydrate targets for athletes to provide high carbohydrate
availability (61)
Training volume
Carbohydrate target range
(g·kg-1·day-1)
Light
Low-intensity or skill-based activities
3.0-5.0
Moderate
~1 hr·day-1 moderate exercise
5.0-7.0
High
1-3 hr·day-1 moderate-high intensity
exercise
6.0-10.0
Very high
>4-5 hr·day-1 moderate-high intensity
exercise
8.0-12.0
Table 3 summarises the acute carbohydrate fuelling strategies for before, during, and
after exercise, to promote high carbohydrate availability for optimal performance
during competition or key training sessions. Carbohydrate intake before exercise helps
maintain blood glucose concentration and optimise glycogen stores (particularly liver
glycogen, which becomes depleted after an overnight fast) (70). Consuming
1.0-4.0 g·kg-1 of carbohydrate in the 1-4 hours before exercise is recommended, while a
carbohydrate loading protocol of 10.0-12.0 g·kg-1·day-1 during the 24-36 hours before
competition or a key training session can supercompensate muscle glycogen stores
(61). Despite early research suggesting a reduced capacity for glycogen storage in
endurance-trained females compared with males (71), later research showed that
differences were related to energy intake, not the physiological capacity to “load”
muscle glycogen (72). While practically, females may have more trouble achieving
carbohydrate loading guidelines, provided they are met, the capacity to carbohydrate
load does not appear to differ by sex (72, 73) (although female participants in the study
by James et al. were oral contraceptive users (73)).
15
Table 3 Acute carbohydrate fuelling strategies to promote high carbohydrate
availability for optimal performance during competition or key training sessions
(61)
Situation
Carbohydrate target range
Before exercise:
24 hrs before event or key session <90min
7.0-12.0 g·kg-1·day-1
36-48 hrs before event or key session
>90min
10.0-12.0 g·kg-1·day-1
1-4 hrs before event or key session
>60min
1.0-4.0 g·kg-1
During exercise:
<45 mins
Not required
45-75 mins sustained high intensity
30.0 g·hr-1 or mouth rinse
1-3 hrs moderate intensity or intermittent
high-intensity
30.0-60.0 g·hr-1
>2.5 hrs
60.0-90.0 g·hr-1 (dual source)
After exercise:
Fast refuelling (<8 hrs between two fuel-
demanding sessions)
1.0-1.2 g·kg-1·hr-1 for first
4 hrs
However, glycogen storage may be altered during different phases of the menstrual
cycle (74-76). In moderately-trained females who underwent glycogen depletion
exercise followed by three days of a 4 g·kg-1·day-1 carbohydrate diet (considered
suboptimal for maximum glycogen resynthesis (61)), glycogen content was greater in
the mid-luteal phase compared with the mid-follicular phase (383 vs. 313 mmol·kg-1dw)
(75). McLay et al. and Hackney et al. also observed resting glycogen storage to be
greater in the mid-luteal phase compared with the mid-follicular phase (74, 76).
However, McLay et al. found that the follicular phase impairment in glycogen storage
could be overcome by carbohydrate loading: following a glycogen depletion protocol
and three days of a normal carbohydrate (5.2 g·kg-1·day-1) diet, resting glycogen content
was markedly lower in the mid-follicular phase compared with the mid-luteal phase
(575 vs. 756 mmol·kg-1dw), but after three days of a high carbohydrate diet
(8.4 g·kg-1·day-1), there was no longer evidence of a difference between phases (728 vs.
771 mmol·kg-1dw) (76). Taken with the fact that exercise performance may be trivially
reduced in the early-follicular phase (15), these results suggest that carbohydrate
intake may be particularly important during the follicular phase, and that female
16
athletes may benefit from a slightly higher intake (or at least towards the upper end of
the recommended ranges) at this time, especially if key training sessions or competition
are limited by glycogen content (19).
Consuming carbohydrate during periods of exercise greater than 45 minutes enhances
performance by maintaining blood glucose concentration, providing an exogenous
muscle substrate, sparing liver glycogen, increasing carbohydrate oxidation, and acting
on reward centres of the central nervous system (77). Current guidelines recommend
consuming 30.0 g·hr-1 of carbohydrate or a carbohydrate mouth rinse during exercise
lasting less than 75 minutes, 30.0-60.0 g·hr-1 for exercise 1-3 hours in duration, and up
to 90.0 g·hr-1 (via multiple transportable carbohydrates, i.e., glucose and fructose
mixture) for exercise greater than 2.5 hours in duration (61). While these guidelines
were developed based on studies primarily conducted in male populations, no
conclusive evidence exists to suggest that in-exercise feeding strategies should be
different for females (19). For example, Wallis et al. found similar metabolic responses
to 90 g·hr-1 of carbohydrate during a 2-hour cycling protocol at 67% VO2max in trained
males and females (78).
In terms of menstrual phase-related discrepancies in substrate utilization during
exercise, the data are conflicting: some studies suggest a glycogen-sparing effect, and
greater fat oxidation in the luteal phase compared with the follicular phase, whilst
others find no evidence of a difference (62). Inconsistent findings are likely partly due
to differences in exercise intensity (the major determinant of substrate oxidation)
between studies, as well as potential inter-individual variability in sex hormone
concentrations (62). However, Campbell et al. demonstrated a similar ergogenic effect
of 67 g·hr-1 of carbohydrate during a 2-hour cycling protocol at 70% VO2max followed by
a 4 kJ·kg-1 time trial in the follicular and luteal phases in endurance-trained athletes,
with carbohydrate ingestion negating the menstrual cycle-related effects on blood
glucose kinetics and time trial performance observed with a carbohydrate-free placebo
(79). Bailey et al. also showed that carbohydrate supplementation (0.6g·kg-1·hr-1) whilst
cycling at 70% VO2max increased time to exhaustion in the follicular and luteal phases in
equal measure compared to a placebo in moderately-trained females (80). Based on the
existing literature, any potential differences in substrate utilization across the cycle do
17
not yet warrant phase-specific in-exercise fuelling strategies for female athletes beyond
those of current guidelines (19, 62).
Evidence is also emerging from studies in males to indicate that intakes as high as
120.0 g·hr-1 may be favourable for performance in ultra-endurance events such as ultra-
running and cycling, and can be tolerated by the gut (81, 82). While intakes this high are
less studied in females, and several methodological caveats exist to those that have
investigated this (such as lack of control for menstrual cycle phase and hormonal
contraception use), peak rates of exogenous carbohydrate oxidation may be slightly
lower in females (83-85). However, Pettersson et al. do suggest that highly-trained
female athletes can tolerate high carbohydrate intakes (2.2 g·min-1 of an 18% 1:0.8
maltodextrin and fructose solution) during sub-maximal exercise without
gastrointestinal distress (at least in cold conditions) (85). Nonetheless, further research
is required to determine maximal rates of exogenous carbohydrate oxidation and
corresponding gut tolerance in female athletes (19).
After exercise, carbohydrate is required to replenish muscle glycogen; this becomes
particularly important when key training sessions or events are close together (86).
Maximal rates of glycogen resynthesis are achieved by consuming 1.0-1.2 g·kg-1·hr-1 of
carbohydrate in the four hours following exercise (61). Similar rates of post-exercise
muscle glycogen resynthesis have been reported between males and females (87, 88),
although the effect of the menstrual cycle has not yet been investigated (19).
The controlled restriction of dietary carbohydrate before, during, or after training
within a periodized training programme (e.g., “train low” or “recover low” strategies) is
also becoming recognized as a potential approach to enhance cellular adaptations (89).
However, a consensus on the performance benefits of such strategies is still unclear, and
possible risks such as symptoms associated with RED-S, as well as the sacrifice of high-
quality training, must be weighed against any potential advantages (19, 63, 89).
Furthermore, research on manipulating carbohydrate availability has primarily been
conducted in males; less is known about the efficacy in females (19).
18
2.3.1.2 Carbohydrate Intakes of Female Athletes
Despite the wealth of evidence supporting these guidelines, it appears that female
athletes often don’t achieve carbohydrate intake targets (90). A 2001 review on the self-
reported carbohydrate intakes of high-level (collegiate, national-level, elite and
professional) athletes from studies published between 1971 and 1999 revealed that
whilst males generally met the recommended carbohydrate targets, females often fell
short (90). Mean carbohydrate intake was 5.5 g·kg-1·day-1 for female endurance athletes,
and 4.7 g·kg-1·day-1 for non-endurance athletes, which are below the respective
recommendations of at least 6-10 g·kg-1·day-1 and 5-7 g·kg-1·day-1 for the expected
training loads (61, 90). More recently, a 2021 systematic review of the dietary intakes of
female field-based team sport athletes reported a mean carbohydrate intake of
4.3 ± 1.2 g·kg-1·day-1, with the majority of studies (n=13) observing mean intakes below
recommendations (91). Of the remaining studies, four reported mean intakes at the
lower end of the recommendations (5.0-5.4 g·kg-1·day-1), while only two studies
reported mean intakes that met the daily carbohydrate recommendations for moderate-
high intensity exercise (7.0 and 8.3 g·kg-1·day-1) (91).
Table 4 summarises the carbohydrate and protein intakes of adult female athletes
across a range of sports, from studies published between 2017 and 2022. Except for
Noh et al. who reported a mean carbohydrate intake of 7.6 g·kg-1·day-1 in a population of
endurance athletes (92), the majority of studies reported mean carbohydrate intakes
for female athletes below recommendations (61, 93-106). Generally, mean
carbohydrate intakes ranged between 3.0-5.0 g·kg-1·day-1 (93-103, 105, 106), with a
select few studies reaching or exceeding the 5.0 g·kg-1·day-1 threshold (at least at certain
times in their sporting seasons) (92, 98, 101, 102, 106), and one study in professional
Australian Football League players reporting a mean intake below 3.0 g·kg-1·day-1 (104).
Either reported carbohydrate intakes inaccurately reflect habitual intakes, the
carbohydrate guidelines for females require re-evaluation, or a vast number of female
athletes could benefit from improving their carbohydrate intake (90). The limitations of
dietary assessment are discussed in depth in section 2.4.1, but under-reporting of
dietary intake likely plays a role. But even a crude assumption of underreporting by
19
20% (as the results of Capling et al. might suggest (107), and assuming uniform under-
reporting across macronutrients) would, at the very least, put intakes at the lower end
of the recommended ranges, if not still outside of the lower bounds. It is likely that
significant periods of insufficient intake, if not chronic under-consumption, still occur in
female athletes (90). Reasons for this may include: restricting energy intake to achieve
body composition goals, inadequate nutrition knowledge, background dietary practices
and food culture, poor food availability, gastrointestinal challenges of a high-fibre
intake, promotion of low carbohydrate diets, and lifestyle and travel commitments (90).
While a large quantity of the research on carbohydrate requirements for exercise is
derived from male populations, to date, no strong evidence exists to suggest that these
are drastically different for female exercisers (19). The success of even high-level female
athletes who apparently fail to meet daily carbohydrate recommendations is likely not
because of, but rather despite, suboptimal intakes, and their failure to achieve daily
targets doesn’t necessarily nullify the advantages of meeting such guidelines, which are
based on abundant evidence (90).
However, previous research on the habitual carbohydrate intakes of female athletes has
largely failed to consider any potential effects of menstrual cycle phase (or hormonal
contraceptive use). Changes in appetite, gastrointestinal symptoms, and food cravings
across the menstrual cycle may influence dietary patterns and absolute carbohydrate
intakes, in turn affecting the “real world” fuelling strategies of exercising females (19).
To date, only one study has quantified the macronutrient intakes of athletes in different
phases of the menstrual cycle: Ihalainen et al. reported mean carbohydrate intakes in
the early-follicular, mid-follicular, ovulatory, and luteal phases of 255 ± 80, 260 ± 77,
247 ± 67 and 250 ± 55 g·day-1 respectively, in 15 eumenorrheic recreational athletes
(defined as strength training three times per week and endurance training three times
per week) (108). While there was no evidence of a difference between phases, in
relative terms this corresponded to ~3.7 g·kg-1·day-1 (based on the reported mean
weight) (108), suggesting that carbohydrate guidelines were not being met in any phase
of the menstrual cycle (61). However, as this is only one study, further research is
required to determine whether the habitual macronutrient intakes of female athletes
differ by menstrual cycle phase, and practically, how this might influence fuelling
strategies.
20
Table 4 Carbohydrate and protein intakes of adult female athletes from studies published between 2017-2022
Author & Year
Population
Dietary assessment
method
Carbohydrate intake
(g·kg-1·day-1)1
Protein intake
(g·kg-1·day-1)1
Traversa et al.
2022 (93)
15 varsity rugby union
players
7-day weighed diet
record
3.4 ± 0.2
(12.0% meeting
recommendations)
1.4 ± 0.1
(55.0% meeting
recommendations)
Hoshino et al.
2022 (94)
112 national-level
university athletes
(swimming, track & field,
team sport & dance)
1x FFQ
4.0 ± 1.4
1.2 ± 0.4
Vermeulen et
al. 2021 (95)
23 varsity ice hockey
players
7-day weighed or
image-assisted2 diet
record
4.6 ± 1.0
(39.0% meeting
recommendations)
1.4 ± 0.5
(87.0% meeting
recommendations)
Danh et al.
2021 (96)
14 collegiate indoor
volleyball athletes
Pre-season: 3-day
estimated diet record
Off-season: 1x 24-
hour recall
Pre-season: 3.03
Off-season: 4.23
Pre-season: 1.33
Off-season: 1.73
Zanders et al.
2021 (97)
13 collegiate basketball
players
5x 4-day estimated
diet records
Phase 1: 3.8 ± 0.7
Phase 2: 3.7 ± 1.1
Phase 3: 3.3 ± 0.7
Phase 4: 4.0 ± 0.4
Phase 5: 3.6 ± 0.74
Phase 1: 1.3 ± 0.2
Phase 2: 1.2 ± 0.2
Phase 3: 1.2 ± 0.3
Phase 4: 1.1 ± 0.2
Phase 5: 1.2 ± 0.34
Beerman et al.
2020 (98)
20 collegiate distance
runners
1x semi-quantitative
FFQ5
4.7 ± 1.9
(15.0% meeting
recommendations)
1.4 ± 0.6
(65.0% meeting
recommendations)
21
Gibson-Smith
et al. 2020 (99)
20 experienced & elite
climbers
3-day weighed diet
record
3.8 ± 0.9
1.6 ± 0.5
Gogojewicz et
al. 2020 (100)
31 CrossFit athletes
3-day estimated diet
record
3.9 ± 1.3
1.6 ± 0.4
Noh et al. 2020
(92)
15 endurance athletes
3-day diet record
7.63
1.9 ± 0.8
Kumahara et
al. 2020 (101)
17 collegiate lacrosse
players
2x 7-day image-
assisted diet records
Transition phase: 4.9 ± 0.9
Preparatory phase: 5.2 ± 1.1
Transition phase: 1.1 ± 0.3
Preparatory phase: 1.2 ± 0.3
Baranauskas et
al. 2020 (102)
54 elite “anaerobic”
athletes
76 elite “mixed” athletes
117 elite “aerobic”
athletes
1x 24-hour recall
3.7 ± 2.2
4.7 ± 1.5
5.0 ± 2.4
1.2 ± 0.4
1.4 ± 0.6
1.5 ± 0.7
Brown et al.
2020 (103)
17 collegiate dancers
3-day diet record
3.7 ± 1.6
1.1 ± 0.5
Jenner et al.
2019 (104)
23 professional
Australian Football
League players
3-day estimated diet
record
2.7 ± 0.7
(4.0% meeting
recommendations)
1.6 ± 0.5
(74.0% meeting
recommendations)
Condo et al.
2019 (105)
30 Australian rules
football players
3x web-based 24-hour
recalls6
3.0 ± 0.8
(0.0% meeting
recommendations)
1.5 ± 0.5
(77.8% meeting
recommendations)
Gillen et al.
2017 (109)
39 strength athletes
104 team sport athletes
83 endurance athletes
3x web-based 24-hour
recalls7
Not reported
1.5 ± 0.4
1.3 ± 0.3
1.5 ± 0.4
22
Wardenaar et
al. 2017 (106)
83 elite & sub-elite
endurance athletes
104 elite & sub-elite
team-sport athletes
39 elite & sub-elite
strength athletes
3x 24-hour recalls
5.0 ± 0.9
(45.6% meeting
recommendations)
3.7 ± 0.6
(19.5% meeting
recommendations)
4.3 ± 1.1
(27.0% meeting
recommendations)
1.5 ± 0.2
(78.2% meeting
recommendations)
1.3 ± 0.2
(58.2% meeting
recommendations)
1.5 ± 0.4
(29.1% meeting
recommendations)
1Values represent mean ± SD unless stated otherwise
2Participants were asked to take photos of their food where weighing was not possible (e.g., eating out at a restaurant),
3Value represents mean intake, calculated from reported mean absolute intake (g·day-1) and mean body weight (kg)
4Phase 1: heavy practicing for 2-3 hrs·day-1, separated by 1-2 game days of nonconference basketball games, with one rest day per week;
Phase 2: similar to phase 1, except all but one game was in conference league play; Phase 3: decline in practice time to a maximum of 1.5
hrs·day-1, Phase 4 & 5: off-season, with weekdays consisting of one basketball-specific training day, paired with an aerobic or anaerobic
conditioning workout, with one hour of resistance training the following day, weekends off, and no games.
5Block 2014 Food Frequency Questionnaire (110)
6Automated Self-Administered 24-hour Dietary Assessment Tool (ASA24) (111)
7Compl-eatTM (Validation study: (112))
Abbreviations: FFQ food frequency questionnaire
23
2.2.2 Protein
2.2.2.1 Protein Intake Guidelines for Female Athletes
Protein is often referred to as a ‘building block’ nutrient, for its role in repairing and
enhancing structural adaptations in muscle and connective tissue and synthesising
metabolic proteins (113). Dietary protein stimulates muscle protein synthesis (MPS)
and provides exogenous amino acids to incorporate into new proteins (114, 115).
Specific requirements differ by exercise type, volume, and intensity, as well as training
status, but current guidelines recommend that athletes consume 1.2-2.0 g·kg-1·day-1 of
protein, spread throughout the day (61). During periods of energy restriction, protein
intakes exceeding 2.0 g·kg-1·day-1 may be beneficial to minimise the loss of fat-free mass
(FFM) (61, 116, 117). Furthermore, training with low carbohydrate availability can
increase protein requirements due to greater exercise-induced amino acid oxidation
(118).
Essential amino acids are uniquely responsible for increasing MPS, with the branched-
chain amino acid leucine acting as the “trigger” (115, 119, 120). While total daily protein
intake appears to be the most important dietary factor for stimulating muscle
development, how this is distributed across the day, and relative to training sessions is
also important for optimizing metabolic adaptations to exercise and supporting tissue
turnover (61, 121). Moderate protein intake shortly after exercise, and the continued
intake across several meals in the extended recovery period, act synergistically with the
stimulatory effects of exercise to activate myofibrillar and mitochondrial protein
synthesis and enhance training adaptations (122-127). Sports nutrition guidelines
recommend consuming 0.25-0.30 g·kg-1 of protein, comprising approximately 10 g of
essential amino acids and 1-3 g of leucine, in the two hours immediately following
exercise, and then every 3-5 hours thereafter (61).
Predictably, all the aforementioned research on optimal protein intakes has been
conducted in male populations, yet oestrogen and progesterone can influence protein
metabolism (62), raising the questions of whether the protein requirements of female
athletes differ: a) from their male counterparts, and b) across the menstrual cycle. A
24
2020 systematic review of 14 studies aiming to determine the protein requirements of
premenopausal female athletes concluded that the estimated average requirements
(EAR) were 1.28-1.63 g·kg-1·day-1 for aerobic endurance athletes, 1.49 g·kg-1·day-1 for
resistance athletes, and 1.41 g·kg-1·day-1 for athletes undertaking intermittent exercise
of ~60-90 minutes in duration (128) - values that lie within current guidelines (61). An
EAR is the amount estimated to meet the requirements of 50% of healthy people in a
particular population group: extrapolation to recommended daily intakes (RDI), which
represents the amount estimated to meet the requirements of 97-98% of healthy people
in a population group (and are more appropriate at an individual level) (129), yields
estimates closer to the upper end of current guidelines (61) of 1.59-2.02 g·kg-1·day-1,
1.85 g·kg-1·day-1 and 1.75 g·kg-1·day-1 for aerobic endurance, resistance, and
intermittent exercise athletes respectively (128). However, the protein requirements of
endurance athletes were derived from three studies using nitrogen balance, a method
which may not be suitable for estimating the protein requirements of athletic
populations (130). Similarly, the indicator of amino acid oxidation technique used by
two other studies in the review may overestimate requirements at high habitual protein
intakes without a sufficient dietary adaptation period (131).
Research investigating optimal acute protein intakes or post-exercise protein ingestion
rates in a dose-response manner in females is lacking. West et al. found no evidence of a
difference in the MPS response to resistance exercise in the fed state between males and
females when 25 g of whey protein was ingested immediately after a bout of high-
intensity resistance exercise (132). Given differences in body mass, this equated to
~0.37 g·kg-1 for females, and ~0.32 g·kg-1 for males (132). Several other studies
observed favourable physiological responses with 0.32-0.38 g·kg-1 of protein following
resistance (133-136) and repeated-sprint exercise (137) in trained females. While these
doses are slightly greater than what current guidelines indicate (61), lesser amounts
may have elicited the same response, and the optimal protein dosage required to
achieve maximal MPS in females is yet to be determined (128).
In their systematic review, Mercer et al. were unable to determine the effect of
menstrual cycle phase or hormonal contraceptive use on daily protein requirements
due to a lack of research investigating or controlling for such factors (128). However, it
25
is plausible that protein requirements may be slightly greater in the luteal phase
compared with the follicular phase (128). Several studies have observed greater protein
catabolism, leucine flux, and nitrogen utilisation in the luteal phase compared with the
follicular phase, both at rest and following aerobic exercise (138-141), presumably
attributable to the luteal phase increase in progesterone (142). However, the influences
of any differences in protein metabolism across the menstrual cycle on actual protein
requirements, or the extent to which such effects may differ by exercise type, have not
been quantified (128).
While sports nutrition research in female athletic populations is incredibly limited,
particularly those investigating the effects of the menstrual cycle, no convincing
evidence currently exists to suggest that the ranges provided by current sports
nutrition guidelines are inadequate or inappropriate to meet the needs of female
exercisers (19).
2.3.2.2 Protein Intakes of Female Athletes
Cross-sectional research indicates that daily protein intakes of female athletes generally
fall within the 1.2-2.0 g·kg-1·day-1 range (61, 91-96, 98-100, 102, 104-106, 109). Across
20 studies included in a systematic review of field-based team sport athletes, mean
protein intake was 1.4 ± 0.3 g·kg-1·day-1, with 16 studies reporting mean intakes that
met daily recommendations (91). Table 4 shows that in the last five years, mean daily
protein intakes for females across a range of sports were mostly within the
recommended range (61), with the majority falling between 1.2-1.6 g·kg-1·day-1 (93-100,
102, 104-106, 109). While such intakes would be considered adequate in terms of
overall daily recommendations, they do fall closer to the lower end of the range, with no
study reporting a daily mean intake above 2.0 g·kg-1·day-1 (although underreporting is
likely to have occurred (107)). Furthermore, given that many of the same studies
reported sub-optimal carbohydrate intakes (and given the high prevalence of LEA
amongst exercisers (29)), protein requirements may indeed be slightly higher in these
populations, and daily intakes closer to 2.0 g·kg-1·day-1 may be beneficial (61, 116-118).
Only three studies reported mean protein intakes below 1.2 g·kg-1·day-1 (during at least
one time point) (97, 101, 103), while Noh et al. recorded the highest intakes of
26
1.9 g·kg-1·day-1 in a cohort of endurance athletes (92).
However, even if total daily protein intake is adequate, distribution across the day may
be suboptimal to maximise training adaptations (61, 109). Gillen et al. found that well-
trained female athletes met total daily protein requirements (mean intake:
1.4 ± 0.4 g·kg-1·day-1), but when pooled with male data, distribution was skewed
towards the end of the day, with 19% of total daily protein consumed at breakfast
(19 ± 8 g), 24% at lunch (25 ± 13 g) and 38% at dinner (38 ± 15 g) (109). Protein intake
was less than 20 g for 58% of athletes at breakfast, 36% at lunch, and 8% at dinner
(109). Regarding protein intake across the menstrual cycle, Ihalainen et al. found no
evidence of a difference in mean intakes between the early-follicular (112 ± 40 g·day-1),
mid-follicular (107 ± 30 g·day-1), ovulatory (110 ± 27 g·day-1), and luteal (105 ± 31
g·day-1) phases, all of which equated to ~1.6 g·kg-1·day-1 (108).
2.2.3 Fat
Fat is a requisite part of a healthy diet, but the guidelines for athletes are less specific
than those for carbohydrate and protein (61). Fat intake should be sufficient to provide
the required essential fatty acids, fat-soluble vitamins, and energy to maintain weight,
but should be individualized based on training status, energy requirements, and
performance and body composition goals (61). It is suggested that an athlete’s fat intake
should still align with public health guidelines (61), but may need to be slightly higher
in the case of extreme energy requirements (143). The Eating and Activity Guidelines
for New Zealand Adults recommend that fat intake comprises 20-35% of total energy
intake (TEI) and saturated fat and trans-fat together be limited to no more than 10% of
TEI (69).
To date, evidence is lacking to support any ergogenic benefits of a chronic high-fat diet
for athletes (144, 145). Whilst some research has demonstrated cellular adaptations
that appear theoretically advantageous (146, 147), these fail to consistently translate
into meaningful performance improvements (144, 145). Conversely, chronic very low-
fat diets (fat intake below 20% of TEI) are likely inadequate to provide sufficient
essential fat-based nutrients (61). In some unique situations, such as acute periods of
27
carbohydrate loading before competition, or multi-day self-supported ultra-endurance
events, either extreme may be unavoidable, or indeed favourable, but these are very
much short-term strategies (61, 145).
2.2.4 Energy Availability
Energy availability is defined as the energy remaining from dietary intake after
subtracting the energy expenditure of exercise (EEE), expressed relative to FFM, and
refers to the dietary energy ‘available’ to maintain physiological functioning of tissues
and organs (148). It is distinct from the concept of energy balance, which simply refers
to energy intake minus total energy expenditure (148). In athletes or highly active
individuals, energy availability is considered a more appropriate index for quantifying
nutritional status and is emerging as an important concept for guiding dietary practices
in these populations (29).
Low Energy Availability (LEA) arises from a decrease in energy intake, an increase in
EEE, or both, and is a potent physiological stressor that can disrupt endocrine and
metabolic processes in a matter of days (33, 117, 149, 150). The female athlete triad
was a term first coined in 1992 to describe the concurrent amenorrhoea, low bone
density, and disordered eating observed in female athletes (151), ultimately as a result
of LEA (151, 152). In 2014 the IOC released a consensus statement titled “Beyond the
Female Triad – Relative Energy Deficiency in Sport (RED-S)”, proposing RED-S as an
alternative model to capture the wider range of physiological consequences associated
with LEA and recognise that males can be affected as well (33). The consensus
statement defined RED-S as “impaired physiological functioning caused by relative
energy deficiency and includes, but is not limited to, impairments of metabolic rate,
menstrual function, bone health, immunity, protein synthesis, and cardiovascular
health” (33). The following potential effects on athletic performance are also described:
decreased muscle strength, endurance performance, training response, glycogen stores,
coordination, judgement and concentration; increased injury risk and irritability; and
depression (33). Understanding of both models is still developing, and while some
debate exists around the overlap of each, and whether RED-S can truly replace the
28
female athlete triad, they are both aetiologically underpinned by, and provide a
framework for understanding, LEA (33, 153-155).
In healthy adults, 45 kcal·kgFFM·day-1 equates to an energy balance of approximately
zero, and is considered ‘adequate energy availability’ (148). While the specific
thresholds and durations of LEA required to elicit physiological impairments are not yet
well understood, short-term (five days) well-controlled laboratory studies in females
have shown that energy availability of 30 kcal·kgFFM·day-1 reduces resting MPS (117)
and decreases markers of bone formation (149), while energy availability below
30 kcal·kgFFM·day-1 reduces LH pulsatility (150). Of course, these studies only capture
the initial metabolic and endocrine disruptions with relatively dramatic “doses” of LEA.
In the current absence of longer-term interventions, how these perturbations may
change over time, and the effects of prolonged, but less severe reductions in energy
availability remain unclear. Indeed, it is likely that the duration, magnitude, and
frequency of periods of LEA all have a role to play in physiological outcomes (156).
But whether the 30 kcal·kgFFM·day-1 threshold can be extrapolated to free-living
individuals is debatable, and complicated by the challenges of accurately measuring
energy availability and quantifying the duration of LEA exposure in these populations
(32, 156). Lieberman et al. and Williams et al. both showed that menstrual dysfunction
at least, wasn’t correlated with a particular threshold of energy availability, but that the
rate of menstrual disturbances (but not necessarily the severity) did increase as energy
availability decreased (157, 158). Similarly, Reed et al. found mean energy availability
to be greater than 30 kcal·kgFFM·day-1 (but lower than 45 kcal·kgFFM·day-1) in
exercising females with eumenorrhoea, subclinical menstrual disturbances, and
amenorrhoea (159). Energy availability did not differ between different subclinical
forms of menstrual disturbances but was significantly lower in amenorrhoeic versus
eumenorrhoeic individuals (mean: 30.9 vs. 36.9 kcal·kgFFM·day-1) (159). But the
recovery of biomarkers and reversal of RED-S symptoms may be quicker following
shorter periods of LEA (i.e., weeks to several months), as opposed to more prolonged
durations, and intermittent bouts may offer less opportunity for metabolic disruptions
and performance deficits to manifest (156).
29
The failure of observational research to establish a threshold below which
endocrinological and metabolic disturbances consistently arise is likely due to actual
differences between energy availability and outcomes, duration of exposure, and the
lack of a standardized protocol for measuring energy availability (32). Indeed, health
impairments appear to occur across a continuum of decreases in energy availability,
rather than at a single identifiable threshold, with various body systems differing in the
nature of their functional decline (32). But methodological inconsistencies in the
definition and quantification of energy availability cannot be overlooked. Guebels et al.
demonstrated how different definitions of EEE in the same population can lead to
considerably different estimates of energy availability and interpretations of adequacy:
mean energy availability was 28.2 ± 9.0 kcal·kgFFM·day-1 when EEE was defined as “all
planned exercise, bike commuting, and all walking”, compared with 36.7 ± 10.2
kcal·kgFFM·day-1 when EEE defined as “all exercise >4.0 metabolic equivalents of task
(MET)” (160) (more detail in section 2.4.2).
Furthermore, interactions with additional factors such as dietary composition (66, 67,
161, 162), within-day (163) and between-day energy deficits (164), psychological stress
(165), and potentially gynaecological age (156) may modify the effects of LEA on body
systems. An important issue with the real-world application of controlled laboratory
studies is the consistent level of energy availability typically applied throughout the
intervention (117, 149, 150, 164). In reality, differences in dietary intake and training
load mean that energy availability tends to fluctuate on a day-to-day basis (164), but
how the magnitude and frequency of such differences affect physiological functioning is
unclear: large between-day differences that span the continuum from low to high
energy availability may impose greater metabolic stress that a more consistent
adequate energy availability does not (32). Whilst there is no established optimal
recording period for measuring energy availability, calculating mean energy availability
over multiple days may mask such variations (32).
However, while questions remain about the degree and duration of LEA required to
elicit metabolic and endocrine disruption, and the influence of other factors, the
negative consequences LEA and its sequelae pose to health and performance are
evident (153, 155, 166). Critically, an alarming number of female exercisers appear to
30
be at risk, preventing them from not only reaching peak performance but achieving
optimal health throughout life (29). Estimates of prevalence range widely between 14%
and 63% across different sports (29). Traditionally there has been a greater focus on
females in this field of research, and it appears females may be more vulnerable to the
effects of LEA than males (166). Both recreational and elite-level athletes across a range
of sports are affected, but endurance, aesthetic, and weight-category sport athletes may
be at greater risk (33, 166). Energy availability has established itself as a key concept in
sports nutrition, but the understanding of its complexities is still developing (29). How
energy availability should be measured and optimised in practice requires further
attention, especially given the prevalence of LEA amongst female exercisers.
2.3 Measuring Energy Availability
Further to the discussion in section 2.2.4, researchers and nutrition professionals face
the challenges of accurately measuring energy availability and the lack of a standardized
protocol for doing so in free-living individuals (32). Significant sources of error exist
when quantifying dietary intake, EEE, and FFM, and any estimates of energy availability
should always be interpreted in the context of the methods used (32). Due to these
challenges, focus has recently shifted from quantitative measures of energy availability
towards more “objective parameters” of LEA, based on physiological symptoms and
“endpoint” biomarkers (29, 32, 156). However, further research is still required to
identify appropriate biomarkers and athlete-specific reference ranges for short-,
medium-, and long-term LEA (156).
2.3.1 Dietary Assessment
Useful estimates of dietary intake rely on accurate methods of dietary assessment. Food
frequency questionnaires, 24-hour recalls, and diet records are the methods most
typically used in research, but the best method will depend on the research question
and the study objective (167).
31
2.3.1.1 Food Frequency Questionnaires
Food frequency questionnaires consist of a finite list of foods and beverages
accompanied by ‘frequency of consumption’ categories and provide descriptive data on
usual intakes over a long period (167). Some may ask the respondent to also indicate
the amount of food or beverage that they usually consume, but at best they are semi-
quantitative (167). Food frequency questionnaires can be used in epidemiological
studies to rank intakes of specific foods, food components, or nutrients to compare with
disease prevalence and mortality, or to identify dietary patterns associated with
inadequate nutrient intakes (167). Whilst they are quick and have a low respondent
burden, they are less accurate than fully quantitative methods (167). Because the
results represent average intake over an extended period (eliminating the influence of
daily variability) (167), they are less appropriate for quantifying variation between
specific short-term periods, such as phases of the menstrual cycle.
2.3.1.2 24-Hour Recalls
Quantitatively assessing dietary energy and nutrient intakes therefore relies on 24-hour
recalls or diet records (167). Twenty-four-hour recalls retrospectively assess an
individual’s actual dietary intake and involve a trained interviewer asking the
respondent about everything they consumed over the past 24 hours or previous day,
using a standardized interview protocol (167). Multiple 24-hour recalls are required to
obtain estimates of an individual’s usual intake of food and nutrients, with the number
depending on the nutrient(s) of interest, seasonal variation, study population, and
within-person variations in intake (167). Some of the main advantages of 24-hour
recalls are their low respondent burden (yielding high compliance) and the low levels of
literacy and numeracy required (167). Given they are typically administered on random
days, the element of surprise is also considered a strength (167), however, this is
perhaps less applicable when studying the menstrual cycle.
32
2.3.1.3 Diet Records
Diet records, on the other hand, are prospective, and involve the respondent recording
all food and drink at the time of consumption for a specified period (167). The
respondent includes a description of the food or beverage, preparation method, and
time of consumption, with amounts either given as household measurements or counts
(estimated diet record) or by weight (weighed diet record) (167). Prospective dietary
assessment can decrease recall error, but it also introduces the possibility of reactivity,
whereby the respondent modifies their diet to simplify the recording process, or to
comply with social desirability (167). When comparing differences in dietary intake
within the same person – differences that may be subconscious, and not apparent
retrospectively – prospective dietary assessment becomes particularly important. While
estimated diet records tend to be less burdensome for the respondent, weighed records
are the most precise method for estimating usual intake of food and nutrients for
individuals as they decrease the chance of over- and under-estimation of portion sizes
and show better reproducibility (167). However, multi-day weighed diet records do
require a certain level of literacy and numeracy as well as a high level of participant
motivation, so are not appropriate for use in all populations (167).
As with 24-hour recalls, the ideal number of recording days depends on several factors,
but the additional respondent burden of keeping a diet record must also be considered:
as the recording period increases, the quality of information generally declines (168).
Diet records typically range between 1-7 days in duration: a single diet record is
sufficient for quantifying the average intake of a group, whereas multiple days are
required to quantify the usual intake of an individual (167, 168). While 3-4 days is fairly
typical for assessing habitual energy intake in sedentary populations (167-169), the
most appropriate recording period for athletic populations may vary between sports
and the nature of training micro-cycles (170, 171). For both 24-hour recalls and diet
records, in sedentary populations at least, weekend days should be included
proportional to weekdays to capture any differences in weekend dietary patterns, and
depending on the study objective, seasonality may also need to be accounted for (167).
33
However, the nature of the menstrual cycle does impose limitations on the duration of
the recording period, the ability to capture training micro-cycles, and the inclusion of
weekends. While several studies measuring dietary intake across the menstrual cycle
actively sought to include both weekdays and weekends in their multi-day diet records
(172-175), this can be difficult and often unrealistic. Furthermore, intentionally trying
to capture a weekend day could lead to inconsistent phase time points between
participants. A more appropriate approach may be to measure across multiple cycles
and ensure a large enough sample size to reach a sufficient balance between phases, or
to control for any day-of-the-week effects a-posteriori. At the very least, researchers
should report the ratio of weekdays to weekend days in the final sample. Several studies
obtained self-reported daily diet records for the duration of one or more complete
menstrual cycles (20, 22, 176, 177), however, this places an enormous burden on
participants, and likely compromises the quality of information collected (168). The
number of recording days is perhaps most appropriately guided by the bounds of the
phase definitions proposed by Elliott-Sale et al.: days 1-5 for the early follicular phase,
and 7-9 days following a positive ovulation test for the mid-luteal phase (8).
2.3.1.4 Image-Based and Image-Assisted Methods
Mobile technology that incorporates food photography has the potential to improve
accuracy and compliance in dietary assessment (178). Several studies have investigated
using images captured from handheld devices or wearable cameras during eating
occasions to supplement traditional dietary assessment methods (image-assisted), or to
provide the primary record of intake (image-based) (178). These methods can reveal
unreported foods, cooking methods, condiments, and self-reporting errors not captured
by conventional methods, and participants tend to prefer them (179). However, the
technology is still developing, and several barriers to implementing such methods still
exist, including technical limitations, financial and time costs to researchers, and lack of
thorough validation (178). But incorporating digital photography appears to be a viable
way of improving the accuracy of diet records, especially where weighing food may not
be practical or possible (180).
34
Simpson et al. tested the feasibility of the smartphone application MealLogger® to
monitor dietary intake and improve nutritional knowledge in a group of young athletes,
and the results provide insight into using the application for dietary assessment (181).
During a six-week intervention, participants logged images and descriptions of their
dietary intake three days per week, which researchers could view and comment on in
real-time to obtain further information, or provide feedback (181). All participants
preferred MealLogger® to traditional pen-and-paper diet records and indicated a
relatively low burden of using the app (181), corroborating the findings of other studies
that have shown participant preference towards electronic methods of dietary
assessment (182, 183).
Currently, weighed diet records still appear to provide the most objective and accurate
dietary information, but respondents tend to prefer recording these digitally (especially
younger populations) (179, 181-183), which likely improves compliance. Applications
such as MealLogger® also allow researchers to probe for further detail in close
proximity to the eating occasion (181), minimising potential recall error. Using digital
photos to supplement weighed diet records may enhance the quality of information
collected, especially in situations where the participant is unable to weigh food or fluid,
for example, eating out at a restaurant (179, 180).
2.3.1.5 Underreporting
The very act of measuring the diet can change dietary intake, so while recorded intake
may be accurate, it may not reflect usual intake (107, 167). Underreporting occurs in
both general and athletic populations and is a critical barrier to obtaining accurate
habitual dietary intake data (107, 184-186). A 2017 meta-analysis compared the self-
reported energy intakes of athletes with total energy expenditure (TEE) based on
doubly labelled water and found that participants underreported mean energy intake
by 19%, representing a discrepancy of 2793 ± 1134 kJ·day-1 (668 ± 271 kcal· day-1)
(107). Underreporting of habitual energy intake can result from both undereating
(eating less than usual, or insufficient to maintain body weight) and under-recording
(failing to record all items consumed, or underestimating amounts, without a change in
body weight) (187). The relative influence of each appears to differ by population:
35
although total under-reporting tends to be more common in females than males,
undereating may play a more important role than under-recording in highly motivated
lean women (187). In sedentary populations, under-reporting tends to increase as body
mass index (BMI) increases (186, 188), but also appears to increase with greater energy
requirements, potentially relating to the increased burden of recording large volumes of
food and frequent eating occasions (171, 189). Indeed, assessing the diets of athletes
presents the unique challenges of large portion sizes, frequent snacking, irregular meal
patterns, eating outside the home, and consuming sports foods and supplements (171).
While selective underreporting of food or macronutrients is less well described, under-
reporters appear to overreport carbohydrate and protein intake and underreport fat as
percentages of TEI (186). This may be due to overreporting foods deemed to be
“healthy”, and underreporting those deemed to be “unhealthy” (167). Over-reporting of
energy intake tends to be less problematic in prospective dietary assessment, but can
still occur by incorrectly estimating portion size (167).
In the absence of expensive techniques like doubly labelled water, Goldberg cut-offs can
be used to assess the validity of self-reported dietary energy intakes and identify
underreporting (190). For small studies (n<100), it is beneficial to measure daily energy
expenditure (via activity diaries, accelerometers, or heart rate monitors, for example),
which can then be compared to energy intake (190). Even when using weighed records,
underreporting may still occur in some populations (185), and only so much can be
done to eliminate recording errors of behavioural and psychological origin in dietary
assessment. While no evidence exists to suggest that underreporting changes across the
menstrual cycle, it is worth noting that any observations of absolute or phase-related
differences in dietary intake could be underestimating the actual difference.
2.3.2 Measuring Energy Expenditure of Exercise
Energy availability differs from energy balance as it doesn’t take into account TEE, only
EEE, eliminating the need to quantify the other components of TEE such as: resting
metabolic rate (RMR), the thermic effect of food, non-exercise activity thermogenesis,
and growth (32). However, accurately estimating even just EEE is challenging: a variety
of different methods exist for doing so, but no accepted gold standard method (32, 156).
36
The doubly-labelled water method is typically considered the gold standard for
measuring energy expenditure in free-living individuals (191), but aside from being
expensive and often unfeasible, it does not distinguish EEE from TEE (156). Some
studies have used accelerometers (which monitor body movements) (192-195), whilst
others have used the relationship between heart rate and oxygen consumption and
applied this to heart rate data from training logs (196-200). Heart rate monitors, power
meters, and Global Positioning System units can help estimate an individual’s energy
expenditure for simple activities such as running and cycling, but quantifying the energy
cost of more complex activities such as team sports or resistance training is difficult
(32). Some studies have used a combination of methods for estimating EEE (201, 202),
although Burke et al. recommend using the same method across all forms of exercise
(32). Commercially available wearable devices are generally unreliable for quantifying
EEE (203), and even accelerometers that are commonly used in research tend to
underestimate EEE for vigorous exercise (204). While perhaps less precise, using tables
of energy cost of exercise, such as METs, is also common throughout the literature (205-
218). Metabolic Equivalents of Task (METs) cover a wide range of physical activities and
can therefore be useful for studies in athletes that participate in a variety of types of
exercise (219).
Furthermore, there is no consensus on what should be considered ‘exercise’ (32).
Whether EEE only encompasses purposeful training, or further extends to any physical
activity such as transport, leisure, or daily living activities is unclear, and definitions
vary throughout the literature (32). In the idealistic case of an athlete who is largely
sedentary outside of measurable bouts of purposeful training, the definition of ‘exercise’
remains simple, but in free-living individuals, the reality tends to be less clear-cut. The
Eating and Activity Guidelines for New Zealand Adults define moderate physical activity
as 3.0-5.9 METs (69), so in the absence of an accepted definition, a minimum of 3.0
METs would seem an appropriate threshold to define ‘exercise’ in the context of energy
availability.
The original definition of EEE in the energy availability equation also involves
subtracting the habitual waking energy expenditure that would have occurred during
the exercise period, from the total energy cost of the session (36), but this is rarely done
37
in practice (32). Failing to perform this adjustment may overestimate EEE, and
therefore underestimate energy availability (particularly for those who undertake
prolonged periods of moderate-intensity exercise) but doing so correctly introduces yet
another source of error (32). Accurately measuring an individual’s resting energy
expenditure is time-consuming, and whilst several standard equations can be used to
predict RMR, they have limitations, particularly in athletic populations (such as
overestimating energy expenditure in metabolically adapted athletes, or failing to take
into account FFM) (32, 220). However, the Cunningham equation has been shown to
best predict RMR in female collegiate athletes (221), male and female endurance
athletes (222), and male and female recreational athletes (223).
Cunningham Equation (224):
𝑅𝑀𝑅 = 500 +22 × 𝐹𝐹𝑀(𝑘𝑔)
2.3.3 Measuring Fat-Free Mass
Discrepancies in estimates of energy availability attributable to measurement errors of
FFM are relatively small compared to the other two inputs (TEI and EEE) (32).
Nevertheless, important limitations still exist with the variety of methods available for
measuring body composition, and there is currently no universally accepted method for
doing so in free-living individuals (32, 225). Hydro-densitometry, ultra-sound, three-
dimensional scanning, and air-displacement plethysmography are among the less
frequently used methods in athletic populations, for reasons including access to
expensive specialist equipment, impracticality, and insufficient validation (225). Dual-
energy X-ray absorptiometry (DEXA), bioelectrical impedance analysis (BIA), and
skinfold thickness are more commonly used in sports nutrition settings (225). All three
methods, however, still rely on certain assumptions or constants and require
appropriate standardisation to ensure reliable assessment (225).
38
DEXA works by passing beams of low and high energy X-ray photons through the body,
which are attenuated according to the density and thickness of different tissues: soft
tissues like fat mass (FM) and FFM allow greater passage of photons compared to
denser tissues like bone (226). Despite its original intended use for measuring bone
mineral density, DEXA is now often considered the criterion standard for assessing
body composition (225). However, several technical considerations (such as body size
limitations, greater errors in those with high levels of FFM, and machine and software
differences) and standardisation factors (such as recent exercise, fasting status, muscle
glycogen content, hydration status, and time of day) can affect the reliability of FFM
measurements (225). Furthermore, whilst the amount of ionising radiation exposure is
considered low, there is some debate over how regularly DEXA should be used (225).
But perhaps one of the biggest barriers to using DEXA is access and affordability (225).
Skinfold measurements have long been used in applied sports settings for assessing
body composition due to their convenience and affordability (225). Skinfold thickness
assessment involves using calipers to measure the thickness of a double fold of pinched
skin and the underlying subcutaneous fat at specific anthropometric sites across the
body by an accredited International Society for the Advancement of Kinanthropometry
technician (225). Regression equations are then used to estimate body density and fat
percentage from the sum of the skinfolds but rely on several assumptions, such as
uniform skinfold thickness and compressibility, consistent fat patterning across the
body, and a fixed relationship between subcutaneous fat and visceral fat (227). It is
important to select an appropriate equation that has ideally been validated in the target
population (227). Furthermore, the accuracy of skinfold measurements greatly depends
on the technique of the measurer, with inter-individual discrepancies common (225).
While assessing body composition via skinfolds appears to be the least vulnerable to
everyday activities, hydration status, and recent food consumption (225), it is perhaps
more appropriate for tracking changes in discrete physical characteristics (for example,
in response to training or a nutritional intervention), rather than for specifically
estimating FFM (32, 227). Finally, taking skinfold measurements can be an
uncomfortable or intrusive experience, particularly for those with body image issues
(227).
39
BIA involves passing a small electric current through the body to indirectly measure
total body water volume (228). Calculating the impedance (resistance) of the current
enables inferences about body composition to be made: FFM has a higher water content
and therefore less impedance than FM (228). BIA devices exist as ‘single-frequency’, or
‘multiple-frequency bioelectrical impedance analysers’ and use regression equations to
estimate body composition (which, like skinfolds, should be population-specific) (228).
However, as these are based on body water volume, factors such as hydration status
(229), recent food consumption (230), and exercise (231), can influence estimates of
FFM. Other limitations include the considerable variability between devices and
equations (232), the large proportion of whole-body impedance attributable to the
limbs, despite a relatively low contribution to overall body mass (233), and limited
validation in athletic populations (225). However, BIA is incredibly quick, easy, and
cheap, and requires minimal expertise to administer (225). Similar to skinfold
assessment, BIA is considered a useful tool for tracking changes over time, provided the
same device, population-specific equations, and appropriate standardisation are
implemented (232).
As fluid retention can potentially fluctuate across the menstrual cycle (234), the effect
that this might have on estimates of FFM, and thus energy availability, needs to be taken
into account when considering methods influenced by hydration status. Historically it
was recommended that females avoid body composition assessment when experiencing
any menstrual cycle-related water retention (235), however, recent research indicates
that measures of body composition obtained via DEXA, skinfolds, and single- and multi-
frequency BIA remain consistent across the menstrual cycle (236, 237).
When the purpose of measuring energy availability is to assess nutritional status or
identify those at risk of LEA, accurate and precise measurement of each input is highly
important. However, when the purpose is rather to quantify differences in energy
availability between time points (for example, between different phases of the
menstrual cycle), using consistent definitions and methods can reduce, or altogether
eliminate, many of the aforementioned limitations (for example, selecting the correct
equations, how exercise is defined, and even to a certain extent, under-reporting of
energy intake).
40
2.4 Dietary Intake Across the Menstrual Cycle
2.4.1 Energy Intake
2.4.1.1 Energy Intake in the Follicular versus Luteal Phases
While research is lacking in athletic populations, studies in the general population
suggest that daily energy intake is greater in the luteal phase of the menstrual cycle
compared with the follicular phase in naturally menstruating females (20, 25, 26, 172-
174, 238-243). Appendices A and B summarise the studies that have compared energy
intake in different phases of the menstrual cycle. A study in 30 Brazilian females found
mean energy intake to be 1730 ± 254 kcal·day-1 in the follicular phase, compared with
2259 ± 375 kcal·day-1 in the luteal phase, representing a mean increase of 529 kcal·day-1
(173). Similarly, Reimer et al. reported a mean increase of 337 kcal·day-1 from the
follicular phase to the luteal phase (240), while Martini et al. reported a mean increase
of 159 kcal·day-1 (26). Many others have observed a similar pattern with mean
increases from the follicular to luteal phase ranging from 90 to 504 kcal·day-1 (20, 25,
26, 172-174, 176, 177, 238-242), yet several studies have found no evidence of a
difference in phase-related energy intake (23, 24, 108, 175, 244).
Differences in study quality may in part explain conflicting findings: failing to verify
menstrual cycle phases according to recommended methodology risks collecting data
outside of target phases or including cycles with abnormal hormonal profiles (8). For
example, Chung et al. observed a mean increase in energy intake from the follicular to
the luteal phase of 160 kcal·day-1, but after excluding 17 of the 39 participants with
incorrectly identified cycles phases (via retrospective analysis of serum hormones), the
difference increased to 180 kcal·day-1 (174). This not only highlights the large number
of cycles that could be misclassified without hormonal verification (~44%), but shows
that when such cycles are included, the reported phase-related difference in energy
intake may not reflect the actual difference for eumenorrheic females. Tarasuk & Beaton
reported one of the smaller mean increases in energy intake between the follicular and
luteal phases of 90 kcal·day-1, however phases were based on retrospective accounts of
menses reported after the year-long study (25). Of those that adhered to recommended
41
phase verification methods, five studies observed greater energy intakes in the luteal
phase compared with the follicular phase of between 159 and 529 kcal·day-1 (20, 26,
173, 174, 240), whilst Gorczyca et al. found no evidence of a difference in energy intake
between the early-follicular, mid-follicular, ovulatory, or luteal phases (24).
Several studies that verified phases with serum hormones compared energy intake
between cycle phases based on one meal alone: Brennan et al. reported a greater mean
energy intake during the luteal phase compared with the follicular phase (243), while
McNeil et al. reported a trend for menstrual cycle phase, with intakes in the early-
follicular, late-follicular, and mid-luteal phases of 670 ± 293, 525 ± 289, and
711 ± 334 kcal respectively (245). However, data from a single meal may be insufficient
to capture an overall pattern of energy intake, and extrapolating results to usual daily
intake may be less appropriate. Similarly, the buffet or menu-style meals provided in
these studies may not accurately reflect food choices made in free-living conditions (and
their relationship to cycle phase).
2.4.1.2 The Late-Follicular and Ovulatory Phases
Simply comparing the follicular phase with the luteal phase fails to acknowledge
potential changes in dietary intake within these broader umbrellas. Certainly, the high
oestrogen and low progesterone concentrations of the late-follicular phase, and the
phenomenon of ovulation present unique hormonal environments (7, 46). Given the
large inter-individual variation in follicular phase duration (44), capturing these
relatively transient phases is particularly unreliable without hormonal verification (46).
Yet doing so can be challenging and burdensome, such that they are often neglected (23,
25, 26, 172, 173, 175, 176, 238-243, 246), or merely approximated (22, 177, 244).
Several studies observed a decrease in energy intake during the days leading up to and
including ovulation (20, 22, 245, 247), supporting the argument of Fessler for the
evolutionary advantage of prioritising reproduction over food intake during this time
(248). Roney & Simmons reported a decrease in food intake in the days approaching
ovulation, with the lowest intake coinciding with the peak in salivary oestrogen
indicative of the late-follicular phase (247). Food intake then increased in the luteal
42
phase, tracking closely with the rise in salivary progesterone (247). However, reported
intake was based on a questionnaire about the previous day’s food consumption, meal
size, and hunger, relative to usual consumption, rather than quantitative dietary
assessment (247). Therefore, these findings only provide evidence that hormonal
fluctuations across the menstrual cycle may be associated with changes in perceptions of
food intake, rather than actual energy intake.
Johnson et al. provided stronger evidence to support the hypothesis of Fessler (248):
when cycle phases were verified by serum hormones, the greatest increase in energy
intake was recorded between the ovulatory phase (defined by increasing oestrogen, low
progesterone, peaks in FSH and LH, and a rise in BBT) and the luteal phase, of 166
kcal·day-1 (20). Using a similar hormonal and oral temperature profile to define the late-
follicular/ovulatory phase, McNeil et al. reported the same pattern of energy intake
from a lunch meal alone (the limitations of which have been previously discussed)
(245). This pattern was also observed by Lyons et al. who measured the dietary intakes
of 18 Australian females over one complete menstrual cycle via daily weighed diet
records (22). Using urinary LH tests to detect ovulation, the menstrual cycle was then
divided into five phases: menses, post-menses, ovulatory, post-ovulatory, and pre-
menses (22). Mean daily energy intake was lower during the ovulatory phase compared
with menses and both luteal time points, with the maximum effect observed between
the four days leading up to and including ovulation, and the four days immediately after
(324 kcal·day-1 mean difference) (22). However, when comparing energy intake
between the 10 days pre-menstruation and the 10 days post-menstruation, there was
no longer evidence of a difference (22). Whilst Lyons et al. did not verify phases with
serum oestrogen and progesterone (22), these results do highlight the possibility of
more specific within-phase fluctuations in energy intake.
2.5.1.3 Measuring Across Multiple Cycles
Any phase-related differences in energy intake that may exist, likely vary from cycle to
cycle. To reduce the influence of such variability, recent guidelines now encourage
collecting data across at least two menstrual cycles (8). Despite the evidence in favour
of the aforementioned trend in energy intake across the cycle, several studies failed to
43
observe any phase-related differences (23, 24, 108, 175, 244). Of particular interest are
the findings of Gorczyca et al., who found no evidence of a difference in energy intake
(as measured by 24-hour recalls) between the early-follicular, mid-follicular, ovulatory,
and luteal phases in a sample of 259 American females (24). These results are
interesting given not only the large sample size, but the fact that cycle phases were
verified by serum oestrogen and progesterone, and that all but nine participants were
followed for two cycles (24). Elliot et al. also reported that tracking participants over
two cycles failed to provide evidence for a difference in phase-related energy intake
(175). While energy intake was greater in the follicular phase compared with the luteal
phase based on one cycle, when 13 of the 31 Chinese participants were followed for an
additional cycle, there was no longer evidence of a difference in energy intake (175).
However, several other studies that tracked more than one cycle still found increases in
energy intake from the follicular to the luteal phase of between 90 and 605 kcal·day-1
(25, 26, 176, 238, 242).
It is possible that other factors, such as cultural or societal influences, could attenuate
any potential dietary changes of endocrinological origin, and may play a more important
role in energy intake than the menstrual cycle. Such factors may have varying degrees of
influence across populations differing by age, lifestyle, environment, country, and
culture. But alternatively, the large sample size and favourable methodology of Gorczyca
et al. could suggest that significant phase-related differences in energy intake may not
exist, especially given the methodological biases frequently unaccounted for in the
published literature, which are discussed in subsequent sections (8, 24).
2.4.1.4 Anovulatory Cycles
The presence of menstrual bleeding alone does not guarantee regular hormonal
fluctuations, so without measuring sex hormone concentrations, one is unable to
accurately identify anovulatory cycles or other menstrual irregularities (46). Barr et al.
found that whilst energy intake was on average, 303 kcal·day-1 greater during the luteal
phase compared with the follicular phase for ovulatory cycles, there was no evidence of
a difference in energy intake for anovulatory cycles (238). While these findings were
based on BBT (a less reliable method that may not always correlate well with hormone
44
levels (46)), they do highlight a potential discrepancy between ovulatory and
anovulatory cycles, which needs to be addressed in research. This is particularly
important when considering underlying mechanisms: the rise in luteal progesterone,
which is hypothesised to influence energy intake, necessitates ovulation (249). Studies
that included cycles failing to reach a luteal progesterone threshold of 16.00 nmol. L-1
(46) may have underestimated, or indeed failed to observe, a difference in energy intake
between the luteal phase and other phases.
2.4.1.5 Weekday versus Weekend Dietary Intake
Li et al. found that the increase in energy intake observed during the mid-luteal phase
compared with the mid-follicular phase was primarily driven by weekend intake (172).
Dietary energy intake was 23% greater in the luteal phase when comparing weekend
intakes, but there was no evidence of a difference in phase-related energy intake based
on weekday intakes alone (172). Weekends may offer more freedom to modulate
dietary intake based on appetite and preference, while routine, time constraints, and
food availability may leave weekday dietary patterns less vulnerable to fluctuation.
However, Li et al. considered the weekend to be Thursday to Sunday, which represents
over half the week and may not reflect what is typically considered the weekend. This
phenomenon has not been replicated by any other studies: while Martini et al. observed
mean energy intake to be higher on the weekend (Saturday and Sunday) regardless of
menstrual cycle phase, and energy intake to be higher in the mid-luteal phase compared
with the mid-follicular phase, there was no interaction found between cycle phase and
day of the week (26).
2.4.1.6 Physical Activity
Energy intake and energy expenditure are fundamentally related, yet much of the
research neglects to measure physical activity across the menstrual cycle (22, 23, 25,
172, 173, 176, 239-243, 246). Several studies collected information on exercise habits at
baseline but not during dietary assessment periods (24, 177, 238, 247) (with Barr et al.
excluding those who exercised over seven hours a week (238)), whilst others required
abstinence from heavy exercise during the study (26, 243). When Fong & Kretsch kept
45
physical activity levels consistent across phases in their metabolic ward study (244), no
differences in energy intake across the menstrual cycle were observed, however, the
applicability of these results to free-living conditions makes inferences difficult.
Levels of physical activity could plausibly fluctuate across the cycle for some individuals
(59, 250-252) - such that EEE would become a mediator in the relationship between
cycle phase and energy intake - particularly considering the recent media commentary
on modulating exercise across the menstrual cycle (11-13). Several studies that tested
for phase-related changes in physical activity found no evidence of a difference across
the cycle (20, 26, 174), but most of these were based on relatively crude estimations of
exercise (such as the hours per day of exercise without collecting information on
intensity (174), or only asking participants to report daily physical activity above their
subjective normal levels (26)), and for Chung et al. this only included the four
participants who regularly exercised (174).
2.4.1.7 Energy Availability Across the Menstrual Cycle
While it is important to recognise the potential influence of physical activity on dietary
intake when considering changes in energy intake across the menstrual cycle,
accounting for physical activity in largely sedentary populations may not be as
important as it is for athletes (or those who are very physically active), for whom daily
energy intake may vary dramatically depending on training load. In these individuals,
energy availability is considered a more appropriate index for quantifying nutritional
status (29). Whilst recent research has primarily focused on how energy availability
affects menstrual function (see section 2.3.4), there has been less attention in the
opposite direction: how the menstrual cycle affects energy availability. As energy
availability consists of three components: energy intake, EEE, and FFM (148), for any
differences in energy availability to exist between menstrual cycle phases, at least one of
these factors would have to fluctuate, but such that the effect of more than one changing
would not cancel the other out.
46
To date, only one pilot study has investigated the relationship between the menstrual
cycle and energy availability in an active population (108). Ihalainen et al. found no
evidence of a difference in either energy intake, or energy availability, between the
early-follicular, mid-follicular, ovulatory, or mid-luteal phases of the cycle in
recreationally active females (n=15) when cycle phases were confirmed by serum
oestrogen and progesterone (108). Although this is only one study, free-living
exercising females present a unique population: dietary patterns to meet the needs of
exercise may not always correspond to hunger or desirability of food, and exercise can
have variable influences on appetite (27, 28). Exercise-induced energy deficits often fail
to produce a compensatory appetite response (253) and several studies have shown
that individuals are unable to sufficiently adjust ad-libitum energy intake in response to
increased EEE over periods of two (254), three (255), and seven days (256). Recently,
Taylor et al. demonstrated that elite male cyclists were unable to match energy intake to
changes in training load (164). In females, Howe et al. showed that running induced a
negative energy balance, and elicited a greater response in satiety hormones than did
walking, suggesting differential appetite responses to exercise intensity (257).
Additionally, Larson-Meyer et al. showed that in an endurance-trained female
population, single bouts of both moderate- and high-intensity exercise transiently
suppressed appetite and stimulated satiety hormones (258).
Other practical factors may also influence an athlete’s ability to match energy intake to
the energy requirements of exercise such as training-induced fatigue reducing
motivation to prepare food, limited time for eating around training (90), and the sheer
volume of food required, especially with large intakes of high-fibre and low-energy-
density foods (259, 260). Combined with the high prevalence of LEA and menstrual
disturbances in exercising females (29-31), these factors limit the applicability of
research conducted in the general population and warrant further investigation into
energy intake and availability across the menstrual cycle in female exercisers.
47
2.4.1.8 Inter-Individual Variation
Inter-individual variation is important to consider, both in terms of the wide range of
mean effects reported, as well as the practical application of menstrual cycle research.
Whilst Ihalainen et al. observed no evidence of a difference in neither mean energy
intake, nor energy availability across the cycle, the authors noted large inter-individual
differences in their data (108). Gil et al. showed that plasma concentration of
progesterone in the luteal phase was positively correlated with energy intake (r=0.38,
P=0.036) (173), highlighting not only the importance of measuring serum sex
hormones, but the variability in these hormones than can occur even within normal
ranges, and the subsequent influence on dietary intake. As the field of menstrual cycle
research grows, the individual variability in responses to hormonal fluctuations is
becoming clearer (15, 19, 108, 234). Yet this may be lost when studies report mean
effects, with the range of individual responses, or potential “non-responders” being
overlooked. Indeed, some females may be more likely to experience fluctuations in
energy intake across the cycle than others, for example, those that concomitantly
experience premenstrual syndrome (PMS) or premenstrual dysphoric disorder (PMDD)
(246, 261). The infrequent reporting of variance measures of mean differences only
obscures this inter-individual variability.
2.4.1.9 Mechanisms
Changes in the controls of energy intake across the menstrual cycle, such as subjective
hunger (22, 24), food cravings (24, 245 262), transit time (243), and appetite hormones
(249, 263-266), which mirror the reported changes in energy intake have also been
observed. Eating behaviour and appetite regulation are closely linked to the functioning
of the hypothalamic-pituitary-gonadal axis, and oestrogen receptors in the
hypothalamus interact with gastrointestinal peptides (e.g., cholecystokinin (CKK) and
ghrelin), neurotransmitters, and adipocytes (267). Generally, it appears that oestrogen
inhibits appetite, whereas progesterone has a stimulatory effect, at least when provided
in pharmacological ranges (249, 265). The association between oestrogen levels and
energy intake across the menstrual cycle has been documented in rodent models since
the 1920s (263): oestrogen mediates the release of satiating CCK from the small
48
intestine (264), which decreases meal size (rather than frequency of eating (266)) and
attenuates the release of appetite-stimulating hormone, ghrelin. In humans, Lyons et. al
reported hunger to be, on average, lower during the “fertile window” than on other days
during the same cycle (P=0.012) (22). These results may explain the particular decrease
in energy intake in the late-follicular and ovulatory phases (where oestrogen levels are
elevated, while progesterone remains low) observed by several studies (20, 22, 245,
247).
In rats, progesterone slows intestinal transit (268) and inhibits gastric emptying (269),
but whether the same effects are present in humans across the menstrual cycle is
unclear. Brennan et al. found that in response to a preload glucose drink, the rate of
gastric emptying was lower during two follicular phase timepoints, as well as hunger
scores and energy intake from a buffet-style meal (243), supporting/in accordance with
the general trend for an increase energy intake in the luteal phase. However, Jung et al.
reported a longer colonic transit time in the luteal phase in 11 females (coinciding with
high progesterone levels), compared with the follicular phase in 10 different females
(270), while Degen et al. found no evidence of a difference between phases in a group of
12 females (271). These conflicting results could reflect differences in study design,
including the dietary composition of test meals, as well as control for physical activity,
smoking history, body composition (270, 271). Furthermore, neither Jung et al. nor
Degen et al. verified menstrual cycle phases with serum hormones (270, 271). Slower
transit time during the luteal phase would be expected to increase satiety and decrease
energy intake (in contrast to the phase-related trend for energy intake), but this
highlights the fact that several factors likely influence dietary intake across the
menstrual cycle.
Indeed, it is not only the volume of food ingested that influences energy intake, but
energy density. McNeil et al. found that females displayed greater explicit wanting for
high-fat foods during the mid-luteal phase compared with the late-follicular phase,
manifesting as an increase in fat intake (from a single meal) (245). While there was no
evidence of a difference in explicit liking for high fat foods, Robinson et al. suggest
wanting as opposed to liking may play a more important role in influencing ingestive
behaviour (272). Cohen et al. showed that cravings for pastries, fried snacks, desserts
49
and sweets, sandwiches and hot dogs, sausages, chocolate, and “brigadiero” (a typical
Brazilian dessert made of chocolate and condensed milk) were greater in the luteal
phase than the follicular phase (262). Similarly, Gorczyca et al. reported higher craving
scores for chocolate, sweets, salty flavour, and other cravings (as well as overall
appetite) during the late-luteal phase compared with all other phases (24).
Interestingly, this was not reflected in any changes in energy or macronutrient intakes,
however, the 24-hour recalls were collected in the mid-luteal phase, while the cravings
questionnaires were administered in the late-luteal phase, so the true relationship may
not have been captured (24). Furthermore, the translation of cravings to dietary intake
is complex and influenced by many factors including access, affordability, and cultural
perceptions of diet and body image (273).
2.4.1.10 Summary
Methodological differences and poor compliance with menstrual research guidelines
make confidently interpreting the existing literature difficult. Despite this, energy intake
does appear to be greater in the luteal phase of the menstrual cycle compared with the
follicular phase in the general population, with the lowest intake likely occurring during
the late-follicular and ovulatory phases (see Figure 2) (however the number of studies
that have specifically researched these phases is limited: see Appendices A and B). Given
the amount of methodological heterogeneity, the exact magnitude of change across the
cycle is still unclear. While most studies reported an increase of approximately
200-350 kcal.day-1 from timepoints in the follicular to the luteal phase, for those that
observed an increase when cycle phases were hormonally verified, the actual range was
159-529 kcal.day-1 (Appendix A). However, phase-related differences in energy intake
most likely vary both between individuals, and from cycle to cycle. This lack of
generalizability means an individualised approach to interventions may be more
appropriate where dietary intake is concerned. Furthermore, there is a dearth of
research in active individuals, and the complex interplay between exercise, appetite,
dietary intake, and other factors unique to athletes means that patterns of energy intake
in the general population may not reflect those of exercising females.
50
Figure 2 Schematic diagram of the hypothesised changes in dietary energy intake
across an idealised 28-day menstrual cycle with ovulation occurring on day 14 (a),
and the corresponding relative oestrogen, progesterone and luteinising hormone
fluctuations (b). Both (a) and (b) are superimposed over the menstrual cycle
phases representing four distinct hormonal environments: the early-follicular
phase, the late-follicular phase, the ovulatory phase, and the mid-luteal phase1,2
1In (a) the dash-dotted pink line represents dietary energy intake, and in (b) the
solid orange line represents oestrogen, the dashed blue line represents
progesterone, and the dotted green line represents luteinising hormone
2Figure was created with Adobe Photoshop 2020 (Adobe Inc., San Jose, United States)
2.4.2 Macronutrient Intake
While the data on energy intake across the cycle in sedentary populations appear to be
reasonably consistent (at least in direction of effect, and despite the methodological
issues), the data on macronutrient intakes are mixed. Appendices A and B summarise
the studies that have investigated macronutrient intakes in different phases of the
menstrual cycle. Of course, the same methodological critiques discussed previously
apply, a fact that, again, may be partly responsible for the conflicting results.
51
2.4.2.1 Macronutrient Intake in the Absence of Differences in Energy Intake
Firstly, the few studies that found no evidence of a difference in energy intake across the
menstrual cycle rarely observed meaningful changes in macronutrient intakes between
phases either: de Souza et al. found no evidence of a difference in carbohydrate, fat, or
protein intakes (g·day-1) in a group of 27 Brazilian university students (23), and
Ihalainen et al. reported similar macronutrient intakes (g·day-1) in the early-follicular,
mid-follicular, ovulatory, and luteal phases in recreationally active females (108).
In the absence of any significant differences in energy intake across the cycle, Gorczyca
et al. did however observe a small increase in protein intake in the mid-luteal phase
compared with the peri-ovulatory phase when adjusted for energy-intake (65 ± 1.0 vs.
61 ± 0.9 g·day-1), representing a ~1% increase in TEI (16.4 ± 0.3 vs. 15.4 ± 0.2 %) (24).
Whilst statistically significant (P=0.04), and certainly interesting given the large sample
size and favourable methodology (8), this discrepancy in protein intake is minor and
hardly provides compelling evidence for a practically meaningful difference in
macronutrient intake (fat and carbohydrate intakes remained consistent across phases)
(24). On the other hand, Fong & Kretsch et al. observed a slightly more substantial
difference in phase-related carbohydrate intake despite consistent energy intakes: total
carbohydrate intake during menses was 303 ± 59 g·day-1 compared with
267 ± 81 g·day-1 in the peri-ovulatory phase, whilst fat and protein remained unchanged
across the cycle (244). But overall, when energy intake is consistent across the
menstrual cycle, macronutrient intakes appear to be consistent between phases
(Appendices A and B).
2.4.2.2 Macronutrient Intake in the Presence of Differences in Energy Intake
However, for those studies that did observe differences in energy intake between
phases, evidence for changes in macronutrient intakes across the menstrual cycle
becomes more apparent. Phase-related differences in energy intake by very principle
necessitate a change in the absolute intake (g·day-1) of at least one macronutrient; the
question therefore being, whether or not carbohydrate, protein, and fat intakes
fluctuate proportional to energy intake from phase to phase. Johnson et al. (who
52
observed an increase in energy intake from the ovulatory to the luteal phase) found
carbohydrate intake to be greater in the luteal (227.4 ± 60.8 g·day-1) and perimenstrual
phases (229.8 ± 60.9 g·day-1) compared with the ovulatory phase (200.1 ± 45.5 g·day-1),
and fat intake to be greater in the luteal versus ovulatory phase (79.4 ± 23.4 vs.
70.2 ± 23.3, g·day-1) (20). However, there was no evidence of a difference in %TEI for
carbohydrate between phases, but %TEI from fat increased in the luteal phase
compared with the ovulatory phase (37.4 ± 5.1 vs. 35.4 ± 5.5 %) (20).
Whilst %TEI is not often used to describe the dietary patterns of athletes (rather, g·kg-
1·day-1) (61, 90), studies in the general population tend to report macronutrient intake
this way, and it does offer a lens through which to compare macronutrient distribution,
especially in the presence of differences in energy intake. Martini et al. and Elliot et al.
also found that absolute intakes of carbohydrate, fat, and protein were all greater in the
luteal phase compared to the follicular phase, but %TEI for all three remained the same,
suggesting a proportional increase in macronutrient intake between phases (26, 175).
Similarly, Barr et al. found no evidence of a difference in dietary macronutrient
composition for either ovulatory or anovulatory cycles (238). Lyons et al., who observed
a lower energy intake during the ovulatory versus post-ovulatory phase, found that this
was due to lower protein (63 ± 4 vs. 73 ± 2 g·day-1) and carbohydrate (235 ± 14 vs. 270
± 12 g·day-1) intakes (22). Carbohydrate intake in the ovulatory phase was also lower
than in the pre-menses phase (263 ± 13 g·day-1), but none of these changes translated
into significant differences in %TEI for any macronutrient across the five phases
measured (22).
However, many studies did observe differences in the contribution of each
macronutrient to changes in energy intake between phases (21, 22, 25, 172-174, 240,
245). Dalvit-McPhillips found that the ~500 kcal·day-1 increase in energy intake from
the follicular to the luteal phase came almost exclusively from carbohydrates: absolute
intakes of protein and fat remained similar, whilst carbohydrate intake increased from
133.3 ± 28.1 g·day-1 in the follicular phase, to 256.9 ± 51.7 g·day-1 in the luteal phase
(21). On the contrary, Tarasuk & Beaton found that the increase in energy intake from
the follicular to luteal phase was primarily due to changes in fat (72.9 vs. 79.8 g·day-1;
53
41.3 vs. 39.5 g·1000kcal-1) (25), while Gil et al. and Li et al. observed increases in both
fat (26 and 10 g·day-1 increases respectively) and carbohydrate intake (67 and
29 g·day-1 increases respectively), while protein intakes remained consistent between
the follicular and luteal phases for all three studies (25, 172, 173). Similar to the trend
observed with energy intake, Gil et al. also found that plasma progesterone
concentration positively correlated with carbohydrate intake in the luteal phase
(r=0.38, P=0.035) (173).
McNeil et al. also found differences in fat intake across the cycle based on a single meal,
with intakes in the follicular, late-follicular, and mid-luteal phases of 21.88 ± 18.16,
18.73 ± 14.25 and 30.94 ± 21.58 g respectively (245). Reimer et al. observed
carbohydrate intake to remain consistent between the follicular and luteal phases, and
that moderate increases in protein (60.7 ± 5.3 vs. 80.6 ± 7.3 g·day-1) and fat (56.9 ± 5.2
vs. 81.6 ± 8.9 g·day-1) accounted for the increase in energy intake (240). Consequently,
%TEI from carbohydrate was lower in the luteal phase (53.2 ± 2.2 vs. 48.1 ± 2.6 %),
%TEI from fat was greater (28.8 ± 2.1 vs. 34.4 ± 2.5 %), whilst %TEI from protein was
unchanged (14.4 ± 1.6 vs. 15.3 ± 0.8 %) (240).
Differences in macronutrient intakes, proportional or otherwise, may have implications
for achieving sports nutrition guidelines across the menstrual cycle. Put into the relative
terms used in an athletic context (61), the difference in mean carbohydrate intake
observed by Dalvit-McPhillips between the follicular and luteal phases represents an
increase from ~2.0 to ~4.0 g·kg-1·day-1 (based on an estimated mean weight of 65kg),
bringing carbohydrate intake up to within the recommended range for light activity
(61). Similarly, the increases in total carbohydrate intake observed by Gil et al., Li et al.,
and Barr et al. represent increases from 4.0 to 5.1 g·kg-1·day-1 (173), 3.8 to 4.4 g·kg-
1·day-1 (172), and 4.5 to 5.1 g·kg-1·day-1 (238) respectively (calculated based on reported
mean body weight), which, though small, are not entirely dismissible. Of course, such
extrapolations from sedentary populations to an athletic context are hardly appropriate
but could suggest that carbohydrate guidelines may be easier to meet in the luteal
phase. This is interesting given the fact that achieving sufficient carbohydrate intake
may be particularly important during the follicular phase, in terms of glycogen storage
(76), and potential effects on performance (15).
54
Although studies rarely reported disproportionate changes in protein intake across the
menstrual cycle, as protein intake is closely related to TEI in athletes (109), the
potentially greater protein requirements in the luteal phase (19) may be satisfied
through a subconscious increase in energy intake anyway. While %TEI from protein
was unchanged, owing to an increase in energy intake, Reimer et al. reported
differences in actual protein intake equating to ~0.9 and ~1.2 g·kg-1·day-1 in the
follicular and luteal phases respectively (240). Although the study population wasn’t
specifically athletic, in terms of sports nutrition guidelines, this represents a shift from a
suboptimal intake in the follicular phase, to meeting the minimum recommendation in
the luteal phase (61). However, the difference in mean protein intake reported by Gil et
al. in the follicular and luteal phases of 1.3 ± 0.3 and 1.4 ± 0.3 g·kg-1·day-1 respectively,
were not significantly different from each other (173). Differences in actual protein
intake between phases in several other studies (where changes in energy intake were
present) were similarly minor when expressed in relative terms, typically equating to a
mere ~0.1 g·kg-1·day-1 increase in the luteal phase (when calculated based on absolute
intakes and reported weight) (20, 22, 26, 172, 174, 238-240, 244). While potential
differences in the protein requirements of athletes between menstrual cycle phases
have not been quantified (128), whether such changes in intake are meaningful is
debatable, both statistically, and in terms of translating data from the general
population to exercising populations.
Predominantly sedentary populations likely have different dietary practices and
patterns of macronutrient intake from those of active individuals. For example, daily
mean protein intakes in many of the aforementioned studies were quite low
(~0.8-1.2 g·kg-1·day-1) (20, 22, 26, 174, 238-240) compared to those observed in female
athletes (see Table 4). Active individuals concerned with maximising physical
performance may be more intentional with their nutrition and macronutrient intakes
during and around training, and consequently, potential hormonal influences on dietary
intake may be less pronounced. However, given the scarcity of research in active
populations, this remains speculative.
55
2.4.2.3 Methodological Considerations
Overall, the research on macronutrient intakes across the cycle is fairly mixed, but many
of these studies are methodologically flawed. Again, Chung et al. highlight the potential
discrepancy between correctly and incorrectly identified cycle phases (or irregular
cycles): carbohydrate accounted for approximately 5% greater %TEI in the follicular
versus luteal phase, and 4% greater %TEI in the follicular versus ovulatory phase, but
for those with correctly identified cycle phases (n=22), there was no longer evidence of
a difference in %TEI from carbohydrate across the menstrual cycle (even in the
presence of a difference in energy intake) (174). Protein intake was, however, slightly
greater in the luteal phase compared to both other phases (57 vs. 55 and 49 g·day-1), but
this difference is relatively trivial (174).
2.4.2.4 Mechanisms
Bearing all of this in mind, one might cautiously observe the fact that carbohydrate and
fat intakes appear more likely to fluctuate to account for differences in energy intake
across the menstrual cycle, while protein intake may be less vulnerable to change, at
least in the general population. This may be a consequence of firstly, the fact that
protein tends to contribute the least amount to energy intake of the three
macronutrients: similar differences in %TEI for protein yield smaller absolute amounts
in g·day-1, so differences may be less likely to appear nominally significant (at the same
threshold), and larger sample sizes may be required to detect differences. Secondly, and
perhaps more likely, is the fact that typical snack foods tend to be largely carbohydrate-
and fat-based, and may be what people reach for to satisfy increases in appetite, outside
of meals of consistent size and composition. Another factor that may be at play is pre-
menstrual cravings: whether entirely innate or culturally perpetuated, cravings for
“junk” foods, which tend to be high in both fat and simple carbohydrate in the luteal
phase of the menstrual cycle (famously, chocolate), is a familiar sentiment. As discussed
in section 2.4.1.9, these cravings may manifest as an increase in fat or carbohydrate
intake in the luteal phase.
56
On the other hand, differences in macronutrient intake may not be anything beyond a
simple increase in appetite and a consequent desire to satiate that with the most
convenient, cost-effective, or readily-available options. There may not be any specific
effect of progesterone or oestrogen (or any physiological factor relating to the
menstrual cycle), that influences carbohydrate, fat, or protein intake - no biological
drive for a particular nutrient during a particular phase - rather just a desire for energy
that happens to be fulfilled by certain macronutrients as a result of existing dietary
habits.
2.4.2.5 Summary
A consensus is lacking on the existence, let alone the nature, of any meaningful
differences in macronutrient intakes across the menstrual cycle. It appears that in the
absence of any differences in energy intake, differences in macronutrient intakes are
non-existent, or at the very least trivial. For those studies that did show an increase in
energy intake in the luteal phase, some showed proportional increases across all three
macronutrients, others were more heavily weighted towards carbohydrate and fat,
whilst a small handful found a disproportionate increase in protein. Whilst research in
athletic populations is lacking, methodological differences between studies and inter-
individual variability may be contributing to this conflict. However, if differences in
absolute macronutrient intakes across the menstrual cycle are present, this may have
implications for achieving sports nutrition guidelines in different phases.
57
2.5 Conclusion
In the general population, energy intake appears to be greater in the luteal phase of the
menstrual cycle compared with the follicular phase, with a particular decrease in the
days leading up to and including ovulation (although studies investigating the late-
follicular and ovulatory phases are limited) (18). The consensus on macronutrient
intake across the cycle is less clear (see Appendices A and B). However, much of the
literature is methodologically flawed, and to date, no study investigating dietary intake
across the menstrual cycle has adhered to the recently proposed guidelines for
menstrual cycle research (18). Furthermore, existing research has almost exclusively
focused on inactive populations (18).
Energy availability is an important concept in sports nutrition and is considered a more
appropriate index for quantifying nutritional status in exercising individuals (29).
Cross-sectional studies investigating the dietary intake of female athletes rarely take
into account the menstrual cycle (see Table 4), and to date, only one study has
investigated energy availability and macronutrient intakes across the menstrual cycle in
recreational athletes (108). Whether energy availability needs to be measured during
different phases of the menstrual cycle to obtain an accurate picture of nutritional
status, and the implications of any phase-related differences in dietary intake, have
received little attention. Understanding the habitual dietary practices of female
exercisers across the cycle may help identify phases where additional nutritional
support is required, and appropriately tailor dietary strategies to maximise health and
performance. A better understanding of the multi-directional relationships between the
menstrual cycle, exercise, and dietary intake is required to better design future studies.
Therefore, this pilot study aims to provide data on: 1) the variability of habitual energy
availability, energy intake, and macronutrient intake, across the menstrual cycle in
female exercisers, 2) phase-related dietary intake in the post-exercise period, and 3)
retention rates to aid future research.
58
3. Methods
3.1 Ethical Approval and Informed Consent
The University of Otago Ethics Committee (Health) granted ethical approval for the
present study on the 17th of January 2022 (reference number: H21/175) before
participant recruitment began.
Potential participants were informed of the study procedures and any associated risks
in writing (Appendix C) and were given the opportunity to ask questions before
completing a screening questionnaire to assess eligibility (Appendix D). Participants
were made aware that they could withdraw from the study at any time without any
disadvantage to themselves. All participants provided informed written consent
(Appendix E) before beginning the study.
3.2 Participants
This study was conducted at the University of Otago Human Nutrition Clinic in Dunedin,
New Zealand. Recruitment began in February 2022 via word of mouth, social media, and
posters displayed at a local gym and swimming pool, and around University of Otago
campus (Appendix F), and is ongoing until a sample size of 12 is achieved. Data
collection for the six participants included in this thesis occurred from February to
October 2022.
Individuals were eligible for participation in the study if they:
• Were of the female sex
• Were aged 18-45 years
• Had a regular menstrual cycle (defined as a cycle length of 21-35 days, and at
least nine menstrual periods per year)
• Were not using hormonal contraception, or had used hormonal contraception
within the past three months
59
• Reported to be healthy (as assessed by the screening questionnaire, see
Appendix D)
• Exercised at or above the current recommendations for New Zealand adults, i.e.,
completed at least 2.5 hours of moderate-intensity (3.0-5.9 METs), or 1.25 hours
of vigorous-intensity (>6.0 METs) exercise per week (69), and were at least Tier
1 (recreationally active individuals who achieve physical activity guidelines, and
may participate in a variety of activities and/or sports without a specific
commitment of focus on competition) in the participant training status and
performance calibre classification framework proposed by McKay et al. (274)
The exclusion criteria were anyone who:
• Had menstrual irregularities including: amenorrhoea (defined as the cessation of
menses for three consecutive months), oligomenorrhoea (defined as irregular
periods, and/or a cycle length >35 days, with fewer than nine menstrual cycles in
the last year), or anovulation (defined as no positive urinary ovulation test after
18 days of testing beginning on day eight of the cycle, for at least two cycles)
• Was pregnant or lactating
• Was currently using, or had used within the past three months, any oestrogen or
progesterone medication, including hormonal contraception
• Had a food allergy or intolerance
• Was on a weight control diet
• Had a blood disorder
• Had a diagnosed endocrine disorder such as polycystic ovarian syndrome or
endometriosis
3.3 Study Design
This was an observational study that followed participants across four menstrual cycles.
Figure 3 gives a general overview of the study protocol. Briefly, the purpose of cycles
one and two was to confirm the presence of a regular menstrual cycle, with cycle two
also acting as a familiarisation cycle. Data collection occurred during cycles three and
four, in which participants kept three-day weighed diet records and training diaries
60
during the early-follicular and mid-luteal phases of the menstrual cycle. On one day
during each three-day recording period, participants also came into the clinic to have
their body composition measured, provide a urine and blood sample, complete a fasted
exercise session, consume a post-exercise ad-libitum meal, and complete a well-being
questionnaire at various time points. In total there were five data collection periods:
one during cycle two in the mid-luteal phase, and two each during cycles three and four
in the early-follicular and mid-luteal phases.
The true aim of the study was undisclosed to participants to minimise the influence of
intentional manipulation of dietary intake in different phases of the menstrual cycle:
participants were instead informed that the aim was to investigate the effect of
menstrual cycle phase on hydration status and sweat losses during exercise.
61
Figure 3 Schematic diagram of the general study protocol1
1Figure was created with Microsoft PowerPoint 2016 (Microsoft Corporation, Redmond, United States)
62
3.3.1 Cycle One
When completing the screening questionnaire, participants had the opportunity to
provide the date of their last menstrual period, which, if known, was used to represent
“day one” of cycle one. If participants could not recall the date of this, they informed
researchers of the first day of their next menstrual period.
3.3.2 Cycle Two
Participants informed researchers of the first day of their menstrual period,
representing “day one” of cycle two, at which point they were provided with kitchen
scales (Salter AquatronicTM Electronic Kitchen Scale Model 1023, Salter Housewares,
Tonbridge, United Kingdom), urinary ovulation tests (Clearblue® Digital Ovulation Test,
Swiss Precision Diagnostics, Geneva, Switzerland), written instructions on how to
complete the weighed diet record, training diary, and accelerometer wear time diary
(Appendix G), an electronic invite to join a private group on the MealLogger-Photo Food
Journal® mobile application (version 4.7.4, Wellness Foundry, Helsinki, Finland), and
instructions on how to download and use the application (Appendix H). On day eight of
their menstrual cycle, participants began taking an ovulation test with a first void urine
sample every day until a positive result was achieved, indicated by a static smiley face. A
positive result signified the surge in urinary LH characteristic of ovulation, and the tests
were able to detect a minimum concentration of 40 mIU·ml-1 (275). Participants were
asked to upload a photo of the positive result to MealLogger® for researchers to visually
confirm. If no positive result was detected after 18 days of testing (the maximum
interval to ensure three days of data collection in the mid-luteal phase for a 35-day
cycle), the process was repeated in the following cycle, and if this didn’t yield a positive
result, the participant was excluded from the study.
Six days following a positive ovulation test result, participants wore an accelerometer
(Actigraph wGT3X-BT, Actigraph, Pensacola, United States) on their non-dominant wrist
before going to bed to monitor 24-hour activity patterns. This was worn for the next
three full days (seven, eight, and nine days after the positive test), and removed upon
waking 10 days after the positive test. Participants were asked to take the
63
accelerometer off when showering, bathing, swimming, or during periods of very heavy
sweating, recording any times the accelerometer was removed on MealLogger®, as well
as the time they attempted to fall asleep each night it was worn. On the seventh, eighth,
and ninth days following the positive ovulation test, participants kept a weighed image-
assisted diet record and training diary via MealLogger® to familiarise themselves with
the process and the level of detail required for cycles three and four. On the middle day
of the three-day recording period, participants came into the clinic for a familiarisation
trial.
3.3.2.1 Familiarisation Trial
All procedures outlined in the exercise session (section 3.3.3.1) were completed during
the familiarisation trial, except no urine, blood, or expired air samples were collected
(participants did wear the respiration mask for one five-minute period to become
familiar with the sensation), nor was the post-exercise ad-libitum meal provided.
Instead, participants were offered muesli bars and Up & Go drinks to take away (to
minimise time in the clinic as a COVID-19 precaution, when physical distancing and
mask mandates were in place). Height was measured in duplicate to the nearest 0.1cm
using a free-standing stadiometer (Marsden HM-250P Leicester Portable Height
Measure, Marsden Weighing, Rotherham, United Kingdom) during the familiarisation
session only. The stationary bike (BikeERG, Concept2, Vermont, United States) was set
up to the participants’ desired measurements and recorded for future trials. The
stationary bike had 10 levels of resistance and participants were free to adjust this
throughout the 60 minutes until they found a suitable level, which was maintained
during subsequent trials. Participants had the choice of flat pedals or Shimano SPD-SL
Road pedals with their own cycling shoes: this remained consistent for all trials.
3.3.3 Cycles Three and Four
Based on menstrual history, the accelerometers were initialised to begin several days
before the expected start date of the participant’s next menstrual period. As soon as
participants reported their next menstrual period (denoting day one of cycle three),
they put on the accelerometer that evening before going to sleep. The accelerometer
64
was worn for the next three days (days two, three, and four of the menstrual cycle), and
removed on the fifth day and collected by researchers. A weighed image-assisted diet
record, training diary and accelerometer wear time diary (including attempted sleep
time) were recorded via MealLogger® on days two, three, and four of the menstrual
cycle, representing the early-follicular phase. On the middle day of the three-day
recording period, participants came into the clinic for a fasted exercise session, followed
by an ad-libitum breakfast (see section 3.3.3.1). If it was not possible to run the clinic
session on the middle of the three days (i.e., day three for the early-follicular phase, or
eight days after the positive test for the luteal phase) due to participant commitments,
or multiple sessions falling on the same day, occasionally the clinic session was run on
an alternative day within the three-day recording period.
On day eight of their menstrual cycle, participants repeated the ovulation test protocol
from cycle two. If no positive result was detected after 18 days of testing, cycle three
was repeated in a following cycle, and if this didn’t yield a positive result, the participant
was excluded from the study. The same protocol as described for the early-follicular
phase was repeated in the mid-luteal phase, with the three-day recording period taking
place on the seventh, eighth, and ninth days after the positive result, the clinic session
on the middle of these three days, and the accelerometer worn from the evening of the
sixth day, until the morning of the tenth day after the positive result.
3.3.3.1 Exercise Session
Participants remained fasted from 10:00 pm the evening prior to the exercise session.
Before arriving at the clinic, they collected a first void urine sample and completed a
COVID-19 screening questionnaire (Appendix I): if any questions were answered ‘yes’,
data collection for that phase was repeated in a following cycle. Figure 4 gives a general
overview of the exercise session protocol. Upon arriving at the clinic, participants
completed a questionnaire assessing subjective well-being, muscle soreness, thirst,
hunger, and gastrointestinal symptoms (“well-being questionnaire”, Appendix J). Body
composition was measured via BIA (Tanita Body Composition Analyzer Model BC-418,
Tanita Corporation, Tokyo, Japan), accounting for 0.5kg of clothing. To assess sweat
rate, weight was measured in exercise clothing (Wedderburn Scales, Dunedin, New
65
Zealand) to the nearest 0.01kg, and water bottle weight was measured with kitchen
scales (Salter AquatronicTM Electronic Kitchen Scale Model 1023, Salter Housewares,
Tonbridge, United Kingdom) to the nearest 1g, before and after the 60-minute cycle. The
forearm, thigh, chest, and shoulder on the right-hand side of the body were cleaned with
de-ionised water and dried before sweat patches (TegadermTM +Pad, 3M, Saint Paul,
United States) were applied to each site, which remained in place throughout the
exercise session. If the participant was wearing long pants, a sweat patch was not
applied to the thigh. Baseline measurements of capillary blood glucose and lactate were
taken before the participant began exercising: the finger was cleaned with an alcohol
swab, then punctured with a lancet (Unistik® 3 Normal, Owen Mumford, Woodstock,
United Kingdom). Blood glucose was measured using a FreeStyle Optium glucose
monitor (Abbot Diabetes Care Inc., California, United States), and blood lactate with a
Stat Strip XpressTM lactate monitor (Nova Biomedical, Waltham, United States) from the
same site. Any excess blood was wiped off with a gauze swab (Multisorb® non-woven
swab, BSN Medical, Hamburg, Germany) before a plaster (Leukoplast® soft white, Essity
Medical, Stockholm, Sweden) was applied.
A wet heart rate strap (Garmin HRM1G Heart Rate Monitor, Garmin Ltd., Olathe, United
States) was fitted around the chest and paired with a watch (Garmin Forerunner 110,
Garmin Ltd., Olathe, United States). The participant then sat at rest on the stationary
bike while a respiration mask (Cortex Medical, Leipzig, Germany) was placed over the
mouth and nose, and a five-minute expired air sample was collected via a Cortex
MetaLyzer 3B (Cortex Medical, Leipzig, Germany), and recorded using MetaSoft® Studio
software (Version 5.1.3, Cortex Medical, Leipzig, Germany). During this five-minute
period, room temperature and relative humidity were measured via a Kestrel® 3000
Pocket Weather Meter (Nielsen-Kellerman Company, Boothwyn, United States), and
heart rate was recorded. After the mask was removed, participants were given five
minutes to warm up on the bike at a self-selected effort. The timer on the bike was then
reset, and participants began a 60-minute time trial at the resistance selected during the
familiarisation session.
66
During the time trial participants were allowed to drink from their pre-weighed bottle
as desired and permitted to listen to music if they wanted. Elapsed time, distance, and
power were not visible to the participant during the trial. For the last five minutes of
every 15-minute block, the respiration mask was placed over the mouth and nose, and
VO2, VCO2, and RER were measured from the five-minute expired air sample.
Participants were then asked to rest their hand face-up on the handlebars, their finger
was cleaned, and blood glucose and blood lactate were measured according to the
protocol detailed previously. Room temperature and relative humidity were also
measured with a Kestrel® 3000 Pocket Weather Meter during the five-minute period,
and heart rate was recorded immediately before the mask was removed at the
completion of the five-minute block. After the mask was removed, participants were
asked to indicate their RPE on the Borg scale (276) presented to them (Appendix K).
Distance and average power were recorded at 15, 30, 45, and 60 minutes.
Upon completing the time trial, and following reweighing of the participant and their
bottle, the well-being questionnaire was administered within five minutes. Participants
were given a hand warmer to hold whilst completing the questionnaire and the sweat
patches were removed using tweezers and placed in sterile sealed tubes. A 1.0mL blood
sample was collected via finger prick using a BD Microtainer® Contact-Activated Lancet
(Becton, Dickinson and Company; Franklin Lakes, United States) into a 1.5mL
Eppendorf Safe-Lock tube (Eppendorf, Hamburg, Germany) containing EDTA and
immediately centrifuged at 3000 Relative Centrifugal Force for five minutes (MiniSpin®
plus, Eppendorf, Hamburg, Germany). The plasma was pipetted into a separate 1.5mL
Eppendorf Safe-Lock tube and stored at -70°C in the freezer room of the Human
Nutrition department for later analysis, along with the urine sample and sweat patches.
Participants were given the opportunity to shower at the clinic before being left alone in
the kitchen to consume an ad-libitum post-exercise breakfast meal (see Appendix L for
the list of foods). Participants were informed that they were free to consume any food
or drink on offer in their own time and that everything had been pre-weighed and
would be recorded by researchers. When the participant had finished their meal, they
completed the well-being questionnaire once more. Participants were provided with
any equipment needed for the following phase (urine collection containers, ovulation
67
tests, etc.) before leaving. Participants continued with their diet record and training
diary for the remainder of the day but were not required to record the clinic exercise
session or ad-libitum meal in their entries.
After completing cycle three, participants repeated the cycle three procedures in a
fourth cycle. Cycles one through four were not necessarily consecutive, nor were the
follicular and luteal phase measurements always taken during the same cycle. Upon
their completion in the study, participants completed an exit questionnaire (Appendix
M) to gather information on personal perceptions of dietary and exercise changes
across the menstrual cycle, burdensome aspects of the study, and how participation and
retention rates might be improved for the future. The screening, COVID-19, well-being,
and exit questionnaires were all administered online using Research Electronic Data
Capture (REDCap) software (Version 11.0.3, Research Electronic Data Capture,
Nashville, United States).
68
Figure 4 Schematic diagram of the exercise session protocol1
1Figure was created with Microsoft PowerPoint 2016 (Microsoft Corporation, Redmond, United States)
69
3.3.4 Dietary Assessment
Participants kept a series of three-day weighed image-assisted diet records in the mid-
luteal phase of cycle two for familiarisation purposes (beginning seven days after a
positive ovulation test result), and in the early-follicular (beginning on day two of the
menstrual cycle) and mid-luteal phases of cycles three and four. Participants were
provided with kitchen scales (Salter AquatronicTM Electronic Kitchen Scale Model 1023,
Salter Housewares, Tonbridge, United Kingdom) and written instructions on how to
keep a diet record using MealLogger® (Appendices G and H), and were instructed to
maintain their typical diet throughout the study. Given the strict menstrual cycle phase
definitions (8), there was no intentional effort to capture an equal ratio of weekdays to
weekend days in each phase, however, the number of weekends and weekdays in the
early-follicular and mid-luteal phases was reported.
Diet records were kept via the MealLogger-Photo Food Journal® mobile application
(version 4.7.4, Wellness Foundry, Helsinki, Finland) and each participant was invited to
join a private group, which researchers could access. Alongside a detailed description of
the type, brand, amount, and cooking method of food or beverages consumed,
participants were also instructed to upload an image of the eating occasion and/or the
nutrition information panel of packaged food items, especially where weighing
ingredients was not possible or practical (for example, eating out at a restaurant).
Researchers could comment on the uploaded posts in real time to obtain further
information or clarify details.
For the post-exercise clinic meal during cycles three and four, researchers weighed each
food item before and after the participant consumed their meal to calculate intake.
Participants were not required to upload this meal to MealLogger®, but it was included
in the three-day diet record by researchers, as well as being reported as an individual
outcome. The diet records were entered into the nutrient analysis software, FoodWorks
10 (Xyris Software, Brisbane, Australia) using the 2018 New Zealand FOODfilesTM
database (Version 01) (277), and analysed for energy and macronutrient intake.
70
3.3.5 Energy Expenditure of Exercise
For the three consecutive days of data collection in the early-follicular and mid-luteal
phases, participants kept a training diary via MealLogger®. They were instructed to
provide as much information on the exercise session as possible (Appendix G), including
the type, duration, and intensity of exercise, distance covered (if applicable), and if
available, heart rate and/or power data. Researchers could comment on the uploaded
posts to obtain further information if required. Information from the training diaries
was converted to METs using the 2011 Adult Compendium of Physical Activities (219).
Participants were not required to include the clinic cycling time trial in their training
diaries, but this was included by researchers in the calculation of daily EEE and energy
availability. Exercise was defined as any physical activity 3.0 METs or greater.
3.3.6 Body Composition
Weight, FM, FFM, and body fat percentage were measured via BIA (Tanita Body
Composition Analyzer Model BC-418), accounting for 0.5kg of clothing and using the
‘Female Athlete’ setting, at the beginning of each clinic session. Participants stood on the
machine in bare feet in the clothing they planned to exercise in, and held the handles
away from their body until a reading was obtained. Weight, FM and FFM, were
measured to the nearest 0.1kg, and percentage body fat to the nearest 0.1%.
Participants had fasted from 10:00 pm the evening before, but exercise in the previous
24 hours and hydration status were not standardised.
3.3.7 Energy Availability Calculation
Equation 1 shows the calculation for energy availability (expressed in
kcal·kgFFM-1·day-1), where TEI (kcal·day-1) and EEE (kcal·day-1) represent the mean
values obtained from the three-day recording periods, and FFM (kg) represents the
value obtained from the clinic session via BIA. In accordance with the original definition
of energy availability proposed by Loucks et al. in 1998 (36), the habitual waking energy
expenditure that would have occurred during the exercise period was subtracted from
the total energy cost of the session. The Cunningham equation (224) was used to
71
calculate 24-hour RMR (using the FFM value obtained from the clinic session), divided
by 24 to give an hourly rate, and then multiplied by daily exercise duration (in hours).
This value was then subtracted from the EEE calculated from the training logs and METs
to give the true energy cost of daily exercise – “adjusted EEE” – above that of resting
energy expenditure.
Energy Availability Equation (36):
𝐸𝑛𝑒𝑟𝑔𝑦 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑖𝑙𝑖𝑡𝑦 = 𝑇𝐸𝐼 − (𝐸𝐸𝐸 −𝑅𝑀𝑅
24 𝑥 𝑒𝑥𝑒𝑟𝑐𝑖𝑠𝑒 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛)
𝐹𝐹𝑀
Cunningham Equation (224):
𝑅𝑀𝑅 = 500 +22 × 𝐹𝐹𝑀(𝑘𝑔)
3.3.8 Study Size
This was part of a pilot study with an intended sample size of 12, to investigate
recruitment, retention, and feasibility of conducting research of this nature, as well as to
provide information on the variability of outcome measures to inform sample size
calculations for future research. However, given time constraints, only data on daily
dietary intake, energy availability, and post-exercise ad-libitum dietary intake from six
participants were included in this thesis. Other data, and the oestrogen and
progesterone concentrations from the blood samples were not available in time for
inclusion, or were beyond the scope of this thesis.
72
3.3.9 Remuneration
Participants received a $10 grocery voucher for completing cycle one, an additional $20
voucher for completing cycle two, and additional $60 vouchers each for completing
cycles three and four to cover expenses associated with participation in the study.
3.3.10 Statistical Methods
Participant age was defined as age at recruitment, and ethnicity and training volume
were self-reported from the screening questionnaire. Where more than one ethnicity
was reported, Māori and Pasifika ethnicities were prioritised over other ethnicities and
New Zealand European was prioritised over any ethnicity other than Māori and Pasifika.
Training status (according to the classification framework of McKay et al. (274)) was
determined from the screening questionnaire, training diaries, and conversation with
participants. Descriptive anthropometric baseline characteristics (height and weight)
were measured during the luteal phase familiarisation session of cycle two. Continuous
variables were reported as the mean ± SD, whilst categorical variables were reported as
counts (%).
Daily energy intake, absolute and relative macronutrient intakes, adjusted EEE, and
energy availability in the early-follicular and mid-luteal phases were presented as the
mean ± SD (range) of cycles three and four. Energy and macronutrient intake (absolute
and relative) from the post-exercise clinic meal in each phase were also presented as
the mean ± SD (range) of cycles three and four, while body composition data (weight,
FM, FFM and percentage body fat) were presented as the mean ± SD of cycles three and
four. Due to the small sample size, there was insufficient power to undertake statistical
tests to assess the significance of any differences between phases. For variables where
there appeared to be a potentially meaningful difference between phases, the mean
differences between cycles three and four for the early-follicular and mid-luteal phases
were compared with the mean differences between each phase for cycles three and four
combined. The proportions of those meeting current sports nutrition guidelines (61) for
daily relative carbohydrate and protein intakes, energy availability, and post-exercise
relative carbohydrate and protein intakes in each phase were reported as counts (%).
73
For carbohydrate, these took into account daily exercise intensity and duration from
training logs and were reported as counts (%) of individual daily records meeting
guidelines in each. For protein intake and energy availability, counts (%) were based on
three-day mean values.
For participants with one day of obviously incomplete diet records per phase, the
available two-day diet record and training diary were used to determine daily dietary
intake, adjusted EEE, and energy availability for that particular phase: participants with
more than one day of incomplete diet records per phase were excluded from the
analysis, however, their ad-libitum post-exercise clinic meal data were still included.
The frequencies of weekend days included in the three-day recording periods in the
early-follicular and mid-luteal phases were presented as counts (%). All analyses were
performed using Stata Statistical Software (StataCorp, release 17, College Station,
United States).
Individual data for daily energy and macronutrient intakes, adjusted EEE, body
composition, energy availability, and ad-libitum post-exercise energy and macronutrient
intakes were also presented as scatterplots for cycles three, cycle four, and the mean of
cycles three and four, to visualise individual differences between phases and cycles.
Scatterplots were constructed using Microsoft Excel 2016 (Microsoft Corporation,
Redmond, United States) using the templates for creating univariate scatterplots for
paired data from Weissgerber et al. (278), and then colour-coded by participant.
74
4. Results
4.1 Participants
Figure 5 shows participant recruitment and retention throughout the study. A total of
26 individuals expressed interest in the study, and 19 completed the screening
questionnaire. All 16 participants who were eligible for participation after screening
completed cycle one. Three participants were excluded before the completion of cycle
two due to menstrual irregularities (n=2) or starting hormonal contraception (n=1). Of
the 13 participants who completed cycle two, five were excluded before the completion
of cycle three due to menstrual irregularities. One participant who had an anovulatory
cycle during cycle three declined to continue participating due to the time burden of
having to repeat the cycle, whilst another participant dropped out after the completion
of cycle three for mental health reasons. Of the seven participants who began cycle four,
data collection was incomplete for one participant in time for inclusion in this thesis,
therefore full data were available for six participants (see Appendix N for a timeline of
movement throughout the study for the six included participants).
75
Figure 5 STROBE flow diagram of participant recruitment and follow-up1
1Figure was created with Microsoft Word 2016 (Microsoft Corporation, Redmond,
United States)
Abbreviations: STROBE strengthening the reporting of observational studies in
epidemiology
76
Table 5 shows the demographic and anthropometric data of participants. The study
population was relatively young, with a mean age of 24.5 ± 2.5 years. Mean height and
weight were 165.9 ± 5.1 cm and 67.5 ± 8.4 kg respectively (as measured in the luteal
phase of cycle two). Most participants (n=4) were classified as Tier 1 (recreationally
active) according to the training status and performance calibre classification
framework of McKay et al. (274).
Table 5 Participant characteristics1
Participants
(n=6)
Age (years)
24.5 ± 2.5
Height (cm)
165.9 ± 5.1
Weight (kg)
67.5 ± 8.4
Ethnicity2
New Zealand European
Māori
British
Indian
3 (50)
1 (16.7)
1 (16.7)
1 (16.7)
Training volume (hours·week-1)
5.6 ± 1.4
Training status2, 3
Tier 1
Tier 2
4 (66.7)
2 (33.3)
1Values represent mean ± SD unless otherwise stated
2Values represent count (%)
3Tier 1: recreationally active, Tier 2: trained/developmental (based on the participant
training status and performance calibre classification framework of McKay et al. (274))
4.2 Daily Dietary Intake, Exercise Energy Expenditure, Energy Availability, and
Body Composition
Table 6 shows mean daily energy and macronutrient intakes, adjusted EEE, and energy
availability in the early-follicular and mid-luteal phases. One participant had an
incomplete diet record for the luteal phase of cycle four, so a two-day record was used
instead of a three-day record. Mean energy intake was very similar between phases,
however, mean adjusted EEE was lower in the mid-luteal phase compared with the
early-follicular phase, resulting in greater energy availability in the mid-luteal phase.
Energy availability increased by an average of 3.1 kcal·kgFFM-1·day-1 in the mid-luteal
77
phase, which exceeded the mean difference between the early-follicular phase of cycles
three and four (1 kcal·kgFFM-1·day-1) and the mid-luteal phase of cycles three and four
(0.7 kcal·kgFFM-1·day-1). Using 30 kcal·kgFFM-1·day-1 as the threshold for LEA (153),
66.7% (n=4) of participants were classified as being in a state of LEA in the early-
follicular phase, compared with 16.7% (n=1) in the mid-luteal phase.
Figure 6 shows energy intake, EEE, and energy availability of individual participants in
the early-follicular and mid-luteal phases of each cycle, and for the mean of both cycles.
Adjusted EEE displayed inter-individual variability in the direction of effect across the
menstrual cycle: three participants showed a decrease in adjusted EEE in the mid-luteal
phase for both cycles, one participant showed a small increase, and one participant had
a similar level between the early-follicular and mid-luteal phases of both cycles.
However, energy availability did not appear to show the same level of inter-individual
variation, with five participants showing an increase in the mid-luteal phase compared
with the early-follicular phase, based on the mean of both cycles.
While the mean difference in adjusted EEE between phases of 140 kcal·day-1 exceeded
the magnitude of difference between cycles three and four for the early-follicular phase
(16.6 kcal·day-1) and the mid-luteal phase (26.7 kcal·day-1), one participant had a
dramatic decrease in adjusted EEE of ~600kcal·day-1 in the mid-luteal phase of both
cycles (see Figure 6). When this participant was excluded, the difference in mean
adjusted EEE between the early-follicular phase (258.0 ± 160.0 kcal·day-1) and the mid-
luteal phase (212.8 ± 117.1 kcal·day-1) was far smaller. However, mean daily energy
intake became 1767.7 ± 539.4 kcal·day-1 in the early-follicular phase and
1870.0 ± 341.7 kcal·day-1 in the mid-luteal phase, such that energy availability in each
phase remained similar to that of the full dataset: 31.5 ± 12.5 kcal·kgFFM-1·day-1 in the
early-follicular phase and 34.7 ± 7.5 kcal·kgFFM-1·day-1 in the mid-luteal phase.
78
Table 6 Daily energy and macronutrient intake, exercise energy expenditure and
energy availability from three-day weighed image-assisted diet records and
training diaries in the early-follicular and the mid-luteal phases across two
menstrual cycles1
Early-follicular phase
(n=6)
Mid-luteal phase
(n=6)
Energy (kcal·day-1)
2007.9 ± 760.8
(1260.8 – 3208.7)
2019.3 ± 476.7
(1429.1 – 2766.2)
Carbohydrate
g·day-1
g·kg-1·day-1
%TEI
233.8 ± 104.0
(128.2 – 401.5)
3.4 ± 1.2
(1.9 – 4.9)
46.3 ± 6.6
(37.3 – 53.6)
224.5 ± 61.6
(141.9 – 305.8)
3.3 ± 0.8
(2.2 – 4.1)
45.0 ± 4.1
(40.7 – 51.0)
Protein
g·day-1
g·kg-1·day-1
%TEI
83.0 ± 29.1
(57.3 – 122.4)
1.2 ± 0.4
(0.9 – 2.0)
17.4 ± 3.6
(12.8 – 23.7)
84.2 ± 27.0
(58.9 – 126.6)
1.2 ± 0.3
(0.9 – 1.7)
17.0 ± 3.8
(12.7 – 22.8)
Fat g·day-1
g·kg-1·day-1
%TEI
73.4 ± 26.7
(45.0 – 115.0)
1.1 ± 0.3
(0.7 – 1.5)
32.8 ± 6.0
(26.4 – 44.3)
80.4 ± 16.7
(61.8 – 108.3)
1.2 ± 0.2
(0.9 – 1.5)
35.3 ± 2.9
(32.1 – 40.5)
Adjusted EEE2 (kcal·day-1)
397.1 ± 369.5
(55.5 – 1092.6)
257.1 ± 150.8
(56.1 – 478.7)
Energy availability (kcal·kgFFM-1·day-1)
32.6 ± 11.5
(19.2 – 52.2)
35.7 ± 7.1
(26.4 – 46.1)
Weekend days 3
9 (25.0)
9 (25.7)
1Values represent mean ± SD (range) of cycles three and four
2Adjusted EEE represents EEE minus the habitual waking energy expenditure that
would have occurred during the exercise period (see section 3.3.8)
3Values represent count (%)
Abbreviations: EEE exercise energy expenditure
79
Figure 6 Scatterplots showing energy intake, exercise energy expenditure, and
energy availability of participants in the early-follicular and mid-luteal phases of a)
cycle three, b) cycle four, and c) the mean of cycles three and four1,2
1Each colour represents a different participant
2Scatterplots were created with Microsoft Excel 2016 (Microsoft Corporation, Redmond,
United States)
Abbreviations: EEE exercise energy expenditure, FFM fat-free mass
80
Mean protein and carbohydrate intakes were similar between the early-follicular and
mid-luteal phases, whilst there was a small increase in fat intake from
73.4 ± 26.7 g·day-1 in the early-follicular phase, to 80.4 ± 16.7 g·day-1 in the mid-luteal
phase. This mean difference was similar in magnitude to the difference between the
early-follicular phase of cycles three and four (5.6 g·day-1), and less than the mean
difference between the mid-luteal phase of cycles three and four (12.4 g·day-1). Figure 7
shows the macronutrient intakes of individual participants in the early-follicular and
mid-luteal phases of each cycle, and for the mean of both cycles. Participants rarely
showed a consistent direction of effect between phases for carbohydrate, fat, or protein
intakes across both cycles.
Mean protein intakes in both phases were just within the lower bounds of the
recommended range (61), with 33.3% (n=2) meeting the guidelines in the early-
follicular phase, and 50% (n=3) meeting the guidelines in the mid-luteal phase. Mean
carbohydrate intake was at the lower end of the range suggested for “light training
volume: low-intensity or skill-based activities” (61) in both phases. As many
participants undertook daily exercise in excess of this, only 27.8% (n=10) of diet
records met the guidelines for daily carbohydrate intake in the early-follicular phase,
and 31.4% (n=11) in the mid-luteal phase. The same number of weekend days was
represented in both phases.
Macronutrient intakes were similar between the early-follicular and mid-luteal phases
in the reduced sample (n=5), with no apparent difference in mean protein intake
(75.1 ± 24.4 vs. 75.7 ± 19.2 g·day-1) and a negligible difference in carbohydrate intake
(200.3 ± 71.3 vs. 208.2 ± 52.6 g·day-1). Mean fat intake was slightly greater in the mid-
luteal phase (74.8 ± 10.8 g·day-1) compared with the early-follicular phase
(65.1 ± 19.2 g·day-1), but again, was similar to the mean differences between phases of
cycles three and four.
81
Figure 7 Scatterplots showing macronutrient intake of participants in the early-
follicular and mid-luteal phases of a) cycle three, b) cycle four, and c) the mean of
cycles three and four1,2
1Each colour represents a different participant
2Scatterplots were created with Microsoft Excel 2016 (Microsoft Corporation, Redmond,
United States)
82
Table 7 shows the mean body composition values of participants in the early-follicular
and mid-luteal phases, whilst Figures 8 and 9 show this data for individual participants
for each cycle, and for the mean of both cycles. Weight, FM, FFM, and body fat
percentage were very similar in the early-follicular and mid-luteal phases.
Table 7 Body composition measurements obtained via bioelectrical impedance
analysis during the early-follicular and mid-luteal phases across two menstrual
cycles1
Early-follicular phase
(n=6)
Mid-luteal phase
(n=6)
Weight (kg)
67.6 ± 8.1
67.7 ± 8.3
Fat mass (kg)
17.9 ± 5.0
18.2 ± 4.7
Fat-free mass (kg)
49.8 ± 4.1
49.4 ± 4.7
Body fat (%)
26.1 ± 4.5
26.7 ± 4.2
1Values represent mean ± SD of cycles three and four
Figure 8 Scatterplots showing participant weight in the early-follicular and mid-
luteal phases of a) cycle three, b) cycle four, and c) the mean of cycles three and
four1,2
1Each colour represents a different participant
2Scatterplots were created with Microsoft Excel 2016 (Microsoft Corporation, Redmond,
United States)
83
Figure 9 Scatterplots showing participant body composition in the early-follicular
and mid-luteal phases of a) cycle three, b) cycle four, and c) the mean of cycles
three and four1,2
1Each colour represents a different participant
2Scatterplots were created with Microsoft Excel 2016 (Microsoft Corporation, Redmond,
United States)
84
4.3 Post-Exercise Ad-Libitum Meal Intake
Table 8 shows mean energy and macronutrient intakes from the post-exercise ad-
libitum meal, while Figure 10 shows the energy intakes of individual participants in the
early-follicular and mid-luteal phases of each cycle, and for the mean of both cycles.
Energy intake was slightly greater in the mid-luteal phase compared with the early-
follicular phase, with 66.7% (n=4) of participants consuming more energy in the mid-
luteal phase. The magnitude of this mean difference (53.9 kcal) was in between the
mean inter-cycle difference for the early-follicular phase (10.7 kcal) and the mid-luteal
phase (73.1 kcal).
Table 8 Energy and macronutrient intake from an ad-libitum post-exercise meal in
the early-follicular and the mid-luteal phases across two menstrual cycles1
Early-follicular phase
(n=6)
Mid-luteal phase
(n=6)
Energy (kcal)
659.4 ± 305.8
(231.9 – 991.6)
713.3 ± 335.1
(343.1 – 1118.2)
Carbohydrate
g
g·kg-1
89.2 ± 39.6
(35.1 – 132.3)
1.3 ± 0.6
(0.5 – 2.2)
93.1 ± 46.9
(38.9 – 161.6)
1.4 ± 0.8
(0.6 – 2.7)
Protein
g
g·kg-1
23.3 ± 10.0
(10.9 – 33.0)
0.3 ± 0.2
(0.2 – 0.5)
22.2 ± 10.5
(10.3 – 37.7)
0.3 ± 0.1
(0.2 – 0.5)
Fat g
g·kg-1
20.7 ± 12.0
(4.0 – 36.9)
0.3 ± 0.1
(0.1 – 0.5)
25.5 ± 12.5
(11.0 – 44.1)
0.4 ± 0.2
(0.2 – 0.6)
1Values represent mean ± SD (range) of cycles three and four
85
Figure 10 Scatterplots showing energy intake of participants from a post-exercise
ad-libitum meal in the early-follicular and mid-luteal phases of a) cycle three, b)
cycle four, and c) the mean of cycles three and four1,2
1Each colour represents a different participant
2Scatterplots were created with Microsoft Excel 2016 (Microsoft Corporation, Redmond,
United States)
Figure 11 shows the post-exercise macronutrient intakes of individual participants in
the early-follicular and mid-luteal phases of each cycle, and for the mean of both cycles.
Carbohydrate and protein intakes were fairly similar between phases, but fat intake
increased from the early-follicular phase to the mid-luteal phase for 83.3% (n=5) of
participants, a mean difference of 4.8g, or an increase of 23%. This mean difference
between phases well exceeded the mean difference between cycles for the early-
follicular phase (0.4g) and the mid-luteal phase (0.2g). Whilst there was no obvious
pattern to the foods participants selected in each phase, the higher fat intake may partly
be due to a greater tendency to select high-fat spreads such as peanut butter, Nutella
and table spread in the mid-luteal phase, as opposed to low-fat options like jam and
honey, which were consumed more frequently in the early-follicular phase. The
consumption of greater quantities of milk, yoghurt, and toasted muesli may have also
contributed to the increase in fat intake in the mid-luteal phase, as well as the more
frequent consumption of chocolate chips and seeded bread.
86
While mean carbohydrate and protein intakes satisfied post-exercise guidelines in both
phases (in terms of absolute amount) (61), 66.7% (n=4) of participants met
carbohydrate guidelines in the early-follicular phase, compared with 50% (n=3) in the
mid-luteal phase, while the same four participants (66.7%) met protein guidelines in
both phases.
87
Figure 11 Scatterplots showing macronutrient intake of participants from a post-
exercise ad-libitum meal in the early-follicular and mid-luteal phases of a) cycle
three, b) cycle four, and c) the mean of cycles three and four1,2
1Each colour represents a different participant
2Scatterplots were created with Microsoft Excel 2016 (Microsoft Corporation, Redmond,
United States)
88
5. Discussion
This pilot study suggests that in regularly menstruating female exercisers, daily mean
energy and macronutrient intakes were similar between the early-follicular and mid-
luteal phases, while mean energy availability was slightly greater in the mid-luteal
phase, owing to an increase in mean adjusted EEE. However, one participant had a
dramatic ~600 kcal·day-1 decrease in adjusted EEE in the mid-luteal phase of both
cycles. After excluding this participant, adjusted EEE became similar between phases,
whilst energy intake was slightly greater in the mid-luteal phase, but such that the trend
for energy availability between phases remained very similar to that of the full dataset.
Mean energy, carbohydrate, and protein intakes from an ad-libitum post-exercise meal
were similar between phases, but mean fat intake was greater in the mid-luteal phase
compared with the early-follicular phase. The rate of menstrual irregularities and
subsequent participant exclusion was incredibly high, raising questions about the
feasibility of conducting large-scale research in eumenorrheic exercising females
according to gold standard guidelines (8), at least in the present location.
5.1 Daily Energy Intake, Exercise Energy Expenditure, and Energy Availability
Energy availability increased from 32.6 ± 11.5 kcal·kgFFM-1·day-1 in the early-follicular
phase to 35.7 ± 7.1 kcal·kgFFM-1·day-1 in the mid-luteal phase, an increase greater than
the mean difference between phases of different cycles. The clinical significance of this
is difficult to ascertain with such a small sample size, but could potentially have
implications for how energy availability is measured in research and the field. While the
very presence of a regular menstrual cycle with normal hormonal fluctuations is an
important indicator of adequate energy availability, optimising energy availability
above the bare minimum required for eumenorrhoea is another matter: performance
and health impairments appear to differ in the nature of their functional decline across
the continuum of energy availability (32).
89
Of course, many limitations exist with translating energy availability thresholds from
short-term controlled laboratory studies to free-living individuals (see section 2.3.4).
Nonetheless, using the traditional 30 kcal·kgFFM-1·day-1 threshold for LEA (153), 66.7%
(n=4) of participants would be classed as being in a state of LEA in the early-follicular
phase, compared with only 16.7% (n=1) in the mid-luteal phase. A difference in habitual
energy availability between menstrual cycle phases may be yet another variable to
consider when quantifying energy availability, and the historical failure to take this into
account may contribute to the inconclusive thresholds from observational research and
the wide range of prevalence estimates for LEA (29). Furthermore, while serum
oestrogen and progesterone were not analysed for this thesis, all participants presented
with regular menstrual cycles and positive urinary ovulation tests, yet only one
participant had a daily energy availability of 45 kcal·kgFFM-1·day-1 or greater (in either
phase) - the threshold reportedly associated with adequate energy availability (148).
Although this is not surprising, given that previous cross-sectional studies have failed to
find a consistent correlation between energy availability and menstrual function (157-
159), the likely presence of under-reporting (107), and the myriad of other factors with
the potential to influence physiological outcomes of LEA (see section 2.3.4).
On one hand, the early-follicular phase could be a target period for additional
nutritional support, yet on the other hand, seemingly “adequate” energy availability in
the mid-luteal phase may only be “adequate” by follicular phase standards. Previous
laboratory studies on energy availability typically standardised interventions to the
follicular phase (36, 149, 150), but it would not be unreasonable to question whether
the thresholds derived from such research are equally applicable to the luteal phase. A
2020 systematic review and meta-analysis of 26 studies found a small effect favouring
an increase in RMR during the luteal phase compared with the follicular phase
(ES: 0.33; 95% CI: 0.17, 0.49; P<0.001) (279), posing the question of whether “the
dietary energy required to sustain physiological functioning after subtracting the
energetic cost of exercise” (155) may be greater in the luteal phase. Unfortunately, the
present study did not measure RMR due to the standardisation protocols required and
the additional participant burden.
90
Many female exercisers perceive the menstrual cycle to affect their training (59, 250-
252) and athletes tend to identify the mid- to late-follicular, or ovulatory phases as the
most preferred time to exercise or compete (251). Both the early-follicular and late-
luteal phases have been associated with negative perceptual responses to train and
compete, relating to menstrual symptoms and decreased vigour in the early-follicular
phase, and mood disturbances and pre-menstrual symptoms in the late-luteal phase
(280). However, the present study specifically investigated the early-follicular and mid-
luteal phases (although for some participants this ended up being very close to the
beginning of their next menstrual period based on the timing of their positive ovulation
test) and did not include the late-follicular or ovulatory phases. Individual experiences
of the menstrual cycle probably also play a role in perceptions of exercise: those with
higher pain scores report their training to be more affected by the menstrual cycle
(251). A survey of 1250 sportswomen found that a considerable proportion of low- and
high-level female athletes adapted their training volume (37.9% and 26.4%
respectively) and intensity (48.6% and 30.8% respectively) during the menstrual phase,
with 47.3% of low-level athletes, and 16.3% of high-level athletes stopping training
entirely (250). This disparity between low-level and high-level athletes is not surprising
given that high-level athletes may have more rigid training schedules.
Another potential influencing factor for participation in exercise across the menstrual
cycle is the recent swell of media commentary on training in alignment with cycle
phase, which typically advocates for high-intensity and strength training in the follicular
phase, and aerobic and skill-based training in the luteal phase (12-14). In reality, high-
quality research to support such general recommendations is incredibly limited, and
current messaging from the scientific community is based on an individualised
approach to training (15). But pre-existing interests or conceptions regarding the
menstrual cycle and exercise may have influenced participants’ decision to partake in
the present study.
Participants did display inter-individual variation in the direction of effect for adjusted
EEE across the menstrual cycle: three participants showed a decrease in the mid-luteal
phase for both cycles, one participant showed a small increase, whilst another
participant maintained the same level of exercise between phases across both cycles
91
(Figure 6). This variation could be due to individual differences in menstrual cycle
symptoms and hormonal fluctuations. Although mean adjusted EEE was greater in the
early-follicular phase compared with the mid-luteal phase in the present study – a trend
perhaps in the opposite direction to expected – this was largely driven by a dramatic
~600 kcal·day-1 decrease in the mid-luteal phase by one participant in both cycles.
While this may demonstrate the inter-individual variation in menstrual cycle effects, it
more likely relates to the greater number of weekend days recorded during the early-
follicular phase than the mid-luteal phase for that participant. Weekends generally
allow greater time for exercise compared with the typical working week, particularly
for recreational athletes, and are often when longer endurance training or racing takes
place. Even though there was an equal number of weekend days between phases in the
sample overall, this particular participant had the largest training volume of the study
population, so this phenomenon may have been disproportionately reflected in the
mean values.
When the data from this participant were excluded, adjusted EEE was far more similar
between the early-follicular (258.0 ± 160.0 kcal·day-1) and mid-luteal phase
(212.8 ± 117.1 kcal·day-1). However, this was likely due to the inter-individual
variability present, which is perhaps unsurprising as the majority of participants were
Tier 1 ‘recreationally active’ (274): should voluntary exercise levels fluctuate as a result
of the menstrual cycle, we would expect this to be more evident in recreational-level
athletes compared with elite-level athletes, due to presumably greater training
flexibility amongst recreational athletes. While the present study only measured the
early-follicular and mid-luteal phases, Ihalainen et al. found no evidence of a difference
in EEE between the early-follicular, late-follicular, ovulatory, and mid-luteal phases in
15 recreationally active eumenorrheic females (although their inclusion criteria were
slightly stricter than the present study: recreational athletes were defined as strength
training and endurance training three times per week each), but they too reported large
inter-individual variability in their data (108).
Yet in the present study, despite the inter-individual variation in patterns of adjusted
EEE between phases, and the lack of a notable difference in adjusted EEE between
phases in the reduced sample, a very similar trend for mean energy availability
92
compared with the full sample was still observed (31.5 ± 12.5 kcal·kgFFM-1·day-1 in the
early-follicular phase, and 34.7 ± 7.5 kcal·kgFFM-1·day-1 in the mid-luteal phase), as
mean energy intake was slightly greater in the mid-luteal phase compared with the
early-follicular phase (1870.0 ± 341.7 vs. 1767.7 ± 539.4 kcal·day-1). The inter-
individual variation in phase-related adjusted EEE patterns did not appear to carry
through to energy availability: two participants consistently showed an increase in
energy availability from the early-follicular phase to the mid-luteal phase across both
cycles, and whilst the remaining four varied in their direction of effect by cycle, based on
the mean of both cycles, five out of six showed an increase in the mid-luteal phase
(Figure 6). Regardless of the direction or magnitude of phase-related changes in
adjusted EEE, on average, participants appeared to adjust their energy intake
accordingly to achieve a similar pattern of energy availability. Of course, one could
argue that ~50-100 kcal·day-1 differences in energy intake and adjusted EEE are easily
within the bounds of measurement error. But the fact that the same trend, of a very
similar magnitude, was observed across both cycles, of both the full and reduced dataset
(albeit slightly different in the contribution of their drivers), warrants further
investigation. Greater mean energy availability in the mid-luteal phase compared with
the early-follicular phase is in line with what previous research in the general
population indicates based on dietary energy intake alone (see section 2.5.1).
While urinary LH tests reduce the possibility of including anovulatory cycles and ensure
appropriate luteal phase timing (8, 46), it would of course be remiss not to address the
absence of serum hormone analysis in the present study. However, the fact that such
trends were observed perhaps only strengthens the case, as the potential exclusion of
those with abnormal hormonal profiles would be expected to increase the magnitude of
the mean difference.
By contrast, Ihalainen et al. reported similar energy intake and energy availability
between the early-follicular, late-follicular, ovulatory, and mid-luteal phases (108).
However, based on mean values, energy availability was greater in the early-follicular
phase (40.0 ± 11.1 kcal·kgFFM-1·day-1) compared with the mid-luteal phase
(37.6 ± 7.2 kcal·kgFFM-1·day-1) (108), the opposite direction of effect to what trends in
the general population (see section 2.5.1), and the findings of the present study indicate.
93
While the differences in energy availability across the cycle were not statistically
significant (P=0.465), and phases were verified by serum hormones, participants were
only followed for one cycle, and given it was a pilot study, it may have been
underpowered to detect differences (108). Although Ihalainen et al. didn’t clarify the
nature of the inter-individual variation in their data, it may have contributed to the lack
of evidence for a difference in mean energy availability between phases (108): so-called
“responders” and “non-responders” to menstrual cycle influences may have cancelled
each other out. The present study had a very small sample size, so by chance, may have
captured a greater proportion of “responders” (of the same direction). Nonetheless,
inter-individual variation is common in menstrual cycle research, hence the advocacy
for individualised approaches based on personal experiences of the menstrual cycle
rather than generalised recommendations (15, 19, 108, 234).
5.2 Daily Macronutrient Intakes
Given the lack of a difference in energy intake between the early-follicular and mid-
luteal phases, the similar macronutrient intakes between phases are unsurprising:
previous research has shown that meaningful differences in macronutrient intakes tend
only to occur as a result of differences in energy intake (Appendices A and B). While
there was a slight increase in fat intake in the mid-luteal phase, which could potentially
be explained by “pre-menstrual cravings” satisfied through foods high in fat (245, 262),
this difference was within the bounds of variability displayed from cycle to cycle, so is
probably inconsequential. As participants rarely showed the same direction of effect for
carbohydrate, protein, and fat intakes between phases for both cycles, this likely speaks
to the true absence of phase-related differences in macronutrient intake and doesn’t
appear to indicate significant inter-individual differences in menstrual cycle effects
(Figure 7).
Considering the reduced sample of five participants, the ~100 kcal·day-1 increase in
mean energy intake from the early-follicular to the mid-luteal phase came from mean
increases in carbohydrate of ~8 g·day-1, and fat of ~10 g·day-1, whilst mean protein
intakes remained almost identical. Although the minor increase in energy intake may
not be meaningful, and changes in macronutrient intakes were comparable to mean
94
differences between cycles, the tendency for carbohydrate and fat to be responsible for
increases in energy intake across the menstrual cycle, rather than protein, aligns with
what several studies in the general population have shown (20, 25, 172, 173, 245).
Regardless of any minor differences in absolute macronutrient intakes, a considerable
proportion of participants failed to meet sports nutrition guidelines in either phase,
especially for carbohydrate. Under-reporting may have played a role (107), but given
the degree of inadequacy, may not be entirely responsible. While perhaps unlikely
based on the athletic calibre of the cohort, participants may have been intentionally
periodising carbohydrate intake, or more plausibly, attempting to adhere to lower-
carbohydrate diets due to the media attention these have received in recent years (90).
But as fat intakes were not extortionately high, and estimates of energy availability
were generally suboptimal (caveats aside (32)), inadequate carbohydrate and protein
intakes in this group of female exercisers were likely secondary to suboptimal energy
intakes. Reasons for this may include poor nutrition knowledge, limited food
availability, lifestyle factors and time constraints, cultural dietary practices,
gastrointestinal challenges or changes in appetite related to exercise and/or high-fibre
diets, or more intentional dietary restriction due to body image concerns or body
composition goals (27, 28, 90, 253, 254). It is unlikely that many of these factors would
differ by menstrual cycle phase, but evidence does exist for fluctuations in
gastrointestinal symptoms (281, 282) and perceptions of body image across the cycle
(283): whether or not these meaningfully alter dietary habits and fuelling strategies is
unclear. Data on gastrointestinal symptoms and appetite were collected via the well-
being questionnaire (Appendix J), however, are beyond the scope of this thesis. While
there has been some media commentary in recent years about fuelling according to
menstrual cycle phase (12, 14), any influence of this on dietary patterns was not evident
in the present study.
Similarly, Ihalainen et al. found no evidence of a difference in macronutrient intake
across the menstrual cycle (108). Even with slightly higher energy intakes compared to
the present study, mean absolute carbohydrate intakes were only marginally higher, at
~250 g·day-1, or ~3.7 g·kg-1·day-1 across phases (compared with 233.8 ± 104.0 and
224.5 ± 61.6 g·day-1 in the early-follicular and mid-luteal phases of the present study,
95
respectively), and were still suboptimal based on participants’ exercise levels (61, 108).
Mean fat and protein intakes were slightly greater as well, with protein intakes falling in
the mid-range of recommendations at ~1.6 g·kg-1·day-1 across phases (61, 108) (as
opposed to the mean intakes of 1.2 ± 0.4 and 1.2 ± 0.3 g·kg-1·day-1 in the early-follicular
and mid-luteal phases of the present study). While Ihalainen et al. reported slightly
greater energy intakes, as well as EEE, they also had a stricter definition of
“recreationally active” than the present study (108), so participants may have had
greater sports nutrition knowledge.
5.3 Post-Exercise Ad-Libitum Meal
Dietary intake from the post-exercise ad-libitum meal was fairly similar between
phases, with the most notable difference being an increase in mean fat intake from the
early-follicular to the mid-luteal phase of 4.8g (a difference that was well in excess of
the mean difference between cycles). Mean carbohydrate and protein intakes were in
line with post-exercise recommendations (for absolute amounts) in both phases, while a
similar number of participants met the guidelines in each phase as well (61). This
suggests that menstrual cycle phase may not influence the ability to achieve post-
exercise macronutrient targets following one hour of moderate-intensity fasted cycling
exercise (at least when comparing the early-follicular and mid-luteal phases).
Kamemoto et al. also investigated dietary intake from an ad-libitum meal following an
acute bout of fasted exercise (60 minutes of cycling at 70% of “HR reserve”) in the early-
follicular and mid-luteal phases of the menstrual cycle, in 10 physically active Japanese
females (284). Overall energy and macronutrient intakes were higher than those of the
present study (likely owing to the 90-minute interval between completing exercise and
consuming the ad-libitum meal), but similarly, there was no evidence of a difference in
energy, protein, or carbohydrate intakes between phases (284). However, in contrast to
the present study, fat intake was also similar between phases (284). This discrepancy
could be due to the different selection of foods provided: Kamemoto et al. offered a far
smaller variety of foods, many of which were provided in standardised portions that
may not have allowed participants to sufficiently exercise variation in dietary intake, for
example, individually portioned spreads, rice balls, and boiled eggs. Furthermore,
96
results may have been influenced by not only different dietary patterns, but different
cultural perceptions of the menstrual cycle.
5.4 Participant Retention
Of the 16 participants who enrolled in the study, only six (37.5%) completed the full
four cycles. The high rate of menstrual irregularities (43.8%) severely reduced the
study population, even before serum oestrogen and progesterone analyses, which could
lead to further exclusions. This is similar to the previously reported rates of menstrual
irregularities in exercising populations of ~30-50% (30, 31, 46). Future research needs
to take this into account when performing sample size calculations and should assume
an exclusion rate due to menstrual irregularities alone of at least 40-45%. However,
participants were generally highly motivated and eager to participate in research of this
kind. We hypothesise that the main challenge to recruitment came from the exclusion
criteria of hormonal contraception use, especially given the high rates of use in
exercising populations (59, 285), whilst the main challenge to retention was menstrual
irregularities.
Recruiting and retaining a sufficient sample of physically active eumenorrheic females
for a large-scale study in the present location, whilst strictly adhering to gold standard
protocols for menstrual cycle research (8), is challenging. Without greater collaboration
between research centres and increased funding for female-specific research, a
potentially wider scope may be required.
5.5 Strengths and Limitations
Tracking participants for two cycles before data collection helped ensure the presence
of a regular menstrual cycle, and using urinary ovulation tests reduced the likelihood of
including participants with anovulatory cycles (8). This is important as Barr et al. found
that the greater energy intake observed during the luteal phase compared with the
follicular phase was only present for ovulatory cycles (238). The ovulation tests also
allowed the mid-luteal phase to be defined relative to ovulation (as recommended by
Elliott-Sale et al. (8)), as opposed to an arbitrary number of days from the onset of
97
menstruation, which may not capture the mid-luteal phase for every female (44, 46).
Furthermore, data were collected over two menstrual cycles to reduce the influence of
inter-cycle variability, per recent guidelines (8). Elliot et al. highlight the importance of
this: while energy intake was greater in the follicular phase compared with the luteal
phase based on one cycle, when 13 of the 31 participants were followed for an
additional cycle, there was no longer evidence of a difference (175).
However, because serum oestrogen and progesterone analyses were unavailable in time
for the completion of this thesis, the possibility that participants with abnormal
hormonal profiles, and/or a luteal phase progesterone concentration of less than
16.00 nmol·L-1 were included, cannot be eliminated. Therefore, the differences in
energy availability and dietary intake as a result of fluctuations in endogenous sex
hormones may be underestimating true differences. The present study only collected
data during the early-follicular and mid-luteal phases of the menstrual cycle: the late-
follicular and ovulatory phases were not included, but present unique hormonal
environments (8). However, identifying the late-follicular phase would involve
prospectively collecting and immediately analysing blood samples in the days
approaching anticipated ovulation, which was not financially possible and places an
even greater burden on participants. Ovulatory phase measurement would have to take
place on the day of the first positive urinary ovulation test, which would have been
logistically challenging for both participants and researchers. Furthermore, quantifying
dietary patterns for the short-lived duration of the ovulatory phase (essentially a single
day of waking hours (8)) may not be that clinically relevant. The results of the present
study are therefore restricted to comparisons specifically between the early-follicular
and mid-luteal phases, rather than the historically broader definition of the ‘follicular’
and ‘luteal’ phases, or other time points within each phase. Furthermore, the present
study did not investigate dietary intakes or energy availability in hormonal
contraception users, due to the additional anticipated challenges of recruiting users of
the same brand and formulation of OCP.
While self-reported dietary intake will rarely be an entirely accurate reflection of
habitual intake (107, 167), using three-day weighed image-assisted diet records is an
important strength. Weighed records are the most precise method for estimating
98
dietary energy and nutrient intakes, and their prospective nature reduces recall error,
which may be particularly important if differences in intake across the menstrual cycle
are subconscious (167). Images uploaded to the MealLogger® application provided
researchers with valuable information, and reduced reliance on assumptions, especially
where weighing foods was not possible (for example, eating out at a restaurant). The
coincidental inclusion of an equal amount of weekend days in each phase is also
advantageous. Furthermore, the true aim of the study was undisclosed from
participants to minimise the influence of intentional manipulation of dietary intake in
different phases of the menstrual cycle (participants were instead informed that the aim
was to investigate the effect of menstrual cycle phase on hydration status and sweat
losses during exercise).
The quantification of energy expenditure is also a limitation: estimating EEE via MET
values is perhaps less precise than alternative methods (32, 156) and depends greatly
on the level of information the participant provides. However, for a population that
engaged in a range of physical activities and sports, MET values allowed the same
method to be used across all types of exercise, as Burke et al. recommend (32).
Unfortunately, accelerometer data were unavailable in time for the completion of this
thesis. Whilst doubly-labelled water provides the most accurate estimation of energy
expenditure in free-living individuals (191), this was not financially feasible, especially
given the lack of predictability of the menstrual cycle, and the possibility of anovulation.
Furthermore, doubly-labelled water doesn’t directly distinguish between total daily
energy expenditure and EEE (156). However, the present study did use the original
definition of EEE proposed by Loucks et al. in 1998 (36), using the Cunningham
equation which appears to be the best predictive RMR equation for female athletes
(221-223).
Body composition measurements were obtained via BIA, and while participants were
fasted, no attempt was made to control for hydration status or recent exercise activity.
However, as this study collected information on habitual exercise patterns, dietary
intake, and hydration status, imposing such rigid standardisation protocols would
partially eliminate any habitual fluctuations. Even though DEXA is often considered the
criterion standard for measuring body composition, it is still vulnerable to changes in
99
these variables (225). However, use in this study was restricted by access and
affordability, especially with the short notice that would have been given due to the
nature of the menstrual cycle.
Few studies have assessed dietary energy and macronutrient intake during the post-
exercise period in a controlled setting in exercising females in different phases of the
menstrual cycle. Participants each completed the same exercise protocol (five-minute
warm up and 60-minute cycling time trial) and were provided with a standardised
selection of breakfast foods of varying nutritional composition after each session. While
attempts were made to include a range of breakfast options typical of the New Zealand
diet, they may not have reflected the habitual dietary choices of participants following
exercise. It should also be noted that the exercise session was undertaken fasted, which
may not reflect the typical exercise practices of participants either. While it appears that
a considerable number of female athletes do partake in fasted training (at least in
endurance sports) (286), this is not in line with current sports nutrition guidelines (61).
Another important limitation is the small sample size. Only data from six participants
were available in time for the completion of this thesis, which makes identifying trends
difficult, and the data vulnerable to the influence of more extreme observations.
However, this pilot study aimed to provide data on the variability of outcome variables
to inform sample size calculations for future research on dietary intake and energy
availability across the menstrual cycle, and recruitment is ongoing.
Furthermore, the present study did not assess PMS symptoms, which may play a role in
the relationship between dietary intake and menstrual cycle phase. Premenstrual
Syndrome (PMS), and the more severe form, PMDD, are characterised by a range of
distressing physical, emotional, and behavioural symptoms during the luteal phase
(287), and may also influence dietary patterns. Reed et al. compared energy intake of a
lunch meal, self-selected from a menu, between those with PMDD and those without,
during the follicular and luteal phases (261). The PMDD group consumed 16% more
energy during the luteal phase, primarily driven by an increase in fat, whilst there was
no evidence of a difference in energy nor macronutrient intake between phases for the
control group (see Appendix A) (261). Similarly, in a study of overweight females with
100
and without PMS, Cross et al. showed that the increase in energy intake from the
follicular phase to the luteal phase was even more profound for those with PMS (605
kcal·day-1 mean difference) compared with those without (112 kcal·day-1 mean
difference) (246). They also observed a weak correlation between the severity of
symptoms and energy intake (246). Absolute intake of all three macronutrients
increased from the follicular phase to the luteal phase in those with PMS, but in such a
way that %TEI from carbohydrate increased, and %TEI from protein decreased, whilst
the control group experienced an increase in absolute fat intake in the luteal phase, but
no change in %TEI of any macronutrient between phases (see Appendix B) (246).
Conversely, Bryant et al. found no evidence of a difference in energy nor macronutrient
intakes across the menstrual cycle for neither those with PMS nor those without (see
Appendix B) (288).
But comparing these studies is difficult as Reed et al. investigated females with the more
severe condition of PMDD (261), whilst Cross et al. and Bryant et al. used different
diagnostic criteria for PMS (246, 288). This could potentially contribute to the
discrepancy present, especially if symptom severity is related to changes in energy
intake, as the findings of Cross et al. may suggest (246). Furthermore, Cross et al.
studied an overweight population (246), raising the question of whether body size
affects phase-related changes in energy intake, whilst neither Bryant et al. nor Cross et
al. confirmed cycle phase with serum hormones (246, 288). While the underlying causes
of PMS and PMDD are not well understood (287), any potential influences on dietary
intake may be of emotional or behavioural origin, rather than strictly endocrinological.
Nonetheless, the present study failed to address and investigate PMS and PMDD as
potential effect modifiers on the relationship between dietary intake and menstrual
cycle phase. As competitive sport may be associated with a greater incidence of PMS
(with greater training volume and a longer sporting career being additional risk
factors), this may be particularly relevant to exercising populations (289). Furthermore,
PMS symptoms may also influence exercise habits, as discussed in section 4.1.
101
Participants in the present study were classified as either Tier 1 ‘recreationally active’
(n=4), or Tier 2 ‘trained/developmental’ (n=2) in the classification framework of McKay
et al. (274). Caution must therefore be taken in extrapolating any trends to more highly
trained populations of Tier 3 or above, for whom patterns of nutritional intake and
training may be unique again from those of more recreational exercisers.
102
6. Conclusions and Future Research
This pilot study suggests that energy availability was greater in the mid-luteal phase of
the menstrual cycle, compared with the early-follicular phase in six female exercisers.
Dietary intake from an ad-libitum meal following moderate-intensity cycling exercise
was similar between the early-follicular and mid-luteal phases, apart from fat, which
was slightly greater in the mid-luteal phase. The menstrual cycle did not appear to
influence the ability to meet carbohydrate and protein guidelines for exercising females,
however daily intakes tended to be suboptimal in both phases, whilst post-exercise
intakes were generally more adequate. The ~45% exclusion rate due to menstrual
irregularities severely reduced the sample size, which presents challenges for
conducting large-scale research in eumenorrheic exercising females.
Given the small sample size and other limitations, it would be inappropriate to draw
major conclusions from this study, but the trends observed are worth further
exploration. Studies with larger sample sizes that adhere to best practice guidelines for
menstrual cycle research (8) and also include the late-follicular phase, if not the
ovulatory phase as well, would provide valuable further insight into the relationships
between energy availability, dietary intake, and the menstrual cycle in female
exercisers. However, without greater collaboration between research centres and
increased funding for female-specific research, the feasibility of doing so is certainly
challenging. Ideally, future research should also focus on investigating more diverse
populations, such as: more highly trained or elite athletes; different ethnic groups and
how cultural perceptions of the menstrual cycle may influence dietary and physical
activity habits; hormonal contraceptive users and the potential influence of different
phases, especially given high rates of use in exercising populations (59, 285); and those
with PMS and PMDD and whether these conditions act as effect modifiers on
relationships between the menstrual cycle, dietary intake, physical activity, and energy
availability.
103
Energy availability has become a critical concept in optimising the health and
performance of physically active individuals, but its intricacies are not yet fully
understood (29). Future research may need to consider the menstrual cycle when
developing a standardised protocol for measuring energy availability, investigating
rates of LEA, determining thresholds for metabolic and endocrine impairments, and
designing strategies to optimise energy availability, beyond conducting research
exclusively in the follicular phase. However, the additional financial, logistical, and
compliance challenges this imposes on researchers and participants must be
recognised.
104
7. References
1. Cowley ES, Olenick AA, McNulty KL, Ross EZ. “Invisible sportswomen”: the sex data
gap in sport and exercise science research. Women Sport Phys Act J
2021;29(2):146-51.
2. Melloni C, Berger JS, Wang TY, Gunes F, Stebbins A, Pieper KS, Dolor RJ, Douglas PS,
Mark DB, Newby LK. Representation of women in randomized clinical trials of
cardiovascular disease prevention. Circ Cardiovasc Qual Outcomes 2010;3(2):135-
42.
3. Costello JT, Bieuzen F, Bleakley CM. Where are all the female participants in sports
and exercise medicine research? Eur J Sport Sci 2014;14(8):847-51.
4. Kong BY, Haugh IM, Schlosser BJ, Getsios S, Paller AS. Mind the gap: sex bias in
basic skin research. J Invest Dermatol 2016;136(1):12-4.
5. Beery AK, Zucker I. Sex bias in neuroscience and biomedical research. Neurosci
Biobehav Rev 2011;35(3):565-72.
6. Lee SK. Sex as an important biological variable in biomedical research. BMB Rep
2018;51(4):167-73.
7. Owen Jr JA. Physiology of the menstrual cycle. Am J Clin Nutr 1975;28(4):333-8.
8. Elliott-Sale KJ, Minahan CL, de Jonge X, Ackerman KE, Sipilä S, Constantini NW,
Lebrun CM, Hackney AC. Methodological considerations for studies in sport and
exercise science with women as participants: a working guide for standards of
practice for research on women. Sports Med 2021;51(5):843-61.
9. The International Olympic Committee. IOC gender equality and inclusion report
2021. Version current 2021. Internet: https://olympics.com/ioc/gender-equality
(accessed 11 March 2022).
10. Women’s Sport Foundation. Chasing equity: the triumphs, challenges, and
opportunities in sports for girls and women. Version current January 2020.
Internet:
https://www.womenssportsfoundation.org/articles_and_report/chasing-equity-
the-triumphs-challenges-and-opportunities-in-sports-for-girls-and-women/
(accessed 11 March 2022).
11. Saner E. The menstrual month: how to exercise effectively at every stage of your
cycle. Version current 2 February 2021. Internet:
https://www.theguardian.com/lifeandstyle/2021/feb/02/the-menstrual-month-
how-to-exercise-effectively-at-every-stage-of-your-cycle (accessed 25 February
2022).
105
12. Whitehead J. Make the most of menstruation: when and how to exercise and eat to
feel your best at each stage of your cycle. Version current 3 October 2019. Internet:
https://www.independent.co.uk/life-style/women/menstrual-cycle-period-
syncing-best-exercise-what-to-eat-hormones-rest-a9121276.html (accessed 25
February 2022).
13. Landsverk G. Plan your workouts around your menstrual cycle to maximize
muscle gains, according to an expert. Version current 10 July 2021. Internet:
https://www.insider.com/how-to-exercise-around-your-menstrual-cycle-period-
2021-7 (accessed 25 February 2022).
14. Kitchen S. Nutrition, training and performance for the menstruating female athlete.
Version current 3 March 2021. Internet:
https://www.fleetfeet.com/blog/nutrition-training-and-performance-for-the-
menstruating-female-athlete (accessed 25 February 2022).
15. McNulty KL, Elliott-Sale KJ, Dolan E, Swinton PA, Ansdell P, Goodall S, Thomas K,
Hicks KM. The effects of menstrual cycle phase on exercise performance in
eumenorrheic women: a systematic review and meta-analysis. Sports Med
2020;50(10):1813-27.
16. Thompson B, Almarjawi A, Sculley D, Janse de Jonge X. The effect of the menstrual
cycle and oral contraceptives on acute responses and chronic adaptations to
resistance training: a systematic review of the literature. Sports Med
2020;50(1):171-85.
17. Helm MM, McGinnis GR, Basu A. Impact of nutrition-based interventions on
athletic performance during menstrual cycle phases: a review. Int J Environ Res
Public Health 2021;18(12):6294.
18. Rogan MM, Black KE. Dietary energy intake across the menstrual cycle: a narrative
review. Nutr Rev 2022 Nov 11: (Epub ahead of print; doi:
10.1093/nutrit/nuac094).
19. Moore DR, Sygo J, Morton JP. Fuelling the female athlete: carbohydrate and protein
recommendations. Eur J Sport Sci 2022;22(5):684-96.
20. Johnson WG, Corrigan SA, Lemmon CR, Bergeron KB, Crusco AH. Energy regulation
over the menstrual cycle. Physiol Behav 1994;56(3):523-7.
21. Dalvit-McPhillips SP. The effect of the human menstrual cycle on nutrient intake.
Physiol Behav 1983;31(2):209-12.
22. Lyons PM, Truswell AS, Mira M, Vizzard J, Abraham SF. Reduction of food intake in
the ovulatory phase of the menstrual cycle. Am J Clin Nutr 1989;49(6):1164-8.
23. de Souza LB, Martins KA, Cordeiro MM, Rodrigues YS, Rafacho BPM, Bomfim RA.
Do food intake and food cravings change during the menstrual cycle of young
women? Rev Bras Ginecol 2018;40(11):686-92.
106
24. Gorczyca AM, Sjaarda LA, Mitchell EM, Perkins NJ, Schliep KC, Wactawski-Wende J,
Mumford SL. Changes in macronutrient, micronutrient, and food group intakes
throughout the menstrual cycle in healthy, premenopausal women. Eur J Nutr
2016;55(3):1181-8.
25. Tarasuk V, Beaton GH. Menstrual-cycle patterns in energy and macronutrient
intake. Am J Clin Nutr 1991;53(2):442-7.
26. Martini MC, Lampe JW, Slavin JL, Kurzer MS. Effect of the menstrual cycle on
energy and nutrient intake. Am J Clin Nutr 1994;60(6):895-9.
27. King NA, Tremblay A, Blundell JE. Effects of exercise on appetite control:
implications for energy balance. Med Sci Sports Exerc 1997;29(8):1076-89.
28. Dorling J, Broom DR, Burns SF, Clayton DJ, Deighton K, James LJ, King JA, Miyashita
M, Thackray AE, Batterham RL, et al. Acute and chronic effects of exercise on
appetite, energy intake, and appetite-related hormones: the modulating effect of
adiposity, sex, and habitual physical activity. Nutrients 2018;10(9):1140.
29. Logue DM, Madigan SM, Melin A, Delahunt E, Heinen M, McDonnell SJ, Corish CA.
Low energy availability in athletes 2020: an updated narrative review of
prevalence, risk, within-day energy balance, knowledge, and impact on sports
performance. Nutrients 2020;12(3):835.
30. Schaumberg MA, Jenkins DG, Janse de Jonge XAK, Emmerton LM, Skinner TL.
Three-step method for menstrual and oral contraceptive cycle verification. J Sci
Med Sport 2017;20(11):965-9.
31. de Souza MJ, Toombs RJ, Scheid JL, O'Donnell E, West SL, Williams NI. High
prevalence of subtle and severe menstrual disturbances in exercising women:
confirmation using daily hormone measures. Hum Reprod 2010;25(2):491-503.
32. Burke LM, Lundy B, Fahrenholtz IL, Melin AK. Pitfalls of conducting and
interpreting estimates of energy availability in free-living athletes. Int J Sport Nutr
Exerc Metab 2018;28(4):350-63.
33. Mountjoy M, Sundgot-Borgen J, Burke L, Carter S, Constantini N, Lebrun C, Meyer
N, Sherman R, Steffen K, Budgett R, et al. The IOC consensus statement: beyond the
female athlete triad—relative energy deficiency in sport (RED-S). Br J Sports Med
2014;48(7):491-7.
34. Kerksick CM, Arent S, Schoenfeld BJ, Stout RJ, Campbell B, Wilborn CD, Taylor L,
Kalman D, Smith-Ryan AE, Krieder RB, et al. International Society of Sports
Nutrition position stand: nutrient timing. J Int Soc Sports Nutr 2017;14(1):33.
35. Institute of Medicine (US) Committee on Understanding the Biology of Sex and
Gender Differences. Exploring the biological contributions to human health: does
sex matter? Washington, DC: National Academies Press, 2001.
107
36. Loucks AB, Verdun M, Heath EM. Low energy availability, not stress of exercise,
alters LH pulsatility in exercising women. J Appl Physiol 1998;84(1):37-46.
37. Speroff L, Fritz MA. Clinical gynecologic endocrinology & infertility. 7th ed.
Philadelphia, PA: Lippincott Williams & Wilkins, 2005.
38. Bjelland EK, Hofvind S, Byberg L, Eskild A. The relation of age at menarche with
age at natural menopause: a population study of 336 788 women in Norway. Hum
Reprod 2018;33(6):1149-57.
39. Park SJ, Goldsmith LT, Skurnick JH, Wojtczuk A, Weiss G. Characteristics of the
urinary luteinizing hormone surge in young ovulatory women. Fertil Steril
2007;88(3):684-90.
40. Hoff JD, Quigley ME, Yen SS. Hormonal dynamics at midcycle: a reevaluation. J Clin
Endocrinol Metab 1983;57(4):792-6.
41. World Health Organization, Task Force on Methods for the Determination of the
Fertile Period, Special Programme of Research, Development and Research
Training in Human Reproduction. Temporal relationships between ovulation and
defined changes in the concentration of plasma estradiol-17 beta, luteinizing
hormone, follicle-stimulating hormone, and progesterone. Am J Obstet Gynecol
1980;138(4):383-90.
42. Behre HM, Kuhlage J, Gassner C, Sonntag B, Schem C, Schneider HP, Nieschlag E.
Prediction of ovulation by urinary hormone measurements with the home use
ClearPlan® Fertility Monitor: comparison with transvaginal ultrasound scans and
serum hormone measurements. Hum Reprod 2000;15(12):2478-82.
43. Fehring RJ, Schneider M, Raviele K. Variability in the phases of the menstrual cycle.
J Obstet Gynecol Neonatal Nurs 2006;35(3):376-84.
44. Sherman BM, Korenman SG. Hormonal characteristics of the human menstrual
cycle throughout reproductive life. J Clin Invest 1975;55(4):699-706.
45. Wilcox AJ, Dunson D, Baird DD. The timing of the "fertile window" in the menstrual
cycle: day specific estimates from a prospective study. Br Med J (Clin Res Ed)
2000;321(7271):1259-62.
46. Janse De Jonge X, Thompson B, Han A. Methodological recommendations for
menstrual cycle research in sports and exercise. Med Sci Sports Exerc
2019;51(12): 2610-7.
47. Bauman JE. Basal body temperature: unreliable method of ovulation detection.
Fertil Steril 1981;36(6):729-33.
108
48. Forman RG, Chapman MC, Steptoe PC. The effect of endogenous progesterone on
basal body temperature in stimulated ovarian cycles. Hum Reprod 1987;2(8):631-
4.
49. Barron ML, Fehring RJ. Basal body temperature assessment: is it useful to couples
seeking pregnancy? MCN Am J Matern Child Nurs 2005;30(5):290-6.
50. Moghissi KS. Accuracy of basal body temperature for ovulation detection. Fertil
Steril 1976;27(12):1415-21.
51. Miller PB, Soules MR. The usefulness of a urinary LH kit for ovulation prediction
during menstrual cycles of normal women. Obstet Gynecol 1996;87(1):13-7.
52. Wood P. Salivary steroid assays - research or routine? Ann Clin Biochem
2009;46(3):183-96.
53. Chatterton RT, Mateo ET, Hou N, Rademaker AW, Acharya S, Jordan VC, Morrow M.
Characteristics of salivary profiles of oestradiol and progesterone in
premenopausal women. J Endocrinol 2005;186(1):77-84.
54. Delfs TM, Klein S, Fottrell P, Naether OG, Leidenberger FA, Zimmermann RC. 24-
hour profiles of salivary progesterone. Fertil Steril 1994;62(5):960-6.
55. Gandara BK, Leresche L, Mancl L. Patterns of salivary estradiol and progesterone
across the menstrual cycle. Ann N Y Acad Sci 2007;1098:446-50.
56. Prior JC, Naess M, Langhammer A, Forsmo S. Ovulation prevalence in women with
spontaneous normal-length menstrual cycles - a population-based cohort from
HUNT3, Norway. PLoS One 2015;10(8):e0134473.
57. Schliep KC, Mumford SL, Hammoud AO, Stanford JB, Kissell KA, Sjaarda LA, Perkins
NJ, Ahrens KA, Wactawski-Wende J, Mendola P, et al. Luteal phase deficiency in
regularly menstruating women: prevalence and overlap in identification based on
clinical and biochemical diagnostic criteria. J Clin Endocrinol Metab
2014;99(6):e1007-14.
58. Chesang J, Richardson A, Potter J, Coope P. Prevalence of contraceptive use in New
Zealand women. N Z Med J 2016;129(1444):58-67.
59. Heather AK, Thorpe H, Ogilvie M, Sims ST, Beable S, Milsom S, Schofield KL,
Coleman L, Hamilton B. Biological and socio-cultural factors have the potential to
influence the health and performance of elite female athletes: a cross sectional
survey of 219 elite female athletes in Aotearoa New Zealand. Front Sports Act
Living 2021;3(27):601420.
60. Elliott-Sale KJ, Smith S, Bacon J, Clayton D, McPhilimey M, Goutianos G, Hampson J,
Sale C. Examining the role of oral contraceptive users as an experimental and/or
control group in athletic performance studies. Contraception 2013;88(3):408-12.
109
61. Academy of Nutrition and Dietetics, Dietitians of Canada, American College of
Sports Medicine. Nutrition and Athletic Performance. Med Sci Sports Exerc
2016;48(3):543-68.
62. Boisseau N, Isacco L. Substrate metabolism during exercise: sexual dimorphism
and women’s specificities. Eur J Sport Sci 2022;22(5):672-83.
63. Burke LM, Hawley JA, Wong SHS, Jeukendrup AE. Carbohydrates for training and
competition. J Sports Sci 2011;29(sup1):17-27.
64. Hargreaves M. Metabolic responses to carbohydrate ingestion: effect on exercise
performance. In: Perspectives in exercise science and sports medicine vol. 12: the
metabolic basis of performance in exercise and sport. Carmel, IN: Cooper
Publishing Company, 1999:93-124.
65. Li ETS, Anderson GH. Dietary carbohydrate and the nervous system. Nutr Res
1987;7(12):1329-39.
66. Hammond KM, Sale C, Fraser W, Tang J, Shepherd SO, Strauss JA, Close GL, Cocks
M, Louis J, Pugh J, et al. Post-exercise carbohydrate and energy availability induce
independent effects on skeletal muscle cell signalling and bone turnover:
implications for training adaptation. J Physiol 2019;597(18):4779-96.
67. Heikura IA, Burke LM, Hawley JA, Ross ML, Garvican-Lewis L, Sharma AP, McKay
AKA, Leckey JJ, Welvaert M, McCall L, et al. A short-term ketogenic diet impairs
markers of bone health in response to exercise. Front Endocrinol 2019;10:880.
68. McKay AKA, Peeling P, Pyne DB, Tee N, Whitfield J, Sharma AP, Heikura IA, Burke
LM. Six days of low carbohydrate, not energy availability, alters the iron and
immune response to exercise in elite athletes. Med Sci Sports Exerc
2022;54(3):377-87.
69. Ministry of Health. Eating and activity guidelines for New Zealand Adults.
Wellington: Ministry of Health, 2020.
70. Hargreaves M, Hawley JA, Jeukendrup A. Pre-exercise carbohydrate and fat
ingestion: effects on metabolism and performance. J Sports Sci 2004;22(1):31-8.
71. Tarnopolsky MA, Atkinson SA, Phillips SM, MacDougall JD. Carbohydrate loading
and metabolism during exercise in men and women. J Appl Physiol
1995;78(4):1360-8.
72. Tarnopolsky MA, Zawada C, Richmond LB, Carter S, Shearer J, Graham T, Phillips
SM. Gender differences in carbohydrate loading are related to energy intake. J Appl
Physiol 2001;91(1):225-30.
73. James AP, Lorraine M, Cullen D, Goodman C, Dawson B, Palmer TN, Fournier PA.
Muscle glycogen supercompensation: absence of a gender-related difference. Eur J
Appl Physiol 2001;85(6):533-8.
110
74. Hackney AC. Effects of the menstrual cycle on resting muscle glycogen content.
Horm Metab Res 1990;22(12):647.
75. Nicklas BJ, Hackney AC, Sharp RL. The menstrual cycle and exercise: performance,
muscle glycogen, and substrate responses. Int J Sports Med 1989;10(4):264-9.
76. McLay RT, Thomson CD, Williams SM, Rehrer NJ. Carbohydrate loading and female
endurance athletes: effect of menstrual-cycle phase. Int J Sport Nutr Exerc Metab
2007;17(2):189-205.
77. Karelis AD, Smith JW, Passe DH, Péronnet F. Carbohydrate administration and
exercise performance: what are the potential mechanisms involved? Sports Med
2010;40(9):747-63.
78. Wallis GA, Dawson R, Achten J, Webber J, Jeukendrup AE. Metabolic response to
carbohydrate ingestion during exercise in males and females. Am J Physiol
Endocrinol Metab 2006;290(4):e708-15.
79. Campbell SE, Angus DJ, Febbraio MA. Glucose kinetics and exercise performance
during phases of the menstrual cycle: effect of glucose ingestion. Am J Physiol
Endocrinol Metab 2001;281(4):e817-25.
80. Bailey SP, Zacher CM, Mittleman KD. Effect of menstrual cycle phase on
carbohydrate supplementation during prolonged exercise to fatigue. J Appl Physiol
2000;88(2):690-7.
81. Viribay A, Arribalzaga S, Mielgo-Ayuso J, Castañeda-Babarro A, Seco-Calvo J,
Urdampilleta A. Effects of 120 g/h of carbohydrates intake during a mountain
marathon on exercise-induced muscle damage in elite runners. Nutrients
2020;12(5):1367.
82. Hearris MA, Pugh JN, Langan-Evans C, Mann SJ, Burke L, Stellingwerff T, Gonzalez
JT, Morton JP. 13c-glucose-fructose labeling reveals comparable exogenous CHO
oxidation during exercise when consuming 120 g/h in fluid, gel, jelly chew, or
coingestion. J Appl Physiol 2022;132(6):1394-406.
83. O'Hara JP, Woods DR, Mellor A, Boos C, Gallagher L, Tsakirides C, Arjomandkhah
NC, Holdsworth DA, Cooke CB, Morrison DJ. A comparison of substrate oxidation
during prolonged exercise in men at terrestrial altitude and normobaric normoxia
following the coingestion of 13c glucose and 13c fructose. Physiol Rep
2017;5(1):e13101.
84. O'Hara JP, Duckworth L, Black A, Woods DR, Mellor A, Boos C, Gallagher L,
Tsakirides C, Arjomandkhah NC, Morrison DJ, et al. Fuel use during exercise at
altitude in women with glucose-fructose ingestion. Med Sci Sports Exerc
2019;51(12):2586-94.
111
85. Pettersson S, Edin F, Bakkman L, McGawley K. Effects of supplementing with an
18% carbohydrate-hydrogel drink versus a placebo during whole-body exercise in
−5 °C with elite cross-country ski athletes: a crossover study. J Int Soc Sports Nutr
2019;16(1):46.
86. Burke LM, Kiens B, Ivy JL. Carbohydrates and fat for training and recovery. J Sports
Sci 2004;22(1):15-30.
87. Tarnopolsky MA, Bosman M, Macdonald JR, Vandeputte D, Martin J, Roy BD.
Postexercise protein-carbohydrate and carbohydrate supplements increase
muscle glycogen in men and women. J Appl Physiol 1997;83(6):1877-83.
88. Flynn S, Rosales A, Hailes W, Ruby B. Males and females exhibit similar muscle
glycogen recovery with varied recovery food sources. Eur J Appl Physiol
2020;120(5):1131-42.
89. Impey SG, Hearris MA, Hammond KM, Bartlett JD, Louis J, Close GL, Morton JP. Fuel
for the work required: a theoretical framework for carbohydrate periodization and
the glycogen threshold hypothesis. Sports Med 2018;48(5):1031-48.
90. Burke LM, Cox GR, Culmmings NK, Desbrow B. Guidelines for daily carbohydrate
intake: do athletes achieve them? Sports Med 2001;31(4):267-99.
91. Renard M, Kelly DT, Chéilleachair NN, Catháin CÓ. How does the dietary intake of
female field-based team sport athletes compare to dietary recommendations for
health and performance? A systematic literature review. Nutrients
2021;13(4):1235.
92. Noh N, Shuhaimi F, Md Nor N. Carbohydrate intakes and preferences among
endurance athletes in Universiti Teknologi MARA Selangor. J Phys Conf Ser
2020;1496(1):012010.
93. Traversa C, Nyman DLE, Spriet LL. Dietary intake over a 7-day training and game
period in female varsity rugby union players. Nutrients 2022;14(11):2281.
94. Hoshino F, Inaba H, Edama M, Saya N, Maruyama S, Omori G. Sleep quality and
nutrient intake in Japanese female university student-athletes: a cross-sectional
study. Healthcare 2022;10(4):663.
95. Vermeulen TF, Boyd LA, Spriet LL. Dietary macronutrient and micronutrient
intake over a 7-day period in female varsity ice hockey players. Nutrients
2021;13(7):2262.
96. Danh JP, Nucci A, Andrew Doyle J, Feresin RG. Assessment of sports nutrition
knowledge, dietary intake, and nutrition information source in female collegiate
athletes: a descriptive feasibility study. J Am Coll Health 2021 Nov 4: (Epub ahead
of print; doi: 10.1080/07448481.2021.1987919).
112
97. Zanders BR, Currier BS, Harty PS, Zabriskie HA, Smith CR, Stecker RA, Richmond
SR, Jagim AR, Kerksick CM. Changes in energy expenditure, dietary intake, and
energy availability across an entire collegiate women's basketball season. J
Strength Cond Res 2021;35(3):804-10.
98. Beermann BL, Lee DG, Almstedt HC, McCormack WP. Nutritional intake and energy
availability of collegiate distance runners. J Am Coll Nutr 2020;39(8):747-55.
99. Gibson-Smith E, Storey R, Ranchordas M. Dietary intake, body composition and
iron status in experienced and elite climbers. Front Nutr 2020;7:122.
100. Gogojewicz A, Śliwicka E, Durkalec-Michalski K. Assessment of dietary intake and
nutritional status in crossfit-trained individuals: a descriptive study. Int J Environ
Res Public Health 2020;17(13):4772.
101. Kumahara H, Ohta C, Nabeshima E, Nakayama A, Mine S, Yamato T. Dietary intake
and energy expenditure during two different phases of athletic training in female
collegiate lacrosse players. J Strength Cond Res 2020;34(6):1547-54.
102. Baranauskas M, Jablonskienė V, Abaravičius JA, Stukas R. Actual nutrition and
dietary supplementation in Lithuanian elite athletes. Medicina 2020;56(5):247.
103. Brown AF, Brooks SJ, Smith SR, Stephens JM, Lotstein AK, Skiles CM, Alfiero CJ,
Meenan MJ. Female collegiate dancers body composition, macronutrient and
micronutrient intake over two academic years: a longitudinal analysis. J Funct
Morphol Kinesiol 2020;5(1):17.
104. Jenner SL, Devlin BL, Forsyth AK, Belski R. Dietary intakes of professional
Australian Football League Women’s (AFLW) athletes during a preseason training
week. J Sci Med Sport 2019;22(11):1266-71.
105. Condo D, Lohman R, Kelly M, Carr A. Nutritional intake, sports nutrition knowledge
and energy availability in female Australian Rules Football players. Nutrients
2019;11(5):971.
106. Wardenaar F, Brinkmans N, Ceelen I, Van Rooij B, Mensink M, Witkamp R, de Vries
J. Macronutrient intakes in 553 Dutch elite and sub-elite endurance, team, and
strength athletes: does intake differ between sport disciplines? Nutrients
2017;9(2):119.
107. Capling L, Beck KL, Gifford JA, Slater G, Flood VM, O'Connor H. Validity of dietary
assessment in athletes: a systematic review. Nutrients 2017;9(12):1313.
108. Ihalainen JK, Löfberg I, Kotkajuuri A, Kyröläinen H, Hackney AC, Taipale-Mikkonen
RS. Influence of menstrual cycle or hormonal contraceptive phase on energy intake
and metabolic hormones—a pilot study. Endocrines 2021;2(2):79-90.
109. Gillen JB, Trommelen J, Wardenaar FC, Brinkmans NY, Versteegen JJ, Jonvik KL,
Kapp C, de Vries J, van den Borne JJGC, Gibala MJ, et al. Dietary protein intake and
113
distribution patterns of well-trained Dutch athletes. Int J Sport Nutr Exerc Metab
2017;27(2):105-14.
110. NutritionQuest. List of questionnaires and screeners. Internet:
https://www.nutritionquest.com/assessment/list-of-questionnaires-and-
screeners/ (accessed 15 October 2022).
111. Subar AF, Kirkpatrick SI, Mittl B, Zimmerman TP, Thompson FE, Bingley C, Willis G,
Islam NG, Baranowski T, McNutt S, et al. The Automated Self-Administered 24-
hour dietary recall (ASA24): a resource for researchers, clinicians, and educators
from the National Cancer Institute. J Acad Nutr Diet 2012;112(8):1134-7.
112. Wardenaar FC, Steennis J, Ceelen IJ, Mensink M, Witkamp R, de Vries JH. Validation
of web-based, multiple 24-h recalls combined with nutritional supplement intake
questionnaires against nitrogen excretions to determine protein intake in Dutch
elite athletes. Br J Nutr 2015;114(12):2083-92.
113. Jäger R, Kerksick CM, Campbell BI, Cribb PJ, Wells SD, Skwiat TM, Purpura M,
Ziegenfuss TN, Ferrando AA, Arent SM, et al. International Society of Sports
Nutrition Position stand: protein and exercise. J Int Soc Sports Nutr 2017;14(1):20.
114. Rodriguez NR, Vislocky LM, Gaine PC. Dietary protein, endurance exercise, and
human skeletal-muscle protein turnover. Curr Opin Clin Nutr Metab Care
2007;10(1):40-5.
115. Anthony JC, Anthony TG, Kimball SR, Jefferson LS. Signaling pathways involved in
translational control of protein synthesis in skeletal muscle by leucine. J Nutr
2001;131(3):856-60.
116. Mettler S, Mitchell N, Tipton KD. Increased protein intake reduces lean body mass
loss during weight loss in athletes. Med Sci Sports Exerc 2010;42(2):326-37.
117. Areta JL, Burke LM, Camera DM, West DW, Crawshay S, Moore DR, Stellingwerf T,
Phillips SM, Hawley JA, Coffey VG. Reduced resting skeletal muscle protein
synthesis is rescued by resistance exercise and protein ingestion following short-
term energy deficit. Am J Physiol Endocrinol Metab 2014;306(8):e989-97.
118. Gillen JB, West DWD, Williamson EP, Fung HJW, Moore DR. Low-carbohydrate
training increases protein requirements of endurance athletes. Med Sci Sports
Exerc 2019;51(11):2294-301.
119. Børsheim E, Tipton KD, Wolf SE, Wolfe RR. Essential amino acids and muscle
protein recovery from resistance exercise. Am J Physiol Endocrinol Metab
2002;283(4):e648-57.
120. Tipton KD, Gurkin BE, Matin S, Wolfe RR. Nonessential amino acids are not
necessary to stimulate net muscle protein synthesis in healthy volunteers. J Nutr
Biochem 1999;10(2):89-95.
114
121. Schoenfeld BJ, Aragon AA. Is there a postworkout anabolic window of opportunity
for nutrient consumption? Clearing up controversies. J Orthop Sports Phys Ther
2018;48(12):911-4.
122. Areta JL, Burke LM, Ross ML, Camera DM, West DW, Broad EM, Jeacocke NA, Moore
DR, Stellingwerff T, Phillips SM, et al. Timing and distribution of protein ingestion
during prolonged recovery from resistance exercise alters myofibrillar protein
synthesis. J Physiol 2013;591(9):2319-31.
123. Breen L, Philp A, Witard OC, Jackman SR, Selby A, Smith K, Baar K, Tipton KD. The
influence of carbohydrate-protein co-ingestion following endurance exercise on
myofibrillar and mitochondrial protein synthesis. J Physiol 2011;589(16):4011-25.
124. Ferguson-Stegall L, McCleave EL, Ding Z, Doerner 3rd PG, Wang B, Liao YH,
Kammer L, Liu Y, Hwang J, Dessard BM, et al. Postexercise carbohydrate-protein
supplementation improves subsequent exercise performance and intracellular
signaling for protein synthesis. J Strength Cond Res 2011;25(5):1210-24.
125. Hulmi JJ, Kovanen V, Selänne H, Kraemer WJ, Häkkinen K, Mero AA. Acute and
long-term effects of resistance exercise with or without protein ingestion on
muscle hypertrophy and gene expression. Amino Acids 2009;37(2):297-308.
126. Farnfield MM, Breen L, Carey KA, Garnham A, Cameron-Smith D. Activation of
mTOR signalling in young and old human skeletal muscle in response to combined
resistance exercise and whey protein ingestion. Appl Physiol Nutr Metab
2012;37(1):21-30.
127. Burd NA, West DW, Moore DR, Atherton PJ, Staples AW, Prior T, Tang JE, Rennie
MJ, Baker SK, Phillips SM. Enhanced amino acid sensitivity of myofibrillar protein
synthesis persists for up to 24 h after resistance exercise in young men. J Nutr
2011;141(4):568-73.
128. Mercer D, Convit L, Condo D, Carr AJ, Hamilton DL, Slater G, Snipe RMJ. Protein
requirements of pre-menopausal female athletes: systematic literature review.
Nutrients 2020;12(11):3527.
129. National Health and Medical Research Council, Australian Government
Department of Health and Ageing, New Zealand Ministry of Health. Nutrient
reference values for Australia and New Zealand. Version 1.2. Canberra: National
Health and Medical Research Council, 2017.
130. Tipton KD, Witard OC. Protein requirements and recommendations for athletes:
relevance of ivory tower arguments for practical recommendations. Clin Sports
Med 2007;26(1):17-36.
131. Tinline-Goodfellow CT, West DWD, Malowany JM, Gillen JB, Moore DR. An acute
reduction in habitual protein intake attenuates post exercise anabolism and may
bias oxidation-derived protein requirements in resistance trained men. Front Nutr
2020;7:55.
115
132. West DW, Burd NA, Churchward-Venne TA, Camera DM, Mitchell CJ, Baker SK,
Hawley JA, Coffey VG, Phillips SM. Sex-based comparisons of myofibrillar protein
synthesis after resistance exercise in the fed state. J Appl Physiol
2012;112(11):1805-13.
133. Taylor LW, Wilborn C, Roberts MD, White A, Dugan K. Eight weeks of pre- and
postexercise whey protein supplementation increases lean body mass and
improves performance in Division III collegiate female basketball players. Appl
Physiol Nutr Metab 2016;41(3):249-54.
134. Wilborn CD, Taylor LW, Outlaw J, Williams L, Campbell B, Foster CA, Smith-Ryan A,
Urbina S, Hayward S. The effects of pre- and post-exercise whey vs. casein protein
consumption on body composition and performance measures in collegiate female
athletes. J Sports Sci Med 2013;12(1):74-9.
135. Wilborn CD, Outlaw JJ, Mumford PW, Urbina SL, Hayward S, Roberts MD, Taylor
LW, Foster CA. A pilot study examining the effects of 8-week whey protein versus
whey protein plus creatine supplementation on body composition and
performance variables in resistance-trained women. Ann Nutr Metab 2016;69(3-
4):190-9.
136. Pihoker AA, Peterjohn AM, Trexler ET, Hirsch KR, Blue MNM, Anderson KC, Ryan
ED, Smithy-Ryan AE. The effects of nutrient timing on training adaptations in
resistance-trained females. J Sci Med Sport 2019;22(4):472-7.
137. Brown MA, Stevenson EJ, Howatson G. Whey protein hydrolysate supplementation
accelerates recovery from exercise-induced muscle damage in females. Appl
Physiol Nutr Metab 2018;43(4):324-30.
138. Lamont LS, Lemon PW, Bruot BC. Menstrual cycle and exercise effects on protein
catabolism. Med Sci Sports Exerc 1987;19(2):106-10.
139. Landau RL, Lugibihl K. The effect of dietary protein on the catabolic influence of
progesterone. J Clin Endocrinol Metab 1961;21(11):1345-54.
140. Calloway DH, Kurzer MS. Menstrual cycle and protein requirements of women. J
Nutr 1982;112(2):356-66.
141. Lariviere F, Moussalli R, Garrel D. Increased leucine flux and leucine oxidation
during the luteal phase of the menstrual cycle in women. Am J Physiol Endocrinol
Metab 1994;267(3):e422-8.
142. Kriengsinyos W, Wykes LJ, Goonewardene LA, Ball RO, Pencharz PB. Phase of
menstrual cycle affects lysine requirement in healthy women. Am J Physiol
Endocrinol Metab 2004;287(3):e489-96.
116
143. Kerksick CM, Wilborn CD, Roberts MD, Smith-Ryan A, Kleiner SM, Jäger R, Collins
R, Cooke M, Davis JN, Galvan E, et al. ISSN exercise & sports nutrition review
update: research & recommendations. J Int Soc Sports Nutr 2018;15(1):38.
144. Burke LM, Ross ML, Garvican-Lewis LA, Welvaert M, Heikura IA, Forbes SG,
Mirtschin JG, Cato LE, Strobel N, Sharma AP, et al. Low carbohydrate, high fat diet
impairs exercise economy and negates the performance benefit from intensified
training in elite race walkers. J Physiol 2017;595(9):2785-807.
145. Burke LM. Re-examining high-fat diets for sports performance: did we call the 'nail
in the coffin' too soon? Sports Med 2015;45(s1):33-49.
146. Yeo WK, Carey AL, Burke L, Spriet LL, Hawley JA. Fat adaptation in well-trained
athletes: effects on cell metabolism. Appl Physiol Nutr Metab 2011;36(1):12-22.
147. Leckey JJ, Hoffman NJ, Parr EB, Devlin BL, Trewin AJ, Stepto NK, Morton JP, Burke
LM, Hawley JA. High dietary fat intake increases fat oxidation and reduces skeletal
muscle mitochondrial respiration in trained humans. FASEB J 2018;32(6):2979-
91.
148. Loucks AB, Kiens B, Wright HH. Energy availability in athletes. J Sports Sci
2011;29(s1):7-15.
149. Ihle R, Loucks AB. Dose-response relationships between energy availability and
bone turnover in young exercising women. J Bone Miner Res 2004;19(8):1231-40.
150. Loucks AB, Thuma JR. Luteinizing hormone pulsatility is disrupted at a threshold
of energy availability in regularly menstruating women. J Clin Endocrinol Metab
2003;88(1):297-311.
151. Yeager KK, Agostini R, Nattiv A, Drinkwater B. The female athlete triad: disordered
eating, amenorrhea, osteoporosis. Med Sci Sports Exerc 1993;25(7):775-7.
152. Nattiv A, Loucks AB, Manore MM, Sanborn CF, Sundgot-Borgen J, Warren MP.
American College of Sports Medicine position stand. The female athlete triad. Med
Sci Sports Exerc 2007;39(10):1867-82.
153. Mountjoy M, Sundgot-Borgen JK, Burke LM, Ackerman KE, Blauwet C, Constantini
N, Lebrun C, Lundy B, Melin AK, Meyer NL, et al. IOC consensus statement on
relative energy deficiency in sport (RED-S): 2018 update. Br J Sports Med
2018;52(11):687-97.
154. de Souza MJ, Williams NI, Nattiv A, Joy E, Misra M, Loucks AB, Matheson G,
Olmstead MP, Barrack M, Mallinson RJ, et al. Misunderstanding the female athlete
triad: refuting the IOC consensus statement on relative energy deficiency in sport
(RED-S). Br J Sports Med 2014;48(20):1461-5.
117
155. Areta JL, Taylor HL, Koehler K. Low energy availability: history, definition and
evidence of its endocrine, metabolic and physiological effects in prospective
studies in females and males. Eur J Appl Physiol 2021;121(1):1-21.
156. Heikura IA, Stellingwerff T, Areta JL. Low energy availability in female athletes:
from the lab to the field. Eur J Sport Sci 2022;22(5):709-19.
157. Lieberman JL, MJ DES, Wagstaff DA, Williams NI. Menstrual disruption with
exercise is not linked to an energy availability threshold. Med Sci Sports Exerc
2018;50(3):551-61.
158. Williams NI, Leidy HJ, Hill BR, Lieberman JL, Legro RS, de Souza MJ. Magnitude of
daily energy deficit predicts frequency but not severity of menstrual disturbances
associated with exercise and caloric restriction. Am J Physiol Endocrinol Metab
2015;308(1):e29-39.
159. Reed JL, de Souza MJ, Mallinson RJ, Scheid JL, Williams NI. Energy availability
discriminates clinical menstrual status in exercising women. J Int Soc Sports Nutr
2015;12(1):11.
160. Guebels CP, Kam LC, Maddalozzo GF, Manore MM. Active women before/after an
intervention designed to restore menstrual function: resting metabolic rate and
comparison of four methods to quantify energy expenditure and energy
availability. Int J Sport Nutr Exerc Metab 2014;24(1):37-46.
161. Hilton LK, Loucks AB. Low energy availability, not exercise stress, suppresses the
diurnal rhythm of leptin in healthy young women. Am J Physiol Endocrinol Metab
2000;278(1):e43-9.
162. Spaulding SW, Chopra IJ, Sherwin RS, Lyall SS. Effect of caloric restriction and
dietary composition of serum T3 and reverse T3 in man. J Clin Endocrinol Metab
1976;42(1):197-200.
163. Fahrenholtz IL, Sjödin A, Benardot D, Tornberg Å B, Skouby S, Faber J, Sundgot-
Borgen KJ, Melin AK. Within-day energy deficiency and reproductive function in
female endurance athletes. Scand J Med Sci Sports 2018;28(3):1139-46.
164. Taylor HL, Garabello G, Pugh J, Morton J, Langan-Evans C, Louis J, Borgersen R,
Areta JL. Patterns of energy availability of free-living athletes display day-to-day
variability that is not reflected in laboratory-based protocols: insights from elite
male road cyclists. J Sports Sci 2022;40(16):1849-56.
165. Pauli SA, Berga SL. Athletic amenorrhea: energy deficit or psychogenic challenge?
Ann N Y Acad Sci 2010;1205:33-8.
166. Wasserfurth P, Palmowski J, Hahn A, Krüger K. Reasons for and consequences of
low energy availability in female and male athletes: social environment,
adaptations, and prevention. Sports Med 2020;6(1):44.
118
167. Gibson RS. Principles of nutritional assessment. 2nd ed. New York: Oxford
University Press, 2005.
168. Bailey RL. Overview of dietary assessment methods for measuring intakes of foods,
beverages, and dietary supplements in research studies. Curr Opin Biotechnol
2021;70:91-6.
169. Marr JW, Heady JA. Within- and between-person variation in dietary surveys:
number of days needed to classify individuals. Hum Nutr Appl Nutr
1986;40(5):347-64.
170. Braakhuis AJ, Meredith K, Cox GR, Hopkins WG, Burke LM. Variability in estimation
of self-reported dietary intake data from elite athletes resulting from coding by
different sports dietitians. Int J Sport Nutr Exerc Metab 2003;13(2):152-65.
171. Magkos F, Yannakoulia M. Methodology of dietary assessment in athletes: concepts
and pitfalls. Curr Opin Clin Nutr Metab Care 2003;6(5):539-49.
172. Li ET, Tsang LB, Lui SS. Menstrual cycle and voluntary food intake in young
Chinese women. Appetite 1999;33(1):109-18.
173. Gil YRC, Fagundes RLM, Santos E, Calvo MCM, Bernardine JD. Relation of menstrual
cycle and alimentary consumption of women. Clin Nutr ESPEN 2009;4(5):e257-60.
174. Chung SC, Bond EF, Jarrett ME. Food Intake Changes Across the Menstrual Cycle in
Taiwanese Women. Biol Res Nurs 2010;12(1):37-46.
175. Elliott SA, Ng J, Leow MK-s, Henry CJ. The influence of the menstrual cycle on
energy balance and taste preference in Asian Chinese women. Eur J Nutr
2015;54(8):1323-32.
176. Manocha S, Choudhuri G, Tandon BN. A study of dietary intake in pre- and post-
menstrual period. Hum Nutr Appl Nutr 1986;40(3):213-6.
177. Gong EJ, Garrel D, Calloway DH. Menstrual cycle and voluntary food intake. Am J
Clin Nutr 1989;49(2):252-8.
178. Boushey CJ, Spoden M, Zhu FM, Delp EJ, Kerr DA. New mobile methods for dietary
assessment: review of image-assisted and image-based dietary assessment
methods. Proc Nutr Soc 2017;76(3):283-94.
179. Gemming L, Utter J, Ni Mhurchu C. Image-assisted dietary assessment: a systematic
review of the evidence. J Acad Nutr Diet 2015;115(1):64-77.
180. Williamson DA, Allen HR, Martin PD, Alfonso AJ, Gerald B, Hunt A. Comparison of
digital photography to weighed and visual estimation of portion sizes. J Am Diet
Assoc 2003;103(9):1139-45.
119
181. Simpson A, Gemming L, Baker D, Braakhuis A. Do image-assisted mobile
applications improve dietary habits, knowledge, and behaviours in elite athletes? A
pilot study. Sports 2017;5(3):60.
182. Rollo ME, Ash S, Lyons-Wall P, Russell A. Trial of a mobile phone method for
recording dietary intake in adults with type 2 diabetes: evaluation and
implications for future applications. J Telemed Telecare 2011;17(6):318-23.
183. Martin CK, Han H, Coulon SM, Allen HR, Champagne CM, Anton SD. A novel method
to remotely measure food intake of free-living individuals in real time: the remote
food photography method. Br J Nutr 2008;101(3):446-56.
184. Macdiarmid J, Blundell J. Assessing dietary intake: who, what and why of under-
reporting. Nutr Res Rev 1998;11(2):231-53.
185. Livingstone MB, Black AE. Markers of the validity of reported energy intake. J Nutr
2003;133(s3):895-920.
186. Briefel RR, Sempos CT, McDowell MA, Chien S, Alaimo K. Dietary methods research
in the third National Health and Nutrition Examination Survey: underreporting of
energy intake. Am J Clin Nutr 1997;65(s4):1203-9.
187. Goris AH, Westerterp KR. Underreporting of habitual food intake is explained by
undereating in highly motivated lean women. J Nutr 1999;129(4):878-82.
188. Johansson L, Solvoll K, Bjørneboe GE, Drevon CA. Under- and overreporting of
energy intake related to weight status and lifestyle in a nationwide sample. Am J
Clin Nutr 1998;68(2):266-74.
189. Barnard JA, Tapsell LC, Davies PS, Brenninger VL, Storlien LH. Relationship of high
energy expenditure and variation in dietary intake with reporting accuracy on 7
day food records and diet histories in a group of healthy adult volunteers. Eur J
Clin Nutr 2002;56(4):358-67.
190. Black AE. Critical evaluation of energy intake using the Goldberg cut-off for energy
intake: basal metabolic rate. A practical guide to its calculation, use and limitations.
Int J Obes Relat Metab Disord 2000;24(9):1119-30.
191. Westerterp KR. Doubly labelled water assessment of energy expenditure:
principle, practice, and promise. Eur J Appl Physiol 2017;117(7):1277-85.
192. Brown MA, Howatson G, Quin E, Redding E, Stevenson EJ. Energy intake and
energy expenditure of pre-professional female contemporary dancers. PLoS One
2017;12(2):e0171998.
193. Hoch AZ, Papanek P, Szabo A, Widlansky ME, Schimke JE, Gutterman DD.
Association between the female athlete triad and endothelial dysfunction in
dancers. Clin J Sport Med 2011;21(2):119-25.
120
194. Woodruff SJ, Meloche RD. Energy availability of female varsity volleyball players.
Int J Sport Nutr Exerc Metab 2013;23(1):24-30.
195. Amorim T, Freitas L, Metsios GS, Gomes TN, Wyon M, Flouris AD, et al. Associations
between nutrition, energy expenditure and energy availability with bone mass
acquisition in dance students: a 3-year longitudinal study. Arch Osteoporos
2021;16(1):141.
196. Lagowska K, Kapczuk K, Friebe Z, Bajerska J. Effects of dietary intervention in
young female athletes with menstrual disorders. J Int Soc Sports Nutr 2014;26:11-
21.
197. Łagowska K, Kapczuk K. Testosterone concentrations in female athletes and ballet
dancers with menstrual disorders. Eur J Sports Sci 2016;16(4):490-7.
198. Melin A, Tornberg Å B, Skouby S, Møller SS, Sundgot-Borgen J, Faber J, Sidelmann
JJ, Aziz M, Sjödin. Energy availability and the female athlete triad in elite endurance
athletes. Scand J Med Sci Sports 2015;25(5):610-22.
199. Schaal K, Loan MDV, Casazza GA. Reduced catecholamine response to exercise in
amenorrheic athletes. Med Sci Sports Exerc 2011;43(1):34-43.
200. Kinoshita N, Uchiyama E, Ishikawa-Takata K, Yamada Y, Okuyama K. Association of
energy availability with resting metabolic rates in competitive female teenage
runners: a cross-sectional study. J Int Soc Sports Nutr 2021;18(1):70.
201. Cialdella-Kam L, Guebels CP, Maddalozzo GF, Manore MM. Dietary intervention
restored menses in female athletes with exercise-associated menstrual
dysfunction with limited impact on bone and muscle health. Nutrients
2014;6(8):3018-39.
202. Reed JL, de Souza MJ, Williams NI. Changes in energy availability across the season
in Division I female soccer players. J Sports Sci 2013;31(3):314-24.
203. Murakami H, Kawakami R, Nakae S, Nakata Y, Ishikawa-Takata K, Tanaka S,
Miyachi M. Accuracy of wearable devices for estimating total energy expenditure:
comparison with metabolic chamber and doubly labeled water method. JAMA
Intern Med 2016;176(5):702-3.
204. Abel MG, Hannon JC, Sell K, Lillie T, Conlin G, Anderson D. Validation of the Kenz
Lifecorder EX and ActiGraph GT1M accelerometers for walking and running in
adults. Appl Physiol Nutr Metab 2008;33(6):1155-64.
205. Hoch AZ, Pajewski NM, Moraski L, Carrera GF, Wilson CR, Hoffmann RG, Schimke
JE, Gutterman DD. Prevalence of the female athlete triad in high school athletes
and sedentary students. Clin J Sports Med 2009;19(5):421-8.
121
206. Doyle-Lucas AF, Akers JD, Davy BM. Energetic efficiency, menstrual irregularity,
and bone mineral density in elite professional female ballet dancers. J Dance Med
Sci 2010;14(4):146-54.
207. Koehler K, Achtzehn S, Braun H, Mester J, Schaenzer W. Comparison of self-
reported energy availability and metabolic hormones to assess adequacy of
dietary energy intake in young elite athletes. Appl Physiol Nutr Metab
2013;38(7):725-33.
208. Silva AM, Matias CN, Santos DA, Thomas D, Bosy-Westphal A, Müller MJ,
Heymsfield SB, Sardinha LB. Compensatory changes in energy balance regulation
over one athletic season. Med Sci Sports Exerc 2017;49(6):1229-35.
209. Silva MR, Paiva T. Low energy availability and low body fat of female gymnasts
before an international competition. Eur J Sports Sci 2015;15(7):591-9.
210. Silva MR, Paiva T. Comparison of body composition and nutrients’ deficiencies
between Portuguese rink-hockey players. Eur J Pediatr 2017;176(1):41-50.
211. Viner RT, Harris M, Berning JR, Meyer NL. Energy availability and dietary patterns
of adult male and female competitive cyclists with lower than expected bone
mineral density. Int J Sports Nutr Exerc Metab 2015;25(6):594-602.
212. Muia EN, Wright HH, Onywera VO, Kuria EN. Adolescent elite Kenyan runners are
at risk for energy deficiency, menstrual dysfunction and disordered eating. J Sports
Sci 2016;34(7):598-606.
213. Heikura IA, Uusitalo ALT, Stellingwerff T, Bergland D, Mero AA, Burke LM. Low
energy availability is difficult to assess but outcomes have large impact on bone
injury rates in elite distance athletes. Int J Sport Nutr Exerc Metab 2018;28(4):1-9.
214. Black K, Slater J, Brown RC, Cooke R. Low energy availability, plasma lipids, and
hormonal profiles of recreational athletes. J Strength Cond Res 2018;32(10):2816-
24.
215. Braun H, Andrian-Werburg Jv, Schänzer W, Thevis M. Nutrition status of young
elite female German football players. Pediatr Exerc Sci 2018;30(1):157-67.
216. Vanheest JL, Rodgers CD, Mahoney CE, de Souza MJ. Ovarian suppression impairs
sport performance in junior elite female swimmers. Med Sci Sports Exerc
2014;46(1):156-66.
217. Chen Z, Sherk VD, Sharma-Ghimire P, Bemben MG, Bemben DA. Site-specific bone
differences and energy status in male competitive runners and road cyclists. J Clin
Densitom 2022;25(2):150-9.
218. McCormack WP, Shoepe TC, LaBrie J, Almstedt HC. Bone mineral density, energy
availability, and dietary restraint in collegiate cross-country runners and non-
running controls. Eur J Appl Physiol 2019;119(8):1747-56.
122
219. Ainsworth BE, Haskell WL, Herrmann SD, Meckes N, Bassett Jr DR, Tudor-Locke C,
Greer JL, Vezina J, Whitt-Glover MC, Leon AS. The compendium of physical
activities tracking guide: Version current 2011. Internet:
https://sites.google.com/site/compendiumofphysicalactivities/ (accessed 21
October 2022).
220. Schofield KL, Thorpe H, Sims ST. Resting metabolic rate prediction equations and
the validity to assess energy deficiency in the athlete population. Exp Physiol
2019;104(4):469-75.
221. Jagim AR, Camic CL, Kisiolek J, Luedke J, Erickson J, Jones MT, Oliver JM. Accuracy
of resting metabolic rate prediction equations in athletes. J Strength Cond Res
2018;32(7):1875-81.
222. Thompson J, Manore MM. Predicted and measured resting metabolic rate of male
and female endurance athletes. J Am Diet Assoc 1996;96(1):30-4.
223. ten Haaf T, Weijs PJ. Resting energy expenditure prediction in recreational athletes
of 18-35 years: confirmation of Cunningham equation and an improved weight-
based alternative. PLoS One 2014;9(9):e108460.
224. Cunningham JJ. A reanalysis of the factors influencing basal metabolic rate in
normal adults. Am J Clin Nutr 1980;33(11):2372-4.
225. Kasper AM, Langan-Evans C, Hudson JF, Brownlee TE, Harper LD, Naughton RJ,
Morton JP, Close GL. Come back skinfolds, all is forgiven: a narrative review of the
efficacy of common body composition methods in applied sports practice.
Nutrients 2021;13(4):1075.
226. Bazzocchi A, Ponti F, Albisinni U, Battista G, Guglielmi G. DXA: technical aspects
and application. Eur J Radiol 2016;85(8):1481-92.
227. Larson-Meyer DE, Woolf K, Burke L. Assessment of nutrient status in athletes and
the need for supplementation. Int J Sport Nutr Exerc Metab 2018;28(2):139-58.
228. Moon JR. Body composition in athletes and sports nutrition: an examination of the
bioimpedance analysis technique. Eur J Clin Nutr 2013;67(s1):54-9.
229. Ugras S. Evaluating of altered hydration status on effectiveness of body
composition analysis using bioelectric impedance analysis. Libyan J Med
2020;15(1):1741904.
230. Slinde F, Rossander-Hulthén L. Bioelectrical impedance: effect of 3 identical meals
on diurnal impedance variation and calculation of body composition. Am J Clin
Nutr 2001;74(4):474-8.
123
231. Saunders MJ, Blevins JE, Broeder CE. Effects of hydration changes on bioelectrical
impedance in endurance trained individuals. Med Sci Sports Exerc
1998;30(6):885-92.
232. Carrion BM, Wells A, Mayhew JL, Koch AJ. Concordance among bioelectrical
impedance analysis measures of percent body fat in athletic young adults. Int J
Exerc Sci 2019;12(4):324-31.
233. Buchholz AC, Bartok C, Schoeller DA. The validity of bioelectrical impedance
models in clinical populations. Nutr Clin Pract 2004;19(5):433-46.
234. Rodriguez-Giustiniani P, Rodriguez-Sanchez N, Galloway SDR. Fluid and electrolyte
balance considerations for female athletes. Eur J Sport Sci 2022;22(5):697-708.
235. Heyward VH, Wagner DR. Bioelectrical impedance analysis method. 2nd ed. In:
Applied body composition assessment. Champaign, IL: Human Kinetics, 2004:87-
98.
236. Ong JN, Ducker KJ, Furzer BJ, Dymock M, Landers GJ. Measures of body
composition via dual-energy X-ray absorptiometry, ultrasound and skinfolds are
not impacted by the menstrual cycle in active eumenorrheic females. J Sci Med
Sport 2022;25(2):115-21.
237. Cumberledge EA, Myers C, Venditti JJ, Dixon CB, Andreacci JL. The effect of the
menstrual cycle on body composition determined by contact-electrode
bioelectrical impedance analyzers. Int J Exerc Sci 2018;11(4):625-32.
238. Barr SI, Janelle KC, Prior JC. Energy intakes are higher during the luteal phase of
ovulatory menstrual cycles. Am J Clin Nutr 1995;61(1):39-43.
239. Cheikh Ismail LI, Al-hourani H, Lightowler HJ, Aldhaheri AS, Henry CJK. Energy and
nutrient intakes during different phases of the menstrual cycle in females in the
United Arab Emirates. Ann Nutr Metab 2009;54(2):124-8.
240. Reimer RA, Debert CT, House JL, Poulin MJ. Dietary and metabolic differences in
pre- versus postmenopausal women taking or not taking hormone replacement
therapy. Physiol Behav 2005;84(2):303-12.
241. Pliner P, Fleming AS. Food intake, body weight, and sweetness preferences over
the menstrual cycle in humans. Physiol Behav 1983;30(4):663-6.
242. Dalvit SP. The effect of the menstrual cycle on patterns of food intake. Am J Clin
Nutr 1981;34(9):1811-5.
243. Brennan IM, Feltrin KL, Nair NS, Hausken T, Little TJ, Gentilcore D, et al. Effects of
the phases of the menstrual cycle on gastric emptying, glycemia, plasma GLP-1 and
insulin, and energy intake in healthy lean women. Am J Physiol Gastrointest Liver
Physiol 2009;297(3):602-10.
124
244. Fong AK, Kretsch MJ. Changes in dietary intake, urinary nitrogen, and urinary
volume across the menstrual cycle. Am J Clin Nutr 1993;57(1):43-6.
245. McNeil J, Cameron JD, Finlayson G, Blundell JE, Doucet É. Greater overall olfactory
performance, explicit wanting for high fat foods and lipid intake during the mid-
luteal phase of the menstrual cycle. Physiol Behav 2013;112-113:84-9.
246. Cross GB, Marley J, Miles H, Willson K. Changes in nutrient intake during the
menstrual cycle of overweight women with premenstrual syndrome. Br J Nutr
2001;85(4):475-82.
247. Roney JR, Simmons ZL. Ovarian hormone fluctuations predict within-cycle shifts in
women's food intake. Horm Behav 2017;90:8-14.
248. Fessler DM. No time to eat: an adaptationist account of periovulatory behavioral
changes. Q Rev Biol 2003;78(1):3-21.
249. Hirschberg AL. Sex hormones, appetite and eating behaviour in women. Maturitas
2012;71(3):248-56.
250. García-Pinillos F, Bujalance-Moreno P, Jérez-Mayorga D, Velarde-Sotres Á, Anaya-
Moix V, Pueyo-Villa S, Largo-Fuentes C. Training habits of eumenorrheic active
women during the different phases of their menstrual cycle: a descriptive study.
Int J Environ Res Public Health 2021;18(7):3662.
251. McNamara A, Harris R, Minahan C. ‘That time of the month’ … for the biggest event
of your career! Perception of menstrual cycle on performance of Australian
athletes training for the 2020 Olympic and Paralympic Games. BMJ Open Sport
Exerc Med 2022;8(2):e001300.
252. Findlay RJ, Macrae EHR, Whyte IY, Easton C, Forrest LJ. How the menstrual cycle
and menstruation affect sporting performance: experiences and perceptions of
elite female rugby players. Br J Sports Med 2020;54(18):1108-13.
253. Thivel D, Metz L, Julian V, Isacco L, Verney J, Ennequin G, et al. Diet- but not
exercise-induced iso-energetic deficit induces compensatory appetitive responses.
Eur J Clin Nutr 2021;75(10):1425-32.
254. King NA, Lluch A, Stubbs RJ, Blundell JE. High dose exercise does not increase
hunger or energy intake in free living males. Eur J Clin Nutr 1997;51(7):478-83.
255. Cameron JD, Goldfield GS, Riou M, Finlayson GS, Blundell JE, Doucet É. Energy
depletion by diet or aerobic exercise alone: impact of energy deficit modality on
appetite parameters. Am J Clin Nutr 2016;103(4):1008-16.
256. Stubbs RJ, Hughes DA, Johnstone AM, Whybrow S, Horgan GW, King N, Blundell J.
Rate and extent of compensatory changes in energy intake and expenditure in
response to altered exercise and diet composition in humans. Am J Physiol Regul
Integr Comp Physiol 2004;286(2):350-8.
125
257. Larson-Meyer DE, Palm S, Bansal A, Austin KJ, Hart AM, Alexander BM. Influence of
running and walking on hormonal regulators of appetite in women. J Obes
2012;2012:730409.
258. Howe SM, Hand TM, Larson-Meyer DE, Austin KJ, Alexander BM, Manore MM. No
effect of exercise intensity on appetite in highly-trained endurance women.
Nutrients 2016;8(4):223.
259. Reed JL, Bowell JL, Hill BR, Williams BA, de Souza MJ, Williams NI. Exercising
women with menstrual disturbances consume low energy dense foods and
beverages. Appl Physiol Nutr Metab 2011;36(3):382-94.
260. Melin A, Tornberg ÅB, Skouby S, Møller SS, Faber J, Sundgot-Borgen J, Sjödin A.
Low-energy density and high fiber intake are dietary concerns in female
endurance athletes. Scand J Med Sci Sports 2016;26(9):1060-71.
261. Reed SC, Levin FR, Evans SM. Changes in mood, cognitive performance and
appetite in the late luteal and follicular phases of the menstrual cycle in women
with and without PMDD (premenstrual dysphoric disorder). Horm Behav
2008;54(1):185-93.
262. Cohen IT, Sherwin BB, Fleming AS. Food cravings, mood, and the menstrual cycle.
Horm Behav 1987;21(4):457-70.
263. Slonaker JR. The effect of copulation, pregnancy, pseudopregnancy and lactation
on the voluntary activity and food consumption of the albino rat. AJP-Legacy.
1925;71(2):362-94.
264. Eckel LA, Geary N. Endogenous cholecystokinin’s satiating action increases during
estrus in female rats. Peptides. 1999;20(4):451-6.
265. Eckel LA. The ovarian hormone estradiol plays a crucial role in the control of food
intake in females. Phsyiol Behav. 2011;104(4):517-24.
266. Blaustein JD, Wade GN. Ovarian influences on the meal patterns of female rats.
Physiol Behav. 1976;17(2):201-8.
267. Frank A, Brown LM, Clegg DJ. The role of hypthalamic estrogen receptors in
metabolic regulation. Front Neuroendocrinol. 2014;35(4):550-7.
268. Wira CR, Merritt K. Effect of the estrous cycle, castration and pseudopregnancy on
E. coli in the uterus and uterine secretions of the rat. Biol Reprod. 1977;17(4);519-
22.
269. Liu CY, Chen LB, Liu PY, Xie DP, Wang PS. Effects of progesterone on gastric
emptying and intenstinal transit time in healthy subjects. World J Gastroenterol.
2002;8(2):338-41.
126
270. Jung HK, Kim DY, Moon IH. Effects of gender and menstrual cycle on colonic transit
time in healthy subjects. Korean J Intern Med. 2003;18(3):181-6
271. Degen LP, Phillips SF. Variability of gastrointestinal transit in healthy women and
men. Gut. 1996;39(2):299-305.
272. Robinson TE, Berridge KC. The psychology and neurobiology of addiction: an
incentive-sensitization view. Addiction 2000;95(s2):91-117.
273. Leng G, Adan RAH, Belot M, Brunstrom JM, de Graaf K, Dickson SL, Hare T, Maier S,
Menzies J, Priessl H, et al. The determinants of food choice. Proc Nutr Soc
2017;76(3):316-27.
274. McKay AKA, Stellingwerff T, Smith ES, Martin DT, Mujika I, Goosey-Tolfrey VL,
Sheppard J, Burke LM. Defining training and performance caliber: a participant
classification framework. Int J Sports Physiol Perform 2022:17(2):317-31.
275. Clearblue. Digital ovulation test. Version current 3 August 2022. Internet:
https://nz.clearblue.com/ovulation-tests/digital (accessed 17 November 2022).
276. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc
1982;14(5):377-81.
277. The New Zealand Institute for Plant and Food Research Limited, New Zealand
Ministry of Health. New Zealand FOODfilesTM 2018 Version 01. Version current
2018. Internet: https://www.foodcomposition.co.nz/foodfiles/ (accessed 8 April
2022).
278. Weissgerber TL, Milic NM, Winham SJ, Garovic VD. Beyond bar and line graphs:
time for a new data presentation paradigm. PLoS Biol 2015;13(4):e1002128.
279. Benton MJ, Hutchins AM, Dawes JJ. Effect of menstrual cycle on resting
metabolism: a systematic review and meta-analysis. PLoS One
2020;15(7):e0236025.
280. Paludo AC, Paravlic A, Dvořáková K, Gimunová M. The effect of menstrual cycle on
perceptual responses in athletes: a systematic review with meta-analysis. Front
Psychol 2022;13:926854.
281. Bernstein MT, Graff LA, Avery L, Palatnick C, Parnerowski K, Targownik LE.
Gastrointestinal symptoms before and during menses in healthy women. BMC
Womens Health 2014;14:14.
282. Moore J, Barlow D, Jewell D, Kennedy S. Do gastrointestinal symptoms vary with
the menstrual cycle? Br J Obstet Gynaecol 1998;105(12):1322-5.
283. Carr-Nangle RE, Johnson WG, Bergeron KC, Nangle DW. Body image changes over
the menstrual cycle in normal women. Int J Eat Disord 1994;16(3):267-73.
127
284. Kamemoto K, Yamada M, Matsuda T, Ogata H, Ishikawa A, Kanno M, Miyashita M,
Sakamaki-Sunaga M. Effects of menstrual cycle on appetite-regulating hormones
and energy intake in response to cycling exercise in physically active women. J
Appl Physiol 2021;132(1):224-35.
285. Martin D, Sale C, Cooper SB, Elliott-Sale KJ. Period prevalence and perceived side
effects of hormonal contraceptive use and the menstrual cycle in elite athletes. Int J
Sports Physiol Perform 2018;13(7):926-32.
286. Rothschild JA, Kilding AE, Plews DJ. Prevalence and determinants of fasted training
in endurance athletes: a survey analysis. Int J Sport Nutr Exerc Metab
2020;30(5):345-56.
287. Reid RL. Premenstrual dysphoric disorder (formerly premenstrual syndrome).
South Dartmouth, MA: NIH/NLM/NCBI Books, 2000.
288. Bryant M, Truesdale KP, Dye L. Modest changes in dietary intake across the
menstrual cycle: implications for food intake research. Br J Nutr 2006;96(5):888-
94.
289. Czajkowska M, Drosdzol-Cop A, Gałązka I, Naworska B, Skrzypulec-Plinta V.
Menstrual cycle and the prevalence of premenstrual syndrome/premenstrual
dysphoric disorder in adolescent Athletes. J Pediatr Adolesc Gynecol
2015;28(6):492-8.
128
8. Appendices
Appendix A
Table 9 Studies assessing dietary energy intake of naturally menstruating females across different phases of the menstrual cycle with phases verified by
serum oestrogen and progesterone concentrations
Author &
Year
Population & Countrya
Menstrual cycle phases
and quantification
method
Number
of cycles
Dietary
assessment
method
Energy intakea
Macronutrient intake
Johnson
et al. 1994
(20)
n=26
Age: 32.3 ± 4.3 years
BMI: 21.5 kg/m2 b
United States
• Perimenstrual
• Ovulatory
• Luteal
BBT also used.
1
Daily
estimated
diet records
over one
cycle.
EI was greater in the
luteal vs. ovulatory phase
(1902 ± 452 vs. 1736 ±
427 kcal·day-1, mean
difference: 166 kcal·day-1
c, P<0.05d).
Absolute carbohydrate intake was
greater in the luteal (227.4 ± 60.8
g·day-1) and perimenstrual phases
(229.8 ± 60.9 g·day-1) compared
with the ovulatory phase (200.1 ±
45.5 g·day-1, P<0.05d), however
%TEI from carbohydrate remained
unchanged. Absolute fat intake was
greater in the luteal vs. ovulatory
phase (79.4 ± 23.4 vs. 70.2 ± 23.3,
g·day-1, P<0.05d), as well as %TEI
from fat (37.4 ± 5.1 % vs. 35.4 ± 5.5
%, P<0.05d). No evidence of a
difference in protein intake
between phases. Alcohol intake not
reported.
Martini et
al. 1994
(26)
n=18
Age: 26.9 ± 4.9 years
BMI: 22.2 ± 2.0 kg/m2
United States
• Mid-follicular: days
7-9
• Mid-luteal: 7-9 days
after a positive
urinary LH test
BBT also used. Only
serum progesterone was
measured in the luteal
4 (n= 2)
5 (n= 4)
6 (n=11)
3-day
estimated
diet records.
EI was greater in the mid-
luteal vs. mid-follicular
phase (1908 ± 38 vs. 1749
± 37 kcal·day-1 e, mean
difference: 159 kcal·day-1
c, P=0.003).
Absolute intakes of all
macronutrients were greater in the
mid-luteal vs. mid-follicular phase
(carbohydrate: 241.6 ± 5.4 vs. 226.3
± 5.3 g·day-1, P=0.04; fat: 74.0 ± 1.9
vs. 65.4 ± 1.9 g·day-1, P=0.002;
protein: 73.5 ± 1.8 vs. 67.4 ± 1.8
g·day-1, P=0.02), but there was no
129
phase, no oestrogen
measurements taken.
evidence of any differences in
%TEI. Participants were required
to abstain from alcohol
consumption for the duration of the
study.
Reimer et
al. 2005
(240)
n=9
Age: 27.9 ± 2.4 years
BMI: 23.9 ± 0.8 kg/m2
Canada
• Follicular
• Luteal
1
3-day
estimated
diet records.
EI was greater in the
luteal vs. follicular phase
(2089 ± 178 vs. 1752 ±
158 kcal·day-1, mean
difference: 337 kcal·day-1
c, P<0.05d).
No evidence of a difference in
absolute carbohydrate intake
between phases, however, %TEI
was lower in the luteal phase (48.1
± 2.6 vs. 53.2 ± 2.2 g·day-1, P<0.05d).
Both absolute fat intake (81.6 ± 8.9
vs. 56.9 ± 5.2 g·day-1, P<0.05d) and
%TEI fat (34.4 ± 2.5 vs. 28.8 ± 2.1
%, P<0.05d) were greater in luteal
vs. follicular phase. Absolute
protein intake was greater in the
luteal vs. follicular phase (80.6 ± 7.3
vs. 60.7 ± 5.3 g·day-1, P<0.05d), but
there was no evidence of a
difference in %TEI. No evidence of a
difference in alcohol intake
between phases. (g·day-1 or %TEI).
Reed et al.
2008
(261)
n=14 with PMDD
Age: 30.0 ± 6.7 years
BMI: 22.0 ± 2.1 kg/m2
n=15 control
Age: 30.0 ± 6.1 years
BMI: 22.0 ± 2.0 kg/m2
United States
• Follicular: days 6-10
• Luteal: 1-5 days
before menstruation
onset
Urinary LH used to
identify ovulation.
2
Intake based
on single ad-
libitum meal
(weighed by
researchers).
EI was 16% greater in the
luteal vs. follicular phase
for the PMDD group
(P=0.04), but there was
no evidence of a
difference in EI for the
control group (P>0.05d).
Overall, fat (P=0.008)f and protein
(P=0.03)f intakes were greater in
the luteal vs. follicular phase, driven
by changes in the PMDD group, who
consumed more fat (+7.2gg,
P≤0.05)f in the luteal phase. Alone,
there was no evidence of a
difference in protein or
carbohydrate intake for the PMDD
group, nor for carbohydrate, fat, or
protein in the control group.
Alcohol intake not applicable to the
ab-libitum meal, but no evidence of
a difference in “self-reported
alcohol use” between phases for
either group.
130
Brennan
et al. 2009
(243)
n=9
Age: 31.0 ± 1.0 years
BMI: 21.0 ± 0.5 kg/m2
Australia
• Follicular: twice
between days 6-12
• Luteal: days 18-24
1
Intake based
on single
provided
meal
(weighed by
researchers).
EI was greater in the
luteal phase (914 ± 113
kcal) compared with the
first and second follicular
phase time points (771 ±
84 kcal & 728 ± 82 kcal,
mean difference: 143 &
186 kcal respectivelyc,
P<0.05d).
Not reported.
Gil et al.
2009
(173)
n=30
Age: 27.7 ± 4.1 years
BMI: 22.2 ± 2 kg/m2
Brazil
• Follicular: days 5-9
• Luteal: days 20-25,
or days 23-28 for
those with cycles
>30 days
1
3-day
estimated
diet records
(including
one weekend
day).
EI was greater in the
luteal vs. follicular phase
(2259 ± 375 kcal·day-1 vs.
1730 ± 254 kcal·day-1,
mean difference: 529
kcal·day-1 c, P<0.001).
Absolute carbohydrate and fat
intakes were greater in the luteal
vs. follicular phase (carbohydrate:
298 ± 60 vs. 231± 45 g·day-1,
P<0.001; fat: 81 ± 18 vs. 55 ± 12
g·day-1, P<0.001). No evidence of a
difference in protein intake. Alcohol
intake not reported.
Chung et
al. 2010
(174)
n=39
Age: 20-40 yearsf
BMI: 21.1 ± 4.0 kg/m2
Taiwan
• Follicular: 3 days
after menses
• Ovulation: positive
LH test
• Luteal: 7-8 days after
positive LH test
Cycle history used to help
identify phase timing.
1
3-day
estimated
diet records
(one
weekend day
included
where
possible).
EI was greater during the
luteal vs. follicular phase
(1753 vs. 1593 kcal·day-1
b, mean difference: 160
kcal·day-1 c, P=0.03). For
those with correctly
identified cycle phases
(n=22), EI was greater in
the luteal vs. follicular
phase by 180 kcal·day-1 c
(P=0.04).
No evidence of a difference in
absolute carbohydrate between
phases, however %TEI from
carbohydrate was greater in the
follicular phase vs. luteal (+5%)f
and ovulatory phases (+4%)f. But
for those with correctly identified
cycle phases (n=22), %TEI
differences were no longer
significant. No evidence of a
difference in absolute fat intake or
%TEI from fat between phases.
Absolute protein intake was greater
in the luteal phase compared to
both other phases (57 vs. 55 and 49
g·day-1, P=0.01) for the whole
sample, with an increase of 8 g·day-
1 (P=0.01)f from the follicular to the
luteal phase for the subsample with
correctly identified cycle phases.
Alcohol intake not reported.
131
McNeil et
al. 2013
(245)
n=17
Age: 22.4 ± 3.2 years
BMI: 22.3 ± 1.7 kg/m2
Canada
• Early-follicular
• Late-
follicular/ovulatory
• Mid-luteal
Cycle history used to help
identify phase timing.
1
Intake based
on single ad-
libitum meal
(weighed by
researchers).
EI in the early-follicular,
late-follicular, and mid-
luteal phases were 670 ±
293 kcal, 525 ± 289 kcal,
and 711 ± 334 kcal
respectively (P=0.05 for
trend).
No evidence of a difference in
energy intake from carbohydrate or
protein between phases, but trend
for fat across the follicular, late-
follicular, and mid-luteal phases
(21.9 ± 18.1 vs. 18.7 ± 14.2 vs. 30.9
± 21.6 g, P=0.03 for trend)g. Alcohol
intake not applicable.
Gorczyca
et al. 2016
(24)
n=259
Age: 27.3 ± 8.2 years
BMI: 24.1 ± 3.9 kg/m2
United States
• Menstrual
• Mid-follicular
• Ovulation: estimated
using urinary LH
and estrone-3-
glucuronide
• Mid-luteal
1 (n=9)
2 (n=250)
24-hour
recalls.
No evidence of a
difference in EI between
the menstrual, mid-
follicular, ovulatory, or
mid-luteal phases for
neither ovulatory (1596 ±
28 vs. 1599 ± 28 vs. 1594
± 28 vs. 1656 ± 29
kcal·day-1 e, P=0.35) nor
anovulatory cycles (1641
± 100 vs. 1500 ± 95 vs.
1543 ± 100 vs. 1749 ±
117 kcal·day-1 e, P=0.36).
For ovulatory cycles, absolute
“energy-adjusted” protein intake
was greater in the mid-luteal phase
vs. peri-ovulatory phase (65 ± 1.0
vs. 61 ± 0.9 g·day-1, P=0.04)e, as well
as %TEI from protein (16.4 ± 0.3 vs.
15.4 ± 0.2 %TEI, P=0.02) e. For
anovulatory cycles, there was no
evidence of a difference in absolute
“energy-adjusted” intakes or %TEI
from carbohydrate, fat, or protein.
No evidence of a difference in
alcohol intake between phases for
neither ovulatory nor anovulatory
cycles.
Ihalainen
et al. 2021
(108)
n=15 recreational
athletes
Age: 26 ± 4 years
Body fat: 22.1 ± 6.7%
Finland
• Early-follicular: days
2-4
• Mid-follicular: 7-11
days post
menstruation onset
• Ovulation: positive
urinary LH test
• Mid-luteal phase: 7
days after positive
urinary LH test
1
3-day
estimated
diet records.
No evidence of a
difference in EI between
the early-follicular, mid-
follicular, ovulatory, or
mid-luteal phases (2340 ±
660 vs. 2340 ± 540 vs.
2280 ± 510 vs. 2270 ±
370 kcal·day-1, P=0.825).
No evidence of a
difference in energy
availability between
phases (40.0 ± 11.1 vs.
39.9 ± 11.1 vs. 35.9 ± 9.0 ±
37.6 ± 7.2
No evidence of a difference in
absolute carbohydrate, fat, or
protein intake between phases.
Alcohol intake not reported.
132
kcal·kgFFM-1·day-1,
P=0.465)
aValues represent mean ± SD unless otherwise specified
bSD not reported
cMean difference not reported and calculated from absolute values reported in each phase, or mean difference reported without SD of the difference
dExact p-value not reported
eValues represent mean ± SE
fActual values not reported
gValues reported as kcal and converted to grams
hRepresents age range of participants, mean ± SD not reported
Abbreviations: BBT basal body temperature, BMI body mass index, EA energy availability, EI energy intake, LH luteinising hormone, PMS premenstrual syndrome,
PMDD premenstrual dysphoric disorder, TEI total energy intake
133
Appendix B
Table 10 Studies assessing dietary energy intake of naturally menstruating females across different phases of the menstrual cycle with out verification of
phases with serum estrogen and progesterone concentrations
Author &
Year
Populationa
Menstrual cycle phases
and quantification
method
Number of
cycles
Dietary
assessment
method
Energy intakea
Macronutrient intake
Dalvit
1981
(242)
n=8
Age: 18-22 yearsb
United States
• Follicular: the 10
days post-
menstruation
• Luteal: the 10 days
pre-menstruation
2
Daily 24-hour
recalls for 60
days.
EI was greater in the luteal
vs. follicular phase for both
cycle 1 (1935 vs. 1431
kcal·day-1 c, mean
difference: 504 ± 219
kcal·day-1, P=0.0004) and
cycle 2 (1945 vs. 1449
kcal·day-1 c, mean
difference: 496 ± 378
kcal·day-1, P=0.008).
Not reported.
Dalvit-
McPhillips
1983 (21)
n=8
Age: 18-22 yearsb
United States
• Follicular: the 10
days post-
menstruation
• Luteal: the 10 days
pre-menstruation
2
Daily 24-hour
recalls for 60
days.
Not reported.
Absolute carbohydrate intake was
greater in the luteal vs. follicular
phase (256.9 ± 51.7 vs. 133.3 ±
28.1 g·day-1). No evidence of a
difference in absolute fat or
protein intake. Alcohol intake not
reported.
Pliner &
Fleming
1983
(241)
n=41
Age: 20.5 ± 4.0 years
Canada
• Mid-point of the
follicular phase
(phase length
calculated by
subtracting 14 days
from total cycle
duration)
• Mid-point of the
luteal phase
(assumed to be 14
days)
Based on cycle history.
1
24-hour
recalls.
EI was greater in the luteal
vs. follicular phase (2013 ±
533 vs. 1790 ± 642
kcal·day-1, mean
difference: 223 kcal·day-1 d,
P<0.05e)
Not reported.
134
Manocha
et al. 1986
(176)
n=11
Age: 22-30 yearsb
India
• Follicular: 10-days
post-menstruation
• Luteal: 10-days
pre-menstruation
Retrospective recall of
menstruation after 60
days.
2
Daily
estimated
diet records
for 60 days.
EI was greater in the luteal
vs. follicular phase for both
cycle 1 (1620 ± 275 vs.
1300 ± 290 kcal·day-1,
mean difference: 320
kcal·day-1 d, P<0.05e) and
cycle 2 (1605 ± 270 vs.
1300 ± 255 kcal·day-1,
mean difference: 305
kcal·day-1 d, P<0.01e).
Not reported.
Gong et al.
1989
(177)
n=7
Age: 31.4 ± 7.0 years
BMI: 22.4 kg/m2
United States
• Menstrual: days 1-
4
• Follicular: day 5 –
periovulatory
phase
• Periovulatory: 4
days spanning
estimated
ovulation
• Luteal: the day
following the
periovulatory
phase – the day
before menses
1
Daily
weighed food
records for
one cycle.
EI was greater in the luteal
vs. follicular phases (2040
± 156 vs. 1833 ± 146
kcal·day-1, mean
difference: 207 kcal·day-1 d,
P<0.05e), and in the luteal
vs. periovulatory phase
(2040 ± 156 vs. 1766 ±
252 kcal·day-1, mean
difference: 274 kcal·day-1 d,
P<0.05e).
Not reported.
Lyons et
al. 1989
(22)
n=18
Age: 23.6 yearsc
BMI: 22.8 kg/m2 c
Australia
• Menses: days 1-4
• Post-menses: day 5
to ovulatory phase
• Ovulatory: 4 days
surrounding
ovulation (as
determined by
positive urinary LH
test)
• Post-ovulatory: 4
days after the
ovulatory phase
1
Daily
weighed-diet
records for
one cycle.
EI was lower in the
ovulatory (1874 ± 81
kcal·day-1) vs. post-
ovulatory (2198 ± 86
kcal·day-1, mean
difference: 324 kcal·day-1 d,
P<0.05e), pre-menses
(2150 ± 86 kcal·day-1,
mean difference: 276
kcal·day-1 d, P<0.05e) and
menses (2155 ± 100
kcal·day-1, mean difference:
281 kcal·day-1 d, P<0.05e).
Absolute protein intake was lower
in ovulatory vs. post-ovulatory
phase (63.0 ± 4.0 vs. 73.0 ± 2.0
g·day-1, P<0.05e)f, and absolute
carbohydrate intake was lower in
ovulatory vs. post-ovulatory and
pre-menses phases (235.0 ± 14.0
vs. 270 ± 12.0 and 263.0 ± 13.0
g·day-1, P<0.05e)f, however, there
was no evidence of a difference in
%TEI of any macronutrient
between the five phases
measured. No evidence of a
135
• Pre-menses:
between post-
ovulatory and
menses
There was no evidence of a
difference in EI between
10 days pre- and post-
menstruation (2133
kcal·day-1 vs. 2102
kcal·day-1 c, e).
difference in alcohol intake
between phases.
Tarasuk &
Beaton
1991 (25)
n=14
Age: 32 ± 9.5 years
BMI: 23.0 ± 5.7 kg/m2
United States
• Follicular: 10 days
following onset of
menstruation
• Luteal: 10 days
preceding
menstruation
Retrospectively
reported at end of the
year-long study.
1 (n=1)
2 (n=3)
3 (n=1)
4 (n=9)
Daily
weighed diet
records.
EI was greater in the luteal
vs. follicular phase (1912
vs. 1822 kcal·day-1 c, mean
difference: 90 ± 38
kcal·day-1, P=0.03).
Absolute fat intake was greater in
the luteal vs. follicular phase (79.8
vs. 72.9 g·day-1 c, P=0.0102), and
when adjusted for energy intake
(41.3 vs. 39.5 g·1000 kcal·-1day-1 c,
P=0.0040). No evidence of a
difference in absolute or energy-
adjusted carbohydrate or protein
intakes. Alcohol intake not
reported.
Fong &
Kretsch
1993
(244)
n=9
Age: 28.1 ± 4.1 years
BMI: 22.4 ± 2.0 kg/m2
United States
• Menses: days of
menstrual bleeding
• Follicular: between
menses and
periovulatory
• Periovulatory: 2
days either side of
ovulation (as
determined by
BBT)
• Luteal: between
periovulatory and
menses
1
Daily
weighed diet
records for
one cycle (by
researchers
in the
metabolic
ward study).
No evidence of a difference
in EI between menses
(2045 ± 468 kcal·day-1),
follicular (2027 ± 443
kcal·day-1), periovulatory
(1968 ± 516 kcal·day-1),
and luteal phases (2204 ±
475 kcal·day-1 e).
Absolute carbohydrate was
greater during menses vs. peri-
ovulatory phase (303 ± 59 vs. 267
± 81 g·day-1, P=0.06). No evidence
of a difference in absolute fat or
protein intakes between phases.
Alcohol intake not reported.
Barr et al.
1995
(238)
n=29 ovulatory cycles
Age: 27.9 ± 5.3 years
BMI: 21.8 ± 2.1 kg/m2
n=13 anovulatory
cycles
• Follicular: days 3-8
• Luteal: days 20-26
BBT used to identify
anovulatory cycles.
3
3-day
weighed diet
records.
EI was greater during the
luteal vs. follicular phase
for ovulatory cycles (2248
± 652 vs. 1942 ± 572
kcal·day-1, mean
difference: 305 kcal·day-1 d,
No evidence of a difference in
%TEI from carbohydrate, fat, or
protein for neither ovulatory nor
anovulatory cycles. Alcohol intake
not reported.
136
Age: 26.5 ± 4.8 years
BMI: 21.6 ± 2.3 kg/m2
Canada
P<0.001), but there was no
evidence of a difference for
anovulatory cycles (1918 ±
529 vs. 1990 ± 359
kcal·day-1 e).
Li et al.
1999
(172)
n=20
Age: 21.2± 1.3 years
BMI: 19.6 ± 1.4 kg/m2
China
• Mid-follicular: days
6-10
• Mid-luteal: 6-10
days after positive
urinary LH test
1
3-day
estimated
diet records
(not
necessarily
consecutive,
but including
one weekend
day).
EI was greater in the mid-
luteal vs. mid-follicular
phase (1692 ± 448 vs.
1478 ± 285 kcal·day-1,
mean difference: 214
kcal·day-1 d, P=0.02). There
was a 23% increase in EI in
the luteal vs. follicular
phase based on weekend
records (1749 ± 574.4 vs.
1419 ± 417 kcal·day-1,
mean difference: 330
kcal·day-1 d, P=0.018), but
no evidence of a difference
in EI for weekday records
(1663 ± 507 vs. 1508 ±
357 kcal·day-1, mean
difference: 155 kcal·day-1 d,
P=0.149).
Absolute intakes of carbohydrate
and fat were greater in the luteal
vs. follicular phase (carbohydrate:
218 ± 62 vs. 189 ± 40 g·day-1,
P=0.05; fat: 58 ± 23 vs. 48 ± 11
g·day-1, P=0.04) but there was no
evidence of a difference in %TEI.
No evidence of a difference in
absolute intake or %TEI from
protein. Alcohol intake not
reported.
Cross et
al. 2001
(246)
n=82 with PMS
Age: 37.2 ± 5.2 years
BMI: 29 ± 3.6 kg/m2
n=40 control
Age: 37.2 ± 6.9 years
BMI: 29.2 ± 3.1 kg/m2
Australia
• Follicular: day 5-8
• Luteal: the 4 days
pre-anticipated
menstruation
2
4-day
estimated
diet records.
EI was greater in the luteal
vs. follicular phase for both
groups, but the PMS group
had a larger difference
(1467 ± 307 vs. 2097± 419
kcal·day-1, mean
difference: 603 kcal·day-1 d,
P<0.001), compared with
the control group (1802 ±
411 vs. 1914 ± 326
kcal·day-1, mean
difference: 112 kcal·day-1 d,
P<0.05e). There was a
weak correlation between
For the PMS group, absolute
intakes of carbohydrate, fat, and
protein all increased from the
follicular phase to the luteal phase
(carbohydrate: 167 ± 36.3 vs. 245
± 50.1 g·day-1, P<0.01; fat: 53.6 ±
15 vs. 81.6 ± 23.5 g·day-1, P<0.01;
protein: 65.6 ± 16 vs. 78.7 ± 17.6
g·day-1, P<0.01). From the
follicular to the luteal phase, there
was an increase in %TEI from
carbohydrate (44.6c vs. 45.6 ±
5.0%, P<0.05e) and fat (32.6 ± 4.6
vs. 34.6 ± 4.8%, P<0.05e) and
137
the severity of PMS
symptoms and energy
intake changes
(Spearman’s correlation =
0.29).
%TEI from protein decreased
(18.5 ± 3.2 vs. 15.7 ± 2.9%,
P<0.05e). For the control group,
absolute fat intake increased from
follicular to the luteal phase (74.2
± 21 vs. 67 ± 21 g·day-1, P<0.05e),
but there was no evidence of a
difference in %TEI. No evidence of
a difference in absolute intake or
%TEI of carbohydrate or protein
intake. Alcohol intake not
reported.
Bryant et
al. 2006
(288)
n=31 with PMS
Age: 34.5 ± 5.4 years
BMI: 23.8 ± 4.2 kg/m2
n=27 control
Age: 34.0 ± 7.7 years
BMI: 23.1 ± 3.1 kg/m2
United States
• Follicular: days 4-6
• Luteal: 6-4 days
before
menstruation
(based on cycle
history)
1
3-day
estimated
diet records.
No evidence of a difference
in EI between the follicular
and luteal phases for
neither those with PMS
(2032 vs. 2009 kcal·day-1
c,e) nor the control group
(1981 vs. 2080 kcal·day-1
c,e).
No evidence of a difference in
%TEI from carbohydrate, fat or
protein for neither the PMS nor
control group. No evidence of a
difference in alcohol intake
between phases for either group.
Cheikh
Ismail et
al. 2009
(239)
n=43
Age: 22.0 ± 3.0 years
BMI: 22.0 ± 3.9 kg/m2
United Arab Emirates
• Pre-menstrual: 2
days within 10-
days pre-
menstruation
• Menstrual: days 1-
2
• Post-menstrual: 2
days within 10-
days post menses
onset
1
2-day
estimated
diet records.
EI was higher in the pre-
menstrual vs. menstrual
phase (1363 ± 550 vs.
1126 ± 462 kcal·day-1,
mean difference: 237
kcal·day-1 d, P=0.002).
Not reported.
Elliot et al.
2015
(175)
n=31
Age: 23.7 ± 1.3 yearsf
BMI: 20.2 ± 2.8 kg/m2 f
Singapore
• Menstrual: 0-24%
• Follicular: 25-49%
• Luteal: 51-100%
Menstrual cycles were
normalized with phases
expressed as %, where
1 (n=18)
2 (n=13)
3-day
estimated
diet record
(including
one weekend
day).
EI was greater in the
menstrual vs. luteal phase
(1687 ± 419 vs. 1404 ±
311 kcal·day-1, mean
difference: 283 kcal·day-1 d,
P<0.05e). For those who
completed two cycles
Absolute fat intake was lower in
the luteal vs. follicular phase (57.3
vs. 51.0 g·day-1, P=0.029)h across
two cycles, but there was no
evidence of a difference in %TEI of
any macronutrient across two
cycles. Alcohol intake not
138
0% represents the first
day of menstrual
bleeding, 50%
represents ovulation,
and 100% represents
the day before
menstrual bleeding.
(n=13), there was no
evidence of a difference in
EI between the menstrual,
follicular and luteal phases
(1494 ± 352 vs. 1544 ±
326 vs. 1589 ± 354
kcal·day-1, P=0.07).
reported.
Roney &
Simmons
2017
(247)
n=24
Age: 18.9 yearsc
United States
Daily salivary
oestrogen and
progesterone used to
estimate the date of
ovulation.
1
Daily survey
on the
quantity
eaten,
meal size,
restriction,
and hunger
relative usual
on a 1-5 scale
for one cycle.
Decrease in reported food
intake in the days
approaching ovulation,
with the lowest intake
corresponding to peak
salivary oestrogen
(P=0.024). Reported food
intake rose concomitantly
with increasing salivary
progesterone in the luteal
phase (P<0.01e).
Not reported.
de Souza
et al. 2018
(23)
n=27
Age: 21.9 ± 0.5 years
BMI: 23.4 ± 0.9 kg/m2
Brazil
• Follicular: days 5-9
• Luteal: days 20-25
1
24-hour
recalls.
No evidence of a difference
in EI between the luteal
and follicular phases (1738
± 414 vs. 1694 ± 437
kcal·day-1, mean
difference: 44 kcal·day-1 d,
P=0.383).
No evidence of a difference in
absolute carbohydrate, fat, or
protein intake between phases.
Alcohol intake not reported.
aValues represent mean ± SD unless otherwise specified
bRepresents age range of participants, mean ± SD not reported
cSD not reported
dMean difference not reported and calculated from absolute values reported in each phase, or mean difference reported without SD of the difference
eExact p-value not reported
fValues represent mean ± SE
gValues for age and BMI only reported for those who completed two cycles
hActual values not reported, means calculated from reported %TEI and energy intake
Abbreviations: BBT basal body temperature, BMI body mass index, EI energy intake, LH luteinising hormone, PMS premenstrual syndrome, TEI total energy intake
139
Appendix C
140
141
142
143
144
Appendix D
145
146
147
Appendix E
148
149
Appendix F
150
Appendix G
151
152
153
Appendix H
154
Appendix I
155
Appendix J
156
157
158
159
Appendix K
160
Appendix L
Ad-libitum post-exercise clinic meal food list
• Standard full-fat milk
• Trim milk
• Oat milk
• Soy milk
• Apple & orange juice
• Greek-style plain yoghurt
• Coconut yoghurt
• Apples
• Frozen blueberries
• Canola-oil-based table spread
• Raspberry jam
• Crunchy peanut butter
• Honey
• Nutella
• Marmite
• Maple-flavoured syrup
• Chocolate chips
• Pumpkin seeds
• Dates
• Tinned peaches
• Baked beans
• Nutri-grain
• Toasted muesli
• Rolled oats
• All-bran
• Cornflakes
• Blueberry muffins
• Croissants
• White bread
• Wheatmeal bread
• Vogel’s soy & linseed bread
• Crumpets
• Sugar
• Drinking chocolate
• Salt & Pepper
• Instant coffee
• Chocolate Up & Go
• Dairy-free chocolate up & go
161
Appendix M
162
Appendix N
Figure 12 Timeline of participant movement throughout the study1,2
1Each horizontal line represents a different participant
2Figure was created with Microsoft Excel 2016 (Microsoft Corporation, Redmond,
United States) and Adobe Photoshop 2020 (Adobe Inc., San Jose, United States)
Abbreviations: EF early-follicular phase, ML mid-luteal phase
