Phthalate exposure and reproductive hormones in adult men

Abstract

BACKGROUND: Phthalates are used in personal and consumer products, food packaging materials, and polyvinyl chloride plastics and have been measured in the majority of the general population of the USA. Consistent experimental evidence shows that some phthalates are developmental and reproductive toxicants in animals. This study explored the association between environmental levels of phthalates and altered reproductive hormone levels in adult men. METHODS: Between 1999 and 2003, 295 men were recruited from Massachusetts General Hospital. Selected phthalate metabolites were measured in urine. Linear regression models explored the relationship between specific gravity-adjusted urinary phthalate monoester concentrations and serum levels of reproductive hormones, including FSH, LH, sex hormone-binding globulin, testosterone, and inhibin B. RESULTS: An interquartile range (IQR) change in monobenzyl phthalate (MBzP) exposure was significantly associated with a 10% [95% confidence interval (CI): −16, −4.0] decrease in FSH concentration. Additionally, an IQR change in monobutyl phthalate (MBP) exposure was associated with a 4.8% (95% CI: 0, 10) increase in inhibin B but this was of borderline significance. CONCLUSIONS: Although we found associations between MBP and MBzP urinary concentrations and altered levels of inhibin B and FSH, the hormone concentrations did not change in the expected patterns. Therefore, it is unclear whether these associations represent physiologically relevant alterations in these hormones, or whether they represent associations found as a result of conducting multiple comparisons.

Introduction

Evidence of widespread human exposure to phthalates and toxicological studies showing that some phthalates are developmental and reproductive toxicants in animals has raised both scientific and public concern about the ubiquitous use of phthalates in consumer and personal care products (ATSDR, 1997, 2001, 2002; CDC, 2003). Phthalates are industrial chemicals with many uses, including making polyvinyl chloride plastics flexible and holding colour and scent in personal care products. Phthalates can be found in perfumes and cologne, paints, insect repellent, medication coatings, hairspray, shampoo, and medical devices including i.v. bags, infusion tubing, and dialysis bags (Nassberger et al., 1987; ATSDR, 1995, 1997, 2001, 2002). As a result of their ubiquitous use, phthalates are present in food, air and water (ATSDR, 1995, 2001, 2002).

Gestational, lactational or pubertal exposures to di-n-butyl phthalate (DBP), butylbenzyl phthalate (BBzP) and di-2-ethylhexyl phthalate (DEHP) in rodents demonstrate an anti-androgenic mechanism of toxicity, but not at the level of the androgen receptor (Gray et al., 1999, 2000; Foster et al., 2001). Exposure to DEHP from gestational day 14 to post-natal day 3 reduced fetal and neonatal testosterone production in rats, thus suggesting a possible mechanism for the reproductive toxicity of phthalates (Parks et al., 2000).

Limited human data are available on the possible relationship between phthalates and testicular function. In previous studies on men who were partners in infertile relationships, we found an inverse dose–response relationship between monobutyl phthalate (MBP), a metabolite of DBP, and sperm motility and concentration, and between monobenzyl phthalate (MBzP), a metabolite of BBzP, and sperm concentration (Duty et al., 2003a). We also found a relationship between monoethyl phthalate (MEP), a metabolite of diethyl phthalate (DEP), and increased sperm DNA damage measured using the neutral single cell gel electrophoresis (comet) assay (Duty et al., 2003b). In the present study on men from the same study population, we explored the relationship between environmental exposure to phthalates and reproductive hormones.

Materials and methods

Subjects

Between 1999 and 2003, men between 18 and 54 years of age were recruited from the Vincent Burnham Andrology laboratory at Massachusetts General Hospital and invited to participate in a study to assess the effects of environmental exposures on male reproductive health. Approximately 65% of eligible men agreed to participate. The main reason cited by non-participants was lack of time. Exclusionary criteria included prior vasectomy or current use of exogenous hormones. A retrospective review of anonymized clinic records of non-participants, who met the same eligibility criteria as the study subjects, found that there were no differences between participants and non-participants in regards to age or semen parameters. Height and weight were measured, and all men completed a brief nurse-administered questionnaire at the time of recruitment which collected basic demographic, reproductive, and health information. The study was approved by the Harvard School of Public Health (HSPH) and Massachusetts General Hospital (MGH) Human Subjects Committees and all subjects signed an informed consent.

Reproductive hormones

One non-fasting blood sample was drawn between 09:00 and 16:00. Blood samples were centrifuged and serum stored at −20°C until analysis. Serum was analysed for hormones as follows. Testosterone was measured directly using the Coat-A-Count RIA kit (Diagnostic Products, USA), which have inter-assay and intra-assay coefficients of variation (CV) of 12 and 10% respectively with a sensitivity of 4 ng/dl (0.139 nmol/l). The free androgen index (FAI) was calculated as the molar ratio of total testosterone to sex hormone-binding globulin (SHBG) (Wilke and Utley, 1987). SHBG was measured using a fully automated system (Immulite: DPC, Inc.) which uses a solid-phase two-site chemiluminescent enzyme immunometric assay. Inhibin B was measured using a commercially available, double antibody, enzyme-linked immunosorbent assay (Oxford Bioinnovation, UK) with inter-assay and intra-assay CV of 20 and 8% respectively, limit of detection (LOD) of 15.6 pg/ml and a functional sensitivity (20% CV) of 50 pg/ml (Groome et al., 1996). Serum LH and FSH concentrations were determined by microparticle enzyme immunoassay using an automated Abbott AxSYM system (Abbott Laboratories, USA). The Second International Reference Preparation (WHO 71/223) was used as the reference standard. The assay sensitivity for LH and FSH were 1.2 and 1.1 IU/l respectively. The intra-assay CV for LH and FSH were <5 and <3% respectively, with inter-assay CV for both hormones of <9%.

Phthalate monoester metabolites in urine

The phthalate monoester metabolites were measured because of potential sample contamination from the parent diester and because some of the metabolites are believed to be the active toxicant as opposed to the parent diester compounds (Peck and Albro, 1982; Li et al., 1998). Phthalate monoesters were measured in a single spot urine sample collected in a sterile specimen cup on the same day as the blood sample. The analytical approach has been described in detail and adapted recently to enable the detection of additional monoesters and improve efficiency of the analysis (Blount et al., 2000; Silva et al., 2003). Briefly, phthalate metabolite determination in urine involved enzymatic deconjugation of the metabolites from their glucuronidated form, solid-phase extraction followed by reversed-phase high-performance liquid chromatography–atmospheric pressure chemical ionization–tandem mass spectrometry using isotope dilution with ¹³C₄ internal standards. Detection limits were in the low nanogram per millilitre range. One method blank, two quality control samples (human urine spiked with phthalate monoesters), and two sets of standards were analysed along with every 21 unknown urine samples. Analysts at the Centers for Disease Control and Prevention (CDC), Atlanta, Georgia, USA, were blind to all information concerning subjects.

Phthalate monoester levels were normalized for urine dilution by specific gravity (SG) adjustment using the following formula: P_c=P[(1.024–1)/SG–1)] where P_c is the SG-corrected phthalate concentration (ng/ml), P is the observed phthalate concentration (ng/ml) and SG is the specific gravity of the urine sample (Boeniger et al., 1993; Teass et al., 1998). Specific gravity was measured using a hand-held refractometer (National Instrument Company, Inc., USA), which was calibrated with deionized water before each measurement.

Statistical analyses

Statistical Analysis Software (SAS) version 8.1 (SAS Institute Inc., Cary, NC, USA) was used for data analysis. Descriptive and summary statistics were generated, outcomes were assessed for outliers and general distributional shape and the associations between covariates and hormone levels were explored for evidence of non-linearity. In preliminary analyses, scatter-plots and Spearman correlation coefficients were used to explore the association between each hormone concentration and each phthalate metabolite concentration. Multiple linear regression analysis was then performed adjusting for appropriate covariates. Residuals were then checked for normality, homogeneity and lack of pattern with respect to covariates of interest. As possible covariates, we considered smoking status (i.e. current and former versus never), race, age, body mass index (BMI), previous infertility evaluation (i.e. yes or no), prior ability to impregnate a partner (i.e. yes or no), season (spring, summer and fall versus winter) and time of day the blood was drawn [morning (09:00 to 12:59) and afternoon (13:00 to 16:00)]; the inclusion of specific covariates in the multivariate models was based on statistical and biological considerations (Hosmer and Lemeshow, 1989). Age and BMI were modelled as continuous independent variables after evaluating appropriateness using a quadratic term, all others as dummy variables. For the secondary analyses, we excluded urine samples with specific gravity values <1.010 (too diluted) or >1.030 (too concentrated) (Teass et al., 1998).

Results

The urinary specific gravity-adjusted concentrations of five phthalate metabolites [i.e. MBP, MEP, MBzP, mono-(2-ethylhexyl) phthalate (MEHP) and monomethyl phthalate (MMP)] and serum levels of reproductive hormones were available for 295 men, for whom demographic characteristics are described in Table I. The men were predominantly Caucasian (83%) with a mean (SD) age and BMI of 36.0 (5.5) years and 28.0 (4.7) respectively. Most men never smoked (72%), 25% had a prior semen analysis and 42% were of proven fertility, defined as a history of impregnating a partner.

Reproductive hormones

The distribution of serum hormones levels is described in Table II. Concentrations of testosterone and inhibin B closely approximated normality and were used in statistical models untransformed. The mean (SD) concentrations of testosterone and inhibin B were 14.67 (5.1) nmol/ml and 160.3 (69.7) pg/ml. Subsequent examination of model residuals suggested that a normal linear regression model was appropriate. The distributions of FSH, LH, SHBG and FAI concentrations (Table II) were skewed. Because log-transformation approximated normality, concentrations were log-transformed for all statistical analyses. The range of concentrations of each hormone was consistent with the MGH laboratory-defined reference ranges for men with the exception of FSH. FSH values ranged from 2.5 to 75.1 IU/l; however, the reference range at MGH is 1.6–16.8 IU/l. Among the men with FSH above the MGH reference range, 90% had at least one abnormal semen parameter (i.e. sperm concentration, sperm motility, or sperm morphology). In contrast, 50% of men with FSH concentrations within the MGH reference range had at least one abnormal semen parameter. Overall, 25% of men had sperm concentrations <20 × 10⁶/ml, 50% had motility below WHO reference range and 35% had <4% sperm with normal morphology.

Phthalate monoesters

There was a wide distribution of both specific gravity-adjusted and unadjusted phthalate monoester levels (Table III). Five phthalate monoesters were detected in 95 to 100% of subjects. MEP was the most prevalent (100%) followed by MBP (99%), then MBzP, MEHP and MMP which were all found in ∼95% of subjects. Half the LOD was imputed for phthalate concentrations below the LOD (ng/ml) [except when quantification was given] as follows: MEP, 0.605; MBzP, 0.235; MBP, 0.47; MEHP, 0.435; and MMP, 0.355. Phthalate metabolite concentrations are presented both adjusted for specific gravity and unadjusted for comparison with other studies. The highest geometric mean levels were found for MEP (183 ng/ml) followed by MBP (16.8 ng/ml), MBzP (7.5 ng/ml), MEHP (6.8 ng/ml) and lastly MMP (4.5 ng/ml). Median MEP levels were 3–10-fold higher than for any other phthalate with a wide distribution of values. MBP levels were much lower than MEP. However, one man had an extremely high MBP level (16 868 ng/ml) (Hauser et al., 2004) and his sample was excluded in the secondary analyses. Forty-five (15%) samples were excluded in the secondary analysis because of extreme specific gravity values (Teass et al., 1998). The final sample size for the primary analysis was 295 and for the secondary analysis, 250.

Covariates

There were suggestive associations between smoking status and MEP, MBP and MMP concentrations. Median MEP levels were higher in current smokers (236 ng/ml) and former smokers (231 ng/ml) than in never smokers (135 ng/ml). This pattern was similar for MMP and MBP (data not shown). Similarly, we observed suggestive associations between race and levels of MEP, MBzP and MBP. African-Americans and Hispanics had 2–4-fold higher MEP levels (506 and 299 ng/ml respectively) than either Caucasians (153 ng/ml) or ‘other races’ (125.3 ng/ml). This pattern, similar to that observed in the general US population (Silva et al., 2004), was also seen for MBzP and MBP (data not shown). The levels of MEHP differed considerably depending upon the time of collection of the urine sample. The mean (median) MEHP levels in the morning [15.1 ng/ml (4.9 ng/ml)] were ∼50% lower than in the afternoon [35.5 ng/ml (7.5 ng/ml)]. Similar trends have been observed for MBzP, MBP and MEHP in the general US population (Silva et al., 2004). In contrast to the phthalate data, the median hormone parameter values were stable across all levels of these covariates with the following exceptions. Increasing age was associated with higher FSH levels, but with lower FAI and testosterone levels, whereas increasing BMI was associated with lower LH, inhibin B, SHBG, testosterone and FAI levels (Table IV). Almost 90% of blood samples were drawn between 10:00 and 14:00, which should minimize diurnal variability in hormone levels. Only inhibin B and testosterone had a significant association with time of blood sampling; both were higher in the morning than in the afternoon. For example, the mean morning inhibin B level (174 pg/ml) was ∼14% higher than the afternoon level (152 pg/ml). Similarly the mean morning testosterone level (15.7 nmol/ml) was ∼12% higher than the afternoon level (14.0 nmol/ml).

Multiple regression analysis

Presented in Table V are the regression coefficients or multiplicative factors and their 95% confidence intervals (CI) for each hormone for an interquartile range (IQR) change in specific gravity-adjusted phthalate monoester concentrations. For the models where both the exposure (phthalate concentrations) and outcome (FSH, LH, SHBG and FAI) were log-transformed, the multiplicative factor is the change in hormone level for an IQR change in phthalate levels after back-transformation of both hormone and phthalate concentrations. Therefore, for an IQR change in phthalate concentration, a multiplicative factor of 1.0 indicates no change in hormone level, a factor <1.0 indicates a multiplicative decrease in hormone level, and a factor >1.0 indicates a multiplicative increase in hormone level. For models where the exposure (phthalate concentrations) was log-transformed, but not the outcome (testosterone and inhibin B), the regression coefficient refers to the change in the hormone level for an IQR change in phthalate levels after back-transformation of phthalate concentration. Therefore, for an IQR change in phthalate concentration, a coefficient of 0 indicates no change in hormone level, a coefficient <0 indicates a decrease in hormone level, and a coefficient >0 indicates an increase in hormone level. All models were adjusted for age, BMI and time of day on which the blood sample was collected; the testosterone model was additionally adjusted for SHBG.

There was a significant negative association between MBzP and FSH, and a borderline positive association between MBP and inhibin B. For example, an increase in MBzP exposure from the 25th to the 75th percentile (i.e. an IQR change) was associated with a decrease in FSH by the multiplicative factor of 0.9 (95% CI: 0.84, 0.96; P=0.003). For the median value of FSH (7.3 IU/l), this represents a 10% (95% CI: −16, −4) decrease in FSH or a 0.73 IU/l decrease for the IQR increase in MBzP. For an IQR increase in MBP, inhibin B increased by 7.33 pg/ml (95% CI: −0.55, 15.21; P=0.07). Therefore, for the median inhibin B concentration (153.5 pg/ml), this represents a 4.8% (95% CI: 0, 10) increase in inhibin B for the IQR increase in MBP levels. MEHP and testosterone exhibited a negative non-significant association [−0.47 nmol/ml (95% CI: −1.03, 0.10; P=0.10)].

Secondary analysis

In secondary analyses, subjects with urine samples that were too concentrated or too dilute (specific gravity <1.010 or >1.030 respectively) were excluded. The associations between the concentrations of MBzP and FSH and between MBP and inhibin B were essentially unchanged. The association between MEHP and testosterone became weaker [−0.42 (95% CI: −1.05, 0.21); P=0.19]. There were suggestive associations between MEP and testosterone and between MMP and FSH. For an IQR change in MEP and MMP, testosterone increased 0.73 nmol/ml (95% CI: −0.05, 1.52; P=0.07) and FSH increased by a multiplicative factor of 1.09 (95% CI: 0.99, 1.20; P=0.07) respectively. For the median of testosterone (14.2 nmol/ml) and FSH (7.3 IU/l), this represents a 5% (95% CI: 0, 11) and a 9% (95% CI: −1, 20) increase respectively. We also excluded one subject with an extreme MBP value and results were unchanged.

Discussion

The present study suggests that exposure to some phthalates, at environmental levels, may be associated with alterations in select reproductive hormone levels in adult men. Specifically, we found a dose–response relationship between concentrations of MBzP and FSH and between MBP and inhibin B. The robustness of these associations was confirmed by the secondary analyses in which we excluded 45 urine samples for being highly concentrated or very diluted. We did not find strong evidence of an association between MEHP and testosterone levels. In the secondary analyses, MEP levels were associated with an increase in serum testosterone and MMP concentrations were associated with an increase in serum FSH. These associations, which were not found in the primary analyses, are not consistent with toxicological studies.

There are limited toxicological data on exposure of adult animals to phthalates, including BBzP, DBP and DEHP, and effects on reproductive hormones. In a study on adult male Fisher 344 rats, Agarwal et al. (1985) evaluated the potential effects BBzP had on male rats' reproductive health following dietary exposure to this phthalate. Plasma testosterone levels were significantly reduced following BBzP exposure, whereas plasma FSH levels increased in a dose-dependent fashion. LH levels were also increased. Marked degenerative changes in the testicular Leydig cells were not observed (Agarwal et al., 1985). The increased FSH and LH concentrations in rats with reduced serum testosterone indicated responsiveness of the negative feedback system and suggested that pituitary–hypothalamic function was not impaired. BBzP metabolizes primarily to MBzP; MBP is also a minor metabolite of BBzP.

Although we found associations between some phthalate metabolites and serum hormone levels in the present study, there were several potential limitations in our design and methodology. We used a single blood and one-spot urine sample to measure hormone and phthalate metabolite concentrations respectively. Despite the diurnal and pulsatile fluctuations in serum hormone levels, in population studies a single blood sample can be used to provide a reliable measure of testosterone and FSH over both short and long time-periods (Bain et al., 1987; Vermeulen and Verdonck, 1992; Schrader et al., 1993). Additionally, requiring multiple blood samples may limit participation rates in epidemiological studies (Schrader et al., 1993).

A single-spot urine sample was used to measure urinary levels of phthalate monoester metabolites. The development of internal dose biomarkers of phthalate exposure allows for accurate assessments of human exposure since urinary concentrations of these metabolites represent an integrative measure of exposure to phthalates from multiple sources and pathways. Since humans rapidly metabolize phthalate diesters to their respective monoesters, phthalates do not bioaccumulate (Peck and Albro, 1982; ATSDR, 2002; Koo et al., 2002). Biological half-lives of phthalates are on the order of 1 day or less and hence represent exposure for no more than the few days preceding the collection of the urine specimen. Since most health endpoints of interest are likely affected by exposures over time-periods longer than a few days, information on the temporal variability of urinary levels of phthalate monoesters is needed to optimize exposure assessment in epidemiological studies. There are limited published data on the temporal variability of urinary phthalate monoester concentrations. A recent study documented acceptable reproducibility of urinary phthalate monoester levels in two first-morning urine specimens collected for 2 consecutive days; day-to-day intra-class correlation coefficients ranged from 0.5 to 0.8 (Hoppin et al., 2002). Time intervals beyond a couple of days were not explored. If there is substantial temporal variability in urinary phthalate monoester levels, the associations between phthalate metabolite levels in urine and reproductive hormones may be attenuated.

Although the men in the present study may not be representative of men from the general population in Massachusetts, generalizability of the results is not necessarily limited. It is a misconception that generalization from a study group depends on the study group's being a representative subgroup of the target population (Rothman and Greenland, 1998). For generalizability to be limited, the associations between reproductive hormones and phthalates in this clinic population would have to differ from the associations within the larger general population. Therefore, we would need to speculate that, compared to others, men visiting this andrology clinic display an altered hormonal response to phthalates. Currently, there is no reason to suspect that the susceptibility to phthalates in the men who visit this andrology clinic is different to that from men who visit other clinics or men from the general population. However, until the results of the present study are replicated in larger and more diverse populations, the generalizability of our results will remain unclear.

We attempted to synthesize the present findings with the results from our previous work investigating the association between phthalates and several other reproductive endpoints. Although we found evidence that MEP exposure at environmental levels was associated with DNA damage in sperm (Duty et al., 2003b), there was little evidence that MEP was associated with any change in semen parameters (count, motility or morphology) (Duty et al., 2003a), or with changes in hormone levels. The implications of these findings are unclear, but suggest that sperm DNA damage may result from mechanisms unrelated to alterations in semen quality or hormone profiles.

In our earlier studies, we also found inverse relationships between both MBP and MBzP and sperm concentration and between MBP and sperm motility (Duty et al., 2003a). Because these two phthalate metabolites are known Sertoli cell toxicants, we hypothesize that inhibin B, produced by Sertoli cells, would be inversely related to these phthalate monoesters. However, in the present study, inhibin B did not decrease but rather increased with higher MBP levels and there was no concurrent increase in FSH levels. Additionally, higher MBzP exposure was associated with a decrease in FSH but no change in inhibin B level. In short, although higher MBP and MBzP exposures were associated with lower sperm counts and motility in our previous work (Duty et al., 2003a), FSH and inhibin B levels did not change in the expected direction. The pattern of change of inhibin B and FSH are inconsistent with other studies which showed that serum inhibin B, in combination with serum FSH levels, were a sensitive marker of impaired spermatogenesis (Uhler et al., 2003). In one of these studies, an inhibin B level <80 pg/ml, in combination with an FSH level >10 mIU/ml, was 100% predictive for sperm concentrations <20 × 10⁶/ml (Jensen et al., 1997). At present, it is unclear why our results were inconsistent with earlier studies supporting the utility of inhibin B and FSH as markers of impaired spermatogenesis.

In conclusion, although we found associations between urinary concentrations of MBP and MBzP and altered serum levels of inhibin B and FSH, the hormone concentrations did not change in the expected patterns. Therefore, it is unclear whether these associations represent physiologically relevant alterations in these hormone levels, or whether they represent associations found as a result of conducting multiple comparisons. Our current understanding of how phthalate exposure affects the interrelationships between hormones, semen parameters and sperm DNA damage is limited and requires further investigation. Enrollment of additional men is ongoing in our study and we plan to perform further analyses on a larger dataset to revisit the preliminary associations found in this report, as well as the associations reported in our other previous studies on phthalate exposure and semen quality and sperm DNA damage.

The authors thank Drs Dana Barr and John A. Reidy of the Centers for Disease Control and Prevention, the staff of the Vincent Memorial Obstetrics and Gynecology Service Andrology Laboratory and In Vitro Fertilization Unit at Massachusetts General Hospital, Dr Patrick Sluss, Director of the Reproductive Endocrine Unit Reference Laboratory at Massachusetts General Hospital, our Research Nurse, Ms Linda Godfrey-Bailey and Research Assistants Ms Ana Trisini and Mr Ramace Dadd as well as data manager Janna Frelich. This publication was supported by grant numbers ES09718 and ES00002 from the National Institute of Environmental Health Sciences (NIEHS), NIH. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS, NIH. Dr S.Duty was supported by NIH Training grant T32 ES07069.

References

Author notes

1Environmental Health Department, Occupational Health Program and 2Biostatistics Department, Harvard School of Public Health, Boston, MA 02115, 3Department of Nursing, School for Health Studies, Simmons College, Boston, MA 02115 4National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA 30341 5Department of Biostatistical Science, Dana-Farber Cancer Institute, Boston, MA 02115 and 6Vincent Memorial Obstetrics & Gynecology Service, Andrology Laboratory and In Vitro Fertilization Unit, Massachusetts General Hospital, Boston, MA 02114, USA