Key findings & implications
-
There is inconsistency in information recommending nutritional interventions to women experiencing a natural menstrual cycle when aiming to reduce menstrual-related symptoms, resulting from the wide range of symptoms and unknown mechanism of cause. We recommend treating symptoms on an individual case basis, adapting and personalising the nutrition intervention for the individuals’ symptoms and dietary/supplement preferences.
-
Curcumin, vitamin D, calcium, magnesium and zinc may reduce menstrual-related symptoms; however, further investigation is required, considering the mechanisms and reporting individual responses.
-
Future studies should focus on investigating nutritional strategies to manage menstrual-related symptoms, ensuring DRV and safe upper limits of supplementation are adhered to.
Introduction
The menstrual cycle is a biological rhythm, whereby cyclic fluctuations in endogenous sex hormones such as oestrogen and progesterone are observed (Reference Davis, Hackney and Hackney1–Reference Owen3). The changes in hormone levels and associated interactions with the brain and wider bodily functions can cause a variety of menstrual cycle related symptoms.
Worldwide estimations of pre-menstrual symptom prevalence range from 10 to 98%(Reference del Mar Fernández, Saulyte, Inskip and Takkouche4), with up to 80% of women experiencing at least one symptom of menstrual related symptoms during their lifetime(Reference Hofmeister and Bodden5). Pre-menstrual symptoms can reduce quality of life, limit work capacity and cause absenteeism from school, work and social activities(Reference De Sanctis, Soliman and Elsedfy6–Reference Tanaka, Momoeda and Osuga8). The most common symptom is primary dysmenorrhoea, otherwise known as cramping pain in the lower abdomen before or during menstruation without any evident disease or pathology(Reference Schoep, Nieboer, van der Znden, Braat and Nap9). The pain associated with dysmenorrhoea, can also reduce mood and sleep quality compared with the pain-free phase after menstruation(Reference Lacovides, Avidon and Baker10). Other common symptoms include changes to mood, bloating, food cravings, fatigue and breast pain(Reference Bruinvels, Goldsmith and Blagrove11). However, many women do not consult a doctor whilst experiencing menstrual symptoms, which has been attributed to medication avoidance(Reference Schoep, Nieboer, van der Znden, Braat and Nap9).
The exact aetiology of menstrual related symptoms is unclear, there are multiple theories suggesting the involvement of several hormones released with ovulation, diets with nutritional deficiencies, family medical history, progesterone and GABA neurotransmitter aberrations(Reference Itriyeva12) and circulating gonadal steroids(Reference Dilbaz and Aksan13). Alternatively, inflammation has been proposed as a potential mechanism as medications such as non-steroidal anti-inflammatory drugs are typically used to reduce menstrual-related symptoms such as abdominal cramps. High-sensitivity C-reactive protein is an acute phase inflammatory marker which has been positively associated with pre-menstrual mood, abdominal/back pain, food cravings, weight gain, bloating and breast pain but not menstrual headaches(Reference Gold, Wells and O’Neill Rasor14). In some cases, suppression of ovarian hormone secretion attenuates menstrual related symptoms, although differences in ovarian steroid hormones have not consistently been observed between symptomatic and asymptomatic individuals(Reference Gold, Wells and O’Neill Rasor14). It has been acknowledged that biological, social, demographic and behavioural factors have been associated with menstrual related symptoms(Reference Halbreich, Endicott and Lesser15–Reference Gold, Bair and Block17) and due to the broad spectrum of symptoms and potential mechanisms, symptom relief is a main challenge and effective treatments are limited(Reference Siminiuc and Turcanu18).
Hormonal contraceptives (HC) are frequently used to manage symptoms caused by fluctuating reproductive hormones(Reference Schindler19). HC are exogenous steroid hormones that inhibit ovulation and result in consistent low endogenous sex hormones(Reference Wong, Farquhar, Roberts and Proctor20,Reference Imai, Matsunami, Takagi and Ichigo21) . Suppression of ovarian hormone secretion has markedly attenuated menstrual related symptoms(Reference Thys-Jacobs22). HC inhibit hypothalamic gonadotrophin releasing hormone which prevents pituitary secretion of follicle stimulating hormone and luteinising hormone. Contraceptives provide pharmacological control of the reproductive cycle by consistently promoting a negative feedback loop to prevent endogenous oestrogen or progesterone release(Reference Sims and Heather23). As synthetic hormones, their actions differ from endogenous hormones, the differences in mechanisms of which extend beyond the reproductive system and contradict responses and actions of other bodily systems such as cardiovascular and metabolic systems(Reference Sims and Heather23), including but not limited to oxidative stress and inflammatory response(Reference Cauci, Xodo, Buligan, Colaninno, Barbina, Barbina and Francescato24). Therefore, although HC may manage menstrual related symptoms, individuals have been hesitant to use this approach due to the synthetic hormones associated and unknown effects on other systems beyond the reproductive system, or frequently experiencing side effects of the synthetic hormones – commonly mistaken for menstrual related symptoms.
Following a healthy diet and managing stress have been reported as important factors in naturally preventing and managing menstrual related symptoms, specifically fresh, unprocessed foods, avoiding foods rich in refined carbohydrates or fats, salt, alcohol and simulating beverages has been reported(Reference Aguilar25). Whilst other research has found no correlation between macronutrients and menstrual related symptoms(Reference Houghton, Manson, Whitcomb, Hankinson, Troy, Bigelow and Bertone-Johnson26), reports of micronutrients, such as zinc, to provide anti-inflammatory and neurotrophic factors have been stated(Reference Siminiuc and Turcanu18). Further dietary interventions could ameliorate inflammation(Reference Zhu, Du and Xu27) such as fruits, vegetables and food legumes which contain high levels of phytochemicals that show anti-inflammatory effects(Reference Mrityunjaya, Pavithra, Neelam, Janhavi, Halami and Ravindra28), whilst several other nutraceuticals have been investigated that may reduce menstrual related symptoms, these include vitamin D, vitamin C and curcumin(Reference Zhu, Du and Xu27).
Currently there is no treatment universally recognised to prevent or manage menstrual-related symptoms. Given the ease of implementation of non-pharmaceutical strategies to manage menstrual symptoms(Reference Tomlinson29), and potential solution it offers to prevent symptoms disrupting daily life, nutritional interventions may provide a viable option. However, to the authors knowledge, no review exits that provides practical recommendations and guidance to practitioners or individuals on food, nutrition or supplements for the purpose of menstrual cycle symptom management(Reference Taylor30).
Based on the above reasoning, the aim was to systematically review all studies investigating changes in menstrual-related symptoms in eumenorrhoeic women using foods or supplements as interventions for management.
Methods
This review conforms to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement guidelines(Reference Page, McKenzie and Bossuyt31).
Study inclusion and exclusion criteria
Considerations of Population, Intervention, Outcomes and Study design were used to determine the parameters within which the review was conducted(Reference Higgins, Thomas, Chandler, Cumpston, Li, Page and Welch32):
Population
Participants included healthy women who were (a) naturally menstruating (menstrual cycle length >21 d and <35 d resulting in nine or more consecutive periods in a year;(Reference Elliott-Sale, Minahan and Janse de Jonge33) (b) not using HC, (c) free from menstrual-related dysfunctions (such as amenorrhoea, endometriosis, polycystic ovarian syndrome) or any other condition that could affect menstrual function (e.g. pregnancy, eating disorder or disordered eating), (d) not overweight or obese, and (e) not using hormone treatment/therapy.
Individuals using HC were excluded from this review due to differences in mechanisms and action of synthetic hormones that extend beyond the reproductive system(Reference Siminiuc and Turcanu18). The cause of symptoms and therefore management related to a natural menstrual cycle is different to side effects that may be experienced using HC. In the instance side effects of HC are experienced, a different type or brand of HC is prescribed due to the cause being related to the synthetic hormone being ingested/released.
Intervention
The selected studies had to include one or more food, diet, supplement or nutritional intervention, without being focussed on weight loss or changes in energy intake. Each intervention was required to be completed for at least one menstrual cycle.
Outcome
The primary outcome was menstrual-related symptoms. For the purpose of this review, menstrual symptoms included any symptoms on the pre-menstrual syndrome scale (PMSS) or dysmenorrhoea (abdominal/pelvic/lower back pain).
Study design
Experimental studies were considered for analysis if they met the following criteria to ensure sensitivity and robustness of results: (a) published in full in a peer-reviewed journal, (b) had primary or secondary objectives of assessing changes in menstrual-related symptoms due to food, diet, supplement or nutritional intervention, (c) interventions included groups that did not combine the use of pharmaceutical medication with the nutritional intervention, (d) non-steroidal anti-inflammatories were not used unless a statement of control was provided, and (e) Cochrane risk-of-bias 2 categorised as low risk or some concern. Single case studies, review articles, study protocol papers and conference abstracts were excluded. Only full texts that were published in English or had existing translation were examined. There was no limit on the date of publication; the studies returned range in publication date between 1953 and 2020 and within the final review between 1985 and 2020, the last search was conducted 11 February 2021.
Search strategy for identification of studies
A systematic electronic literature search was conducted by N.B. to identify relevant articles using two online databases (PubMed, ScienceDirect). The searches were performed using the following search terms (‘Diet’, OR ‘Supplement’, OR ‘nutrition’, OR ‘food’, OR ‘nutraceuticals’, OR ‘dietary patterns’, OR ‘micronutrients’, OR ‘carbohydrate’, OR ‘protein’, OR ‘fat’, OR ‘alternative medicine’, OR ‘herbal medicine’) AND (‘menstrual cycle symptoms’, OR ‘premenstrual symptoms’, OR ‘premenstrual syndrome’, OR ‘dysmenorrhea’). The reference lists of obtained relevant articles and review articles were hand-searched to identify any further studies and were added manually.
Data selection
Selection of studies
Two reviewers (N.B., D.M.) independently reviewed the titles, abstracts and access to full text paper of the identified articles for inclusion and any duplicates or review articles were removed. Three reviewers (N.B., R.F., L.F.) independently reviewed the full texts against the inclusion and exclusion criteria and any conflicts between the reviewers were resolved in meetings.
Extraction of data
Initial data extraction was conducted by one reviewer (N.B.) and verified independently by another member of the review team (R.F.). Once compiled into a preliminary table, other members of the research team (G.B., M.W., D.M.) confirmed data inclusion in the final analysis and presentation of results.
Study quality assessment
Study quality was assessed by two reviewers (N.B., R.F.) using Cochrane risk-of-bias 2 tool for randomised trials(Reference Sterne, Savović and Page34). This process ensured studies were randomised controlled trials with the purpose of assessing the effect of assignment to the intervention (intention-to-treat effect). Studies were assessed for randomisation, risk of bias due to deviations from the intended interventions, missing outcome data, measurement of the outcome and selection of the reported result. All domains required the studies follow a pre-determined algorithm for suggested judgement of risk of bias, categorising studies as low risk, some concern or high risk. An overall risk of bias judgement was provided and only those studies classified as low risk or some concern were considered. Studies judged to be of some concern were due to the research team of the study in question not being blinded from the intervention, but where due control was detailed and deemed by the reviewers to not affect the outcome results. This risk-of-bias 2 process addressed studies with missing data, only one author was contacted due to missing number of participants in each trial arm. As this data was not available, the study was excluded at this stage. The remaining studies were screened by two reviewers (N.B., R.F.) to ensure participants met the inclusion criteria outlined in the study. In addition, sample size was assessed for power (R.F.) to ensure quality and consistency of studies included for data extraction, a minimum sample size of 20 for each arm (experimental, control, placebo) was applied.
Data synthesis
Studies were grouped on the basis of food/supplement intervention and dosage compared with UK DRV and upper limits for safety of supplementation(35,36) . Data were also extracted to tabulate measurement of pre-menstrual symptom or dysmenorrhoea to allow comparison between interventions whilst informing the quantity of studies investigating each of these menstrual-related symptoms.
Meta-analysis of effect estimates and extensions were not possible due to the small number of studies for each intervention type, leading to limited evidence due to the infancy of robust studies in the research area. Extracted studies also suffered from different interventions and effect measures investigated(Reference McKenzie, Brennan, Higgins, Thomas, Chandler, Cumpston, Li, Page and Welch37). Effect sizes could not be summarised as this data was not available.
Heterogeneity arises in systematic reviews due to a variety of factors pertaining to the participants, methodologies, intervention exposure and outcomes. This review included several different interventions (nutrition, food products, supplements) in the treatment of PMS and dysmenorrhoea measured using different outcome measures, indeed the only clinical heterogeneity was that all participants were female, were naturally menstruating and were not taking any form of HC.
Results
Study selection and characteristics
The literature search and selection of studies are presented in Fig. 1. In total, twenty-eight studies with a total of 2621 participants were included in the review. Details of the included nutritional interventions are shown in Table 1 and individual studies are shown in Table 2.
Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines flow chart for literature search and study selection
Supplement/foods included within the reviewed studies, including associated dietary reference/safe upper limit in the UK
µg/d, microgram per day; IU/d, International Units per day; RNI, Reference Nutrient Intake; EFSA, European Food Safety Authority; PS, Phosphatidylserine.
Summary of reviewed studies
B, baseline; BMI, body mass index; g, grams; mg, milligrams; IU, International Unit; kg/m2, kilograms divided by height in metres squared; PI, post intervention; PMS, premenstrual symptoms; VAS, Visual Analogue Scale
Risk of bias
A proportion of twenty-six of the studies (93%) were classified as low risk. Where assessment resulted in some concerns for the studies included in this review, this was due to assessors knowing the intervention received by the study participants, as it was not possible to blind from the intervention (e.g. increasing whole grains in the diet). These papers were not excluded as precautions were stated which resulted in the review authors concluding the assessment of the outcome was not influenced by knowledge of the intervention received(Reference Esmaeilpour, Ghasemian and Alizadeh52).
All studies were reviewed in accordance with Consolidated Standards of Reporting Trials for randomisation and concealment. Despite all studies stating randomisation with evidence provided within methodological figures, only twenty-one (75%) studies provided details of the method of randomisation(Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38,Reference Moini, Ebrahimi and Shirzad39,Reference Khajehei, Abdali and Tabatabaee41–Reference Charandabi, Mirghafourvand, Nezamivand-Chegini and Javadzadeh43,Reference Jafari, Amani and Tarrahi45,Reference Zekavat, Karimi, Amanat and Alipour46,Reference Schmidt, Weber, Steiner, Meyer, Dubberke, Rutenberg and Hellhammer51,Reference Atallahi, Akbari, Mojab and Majd54–Reference Kashefi, Khajehei, Tabatabaeichehr, Alavinia and Asili57,Reference Younesy, Amiraliakbari, Esmaeili, Alavimajd and Nouraei59,Reference Agha-Hosseini, Kashani, Aleyaseen, Ghorishi, Rahmanpour, Zarrinara and Akhondzadeh60,Reference Canning, Waterman, Orsi, Ayres, Simpson and Dye62–Reference Turner and Mills65,Reference Jenabi and Fereidoony67–Reference Vannabhum, Poopong, Wongwananuruk, Nimmannit, Suwannatrai, Dangrat and Apichartvorakit69) . A total of sixteen studies (57%) provided clear details on randomisation implementation reporting either external, manufacturer or principal investigator not involved in treatment. All except one study(Reference Esmaeilpour, Ghasemian and Alizadeh52) stated both participants and experimenters were blinded and matched for method of administration. A total of five studies did not state, or it was unclear, if the intervention and control were matched on appearance(Reference Yousefi, Kavianpour, Hesami, Nooshabadi and Haghighian49,Reference Esmaeilpour, Ghasemian and Alizadeh52,Reference Atallahi, Akbari, Mojab and Majd54,Reference Agha-Hosseini, Kashani, Aleyaseen, Ghorishi, Rahmanpour, Zarrinara and Akhondzadeh60,Reference Turner and Mills65) , in ten studies it was unclear if groups were matched for taste(Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38,Reference Khajehei, Abdali and Tabatabaee41,Reference Yousefi, Kavianpour, Hesami, Nooshabadi and Haghighian49,Reference Esmaeilpour, Ghasemian and Alizadeh52,Reference Atallahi, Akbari, Mojab and Majd54,Reference Younesy, Amiraliakbari, Esmaeili, Alavimajd and Nouraei59,Reference Agha-Hosseini, Kashani, Aleyaseen, Ghorishi, Rahmanpour, Zarrinara and Akhondzadeh60,Reference Turner and Mills65,Reference Jenabi and Fereidoony67,Reference Raisi Dehkordi, Rafieian-kopaei and Hosseini-Baharanchi68) and twelve studies lacked clarity on matching based on smell(Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38,Reference Khajehei, Abdali and Tabatabaee41,Reference Hagen, Nesheim and Tuntland44,Reference Zekavat, Karimi, Amanat and Alipour46,Reference Yousefi, Kavianpour, Hesami, Nooshabadi and Haghighian49,Reference Esmaeilpour, Ghasemian and Alizadeh52,Reference Atallahi, Akbari, Mojab and Majd54,Reference Kashefi, Khajehei, Tabatabaeichehr, Alavinia and Asili57,Reference Agha-Hosseini, Kashani, Aleyaseen, Ghorishi, Rahmanpour, Zarrinara and Akhondzadeh60,Reference Turner and Mills65,Reference Jenabi and Fereidoony67,Reference Raisi Dehkordi, Rafieian-kopaei and Hosseini-Baharanchi68) .
Results of individual studies
Table 2 presents a summary of results for each study. The studies have been grouped and displayed based on similarities in type of food or supplement intervention.
Results of synthesis
A total of twenty-one food or supplement interventions, including individual and combined supplements were identified in the analysis of studies as interventions for management of menstrual-related symptoms (Table 1). Overall, fourteen studies investigated the effect of a food or supplement intervention on general pre-menstrual symptoms (ten positive results, four no significant effect) and fourteen studies (thirteen positive results, one with no significant effect) investigated use specifically for dysmenorrhoea. The results were variable between interventions, despite some consistencies in the prescribed type of food/supplement. This variability could be attributed to differences in dose, length of intervention, participant characteristics (e.g. deficiency in vitamin under study, age, age at menarche, Table 3) and methodological variation in determination of symptoms. In the latter case there were few consistencies in determination of pre-menstrual symptoms and dysmenorrhoea (Table 4). A total of eighteen different ways of measuring pre-menstrual symptoms were described across the studies and seven different methods to determine dysmenorrhoea pain were extracted from the twenty-eight studies. This prevented any further comparisons between the studies.
Participant characteristics
PAS, Phosphatidic acid complex; PPF, Prasaplai formula; ALA, Alpha-lipoic acid.
Reported measurements from twenty-eight studies to determine pre-menstrual symptoms and primary dysmenorrhoea
* Indicates validated measure.
PMS, pre-menstrual syndrome; PSST, pre-menstrual syndrome screening tool; VAS, visual analogue scale; DSM VI, Diagnostic and Statistical Manual of Mental Disorders 6; DSM IV Diagnostic and Statistical Manual of Mental Disorders 4
Of the twenty-eight studies reviewed, two were crossover study design, whereas twenty-six were parallel controlled trials. None reported a negative effect on symptoms nor investigated any potential placebo effect, twenty-three reported a positive effect (Table 2) and five had no effect or were inconclusive (Table 2). Additional factors such as lack of control of dietary intake when investigating omega-3 or other vitamins and minerals naturally found in food prevented conclusions being drawn of the use of additional supplements. Furthermore, there was variation in dose of intervention used and duration of intervention across the studies preventing further analysis (Table 1).
Where more than one paper was reviewed of the same intervention type, further synthesis of results was completed. This included thirteen studies and six different nutritional practices: vitamin D, calcium (and magnesium) St John’s wort, Vitus agnus-castus, curcumin and zinc. Synthesis highlights seven of these studies report no other form of vitamin or mineral supplement ingested prior to and during the intervention; interventions where this control was not stated were St John’s wort and Vitus agnus-castus. There was variation in identification of deficiency, for instance, serum vitamin D measures included blood ELISA analysis(Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38), electrochemiluminescene(Reference Moini, Ebrahimi and Shirzad39) or self-reported diet lacking consumption of vitamin D(Reference Khajehei, Abdali and Tabatabaee41,Reference Zarei, Charandabi, Mirghafourvand, Javadzadeh and Effati-Daryani42) . In some cases(Reference Khajehei, Abdali and Tabatabaee41,Reference Zarei, Charandabi, Mirghafourvand, Javadzadeh and Effati-Daryani42) , levels were not measured. Self-reported food diaries were also used to establish calcium intake and determination of insufficient quantities consumed (e.g. less than 1000 mg calcium consumed per day deemed as deficient(Reference Zarei, Charandabi, Mirghafourvand, Javadzadeh and Effati-Daryani42)).
When differentiating the outcome measures of the four studies which included vitamin D, only two looked solely at vitamin D supplementation, one for the treatment of dysmenorrhoea(Reference Houghton, Manson, Whitcomb, Hankinson, Troy, Bigelow and Bertone-Johnson26) and one for the treatment of pre-menstrual symptoms(Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38). Disparity existed between the outcome of vitamin D on treatment of pre-menstrual symptoms with one overall report of no effect(Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38). Both studies included participants of the same age range, deficient in vitamin D (<0.03 mg/l or <0.02 mg/l) and not taking any vitamin D or other vitamin/mineral supplements. There was variation in dose and length of time; 50 µg every other day for 12 weeks compared to 530 mg once a week for 12 weeks. Although Adbollahi et al. (Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38) reported results as overall no effect, vitamin D significantly improved anxiety, providing evidence that vitamin D may impact on mood related symptoms; however, consistency in length and dose amount need to be controlled along with outcome measure to determine any effect on symptoms.
A combined intervention of vitamin D and calcium identified dydrogesterone was more efficient to manage menstrual-related symptoms(Reference Khajehei, Abdali and Tabatabaee41) or calcium alone(Reference Zarei, Charandabi, Mirghafourvand, Javadzadeh and Effati-Daryani42) to reduce dysmenorrhoea and pain intensity. However, in another study calcium plus magnesium was reported to have a greater effect of reducing dysmenorrhoea than calcium alone(Reference Charandabi, Mirghafourvand, Nezamivand-Chegini and Javadzadeh43). Across the calcium and vitamin D studies there was variation in dose and duration of intervention; 500 mg of calcium plus 200 mg (200 000 ug) vitamin D were prescribed twice daily from the fifteenth to twenty-fourth day of the menstrual cycle for two consecutive cycles. Whereas one tablet per day of 1000 mg calcium and 5.3 mg vitamin D3, or one tablet per day 1000 mg calcium only were taken from the fifteenth day of menstrual cycle until the disappearance of menstrual pain in the following cycle. Whilst high, this calcium dose is below the likely safe upper limit of 1500 mg/d(36). Calcium doses were consistent across the studies, however vitamin D doses varied. Duration of supplementation did vary between the studies reviewed. Intervention duration ranged from two consecutive cycles(Reference Khajehei, Abdali and Tabatabaee41) or categorised as 8 weeks in another paper(Reference Moini, Ebrahimi and Shirzad39) to three cycles(Reference Zarei, Charandabi, Mirghafourvand, Javadzadeh and Effati-Daryani42) or 12 weeks(Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38).
Zinc reduced physical and psychological symptoms associated with the menstrual cycle(Reference Jafari, Amani and Tarrahi45) and dysmenorrhoea(Reference Zekavat, Karimi, Amanat and Alipour46), with results also highlighting zinc combined with ginger can reduce primary dysmenorrhoea pain(Reference Kashefi, Khajehei, Tabatabaeichehr, Alavinia and Asili57). Studies spanned age ranges of 14–30 years, with interventions ranging from approximately 4 weeks to 12 weeks and variation in zinc dosage of 30 mg/d, 50 mg/d and 220 mg/d (Table 1) taken 4 d prior to onset of menstruation or on the first day of menstruation for 3 d or 3 d prior to the end of menstruation. It should be noted that Kashefi et al. (Reference Kashefi, Khajehei, Tabatabaeichehr, Alavinia and Asili57) provided doses significantly higher than the recommended safe upper limit for zinc of 50 mg/d for adults(36).
Curcumin reduced the severity of menstrual related symptoms in female eumenorrhoeic students with total scores reducing from 102·06 (SD 39·64) at baseline to 42·47 (SD 16·37) post intervention, with placebo reducing to 91·60 (SD 43·56) (p < 0·001). Curcumin was taken in capsules over three menstrual cycles, two capsules were taken daily (100 mg/12 h) for 10 d in total – seven before menstruation and three during menstruation in both studies(Reference Fanaei, Khayat, Kasaeian and Javadimehr55,Reference Khayat, Fanaei, Kheirkhah, Moghadam, Kasaeian and Javadimehr56) .
Studies investigating St John’s wort and Vitus agnus-castus were inconsistent in results with only one study of each showing an effect. There was variation between intervention dose (e.g. St John’s wort) preventing any conclusions being drawn on effectiveness of these nutritional practices.
Overall, vitamin D, calcium (and magnesium), zinc and curcumin were effective in reducing pain or severity of menstrual related symptoms. A maximum of three papers were reviewed under each of these interventions; therefore, despite positive results, caution should still be taken when applying the results widely.
One extreme case which should be interpreted with caution investigating the effect of Psiddi guajavae folium extract for treatment of dysmenorrhoea(Reference Doubova, Morales, Hernández, del Carmen Martínez-García, de Cossío Ortiz, Soto, Arce and Lozoya63). Reported results over emphasized the benefit of the extract compared with ibuprofen. Treatment combined with compliance for ibuprofen was not reported, despite completing this for both 6 mg/d and 3 mg/d Psiddi guajavae folium extract treatment groups. By the end of treatment there was no difference in pain reported between any of the groups. Compliance was reported with similarities between groups, but the number of participants in each group varied for analysis (fifty-seven 6 mg/d extract participants compared with forty-two placebo participants and forty-six ibuprofen participants). All interventions were identical and labelled with codes only known by the investigators who manufactured the products. However, there was no clarification if these were the same investigators completing the research and details of double blinding the interventions to prevent bias lacked clarity.
Eight of the reported interventions involved nutrients with UK DRV and safe upper limits recommended (Table 1). Of note, some of the supplement interventions included lactose in the preparations for instance Agha-Hossini et al. (Reference Agha-Hosseini, Kashani, Aleyaseen, Ghorishi, Rahmanpour, Zarrinara and Akhondzadeh60) and Hicks et al. (Reference Hicks, Walker, Gallagher, Middleton and Wright61), which could be problematic for those with lactose intolerance. In addition, lactose was included as a placebo in some studies(Reference Zarei, Charandabi, Mirghafourvand, Javadzadeh and Effati-Daryani42,Reference Kashefi, Khajehei, Tabatabaeichehr, Alavinia and Asili57) which needs consideration if recruiting for further studies.
Discussion
The aim of this systematic review was to determine the influence of nutritional practices on symptoms related to the menstrual cycle. The results indicated that there are food/supplements which may be used by women to reduce severity of menstrual-related symptoms (based on single studies) including pain caused by primary dysmenorrhoea if guidelines are followed and any contraindications are considered. However, there is a lack of consistency regarding determination of menstrual related symptoms and dose/duration of food or supplement to make recommendations in practice.
Within this review, various foods and supplements have found a positive effect on reducing menstrual related symptoms and dysmenorrhoea including vitamin D, calcium (and magnesium), zinc and curcumin, but in these instances a maximum of two to three papers were reviewed for each treatment, with variations in intervention duration and dose, alongside inconsistent outcome measures, thus preventing direct comparison between the studies. Foods such as whole grains in place of refined grains and omega-3 may contribute to improved menstrual related symptoms, whilst cinnamon may reduce the intensity of dysmenorrhoea; however, these studies did not explore the mechanisms related to these findings and only one study was reviewed on each preventing definitive recommendations from being provided.
Current pharmacological treatments for women suffering from menstrual-related symptoms and primary dysmenorrhea can include non-steroidal anti-inflammatory drugs, HC, antipyretic medication and analgesic medication. However, these medications have a reported failure rate of 20–25%(Reference Edwards, Moore and Mcquay70,Reference Zhu, Proctor, Bensoussan, Smith and Wu71) , and can have associated unwanted outcomes, such as diarrhoea, stomach-ache and nausea(Reference Harel72) along with long-term health implications. Furthermore, some women might not wish to use medication or contraceptives for the relief of symptoms and pain, and some have religious or cultural conflicts with the use of these medications(Reference Durain73). Alternative options are required; despite inconsistencies in current literature reviewed, there may be vitamins, minerals or other nutritional interventions for the treatment of menstrual related symptoms and dysmenorrhoea. Results of this review have informed the specific sections below on vitamin D, calcium and magnesium, zinc and curcumin.
Calcium and vitamin D
From the studies reviewed, one study identified calcium and vitamin D combined were more effective than a placebo in reducing pre-menstrual symptoms in participants self-reporting a deficiency of calcium or vitamin D in their diet(Reference Zarei, Charandabi, Mirghafourvand, Javadzadeh and Effati-Daryani42). No measures determined participants’ serum levels of either vitamin or mineral to recognise previous suboptimal levels prior to intervention and nor were any measures included in the study design to determine the biological mechanism causing a difference in the results. However, in more recent years, it has been noted that blood levels of calcium and vitamin D, the latter facilitating calcium absorption, fluctuate across the menstrual cycle(Reference Thys-Jacobs, McMahon and Bilezikian74). Disruption of calcium regulation has been proposed as an underlying factor for increasing incidence and severity of menstrual pain(Reference Thys-Jacobs22).
From the perspective of a reduction in pain associated with primary dysmenorrhoea, calcium alone was identified as being more effective than when combined with vitamin D in the studies reviewed(Reference Zarei, Charandabi, Mirghafourvand, Javadzadeh and Effati-Daryani42). Differences in protocols may explain these results, with supplements only being taken from the fifteenth day of the menstrual cycle. Again, all measures were self-reported, including determining calcium deficiency using a food diary; those classed as deficient if consuming less than 1000 mg/d. Analysis of each group showed that 93% of participants were deemed to be deficient in the calcium and vitamin D group, 86% deficient in calcium alone group and 89% deficient within the placebo. These results highlight calcium supplementation may be beneficial to women deficient in calcium to reduce dysmenorrhoea; however, there was limited mechanistic insight provided by the study to draw any conclusions.
Two studies considered solely vitamin D supplementation; the studies were inconsistent in outcome measures (dysmenorrhoea or pre-menstrual symptoms) with further disparity between dose and duration. It has been shown that vitamin D has anti-inflammatory effects(Reference Bertone-Johnson and Manson75). It reduces production of prostaglandins(Reference Lasco, Catalano and Benvenga76), which have been found to have a major role in the pathophysiology of PMS(Reference Koshikawa, Tatsunuma, Furuya and Seki77) and dysmenorrhoea(Reference Fajrin, Alam and Usman78). Previous research has shown that vitamin D may influence several different mechanisms related to dysmenorrhoea and PMS, including reduced expression of cyclooxygenase-2 and, accordingly, reduced prostaglandin production, up-regulation of 15-hydroxyprostaglandin dehydrogenase, increased prostaglandin inactivation, regulation of the expression of prostaglandin receptor, and consequently reduced pain intensity(Reference Bertone-Johnson and Manson75). This may highlight the beneficial effects reported in the reviewed studies for women deficient in vitamin D but may be limited to this population.
One study in the present review investigated the use of calcium alone or combined with magnesium(Reference Charandabi, Mirghafourvand, Nezamivand-Chegini and Javadzadeh43). Previous research has reported serum magnesium acts as a transit pathway between electrolyte uptake and excretion, bone stores and actively metabolising tissues; these processes are affected by several hormones, including sex steroids(Reference Bermon, Castell and Calder79). Deficiency can cause muscle cramps, anxiety and signs of inflammation(Reference Bermon, Castell and Calder79). There is evidence of reduced levels of prostaglandin F2α, involved in pain and inflammation(Reference Marjoribanks, Proctor and Farquhar80), which may play an important role in pain modulation of primary dysmenorrhoea. Out of the studies reviewed, one study investigated the use of calcium alone or calcium and magnesium combined, with results identifying the addition of magnesium had an increased effect on reducing pain intensity. However, no details were provided on participant levels of calcium or magnesium prior to the intervention. Consistent with previously reported findings, no measures to determine biological mechanisms affecting the reduction in self-reported pain were included in the study design.
Zinc
Serum zinc concentrations change during the menstrual cycle; zinc deficiency can reduce the zinc serum concentrations and consequently may cause the glucocorticoid’s production to be irregular and lead to some neuropsychological symptoms such as irritability, depression and emotional instability(Reference Sowa-Kućma, Legutko and Szewczyk81), along with being associated with dysmenorrhoea(Reference Nasiadek, Stragierowicz, Klimcxak and Kilanowicz82).
Within this current review, findings of Jafari et al. (Reference Jafari, Amani and Tarrahi45) demonstrated that taking zinc supplements for 12 weeks in women with PMS resulted in significant increase in brain-derived neurotrophic factor (BDNF) compared with the placebo. There has been no other study that assessed the effect of zinc supplementation on serum levels of BDNF in PMS subjects(Reference Jafari, Amani and Tarrahi45). However, it is believed zinc induces the matrix metalloproteinase that activates tropomyosin-related kinase proteinase, and this leads to release of pro-BDNF from cells and then converts to BDNF(Reference Hwang, Park, Choi and Koh83). Previous studies have shown that BDNF has a role in a women’s reproductive physiology and some of the actions of sex hormones are mediated by BDNF(Reference Cubeddu, Bucci and Giannini84). Recent studies in PMS women revealed serum BDNF levels in luteal phase are significantly different than women without PMS, with studies reporting both higher(Reference Oral, Kirkan, Yildirim, Kotan, Cansever, Ozcan, Aliyev and Gulec85) and lower(Reference Cubeddu, Bucci and Giannini84) BDNF levels associated with menstrual-related symptoms. However, there is consensus between studies suggesting that circulating serum BDNF levels are associated with incidence of PMS symptoms and may play a role in the pathophysiology of PMS(Reference Cubeddu, Bucci and Giannini84,Reference Oral, Kirkan, Yildirim, Kotan, Cansever, Ozcan, Aliyev and Gulec85) .
Alternative research has suggested zinc promotes microcirculation and prevents ischaemia(Reference Eby and Halcomb86), inactivates free oxygen radicals by increasing the dismutase enzyme level(Reference Sugino, Karube-Harada, Sakata, Takiguchi and Kato87), down-regulates inflammatory cytokines(Reference Prasad, Bao, Beck, Kucuk and Sarkar88) and decreases and regulates the level of the cyclooxygenase-2 enzyme(Reference Fong, Zhang, Jiang and Farber89) to reduce dysmenorrhoea. Additionally, inflammatory markers are significantly associated with severity of menstrual symptoms in women(Reference Bertone-Johnson, Ronnenberg and Houghton90); since zinc functions as an anti-inflammatory agent(Reference Prasad91), it may affect inflammatory markers, such as high-sensitivity C-reactive protein (Reference Bao, Prasad, Beck, Fitzgerald, Snell, Bao, Singh and Cardozo92) and, therefore, alleviates PMS symptoms as well.
The tolerable uptake level of zinc suggested by the UK safe upper limit is 50 mg/d and popular multi-vitamins supply more than 15–30 mg of zinc as a dietary supplement(93). The recommended daily dietary allowance of elemental zinc for women is 8 mg/d, and long-term supplementation with more than 20 mg elemental zinc per day would be unsafe(Reference Maret and Sandstead94). Within studies included in this current review, the dose of zinc was much greater in the study by Kashefi et al. (Reference Kashefi, Khajehei, Tabatabaeichehr, Alavinia and Asili57), reporting 220 mg zinc sulphate capsules taken for 4 d approximately at the start of menstruation, and Zekavat et al. (Reference Zekavat, Karimi, Amanat and Alipour46) reported an intervention of 50 mg/d of zinc sulphate for 4 d. Although Jafari et al. (Reference Jafari, Amani and Tarrahi45) reported use of a lower dosage (30 mg/d zinc gluconate) this was prescribed every day for 12 weeks. The studies included in this review reporting use of zinc supplements should be applied with caution, checking the dose first, and further research is required to determine dose quantity of zinc to improve PMS and dysmenorrhoea experienced by women.
Curcumin
The results from the studies included in this systematic review(Reference Fanaei, Khayat, Kasaeian and Javadimehr55,Reference Khayat, Fanaei, Kheirkhah, Moghadam, Kasaeian and Javadimehr56) both reported a beneficial effect of curcumin. Fanaei et al. (Reference Fanaei, Khayat, Kasaeian and Javadimehr55) identified that BDNF levels were significantly higher and mean scores of PMS were significantly lower following supplementation of curcumin for 10 d repeated for three menstrual cycles; a daily dose of 100 mg/12 h was prescribed. This study supported the proposed mechanisms related to BDNF and subsequent impact on mood and behaviour regulation(Reference Fanaei, Khayat, Kasaeian and Javadimehr55).
Curcumin, otherwise known as turmeric, has recently been involved in research indicating its medicinal, health(Reference Suhett, de Miranda Monteiro Santos, Souza Silveira, Gualandi Leal, Melo de Brito, Farias de Novaes and Lucia95) and anti-inflammatory effects (Reference Ghandadi and Sahebkar96). Studies have demonstrated that curcumin can have beneficial effects on physiological and pathological conditions(Reference Suhett, de Miranda Monteiro Santos, Souza Silveira, Gualandi Leal, Melo de Brito, Farias de Novaes and Lucia95). Curcumin is proposed to have a neuroprotective effect, it is observed that it acts through modulating the release of certain neurotransmitters, such as BDNF(Reference Liu, Wang, Gao, Xie, Zhang, Jiang and Pang97).
Both reviewed studies were consistent in protocol, Khayat et al. (Reference Khayat, Fanaei, Kheirkhah, Moghadam, Kasaeian and Javadimehr56) prescribed 100 mg/12 h over 10 d and repeated for three menstrual cycles. There was consistency in the results and protocol implemented and there is no daily recommended dose or safe upper limit for this plant-based extract which may therefore be deemed safe to supplement. Similarly, cinnamon was reported to have a beneficial effect on pain intensity of primary dysmenorrhoea and requires further research to support the one study reviewed(Reference Jahangirifar, Taebi and Dolatian58).
Measurement of menstrual symptoms and outcome measures
To date, although over 200 menstrual cycle-related symptoms have been described, a universally accepted definition and/or conventional diagnostic criteria for PMS are still lacking(Reference Halbreich98). Although no specific physical finding or laboratory test is currently available for the diagnosis of PMS, the American College of Obstetricians and Gynecologists published a list of diagnostic criteria for PMS(99,Reference Steiner, Streiner, Steinberg, Stewart, Carter, Berger, Reid and Grover100) , establishing that PMS could be diagnosed if at least one of the affective and one of the somatic symptoms were reported 5 d before the onset of menses in the three previous menstrual cycles(Reference Tartagni, Cicinelli and Tartagni101). Symptoms must be prospectively recorded in at least two cycles, must cease within 4 days from the onset of menses, and should not recur before day twelve of the next cycle(Reference Tartagni, Cicinelli and Tartagni101). In addition, symptoms must be recorded in the absence of pharmacologic therapy in patients not taking hormones, drugs and alcohol(Reference Tartagni, Cicinelli and Tartagni101). However, dysmenorrhoea alone may be defined as a menstrual cycle-related symptom, but under the diagnostic criteria this would exclude a participant from being diagnosed with PMS.
Criteria were not consistent between the studies reviewed, there were instances when multiple diagnostic criteria were included. For instance, Abdollahi et al. (Reference Abdollahi, Abiri, Sarbakhsh, Kashanian and Vafa38) recorded the cycles for 3 months; the individual presented with one or more somatic symptoms for 5 d before a period and symptoms did not start until day thirteen of the cycle. Future research should determine PMS diagnostic criteria and studies should implement these criteria as a minimum when examining the efficacy of a food or supplement on menstrual-related symptoms.
To further improve the quality of the studies, longitudinal tracking of symptoms over time to allow for other lifestyle influences which may affect symptomology should be considered. Where possible, sex hormones should be measured to confirm the changes in hormones across menstrual cycle phases coinciding with presentation of symptoms, along with physiological markers, to assist with investigation of underlying mechanisms of interventions to support self-reported measures.
Participant characteristics
The age of participants varied in the review (Table 3) with one study(Reference Zekavat, Karimi, Amanat and Alipour46) investigating the effect of intervention on participants as young as 15 years with no information provided in the methodology stating eligibility to participate based on age of menarche. Peri-pubertal girls can be more symptomatic with hormone secretions more likely to be atypical(102); this may add variability to the effect of the intervention on outcomes measures and findings reported. Vitamin and mineral deficiencies were considered in some studies, but not all; the level of determination of deficiency varied as well. Future studies need to measure and report method of measurement of deficiency status in participants to understand if supplementing elevated to a ‘normal’ level and the consequential impact on symptoms.
Limitations and future research
Limitations of evidence reported within this review should be considered, there was consistently a lack of alignment in outcome measures with further variation noted between studies regarding intervention in terms of food or supplement investigated and dose and duration of interventions. Furthermore, the majority of studies did not investigate physiological mechanisms associated with symptoms related to the menstrual cycle. Self-reported outcome measures are important in understanding severity and frequency of symptoms experienced, but biological mechanisms are an additional requirement in understanding nutritional practices and interventions in management of menstrual cycle symptoms.
Another limitation of reviewed studies is the lack of consideration or discussion relating to the placebo effect, despite research identifying many types of placebo responses driven by different mechanisms depending on the particular context(Reference Price, Finniss and Benedetti103). As placebo factors have neurobiological underpinnings and affects the brain and body, placebo mechanisms should be considered in clinical trial design(Reference Price, Finniss and Benedetti103). Research has evidenced that placebo effects can be strong in the first cycle of intervention(Reference Freeman and Rickels104) but should begin to decline to baseline as the study continues(Reference Bryant, Cassidy, Hill, Powell, Talbot and Dye105). Therefore, study design should consider this with interventions completed for at least two cycles. Conversely, not all studies provided clear detail on how the placebo/control was equally matched to the active intervention(Reference Atallahi, Akbari, Mojab and Majd54,Reference Canning, Waterman, Orsi, Ayres, Simpson and Dye62) which may influence the study findings. Future research should consider the placebo/control within methodological design to ensure it is evenly matched on the basis of appearance, taste and smell and ensure more than one cycle of intervention is completed.
The reviewed studies included individuals naturally menstruating, excluding those using HC and eliminating multiple studies from the review. It is acknowledged that this reduces the practical application of the review to only those naturally menstruating; however, the mechanisms causing side effects of HC vary significantly due to the differences in exogenous versus endogenous hormones and are not comparative and therefore are beyond the scopes of this review. Furthermore, individuals experiencing side effects of HC can change the type and brand, associated with different levels of synthetic hormone and composition, to actively manage side effects. Future research is needed to explore the mechanisms of symptoms related to the menstrual cycle and HC to enable application of interventions to manage these two groups.
Future research specifically should investigate the effect of vitamin D, calcium, magnesium and zinc prescribed within safe upper limits and measured physiological changes in both normal and deficient women.
Conclusion
This is the first systematic review to examine the use of food or dietary supplements to help with management of menstrual related symptoms. These data provide some evidence for the use of vitamin D, calcium (and magnesium), zinc and curcumin to reduce pre-menstrual symptoms. However, further investigation is required to understand the physiological mechanism relating to these nutritional interventions and how specific symptoms related to the menstrual cycle may be improved rather than grouped as PMS. Future studies should focus on the general quality of the methods used, to allow direct comparisons between studies to be made and conclusions to be drawn from studies and implemented in practice.
Availability of data materials
Not applicable.
Acknowledgements
The authors would like to thank Dr Paul Hewlett, Department of Applied Psychology, Cardiff School of Sport and Health Sciences, Cardiff Metropolitan University for statistical advice relating to assessment of the studies.
Authorship
N.B., R.F., D.M., M.W. and G.B. contributed to conceptualisation; N.B., D.M., R.F., L.F., M.W. and G.B. were involved in review; N.B. and R.F. wrote the original draft; N.B., R.F., D.M., M.W. and G.B. reviewed and edited subsequent drafts. All authors read and approved the final manuscript.
Financial support
No sources of funding were used to assist in the preparation of this article.
Competing interests
All authors declare that they have no conflicts of interest relevant to the content of this review.
Ethical statement
Not applicable.
Consent for publication
Not applicable.
Open access
