Urinary Phthalate Metabolite Concentrations and Hot Flashes in Women from an Urban Convenience Sample of Midlife Women

Abstract

Context:

Phthalate exposure is associated with altered reproductive function, but little is known about associations of phthalate exposure with risk of hot flashes.

Objective:

To investigate associations of urinary phthalate metabolite levels with four hot flash outcomes in midlife women.

Design and Setting:

A cross-sectional study of the first year of a prospective cohort of midlife women, the Midlife Women’s Health Study (2006–2015), a convenience sample from an urban setting.

Participants:

728 multi-racial/ethnic pre- and perimenopausal women aged 45–54 years.

Outcome Measures:

Women completed questionnaires about hot flash experience and provided 1–4 urine samples over four consecutive weeks that were pooled for analysis. Phthalate metabolites were assessed individually and as molar sums representative of common compounds and exposure sources (plastics, ΣPlastic). Covariate-adjusted logistic regression models were used to assess associations of continuous natural log-transformed phthalate metabolite concentrations with hot flash outcomes. Analyses were conducted to explore whether associations differed by menopause status, body mass index (BMI), race/ethnicity, and depressive symptoms.

Results:

Overall, 45% of women reported a history of hot flashes. Compared to women who never experienced hot flashes, every two-fold increase in ΣPlastic was associated with 18% (OR: 1.18; 95%CI: 0.98, 1.43) and 38% (OR: 1.38; 95%CI: 1.11, 1.70) higher odds of experiencing hot flashes in the past 30 days and experiencing daily/weekly hot flashes, respectively. Some associations of phthalates with certain hot flash outcomes differed by menopause status, BMI, race/ethnicity, and depressive symptoms.

Conclusions:

This study suggests that phthalates are associated with hot flash experience and may impact hot flash risk in women who are susceptible to experiencing hot flashes.

Graphical Abstract

1. Introduction

Hot flashes are one of the most common symptoms of menopause, but little is known about the risk factors associated with increased risk of hot flashes. Hot flashes are characterized by sudden and transient periods of intense body heat accompanied by flushing, sweating, chills, and anxiety (1). Experiencing hot flashes can impact daily life for symptomatic women for years and results in estimated medical costs of $340 million in the U.S. each year and an additional $27 million in lost work (2,3). Although the majority of peri- and postmenopausal women experience hot flashes, the dynamics of hot flashes during the menopause transition (such as age of onset, duration, intensity, and risk factors) are not well understood (2). Environmental factors (e.g. smoking), physiological factors (e.g. later stage of menopause), and decreasing estrogen levels are known to be associated with increased risk of hot flashes (4). Our own small cross-sectional analysis of a representative sample of 195 white midlife women from the Midlife Women’s Health Study (MWHS) indicated that exposure to phthalates may be associated with an increased odds of hot flashes in midlife women (5).

Phthalates are a class of synthetic chemicals composed of esters of ortho-phthalic acid with hydrocarbon side chains of varying lengths. Phthalates are used in a wide variety of consumer products, including food contact materials, medical equipment, car interiors, shower curtains, synthetic leather, and children’s toys, as well as fragranced cleaning and personal care products (6). Humans are ubiquitously and unavoidably exposed to phthalates, with 99% of urine samples from the general U.S. population containing phthalate metabolites (7). Due to greater use of personal care products by women compared to men, women typically have higher concentrations of phthalates than men (8,9).

Importantly, studies show that some phthalate metabolites exert toxicity on biological systems including the reproductive system (10–15). In both experimental and observational studies, phthalates have been shown to alter estradiol levels (13,16–18). Animal studies have also shown that mixtures of phthalates can impact hormone levels and steroidogenesis in the ovary (18–20). This is concerning given that midlife women are widely exposed to phthalates through diet and personal care products (11,14,18,21). The molecular mechanisms of action of phthalates to cause hormone disruption are hypothesized to occur through activation of peroxisome proliferator-activated receptors (13,18).

In general, factors that decrease estrogen levels in women are strongly associated with increased incidence of hot flashes (reviewed in 4). Extensive literature suggests associations between estrogen levels and hot flashes in conditions that cause acute drops in estrogen levels, such as oophorectomy (22). As phthalates are associated with decreased estrogen levels in human and animals, we hypothesized that phthalate exposure may contribute to hot flash experience in women. To our knowledge, no studies have evaluated the impact of phthalate exposure on hot flash risk in women in detail. However, this study expands upon our previous pilot study (5) to include the entire cohort of the MWHS, which contains multiple racial/ethnic groups, and includes additional analyses based on participant characteristics. Therefore, the primary objective of this study was to assess associations of urinary phthalate metabolite levels with hot flash occurrence, frequency, and severity in midlife women enrolled in the first year of the MWHS. Because risk of hot flashes may differ in women based on menopausal status, midlife body mass index (BMI), race/ethnicity, and depression status (4,23), the secondary objective of this study was to evaluate differences in associations of urinary phthalate metabolite levels with hot flash risk by these characteristics.

2. Materials and Methods

2.1. Ethical approval

All participants gave written informed consent according to procedures approved by the University of Illinois and Johns Hopkins University Institutional Review Boards (file number: 06741).

2.2. Study population

This study was a cross-sectional analysis of data collected in the first year of the MWHS, a prospective cohort study with the overall goal of evaluating risk factors of hot flashes in midlife women. A detailed study protocol of the MWHS has been published previously (24). Briefly, participants were recruited from the city of Baltimore, MD (USA) and surrounding counties from 2006 to 2015. Women were eligible to participate in the study if they were 45–54 years old and pre- or perimenopausal with or without natural hot flashes. Women were excluded if they had a history of hysterectomy or oophorectomy, were currently pregnant, were taking hormone therapy or herbal/other agents for treatment of menopause symptoms, were taking oral contraceptives, were undergoing cancer treatment, or were postmenopausal. Menopausal status was defined using the Stages of Reproductive Aging Workshop + 10 (STRAW+10) criteria (25). Briefly, menopausal status was defined as follows: pre-menopausal women were those who experienced their last menstrual period within the past 3 months and reported 11 or more periods within the past year. Perimenopausal women were those who experienced their last menstrual period within the past year, but not within the past 3 months, or their last menstrual period within the past 3 months and experienced 10 or fewer periods within the past year. Postmenopausal women were those who had not experienced a menstrual period within the past year. A total of 780 women enrolled in the study during year 1.

2.3. Collection of demographic and lifestyle characteristics

At the baseline clinic visit, women completed a detailed questionnaire and had anthropometrics measured by trained staff. Each woman’s weight and height (without shoes) were measured by trained clinic staff, and values were rounded to the nearest 0.5 pound and 0.5 inch, respectively. The baseline questionnaire collected detailed information on demographics, reproductive history, menstrual cycle characteristics, menopausal symptoms, and medical history, as well as physical activity, smoking status, and alcohol use. Women self-reported their age in years and listed the types of prescription medications used. Each woman’s racial/ethnic background was determined using the question “What is your main ethnic/racial background? (Answer, mark only one: (1) Caucasian/White, (2) African American/Black, (3) Hispanic/Latino, (4) Asian, (5) Other)”. Women reported their highest completed grade or year of schooling using the following options: (1) elementary, (2) high school, (3) technical school, (4) college training, or (5) postgraduate. Smoking status was ascertained using the questions “Have you ever smoked cigarettes?” and “Do you still smoke cigarettes?” whereas the question “In the last 12 months have you had at least 12 drinks of any kind of alcoholic beverage?” was used to determine women’s most recent alcohol consumption status. Leisure physical activity was assessed with the question: “In comparison with others my own age, I think my physical activity leisure time is” [choices: much more, more, as much, less, and much less]. Women’s depression status was assessed using the Centers for Epidemiologic Studies Depression Scale (CESD) (26), which is a validated depression score that was calculated using 20 questions that asked about how the women were feeling during the past week.

2.4. Collection and assessment of hot flash outcomes

At baseline, a detailed history of hot flashes was collected using a series of validated questions that have been used in the MWHS for over 10 years (23,24,27–30). The current study evaluated four hot flash outcomes that were obtained from the following four questions: 1) whether the woman had ever experienced hot flashes, 2) whether she experienced hot flashes in the past 30 days, 3) the usual severity of her hot flashes, and 4) the usual frequency of her hot flashes. Women were first asked “Have you ever had hot flashes?” where hot flashes were defined as “a sudden feeling of heat in the face, neck, or upper part of the chest” with accompanied “reddening or flushing of the skin followed by sweating and chills.” Women who responded “no” to ever experiencing hot flashes were prompted to skip the more detailed hot flash questions and were categorized as “never experiencing hot flashes”. Those who responded “yes” to ever experiencing hot flashes answered whether they experienced hot flashes within the past 30 days (answer: no, yes). Additionally, women who had ever experienced hot flashes were asked (in general) to describe their hot flashes as: mild (sensation of heat without sweating), moderate (sensation of heat with sweating), or severe (sensation of heat with sweating that disrupts usual activity). We categorized severity of hot flashes as either mild or moderate/severe. Similarly, women who had ever experienced hot flashes were asked (in general) to describe their hot flashes as occurring: every hour, every 2–5 h, every 6–11 h, every 12–23 h, 5–6 days per week, 1–2 days per week, 2–3 days per month, 1 day per month, less than 1 day per month, or never. We categorized frequency of hot flashes as either monthly or daily/weekly.

2.4. Assessment of urinary phthalate metabolites

Urinary phthalate metabolite concentrations are the preferred biomarkers of phthalate exposure (31). Humans are exposed to phthalate diesters (i.e. parent compounds), which are rapidly metabolized to monoester metabolites in the body (32). Therefore, epidemiological studies measuring human urinary phthalate metabolites often measure one or more metabolites for each parent phthalate and report as sums of metabolite concentrations based on parent compound, exposure source, or biological activity (33–37).

Participants provided spot urine specimens at the initial baseline clinic visit and at visits during the next three consecutive weeks, which were used for urinary phthalate metabolite assessment. Each woman provided samples at 1–4 visits in the 4 week timeframe, which were pooled due to the short half-lives of phthalates in the body and the high daily and weekly intra-variability of measured concentrations (38). Pooled samples were analyzed for the following 9 phthalate metabolites: mono-2-ethylhexyl phthalate (MEHP), mono-(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), mono-(2-ethyl-5 -carboxypentyl) phthalate (MECPP), mono-(2-ethyl-5-oxohexyl)phthalate (MEOHP), mono-(3-carboxypropyl) phthalate (MCPP), monobenzyl phthalate (MBzP), monoethyl phthalate (MEP), monobutyl phthalate (MBP), and monoisobutyl phthalate (MiBP). Analyses were performed using isotope dilution high-performance liquid chromatography negative-ion electrospray ionization-tandem mass spectrometry (HPLC–MS/MS) at the Metabolomics Lab of the Roy J. Carver Biotechnology Center, University of Illinois at Urbana-Champaign, using methods adapted from the Center for Disease Control and Prevention (39) and are described in the supporting information (40).

2.5. Statistical analysis

Out of 780 women enrolled during year 1 of the MWHS, 10 were missing information about hot flashes, an additional 10 were missing information about urinary phthalate metabolite concentrations and/or specific gravity, and an additional 32 women were missing information about covariates (which are described below). Therefore, the current study included a total of 728 women with information about baseline covariates, urinary phthalate metabolite concentrations, and hot flashes. Covariates for associations of midlife urinary phthalate metabolite concentrations with risk of hot flashes were chosen a priori and using previous literature that informed a directed acyclic graph (4,23). To reduce potential for multicollinearity issues, we assessed correlations among all selected covariates (with none being strongly correlated). Therefore, final statistical models evaluating overall associations of urinary phthalate metabolite concentrations with four hot flash outcomes (objective 1) were adjusted for age, race, education, current drinking status, smoking status, medication use, menopause status, BMI, and CESD score. Age and CESD score were included as continuous variables, whereas the other variables were categorized with reference group set as shown in Table 1. Our second objective was to assess differences in associations of urinary phthalate metabolite concentrations with 4 hot flash outcomes by menopausal status, BMI, race/ethnicity, and CESD score (objective 2)—factors that may influence the experience of hot flashes in midlife women (4). In addition to including the previously listed covariates in all stratified models, we a priori stratified our analyses as follows: pre- versus perimenopausal women, under-/normal weight (BMI<25kg/m²) versus overweight/obese (BMI≥25kg/m²) women, non-Hispanic white versus Black/other women, and women with fewer (CESD<16) versus more (CESD≥16) depressive symptoms (26).

Table 1.

Demographic and lifestyle characteristics of 45-54 year-old women from the Midlife Women’s Health Study (n=728).

Demographic or Lifestyle Characteristic	n (%)2
Age (years)1
45 to 49	477 (65.5)
50 to 54	251 (34.5)
Race1
Non-Hispanic White (ref)	484 (66.5)
Black	215 (29.5)
Other	29 (4.0)
Employment status
Unemployed	146 (20.1)
Employed	582 (79.9)
Education1
Some college or less	255 (35.0)
College graduate or higher (ref)	473 (65.0)
Annual family income ($)
<20,000	45 (6.2)
20,000 to 39,999	113 (15.5)
40,000 to 99,999	241 (33.1)
≥100,000	308 (42.3)
Marital status
Single	133 (18.3)
Married/Living with Partner	476 (65.4)
Widowed/divorced/separated	118 (16.2)
Menopausal status1
Premenopausal (ref)	468 (64.3)
Perimenopausal	260 (35.7)
Alcohol consumption status (>1 drink/month on average)1
No	252 (34.6)
Yes (ref)	476 (65.4)
Smoking status1
Current	69 (9.5)
Former	256 (35.2)
Never (ref)	403 (55.4)
Leisure physical activity compared to others
Much more/more	257 (35.3)
As much	230 (31.6)
Less/much less	235 (32.3)
Body mass index (kg/m2)1
<25	290 (39.8)
≥25 (ref)	438 (60.2)
Current medication use1
No	311 (42.7)
Yes (ref)	417 (57.3)
CES depression score1
Fewer depressive symptoms (<16)	581 (79.8)
More depressive symptoms (≥16)	147 (20.2)

¹Variables included in logistic regression models.

²Percentages may not add up to 100% due to missing values.

Nine urinary phthalate metabolites were assessed from pooled urine samples (Table 2). Urinary phthalate metabolite concentrations below the level of detection (LOD) were converted to the LOD/√(2). To account for urine dilution, we used the following formula to adjust all urinary phthalate metabolite concentrations: P_c = P[(1.018 − 1)/(SG_i − 1)], where P_c is the specific gravity adjusted phthalate metabolite concentration, P is the measured phthalate metabolite concentration (ng/mL), 1.018 is the median specific gravity of MWHS population included in this analysis, and SG_i is the specific gravity of each woman’s pooled urine sample (41). Specific gravity-adjusted urinary phthalate metabolite concentrations were used to approximate women’s midlife exposure to phthalate parent compounds. Exposure to DEHP was approximated as the molar converted sum of 4 urinary metabolites using the following equation: ΣDEHP = (MEHP/278) + (MEHHP/294) + (MEOHP/292) + (MECPP/308). Exposure to DOP, BzBP, DEP, DBP, and DiBP was approximated directly using the non-molar converted concentrations of their major urinary metabolites MCPP, MBzP, MEP, MBP, and MiBP, respectively. Additional phthalate sums were created based on primary sources of phthalate exposure, and sum of plasticizer (ΣPlastic) and personal care product (ΣPCP) phthalate metabolites were estimated as follows: ΣPlastic = (MEHHP/294) + (MEHP/278) + (MEOHP/292) + (MECPP/308) + (MCPP/252) + (MBzP/256) and ΣPCP = (MEP/194) + (MBP/222) + (MiBP/222). Previous experimental and some epidemiological studies suggest that certain phthalate metabolites have anti-androgenic activity in the body (18,42,43). Therefore, the sum of anti-androgenic phthalate metabolites (ΣAA) was calculated as (MEHP/278) + (MEHHP/294) + (MEOHP/292) + (MECPP/308) + (MBzP/256) + (MBP/222) + (MiBP/222). Lastly, all 9 urinary phthalate metabolites were molar converted and summed to approximate total midlife phthalate exposure (ΣPhthalates).

Table 2.

Concentrations of individual urinary phthalate metabolites and molar sums from 45-54-year-old women from the Midlife Women’s Health Study and the National Health and Nutrition Examination Survey.

Name	Abbreviation	MWHS (2006–2015) n=728	NHANES (2005–2016) n=7571
	Phthalate metabolite	Median (25th, 75th percentiles) in ng/mL
Monoethyl phthalate	MEP	97.3 (48.2, 192.0)2	58.9 (20.0, 179.8)
Mono-n-butyl phthalate	MBP	19.8 (13.0, 32.8)	11.6 (5.4, 25.3)
Mono-isobutyl phthalate	MiBP	16.4 (10.0, 26.1)	5.8 (2.6, 13.2)
Mono(3-carboxypropyl) phthalate	MCPP	2.5 (1.3, 5.4)	1.5 (0.6, 3.4)
Monobenzyl phthalate	MBzP	9.4 (5.4, 16.1)	4.2 (1.8, 10.4)
Mono(2-ethylhexyl) phthalate	MEHP	4.5 (2.7, 9.3)	1.2 (0.6, 3.1)
Mono(2-ethyl-5-hydroxyhexyl) phthalate	MEHHP	33.5 (20.5, 58.7)	9.1 (3.5, 22.6)
Mono(2-ethyl-5-oxohexyl) phthalate	MEOHP	12.0 (7.3, 22.4)	5.6 (2.1, 13.3)
Mono(2-ethyl-5-carboxypentyl) phthalate	MECPP	25.9 (15.9, 48.0)	13.4 (5.6, 31.7)
	Phthalate molar-converted sum	Median (25th, 75th percentiles) in nmol/mL
Di(2-ethylhexyl) phthalate	DEHP	0.3 (0.2, 0.5)	0.1 (0.04, 0.2)
Sum of all phthalate metabolites	ΣPhthalates	1.2 (0.7, 2.1)	0.7 (0.3, 1.8)
Sum of all personal care product phthalate metabolites	ΣPCP	0.7 (0.4, 1.3)	0.4 (0.2, 1.2)
Sum of all plastic phthalate metabolites	ΣPlastic	0.3 (0.2, 0.6)	0.1 (0.1, 0.3)
Sum of anti-androgenic phthalate metabolites	ΣAA	0.5 (0.3, 0.8)	0.2 (0.1, 0.5)

¹Weighted phthalate metabolite concentrations for 45-54-year-old US women from combined NHANES survey years 2005–06, 2007–08, 2009–10, 2011–12, 2013–14, and 2015–16.

²Two samples (0.3%) were <LOD for MEP.

We used logistic regression models to evaluate overall and stratified associations of midlife urinary phthalate metabolite concentrations with the following 4 hot flash outcomes: 1) ever experiencing hot flashes, 2) experiencing hot flashes in the past 30 days, 3) experiencing daily/weekly or monthly hot flashes, and 4) experiencing moderate/severe or mild hot flashes (Table 3). All statistical analyses were conducted in SAS 9.4 (version 14.3, SAS Institute) using PROC LOGISTIC. Specifically, binary logistic regression models assessed overall and stratified associations of continuous midlife urinary phthalate levels with the odds of ever experiencing and experiencing hot flashes in the past 30 days compared to never experiencing hot flashes. Women who did experience hot flashes at some point, but not in the past 30 days (n=87), were excluded from binary logistic regression models comparing women who experienced hot flashes in the past 30 days to those who never experienced hot flashes. Multinomial logistic regression models assessed overall and stratified associations of continuous midlife urinary phthalate levels with the odds of experiencing daily/weekly or monthly hot flashes and of experiencing moderate/severe or mild hot flashes compared to never experiencing hot flashes. Because all phthalate individual metabolites and molar sums were right-skewed, phthalates were natural log-transformed in all logistic regression models. All odds ratios (ORs) and 95% confidence intervals (CIs) were back transformed using the equation [e^{ln(OR)*ln(2.00)}] and data in Table 4, Figures 1–4, and Supplementary Tables 1–4 are presented as the OR of experiencing these hot flash outcomes with every two-fold increase in phthalate metabolite or molar sum concentration with a priori alpha level of p < 0.05 (40). For logistic regression models evaluating stratified associations of phthalates with hot flashes, we provided formal test for effect modification in Supplementary Tables 1–4 (40), but reported on all relevant associations regardless of the interaction p-value.

Table 3.

Prevalence of hot flashes self-reported by women from the Midlife Women’s Health Study (n=728).

Hot Flashes	n (%)
History of hot flashes
No	399 (54.8)
Yes	329 (45.2)
Hot flashes during past 30 days
Never had hot flashes	399 (54.8)
Had hot flashes and experienced in past 30 days	236 (32.4)
Had hot flashes but did not experience in past 30 days	87 (12.0)
Missing	6 (0.8)
Frequency of hot flashes
Never had hot flashes	399 (54.8)
Monthly hot flashes	139 (19.1)
Daily/weekly hot flashes	159 (21.8)
Missing	31 (4.3)
Severity of hot flashes
Never had hot flashes	399 (54.8)
Mild hot flashes	108 (14.8)
Moderate/severe hot flashes	213 (29.3)
Missing	8 (1.1)

Table 4.

Overall associations of urinary phthalate concentrations with hot flashes.

	Outcomes
	Ever experiencing hot flashes1	Experiencing hot flashes in the past 30 days1	Experiencing daily/weekly or monthly hot flashes2		Experiencing moderate/severe or mild hot flashes2
n	728	635	697		720
Reference	Never experiencing hot flashes	Never experiencing hot flashes	Never experiencing hot flashes		Never experiencing hot flashes
Exposure3			Daily/weekly	Monthly	Moderate/severe	Mild
	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
MEP	1.09 (0.98, 1.22)	1.05 (0.93, 1.18)	1.1 (0.96, 1.27)	1.1 (0.96, 1.25)	1.06 (0.94, 1.2)	1.12 (0.97, 1.3)
MBP	1.01 (0.85, 1.21)	1.1 (0.89, 1.35)	1.17 (0.93, 1.47)	0.92 (0.74, 1.16)	1 (0.81, 1.23)	0.98 (0.76, 1.25)
MiBP	0.98 (0.82, 1.16)	1.03 (0.85, 1.26)	1.03 (0.82, 1.3)	0.98 (0.79, 1.21)	1 (0.82, 1.22)	0.91 (0.72, 1.16)
mCPP	0.92 (0.83, 1.03)	0.94 (0.82, 1.06)	1.02 (0.88, 1.18)	0.87 (0.76, 1)	0.92 (0.81, 1.05)	0.92 (0.79, 1.07)
MBzP	0.95 (0.82, 1.1)	0.93 (0.79, 1.11)	1.08 (0.89, 1.3)	0.94 (0.79, 1.12)	0.95 (0.81, 1.13)	0.94 (0.77, 1.15)
MEHP	1.05 (0.92, 1.19)	1.07 (0.92, 1.24)	1.18 (0.99, 1.4)	0.99 (0.85, 1.17)	1.05 (0.9, 1.22)	1.05 (0.88, 1.25)
MEHHP	1.14 (0.97, 1.33)	1.19 (1, 1.43)	1.34 (1.09, 1.65)	1.09 (0.9, 1.32)	1.17 (0.98, 1.39)	1.1 (0.88, 1.36)
MEOHP	1.07 (0.93, 1.23)	1.11 (0.95, 1.3)	1.23 (1.03, 1.48)	1.04 (0.87, 1.23)	1.07 (0.91, 1.25)	1.08 (0.9, 1.31)
MECPP	1.12 (0.97, 1.3)	1.17 (0.99, 1.38)	1.28 (1.06, 1.54)	1.09 (0.92, 1.3)	1.13 (0.96, 1.33)	1.12 (0.93, 1.37)
ΣDEHP	1.13 (0.96, 1.31)	1.18 (0.99, 1.42)	1.32 (1.08, 1.61)	1.08 (0.89, 1.3)	1.14 (0.96, 1.37)	1.1 (0.89, 1.36)
ΣPhthalates	1.13 (0.96, 1.32)	1.12 (0.93, 1.35)	1.26 (1.03, 1.54)	1.09 (0.89, 1.33)	1.07 (0.89, 1.29)	1.19 (0.97, 1.47)
ΣPCP	1.08 (0.94, 1.23)	1.04 (0.89, 1.22)	1.12 (0.94, 1.33)	1.07 (0.91, 1.27)	1.03 (0.88, 1.2)	1.12 (0.93, 1.34)
ΣPlastic	1.12 (0.95, 1.32)	1.18 (0.98, 1.43)	1.38 (1.11, 1.71)	1.05 (0.85, 1.29)	1.14 (0.95, 1.37)	1.1 (0.88, 1.38)
ΣAA	1.08 (0.9, 1.31)	1.16 (0.93, 1.44)	1.36 (1.07, 1.73)	1 (0.79, 1.27)	1.09 (0.88, 1.35)	1.07 (0.83, 1.39)

¹Binary logistic regression models evaluated associations of every 2-fold increase in urinary phthalate concentrations with the odds of ever experiencing and experiencing hot flashes in the last 30 days compared to never having hot flashes.

²Multinomial logistic regression models evaluated associations of every 2-fold increase in urinary phthalate concentrations with the odds of experiencing daily/weekly or monthly hot flashes and experiencing moderate/severe hot flashes compared to never experiencing hot flashes. All logistic regression models were adjusted for age, race/ethnicity, education, alcohol consumption, smoking status, medication use, menopausal status, body mass index, and CESD score. CI, confidence interval; OR, odds ratio.

³ΣDEHP = (MEHP/278) + (MEHHP/294) + (MEOHP/292) + (MECPP/308); ΣPhthalates = (MEP/194) + (MBP/222) + (MiBP/222) + (MEHP/278) + (MEHHP/294) + (MEOHP/292) + (MECPP/308) + (MBzP/256) + (MCPP/252); ΣPCP = (MEP/194) + (MBP/222) + (MiBP/222); ΣPlastic = (MEHHP/294) + (MEHP/278) + (MEOHP/292) + (MECPP/308) + (MCPP/252) + (MBzP/256); ΣAA = (MEHP/278) + (MEHHP/294) + (MEOHP/292) + (MECPP/308) + (MBzP/256) + (MBP/222) + (MiBP/222)

Figure 1. Associations of urinary phthalate concentrations with hot flashes stratified by menopause status.

Binary logistic regression models evaluated associations of urinary phthalate concentrations with the odds experiencing hot flashes in the last 30 days compared to never having hot flashes (n=635), while multinomial logistic regression models evaluated associations of urinary phthalate concentrations with the odds of experiencing daily/weekly or monthly hot flashes compared to never experiencing hot flashes (n=697). All models were stratified by menopause status and were adjusted for age, race/ethnicity, education, alcohol consumption, smoking status, medication use, body mass index, and CESD score. Data are presented as odds ratio (circles) and 95% confidence interval (solid lines) for every two-fold increase in urinary phthalate concentrations. Confidence intervals that do not cross the null are significant at ^#P<0.10 and *P<0.05.

Figure 4. Associations of urinary phthalate concentrations with hot flashes stratified by CESD score.

Binary logistic regression models evaluated associations of urinary phthalate concentrations with the odds of experiencing hot flashes in the last 30 days compared to never having hot flashes (n=635), while multinomial logistic regression models evaluated associations of urinary phthalate concentrations with the odds of experiencing daily/weekly or monthly hot flashes compared to never experiencing hot flashes (n=697). All models were stratified by CESD score and were adjusted for age, race/ethnicity, education, alcohol consumption, smoking status, medication use, menopause status, and body mass index. Data are presented as odds ratio (circles) and 95% confidence interval (solid lines) for every two-fold increase in urinary phthalate concentrations. Confidence intervals that do not cross the null are significant at ^#P<0.10 and *P<0.05.

3. Results

3.1. Baseline MWHS population characteristics and hot flash prevalence

Baseline characteristics of 728 women in the MWHS are reported in Table 1. Overall, 64% of women were premenopausal, whereas 36% were perimenopausal. Most women were white (67%), whereas 33% were Black or of another race/ethnicity. Most women were employed (80%), college educated (65%), married or cohabiting (65%), and had an annual family income ≥$40,000 (75%). The prevalence of baseline healthy lifestyle characteristics in this study population were as follows: 35% did not regularly consume alcohol within the past year, 55% never smoked, 67% reported leisure time physical activity as more/much more than others, 43% were not taking any medications, and 80% did not meet criteria for depression based on the CESD. Over one-third of the women were under- or normal weight (40%) and over half were overweight or obese (60%).

The self-reported baseline prevalence of hot flashes in the MWHS is presented in Table 3. Out of 728 total women in the current study, all women provided information about ever experiencing hot flashes, 722 had information about experiencing hot flashes within the past 30 days, 697 had information about hot flash frequency, and 720 had information about hot flash severity. Approximately 55% of women never experienced hot flashes, whereas 45% had experienced hot flashes and 32% had experienced hot flashes in the past 30 days. Overall, 22% had daily/weekly hot flashes and 19% had monthly hot flashes, whereas 29% had moderate/severe hot flashes and 15% had mild hot flashes.

3.2. Baseline urinary phthalate metabolite concentrations

Table 2 presents median (25^th, 75^th percentiles) concentrations of individual urinary phthalate metabolites and phthalate molar sums during the first year of the MWHS. Concentrations of most urinary phthalate metabolites measured in the MWHS were ≥ LOD in 100% of women, except for MEP, for which 99.7% of women had concentrations ≥ LOD. Median urinary phthalate metabolite and molar sum concentrations in the MWHS are compared to those measured from a nationally representative sample of 45–54-year-old U.S. women from the 2005–2016 National Health and Nutrition Examination Survey (NHANES). Median metabolite levels are slightly higher in the MWHS than NHANES with overlapping 25–75^th percentiles. Metabolite levels are similar to recently reported results from the Study of Environment, Lifestyle, and Fibroids (SELF) from non-Hispanic Black women (23–35 years old) (44). SELF also reports slightly higher DEHP metabolites than the corresponding NHANES samples (44).

3.3. Overall associations of phthalates with hot flashes

Midlife urinary phthalate metabolite concentrations were not associated with ever experiencing hot flashes or experiencing moderate/severe or mild hot flashes compared to never experiencing hot flashes (Table 4). However, phthalates were associated with experiencing hot flashes in the past 30 days. Specifically, women had 19% higher odds of the experiencing hot flashes in the past 30 days with every two-fold increase in MEHHP (OR: 1.19; 95%CI: 1.00, 1.43). Additionally, phthalates were associated with experiencing daily/weekly, but not monthly hot flashes. Specifically, women had 23–38% higher odds of experiencing daily/weekly hot flashes with every two-fold increase in the DEHP metabolites MEHHP (OR: 1.34, 95%CI: 1.09, 1.65), MEOHP (OR: 1.23, 95%CI: 1.03, 1.48), MECPP (OR: 1.28, 95%CI: 1.06, 1.54) and the summary measures ΣDEHP (OR: 1.32, 95%CI: 1.08, 1.61), ΣPlastic (OR: 1.38; 95%CI: 1.11, 1.71), ΣPhthalates (OR: 1.26; 95%CI: 1.03, 1.54), and ΣAA (OR: 1.36; 95%CI: 1.07, 1.73).

3.4. Associations of phthalates with hot flashes stratified by menopause status

Associations of urinary phthalate metabolite concentrations with ever experiencing hot flashes, experiencing hot flashes in the past 30 days, or experiencing daily/weekly hot flashes did not significantly differ by menopause status in any of the interaction terms (Figure 1A, B and Table S1) (40).

3.5. Associations of phthalates with hot flashes stratified by midlife BMI

Associations of urinary phthalate metabolite concentrations with ever experiencing hot flashes, experiencing hot flashes in the past 30 days, and experiencing moderate/severe or mild hot flashes were not different by midlife BMI. (Figure 2A, Table S2) (40). However, some associations of phthalates with experiencing daily/weekly (but not monthly) hot flashes were only observed in under-/normal weight women (Figure 2B, Table S2) (40), who had 57–107% higher odds of experiencing daily/weekly hot flashes with every two-fold increase in MBP (OR: 1.57; 95%CI: 1.04, 2.38), MBzP (OR: 1.63; 95%CI: 1.18, 2.25), ΣDEHP (OR: 1.77; 95%CI: 1.24, 2.52), ΣPlastic (OR: 2.07; 95% CI: 1.40, 3.05), and ΣAA (OR: 2.05; 95%CI: 1.31, 3.19).

Figure 2. Associations of urinary phthalate concentrations with hot flashes stratified by midlife BMI.

Binary logistic regression models evaluated associations of urinary phthalate concentrations with the odds of experiencing hot flashes in the last 30 days compared to never having hot flashes (n=635), while multinomial logistic regression models evaluated associations of urinary phthalate concentrations with the odds of experiencing daily/weekly or monthly hot flashes compared to never experiencing hot flashes (n=697). All models were stratified by midlife BMI and were adjusted for age, race/ethnicity, education, alcohol consumption, smoking status, medication use, menopause status, and CESD score. Data are presented as odds ratio (circles) and 95% confidence interval (solid lines) for every two-fold increase in urinary phthalate concentrations. Confidence intervals that do not cross the null are significant at ^#P<0.10 and *P<0.05. BMI, body mass index.

3.6. Associations of phthalates with hot flashes stratified by race/ethnicity

Some associations of urinary phthalate metabolite concentrations with ever experiencing hot flashes, experiencing hot flashes in the past 30 days, experiencing daily/weekly or monthly hot flashes, and experiencing mild (but not moderate/severe) hot flashes also differed by race/ethnicity (Figure 3A, B, Table S3) (40). Non-Hispanic white women had 19–27% higher odds of ever experiencing hot flashes with every two-fold increase in MEP (OR: 1.19; 95%CI: 1.03, 1.38) and ΣPhthalates (OR: 1.27; 95%CI: 1.01, 1.58) (Table S3) (40). Additionally, non-Hispanic white women had 37–45% higher odds of experiencing daily/weekly hot flashes with every two-fold increase in ΣDEHP (OR: 1.37; 95%CI: 1.06, 1.76), ΣPlastic (OR: 1.45; 95%CI: 1.11, 1.90), ΣPhthalates (OR: 1.40; 95%CI: 1.06, 1.85), and ΣAA (OR: 1.40; 95%CI: 1.04, 1.89), as well as higher odds of experiencing monthly hot flashes with every two-fold increase in MEP (OR: 1.23; 95%CI: 1.03, 1.47) (Figure 3B, Table S3) (40). Conversely, Black/other women had 20% (OR: 0.80; 95%CI: 0.66, 0.98) and 28% (OR = 0.72, 95% CI 0.55, 0.95) lower odds of ever experiencing or experiencing mild hot flashes, respectively, with every two-fold increase in MCPP, but had 51% higher odds of experiencing hot flashes in the past 30 days with every two-fold increase in MEHHP (OR: 1.51; 95%CI: 1.08, 2.10) (Table S3) (40).

Figure 3. Associations of urinary phthalate concentrations with hot flashes stratified by race/ethnicity.

Binary logistic regression models evaluated associations of urinary phthalate concentrations with the odds of experiencing hot flashes in the last 30 days compared to never having hot flashes (n=635), while multinomial logistic regression models evaluated associations of urinary phthalate concentrations with the odds of experiencing daily/weekly or monthly hot flashes and compared to never experiencing hot flashes (n=697). All models were stratified by race/ethnicity and were adjusted for age, education, alcohol consumption, smoking status, medication use, menopause status, body mass index, and CESD score. Data are presented as odds ratio (circles) and 95% confidence interval (solid lines) for every two-fold increase in urinary phthalate concentrations. Confidence intervals that do not cross the null are significant at ^#P<0.10 and *P<0.05.

3.7. Associations of phthalates with hot flashes stratified by CESD score

Some associations of urinary phthalate metabolite concentrations with ever experiencing hot flashes, experiencing hot flashes in the past 30 days, experiencing daily/weekly or monthly hot flashes, and experiencing mild (but not moderate/severe) hot flashes differed by CESD score (Figure 4A, B, Table S4) (40). For example, women with fewer depressive symptoms (CESD<16) had 24% higher odds of experiencing hot flashes in the past 30 days with every two-fold increase in ΣDEHP (OR: 1.24; 95%CI: 1.02, 1.49) (Table S4) (40), and had 38–45% higher odds of experiencing daily/weekly hot flashes with every two-fold increase in ΣDEHP (OR: 1.38; 95%CI: 1.09, 1.73), ΣPlastic (OR: 1.45; 95%CI: 1.14, 1.86), and ΣAA (OR: 1.42; 95%CI 1.07, 1.88) (Figure 4B, Table S4) (40). Conversely, women with more depressive symptoms (CESD≥16) had 42–71% higher odds of experiencing hot flashes in the past 30 days with every two-fold increase in MEP (OR: 1.42; 95%CI: 1.02, 1.97), ΣPCP (OR: 1.50; 95%CI: 1.03, 2.19), and ΣPhthalates (OR: 1.71; 95%CI: 1.04, 2.79) (Figure 4A, Table S4) (40). Additionally, these women had 47–93% and 60–78% higher odds of experiencing daily/weekly or monthly hot flashes, respectively, for every two-fold increase in MEP (daily/weekly OR: 1.47; 95%CI: 1.03, 2.11; monthly OR: 1.60; 95%CI: 1.10, 2.33), ΣPCP (daily/weekly OR: 1.59; 95%CI: 1.06, 2.41; monthly OR: 1.60; 95%CI: 1.05, 2.46), and ΣPhthalates (daily/weekly OR: 1.93; 95%CI: 1.13, 3.28; monthly OR: 1.78; 95%CI: 1.04, 3.06), but also had 43% lower odds of experiencing monthly hot flashes with every two-fold increase in MBzP (OR: 0.57; 95%CI: 0.35, 0.94) (Figure 4B, Table S4) (40).

4. Discussion

In a cross-sectional analysis of year 1 data from the MWHS, a prospective cohort of pre- and perimenopausal women from Baltimore and its surrounding counties, urinary phthalate metabolite concentrations were associated with experiencing recent and experiencing more frequent hot flashes, but not associated with ever experiencing hot flashes or hot flash severity. Generally, we found that phthalate metabolites of parent compounds found in plastics were associated with higher risk of experiencing hot flashes in the past 30 days and experiencing daily/weekly hot flashes. Some associations of phthalates with certain hot flash outcomes were different by menopause status, midlife BMI, race/ethnicity, and depressive symptoms. Most notably, we found that associations of personal care product phthalates with most hot flash outcomes were strongest in women with more depressive symptoms. Interestingly, MCPP (and to a lesser extent MBzP) was associated with lower risk of hot flashes, especially in perimenopausal and Black/other women.

Overall, these results are consistent with our previous findings that phthalate metabolites are associated with increased frequency of hot flashes in midlife women (5). In our pilot study of a representative sample of 195 white women from the MWHS cohort, ΣPCP was associated with higher odds of ever experiencing hot flashes, experiencing hot flashes in the past 30 days, and experiencing daily/weekly hot flashes (5). In the current study, we found that associations of ΣPCP with risk of hot flashes only emerged when associations were stratified by CESD score. In the pilot study, ΣAA and ΣDEHP were not associated with hot flash frequency. However, in the current study, we found that ΣDEHP and ΣPlastic were consistently associated with experiencing hot flashes in the past 30 days and hot flash frequency, whereas ΣAA was only associated with hot flash frequency. The pilot study did not evaluate associations of individual phthalate metabolites, ΣPhthalates, or ΣPlastic with hot flash experience.

To our knowledge, no other observational studies have reported on the associations between phthalates and hot flashes in midlife women. However, phthalates are associated with disorders of the female reproductive system, including premature reproductive senescence and altered hormone levels, and these disorders may contribute to the risk of hot flashes (4,23,28,45,46). Studies in mice have shown premature reproductive senescence following adult exposure to DEHP and diisononyl phthalate (DiNP) (47,48). Observational studies have identified associations between urinary phthalate metabolite levels and earlier onset of menopause (49) and premature ovarian failure (50,51). Women who undergo earlier menopause report more frequent and severe hot flashes than premenopausal women (46,52). In addition, higher urinary DEHP metabolite levels were associated with decreased levels of the hormones inhibin B and anti-Müllerian hormone, markers of ovarian reserve in women (53).

Hot flashes and declining estradiol levels occur simultaneously during the menopause transition and other physiological transitions such as the postpartum period (4). Clinical trials have shown that estrogen therapy can alleviate hot flashes, suggesting a causal association, although the mechanism remains unknown (54–56). Numerous observational studies have also identified associations of altered estradiol and progesterone levels with risk of hot flashes (4,23,28,45,57). Using the MWHS cohort, the close relationship between hot flashes and estradiol was recently modeled using a Bayesian network (54). Animal studies have demonstrated altered estradiol and progesterone levels following phthalate exposure (11,18,19,21,58), and future studies from the MWHS will investigate the association between urinary phthalate metabolites and hormone levels in midlife women. Other potential mechanisms through which phthalates could be linked to hot flashes include direct disruption of hypothalamic or thyroid function (4,59,60). However, as the etiology of hot flashes is not well understood, it is difficult to speculate on causal links between phthalates and hot flashes.

When we evaluated stratified associations of phthalates with these hot flashes outcomes by menopause status, BMI, and race/ethnicity, the strongest associations were observed in perimenopausal women, non-Hispanic white women, and under-/normal weight women. These sub-group associations were most consistent between DEHP metabolites and experiencing daily/weekly hot flashes. Black women generally have higher phthalate exposure than white women (61,62), which may be due to the phthalate content in personal care products used by Black women (63). In the MWHS, Black women had higher urinary PCP phthalate concentrations than non-Hispanic white women (data not shown). However, we did not identify any consistent associations of ΣPCP with risk of hot flashes in Black/other women despite this being a high-risk exposure group, which may indicate that phthalates in personal care products do not contribute to hot flash experience in non-white women. Previously, we have identified high BMI as a risk factor for experiencing perimenopausal hot flashes (45). However, we found stronger associations between phthalates and risk of hot flashes in women with lower BMI. Interestingly, MBzP was associated with increased risk of hot flashes in normal weight women and trending towards decreased risk of ever experiencing hot flashes in overweight women. These differences suggest that women with higher body weight may be less susceptible to endocrine disruption by phthalates. One possible explanation is that estrogen levels are already decreased in perimenopausal women with high BMI (45).

Previous studies have identified bidirectional associations of depression with risk of hot flashes, and suggest that they may be linked through sleep disruption (64). In addition, phthalates may be associated with depression in adults. MECPP, MBP, MiBP, and MBzP were associated with depression in adults from NHANES (65). In elderly populations (ages 59–93), DEHP metabolites, MCPP, and MBP have also been associated with depression (66,67). When associations of phthalates with risk of hot flashes were stratified by CESD scores in the MWHS, we observed different associations of phthalate metabolites, with risk of hot flashes in women experiencing fewer and more depressive symptoms. DEHP metabolites and phthalates found in plastics were consistently associated with hot flashes in women experiencing fewer depressive symptoms, whereas MEP and phthalates associated with personal care products were associated with hot flashes in women with more depressive symptoms. The different observations in women in more vs. fewer depressive symptoms suggests that depression symptoms may related to hormonal changes or that the physiology of depression may play a role, possibly hormonal, in phthalate mechanism of action. Additional studies are needed to investigate the role of depression in environmental chemical action. In addition, the interaction of each depression group with a different phthalate category may be evidence that the high molecular weight phthalates found in plastics act through different mechanisms than the low molecular weight phthalates found in personal care products.

Across multiple hot flashes measures, MCPP and to a lesser extent MBzP were associated with decreased risk of experiencing hot flashes. MCPP is a downstream oxidized metabolite and may be produced from multiple phthalates including MBP, BzBP, and phthalates with long n-hydrocarbon side chains (68). One hypothesis for the negative association between MCPP and hot flashes is that the presence of highly oxidized metabolites is a marker of the overall efficacy of metabolism and detoxification. Greater capacity to metabolize and excrete phthalates (and other environmental chemicals) could reduce their effects on multiple sensitive endpoints. MBzP, containing a benzyl group on its side chain, has a unique structure for a phthalate that is more similar to steroid hormones. As a result, MBzP may act through different mechanisms than other phthalate metabolites.

This study has several strengths. Of note, four urine samples taken over consecutive weeks at similar times of day were pooled for phthalate measurement. Within-subject pooling has been shown to decrease exposure misclassification for phthalates and improves the credibility of estimated exposure based on urine concentrations (69). Other strengths of this study include the large size of the population and the detailed information collected from each participant on hot flash experience using validated questionnaires that are accepted by the National Institute of Health to assess hot flashes (70).

This study also has several limitations. The MWHS cohort is composed of primarily white (67%) and Black (30%) midlife women. Few women of other races/ethnicities were enrolled in the study; therefore, the results of this study are most applicable to non-Hispanic white and Black women. Due to the cross-sectional analyses performed here, some outcomes may have occurred before phthalate exposure, obscuring temporality. The measure of “ever experiencing hot flashes” is the most imprecise and mostly likely to differ in temporal ordering with respect to phthalate exposure. Therefore, prospective studies are needed to confirm whether phthalates influence risk of experiencing hot flashes. We were unable to adjust for co-pollutants or diet, although we adjusted for a number of relevant and important confounder, including age, race, education, current drinking status, smoking status, medication use, menopause status, BMI, and CESD score. However, other environmental chemical exposures may be correlated with phthalates and hot flashes. Diet quality may be important given that we observed strong associations between plasticizing phthalates and hot flashes risk and exposure to plasticizing phthalates occurs primarily through diet. In addition, we did not evaluate non-linear associations.

5. Conclusions

In midlife women from the MWHS, some urinary phthalate metabolites were associated with higher risk of recently experiencing and experiencing frequent hot flashes, but not of ever experiencing or experiencing severe hot flashes. We observed that urinary phthalate metabolites of plasticizer parent compounds are associated with higher odds of experiencing hot flashes in the past 30 days and experiencing daily or weekly hot flashes. Additionally, we found that some associations of urinary phthalate metabolites with hot flashes were different by menopause status, midlife BMI, race/ethnicity, and CESD score. Although this is one of the first studies to assess the relationship between phthalate exposure and risk of hot flashes, these results are consistent with previous studies showing that phthalates can interfere with normal female reproductive function (37). Our results suggest specific relationships between phthalates from common exposure sources and the evaluated hot flash outcomes. Future studies should investigate the mechanisms through which phthalates may be acting to facilitate the development of interventions to alleviate hot flashes in midlife women.

Highlights.

• The Midlife Women’s Health Study is a cohort of pre- and perimenopausal women

• Phthalate exposure is associated with hot flash experience in midlife women

• Phthalates found in plastics are associated with recent hot flash experience

• Associations differed by race/ethnicity, menopause status, body mass, and depressive symptoms

Abbreviations:

CI

confidence interval

OR

odds ratio

CESD

Center for Epidemiological Studies Depression Score

PCP

personal care products

BMI

body mass index

MWHS

Midlife Women’s Health Study

NHANES

National Health and Nutrition Examination Survey

DEP

diethyl phthalate

DBP

dibutyl phthalate

DiBP

diisobutyl phthalate

BzBP

benzylbutyl phthalate, DEHP, di(2-ethylhexyl)phthalate

DOP

dioctyl phthalate

MECPP

mono-(2-ethyl-5-carboxypentyl) phthalate

MBP

monobutyl phthalate

MEP

monoethyl phthalate

MEOHP

mono-(2-ethyl-5-oxohexyl)phthalate

MBzP

monobenzyl phthalate

MiBP

monoisobutyl phthalate

MCPP

mono-(3-carboxypropyl) phthalate

MEHHP

mono-(2-ethyl-5-hydroxyhexyl) phthalate

MEHP

mono-2-ethylhexyl phthalate

LOQ

limit of quantitation

AA

anti-androgenic

DiNP

diisononyl phthalate