A Pilot Study Associating Urinary Concentrations of Phthalate Metabolites and Semen Quality

Abstract

Phthalates are ubiquitous industrial chemicals that are reported to adversely affect human reproductive outcomes. Divergent effects on semen quality have been reported in a limited number of studies. To assess the possible contribution of regional differences in phthalate exposure to these results, we wished to determine if ambient phthalate exposure of men from the Great Lakes region was associated with human sperm parameters. Male partners (N=45) of subfertile couples presenting to a Michigan infertility clinic were recruited. Urinary concentrations of several phthalate metabolites were measured in these men. Semen parameters, measured according to the World Health Organization [WHO 1999] protocols, were divided into those at or above WHO cutoffs for motility (50% motile), concentration (20thinsp;million/mL) and morphology (4% normal) and those below. Phthalate metabolite concentrations were divided into those concentrations above the median and those at or below the median. Specific gravity was used as a covariate in the regression models to adjust for urine dilution. Low sperm concentration was significantly associated with above median concentrations of monoethyl phthalate (MEP) (OR=6.5, 95% CI: 1.0–43.6) and low morphology with above median concentrations of mono-3-carboxypropyl phthalate (OR=7.6, 95% CI: 1.7–33.3). Increased odds for low concentration and above median concentrations of metabolites of di(2-ethylhexyl) phthalate (DEHP) (OR=5.4, 95% CI: 0.9–30.8) and low morphology and above median concentrations of MEP (OR=3.4, 95% CI: 0.9–13.8) were also found. A significant trend was observed for tertiles of MEP and low sperm concentration (p=0.05). Results suggest that ambient phthalate metabolite concentrations may adversely affect human semen quality.

Keywords:

• biomonitoring
• exposure
• Great Lakes
• phthalate
• sperm

Abbreviations
MEP	=	monoethyl phthalate
DEHP	=	di(2-ethylhexyl) phthalate
DEP	=	diethyl phthalate
DBP	=	di-n-butyl phthalate
BBzP	=	butyl benzyl phthalate
AGD	=	anogenital distance
MBP	=	mono-n-butyl phthalate
MBzP	=	mono-benzyl phthalate
MiBP	=	mono-isobutyl phthalate
MMP	=	mono-methyl phthalate
DnOP	=	di-n-octyl phthalate
LODs	=	limits of detection
SC	=	sport caught
WHO	=	World Health Organization
MEHP	=	mono(2-ethylhexyl) phthalate
MEOHP	=	mono (2-ethyl-5-oxohexyl) phthalate
MEHHP	=	mono (2-ethyl-5-hydroxlhexyl) phthalate
OR	=	odds ratios
TE	=	testosterone

Introduction

Phthalates are a group of chemicals that have been used in a variety of consumer products for over fifty years [Hauser and Calafat [2005]]. Diethyl phthalate (DEP) and di-n-butyl phthalate (DBP) are found mainly in personal care products such as perfumes, shampoos, colognes and deodorants [ATSDR [1995]; 2001; Duty et al. [2005]]. DBP, butyl benzyl phthalate (BBzP) and di(2-ethylhexyl) phthalate (DEHP) are found in building supplies such as floorings, paints, solvents, wallpaper and adhesives [ATSDR [2001]; 2002; NTP-CERHR [2000]; [2005]]. The loosely-bound phthalates are released from their substrates into the environment and have been detected in indoor air and dust [Frome et al. 2004; Wormuth et al. [2006]] and in water near Superfund sites [ATSDR [2001]], so designated by the Environmental Protection Agency (EPA) as sites where toxic wastes were dumped and the EPA is charged with their clean up. Phthalates have been found in food, a portion of which may come from phthalate-containing plastic wrap. Due to their extensive unregulated use, phthalates have become ubiquitous environmental contaminants and widespread human exposure to phthalates has been documented among the general US population [CDC [2005]].

In animals, DBP, BBP and DEHP are considered reproductive and developmental toxicants when administered during critical windows of in utero development [Gray et al. [2000]; Mylchreest et al. [1998]; Parks et al. [2000]]. Malformations, including hypospadias and cryptorchidism, testicular injury leading to lowered sperm counts in adults, and reduced anogenital distance (AGD) have been observed. DBP administered to pregnant rats throughout pregnancy and during lactation reduced male AGD in a dose-dependent fashion with significant reductions at 500 mg/kg/day and 750 mg/kg/day, but not at 250 mg/kg/day [Mylchreest et al. [1998]]. A critical decrease in testosterone biosynthesis involving damage to Leydig cells [Foster [2005]; Parks et al. [2000]] and altered cholesterol metabolism [Barlow et al. [2003]] have been observed. A dose–response relationship between treatment of pregnant rats with 0, 0.1, 1, 10, 30, 50, 100 or 500 mg/kg/day DBP and testosterone concentrations in male offspring was reported, with a maternal dose of 30 mg/kg/day being the lowest dose that produced a significant decrease in testosterone (TE) [Lehmann et al. [2004]]. An age-dependent effect has been reported with less severe effects observed in animals exposed as adults [Nagao et al. [2000]], although increased loss of the seminiferous epithelium and morphologically abnormal sperm were seen after adult exposure to 400 mg/kg/day DBP [Higuchi et al. [2003]]. DEP, however, is not considered a developmental toxin in animals [Fujii et al. 2005; Gray et al. [2000]], although it has been reported to adversely affect Leydig cell ultrastructure in vivo [Jones et al. [1993]].

Evidence is accumulating that certain phthalates may adversely affect the human male reproductive tract. The monoester metabolites, which can be measured in urine, are used as biomarkers of exposure to the diesters [Blount et al. [2000]], and have also been reported to be toxic in rat cell cultures [Li et al. [1998]]. In male babies exposed in utero, prenatal urinary concentrations of mono-ethyl phthalate (MEP), mono-n-butyl phthalate (MBP), mono-benzyl phthalate (MBzP) or mono-isobutyl phthalate (MiBP) were associated with significantly reduced anogenital distance [Swan et al. [2005]], and breast milk concentrations of mono-methyl phthalate (MMP), MEP and MBP have been associated with significantly altered male reproductive hormone concentrations [Main et al. 2006].

In a Boston infertility clinic study, men with above median urinary concentrations of MBP were at increased odds of below the World Health Organization [WHO [1999]] reference level of percent motile sperm (OR=2.4, 95% CI: 1–5.0) [Duty et al. 2003a].A recent re-analysis of their data with 463 men again found significant dose-response associations between MBP concentration and low sperm concentration and low motility [Hauser et al. [2006]]. A Swedish study of military service recruits found that subjects with the highest concentrations of MEP had fewer motile sperm (8.8% mean difference, 95% CI, 0.8–17)[Jonsson et al. 2005]. Two other studies that measured phthalate diesters in human seminal fluid also found adverse effects on sperm parameters [Murature et al. [1987]; Rozati et al. 2002].

Since it is possible that some of the different findings on the effects of phthalates on semen parameters may be due to regional differences in the types and quantities of phthalates used, we conducted a pilot study to determine the relationship between concentrations of phthalate metabolites detected in Great Lakes participants’ urine samples and sperm motility, sperm concentration or sperm morphology.

Results

Demographics and Study Population

The mean age of the participants was 34.8 years (Table 1) and ranged from 23 to 48 years (median 34 years). Most of the men were white (77.8%) and 44.4% had at least 4 years of college. Household income categories ranged from <$15,000 per year to >$90,000 per year with a median household income category of $75,000–<$90,000 per year. The mean BMI of the population was 29.0 (median 27.9) and ranged from 19.0 to 47.5. When categorized by BMI according to guidelines from the Centers for Disease Control and Prevention [CDC 2006] (normal=18.5–24.9, overweight=25–29.9, and obese=30 and above), 24.5% (11) men were of normal weight, 42.2% (19) were overweight and 32.3% (15) were obese. Alcohol consumption ranged from 0 to 35.0 servings per week.

Participant Demographics.

Parameter	Study	Population
		%
Mean age	34.8	–
Racial group
Caucasian	35	77.8
African American	5	11.1
Other race	5	11.1
Education
High school or less	9	20.0
Some post-secondary	16	36.6
At least 4 years of college	20	44.4
Annual income
< $15,000	1	2.27
$15,000–<$45,000	10	22.7
$45,000–<90,000	16	36.4
$90,000 and greater	15	34.1
Not reported	3	6.67
Mean BMI	29.0	–
Smoking status
Current smoker	11	24.4
Former smoker	11	24.4
Lifetime nonsmoker	23	51.1
Alcohol servings/week
None	11	24.5
0.24–3	13	28.9
>3–7.5	10	22.2
>7.5	11	24.4
Sperm parameters (median)
Concentration	42.2	–
Motility	55.0	–
% normal forms	2.0	–
Sperm parameters (mean)
Concentration	50.2	–
Motility	53.5	–
% normal forms	3.2	–

The mean semen concentration was 50.2 million/mL (SD=49.0, median 38.50) ranging from 0.7 to 253.8 million/mL. The mean percent motile sperm was 53.5 (SD=16.8, median 55) ranging from 16 to 93 percent. The mean percent normal morphology was 3.2 (SD=3.1, median 2) with a range of 0 to 10%. The mean days of abstinence from ejaculation was 3.3 (SD=1.4, median 3) and ranged from 0 to 7 days. Sperm concentration differed by race category, with African American men having the highest group mean (p=0.06). The mean sperm concentrations for Caucasian (n=35), African American (n=5) and other group (n=5) were 52.9. million/mL, 61.8 million/mL and 19.3 million/mL, respectively. Subject age was negatively correlated with motility (r=−0.31, p=0.04). Neither smoking status nor BMI was correlated with any semen parameter. However, the lack of effect of smoking or BMI on the outcomes may be due to our small sample size and lack of statistical power.

Phthalate Metabolite Concentrations and Correlations

All phthalate monoester metabolites analyzed were detected in 100% of the participants with the exceptions of MMP which was detected in 72.2%, and MCPP, a metabolite of di-n-octyl phthalate (DnOP), which was detected in 88.9%. The phthalate metabolite with the highest geometric mean concentration was MEP (121.9 μg/L) followed by MBP (26.9 μg/L), MBzP (20.0 μg/L), MEHP (11.5 μg/L), MiBP (6.3 μg/L) and MMP (1.1 μg/L) (Table 2). Mean concentrations of the DEHP oxidative metabolites were MEHHP 66.1 μg/L and MEOHP 39.1 μg/L. The mean concentration of MCPP was 2.5 μg/L. All of the phthalate metabolites were correlated with other metabolites (Table 3). MCPP was significantly correlated with MBP, MEHHP and MEOHP. While MCPP is a major metabolite of DnOP, it is also a minor metabolite of DBP, which explains its correlation with MBP, the major DBP metabolite [Calafat et al. [2006]]. Phthalate metabolites demonstrating significant correlations with semen parameters were MEP with concentration (r=−0.37, p=0.01) and with morphology (r=−0.38, p=0.01).

Distribution of Urinary Phthalate Metabolite Levels.

Phthalate metabolites	Min.	5th	25th	50th	75th	95th	Max.	Geo. mean (μ/L)
MMP	0.0	0.0	1.6	2.4	4.5	14.7	60.4	1.1
MEP	14.8	22.1	43.3	108.0	292.4	1,180.0	2,916.7	121.9
MBP	5.5	7.4	16.0	24.7	44.3	144.5	198.0	26.9
MiBP	1.1	2.3	3.7	5.8	10.0	17.9	61.9	6.3
MBzP	3.8	4.7	11.6	17.4	31.3	166.6	214.1	20.0
MEHP	1.5	2.5	4.6	10.1	22.1	85.4	334.3	11.5
MEOHP	1.8	7.6	20.1	37.6	79.3	263.8	670.0	39.1
MEHHP	3.6	13.9	32.7	56.6	137.1	507.0	1,122.2	66.1
MCPP	0.0	0.0	2.6	4.6	8.3	20.5	40.8	2.5

Spearman's Rank Correlations of Specific Gravity-Adjusted Phthalate Metabolite Levels. P-values are Shown Below the Correlation Coefficients.

	MBzP	MCPP	MEP	MMP	MBP	MiBP	MEHHP	MEHP	MEOHP
MBzP	1.00	0.61***	0.27	0.31	0.50***	0.26	0.33	0.18	0.29
MCPP		1.00	0.07	−0.01	0.34*	0.18	0.34*	0.16	0.33*
MEP			1.00	0.50***	0.46**	0.42**	0.19	0.26	0.17
MMP				1.00	0.49***	0.34*	0.21	0.22	0.23
MBP					1.00	0.52***	0.33*	0.25	0.32*
MiBP						1.00	0.22	0.14	0.24
MEHHP							1.00	0.81***	0.99***
MEHP								1.00	0.81***
MEOHP									1.00

Bolded values are significant: *denotes p≤0.05, **denotes p≤0.01, ***denotes p≤0.001.

Relationship Between Phthalate Metabolite Concentration and Sperm Parameters

Phthalate metabolite concentrations were first assessed for association with the demographic and life style risk factors (Table 4). Education, income, smoking status and BMI were not associated with any phthalate metabolite, but nonwhite race was significantly associated with the combined measure of MiBP and MBP (OR=6.0, 95% CI=1.1–32.5) and alcohol consumption of more than 3 servings per week was associated with high concentrations of he combined measure of MiBP and MBP and with MBzP, although the associations were not significant.

Odds Ratios and 95% Confidence Intervals for High Specific Gravity-Adjusted Phthalate Metabolite Levels and Selected Population Characteristics.

	High phthalate metabolite
Parameter	N	DBP metabolites	DEHP metabolites	MBzP	MCPP	MEP	MMP
Education
≤ Grade 12	9	1.0	1.0	1.0	1.0	1.0	1.0
Some college	16	0.6 (0.1, 3.2)	0.4 (0.1, 2.1)	0.2 (0.0, 1.1)	1.0 (0.2, 5.3)	0.4 (0.1, 2.1)	0.8 (0.1, 3.9)
College graduate	20	0.8 (0.2, 3.9)	0.4 (0.1, 2.1)	0.2 (0.0, 1.4)	0.5 (0.1, 2.6)	0.4 (0.1, 2.1)	1.9 (0.4, 9.2)
Income
≤ $45,000	11	1.0	1.0	1.0	1.0	1.0	1.0
$45,000 to < $90,000	16	0.9 (0.2, 3.9)	2.9 (0.6, 14.3)	1.2 (0.3, 5.6)	2.9 (0.6, 14.3)	0.3 (0.1, 1.3)	1.8 (0.4, 8.4)
≥$90,000	15	1.0 (0.2, 4.5)	1.5 (0.3, 7.5)	1.1 (0.2, 5.0)	0.9 (0.2, 4.5)	0.7 (0.1, 3.2)	3.5 (0.7, 17.9)
Not reported	3	*	*	*	*	*	*
Race
Caucasian	35	1.0	1.0	1.0	1.0	1.0	1.0
Nonwhite	10	6.0 (1.1, 32.5)	0.6 (0.2, 2.6)	0.4 (0.1, 1.6)	0.2 (0.0, 1.0)	3.1 (0.7, 14.1)	0.4 (0.1, 1.6)
Body Mass Index
Normal	11	1.0	1.0	1.0	1.0	1.0	1.0
Overweight	19	0.6 (0.2, 2.9)	0.9 (0.2, 3.9)	1.3 (0.3, 5.9)	2.4 (0.5, 11.1)	0.2 (0.0, 1.0)	0.5 (0.1, 2.4)
Obese	15	0.3 (0.1, 1.5)	1.8 (0.4, 8.7)	1.1 (0.2, 5.0)	1.5 (0.3, 7.5)	0.1 (0.0, 0.7)	0.4 (0.1, 1.9)
Current smoker
No	11	1.0	1.0	1.0	1.0	1.0	1.0
Yes	34	2.2 (0.5, 9.0)	0.9 (0.2, 3.3)	0.8 (0.2, 3.3)	0.9 (0.2, 3.3)	0.9 (0.2, 3.3)	2.2 (0.5, 9.0)
Alcohol servings
≤ 3/week	34	1.0	1.0	1.0	1.0	1.0	1.0
> 3/week	11	3.8 (0.9, 17.0)	0.5 (0.1, 2.1)	2.2 (0.5, 9.0)	0.3 (0.1, 1.3)	1.4 (0.3, 5.3)	0.8 (0.2, 3.3)

*Sample size not sufficient to compute odds ratios. Bolded denotes significant (p<0.05) association.

In multivariate regression models, DEP exposure, assessed by use of MEP concentrations, significantly increased the odds of low sperm concentration (OR=6.5; 95% CI=1.0–43.6) in a model adjusted for nonwhite race (Table 5). MCPP was associated with increased risk of low sperm morphology (OR=7.6; 95% CI=1.7–33.3). The odds for low concentration were increased for DEHP exposure (OR=5.4; 95% CI=0.9–30.8) and the odds for low morphology were increased for DEP exposure (OR=3.4; 95% CI=0.9–13.8), both approaching significance. No significant associations were found between MEHP, MEHHP or MEOHP and any of the sperm parameters (data not shown). All models included specific gravity as a covariate. The wide confidence intervals are to be expected with our small sample size.

Crude and Adjusted Odds Ratios for Phthalate Metabolites and Low Semen Parameters.

	Sperm parameter
	Low sperm concentration		Low sperm motility		Low morphology
	Crude OR	Adjust. OR^a	Crude OR	Adjust. OR^b	Crude OR	Adjust. OR^c
Phthalate metabolite	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)
DBP metabolites	2.5 (0.7, 9.2)	0.5 (0.1, 3.6)	1.1 (0.3, 3.6)	0.8 (0.2, 3.9)	1.6 (0.5, 5.6)	3.3 (0.7, 16.2)
DEHP metabolites	2.5 (0.7, 9.2)	5.4 (0.9, 30.8)	0.7 (0.2, 2.5)	1.6 (0.3, 7.1)	1.1 (0.3, 3.8)	0.9 (0.2, 3.2)
MBzP	1.6 (0.5, 5.8)	1.4 (0.3, 6.3)	0.7 (0.2, 2.5)	1.3 (0.3, 5.5)	1.6 (0.5, 5.6)	0.9 (0.2, 3.2)
MCPP	1.1 (0.3, 3.8)	1.3 (0.3, 5.5)	0.7 (0.2, 2.5)	0.4 (0.1, 1.6)	2.4 (0.7, 3.9)	7.6 (1.7, 33.3)
MEP	2.5 (0.7, 9.2)	6.5 (1.0, 43.6)	0.5 (0.1, 1.7)	0.8 (0.2, 3.3)	2.4 (0.7, 8.5)	3.4 (0.9, 13.8)
MMP	0.5 (0.1, 1.7)	0.8 (0.2, 3.8)	1.1 (0.3, 3.6)	3.2 (0.7, 14.2)	0.8 (0.2, 2.6)	0.9 (0.3, 3.3)

Crude odds ratios were adjusted for specific gravity at the individual subject level. Adjusted odds ratios were developed from multivariable logistic regression models in which specific gravity was adjusted at the group level as an independent variable in the model.

^aAdjusted for nonwhite race.

^bAdjusted for age and servings per week of alcohol.

^cAdjusted only for specific gravity as a separate covariate. Bolded denotes significant (p<0.05) association.

We next explored possible dose–response relationships between the adjusted concentrations of those phthalate metabolites shown to increase risk for low sperm parameters. The trend for tertiles of MEP exposure and low sperm concentration was significant (p=0.05), while that for MCPP tertiles and sperm morphology was not (p=0.78) (Table 6). Neither the trend for the DEHP metabolites (p=0.12) and concentration nor that for MEP and morphology (p=0.14) were significant.

Association of Tertiles of Phthalate Metabolite Concentration with Semen Parameters.

	Sperm concentration^a			Sperm motility^b			Sperm morphology^c
Phthalate tertile	N^d	Crude OR (95% CI)	Adjusted OR (95% CI)	N	Crude OR (95% CI)	Adjusted OR (95% CI)	N	Crude OR (95% CI)	Adjusted OR (95% CI)
DEHP metabolites
1	3	1.0	1.0	4	1.0	1.0	7	1.0	1.0
2	4	1.6 (0.3–8.9)	1.97 (0.2–16.1)	6	0.8 (0.2–3.8)	0.7 (0.1–4.4)	11	1.2 (0.3–5.4)	1.8 (0.4–9.3)
3	7	3.1 (0.6–15.4)	6.65 (0.9–47.3)	6	0.7 (0.2–3.0)	0.7 (0.1–4.4)	10	1.1 (0.3–4.7)	1.2 (0.3, 5.6)
p for trend*		0.16	0.12		0.61	0.50		0.89	0.40
MCPP
1	5	1.0	1.0	5	1.0	1.0	9	1.0	1.0
2	3	0.4 (0.1–2.2)	0.1 (0.0–1.7)	7	1.4 (0.3–6.1)	1.2 (0.2–6.5)	7	0.4 (0.1–1.9)	0.5 (0.1–2.4)
3	6	1.2 (0.3–5.4)	1.5 (0.2–9.4)	4	0.7 (0.1–3.2)	0.4 (0.1–2.2)	12	2.2 (0.4–11.8)	3.1 (0.5–18.8)
p for trend*		0.78	0.78		0.60	0.60		0.37	0.4
MEP
1	2	1.0	1.0	6	1.0	1.0	7	1.0	1.0
2	5	2.7 (0.4–17.0)	3.6 (0.4–34.6)	6	0.80 (0.2–3.5)	0.7 (0.1–3.7)	10	1.67 (0.4–7.2)	4.24 (0.7–24.5)
3	7	5.3 (0.9–32.0)	6.5 (0.6–73.4)	4	0.49 (0.1–2.3)	0.5 (0.1–2.9)	11	2.75 (0.6–13.0)	7.04 (1.0–48.0)
p for trend*		0.06	0.05		0.37	0.45		0.20	0.14
MMP
1	4	1.0	1.0	3	1.0	1.0	10	1.0	1.0
2	6	1.1 (0.2–5.1)	1.3 (0.2–9.2)	6	1.5 (0.3–7.4)	2.3 (0.3–16.3)	9	0.49 (0.1–2.3)	0.71 (0.1–4.1)
3	4	0.7 (0.1–3.2)	0.7 (0.1–5.8)	7	2.1 (0.5–9.7)	3.7 (0.5–28.2)	9	0.47 (0.1–2.1)	0.62 (0.1–3.6)
p for trend*		0.78	0.80		0.33	0.21		0.34	0.40

^aAdjusted for nonwhite race and specific gravity as a separate covariate.

^bAdjusted for age, servings per week of alcohol (≥3 versus ≤3), and specific gravity as a separate covariate.

^cAdjusted only for specific gravity as a separate covariate.

^dIndicates the number of men with below the reference level (“low”) of the indicated sperm parameter and with the indicated phthalate exposure.

*p for trend reflects Mantel–Haenszel Chi-square distribution of tertiles among low semen parameters. Tertile cut points (μg/L): DEHP metabolites: 6.9–76.1, 76.2–157.9, 158.0–2, 126.5; MCPP:0.0–3.1, 3.2–7.1, 7.2–40.8; MEP: 14.8–68.5, 68.6–182.8, 182.7–2, 916.7; MMP: 0.0–1.8, 1.9–3.0, 3.1–60.4.

Discussion

Results from our pilot study suggest that low sperm concentration is associated with high MEP concentration and that low morphology is associated with high MCPP concentration (Table 5). A significant trend for tertiles of MEP and sperm concentration indicates that our results are consistent between statistical approaches (Table 6). Increased odds for low sperm concentration and high concentrations of DEHP metabolites and for low morphology and high MEP concentrations were also observed. Ours is the first study reporting a possible adverse effect of MCPP on a measure of male reproductive health. However, due to our sample size and lack of statistical power, our results must be viewed as suggestive.

Two other studies have examined the effects of urinary phthalate metabolite exposure and human sperm parameters. The study which most closely resembles ours in design initially involved 168 men who were partners of couples attending an infertility clinic in Boston [Duty et al. [2003a]]. Men with above the median concentrations of urinary MBP were at an increased odds of having below the WHO-reference level [[1999]] of percent motile sperm (OR=2.4, 95% CI: 1–5.0) after adjusting for age, abstinence time and smoking. A dose response relationship between tertiles of MBP concentrations and sperm motility (test for trend, p=0.02) and between tertiles of MBzP and sperm concentrations (p=0.02) were also found. A recent re-analysis of the data with 463 men again found significant dose-response associations between MBP concentrations and low sperm concentration and low motility, and a suggestive relationship between MBzP and low sperm concentrations [Hauser et al. [2006]]. In our study, the combined measure of MiBP and MBP was not associated with any sperm parameter, nor were MBP or MiBP when analyzed individually (data not shown). While the unadjusted mean concentration of MBP was higher in our study than that reported by Duty et al. [2003a] (26.9 μg/L vs. 16.1 μg/L, respectively), our samples were collected as first morning voids while their samples were spot collections. It is thus possible that our samples were more concentrated than those collected by Duty et al. [[2003a]], making concentration differences an unlikely explanation for the divergent findings. MiBP concentrations could not be compared as it was not distinguished from MBP in their study. We cannot compare adjusted concentrations even though we both used specific gravity to compensate for urine dilution. In our analysis, specific gravity was added to the models as an independent variable [Schisterman et al. 2005], while in Duty's study the individual phthalate concentrations were standardized for specific gravity. MBzP concentrations were not associated with any of the sperm parameters in our study possibly because we lacked statistical power due to our small sample size. Of note, the Boston group also did not find effects of MBzP on the sperm parameters in their re-analysis [Hauser et al. [2006]]. In contrast, we found a negative association between MEP concentration and sperm concentration, not found by Duty et al. [[2003a]]. They did however, find a negative association between MEP concentration and measures of sperm DNA fragmentation [Duty 2003b], which could lower sperm concentration.

The second study involved 234 young Swedish men who were enrolled during their medical conscript examination [Jonsson et al. [2005]]. Participants in the highest quartile of urinary MEP concentrations had fewer motile sperm (mean difference=8.8%, 95% CI 0.8–17) and more immotile sperm (8.9%, 95% CI 0.8–17). No effect on any other sperm parameter was found for MBP, MBzP or MEHP. We also found an effect of MEP, but on concentration rather than motility. Their study differed in several ways from our study. First, the study populations were quite different, including differences in the median age (median age 18 years, range 18–21 years versus median age 34 years, range 23 to 48 years in our study), and source (general Swedish population versus infertility clients). Increased age has been associated with decreased sperm motility [Zavos et al. [2006]], which may in part explain the different findings. Second, their analytical method for measuring phthalate metabolites differed somewhat from ours as did their statistical analysis. In particular, the limits of detection (LODs) of the phthalate metabolites measured in the Swedish study were higher than the LODs obtained with our method (15 to 30 μg/L vs ∼ 1 μg/L, respectively), and this could have decreased their ability to detect small differences leading to exposure misclassification. Differences in precision between the two analytic methods are more likely to lead to misclassification.

It is also possible that a longer duration of exposure to phthalates, as reflected in the higher mean age of participants in our study compared to the Swedish study, is required to see their adverse effects on adult human sperm. It is interesting to note, however, that all three studies found an adverse effect of MEP on a sperm parameter. Since MEP has not been shown to be a reproductive toxicant in animals, it is also possible that DEP exposure is associated with an unmeasured toxicant and MEP is serving as a marker for exposure to this unknown factor. Other possible explanations for the lack of consistent findings among the three studies include different study populations with potentially different regional phthalate exposures, different sampling designs and different genetic susceptibilities (polymorphisms) to the phthalate effects.

The negative association between sperm morphology and MCPP concentration was unexpected. The parent compound, DnOP, is not produced commercially, but it constitutes approximately 20% of commercial mixtures of C6–C10 phthalates [NTP 2005]. These C6–C10 phthalate mixtures are incorporated into PVC products with many uses, including the manufacturing of flooring and carpet tile, canvas tarps, swimming pool liners, notebook covers, traffic cones, toys, vinyl gloves, shoes, garden hoses, weather stripping and flea collars. DnOP-containing PVC materials are also used in food applications such as seam cements and bottle cap liners and DnOP is approved by the U.S. Food and Drug Administration as an indirect food additive [FDA [2000]]. Human exposure is thought to occur mainly through the consumption of contaminated food. The major metabolite of DnOP is MCPP [Calafat et al. [2006]]. However, if exposure to DBP is much higher than to DnOP, most of the MCPP detected could come from DBP rather than DnOP.

The literature on the health effects of either DnOP or MCPP is sparse. Adult rats fed a diet containing 1,821 DnOP mg/kg body weight per day up to 21 days had decreased serum thyroxine but not serum triiodothyronine concentrations [Mann et al. [1985]], while other rats fed only 350.1 mg/kg body weight per day had decreases in thyroid follicle size and colloid density [Poon et al. [1997]]. No adverse male reproductive effects were seen even at these high doses. It is possible that MCPP is not causing the adverse effects we observed on the sperm parameters, but may be a biomarker for another unmeasured chemical. Clearly MCPP needs to be included in phthalate analyses that are part of studies with lager sample sizes to determine its effects on semen quality.

We used a first morning void urine sample to measure phthalate concentrations. Urinary concentrations of phthalate metabolites have been reported to be fairly stable in several studies [Hauser et al. [2004]; Hoppin et al. [2002]], perhaps reflecting constant exposure to these chemicals. In agreement with other studies [CDC [2005]; Hauser et al. [2006]; Jonsson et al. 2005], we found large inter-individual variation in urinary concentrations of the phthalate metabolites possibly resulting from person-to-person variation in the sources of exposure to the different phthalate diesters and/or to differences in phthalate pharmacokinetics. In our study as well as in other similar studies [Duty et al. 2003a; Jonsson et al. [2005]] and in studies representative of the US population [CDC [2005]], MEP had the highest urinary concentration, followed by MBP, MBzP, MEHP and MMP. MEHHP and MEOHP are oxidative metabolites of DEHP and so were not include in this comparison. The MEP concentration was over five times as high as the next highest metabolite, MBP, in our study (Table 2). A recent report found that men who use cologne or aftershave within 48 hrs of urine collection had higher median concentrations of MEP than did men who did not use these products [Duty et al. [2005]]. It is possible that the everyday use of personal care products, many of which contain DEP [ATSDR [1995]], contributes to the high MEP urinary concentrations.

Caution must be used in interpreting the results of this study since the number of participants was small and it was designed as a pilot study to investigate consumption of sport caught (SC) fish as a phthalate exposure route. Consequently, our questionnaire did not have questions addressing the use of items likely to contain phthalates, such as personal care products, and we could not assess these items as exposure sources. However, to assess other possible confounders, we examined the relationship between potential risk factors recorded in the questionnaires and mean phthalate metabolite concentrations. With the exception of the other race category (correlated with just one of the 6 phthalate metabolites), none of the demographic factors examined in our study were associated with high phthalate metabolite concentrations (Table 4), suggesting that they are not confounding the relationships we observed between the phthalate metabolite concentrations and sperm parameters. It should be kept in mind, however, that study participants were likely exposed to other unmeasured environmental compounds for which phthalates may be a surrogate, or with which phthalates may interact to produce the observed effects. It is unlikely that selection bias occurred in our study, since the original focus was the possible association between SC fish consumption and phthalate concentrations and the participants were selected solely on SC fish consumption levels.

Our study also has several strengths. The same laboratory that measured phthalate metabolite concentrations in the national U.S. surveys [CDC [2005]] as well as in several of the studies reporting human reproductive health effects [Duty et al. 2003a; Hauser et al. [2006]; Swan et al. [2005]] performed the analysis for this study, which facilitates data comparison. We found adverse effects on all three sperm parameters using several analytic techniques, indicating the consistency of our findings. The exposure levels of men with low sperm concentrations were at or slightly lower than those reported for other populations despite differences in study design and/or data analysis. Additionally, the low concentrations at which adverse effects were observed supports the notion that exposure to ambient concentrations of these chemicals may have harmful health effects. Our data supports the observations of Duty et al. [2003a; b] and Jonsson et al. [2005], despite differences in protocols, and extends them to include MCPP, a metabolite of DnOP, not previously associated with human reproductive abnormalities. Studies are needed with greater numbers of participants in order to assess different exposure routes and resolve the associations of specific phthalate metabolites with semen quality.

Materials and Methods

Study Population

Men were recruited for the main study on the effects of environmental contaminants on male fertility when they presented at the University Women's Care clinic in Southfield, Michigan, which is affiliated with Wayne State University, with their partners to seek a diagnosis for failure to conceive. Men with the following medical conditions known to be associated with infertility were excluded from our study: diabetes, thyroid or adrenal disorder, genetic disorders related to fertility, testicular cancer, bilateral orchiectomy, and men using hormonal therapy. Because fertility problems in this population were sometimes attributed to the female partner, the population was comprised of men with normal and subnormal semen parameters. The effects of ambient phthalate exposure on the semen outcomes were determined in 45 participants, 19 of whom had consumed SC fish from the Great Lakes within the past year and 26 of whom had not consumed sport-caught fish within that time period. SC fish consumption was too low and too infrequent (total meals for the past 12 months ranged from 2 to 97; median=4) to enable us to discern its effects on the concentrations of the short-lived phthalates [Peck and Albro 1982], so we wished to combine the two groups. To do so, we first determined if these two groups differed in regards to phthalate metabolite concentrations or study variables possibly associated with phthalate exposure. No differences were found for phthalate concentrations, demographic variables, or semen parameters, so the groups were combined for the analyses. The study protocol and consent documents were described to all eligible men by the recruiters. Men agreeing to participant in the study gave written, informed consent prior to entering into the study. The Internal Review Boards of Michigan State University, Wayne State University, the University of Michigan and the Michigan Department of Community Health reviewed and approved the protocol and instruments for this study.

Specimen Collection and Analysis

Participants collected at home first-morning void urine specimens in sterile cups, pre-tested for phthalate diesters, and brought the cups to the clinic on the day of collection. Specimens were subsequently aliquoted into cryogenic vials and stored at −20°C until they were shipped in batches to the testing laboratory. Phthalate monoester testing was performed at the U.S. Centers for Disease Control and Prevention using previously described methods [Kato et al. [2005]; Silva et al. [2004]]. Briefly, the samples were analyzed by isotope dilution high-performance liquid chromatography coupled with tandem mass spectrometry. Specific gravity, used to correct for urine dilution, was measured by means of a refractometer. No subjects fell outside the recommended range for specific gravity (≥1.01 to ≤1.03) [Teass et al. [1998]].

Semen specimens were collected on-site by masturbation without lubrication and were allowed to liquefy for 30–60 min. Semen analyses, performed by laboratory personnel at University Women's Care clinic, were conducted according to WHO [1999] guidelines. The concentration of immobilized sperm was determined using a counting chamber. Motility was measured on a microscope slide using a microscope fitted with an eyepiece reticle with a 10×10 grid pattern. Motile and immotile sperm were systematically counted in each field in the grid until 200 sperm had been classified, and the average percent motile was calculated. The motile sperm were graded according to the following scale: a=rapid, progressive motility; b=slow or sluggish progressive motility and c=non-progressive motility. The sum of a, b and c equaled the total percent motile sperm reported for data analysis. Sperm morphology was determined using air-dried smears stained with a modified Wright-Giemsa stain. At least 100 sperm in 4 different areas of the slide were evaluated according to Kruger's strict criteria [WHO [1999]]. Results were reported as the percent normal forms among the spermatozoa examined. For the calculation of crude odds ratios and logistic regression analyses, men were divided into those with sperm parameters at or above the WHO cutoffs (concentration 20 million/mL, motility 50%, morphology 4% normal forms) and those below.

Data Collection and Analysis

Participants were administered a questionnaire by the study recruiter, who asked questions on demographic characteristics, life style, occupation, fish consumption, general health and reproductive health. Personnel at the testing laboratory were “blind” as to fertility status of the study participants.

For phthalate metabolites, results were reported using both unadjusted concentrations and concentrations adjusted for specific gravity. Specific gravity adjustment was calculated using the following formula: (phthalate metabolite concentration (μg/L)*[(1.024−1)/sg−1)] where sg is the specific gravity of the specimen. Adjusted values were used to test for associations between phthalate metabolite concentrations and outcomes of interest. For analytical purposes, phthalate monoesters were grouped if they were isomers or were derived from the same parent compound. Thus, MBP and MiBP (isomers) were first converted to molar concentrations and then summed as one variable, as were MEHP, mono(2-ethyl-5-oxohexyl) phthalate (MEOHP) and mono(2-ethyl-5-hydroxlhexyl) phthalate (MEHHP) (all metabolites of di-2-ethylhexyl phthalate) (DEHP) [Barr et al. [2003]; Koch et al. [2003]]. Mono (2-ethyl-5-carboxypentyl) phthalate, a metabolite of DEHP, was only measured in 26 of the samples, and therefore was not included in the analyses. For the regression models, unadjusted phthalate metabolite concentrations were divided into those above the median and those concentrations at or below the median.

Crude odds ratios (OR) of high specific gravity-adjusted phthalate metabolite concentrations (above median) were computed for analysis of possible associations with the semen parameters. Spearman's rank correlations were determined for various concentrations of phthalate metabolites and subject age, body mass index, average servings of alcohol per week, and sperm concentration, motility and morphology. Spearman's correlations also were used to examine the associations among phthalate metabolite concentrations. The Wilcoxon rank sum test was employed to test mean phthalate metabolite concentrations in current smokers and smokers within the past 12 months versus nonsmokers. The chi-square test or Fisher's exact test (where appropriate) was used to test for differences between categorical variables. The Kruskal–Wallis test was used to test for differences in phthalate metabolite concentrations and semen parameters by racial group. To assess potential confounding of phthalate metabolite concentrations by factors associated with life style or socioeconomic status, odds of high phthalate metabolite concentrations were determined for those variables in logistic regression models.

Results of the univariate analyses and information from published reports were used to select candidate variables for multivariable models, using a p value of ≤0.05 as the benchmark for potential inclusion in the model. Using forward stepwise addition of the selected variables to the models, we determined which were significant (p<0.05) and which changed the outcomes by more than 10%. Variables were retained if they were significant, if they were confounders or if their presence helped with model fit even if they weren't a cofounder or an explanatory variable. Logistic regression models were developed for low sperm concentration, low motility and low morphology in which the outcomes of interest were dichotomized according to WHO [1999] criteria for subnormal values. In the models, specific gravity adjustment was performed by including specific gravity as an independent covariate in the model with the unadjusted phthalate metabolite concentration variable (above median). This method of adjustment was chosen to reduce bias produced by automatic specific gravity adjustment of phthalate metabolite concentrations [Barr et al. [2005]; Schisterman et al. [2005]]. In addition, any demographic or life style factor that met the screening criteria for association with either phthalate metabolites concentrations or semen parameters was included in the model. A p value of 0.05 was deemed significant. In order to investigate potential dose-response relationships between the phthalate metabolites concentrations and the semen parameters, phthalate concentrations were divided into tertiles and tested in logistic regression models that controlled for the same factors as the previous models. A p value for trend was determined for the adjusted odds ratios by constructing the same model with the tertiles modeled as a single variable (coded as 0=lowest tertile, 1=intermediate tertile and 2=the highest tertile).

Acknowledgments

The analysis of the phthalate metabolites by the Center for Disease Control and Prevention is gratefully acknowledged. This work is supported by the National Institute of Environmental Health Sciences Grant Number ES11856.