Article Text

Download PDFPDF

Prevalence and frequency of menstrual cycle symptoms are associated with availability to train and compete: a study of 6812 exercising women recruited using the Strava exercise app
  1. Georgie Bruinvels1,2,
  2. Esther Goldsmith2,
  3. Richard Blagrove3,
  4. Andrew Simpkin4,5,
  5. Nathan Lewis1,2,
  6. Katie Morton2,
  7. Ara Suppiah6,
  8. John P Rogers7,8,
  9. Kathryn E Ackerman9,
  10. John Newell4,5,
  11. Charles Pedlar1,2,10
  1. 1 Faculty of Sport, Health and Applied Science, St Mary's University Twickenham, Twickenham, London, UK
  2. 2 Orreco, Galway, Ireland
  3. 3 School of Sport, Exercise and Health Sciences, Loughborough University, Loughborough, UK
  4. 4 School of Mathematics, Statistics and Applied Mathematics, National University of Ireland Galway, Galway, Ireland
  5. 5 Insight Centre for Data Analytics, National University of Ireland, Galway, Ireland
  6. 6 University of Central Florida, Orlando, Florida, USA
  7. 7 Sports Medicine, Manchester University NHS Foundation Trust, Manchester, Greater Manchester, UK
  8. 8 Manchester Institute of High Performance, Manchester, UK
  9. 9 Female Athlete Program, Division of Sports Medicine, Boston Children's Hospital, Boston, Massachusetts, USA
  10. 10 Institute of Sport, Exercise and Health, UCL, London, UK
  1. Correspondence to Dr Georgie Bruinvels, Faculty of Sport, Health and Applied Science, St Mary's University Twickenham, Twickenham, London, UK; georgie.bruinvels{at}


Objectives The menstrual cycle can affect sports participation and exercise performance. There are very few data on specific menstrual cycle symptoms (symptoms during various phases of the cycle, not only during menstruation) experienced by exercising women. We aimed to characterise the most common symptoms, as well as the number and frequency of symptoms, and evaluate whether menstrual cycle symptoms are associated with sporting outcomes.

Methods 6812 adult women of reproductive age (mean age: 38.3 (8.7) years) who were not using combined hormonal contraception were recruited via the Strava exercise app user database and completed a 39-part survey. Respondents were from seven geographical areas, and the questions were translated and localised to each region (Brazil, n=892; France, n=1355; Germany, n=839; Spain, n=834; UK and Ireland, n=1350; and USA, n=1542). The survey captured exercise behaviours, current menstrual status, presence and frequency of menstrual cycle symptoms, medication use for symptoms, perceived effects of the menstrual cycle on exercise and work behaviours, and history of hormonal contraception use. We propose a novel Menstrual Symptom index (MSi) based on the presence and frequency of 18 commonly reported symptoms (range 0–54, where 54 would correspond to all 18 symptoms each occurring very frequently).

Results The most prevalent menstrual cycle symptoms were mood changes/anxiety (90.6%), tiredness/fatigue (86.2%), stomach cramps (84.2%) and breast pain/tenderness (83.1%). After controlling for body mass index, training volume and age, the MSi was associated with a greater likelihood of missing or changing training (OR=1.09 (CI 1.08 to 1.10); p≤0.05), missing a sporting event/competition (OR=1.07 (CI 1.06 to 1.08); p≤0.05), absenteeism from work/academia (OR=1.08 (CI 1.07 to 1.09); p≤0.05) and use of pain medication (OR=1.09 (CI 1.08 to 1.09); p≤0.05).

Conclusion Menstrual cycle symptoms are very common in exercising women, and women report that these symptoms compromise their exercise participation and work capacity. The MSi needs to be formally validated (psychometrics); at present, it provides an easy way to quantify the frequency of menstrual cycle symptoms.

  • female
  • exercise physiology
  • sports and exercise medicine
  • women in sport

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


In the female population, menstrual pain is the most common gynaecological issue,1 reducing quality of life, limiting work capacity, and causing absenteeism from school, work and social activities.2–4 Physical activity provides a means to manage menstrual cycle symptoms (MCS), with aerobic exercise and yoga in particular demonstrably reducing menstrual cycle-related pain.5 Within the athletic population, 51%–93% of athletes self-report performance detriments or negative experiences broadly associated with menstrual cycle.6–8 However, the specific reasons for performance detriments have not been investigated. In the general population, up to 90% of women experience premenstrual symptoms and 20%–90% experience premenstrual syndrome (PMS).9 10 Yet the prevalence and impact of symptoms related to the menstrual cycle (MCS, including symptoms at any time in the cycle) have not previously been evaluated in the athletic population, and there is no existing method for quantifying MCS.

The two most common conditions characterised by pain associated with the menstrual cycle are PMS and primary dysmenorrhoea (PD). Due to the lack of universal consensus on diagnosis, the epidemiology of both conditions varies significantly. PMS involves a range of physical and/or psychological symptoms in the days prior to menstruation.11 PD is a term used to describe chronic and cyclic pain just before and/or during menstruation, where there is no underlying pathology evident.12 Some can also experience pain around ovulation.13

The aims of the present study were to (1) describe MCS among exercising women; (2) evaluate the relationship between MCS and negative events, including the need to miss or change training/competition, use of pain medication and missing work/lectures; and (3) quantify the prevalence and occurrence of symptoms with a novel Menstrual Symptom index (MSi).

Materials and methods

Study population and survey dissemination

The survey was disseminated via email from the Strava membership database (including only women who opted in to receiving emails of this nature) and through the Strava application via a hyperlink. The inclusion criteria were women aged 18 years and older (mean age: 38.3 (8.7) years). The derivation of the sample for this study is shown in figure 1. A total of 425 697 women received an invitation to complete an online survey (180 000 via email, 245 697 via the Strava mobile app). Of this sample, 16 423 (3.9%) initiated the survey and 10 371 (2.4%) completed the survey. From start to end, the completion rate was 63.1%. The survey was open for 25 days and launched on 14 February 2019 until 11 March 2019. The participants reported partaking in a wide range of exercises, including running, swimming, cycling, dance classes, team sports, gym-based classes, racket sports, use of a cross-trainer or similar cardio-based exercise machine, martial arts, weight training, and other forms of prolonged exercise.

Figure 1

Sampling plan and number of participants. Dorado: ‘in app’ communications (via the Strava application). HC, hormonal contraception.

Research design

An online survey was created (; SurveyMonkey, London) to evaluate the menstrual characteristics, symptoms and exercise habits of a cross-sectional sample of exercising women. The survey was designed to take approximately 10 min to complete. Following feedback from a pilot survey (n=20; ages 18–40 years) to help evaluate face validity and general clarity around questions, minor alterations were made to the question order and grammar. The final survey comprised 39 questions.

The survey included questions on the following topics: current and previous exercise behaviours (including self-reported training volume and intensity); current menstrual status; presence and frequency of MCS; use of pain medication for MCS (eg, non-steroidal anti-inflammatory drugs); effects of the menstrual cycle on exercise behaviours and work capacity; hormonal contraception use; effects of exercise on MCS; communication and education around the menstrual cycle; aspects of diet and sleep; and iron status. The survey was translated to French, Spanish, Brazilian Portuguese, German and American English. It was then localised to seven countries (UK and Northern Ireland, Ireland, USA, France, Spain, Italy and Germany) by a native language speaker to ensure correct meaning and colloquialism, and subsequently piloted within these countries for affirmation. All questions were multiple choice, checkboxes with or without a limit on the number of selections allowed, dropdown of answer choices, a matrix or a slider. Free text answers were not requested except for instances where an ‘Other’ was applicable (on nine occasions). Logic was applied to the survey to ensure that only relevant questions were asked.

Participants provided informed consent prior to commencing the survey.

Survey analysis

The raw data were exported from SurveyMonkey directly to Microsoft Excel software for Mac. Participants who were using combined hormonal birth control (containing both exogenous oestrogen and progesterone; n=1285; 12.4%), those who indicated being perimenopausal or postmenopausal (n=1781; 17.2%), and those who indicated that they have been pre or post partum (n=493; 4.8%) in the last year were removed. Those using combined hormonal birth control were excluded from this data set due to the known systemic metabolic and inflammatory effects of exogenous hormones.14 As progestin-only birth control options do not exhibit the same systemic metabolic response, users of this type of hormonal contraception were included in the analysis.14 This resulted in a total subanalysis population of 6812 (figure 1).

Menstrual Symptom index

Participants were asked to report their experience of symptoms. Specific symptoms listed were a combination of those examined in previous work.15 To calculate the MSi, a Likert scale was used based on the frequency of each symptom. Where an individual reported a symptom as ‘often’, it was given 3 points, ‘sometimes’ 2 points, ‘rarely’ 1 point and ‘never’ 0 point. Points were then summed to create the athlete’s MSi. Because there were 18 symptoms, total scores ranged from 0 (minimum) to 54 (every symptom, often). The number of symptoms reported regardless of the frequency was also added to create the ‘total number of symptoms’, a value out of 18. Quantifying menstrual symptom frequency this way was not formally validated.

Women were also asked to indicate if MCS led to them missing training/sporting event, missing work or using pain medication, using the following scale: ‘no, never’, ‘yes, rarely’, ‘yes, sometimes (every few cycle)’, or ‘yes, often (every cycle)’.

Statistical analysis

All analyses were performed using R (V.3.5.1). Numerical and graphical summaries were generated to describe the target population of interest and to estimate the number and frequency of symptoms experienced in the cohort of exercising women by country of origin.

A separate logistic regression model was fitted to evaluate the association between the MSi and the probability of missing/changing training, missing a sporting competition/event, missing work/lectures and use of pain medication, while adjusting for body mass index (BMI), training volume and age. The OR for each variable and the accompanying 95% CI were calculated.


The demographics of the exercising women who met the inclusion criteria are shown in table 1.

Table 1

Demographics of the sample of exercising women (N=6812) by country

Prevalence and frequency of reporting each symptom

The prevalence and reported frequency of each symptom are shown in figure 2. The most common symptom regardless of frequency was ‘mood changes/anxiety’, reported by 90.6% of the total sample. ‘Mood changes/anxiety’ was also the symptom most commonly reported as ‘often’ (44.1% of the sample). The next most common symptoms reported as ‘often’ were ‘cravings/increased appetite’ (37.6%), ‘breast pain/tenderness’ (36.7%) and tiredness/fatigue (33.1%).

Figure 2

Stacked bar chart of the frequency of the different types of symptoms experienced by exercising women (N=6812) and the frequency of the occurrence of symptoms.

A number of symptoms were also cited in the ‘Other’ free text box. Symptoms cited by more than one individual and were not duplicated from the other symptoms included heavy/achy legs (n=9), depression-like symptoms (n=7), pelvic/uterine/ovarian pain (n=6), increased sweating (n=5), loss of appetite (n=3), clumsiness (n=2), feelings of getting cold/influenza-like symptoms (n=2), urinary incontinence (n=2), increased thirst (n=2) and increased heart rate (n=2).

Total number of symptoms and MSi

The mean number of MCS experienced was 11.3 (95% CI 11.21 to 11.41; figure 3A,) while the mean MSi was 22.9 (95% CI 22.66 to 23.18; figure 3B).

Figure 3

(A) Bar chart showing the number of menstrual cycle symptoms reported by each exercising woman (N=6812). (B) Bar chart of the distribution of the Menstrual Symptom index scores based on the self-reported menstrual cycle symptoms and the frequency of experiencing these by each exercising woman (N=6812).

Association between MSi and missing training, missing competition or work, and use of pain medication

Table 2 shows the effect of MSi on the probability of missing/changing training, missing a sporting event/competition or missing work/lectures, while adjusting for a woman’s BMI, age and training volume as covariates. The ORs for the MSi provide an estimate of the change in odds for the corresponding response variable per unit increase in MSi score. A higher MSi (when adjusting for age, BMI and training volume) was significantly associated with missing/changing training, missing a sporting event/competition, missing work/lectures and pain medication use. We estimate that the odds of missing training are multiplied by 1.09 per unit increase in MSi.

Table 2

Estimated OR and 95% CI for the (adjusted) effect of MSi on missing training, a sporting event/competition or work

Figure 4 shows the association between MSi and the probability of missing/changing training, missing a sporting event/competition, missing work/lectures and pain medication use. Women are more likely to miss training or use pain medication when compared with missing a competition or work.

Figure 4

Fitted probabilities of missing/changing training, missing an event/competition, missing work/lectures and using pain medication based on the Menstrual Symptom index (MSi) score. P(outcome), probability.


In our study, the most commonly reported MCS included mood changes/increased anxiety, cravings/increased appetite, breast pain/tenderness and increased tiredness/fatigue. Across the population, 81.1% of exercising women reported at least one symptom ‘often’, on average reported to have experienced 11.3 (4.4) MCS out of a maximum of 18, and had an MSi (factoring in prevalence and occurrence) of 22.9 (11.0) out of a maximum of 54. Experiencing a greater number of MCS more frequently and therefore having higher MSi was associated with a greater likelihood of negative outcomes, including changing/missing exercise training, missing a sporting event/competition, missing work/lectures and needing to use pain medication as a form of treatment.

Previous research in the general female population has found that up to 90% of women experience symptoms around menses,16 and that an increased number of symptoms is associated with loss of productivity.15 However, this is the first study to investigate MCS in the exercising demographics and to ascertain potential repercussions associated with these. A major strength of the present study is the large cohort spread across seven geographical territories.

When compared with non-athletes, athletes are more likely to have an irregular menstrual cycle, while also seemingly more likely to suffer certain types of menstrual dysfunction.17 18 There has been a historical focus on menstrual dysfunctions pertaining to low energy availability in athletes, including amenorrhoea and oligomenorrhoea19; however, little research has evaluated MCS and the impact that these may have on exercise participation and performance among the broader exercising population.

Our data highlight that MCS are common in regularly exercising women, and that MCS are a cause for women to miss or change exercise training and competition, to be absent from work/lectures and to use pain medication for treatment. In the general population, research conducted in a Japanese cohort (n=19 254) found an association between the presence and severity of MCS and both a reduction in productivity and with absenteeism from work.14 15 20 In the present study, 81.1% of women reported experiencing at least one symptom ‘often’, which is comparable with previous data in the general population showing up to 90% experience MCS.21 Indeed, some studies have shown that exercise helps with management of MCS; therefore, it could be hypothesised that those who exercise are at reduced risk of MCS. A pilot study demonstrated that vigorous exercise reduced pain associated with menstruation,22 and practising yoga and aerobic exercise have also been found to reduce menstrual pain and menstrual distress.5 23 24 The mechanism for the potential ‘treatment’ effect of exercise in eumenorrhoeic women has yet to be elucidated and may be multifactorial, but is likely dependent on the aetiology of the specific MCS. Two possible hypotheses for the ‘treatment’ effect include the exercise-induced release of endorphins25 and the known anti-inflammatory effects of regular moderate exercise.26

In light of the high number of exercising women who experience MCS, and the impact that MCS can have on everyday life, it is important to gain a better understanding of the aetiology of MCS and its potential risk factors. This will aid in a more targeted approach to management and treatment strategies. There are a number of different mechanisms for the manifestation of symptoms. Those hypothesised include changes to neurobiology, resulting in alterations to the serotonergic and gamma-aminobutyric acid systems,27 and the impact of changes in the release of inflammatory markers28 29 and reactive oxygen species,30 in addition to hormone fluctuations, sensitivities to hormonal changes31 and an overproduction of prostaglandins.12 32 Previous studies have identified several risk factors that are associated with the prevalence of both PD and PMS.11 These include diet, BMI, lifestyle and particular psychological factors.11 33 While lifestyle, diet and psychological components were beyond the scope of the present study, those with lower BMI did have fewer and less frequent symptoms. However, it is important to note that the symptoms may have been historical, but we do not have corresponding historical BMI data, so this finding must be interpreted with caution. Future research should focus on intervention-based randomised controlled trials to accurately evaluate potential non-pharmacological and medical interventions among exercising women to enable more strategic dietary, lifestyle, psychological and medical interventions.

Disruptions to mood states that are related to the menstrual cycle have been shown widely and were the most common symptoms reported by the exercising women in this study. Recent work has shown increased mood disturbances in the mid-luteal phase when compared with the mid-follicular phase, postulating a number of potential mechanisms, including the impact of progesterone-mediated changes to dopaminergic signalling.34 Interestingly, exercise abolished this disturbance.34 Given most exercising women report mood changes/anxiety, research should focus on the management of these symptoms.

Menstrual Symptom index

We quantified the prevalence and frequency of MCS using the MSi; this tool does not capture the severity of MCS. The association between MSi and negative outcomes provides some content validity, and we report an important first step ahead of formal tool development and evaluation.

Note that women were more likely to miss/change training or use pain medication than they were to miss work or miss a sporting event/competition. Even with the MSi above the 95th percentile, the probability of missing a sporting event or competition was less than 0.5 (50%), yet for the same MSi the probability of missing/changing training or using pain medication is almost 0.9 (90%). This suggests that women are likely to continue to compete even with significant MCS. A recent study examining the experiences of elite athletes concluded that fewer negative experiences occur during competition than during training,6 suggesting that women can over-ride or ignore MCS in certain scenarios. Previous research has also highlighted the reluctance of women to seek medical support to help manage MCS. Given the significant impact these are clearly having on everyday life, more resources and open discussion are clearly needed.4

Harking back to our comment that while this tool measures MCS frequency, it does not measure MCS severity. Future studies should interrogate our tool and consider an extended tool that captures symptom severity as well as frequency. Galileo said: “measure what is measurable and make measurable what is not so.” We appreciate the limitations of MSi version 1.0 and we encourage investigators to develop future iterations.

Clinical implications

Medical professionals and athletic support staff need to be aware of the prevalence and impact that MCS can have on exercising women. In light of the evident potential for associated negative outcomes, screening for MCS is warranted. In an applied setting, given the potential MCS aetiologies, where a large number of MCS are reported by exercising women, extra consideration is warranted when planning training, a need for recovery and overall readiness. Establishing ways to mitigate MCS is also necessary, although efficacy may be individual-specific. The derivation of an MSi to quantify the frequency and number of symptoms is novel and could pave the way for the future development of an MCS screening tool. This could also be used as a means to monitor the efficacy of non-pharmacological and medical treatment options, as frequent pain medication use can have negative long-term effects.35 36 In the present study, the most common symptoms experienced by exercising women were mood changes/increased anxiety, cravings/increased appetite, breast pain/tenderness and increased tiredness/fatigue, so interventions and an understanding of the mechanism behind these should be prioritised.

Strengths and limitations

We believe this to be the largest study to date of exercising women, including women across seven different geographical territories, and this is a clear strength of this research. Considering the individual nature of each menstrual cycle and the different potential aetiologies driving symptoms, key timepoints for symptoms are likely to vary. Therefore, in the present study, we purposefully captured symptoms at all times in the menstrual cycle to identify the need for research and medical focus/raised awareness. However, as a result, we were unable to identify the key times in the menstrual cycle where specific symptoms were experienced. When working on an individual athlete basis, it is important to capture this information to fully support their individual needs.

In the present study, we excluded those using combined hormonal contraceptives (containing an exogenous oestrogen and a progestin). These act by downregulating endogenous hormones through inhibition of gonadotrophin-releasing hormone, luteinising hormone and follicle stimulating hormone,31 having systemic metabolic and inflammatory effects.37 Combined hormonal contraception in use worldwide have many different formulations. These can be very helpful in treating menstrual symptoms, but can also cause unwanted side effects that may be similar to the symptoms evaluated here. Therefore, we excluded these participants to avoid risk of confounding the results. Due to the negligible systemic metabolic disturbances observed with progestin-only options,37 these users were included in the analysis.

There are at least five limitations to our study.

  • All of the data are self-reported, so despite all participants being identified as premenopausal the data are reliant on memory recall.

  • The presence of any underlying medical issue was not controlled for in the analysis. Various medical issues may result in worse symptoms. However, the aim of the current study was to specifically understand the symptoms experienced by exercising women, so this does not negate the significance here.

  • We appreciate that this survey was limited to Strava users, who may be more likely to (1) be involved in recreational to elite-level exercise; (2) be willing to invest in their health and wellness; and (3) have access to suitable technology in order to complete the survey. Such factors may add an element of bias and may suggest that the true burden of MCS is not captured here. However, it could also be argued that those with menstrual symptoms may be more compelled to complete this survey.

  • While the women in this study were deemed to be ‘exercising women’, there were no minimal exercise criteria for inclusion. This was largely due to potential short-term alterations in exercise behaviours that would not capture an accurate picture. The performance level of the participants was not captured, and clearly this has the potential to alter whether a sporting event/competition and/or training will be missed.

  • We note that a very extensive set of MCS were provided for women to choose from and this could affect the validity of MSi.

What are the findings?

  • Menstrual cycle symptoms are reported to have a detrimental effect on exercise and work capacity.

  • The most common symptoms experienced were mood changes/anxiety, tiredness/fatigue and stomach cramps.

  • A scoring system (Menstrual Symptom index; MSi) was created based on the occurrence and frequency of symptoms.

  • A higher MSi was associated with an increased likelihood of negative outcomes, such as missing training or competition.

How might it impact on clinical practice in the future?

  • Medical professionals should consider screening exercising women for menstrual cycle symptoms.

  • Where a large number of menstrual cycle symptoms are present, or the symptoms appear to impact performance, training and work, clinicians could pay more attention to planning training and recovery accordingly.

  • MSi may aid clinicians in their screening of sportswomen and help them evaluate interventions being used to treat symptoms.



  • Twitter @gbruinvels, @EGoldsmithPhys, @rich_blagrove, @drkateackerman

  • Contributors GB, CP, KM and EG contributed to the design of the study. GB, AnS and JN contributed to the statistical analysis of data. GB, CP, KM, EG, AnS, JN, KEA, NL, RB, JPR, KM and ArS all contributed to the interpretation of the results and the writing of the manuscript.

  • Funding A small grant was received from Strava to conduct the survey.

  • Competing interests GB, EG, NL, KM, ArS and CP are employees or consultants for Orreco, creators of the FitrWoman app.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Patient consent for publication Not required.

  • Ethics approval The study protocol was approved by the Ethics Committee of St Mary’s University (SMEC_2018-19_011), Twickenham, UK.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement The data that support the findings of this study are available on request from the corresponding author (GB;