An epidemiologic review of menstrual blood loss as an excretion route for per- and polyfluoroalkyl substances

Abstract

Purpose of review:

Menstrual blood loss, a common physiologic occurrence, provides an excretion route for per- and polyfluoroalkyl substances (PFAS) since these chemicals are bound to proteins in blood. To increase awareness of this relationship in environmental epidemiology, we reviewed the available epidemiologic data on menstrual bleeding and PFAS concentrations.

Recent findings:

Initial epidemiologic studies reported generally higher PFAS concentrations in men, menopausal women, and those with a history of hysterectomy compared to premenopausal women. Although subsequent studies investigating menstrual cycle characteristics observed somewhat discrepant results for menstrual irregularity and cycle length, consistent associations have been observed between heavy menstrual bleeding and lower PFAS concentrations.

Summary:

This review highlights the important role of menstrual bleeding on the excretion of PFAS. Given the high prevalence of menstrual bleeding in the population and the implications for environmental epidemiology, we provide recommendations to move this field forward.

Introduction

Per- and polyfluoroalkyl substances (PFAS) are synthetic fluorinated chemicals with water, grease, and stain repellent properties, often used in consumer products such as nonstick cookware, carpets, leather products, textiles, cardboard packaging, and waterproof clothing [1]. The general population is primarily exposed to PFAS from consumption of contaminated food and drinking water [1]. Since PFAS are resistant to breakdown in the environment, have long-half lives, and bioaccumulate, PFAS exposure is widespread in the general population globally [2]. In a nationally representative sample of U.S. adults, several PFAS, including perfluorooctane sulfonate (PFOS), perfluorooctanoic acid (PFOA), perfluorohexane sulfonate (PFHxS), and perfluorononanoic acid (PFNA) have been detected in the serum of more than 98% of adults [3]. This is concerning as the most frequently studied PFAS, PFOS and PFOA, have demonstrated reproductive, developmental, hepatic, and immune toxic effects in animal studies [1] with associations also observed in human epidemiologic studies [4]. Several manufacturers began voluntarily phasing out the production and use of long-chain PFAS in the early 2000s [1]. However, short-chain replacements may also persist in the environment [5, 6].

From the investigation of pathways of human exposure to PFAS, it is recognized that PFAS are transferred to the fetus during pregnancy and are excreted in breastmilk [1]. However, blood loss is also an excretion route [7] as PFAS are distributed in blood, bound to albumin and other proteins [8, 9]. One source of blood loss and excretion route for PFAS that remains underappreciated is menstrual blood loss. Menstrual bleeding may substantially contribute to the excretion of PFAS given that it can occur on average every 24–38 days over several decades between menarche and menopause [10]. In addition, menstrual blood loss is common; half of the world’s population experiences menstrual bleeding in their lifetimes. Therefore, increased awareness and understanding of menstrual bleeding as an excretion route for PFAS is warranted, including the ramifications of this relationship on environmental health research.

To address this need, the purpose of this review is to summarize the available epidemiologic data on menstrual bleeding and PFAS body burden. We first describe the initial epidemiologic evidence on sex differences in PFAS concentrations that pointed towards menstrual status and the biologic plausibility for PFAS elimination with menstrual blood loss. This is followed by a review of epidemiologic studies on menstrual bleeding characteristics and PFAS concentrations in women and a discussion of the implications of the findings in environmental epidemiology. We conclude by providing recommendations for future research to move this field forward.

For this review, we use the term “menstrual bleeding” broadly. We consider menstrual bleeding to include bleeding from menstruation as well as bleeding and spotting between menstrual periods. It also includes withdrawal bleeding, breakthrough bleeding, and spotting with hormonal contraceptive use [11].

Epidemiologic studies pointing towards sex and menstrual status differences in PFAS exposure

Initial epidemiologic studies of correlates of PFAS concentrations reported generally higher PFAS concentrations among males compared to females, with PFOS and PFOA being the most frequently investigated PFAS [12–19]. A small study that included men ages 20–76 years (n=20), premenopausal women ages 20–50 years (n=20) and postmenopausal women ages 51–76 years (n=8) observed intriguing associations: (1) serum PFOA and PFOS concentrations did not change with age in men [20]; (2) women only had concentrations similar to men if they were postmenopausal; and (3) postmenopausal women had higher serum PFOA and PFOS concentrations compared to premenopausal women [20]. As an explanation for the observed differences, Harada et al (2005) proposed menstrual bleeding as an excretion route for PFAS.

Subsequent studies have reported a positive association between increased PFOS and PFOA concentrations and menopause [21, 22]. Although those two cross-sectional studies evaluated exposure to PFAS in relation to the outcome of menopause, additional data support the alternative explanation that postmenopausal women have higher PFAS concentrations due to a lack of PFAS loss via menstrual bleeding. Higher PFAS has been associated with hysterectomy [21, 22], no menstrual bleeding since last study visit in a longitudinal study of the menopausal transition [23, 24], and increasing years since menopause [22, 25]. In addition, a prospective study of PFOA and age at natural menopause observed no association [26]. The investigators of that study also conducted a cross-sectional analysis of PFOA to demonstrate that the early studies of PFAS and menopause may have been due to reverse causation [27]. In the cross-sectional analysis, they reported a positive association between PFOA and menopause prevalence [27]. The concern for reverse causation – that the lack of menses with menopause allowed for the accumulation of PFAS given its long half-life as opposed to PFAS altering the timing of menopause – was further explored in a quantitative bias analysis using Monte Carlo life-stage physiologically-based pharmacokinetic model simulations [28]. That study also suggested that the prior associations observed in cross-sectional analyses [21, 22] could be explained by menopause leading to lower elimination of PFAS (and higher PFAS concentrations) [28].

Pharmacokinetic modelling has been used to investigate the sex difference in PFOS elimination; data from that research suggest that menstruation explains 60% of the discrepancy in PFOS elimination [29–31]. The authors contend that the contribution of menstruation to PFOS elimination over an average individual’s lifetime is greater than that from blood loss during childbirth and elimination with breastfeeding [29].

Biologic plausibility that PFAS is excreted in menstrual fluid

Along with findings from initial epidemiologic studies, data on PFAS protein-binding in blood provide further biologic support for the excretion of PFAS in menstrual fluid. PFOA and PFOS are highly bound to albumin protein in serum [8, 9]. Although menstrual fluid differs from whole blood as it is comprised of blood, vaginal secretions, and endometrial and immune cells that shed from the uterus [32], some data suggest that the albumin concentration in menstrual supernatant is similar to that in plasma [33]. It has been estimated that, on average, women lose around 30 ml to 40 ml of menstrual fluid with menses, with loss exceeding 80 ml indicating heavy menstrual bleeding [34, 35]. These data were based on Swedish women from the 1960s using objective bleeding measurements, and may underestimate the extent of heavy menstrual bleeding. Current clinical guidelines recommend that heavy menstrual bleeding be defined as excessive menstrual bleeding that interferes with quality of life, affecting physical, emotional and/or social well-being [36]. In our research of reproductive-age women, we have observed that 20% of women had heavy menstrual flow based on menstrual product use at their last menses and 25% of women have ever used NSAIDs for heavy menstrual bleeding [37]. Hence, heavy menstrual bleeding is common, and in some women, menstrual blood loss may be considerable. Furthermore, episodic blood loss in men undergoing venesection treatment is associated with 40–50% lower serum PFAS concentrations [7]. By applying pharmacokinetic modelling, that study demonstrated that blood loss contributed to lower PFAS concentrations and that sex differences could be explained by menses [7]. Thus, given PFAS distribution in blood, contribution of blood loss to PFAS elimination, the composition of menstrual fluid, and the amount of menstrual blood loss with each menses, the elimination of PFAS in menstruators may be substantial.

Menstrual bleeding characteristics and PFAS concentrations in women

In this section, menstrual bleeding characteristics include cycle regularity, cycle length, and menstrual flow heaviness. Cycle regularity is generally regarded as menstrual bleeding frequency that is predictable, or cycle lengths that vary little across cycles [10]. Cycle length is defined as the number of days from the first day of menstrual bleeding to the start of the next menstrual period [38]. Menstrual flow heaviness refers to the volume of menstrual blood loss [10]. With regard to these menstrual bleeding characteristics and PFAS concentrations, frequently occurring menstrual bleeding with heavier flow would be expected to be associated with greater menstrual blood loss and lower PFAS concentrations in menstruators.

Irregular menstrual bleeding

The first menstrual bleeding characteristic reported in relation to PFAS concentrations was irregular menstrual period [39] (Table 1). In a subset of the Danish National Birth cohort of pregnant women, Fei et al (2009) provided descriptive data that irregular menstrual periods were more frequently reported among those with higher PFOA and PFOS concentrations [39]. However, the definition of irregular menses and timeframe considered before pregnancy were not provided. Four subsequent studies of irregular menstrual cycles and PFAS chemicals have reported inconsistent results, with two studies reporting increased concentrations of specific PFAS chemicals, including PFOA, PFOS, PFNA, PFHxS [40, 41] and two studies reporting no association [42, 43]. The studies observing associations defined irregular cycles as cycles varying 7 days or more in cycle length from month-to-month [40, 41]; the studies reporting no association asked each participant whether their period was regular the year before they became pregnant, without defining “regular” [42] or was based on cycle length being “too irregular” to report a usual cycle length [43]. Furthermore, for three of these studies the study population was restricted to pregnant women [39, 40, 42]. This raises concerns as (1) the measurement of PFAS concentrations during pregnancy can be affected by pregnancy-related hemodilution and altered PFAS pharmacokinetics, as well as maternal PFAS transfer to fetus [44], and may not characterize PFAS body burden pre-pregnancy when irregular menstrual bleeding was assessed; and (2) the selection of pregnant participants may be related to cycle regularity, biasing the estimate of the association.

Table 1.

Epidemiologic studies of menstrual bleeding characteristics and perfluoroalkyl substances measured in women.

Author, year [ref]	Location, years	Study design	Study population	Inclusion/exclusion criteria	Menstrual bleeding characteristic(s)	Chemicals investigated	Results
Fei 2009 [39]	Denmark, 1996–2002	Cross-sectional	Danish National Birth Cohort of pregnant women subset (n=1240)	Inclusion: Provided blood sample, gave birth to single live born infant without congenital malformation, completed all 4 interviews	Irregular menstrual periodsa	Plasma PFOA, PFOS at 4–14 weeks of pregnancy	↑PFOA, PFOS and ↑irregular menstrual periodsb
Lyngsø 2014 [40]	Greenland, Poland and Ukraine, 2002–2004	Cross-sectional	Pregnant women INUENDO cohort obtaining prenatal care (n=1623)	Inclusion: Age ≥18 years, born in country where study conducted, Exclusion: Use of OCs within 2 months of becoming pregnant	Irregular, short, and long cyclesc	Serum PFOS, PFOA during pregnancyd	↑PFOS and irregular cycle; ↑PFOA and long cyclee
Lum 2017 [45]	Michigan and Texas, USA, 2005–2009	Prospective cohort (followed until pregnancy or 12 months of trying)	Female partners of couples in LIFE study who discontinued contraception to become pregnant (n=501)	Inclusion: In a committed relationship, ages 18–40, English or Spanish speaking, no use of hormonal birth control injection in past year, off contraception for ≤2 months, cycle lengths 21–42 days Exclusion: Clinically diagnosed infertility in either partner	Cycle length (continuous)	Serum PFOSA, Et-PFOSA-AcOH, Me-PFOSA-AcOH, PFDeA, PFNA, PFOS, PFOA at baseline	↑PFDeA and ↑cycle length; ↑PFOA and ↓cycle lengthe
Zhou 2017 [41]	Shanghai, China 2013–2015	Cross-sectional	Shanghai Birth Cohort Study participants (n=950)	Inclusion: Ages ≥20 years, registered Shanghai residents, no plans to move from Shanghai in next 2 years, discontinued contraception, planning to conceive naturally, planning birth at collaborating hospitals Exclusion: trying to conceive >1 year, sought care to conceive	Irregular, short, long cycles, hypomenorrhagia, menorrhagiaf	Plasma PFOA, PFOS, PFNA, PFHxS, PFDeA, PFUnDA, PFBS, PFDoA, PFHpA, PFOSA at enrollment	↑PFOA, PFOS, PFNA, PFHxS and ↑irregular cycles; ↑PFOA, PFOS, PFNA, PFHxS and ↑long cycles; ↑PFOA, PFOS, PFNA, PFHxS and ↓menorrhagia; ↑PFOA, PFNA, PFHxS and ↑hypomenorrhagiae
Singer 2018 [42]	Norway 1999–2008	Cross-sectional	Pregnant women in MoBa; data from two prior sub-studies (n=1977)	Case-control sub-study: Inclusion: Enrolled in 2003–2004, provided blood sample, delivered live-born child. Case-cohort sub-study: Inclusion: Nulliparous women with singleton pregnancies enrolled in 2003–2007; available plasma sample, no history of chronic hypertension.	Irregular, short, long cyclesg	Plasma PFOA, PFNA, PFDA, PFUnDA, PFHxS, PFHpS, PFOS at around 17–18 weeks pregnant	Data suggested short cycles and↓PFHpS and ↓PFOS, with stronger associations among parous participants. Data also suggested short cycles and ↓PFNA and ↓PFUnDA among parous participants and long cycles and ↑PFNA and ↑PFUnDA among recent OC users.
Wise 2021 [43]	Michigan, USA 2010–2012	Cross-sectional	Black women ages 23–35 years in SELF (n=1499)	Intact uterus, no prior diagnosis of uterine leiomyoma or medically-treated autoimmune disease or cancer	Cycle regularity, cycle length, bleeding duration, flow intensityh	Plasma PFHxS, n-PFOS, Sm-PFOS, n-PFOA, Sb-PFOA, PFNA, PFDA, PFUnDA, MeFOSAA	Heavy menstrual bleeding and ↓PFHxS, PFUnDA, PFOS, PFNA, PFOA, PFDA; generally little association with cycle irregularity, cycle length, and bleeding duration.

Abbreviations: Et-PFOSA-AcOH, 2-(N-ethyl-perfluorooctane sulfonamido) acetate; INUENDO, INUit-ENDOcrine; LIFE, Longitudinal Investigation of Fertility and Environment Study; Me-PFOSA-AcOH, 2-(N-methyl-perfluorooctane sulfonamido) acetate; MoBa, Norwegian Mother and Child Cohort; OC, oral contraceptive; PFBS, perfluorobutanesulfonate; PFDA, perfluorodecanoic acid; PFDeA, perfluorodecanoate; PFDoA, perfluor- ododecanoic acid; PFHpA, perfluoroheptanoicacid; PFHpS, perfluoroheptane sulfonate; PFHxS, perfluorohexanesulfonate; PFNA, perfluorononanoate [perfluorononanoic acid]; PFOA, perfluorooctanoate; PFOS perfluorooctane sulfonate; PFOSA, perfluorooctane sulfonamide; PFUnDA, perfluoroundecanoic acid; SELF, Study of Environment, Lifestyle & Fibroids;

^aIrregular menstrual period not defined.

^bBased on descriptive data only, considering PFAS chemicals as the exposure and irregular menstrual periods as the outcome; primary outcome of study was time-to-pregnancy.

^cIrregular cycle defined as ≥7 days of variation; short and long cycles defined as ≤24 days and ≥32 days in length.

^dSerum samples collected during pregnancy on average at 33 weeks pregnant (Poland) and 24 weeks pregnant (Ukraine, Greenland) [51].

^eInvestigated PFAS chemicals (exposure) in relation to menstrual cycle characteristics (outcome).

^fIrregular cycle defined as ≥7 days of variation from month-to-month; short and long cycles defined as average cycle length in past year <21 days and >35 days, respectively. Interview data collected on amount of menstrual bleeding (light, average, heavy, very heavy) used to define hypomenorrhea (light bleeding) and menorrhagia (heavy and very heavy menstrual bleeding).

^gIrregular cycles assessed by the question, “Were your periods regular the year before you became pregnant?”. Short and long cycle lengths defined as usual cycle length 17–24 days and ≥32 days, respectively.

^hCycle regularity was defined as “too irregular to say” when SELF participants were asked about usual cycle length. Bleeding duration defined as number of days of bleeding, but not spotting, at last menstrual period. Intensity of menstrual flow (light, moderate, heavy) estimated using data on pad and tampon product use and absorbency on heaviest day of bleeding at last menstrual period.

Cycle length

Across studies evaluating cycle length [40, 45, 41–43], four studies reported increased PFAS chemical concentrations with longer cycle length and lower chemical concentrations with shorter cycle length [40, 45, 41, 42] (Table 1). However, one of these studies, a prospective cohort study of women attempting to conceive, also reported increased PFOA concentrations and decreased cycle length [45]. In addition, a cross-sectional study reported little association between cycle length and PFAS concentrations [43]. The inconsistency in results across studies may be due to differences in the hypothesis being tested and accompanying study design; three studies investigated whether PFAS disrupts cycle length [40, 45, 41], whereas two studies postulated that increased frequency of bleeding from shorter cycles leads to increased PFAS elimination through menstrual bleeding [42, 43].

Menstrual flow heaviness

Consistent with menstrual blood loss as an excretion route for PFAS, the two studies to evaluate menstrual flow and PFAS concentrations measured in plasma reported inverse associations between PFAS concentrations and heaviness of menstrual flow [41, 43], with the pattern being lower PFAS concentrations with heavier flow [43] or increased PFAS concentrations with light menstrual bleeding [41] (Table 1).

Comments on epidemiologic studies of menstrual bleeding and PFAS concentrations

If we consider irregular menstrual periods and longer cycle length to indicate less frequent menses and decreased opportunity for PFAS excretion, the results across studies have been somewhat discrepant. However, studies of menstrual flow have reported consistent inverse associations with heavier flow being associated with lower PFAS concentrations, in alignment with menstrual blood loss being an excretion route for PFAS. It is important to emphasize that only two published studies reporting on the association between menstrual cycle characteristics and PFAS concentrations were designed with PFAS as the outcome [42, 43]. The other studies evaluated PFAS as the exposure [39, 40, 45, 41] and were interested in understanding the endocrine disruptive properties of PFAS on the menstrual cycle and fecundity. Although PFAS can both operate as an endocrine disruptor and be excreted by menstrual bleeding, the approach to study design to examine these biologic mechanisms should differ. As such, studies considering PFAS as the exposure were designed in alignment with the endocrine disruptive research focus and were conducted among pregnant women [39, 40] or women attempting to conceive [45, 41]. Only one study has been conducted among a general population of women [43].

Implications of PFAS excretion with menstrual bleeding in environmental epidemiology

Elimination of PFAS with menstrual blood loss has implications for environmental epidemiology, particularly given the amount of menstrual blood loss that may occur from menarche to menopause and the substantial portion of the population affected. A history of menstrual bleeding could affect the interpretation of measured PFAS concentrations. For studies of conditions with a long latency using a single measurement of PFAS to characterize exposure, it raises the question as to whether the measurement reflects PFAS concentrations over the etiologically relevant period for disease development. If the biologic sample used for measurement is collected close to disease diagnosis, the measurement error would be larger for those with a history of greater menstrual blood loss – particularly those with shorter cycle lengths, greater number of years of menstruating, and heavier menstrual flow.

Since menstrual blood loss is not a homogenous phenomenon, it may not be adequate to simply use sex and age to account for differences in measurements. In addition to individual variability in menstrual cycle length and flow, menstrual bleeding history is affected by pregnancy and lactation history as well as medications that alter menstrual flow. Combined oral contraceptives (those containing both estrogen and progestin components) have been observed to reduce menstrual blood loss between 40–50% [46]. Another hormonal contraceptive, the injectable contraceptive depot medroxyprogesterone acetate induces cessation of menses with >50% reporting amenorrhea after 12 months of use [47]. Reduction in menstrual blood loss and menstrual suppression can also occur with extended cycle or continuous use of combined oral contraceptives, progestin-only contraceptive pills, and intrauterine systems [46, 48]. Consistent with the reduction in menstrual blood loss, recent and longer duration of oral contraceptive use has been associated with higher PFAS concentrations [49]. Hence, characterizing menstrual blood loss and PFAS concentrations in menstruators is complicated. Similarly, depending on the research question, adjusting or stratifying on menopausal status may also be an oversimplification and not account for differences in PFAS measurement.

Yet, this information on menstrual bleeding may be critical for addressing measurement error to prevent the underestimation of the health impact of PFAS in those with a history of menstrual bleeding. This measurement error could be substantial if PFAS exposure increases the risk of an outcome that presents with increased menstrual blood loss, such as uterine leiomyomas, endometriosis, and adenomyosis, and specimen collection for PFAS measurement transpires after disease diagnosis.

Recommendations for future research

We propose the following recommendations to move the field forward:

Recommendation #1. Further epidemiologic research on menstrual bleeding and PFAS concentrations. Since only two studies to date have explicitly investigated heavy menstrual flow and PFAS concentrations in menstruators, detailed investigation of the key aspects of menstrual bleeding that contribute to excretion and lower PFAS concentrations is warranted. This detailed investigation should include a thorough examination of menstrual cycle characteristics and medication use that alter menstrual bleeding among a general population of menstruators. It should also include an investigation of the impact of menstrual history over the reproductive years on PFAS body burden.

Recommendation #2. Incorporate menstrual bleeding as a PFAS excretion route in epidemiologic study design. The consideration of menstrual bleeding as a route of PFAS elimination in epidemiologic study design is vitally needed. Careful consideration as to whether the timing of biologic sample collection and measurement of PFAS characterize exposure over the etiologically relevant period for disease development is warranted for those with a history of menstrual bleeding. Given the individual variability in menstrual bleeding, epidemiologic studies would benefit from the collection of detailed menstrual bleeding data, including menstrual characteristics, pregnancy and lactation history, and data on medication use that affects menses. The incorporation of menstrual bleeding as an excretion route is pertinent in the study design of outcomes with a long latency in which specimen collection for PFAS measurement occurs close to diagnosis. It is also critical for epidemiologic studies of gynecologic conditions associated with heavy menstrual bleeding, such as uterine leiomyoma, endometriosis, and adenomyosis [50]. The conduct of simulation studies would be informative to understand the impact of menstrual bleeding on estimated associations between PFAS exposure and health outcomes as well as to assess the direction and magnitude of bias under different scenarios.

Recommendation #3. Understand the extent of PFAS excretion via menstrual bleeding and expand the investigation to other chemicals that may be excreted in menstrual fluid. Empirical data are needed on PFAS concentrations in menstrual fluid to estimate the extent of PFAS elimination with menstrual bleeding. This research should also extend to other chemicals primarily distributed in the blood. We focused on long-chain “legacy” PFAS in this review as most evidence to date has considered the excretion of these chemicals via menstrual blood loss. However, the impact of menstrual bleeding on the excretion of emerging short-chain PFAS and other environmental chemicals is ripe for exploration.

Recommendation #4. Investigate the impact of menstrual blood loss on biomonitoring and internal dose assessment of environmental chemicals. Informed by data on the identified key aspects of menstrual bleeding that contribute to excretion (Recommendation #1) and the extent of excretion of PFAS and other environmental chemicals by menstrual blood loss (Recommendation #3), investigation of the impact of menstrual blood loss on biomonitoring and internal dose assessment of environmental chemicals will be critical to refining our understanding of population exposure, particularly for the substantial portion of the population with a history of menstrual blood loss.

Conclusions

Menstrual bleeding is a common, regular occurrence in the population. Epidemiologic data provide compelling evidence for the role of menstrual bleeding on PFAS elimination. Studies over the past decade have generally reported lower PFAS body burden among individuals currently menstruating and those with heavier menstrual flow. This has implications for epidemiologic research; depending on the outcome of study and timing of exposure measurement, menstrual blood loss may introduce measurement error, leading to a disparity in the estimation of the health impact of PFAS in those with a history of menstrual bleeding. Given these implications, further research is warranted on menstrual bleeding as an elimination route for PFAS in environmental epidemiology, and we provide recommendations to move the field forward.