In-utero exposure to polybrominated biphenyl (PBB) and menstrual cycle function in adulthood.

The Michigan Department of Community Health (now the Michigan Department of Health and Human Services; MDHHS) established the Michigan Long-Term PBB Study (now the Michigan PBB registry) between 1976 and 1977 ( Carter 1976 ; Fries 1985 ). The original cohort enrolled over 4800 individuals who lived on contaminated farms, ate food from contaminated farms or who worked at the chemical plant that produced PBB in Michigan ( Landrigan et al. 1979 ) and their family members. MDHHS ceased management of the registry in 2003 and Emory University took over the management and expanded the original cohort ( Chang et al. 2020 ). Today, the Michigan PBB registry’s research is guided by a community-academic partnership between several community groups who represent those affected and Emory University. The registry includes people exposed to PBBs through occupational settings and/or consumption of contaminated food products and their children and grandchildren ( Chang et al. 2020 ; Fries 1985 ). A subset of women from the Michigan PBB registry were recruited to participate in a longitudinal study on menstrual cycle function ( Howards et al. 2019 ). In brief, women from the Michigan PBB registry who were aged 18–45 years (i.e. born before and after the contamination incident), premenopausal, not pregnant or lactating, not currently taking hormonal medications, and never diagnosed or treated for cancer were eligible. Women were recruited in two phases: between 2004 and 2006 (Phase 1) and between 2013 and 2014 (Phase 2). Women who participated in the study had a non-fasting venous blood draw (10 mL) and completed a health questionnaire that included medical history, current medication use, behaviors, and demographics. Participants also provided daily urine collections and daily diaries for menstrual cycle function monitoring. Phase 1 women completed a computer-assisted telephone interview and Phase 2 women completed a self-administered web-based in-depth health questionnaire for females to gather more information about reproductive history. Both questionnaires collected baseline data on medical history and reproductive function and were used to determine study eligibility. In Phase 1, 479 women of 711 eligible women (18–45 years old) completed the interview and were all screened for the study ( Figure 1 ). Of these, 314 were deemed eligible, and 133 provided sufficient urine and diary data. In Phase 2, 152 women completed the questionnaire of the 172 eligible women for the study. Of these, 87 women were screened and deemed eligible, and 58 provided sufficient data. Of the women who provided sufficient data, 5 participated in both phases with 1 being born after exposure. From this pool of 191 women who provided sufficient data, only 62 women (33 from Phase 1 and 29 from Phase 2) were born after the contamination incident and were included in this analysis. Since the effect of PBB exposure in-utero may differ from exposure through diet, we focused on women prenatally exposed who also had maternal PBB levels available. The average age of the women’s mothers at the time of the PBB contamination event was 19.4 years (Range: 8.2 to 30.6 years). The women who were excluded were potentially exposed directly through consuming contaminated farm products in childhood. The majority of these women were included in a previous analysis from the Michigan PBB registry ( Howards et al. 2019 ) that investigated PBB childhood exposure and menstrual cycle function. All 33 of the women from Phase 1 (6.9%) with complete data were included for our analysis, but of the 32 women in Phase 2, maternal PBB exposure levels were available for only 9 (5.2%). Thus, the final sample size for our analysis was 41 women (32 women from Phase 1, 8 women from Phase 2, and 1 woman who contributed cycles during both phases). This study was approved by the Institutional Review Boards at Emory University and the Michigan Department of Health and Human Services. Study participants completed daily diaries for up to 6 months and collected first morning urine samples for up to four menstrual cycles. Participants answered diary questions at approximately the same time each day and diary cards were mailed to the study site weekly. In Phase 1, participants recorded bleeding or spotting patterns, sexual intercourse (including birth control use), exercise, cigarettes smoked, consumption of alcoholic or caffeinated beverages, and other conditions like stress (using 4-point scale), fever, or any illnesses. A comments section was included for any additional information on medications or vitamins taken, if a pregnancy test was taken, or if any other explanation was required. In Phase 2, a similar diary booklet was used with slight alterations to the questions: inclusion of a question regarding menstrual cramping and the removal of questions regarding caffeinated beverages, fever/illness, sexual intercourse, and birth control use. Because of the differences in the diary cards, only the common questions between the two phases were utilized in the current analysis. Urine samples were collected in the morning in pre-numbered vials and stored immediately in the participants’ personal freezer until sent out for laboratory analysis. Participants were also asked to write a note on the diary card if anything went wrong with the sample, e.g., if the sample was collected late, not frozen immediately, or if the vials were used out of numerical order. Urine samples were assayed for primary estradiol and progesterone metabolites, estrone 3-glucuronide (E 1 3G), and pregnanediol 3-glucuronide (Pd3G). The protocols for the urine sample analyses was slightly different in Phase 1 and 2. In Phase 1, all urine samples in the 17-day window around expected ovulation were analyzed for E 1 3G and Pd3G. We defined day of ovulation as the day of luteal transition, which was determined by an algorithm using the ratio of E13G to Pd3G. We also measured E 1 3G, Pd3G, and FSH in urine samples during the 5 days before menses onset through the first 5 days of the new cycle for consecutive cycles to evaluate the luteal-follicular transition. In Phase 2, all urine samples from 3 menstrual cycles were analyzed for E 1 3G and Pd3G; plus, FSH was measured in a 14-day window (5 days before menses onset through the first 9 days of the new cycle) during the luteal-follicular transition. Urinary E 1 3G and Pd3G were measured in triplicate using competitive double-antibody time-resolved fluoroimmunoassays ( Kesner et al. 1994 ). Urinary FSH was assayed in duplicate using immunofluorometric assays (PerkinElmer, Waltham, MA, USA; Cat. Nos. A031–101 and A017–201, respectively) modified and validated for analyzing urine samples ( Kesner et al. 1994 ). To adjust for the concentration of the urine samples, we measured creatinine in all samples using a Vitros 250 Chemistry Analyzer (Ortho-Clinical Diagnostics, Raritan, NJ). Since PBBs can cross the placenta, we assigned exposure to PBBs in-utero based on the estimated maternal PBB levels at the time of pregnancy using a validated model of PBB elimination ( Terrell et al. 2008 ). In brief, participants in the menstrual cycle function study were connected to their mothers in the PBB registry through a maternal PBB ID. At the mother’s enrollment into the PBB registry (from 1976–1978), she provided a blood sample which was analyzed for serum PBB levels using gas chromatography with electron capture detection (min: non-detectable; median: 2.0 ppb; P95: 120.5 ppb; max: 251.7 ppb). The average time between the exposure event and the mothers’ initial measurement was 3.2 years (Range: 3.0 to 4.0 years) and the average time between the mothers’ initial measurement and their conception date for the participants was 1.1 years (Range: −4.0 to 13.0 years). At that time, the limit of detection (LOD) for PBB was 1.0 part per billion (ppb) and the coefficients of variation for PBB quantification ranged from 7–14% ( Burse et al. 1980 ; Needham et al. 1981 ). A validated mixed effects elimination model was then used to estimate the mother’s serum PBB level at time of pregnancy based on a general linear mixed model, which attributes unique intercept and slope estimates for each woman and has been described elsewhere ( Terrell et al. 2008 ). Briefly, the model calculates a subject-specific rate of elimination of serum PBB levels based on age and BMI at initial measurements as time-independent covariates and times since exposure, smoking history, pregnancy and breastfeeding status as time-dependent covariates. These parameters are then used to estimate serum PBB levels outside of the range of measurements. For example, if a participant’s elimination rate, or half-life, is 12 years, and their observed serum PBB level is 2 ppb in 1990, we would estimate their value in 1978 to be 4 ppb. In our study, 45% of the conception dates occurred before the mother’s initial measurements while the remaining 55% of conceptions occurred after the initial measurements.. The validated elimination model was confirmed by comparing results from a previously developed ordinary least squares (OLS) model ( Terrell et al. 2008 ). For purposes of analysis, we categorized in-utero PBB exposure into low (PBB ≤ 1.0 ppb; values at or below LOD), medium (PBB > 1.0–3.0 ppb) and high exposure (PBB > 3.0 ppb) based on previous analyses in the Michigan PBB registry ( Howards et al. 2019 ). Menstrual cycle function outcomes included cycle-level characteristics: cycle length, menses length, and follicular and luteal phase lengths, and 12 endocrine endpoints. All of these 16 endpoints were determined by a combination of diary data and urinary hormone levels. In this study, onset of menses was defined using an algorithm requiring two consecutive days of bleeding where one of the days may be spotting. The first and last day of menses had to be preceded and followed by at least three days without bleeding, respectively. Bleed length was defined as the onset of menses to the day before the three consecutive days without bleeding. If this three-day rule was broken, the duration for menses was not calculated. This algorithm used to determine menses length reliably distinguishes mid-cycle spotting from onset of menses for most women ( Jukic et al. 2008 ). Cycle length was defined as the number of days from the first day of one menses through the day before the onset of next menses. The follicular phase length was defined as the first day of menses through the day of ovulation. The luteal phase length was the day after ovulation through the day before menses onset. Day of ovulation was based on identifying a day of luteal transition (DLT), which was determined by an algorithm examining changes in the ratio of E 1 3G to Pd3G ( Baird et al. 1991 ). Additional samples were analyzed if DLT was ambiguous to investigate whether ovulation occurred outside the 17-day window and if luteal phase days 5 and 6 were outside this window. If no DLT was able to be identified and there were adequate urine samples, the cycle was classified as anovulatory. Of the 193 contributed menstrual cycles, 76 were missing urine samples that prevented us from determining the DLT. Among the 117 remaining cycles, 2 cycles did not meet the DLT criteria, but had adequate urine samples, and were classified as anovulatory. Both of these cycles belonged to women in the medium exposure group (PBB > 1.0–3.0 ppb) with above average cycle length (41 and 43 days, respectively). Cycle length was classified as missing when the day of one of the bracketing menses onsets was missing, luteal and follicular phase lengths were classified as missing for cycles without a known DLT or known day of menses onset. All of the 12 hormone outcomes were 3-day geometric mean hormone levels, calculated during six timeframes. The maximum geometric mean was calculated by identifying the maximum value in the relevant timeframe and then calculating the geometric mean of that day, the day before, and the day after. Early follicular phase levels were calculated as the geometric mean for cycle days 2–4; preovulatory levels were based on the 3 days prior to the DLT, mid-luteal phase levels were based on days 5–7 of the luteal phase, and late luteal phase levels were based on the last 3 days of the cycle. Geometric means for these 12 hormone outcomes were only calculated when hormone data were available for all 3 days, and the preovulatory and luteal phase variables were only calculated when the cycle had a defined DLT. These hormone outcomes were adapted from definitions proposed by Baird et al. that were shown to be related to conception ( Baird et al. 1991 ). Although we had 41 women in our analytic sample, the sample sizes for each hormone analysis varied mostly due to women missing single days of urine collection. We summarized participant characteristics according to their mother’s estimated serum PBB level when the participant was in-utero. We assessed confounding using a priori knowledge in combination with directed acyclic graphs (DAGs) and descriptive statistics from the Michigan PBB registry. Since our exposure was in-utero PBB exposure, many variables such as current smoking status, gravidity, and body mass index (BMI) at interview, were not identified as potential confounders since they were downstream of exposure and left out of the final multivariable models. In addition, since all our study participants were White, we did not adjust for race. Due to the low number of mothers who reported smoking during pregnancy, we were unable to adjust for this variable in the models. All models, however, were adjusted for age of the study participants because the cycle and hormonal outcomes are known to change with increasing age. We fit linear mixed models with a random effect for woman in order to account for the intra-individual correlations among multiple menstrual cycles per woman. The models included fixed effects for categorized in-utero PBB exposure and age as a continuous variable centered on the mean age of the study population. We output predicted means for each of the outcomes by PBB exposure level for the average age of the women in the study (27.5 years).

The 41 women in our study contributed a total of 193 menstrual cycles of follow-up. The mean number of cycles contributed was 4.7 (range: 1 to 8). The number of women and cycles utilized in the analysis varied by outcome, ranging from 23 women and 42 cycles for average E 1 3G in the follicular phase to 39 women and 143 cycles for bleed length. The majority of women were younger than 35 years (90%), had a normal BMI (56%), were employed at least part-time (73%), had at least some college education (84%), were never smokers (78%), and were nulligravid (54%) ( Table 1 ). Only three women were exposed to maternal smoking in-utero. Women with medium and high exposure to PBB in-utero had, on average, slightly higher BMIs in comparison to the women with low in-utero PBB exposure. There were some other demographic and lifestyle characteristics that varied across categories of in-utero PBB exposure, however, none did so in a consistent direction. Between the two phases, those in phase 1 tended to be more highly exposed (>3.0 ppb: 45.5%) compared to phase 2 participants (>3.0 ppb: 22.2%) and phase 2 participants tended to be older (36–40 years: 44.5%) than phase 1 participants (36–40 years: 3.0%) ( Supplemental Table 1 ). Age- and creatinine-adjusted adjusted Pd3G concentrations were slightly lower among women in the lowest category of in-utero PBB exposure as compared to women with medium or high in-utero PBB exposure for all time windows ( Table 2 ). However, we only observed a meaningful, monotonic trend across categories of in-utero PBB exposure for 3-day mean luteal phase maximum concentrations. Specifically, the age- and creatinine-adjusted 3-day mean luteal phase maximum urinary Pd3G levels (95% CI) in increasing categories of in-utero PBB exposure were 9.2 (4.6, 13.9), 14.8 (11.6, 18.0), and 16.1 (12.9, 19.3) ug/mg creatinine. Women in the medium category for in-utero PBB exposure had a significantly higher age- and creatinine-adjusted mean Pd3G levels during the last 3 days of the menstrual cycle (10.4 ug/mg) as compared to women with low exposure (6.3 ug/mg), while women in the highest exposure category had intermediate levels (8.3 ug/mg). After further adjustment for current BMI, which may or may not be on the causal pathway between in-utero PBB exposure and urinary progesterone levels, we still observed statistically significant associations ( Supplemental Table 2 ). There were no major differences in luteal phase length or menses length by categories of in-utero PBB exposure ( Table 2 ). Cycle length and follicular phase length were slightly shorter among women with low in-utero PBB exposure levels, compared to women with medium exposure, although the 95% CIs were wide. We observed slightly higher age- and creatinine-adjusted 3-day mean follicular and luteal phase maximum E 1 3G concentrations among women with the highest exposure to PBB in-utero as compared to women with medium and low in-utero PBB exposure; however, these differences were imprecise. There were no noticeable differences in age- and creatinine-adjusted mean E 1 3G concentrations during days 2–4 of the cycle, 3 days prior to DLT, during luteal days 5–7, and during the last 3 days of the cycle across categories of in-utero PBB exposure. There were also no differences in 3-day mean urinary FSH concentrations across in-utero PBB exposure levels during days 2–4 and the last 3 days of the menstrual cycle.

In our prospective study of 41 women exposed in-utero to PBB through a food contamination event in Michigan in the 1970s, we found suggestive evidence that higher in-utero exposure to PBB was associated with slightly higher maximum progesterone levels during the luteal phase of the menstrual cycle. Other menstrual cycle characteristics, however, were not statistically significantly associated with in-utero PBB exposure. Observing elevated progesterone levels is interesting given its potential association with increased risk of breast cancer ( Coelingh Bennink et al. 2023 ). In the Michigan PBB registry, we have observed evidence for an association between direct PBB exposure during childhood and adulthood and higher risk of breast cancer ( Henderson et al. 1995 ; Terrell et al. 2016 ). Although speculative, our results could be one of many potential mechanisms through which PBB exposure leads to elevated breast cancer risk. In addition, we observed some differences in adult demographic characteristics by in utero PBB exposure. For example, we observed those with medium and high in utero exposure to PBB had a slighter higher adult BMI compared to those with low in-utero exposure. These findings are particularly interesting given the potential association between obesity and progesterone ( Faulkner et al. 2019 ). To our knowledge, only one other study has investigated PBB exposure and menstrual cycle function ( Howards et al. 2019 ). This previous study came from the same Michigan PBB registry of women eligible for this study, but instead focused on women who had been exposed to PBB through diet during childhood. Howards et al, found that women with high (>3.0 ppb) PBB exposure during childhood had lower E 1 3G and Pd3G levels across the menstrual cycle and lower FSH levels during the follicular phase as compared to women with low childhood PBB exposure (≤1.0 ppb) ( Howards et al. 2019 ). Those findings for Pd3G levels, in particular, are quite different from what we observed in this study. This may not be entirely unexpected as exposures experienced in-utero are often hypothesized to have different biological mechanisms underlying their associations with adult reproductive function as compared to direct exposures experienced during childhood. 17 It’s also worth noting that in addition to the difference in primary route of PBB exposure between the two studies, the women in our study were, on average, 10 years younger than the women exposed during childhood. Therefore, it’s hard to completely rule out differences in results that may be due to effect modification by age. In other words, if PBB exposure (regardless of the timing) has a differential impact on menstrual cycle function as women age, it would be challenging to differentiate this effect from effects due to differing routes of exposure since the range of ages in Phase 1 and 2 did not overlap. There is also a limited, but relevant, literature on the impact of in-utero exposure to similar persistent, endocrine disrupting chemicals such as polychlorinated biphenyls (PCBs) and per- and polyfluoroalkyl substances (PFAS) on menstrual cycle function. For example, a comparable study from Taiwan, which evaluated menstrual cycle function in adolescent daughters of women exposed to PCB-contaminated cooking oil found that higher in-utero PCB exposure was associated with increased estradiol and FSH levels and shortened bleeding periods ( Yang et al. 2005 ). In contrast, when the exposed mothers were examined, the authors found very few differences in menstrual cycle function associated with PCB levels, with the exception of longer bleeding periods ( Yang et al. 2011 ). These two studies, which found differing results following in-utero versus direct exposure to high levels of PCB, provide additional evidence that the route and timing of exposure to persistent endocrine disrupting chemicals may result in differing effects on menstrual function. However, similar to our dataset, these studies were small and the differences could be due to small sample size and the wide natural variability of hormones. There have also been multiple studies on the association between in-utero exposure to PFAS and reproductive function in childhood and adolescence. These studies tended to focus on slightly different outcomes, but the results have shown that higher in-utero PFAS exposure was associated with delayed menarche ( Kristensen et al. 2013 ), increased testosterone concentrations ( Maisonet et al. 2015 ), and reduced DHEA concentrations ( Jensen et al. 2020 ) in girls. Multiple biological explanations have been proposed to explain why reproductive hormones and menstrual cycle function may be affected by in-utero exposure to endocrine disrupting chemicals like PBB. For example, a study in rats found that higher in-utero exposure to brominated flame retardants, the same class of chemicals as PBB, was related to early onset of puberty and increased incidence of multi-oocyte follicles and that this was likely due to the downregulation of pathways that are fundamental for ovarian function like the HIF1A, CREB1, EGF, β- estradiol, and PPAR pathways ( Allais et al. 2020 ). While their study did not show any significant differences in progesterone levels according to in-utero exposure to brominated flame retardants, any exposure impacting ovulatory function would likely have downstream effects on progesterone production. Regarding transgenerational effects of exposure to other endocrine disrupting chemicals, a study in pregnant rats found that maternal exposure to imazalil, a fungicide that is also an androgen receptor antagonist, was associated with increased androgen levels in the mothers, but decreased androgen levels in male offspring ( Jin et al. 2019 ). This finding in animals further supports the notion that exposure to endocrine disrupting chemicals during pregnancy may induce hormonal changes in future generations that could be opposite to the effects observed in the initial generation. However, in contrast to our results, two studies - one focused on prenatal phthalate exposure ( Li et al. 2020 ) and the other on prenatal PCB & DDT exposure in rats ( Jonsson et al. 1975 ), showed a decrease in progesterone concentrations in the F1 and F2 generations with increasing exposure to these chemicals. While we observed the opposite effect, this does provide evidence that exposure to endocrine disrupting chemicals may lead to an alteration in progesterone receptors or the hypothalamic-pituitary-ovarian axis. As these were animal studies, timing and dose of exposure are hard to directly compare between these studies and ours but may be a critical consideration. For example, a study on the action of PCB congeners on proliferation and progesterone secretion in cultured in vitro porcine luteal cells showed a concentration dependent decrease in progesterone secretion after 24 and 48 hours PCB153 exposure and a concentration dependent increase in progesterone secretion after 72 hours of exposure( Augustowska et al. 2001 ), suggesting that duration of EDC exposure may play a pivotal role in the type of hormonal effect it has. One of the primary limitations of our study was the small sample size. Given our strict eligibility criteria and our rigorous study protocol, which required women to complete daily diaries, provide daily urine samples, and show up for an in person visit for a blood draw, we had a limited number of participants that were eligible and willing to participate in the study. Therefore, our results must be interpreted with caution. Women experience natural variation in menstrual cycle characteristics across cycles, so it is difficult to distinguish, in small studies like ours, whether the observed patterns are driven by differences in exposure between women or are merely due to chance (e.g., an artifact of the specific cycles we included for each woman) or unmeasured confounding. While we partially addressed this by including multiple cycles per woman and using marginal repeated measures linear models to account for the inherent variability in menstrual cycle function within a woman, the overall power of our study was still limited. In addition, we utilized strict eligibility criteria because several factors (e.g., hormonal contraception, endometriosis, cancer diagnosis and treatment, etc.) can influence the menstrual cycle. Therefore, our results may not be generalizable to individuals with these conditions or treatments and would miss any potential associations between PBB and menstrual cycle dysfunction mediated by these conditions. Conversely, because of our low power, it is possible that we failed to detect small but clinically meaningful differences in menstrual cycle characteristics by levels of in-utero PBB exposure. An additional weakness was that maternal PBB concentrations were not directly measured during pregnancy but rather estimated using a PBB elimination model ( Terrell et al. 2008 ), which likely led to measurement error of the exposure. Given the prospective nature of our study, however, it is highly unlikely that this error was differential with respect to menstrual cycle function. Our sample size varied due to missing data, most often due to missing urine samples, which further limited the precision of some analyses. Lastly, it is possible that women included in our sample were going through perimenopausal shift which would have alter cycle length or regularity. However, given that 90% of our sample was 35 or younger (n=37) and thus far there is no evidence to suggest that PBB increases the risk of early menopause, it is unlikely that many women, if any, were perimenopausal. Regarding generalizability, the estimated maternal PBB levels for the study participants were, on average, much higher than would be expected in the general population. For example, only ~15% of our study participants’ mothers had PBB levels less than 1.0 ppb, which was the limit of detection at the time the assays were performed. For comparison, the geometric mean PBB level among female participants in the 2003–2004 National Health and Nutrition Examination Survey (NHANES) was 0.012 ppb (95% CI: 0.009 to 0.015), which is well below even the average PBB level in our lowest exposure group (0.58, 95% CI: 0.44 to 0.96, N=10) ( Sjodin et al. 2008 ; Statistics 2018 ). It is possible that heterogeneity in our low exposure group may have masked differences between the lowest and highest exposure groups. Because of the high levels of in-utero PBB concentrations observed in our study population, our results may not be directly generalizable to most populations beyond the daughters of affected residents of Michigan. In conclusion, women who were exposed to higher levels of PBB in-utero had slightly higher urinary progesterone metabolite levels across the luteal phase of the menstrual cycle compared to women with the lowest in-utero exposure to PBB. Most other menstrual cycle characteristics, including cycle length, cycle phase lengths, menses length, and urinary concentrations of estrogen and FSH, however, were largely unassociated with in-utero PBB exposure. Given our modest sample size, our results require cautious interpretation. Furthermore, because there are no standardized norms to compare our urinary hormone concentrations to, it is challenging to discern whether our findings for urinary Pd3G are being driven by abnormally low levels in women in the lowest exposure group or truly higher than average high Pd3G values in the medium and high exposure groups. While the production of PBB has decreased or ceased in most countries, our results may be still relevant due to the continued production of chemically related brominated flame retardants worldwide.

Polybrominated biphenyls (PBBs) are a class of highly stable brominated flame retardants that were once used in the manufacturing of plastics and electronics. Although the production of PBBs has ceased in the United States as of the 1970s, health concerns remain due to their persistence in the environment, their ability to accumulate in food products, and their long biological half-life ( Agency 2017 ; Hood et al. 2023 ; Terrell et al. 2008 ). There is also evidence that PBB has estrogen-mimicking endocrine disrupting properties ( Curtis et al. 2019 ), which could have long-term impacts on individuals exposed in-utero ( Small et al. 2011 ). Previous studies have already demonstrated that those exposed to PBB, and its endocrine disrupting properties, in-utero are at higher risk for lower birthweight ( Givens et al. 2007 ), low 1-minute Apgar score ( Terrell et al. 2015 ), earlier menarche ( Blanck et al. 2000 ), and adverse pregnancy outcomes ( Small et al. 2011 ). The Michigan PBB registry is one of the longest-running cohort studies in the country and was established to examine the health effects of the widespread environmental contamination with PBBs. During the early 1970s, PBBs were accidentally introduced into the food supply when it was shipped in place of an animal feed supplement. Research based on the Michigan PBB registry has shown associations between exposure to PBB and many reproductive health outcomes ( Small et al. 2011 ). In 2005, a menstrual function prevalence study among women exposed to PBBs from their diet suggested that higher exposure was associated with shorter menstrual cycle length and longer bleed length among women with past year weight loss ( Davis et al. 2005 ). A follow-up prospective study published in 2019 showed that PBB concentrations measured during early childhood were associated with lower concentrations of endogenous estradiol metabolites throughout the menstrual cycle in adulthood ( Howards et al. 2019 ). Because the Michigan PBB registry is multigenerational, there is the rare opportunity to use the data collected from the registry to assess health outcomes in individuals exposed in-utero. There is evidence that in-utero exposure to PBBs is associated with health outcomes like increased odds of spontaneous abortion ( Small et al. 2011 ), delayed puberty in males( Small et al. 2009 ), and earlier age at menarche ( Blanck et al. 2000 ). However, there is limited research on the effects of in-utero exposure to PBBs or similar chemicals, specifically for menstrual cycle function and reproductive hormone levels (sex steroids and gonadotropins). Therefore, our objective was to examine the association between in-utero exposure to PBBs and menstrual cycle function in adulthood.