ORIGINAL ARTICLE
Reproductive epidemiology
Association between phthalate
exposure and glutathione S-transferase
M1 polymorphism in adenomyosis,
leiomyoma and endometriosis
Po-Chin Huang
1, Eing-Mei Tsai
2,5, Wan-Fen Li
1, Pao-Chi Liao
3,
Meng-Chu Chung
1, Ya-Hui Wang
1, and Shu-Li Wang
1,4,5,*1Division of Environmental Health and Occupational Medicine, National Health Research Institutes, 35 Keyan Road, Zhunan, Miaoli County
350, Taiwan2Department of Obstetrics and Gynecology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan3Department of
Environmental and Occupational Health, Medical College, National Cheng Kung University, Tainan, Taiwan4Institute of Environmental
Medicine, College of Public Health, China Medical University and Hospital, Taichung, Taiwan5Center of Excellence for Environmental Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan
*Correspondence address. Tel: þ886-37-246166 ext. 36509; Fax: þ886-37-587406; E-mail: [email protected] Submitted on September 4, 2009; resubmitted on December 29, 2009; accepted on January 7, 2010
background:
Phthalates are known to have estrogenic effects in cell models and experimental animals. However, the evidenceregard-ing the effects of phthalates on human reproduction is still limited. We conducted a case – control study to determine whether estrogen-dependent diseases are associated with phthalate exposure and how the glutathione S-transferase M1 (GSTM1; a major detoxification enzyme) genotype modulates the risk.
methods:
We recruited subjects who underwent laparotomy and had pathologic confirmation of endometriosis (EN) (n ¼ 28),adeno-myosis (AD) (n ¼ 16) and leiomyoma (LEI) (n ¼ 36) from the Department of Obstetrics and Gynecology at a medical center in Taiwan between 2005 and 2007. Controls (n ¼ 29) were patients without any of the three aforementioned gynecologic conditions. Urine samples were collected before surgery and analyzed for seven phthalate metabolites using liquid chromatography– tandem mass spec-trometry. Peripheral lymphocytes were used for GSTM1 genotype determination.
results:
Patients with LEIs had significantly higher levels of total urinary mono-ethylhexyl phthalate (SMEHP; 52.1 versus 18.9 mg/gcreatinine, P , 0.05) than the controls, whereas patients with EN had an increased level of urinary mono-n-butyl phthalate (94.1 versus 58.0 mg/g creatinine, P , 0.05). Subjects with GSTM1 null genotype had significantly increased odds for AD relative to those with GSTM1 wild genotype [odds ratio (OR) ¼ 5.30; 95% CI, 1.22 – 23.1], even after adjustment for age and phthalate exposure. Subjects who carried the GSTM1 null genotype and had a high urinary level of SMEHP showed a significantly increased risk for AD (OR ¼ 10.4; 95% CI, 1.26 – 85.0) and LEIs (OR ¼ 5.93; 95% CI, 1.10 – 31.9) after adjustment for age, compared with those with GSTM1 wild-type and low urinary level of SMEHP.
conclusions:
These results suggest that both GSTM1 null and phthalate exposure are associated with AD and LEI. Larger studies arewarranted to investigate potential interaction between GSTM1 null and phthalate exposure in the etiology of estrogen-dependent gyneco-logic conditions.
Key words: adenomyosis / leiomyoma / endometriosis / phthalate monoester / glutathione S-transferase M1
Introduction
The etiologies of some estrogen-dependent gynecologic conditions, such as endometriosis (EN), adenomyosis (AD) and leiomyomas (LEIs), are still unclear. EN is defined as the presence of endometrial tissue outside the uterine cavity, whereas AD refers to the condition when endometrial tissues invade the myometrium with smooth
muscle hyperplasia (Devlieger et al., 2003; Bulun, 2009). LEIs, also known as fibroids, are defined as benign tumors of smooth muscle (Al-Hendy and Salama, 2006) in the uterus. EN, AD and LEIs are common gynecologic disorders presented by prolonged or heavy menstrual bleeding, pelvic pain and infertility. The prevalence of EN has been reported to be 2 – 22% in women of childbearing age, whereas AD and LEIs have a prevalence of 20 – 35% in the infertility
&The Author 2010. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: [email protected]
Human Reproduction, Vol.00, No.0 pp. 1 – 9, 2010 doi:10.1093/humrep/deq015
clinic and 20 – 25% in premenopausal women, respectively, and varies by ethnic groups (Devlieger et al., 2003; Guo, 2005; Al-Hendy and Salama, 2006). Susceptibility to these estrogen-dependent gynecologic conditions depends on complex interactions between immunologic, hormonal, environmental and genetic factors (Giudice and Kao, 2004). Previous studies have revealed that some extensively used plasticizers, known as phthalates, are possibly associated with EN (Cobellis et al., 2003; Reddy et al., 2006a, b). Phthalates are added to plastics to make them soft and flexible, to cosmetics as a vehicle for fragrances and to many other daily products, such as building materials, children’s toys and medical devices (Schettler, 2006). Recent reports have shown that phthalates are widely added in considerable amounts (up to 5%) to cosmetics and personal care products, and that they raise the levels of exposure to urinary phthalate monoesters rapidly when these products are used daily (Houlihan et al., 2002; Duty et al., 2005), particularly in women.
Phthalates are estrogenic and anti-androgenic endocrine disruptors that may prolong menstrual cycles and increase the proportion of pre-mature menopause in animal models (Moore, 2000; Ma et al., 2006). Toxicological evidence has shown that some phthalates, such as butyl benzyl phthalate (BBzP), di-n-butyl phthalate (DnBP) and di-(2-ethylhexyl) phthalate (DEHP), may alter or mimic estradiol (E2)
in vivo and in vitro (Harris et al., 1997; Okubo et al., 2003; Jin et al., 2008). However, whether phthalate exposure results in adverse effects on the human reproductive system is largely unknown.
Several studies have examined a relationship between estrogen-dependent gynecologic conditions and genetic polymorphisms of human detoxification enzymes, including N-acetyltransferase 2, gluta-thione S-transferases (GST) M1 and T1 (Baranova et al., 1997, 1999) and cytochrome P450 (CYP) 1A1 (Juo et al., 2006). Previous studies have suggested that the GSTM1 null mutation is associated with EN in French (Baranova et al., 1997, 1999), Greek (Arvanitis et al., 2003) and Indian (Babu et al., 2005) populations. However, other studies have concluded that neither GSTM1 mutations nor CYP 1A1 is linked to EN in Korean (Hur et al., 2005), Taiwanese (Juo et al., 2006) and English (Hadfield et al., 2001) populations. The inconsistent results could be due to small sample size or different genetic backgrounds between ethnic groups, and therefore whether genetic variability acts as a critical factor in EN is still under debate (Guo, 2005). Furthermore, for complex gynecologic conditions, such as EN and AD, it is likely that gene – environmental interactions are more important and relevant than genetics alone.
The present study was aimed to explore the possible interaction between phthalates and GSTM1 on estrogen-dependent gynecologic conditions, including AD, EN and LEIs. We first examined the relation-ship between gynecologic conditions and phthalate exposure using urinary phthalate monoesters as the exposure biomarkers, and deter-mined whether GSTM1 polymorphisms can modify individual risk for such conditions. The results of this study will also help us further understand the possible adverse effects of phthalates and provide hints on the role of GST enzymes in phthalate detoxification.
Materials and Methods
Subject recruitment
We recruited patients who had undergone laparotomy in the Department of Obstetrics and Gynecology of Kaohsiung Medical University Hospital
(KMUH) in Taiwan between 2005 and 2007, and had pathologic confir-mation of EN, AD and LEIs. The diagnoses of EN and AD were based on the pathologic results of the presence of endometrial tissue outside the uterine cavity and within the myometrium with smooth muscle hyper-plasia, respectively. The diagnosis of LEIs was based on the pathologic finding of a benign tumor within the uterine smooth muscle. In this study, 28 cases of EN, 16 cases of AD and 36 cases of LEIs were included. Women who underwent laparotomy for other clinical reasons and did not have EN, AD or LEIs, as confirmed by pathology, served as the control group. The laparotomy was performed to exclude subjects with pelvic masses, such as uterine and ovarian tumors. Women who had been pre-viously diagnosed with these estrogen-dependent gynecologic conditions were also excluded. All cases and controls were of Chinese descent. This protocol was approved by the Institutional Review Board of KMUH and informed consent was obtained from the participants before the study.
Demographic characteristics
The demographic data of each subject were obtained from an interviewed questionnaire during the recruitment process. The recorded character-istics included age, body mass index (BMI), education, age of menarche, duration of menstrual cycle, history of abnormal uterine bleeding, parity, hormone therapy, intrauterine device (IUD) use, oral contraceptive use, caffeinated drink consumption, cigarette smoking, alcohol consumption, a family history of gynecologic diseases and dietary habits.
Blood and urine collection
Blood samples were collected in tubes with EDTA and stored at 48C for 20 – 30 min. Blood samples were centrifuged at 1620 g for 15 min and lymphocytes obtained from the blood sample (8 ml) were collected and analyzed for the GSTM1 genotype. A 20 – 30 ml urine sample was col-lected in a 250 ml glass vessel and immediately transferred into a 12 ml amber glass bottle for phthalate monoesters and creatinine analysis. To prevent possible contamination of the urine samples, all the glassware had been washed in methanol, acetonitrile and acetone, and then was sealed with aluminum foil before sample collection.
Genotype determination
The GSTM1 genotype was determined by polymerase chain reaction (PCR) based on a method described previously (Shinka et al., 1998). PCR, containing 100 ng of genomic DNA, was incubated at 948C for 5 min and subjected to 35 cycles at 948C for 30 s, 588C for 60 s, 728C for 60 s and a final extension at 728C for 10 min. The presence of the GSTM1 allele was shown as a fragment of 215 bp on 2% agarose gel. The primers used were 50-GAACTCCCTGAAAAGCTAAAGC-30 and
50-GTTGGGCTCAAATATACGGTGG-30.
Urinary phthalate monoesters analysis
Standards of phthalate metabolites, including mono-methyl phthalate (MMP), mono-ethyl phthalate (MEP), mono-n-butyl phthalate (MnBP), mono-benzyl phthalate (MBzP), mono-(2-ethylhexyl) phthalate (MEHP), 5-oxo-hexyl) phthalate (5oxo-MEHP) and mono-(2-ethyl-5-hydroxyhexyl) phthalate (5OH-MEHP), and their corresponding
13C
4-labeled compounds were purchased from Cambridge Isotope
Lab-oratories (Andover, MA, USA). Formic acid (FA), acetic acid and b-glucuronidase (Helix pomatia) were purchased from Sigma-Aldrich (St Louis, MO, USA). Methanol (HPLC grade) was purchased from Merck (Darmstadt, Germany). Deionized water was acquired from a Milli-pore system (Milford, MA, USA). Seven urinary phthalate monoesters were analyzed using on-line solid-phase extraction (SPE) coupled with
liquid chromatography/electrospray ionization tandem mass spectrometry (LC/ESI-MS/MS), which was adapted from a previous study (Lin et al., 2004; Huang et al., 2007). Briefly, 1 ml aliquots of the sample containing 750 ml urine, 50 ml of 2000 ppb 13C4-labeled phthalate monoesters as
internal standards, 200 ml of 100 mM ammonium acetate buffer (pH 6.5) and 10 ml of b-glucuronidase were incubated at 378C for 90 min for deconjugation. The deconjugation reaction was stopped by the addition of 20% acetic acid/can (50 ml). The mixture was passed through a 0.2 mm PVDF membrane filter (MSF-3; Advantec MFS, Inc., Pleasanton, CA, USA) and stored at 48C prior to loading onto the analyti-cal system. Then, the urine mixture was loaded onto the on-line SPE car-tridge (C18 trap carcar-tridge, 2.0 55 mm, 3 mm; Merck) and washed with 2% FA/H2O at a flow rate of 600 ml/min for 10 min before the switching
valves (6-ports; Valco Europe, Schenkon, Switzerland) were triggered to start the LC gradient and initiate chromatography on the Chromolith column Flash RP-18e column (4.6 50 mm; Merck). The gradient program was started with a mobile phase from 0.1% FA/H2O at a flow
rate of 600 ml/min to 100% MeOH in 10 min. The LC eluent was split into a 1:20 ratio before entering the mass spectrometer (API365 triple quadrupole; PE Sciex, Throhill, Ontario, Canada). After gradient analysis was completed, the Chromolith column was washed with MeOH for 5 min and re-equilibrated with 2% FA/H2O for 1 min before the next
injection. The precursor to product ion transitions of seven phthalate monoesters and their corresponding 13C4-labeled compounds were the
same as reported in a previous study (Kato et al., 2005).
The dynamic concentration ranges of the calibration curves for MMP, MEP, MBzP, 5OH-MEHP and 5oxo-MEHP were 0.67 – 1300 ng/ml, whereas the dynamic concentration ranges of the calibration curves for MnBP and MEHP were 0.67 – 670 ng/ml. The intra-day variations of these phthalate monoesters in urine at three different concentrations (25%, 50% and 75%) within the dynamic range of the individual substance were all ,10% with intra-day recoveries at 100 + 20%. The detection limits of MMP, MEP, MnBP, MBzP, MEHP, 5OH-MEHP and 5oxo-MEHP were 3.4, 2.2, 1.6, 0.99, 0.55, 0.23 and 0.26 ng/ml, respectively. The accu-racy of the analytical approach was tested against two reference urine samples with different known phthalate monoester concentrations. The samples were received from the laboratory inter-comparison program (www.g-equas.de) in 2006. The relative errors of the five urinary monoe-sters (MnBP, MBzP, MEHP, 5OH-MEHP and 5oxo-MEHP) were ,16%, whereas MMP and MEP were not included in the reference sample.
Statistical analysis
The Kruskal – Wallis and the Wilcoxon rank sum tests were used to evalu-ate the difference in demographic data among all groups and between groups, respectively. We also used the Wilcoxon rank sum test to assess the difference in phthalate exposure between cases and controls. All phthalate metabolite measurements were log transformed to approxi-mate normal distribution. As 5OH-MEHP and 5oxo-MEHP are both the further metabolites of MEHP, we calculated the sum of MEHP (SMEHP) by combining the levels of 5oxo-MEHP, 5OH-MEHP and MEHP measure-ments to evaluate the total exposure of DEHP.
Deviation from the Hardy – Weinberg equilibrium (HWE) was examined using a x2 test. Genotypic effects were evaluated for three
estrogen-dependent gynecologic conditions. We also categorized the subjects into two groups by the median levels of phthalate metabolites to evaluate the joint effects of GSTM1 polymorphism. Then, we evaluated the odds ratio (OR) of individual case groups carrying the GSTM1 null genotype using logistic regression. We used logistic regression to assess whether there was a significant risk increase between the case group with higher phthalate exposure and the control group. Potential covariance was adjusted in these two simple models. Further, stepwise backward logistic
regression analysis was carried out to adjust for significant covariates, including age and BMI. We used the control group as a reference in the logistic model and a cut-off point of P , 0.2 was applied as the criterion for backward variable selection. In addition, we put only the most sensitive exposure variable (urinary phthalate metabolites) into the final models to avoid confounding co-linearity.
In order to assess the interaction of phthalate exposure and GSTM1 polymorphism, four groups are categorized as follows: those with GSTM1 and below-median SMEHP exposure, those with GSTM1 and above-median SMEHP exposure, those with GSTM1 null type and below-median SMEHP exposure, and those with GSTM1 null type and above-median SMEHP exposure. We used two dummy variables for evaluation of interaction. Dummy variable 1 (DV1): the presence and absence of GSTM1 genotype in our subjects is coded as ‘0’ and ‘1’, respectively. Dummy variable 2 (DV2): the exposure level of sum (MEHP) lower and higher than the median in our subjects is coded as ‘0’ and ‘1’, respectively. Then, we used DV1 DV2 as an interaction term in the model. The level of statistical significance was set at 0.05 for all analyses. All analyses were performed using SAS 9.0 software.
Results
Demographic characteristics of the study
participants
The demographic characteristics of the controls and patients with EN, AD or LEIs are shown (Table I). There was a significant difference in age, using the Wilcoxon rank sum test, between the AD and the LEI groups and the control group, but not the EN group. However, the EN group had a significantly lower BMI (21.5 + 2.9 kg/m2) than the
controls (23.9 + 4.8 kg/m2). No significant differences in menstrual and medical histories, such as age of menarche, abnormal uterine bleeding, menstrual cycle, hormone therapy, IUD use and oral contra-ceptive use, and lifestyles, such as caffeinated drink consumption, ciga-rette smoking and alcohol consumption, were found among the control and disease groups.
Levels of phthalate monoesters
The distribution of seven urinary phthalate monoesters (creatinine-adjusted and non-(creatinine-adjusted) among the disease and control groups is shown in Table II. Among the three disease groups, the most obvious difference in urinary phthalate monoesters when compared with con-trols was for the LEI group. The creatinine-adjusted (mg/g-c) levels of MMP (65.8 mg/g-c), MEHP (6.0 mg/g-c) and SMEHP (52.1 mg/g-c) in the LEI patients were significantly higher than controls (P , 0.05). For the EN group, we found increased median levels of all phthalate metab-olites, except for MEP, but significant differences was found in MnBP (94.1 mg/g-c) and 5oxo-MEHP (19.0 mg/g-c), and SMEHP (42.4 mg/g-c) reached a marginal significance (P , 0.1). The difference in urinary phthalate monoester levels was least evident for the AD group.
ORs for estrogen-dependent gynecologic
conditions with different GSTM1 genotypes
The frequencies and crude- and adjusted OR of the GSTM1 poly-morphisms among all groups are presented in Table III. Genotype of GSTM1 was in HWE in both cases and controls. The frequencies of GSTM1 null type in control, EN, AD and LEIs were 34.5%, 42.9%,68.7% and 47.2%, respectively. Subjects with the GSTM1 null-type had higher crude OR for AD relative to those with wild-type (OR ¼ 4.89; 95% CI, 1.31 – 18.3). After adjustment for age, we found a small
increased risk for subjects with the GSTM1 null type in the AD group (OR ¼ 5.37; 95% CI, 1.25 – 23.0). The OR for the GSTM1 null type in the AD group was significantly higher than in the control ...
Table II Urinary levels of phthalate metabolitesain controls and patients of EN, AD or LEI.
Phthalate monoesters Control (n 5 29) EN (n 5 28) AD (n 5 16) LEI (n 5 36)
Creatinine-unadjusted (ng/ml) MMP 28.1 (1.7 – 657.2) 37.8 (1.7 – 97.7) 25.8 (1.7 – 65.5) 30.8 (1.7 – 113.4) MEP 37.2 (10.6 – 396.2) 31.6 (13.4 – 712.9) 33.8 (9.7 – 96.8) 28.5 (6.7 – 705.9) MnBP 35.4 (5.2 – 247.2) 60.2 (8.2 – 315.2) 30.2 (6.5 – 85.7) 36.2 (5.0 – 323.7) MBzP 5.9 (2.1 – 26.2) 5.6 (2.6 – 46.0) 6.0 (1.7 – 15.2) 5.7 (1.5 – 27.0) 5oxo-MEHP 9.2 (0.1 – 38.6) 12.1 (0.1 – 65.4) 10.2 (0.1 – 803.4) 6.2 (0.1 – 621.4) 5OH-MEHP 5.7 (0.1 – 463.4) 13.6 (0.1 – 269.9) 9.0 (0.1 – 511.1) 8.8 (0.1 – 249.3) MEHP 0.8 (0.8 – 11.3) 2.7#(0.8 – 23.4) 3.2 (0.8 – 145.3) 2.1 (0.8 – 89.3) SMEHPb 28.5 (1.1 – 502.9) 28.7 (1.1 – 305.6) 28.9 (1.1 – 1460) 28.6 (1.1 – 941.3) Creatinine-adjusted (mg/g creatinine) MMP 32.1 (6.9 – 2213) 52.4 (8.3 – 357.6) 42.9 (2.2 – 182.4) 65.8* (14.9 – 442.0) MEP 71.4 (5.6 – 373.3) 58.0 (13.4 – 422.3) 53.4 (13.4 – 147.7) 103.7 (11.2 – 519.0) MnBP 58.0 (9.8 – 479.0) 94.1* (8.9– 669.8) 36.3 (5.7 – 201.1) 75.4#(11.8 – 2346) MBzP 8.9 (2.1 – 38.7) 12.2 (3.0 – 94.7) 10.4 (3.1 – 40.7) 14.5 (2.8 – 112.7) 5oxo-MEHP 7.8 (0.1 – 98.5) 19.0* (0.1– 138.5) 18.5#(0.4 – 860.2) 11.5 (0.2 – 2496) 5OH-MEHP 9.9 (0.1 – 898.1) 16.7 (0.1 – 885.1) 7.2 (0.1 – 621.5) 10.5 (0.1 – 1001) MEHP 3.4 (0.3 – 29.4) 4.2 (0.7 – 76.4) 3.4 (0.5 – 155.6) 6.0* (0.7 – 1263) SMEHPb 18.9 (2.1 – 974.6) 42.4#(1.2 – 1002) 37.9 (3.7 – 1563) 52.1* (4.5 – 3780) a
Median (range); Wilcoxon rank sum test, *P , 0.05;#
P , 0.10; all comparisons are relative to controls. b
SMEHP ¼ MEHP þ 5oxo-MEHP þ 5OH-MEHP, where 5oxo-MEHP and 5OH-MEHP are the metabolites of MEHP.
...
Table I Demographic factors of subjects in controls and patients of EN, AD or LEI.
Parameters Control (n 5 29) EN (n 5 28) AD (n 5 16) LEI (n 5 36) P-valuea
Age (years) 36.2 + 9.0 34.3 + 7.5 43.2 + 6.5b,* 41.1 + 6.8b,* ,0.001*
BMI (kg/m2) 23.9 + 4.8 21.5 + 2.8b,* 24.5 + 4.0 23.5 + 3.7 0.039* Education
Senior high school 22 (75.9) 11 (39.3) 7 (46.7) 24 (66.7)
Technology college 4 (13.8) 10 (35.7) 3 (20.0) 8 (22.2) 0.069#
University 3 (10.3) 7 (25.0) 5 (33.3) 4 (11.1)
Age of menarche (years) 13.4 + 1.3 13.3 + 1.5 12.8 + 0.8 13.2 + 1.8 0.336 Menstrual cycle (days) 28.6 + 3.1 27.6 + 2.5 26.5 + 4.0 30.6 + 11.1 0.255 Abnormal uterine bleedingc 8 (27.6) 10 (35.7) 9 (56.3) 10 (27.8) 0.153
Hormone therapy 6 monthsc 8 (27.6) 8 (28.6) 4 (25.0) 7 (19.4) 0.828 Intrauterine device usec 8 (27.6) 5 (17.9) 6 (37.5) 10 (27.8) 0.550
Oral contraceptive usec 8 (27.6) 7 (25.0) 4 (25.0) 5 (13.9) 0.545
Coffeec 7 (24.1) 15 (53.6) 5 (31.3) 17 (47.2) 0.091# Teac 15 (51.7) 17 (60.7) 5 (31.3) 18 (50.0) 0.313 Cigarette smokingc 3 (10.3) 4 (14.3) 2 (12.5) 7 (19.4) 0.778 Second-hand smokingc 15 (51.7) 13 (46.4) 6 (37.5) 13 (36.1) 0.858 Alcohol consumptionc 3 (10.3) 3 (10.7) 1 (6.3) 5 (13.9) 0.950 a
Kruskall – Wallis test; *P , 0.05;#
P , 0.10; as case number ,5, Fisher’s exact test was applied. b
Wilcoxon rank sum test; *P , 0.05; all comparisons were relative to controls. c
group. However, we did not find any significant ORs for GSTM1 genotypes in the EN and LEI groups, even after adjustment for BMI and age, respectively.
ORs for estrogen-dependent gynecologic
conditions with different phthalates exposure
The crude and adjusted ORs of case status by above-median phthalate exposure are shown in Table IV. After adjustment for the GSTM1 gen-otype and age, OR was 2.87 (P , 0.05) for LEIs in those with MEHP above the median. For those with MnBP and MEHP above the median, the adjusted ORs for EN were 2.77 (0.68 – 11.3) and 2.03 (0.62 – 6.69), respectively, indicating that patients with higher exposure to MnBP or MEHP tended to have increased risk of EN, though without statistical significance.Gene – environmental interaction on
estrogen-dependent gynecologic conditions
The ORs for GSTM1 polymorphisms and urinary phthalate monoe-sters in the disease group using stepwise backward logistic regression are shown in Table V. In the EN group, the OR for BMI was 0.88 (95% CI, 0.75 – 1.04), whereas the OR for the urinary log MnBP was 2.92 (95% CI, 0.92 – 9.2), which only show marginal significant increase in risk. In the AD group, the OR was 1.13 (95% CI, 1.02 – 1.25) for age and 5.30 (95% CI, 1.22 – 23.1) for GSTM1, showing that being elderly or carrying the GSTM1 null type represents a significantly increased risk for AD. For the LEI group, age and the urinary level of SMEHP were associated with an increased risk as the ORs for age and SMEHP were 1.08 (95% CI, 1.01 – 1.16) and 2.64 (95% CI, 0.89 – 7.80), respectively.We found a significant joint effect of phthalates exposure and GSTM1 null type on AD and LEIs (Fig. 1). When compared with those with GSTM1 wild-type and low urinary level of SMEHP, subjects who carried the GSTM1 null type with a high urinary level of SMEHP had a significantly increased risk for AD (crude OR ¼ 8.67; 95% CI, 1.34 – 56.3, Ptrend¼ 0.011) and LEIs (crude OR ¼ 5.57; 95% CI,
1.13 – 27.5, Ptrend¼ 0.037). After adjustment for age, the risk
became even more evident as the OR increased to 10.4 (95% CI, 1.26 – 85.0) for AD and 5.93 (95% CI, 1.10 – 31.9) for LEIs. For inter-action analysis, we found no significant interinter-action between GSTM1 deletion and phthalate exposure for AD (OR ¼ 1.7; 95% CI, 0.09 – 31.2), LEI (OR ¼ 0.6; 95% CI, 0.06 – 5.3) or EN (OR ¼ 3.5; 95% CI, 0.31 – 38.5).
Discussion
This is the first study to explore the gene – environmental interaction of phthalates exposure and GSTM1 polymorphisms on EN, AD and LEI. We found that urinary SMEHP was associated with an increased risk of AD and LEI after adjustment of GSTM1 genotype. No associ-ation was observed between urinary phthalate metabolites and GSTM1 polymorphisms in EN. Previous studies have reported that some phthalates, such as DnBP, BBzP and DEHP, are associated with EN (Cobellis et al., 2003; Reddy et al., 2006a, b). Although signifi-cantly higher plasma phthalates and dose – responses were reported in these studies, recent studies have indicated that urinary phthalate metabolites is a better biomarker of predicting individual phthalates
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... .. T abl e III Fr equencies, crude-and adjus ted-OR for GSTM1 null type in women of EN, AD and LEI. G roups C ontr ol (n 5 29) EN (n 5 28) AD (n 5 16) LEI (n 5 36) GST M1 (n ,%) 19 (65.5 ) 16 (57.1 ) 5 (31.3 ) 19 (52.8 ) GST M1 n ull (n ,%) 10 (34.5 ) 12 (42.9 ) 11 (68.7 ) 17 (47.2 ) OR R efer ence OR 95% C I P-value a OR a 95% CI P-value a OR 95% CI P-value a Crud e — 1.67 (0.56 – 4.94) 0.35 7 4.89* (1.31 – 18.3) 0.01 8 1.99 (0.72 – 5.53) 0.18 8 Adjus ted b — 0.86 (0.27 – 2.71) 0.79 3 5.37* (1.25 – 23.0) 0.02 4 2.30 (0.78 – 6.84) 0.13 3 ax 2tes t; *P , 0.05. bDemographic fac tors wer e adjust ed in the logistic reg res sion models. BMI was adjusted for EN, whereas age was adjusted for AD and LEI, respectively .
exposure due to less contamination during the process of sampling and analysis (Silva et al., 2004; Kato et al., 2005). Determining urinary monoesters and oxidative monoesters of phthalates in our subjects provided the reliable exposure data. Besides, we recruited laparotomy-confirmed controls, which reduced the selection bias from the potential cases in the non-laparotomy-confirmed controls. A few of our controls had gynecologic conditions such as endometrial polyps and ovarian cysts. Although we found no evidence in the litera-ture that phthalate exposures are associated with these diseases, other estrogenic compounds such as bisphenol A (BPA) have been associated with them (Vandenberg et al., 2007). It is thus plausible that phthalates, which are also estrogenic, might be associated with
some of the gynecologic conditions in our control group, despite a paucity of research in the literature. If this was true, the phthalate exposures in our control group would be elevated leading toward the null hypotheses of no association. It is possibly the reason why we did not observe any correlation between phthalate exposure and GSTM1 genotype in EN, which requires a large sample size to clarify. However, the associations between phthalate exposure and GSTM1 polymorphism in AD and LEI still exist.
In the present study, we found that BMI and age are potential risk factors for EN and AD/LEI, respectively, which is consistent with previous studies (Cramer and Missmer, 2002; Vercellini et al., 2006), whereas other demographic characteristics, such as age at ... ... ... ...
Table IV Crude- and adjusted-ORaof case status by above-median phthalate exposure.
Phthalate monoestersb EN (n 5 28) AD (n 5 16) LEI (n 5 36)
OR 95% CI P-value OR 95% CI P-value OR 95% CI P-value
MMP 2.53# 0.87 – 7.39 0.089 1.48 0.42 – 5.16 0.540 2.99* 1.08– 8.26 0.035 MMPadj 2.23 0.73 – 7.15 0.155 1.11 0.26 – 4.73 0.893 2.03 0.68– 6.07 0.203 MEP 0.92 0.32 – 2.63 0.881 0.96 0.28 – 3.27 0.944 1.93 0.72– 5.22 0.193 MEPadj 0.66 0.21 – 2.09 0.476 1.08 0.26 – 4.57 0.916 1.32 0.44– 3.96 0.621 MnBP 3.46* 1.16 – 10.3 0.026 0.98 0.28 – 3.46 0.977 1.83 0.68– 4.95 0.235 MnBPadj 2.93 # 0.92 – 9.31 0.069 0.78 0.18 – 3.33 0.737 1.36 0.46– 4.00 0.581 MBzP 1.23 0.43 – 3.49 0.696 0.74 0.21 – 2.58 0.634 1.72 0.64– 4.62 0.280 MBzPadj 1.07 0.35 – 3.28 0.900 1.33 0.29 – 6.13 0.713 1.40 0.48– 4.05 0.537 5oxo-MEHP 1.90 0.66 – 5.46 0.232 1.58 0.46 – 5.41 0.464 1.23 0.46– 3.28 0.678 5oxo-MEHPadj 2.03 0.64 – 6.37 0.225 1.41 0.34 – 5.86 0.636 1.47 0.51– 4.27 0.479 5OH-MEHP 1.64 0.68 – 4.68 0.354 0.96 0.28 – 3.27 0.945 1.23 0.46– 3.28 0.678 5OH-MEHPadj 1.55 0.51 – 4.77 0.442 1.30 0.30 – 5.63 0.724 1.39 0.48– 4.00 0.542 MEHP 1.64 0.57 – 4.70 0.360 0.98 0.28 – 3.46 0.977 2.90* 1.05– 7.97 0.040 MEHPadj 1.42 0.45 – 4.50 0.547 0.94 0.21 – 4.09 0.929 2.70 # 0.92– 7.89 0.070 SMEHP 2.53# 0.87 – 7.39 0.089 1.90 0.55 – 6.59 0.312 2.66# 0.97– 7.32 0.058 SMEHPadj 2.23 0.71 – 6.95 0.168 1.91 0.44 – 8.25 0.384 2.64 # 0.89– 7.80 0.080 a
We adjusted GSTM1 polymorphism and BMI in EN, whereas age and GSTM1 genotype were adjusted for AD and LEI; control group was used as a reference; *P , 0.05,#P , 0.10. b
We categorized our subjects into two groups according to the median levels of each urinary phthalate metabolite in all subjects.
... ... ... ...
Table V Crude- and adjusted-OR of case status by urinary phthalate monoesters using stepwise backward logistic regressiona.
Parameters EN AD LEI
OR 95% CI P-value OR 95% CI P-value OR 95% CI P-value
BMI 0.88 0.75 – 1.04 0.131 — — — — Age — — 1.13* 1.02 – 1.25 0.020 1.08* 1.01– 1.16 0.023 GSTM1 — — 5.30* 1.22 – 23.1 0.026 2.50 0.81– 7.76 0.112 MnBPb 2.92# 0.92 – 9.29 0.069 — — — — SMEHPb — — 1.91 0.44 – 8.25 0.384 2.64# 0.89– 7.80 0.080 R-squarec 0.179 0.921 0.358 0.112 0.216 0.669 a
Stepwise backward logistic regression, —, excluded from stepwise model; *P , 0.05;# P , 0.10. b
We categorized our subjects into two groups according to the median levels of each urinary phthalate metabolite. c
Nagelkerker R-square (R2
menarche and smoking, did not show statistical difference. The fre-quency of the GSTM1 null genotype has been reported in different ethnic groups, ranging from 21% to 77%, and conflicting conclusions have been offered in different populations. In the current study, the frequency of the GSTM1 null genotype in our controls was 34.5%, which is comparable to the Hans population (35 – 63%; Rebbeck, 1997; Lin et al., 2003). The frequency of GSTM1 null type in EN is 42.9%, which is comparable to the Indian population (39%; Babu et al., 2005), and approximates the Hans population (51%; Ding et al., 2004). Although our results of GSTM1 polymorphisms in EN and controls were different from Japanese and Korean studies (Morizane et al., 2004; Hur et al., 2005), our results are in agreement with most previous studies reporting no association between GSTM1 and EN.
Our data reveal that the risk of SMEHP exposure in AD and LEI with the GSTM1 null genotype was significantly increased. DEHP, the parent compound of MEHP, MEHHP, MEOHP and other phthalates, has estrogenic effects in vivo and in vitro studies (Harris et al., 1997; Jin et al., 2008), although a few studies have conflicting findings (Hong et al., 2005). No human studies have focused on the interaction between phthalates exposure and GSTM1 genotyping in women with AD and LEI. The effects of other estrogenic endocrine disrup-tors on AD may provide some clues. One study showed that low-dose diethylstilbestrol exposure may affect the development of AD in female mice (Huseby and Thurlow, 1982). Another study reported that long-term exposure to BPA, which has similar estro-genic activity to phthalates, can cause AD in female CD-1 mice
(Newbold et al., 2007). Therefore, we suggest that both GSTM1 null and phthalate exposure are associated with AD and LEI, but our study lacked power to test for interactions. The potential effects from phthalates exposure modified by estrogen-receptor-related genotypes or other enzymes on the estrogen-dependent gynecologic diseases are worthy of future investigation. Larger studies to investigate potential interaction between GSTM1 null and phthalate exposure in the etiology of estrogen-dependent gynecologic conditions are warranted.
Interestingly, we also found a significant correlation between urinary MEHP and MnBP (Pearson’s correlation coefficient: r ¼ 0.31, P ¼ 0.001), which is also reported in another study (Becker et al., 2009). Low correlation between urinary MEHP and MnBP may indi-cate that the exposure sources of their parent compounds, DEHP and DnBP, in our subjects could partly come from similar sources. It is reported that high percentages of DEHP and DnBP exposure in humans come from food intake, possibly by contamination in the process of food preparation and leaching out from plastic containers (Schettler, 2006). In addition, DEHP and DnBP are also used in cos-metic and personal care products (DiGangi et al., 2002; Houlihan et al., 2002). The co-exposure of DEHP and DnBP may occur as women tend to use various types of these products, such as deodor-ant, perfume, hair spray, nail polish etc. in their daily routine. However, lacking an integrated toxic equivalence factor of phthalates may underestimate the joint effects of total phthalates exposure and GSTM1 polymorphism in our subjects.
The present study had some limitations. First, although the small sample size may limit our interpretation, our results provide useful data for further investigation. Further, studies with a large sample size are needed. Second, other gynecologic conditions in controls may derive from estrogenic compounds, such as BPA or phthalates, which may lead toward the null result of no association. Third, we did not have age-matched controls for each estrogen-dependent gyne-cologic condition. Because the presenting age of women with EN, AD and LEIs was quiet different, we used laparotomy-confirmed criteria for subject recruitment. Then, we used age as a covariance in the regression models. Fourth, because phthalates are ubiquitous in our environment and consumer products, it is generally hypothesized that phthalate exposure in humans is continual and the intensity is gradually increased, especially in women (DiGangi and Norin, 2002; DiGangi et al., 2002; Houlihan et al., 2002; Duty et al., 2005). In addition, recent studies have also reported the prenatal, post-natal and children’s exposure to phthalate, and its potential effects in early gene expression during embryonic stem cell and neonates (Wolff et al., 2007; Becker et al., 2009; Engel et al., 2009; Huang et al., 2009; van Dartel et al., 2009). However, the missing evidence of phthalate exposure during the critical period for development of measured outcomes is still a limitation of this study. Fifth, because we set 0.05 as the significant level of each statistic testing, it may produce a few false associations by multiple testing. Owing to different phthalates exposure in humans from different sources, it is helpful to identify possible exposure profiles of phthalate in human subject by multiple comparison of phthalate exposure between cases and con-trols. In order to validate our findings, we used stepwise backward logistic regression model to avoid the false association. Therefore, our findings still provide important information for future studies on estrogen-dependent gynecologic conditions.
Figure 1 Joint effects of total MEHP exposure and GSTM1 poly-morphism on adenomyosis (AD) and leiomyoma (LEI) after adjust-ment for age.
Low and high SMEHP groups were divided by the median concentrations of SMEHP (38.9 mg/g c). GSTM1 ( – ) represents the GSTM1 null-type, whereas GSTM1 (þ) represents the GSTM1 wild-type. Asterisk indicates significant difference (P,0.05) as compared to the group of low SMEHP/ wild-genotype group. Significant trend was found for AD (P ¼ 0.011) and LEI (P ¼ 0.037), respectively.
Conclusions
We found a significantly increased risk for AD and LEI in subjects with high MEHP exposure and GSTM1 deletions. However, we did not observe any significant interaction between phthalates exposure and GSTM1 polymorphisms in EN, AD or LEI, probably limited by the sample size. We suggest that both GSTM1 null and phthalate exposure are associated with AD and LEI, but our study lacked power to test for interactions. Larger studies to investigate potential interaction between GSTM1 null and phthalate exposure in the etiology of estrogen-dependent gynecologic conditions are warranted.
Acknowledgements
We are greatly indebted to the gynecologists of the Department of Obstetrics and Gynecology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan, and our colleagues, especially for Ms Yi-Hsin Chang, at the National Health Research Institutes, Miaoli, Taiwan, for subject recruitment and sample collection.
Funding
We are grateful for the financial support from the National Health Research Institutes (grant nos: EO-095-PP-09 and EO-096-PP-09).
References
Al-Hendy A, Salama S. Gene therapy and uterine leiomyoma: a review. Hum Reprod Update 2006;12:385 – 400.
Arvanitis DA, Koumantakis GE, Goumenou AG, Matalliotakis IM, Koumantakis EE, Spandidos DA. CYP1A1, CYP19, and GSTM1 polymorphisms increase the risk of endometriosis. Fertil Steril 2003; 79(Suppl 1):702 – 709.
Babu KA, Reddy NG, Deendayal M, Kennedy S, Shivaji S. GSTM1, GSTT1 and CYP1A1 detoxification gene polymorphisms and their relationship with advanced stages of endometriosis in South Indian women. Pharmacogenet Genomics 2005;15:167 – 172.
Baranova H, Bothorishvilli R, Canis M, Albuisson E, Perriot S, Glowaczower E, Bruhat MA, Baranov V, Malet P. Glutathione S-transferase M1 gene polymorphism and susceptibility to endometriosis in a French population. Mol Hum Reprod 1997; 3:775 – 780.
Baranova H, Canis M, Ivaschenko T, Albuisson E, Bothorishvilli R, Baranov V, Malet P, Bruhat MA. Possible involvement of arylamine N-acetyltransferase 2, glutathione S-transferases M1 and T1 genes in the development of endometriosis. Mol Hum Reprod 1999;5:636 – 641. Becker K, Go¨en T, Seiwert M, Conrad A, Pick-Fuß H, Mu¨ller J, Wittassek M, Schulz C, Kolossa-Gehring M. GerES IV Phthalate metabolites and bisphenol A in urine of German children. Int J Hyg Environ Health 2009;212:685 – 692.
Bulun SE. Endometriosis. New Engl J Med 2009;15:268 – 279.
Cobellis L, Latini G, De Felice C, Razzi S, Paris I, Ruggieri F, Mazzeo P, Petraglia F. High plasma concentrations of di-(2-ethylhexyl)-phthalate in women with endometriosis. Hum Reprod 2003;18:1512 – 1515. Cramer DW, Missmer SA. The epidemiology of endometriosis. Ann N Y
Acad Sci 2002;955:11 – 22.
Devlieger R, D’Hooghe T, Timmerman D. Uterine adenomyosis in the infertility clinic. Hum Reprod Update 2003;9:139 – 147.
DiGangi J, Norin H. Pretty Nasty—Phthalates in European Cosmetics Products. USA: Healthcare Without Harm, and in association with
Women’s Environmental Network and Swedish Society for Nature Conservation, 2002.
DiGangi J, Schettler T, Cobbing M, Rossi M. Aggregate Exposures to Phthalates in Humans. USA: Healthcare Without Harm, and in association with Women’s Environmental Network and Swedish Society for Nature Conservation, 2002.
Ding Y, Chen ZF, Lin RY, Wang XF, Ding JB, Ai XZ, Wen H. Relationship between endometriosis and glutathione S-transferase M1 T1 genes of the Uygurs and Hans in Xinjiang. Zhonghua Fu Chan Ke Za Zhi 2004; 39:101 – 104.
Duty SM, Ackerman RM, Calafat AM, Hauser R. Personal care product use predicts urinary concentrations of some phthalate monoesters. Environ Health Perspect 2005;113:1530 – 1535.
Engel SM, Zhu C, Berkowitz GS, Calafat AM, Silva MJ, Miodovnik A, Wolff MS. Prenatal phthalate exposure and performance on the Neonatal Behavioral Assessment Scale in a multiethnic birth cohort. Neurotoxicology 2009;30:522 – 528.
Giudice L, Kao L. Endometriosis. Lancet 2004;364:1789 – 1799. Guo SW. Glutathione S-transferases M1/T1 gene polymorphisms and
endometriosis: a meta-analysis of genetic association studies. Mol Hum Reprod 2005;11:729 – 743.
Hadfield RM, Manek S, Weeks DE, Mardon HJ, Barlow DH, Kennedy SH, OXEGENE Collaborative Group. Linkage and association studies of the relationship between endometriosis and genes encoding the detoxification enzymes GSTM1, GSTT1 and CYP1A1. Mol Hum Reprod 2001;7:1073 – 1078.
Harris CA, Henttu P, Parker MG, Sumpter JP. The estrogenic activity of phthalate esters in vitro. Environ Health Perspect 1997;105:802 – 811. Hong EJ, Ji YK, Choi KC, Manabe N, Jeung EB. Conflict of estrogenic activity
by various phthalates between in vitro and in vivo models related to the expression of Calbindin-D9k. J Reprod Dev 2005;51:253 – 263.
Houlihan J, Brody C, Schwan B. Not Too Pretty: Phthalates, Beauty Products and the FDA. Environmental Working Group, Coming Clean and Health Care Without Harm. CA, USA. 2002. Available: http://safecosmetics. org/downloads/NotTooPretty_report.pdf [last accessed 26 January 2010].
Huang PC, Kuo PL, Guo YL, Liao PC, Lee CC. Associations between urinary phthalate monoesters and thyroid hormones in pregnant women. Hum Reprod 2007;22:2715 – 2722.
Huang PC, Kuo PL, Ghou YY, Lin SJ, Lee CC. Association between prenatal exposure to phthalates and the health of newborns. Environ Int 2009;35:14 – 20.
Hur SE, Lee JY, Moon HS, Chung HW. Polymorphisms of the genes encoding the GSTM1 GSTT1 and GSTP1 in Korean women: no association with endometriosis. Mol Hum Reprod 2005;11:15 – 19. Huseby RA, Thurlow S. Effects of prenatal exposure of mice to ‘low-dose’
diethylstilbestrol and the development of adenomyosis associated with evidence of hyperprolactinemia. Am J Obstet Gynecol 1982;15:939 – 949. Jin Q, Sun Z, Li Y. Estrogenic activities of di-2-ethylhexyl phthalate. Front
Med China 2008;2:303 – 308.
Juo SH, Wang TN, Lee JN, Wu MT, Long CY, Tsai EM. CYP17, CYP1A1 and COMT polymorphisms and the risk of adenomyosis and endometriosis in Taiwanese women. Hum Reprod 2006;21:1498 – 1502. Kato K, Silva MJ, Needham LL, Calafat AM. Determination of 16 phthalate metabolites in urine using automated sample preparation and on-line preconcentration/ high-performance liquid chromatography/tandem mass spectrometry. Anal Chem 2005;77:2985 – 2991.
Lin J, Zhang X, Qian Y, Ye Y, Shi Y, Xu K, Xu J. Glutathione S-transferase M1 and T1 genotypes and endometriosis risk: a case-controlled study. Chin Med J 2003;116:777 – 780.
Lin LC, Tyan YC, Shih TS, Chang YC, Liao PC. Development and validation of an isotope-dilution electrospray ionization tandem mass
spectrometry method with an on-line sample clean-up device for the quantitative analysis of benzene exposure biomarker SPMA in human urine. Rapid Commun Mass Spectrom 2004;18:1310 – 1316.
Ma M, Kondo T, Ban S, Umemura T, Kurahashi N, Takeda M, Kishi R. Exposure of prepubertal female rats to inhaled di(2-ethylhexyl)phthalate affects the onset of puberty and postpubertal reproductive functions. Toxicol Sci 2006;93:164 – 171.
Moore NP. The oestrogenic potential of the phthalate esters. Reprod Toxicol 2000;14:183 – 192.
Morizane M, Yoshida S, Nakago S, Hamana S, Maruo T, Kennedy S. No association of endometriosis with glutathione S-transferase M1 and T1 null mutations in a Japanese population. J Soc Gynecol Invest 2004; 11:118 – 121.
Newbold RR, Jefferson WN, Padilla-Banks E. Long-term adverse effects of neonatal exposure to bisphenol A on the murine female reproductive tract. Reprod Toxicol 2007;24:253 – 258.
Okubo T, Suzuki T, Yokoyama Y, Kano K, Kano I. Estimation of estrogenic and anti-estrogenic activities of some phthalate diesters and monoesters by MCF-7 cell proliferation assay in vitro. Biol Pharm Bull 2003; 26:1219 – 1224.
Rebbeck TR. Molecular epidemiology of the human glutathione S-transferase genotypes GSTM1 and GSTT1 in cancer susceptibility. Cancer Epidemiol Biomarkers Prev 1997;6:733 – 743.
Reddy BS, Rozati R, Reddy BVR, Raman NVVSS. Association of phthalate esters with endometriosis in Indian women. BJOG 2006a;113:515 – 520. Reddy BS, Rozati R, Reddy S, Kodampur S, Reddy P, Reddy R. High plasma concentrations of polychlorinated biphenyls and phthalate esters in
women with endometriosis: a prospective case – control study. Fertil Steril 2006b;85:775 – 779.
Schettler T. Human exposure to phthalates via consumer products. Int J Androl 2006;29:134 – 139.
Shinka T, Ogura H, Morita T, Nishikawa T, Fujinaga T, Ohkawa T. Relationship between glutathione S-transferase M1 deficiency and urothelial cancer in dye workers exposed to aromatic amines. J Urol 1998;159:380 – 383.
Silva MJ, Barr DB, Reidy JA, Malek NA, Hodge CC, Caudill SP, Brock JW, Needham LL Calafat AM. Urinary level of seven phthalate metabolites in the U.S. population from the national health and nutrition examination survey (NANES) 1999 – 2000. Environ Health Perspect 2004; 112:331 – 338.
van Dartel DA, Pennings JL, Hendriksen PJ, van Schooten FJ, Piersma AH. Early gene expression changes during embryonic stem cell differentiation into cardiomyocytes and their modulation by monobutyl phthalate. Reprod Toxicol 2009;27:93 – 102.
Vandenberg LN, Hauser R, Marcus M, Olea N, Welshons WV. Human exposure to bisphenol A (BPA). Reprod Toxicol 2007;24:139 – 177. Vercellini P, Vigano` P, Somigliana E, Daguati R, Abbiati A, Fedele L.
Adenomyosis: epidemiological factors. Best Pract Res Clin Endocrinol Metab 2006;20:465 – 477.
Wolff MS, Teitelbaum SL, Windham G, Pinney SM, Britton JA, Chelimo C, Godbold J, Biro F, Kushi LH, Pfeiffer CM et al. Pilot study of urinary biomarkers of phytoestrogens, phthalates, and phenols in girls. Environ Health Perspect 2007;115:116 – 121.
