Not all ‘women’ are the same. Women have a variety of reproductive hormonal profiles that change across the lifespan from puberty to the menopause. |
The endogenous hormonal profile of women is frequently influenced by exogenous sources, such as hormonal contraceptives (HC) and hormone replacement therapy (HRT). |
Depending on the research question, women should be recruited on pre-defined, standardised criteria, which, in most cases, should be retrospectively confirmed (i.e., homogenous a priori inclusion and a posteriori exclusion criteria). |
Depending on the research question, the experimental design needs to be adapted in line with the hormonal milieu (e.g., consideration of menstrual cycle phase, type of HC used, stage of menopause). |
1 Introduction
1.1 Statement of the Problem
1.2 Current State of the Art
1.3 The Gap in the Field
1.4 Background to the Statement
1.5 Approach to the Statement
1.6 Purpose of the Statement
2 Considerations
2.1 Participants
Consideration | Rationale (intended to…) | Pros (could…) | Cons (could…) |
---|---|---|---|
Define puberty as the onset of menarche [50] | Increase accuracy and validity of the population definition | Reduce between study variability in describing the population studied | Increase timescale of the study in order to recruit participants who fit the criteria |
Define peri-puberty as the time around puberty | Need to be aware of other physical and endocrine indicators of puberty outside of menarche (e.g., maturation of the genital organs, development of secondary sex characteristics; Tanner stages; oestrogen and progesterone concentrations) | ||
Define adrenarche as activation of adrenal androgen production; usually occurs before gonadarche [51] | Increase homogeneity of hormonal profiles Increase consistency of terminology | Reduce between participant variability in hormone status Reduce between study variability in describing the population studied | Reduce availability of eligible participants Increase timescale of the study if the condition needs to be confirmed prior to the commencement of data collection Increase number of participants who need to be excluded (during or retrospectively) from the study if the condition was not confirmed prior to commencement of the study |
Define gonadarche as activation of reproductive glands leading to menarche [51] | |||
Define eumenorrhea as menstrual cycle lengths ≥ 21 days and ≤ 35 days resulting in 9 or more consecutive periods per year, plus evidence of LH surge, plus correct hormonal profile, plus no HC use 3 months prior to recruitment | |||
Define luteal phase deficiency as cycles with less than 16 nmol·L−1 of progesterone, when a single luteal phase progesterone measurement is taken [48] | |||
Define oligomenorrhea as those with cycle length > 35 days [54] | |||
Define naturally menstruating women as those who experience menstruation, with menstrual cycle lengths ≥ 21 days and ≤ 35 days, but without confirmed ovulation [ovulation was not confirmed by urinary LH surge or verified by hormone concentrations via blood sample analysis] | Increase accuracy and validity of the population definition | Reduce unfounded assumption that ovulation, and thus eumenorrhea, has been established Reduce between study variability describing the population studied | Increase timescale of the study in order to recruit participants who fit the criteria |
A priori exclusion of participants with self-reported or diagnosed menstrual irregularities for eumenorrheic studies. Menstrual irregularities refer to perturbations of the eumenorrheic menstrual cycle, such as amenorrhea, anovulation, oligomenorrhea etc | Increase homogeneity of hormonal profiles | Reduce between participant variability in hormone status Increase validity of the data | Reduce availability of eligible participants for eumenorrheic studies |
A posteriori exclusion of participants with observed or implied menstrual irregularities for eumenorrheic studies | Increase number of participants who need to be excluded (during or retrospectively) | ||
No HC use ≥ 3 months prior to recruitment for study on eumenorrheic participants | Increase likelihood that an eumenorrheic cycle and its typical hormonal profile has been re-established | Reduce occurrence of atypical hormonal profiles not fitting the eumenorrheic definition | Reduce availability of eligible participants for eumenorrheic studies Increase timescale of the study as the condition needs to be met prior to recruitment |
HC use ≥ 3 months prior to recruitment for HC studies | Increase likelihood that eumenorrheic cycle has been removed and replaced by a hormonal profile indicative of HC use | Reduce occurrence of atypical hormonal profiles not fitting the characterisation of HC users | |
Define HC users as those taking any type of contraceptive capable of altering the endogenous hormonal milieu. Please note that HC also influence other aspects of metabolism, which are beyond the scope of this paper | Increase accuracy and validity of the population definition | Reduce between study variability in describing the population studied | N/A |
Report the type (e.g., OCPs, implants, injections, intrauterine devices/coils that are hormone releasing and NOT copper-based, vaginal rings, contraceptive transdermal patches) and formulation (e.g., mono, bi or triphasic; combined or progesterone-only; names and concentration of exogenous hormones) of HC used | Increase reliability of studies Increase validity of findings | Reduce between participant variability in hormone status Reduce between study variability in describing the population studied | Increase number of participants who need to be excluded as they do not know the exact type or formulation of HC used |
One brand/type of OCP per group of participants [56]. Clearly identify exogenous hormone names and concentrations in each OCP, as well as androgenicity | Increase homogeneity of hormonal profiles (both endogenous and exogenous) | Increase timescale of the study whilst trying to recruit a sufficient sample size on the same brand/type of OCP Reduce generalisability of the findings | |
Define the first trimester as first 13 weeks of pregnancy [54] | Increase accuracy and validity of the population definition | Increase timescale of the study in order to recruit participants who fit the criteria | |
Define the second trimester as the time from week 14 to week 27 of pregnancy [54] | Reduce between participant variability in hormone status Reduce between study variability in describing the population studied | Increase timescale of the study in order to recruit participants who fit the criteria | |
Define the third trimester as the time from 27 weeks of pregnancy onwards [54] | |||
Define full term as when a pregnancy is a normal duration (i.e., 37–42 weeks gestation) [54] | |||
Define postpartum as the 12 months following parturition | |||
State gestation (i.e., the length of time, in days or weeks, that a baby is in the uterus) | Reduce between study variability in describing the population studied | Increase number of participants who need to be excluded as they do not know their exact gestational stage | |
State gravidity (i.e., number of times that a woman has been pregnant, including miscarriages and abortions) | Increase timescale of the study in order to recruit women willing to state this number (can be a sensitive issue) | ||
State parity (i.e., number of times that a woman has given birth to a foetus with a gestational age of 24 weeks or more, regardless of whether the child was born alive or was stillborn) | |||
State singleton or multiple pregnancy | |||
Define peri-menopause as the time around the occurrence of the menopause when the ovaries gradually produce less oestrogen [57] | Increase accuracy and validity of the population definition | Reduce between study variability in describing the population studied | Need to be aware of other physical indicators of menopause (e.g., hot flushes, vaginal dryness, emotional changes) |
Define menopause as the time when menstruation surceases; i.e., characterised by sporadic amenorrhea. *Please note that this is the transitional time between peri- and post-menopause | Increase timescale of the study in order to recruit participants who fit the criteria | ||
Define post-menopause as the time after which a woman has experienced 12 consecutive months of amenorrhea and is characterised by < 118 pmol∙L−1 of oestrogen, < 4.4 nmol∙L−1 of progesterone and follicle stimulating hormone > 25 IU∙L−1 [57] | Increase homogeneity of hormonal profiles Increase consistency of terminology | Reduce between participant variability in hormone status Reduce between study variability in describing the population studied | Reduce availability of eligible participants Increase timescale of the study if the condition needs to be confirmed prior to the commencement of data collection Increase number of participants who need to be excluded (during or retrospectively) if the condition was not confirmed prior to commencement of the study |
Consider menopausal symptoms [58] | To limit the effects of menopausal symptoms on outcomes | Reduce the likelihood that effects are indirectly due to symptoms rather than directly due to changes in hormones | Increase the time burden to identify participants with no, or a consistent set of, menopausal symptoms capable of affecting the intended outcome |
Define HRT users as those taking any type of HRT capable of altering the endogenous hormonal milieu (e.g., tablets, skin patches, gels, implants, vaginal creams, pessaries or rings; combined or oestrogen only; cyclical or continuous) | Increase accuracy and validity of the population definition | Reduce between study variability in describing the population studied | N/A |
Report the type and formulation of HRT used | Increase reliability of studies Increase validity of findings | Reduce between participant variability in hormone status Reduce between study variability in describing the population studied | Increase number of participants who need to be excluded as they do not know the exact type or formulation of HRT used |
2.2 Experimental Design
Consideration | Rationale (intended to…) | Pros (could…) | Cons (could…) |
---|---|---|---|
Take into account the changes in androstenedione prior to the onset of puberty (e.g., the initial increase in androstenedione has been noted 18 to 12 months prior to the onset of puberty) [60] | Increase breadth of data on reproductive ageing by considering the peri-pubertal period | Increase understanding of the peri-pubertal changes in physiological functioning and athletic performance | Increase timescale of the study in order to identify and group participants along these spectrums |
Take into account the changes in oestrogen prior to the onset of puberty (e.g., the initial increase in oestrogen has been noted 12 and 6 months prior to the onset of puberty) [60] | |||
Take into account the time scale of establishing a eumenorrheic cycle: menarche follows an anovulatory cycle; menstrual cycles during the 1st year after menarche are typically irregular and anovulatory, ranging in duration from 21 to 45 days; by 3 years post-menarche, > 90% of girls have ≥ 10 menstrual cycles per year with an average menstrual interval of 36.5 days; cycles can remain irregular until the 5th year post-menarche [61] | Reduce assumption that once menarche has been initiated all girls have fully eumenorrheic cycles | Reduce between participant variability in hormone status | |
Increase reliability of studies Increase validity of findings | Reduce likelihood of grouping non-homogenous hormonal profiles Reduce inconsistency in phase definitions between studies | Increase timescale of the study in order to recruit participants who are willing to undertake blood sampling Increase cost of the study | |
Track and establish menstrual cycle characteristics for ≥ 2 months prior to testing. Tracking can be achieved by denoting the first and the last day of menstruation on a calendar for each cycle. Corroboration can be achieved by confirmation of ovulation and hormone concentrations | Reduce within participant variation in menstrual cycle characteristics Reduce likelihood of including participants with menstrual irregularities in eumenorrheic studies Increase ability to accurately predict testing timepoints (i.e., phases) | Increase timescale of the study due to the long lead-in time Increase burden on participants to track their cycles before the experimental aspect of the study | |
Outcome measures should be repeated in a second cycle | Reduce variability of the data | Increase timescale of the study due to the repeated measures Increase burden on participants to repeat all of the testing sessions | |
Reduce risk of including anovulatory women in eumenorrheic studies Reduce chance of a false positive result by the participant from an at-home interpretation | Increase chance of missing a positive ovulation result in participants who do not comply or adhere to conducting the test at the same time of day Increase likelihood of overlooking LPD as this method does not exclude LPD cycles | ||
Ensure that the intended reproductive profiles were assessed | Reduce likelihood of grouping non-homogenous hormonal profiles | Increase number of participants who need to be excluded (retrospectively) as a result of not fitting the inclusion criteria | |
Stipulate and take into account OCP-taking (i.e., active OCP) days and OCP-free (i.e., inactive/placebo OCP) days: (i) The endogenous concentration of oestrogen and progesterone rises during the OCP free/inactive/placebo days [65, 66] (ii) The concentration of exogenous hormones increases during active OCP intake: for example, for a combined monophasic OCP ethinyl estradiol (a type of exogenous oestrogen) increases twofold from day 1 of active OCP to day 21 [66, 67] and progestin increases threefold from day 1 of active OCP to day 8–11 and then maintains that level [67, 68] | Increase homogeneity of hormonal profiles | Reduce between participant variability in hormone status | Increase timescale of the study if several conditions need to be assessed |
Take into account the rising concentrations of oestrogen and progesterone throughout each trimester of pregnancy | Increase timescale of the study in order to identify and group participants along this spectrum | ||
Take into account the large variation in hormonal profiles associated with the peri-menopause, menopause and post-menopause, thus treating these as separate categories of women based on the criteria outlined in Table 1 |
Recommendation | Rationale (intended to…) | Pro | Con |
---|---|---|---|
Phase 1: indicated by the onset of bleeding until day 5 Oestrogen and progesterone levels are low | Capture the lowest concentrations of oestrogen and progesterone | Easy to determine due to obvious physical cue (i.e., bloody discharge) | Can be difficult to predict in those with variable cycle length therefore requiring reactive testing sessions (i.e., participant alerting the researcher on day 1 of bleeding and then both parties having availability for testing within the next 4 days) |
Phase 2: occurs in the 14–26 h prior to ovulation and the LH surge Oestrogen higher than during phase 1, 3 and 4 and progesterone higher than during phase 1, but lower than 6.36 nmol·L−1 | Capture the highest oestrogen concentration, while progesterone remains low | Enables the biggest difference between oestrogen and progesterone to be investigated | Difficult to predict without daily blood samples for the determination of oestrogen and progesterone |
Phase 3: indicated by a positive urinary ovulation kit and lasts 24–36 h Oestrogen higher than phase 1 but lower than phase 2 and 4 and progesterone higher than phase 1 but lower than 6.4 nmol·L−1 | Capture a medium oestrogen concentration, while progesterone remains low | Easy to establish due to the positive LH surge captured by the ovulation kit | Relies on having multiple ovulation kits available for each participant (cost) and requires reactive testing sessions (i.e., participant alerting the researcher to the positive result and then both parties having availability for testing within the next 24–36 h) |
Phase 4: + 7 days after ovulation has been confirmed Oestrogen higher than phase 1 and 3 but lower than phase 2 and progesterone > 16 nmol·L−1 | Capture the highest concentration of progesterone and a high concentration of oestrogen | Easy to establish in those with eumenorrheic cycles as it typically occurs within 7 days of confirmed ovulation | Relies on the confirmation of ovulation |
2.3 General Guidance
Consideration | Rationale (intended to…) | Pros (could…) | Cons (could…) |
---|---|---|---|
Employ a single-blind design; although participants cannot be blinded, the researcher can be blind to the intended testing timepoint | Protect against bias | Reduce unintentional bias from the researcher to the participant | Increase staffing as an independent person is needed to undertake the blinding process |
Do not use unclassified (i.e., not stating reproductive status) women as participants even if you are not concerned by the potential influence of reproductive hormones on your outcome measure | Increase homogeneity of hormonal profiles | Reduce between participant variability in hormone status Increase validity of the data | Reduce the availability of eligible participants |
Standardise, by quantitative means, time of day, prior exercise, caffeine ingestion, dietary intake and nutritional supplementation, alcohol consumption and smoking as these have been shown to affect the concentration of reproductive hormones | Control extraneous variables | Allow the investigation of X on Y, without the influence of Z | Adds more requirements on the participants by asking them to standardise a large number of variables over a set period of time |
Use the term “withdrawal bleed” rather than “period” when referring to the bleed experienced by OCP users | Stop OCP users misidentifying themselves as eumenorrheic based on bleeding patterns | Education; informing women and researchers about the differences between hormonal contraceptive users and non-users | |
Do not use the terms “menstrual cycle” and “periods” synonymously | Dispel the myth that they are the same things and that these terms can be used interchangeably | Education; informing women and researchers about menstrual cycles and how periods are just one aspect of that cycle | |
Do not impose any menstrual cycle language upon HC users (i.e., trying to match certain days of the OCP cycle with eumenorrheic phase descriptions, e.g., calling the first five days of OCP taking as the early follicular phase of the menstrual cycle) | Prevent confusion between hormonal contraceptive users and non-users | Allows HC users to describe their own status without adding unnecessary complexity or without misperception | |
Report gestational age based on ultrasound dating rather than on last menstrual period | Increase the accuracy of reporting of gestational age | Reduce the ambiguity in defining participants in studies involving pregnant women | |
Consider the timescale for resumption of eumenorrheic cycles following childbirth (i.e., in the postpartum period), given that this varies considerably between women | Reduce the assumption that all postpartum women who do not use hormonal contraceptives have eumenorrheic cycles | Reduce ambiguity in defining participants in studies involving postpartum women | |
Do not use the term post-menopausal based on participants’ age solely | Protect against including irrelevant participants and to increase the homogeneity of hormonal profiles | Reduce between participant variability in hormone status Increase validity of the data | Increase the time and economic burden to correctly identify and confirm post-menopausal status |
Do not report gynaecological age (i.e., number of years from menarche to recruitment in the study) as a characteristic of menstrual function (i.e., to illustrate the number of years with eumenorrheic menstrual cycles) | Reduce the assumption that the time between menarche and recruitment is filled with eumenorrheic cycles | Reduce ambiguity in defining participants in studies involving eumenorrheic women | |
Include an online supplement with additional in-depth information about reproductive status; e.g., data from questionnaires on menstrual cycle status or hormonal contraceptive use, blood marker data, etc | Provide data that can be used for future meta-analyses in studies with women as participants | Quickly increase our understanding of female physiology in relation to sport and exercise science |
3 Discussion
3.1 Definition of ‘Woman’
-
Example 1; a researcher wishing to investigate potential effects of the menstrual cycle on exercise performance needs to decide: if they will use eumenorrheic or naturally menstruating women (see Table 1 for delineation of terminology); how they will confirm eumenorrhea; when in the cycle they will test to ensure different hormonal environments.
-
Example 2; a researcher wishing to use oral contraceptive pill (OCP) users will need to decide which role they will fulfil (i.e., control or experimental sample [56]):
-
OCP users can be used as a control group, wherein cyclical fluctuations in oestrogen and progesterone experienced during the menstrual cycle can be compared to consistently downregulated sex hormone levels. The researcher then needs to decide whether to compare with the active pill phase or the non-active pill phase. For example, if the non-active pill phase is compared with the early follicular phase, then it is likely that the hormonal environments will be very similar between the two groups. If the mid-luteal phase is compared with the active pill phase, the hormonal environments between groups will be different. The researcher will also need to consider the type of OCP being used (i.e., oestrogen and progestin formulations and concentrations; monophasic, biphasic, triphasic) and the duration of usage, ensuring this is similar for the whole control group of OCP users.
-
OCP users can also be considered as an experimental group, wherein (i) the effects of low endogenous oestrogen and progesterone levels can be investigated and/or (ii) the effects of exogenous oestrogen and progestin can be examined. In this case the researcher needs to decide which hormonal environments to compare, early in the non-active pill phase (with low endogenous and exogenous hormone levels), late in the non-active pill phase (with increasing endogenous oestrogen and low exogenous hormones), early in the active pill phase (with increasing exogenous hormone levels) or late in the active pill phase (with highest exogenous hormone levels). Again, the researcher will need to consider the type of OCP being used and ensure this is similar for the whole control group of OCP users and/or consider comparison between different types of OCP.