Urinary concentrations of benzophenone-type ultra violet light filters and reproductive parameters in young men

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

International Journal of Hygiene and Environmental Health journal homepage: www.elsevier.com/locate/ijheh

Urinary concentrations of benzophenone-type ultra violet light filters and reproductive parameters in young men ⁎

Evdochia Adoamneia, Jaime Mendiolaa,b,c, , Miriam Moñino-Garcíaa, Fernando Vela-Soriac,d, Luz M. Iribarne-Duránc,d, Mariana F. Fernándezc,d, Nicolás Oleac,d, Niels Jørgensene, Shanna H. Swanf, Alberto M. Torres-Canteroa,b,c,g a

Division of Preventive Medicine and Public Health, Department of Public Health Sciences, University of Murcia School of Medicine, 30100 Murcia, Spain Health Research Methodology Group, Biomedical Research Institute of Murcia (IMIB-Arrixaca), 30120 El Palmar, Murcia, Spain c CIBER de Epidemiología y Salud Pública (CIBERESP), 28029 Madrid, Spain d Instituto de Investigación Biosanitaria (ibs. GRANADA), Hospitales Universitarios de Granada, Departamento de Radiología y Medicina Física, Universidad de Granada, 18010 Granada, Spain e Department of Growth and Reproduction, and International Center for Research and Research Training in Endocrine Disruption of Male Reproduction and Child Health (EDMaRC), Rigshospitalet, University of Copenhagen, DK-2100 Copenhagen, Denmark f Department of Environmental Medicine and Public Health, Icahn School of Medicine at Mount Sinai, 10029 New York, NY, USA g Department of Preventive Medicine, “Virgen de la Arrixaca” University Clinical Hospital, 30120 El Palmar, Murcia, Spain b

A R T I C L E I N F O

A B S T R A C T

Keywords: Benzophenone-type ultra violet light filters Endocrine disruptors Reproductive hormones Semen quality

Background: Benzophenone (BP)-type ultraviolet (UV) light filters are chemicals frequently added to personal care products, insect repellents, sunscreens, and beverage and food packaging to diminish the harmful effects of UV sunlight on human skin or foodstuffs. BP-type UV filters have shown negative effects on male reproduction function in in vitro and animal models, but human epidemiologic studies are limited. The goal of this study was to examine associations between urinary concentrations of BP-type UV filters and semen quality and reproductive hormone levels. Methods: This is a cross-sectional study with 215 young university students (18–23 years old) recruited between 2010 and 2011 in Southern Spain (Murcia Region). All men provided a urine, blood and semen sample on a single day. Urinary concentrations of 2,4-dihydroxybenzophenone (BP-1); 2,2′,4,4′-tetrahydroxybenzophenone (BP-2); 2-hydroxy-4-methoxybenzophenone (BP-3); 2,2′-dihydroxy-4-methoxybenzophenone (BP-8) and 4-hydroxybenzophenone (4OH-BP) were measured by dispersive liquid–liquid microextraction and ultra-high performance liquid chromatography with tandem mass spectrometry detection. Semen quality was evaluated by measuring volume, sperm counts, motility and morphology. Serum samples were analyzed for reproductive hormones, including follicle-stimulating hormone (FSH), luteinizing hormone (LH), testosterone (T), inhibin B and estradiol (E2). Associations between urinary concentrations of BP-type UV filters and semen quality parameters and reproductive hormone levels were examined using linear regression, adjusting for potential confounders. Results: Ninety-seven percent of the men had detectable urinary concentrations of at least one of the five BP-type UV filters quantified. After adjustment for important covariates (body mass index, smoking status and time of blood sample collection), there was a significant positive association between urinary BP-1 and BP-3 concentrations and serum FSH levels (β = 0.08, 95%CI: 0.009; 0.15 and β = 0.04, 95%CI: 0.0002; 0.08, respectively). Urinary BP-1 concentration was also significantly positively associated with T/E2 (β = 0.04, 95%CI: 0.002; 0.07) and negatively with inhibin b/FSH (β = −0.11, 95%CI: −0.21; −0.006) ratio. No significant associations were found between other urinary BP-type UV filters and other reproductive hormone levels or between any semen parameters and any of the urinary BP-type UV filters quantified. Conclusions: Our results suggest that, in young men, urinary BP-type UV filters may be associated with a modest alteration of some reproductive hormones, but the effects we report on reproductive function are likely to be

⁎

Corresponding author at: Division of Preventive Medicine and Public Health, University of Murcia School of Medicine, IMIB-Arrixaca, 30100 Espinardo (Murcia), Spain. E-mail address: [email protected] (J. Mendiola).

https://doi.org/10.1016/j.ijheh.2018.02.002 Received 15 November 2017; Received in revised form 5 February 2018; Accepted 6 February 2018 1438-4639/ © 2018 Elsevier GmbH. All rights reserved.

Please cite this article as: Adoamnei, E., International Journal of Hygiene and Environmental Health (2018), https://doi.org/10.1016/j.ijheh.2018.02.002

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

E. Adoamnei et al.

small, and of unclear clinical significance. Further research is needed to replicate these findings in other male populations.

1. Introduction

215 students (90% of the total contacted) agreed to participate and completed the study visit between October 2010 and November 2011. At the study visit men underwent an andrological examination (assessment of breast, lower abdomen, testicles and penis), provided semen, urine and blood samples and completed questionnaires on general health and lifestyles. The Research Ethics Committee of the University of Murcia approved this study and written informed consent was obtained from all subjects.

Benzophenone (BP)-type ultraviolet (UV) light filters are chemicals coming from different sources generally added to personal care products, insect repellents and sunscreens to diminish the harmful effects of UV sunlight on human skin (Pillai et al., 2005). These compounds are also used for coating surfaces exposed to sunlight due to their UV absorption properties, including protecting colors from bleaching, photoinitiators in ink and adhesives, or food packaging and water containers, where they can leach to foodstuffs (Kawamura et al., 2003; Muncke, 2011; Suzuki et al., 2005). Widespread human exposure to BP-type UV filters is mainly through dermal contact (Jiang et al., 1999; Schlumpf et al., 2004) and dietary ingestion (Muncke, 2011), and are excreted via urine (Janjua et al., 2004, 2008). International biomonitoring studies have reported that BP-type UV filters exposure is common among different populations (USA, Germany, Switzerland), with detectable concentrations in the urine of most of the study participants (Calafat et al., 2008; CDC, 2015; Moos et al., 2014; Philippat et al., 2015; Schlumpf et al., 2010). BP-type UV filters represent around twenty-nine different chemicals, but only a few have been examined with regard to endocrinedisrupting properties. Some BP-type UV filters have been reported to show estrogenic, anti-estrogenic and anti-androgenic effects in in vitro and in vivo studies (Ma et al., 2003; Nakagawa and Tayama, 2011; Kawamura et al., 2005; Schlumpf et al., 2001; Schreurs et al., 2005; Suzuki et al., 2005). In animal models, mainly fish, some studies have demonstrated (Blüthgen et al., 2012; Kim et al., 2014; Weisbrod et al., 2007) a negative impact of BP-type UV filters on male reproductive function, although not all (Daston et al., 1993). However, studies examining the relationship between BP-type UV filters exposure and human reproductive function are limited (Ghazipura et al., 2017). Janjua et al. (2004) observed minor differences in testosterone levels, but not in gonadotrophins (FSH or LH), between levels measured before and two weeks after whole-body topical application of selected sunscreens in healthy volunteers. In a case-control study on male infertility, Chen et al. (2013) found no association between urinary concentrations of 2hydroxy-4-methoxybenzophenone (BP-3) and idiopathic male infertility. Buck Louis et al. (2015), in a cohort study following US couples attempting pregnancy, reported significant inverse associations between urinary concentrations of 2,2′,4,4′-tetrahydroxybenzophenone (BP-2) and sperm concentration, motility and morphology. Recently, urinary BP-3 has been associated with significantly lower total testosterone in male adolescents (12–19 years) participating in the National Health and Nutrition Examination Survey (NHANES) 2011–2012 (Scinicariello and Buser, 2016). However, this is the first study to explore urinary concentrations of BP-type UV filters in relation to semen quality and serum reproductive hormone levels in unselected young men.

2.2. Physical examination and semen analysis Body weight and height were measured using a digital scale (Tanita SC 330-S, London, UK). Body mass index (BMI) was calculated as weight in kilograms divided by squared height in meters. The presence of varicocele or other scrotal abnormalities was evaluated and recorded. Semen analyses were carried out as described in detail elsewhere (Mendiola et al., 2013). Briefly, men were asked to abstain from ejaculation for at least 48 h before sample collection by masturbation without lubrication. Abstinence time was recorded as the time between current and previous ejaculation as reported by the study subject. Ejaculate volumes were estimated by specimen weight, assuming a semen density of 1.0 g/ml. Sperm concentration was evaluated by hemocytometer (Improved Neubauer; Hauser Scientific, Inc., Horsham, PA, USA). The spermatozoa were classified as either motile or immotile (WHO, 2010) to report the percentage of motile spermatozoa [progressive (PR) and non-progressive (NP)]. Total sperm count (TSC) (volume × sperm concentration) was also calculated. Smears for morphology were made, air-dried, fixed, Papanicolaou stained and assessed using strict criteria (Menkveld et al., 1990). The same specialized biologist, that was unaware of the men’s exposure status, carried out all the semen analyses. An external quality control on semen samples throughout the study period was carried out in collaboration with the University of Copenhagen’s Department of Growth and Reproduction. No systematic difference was shown and the mean inter-examiner coefficient of variation was 4.0% with a range for the sets between 1.7 and 7.1%.

2.3. Hormonal analyses Hormone analysis methods have been described previously (Asklund et al., 2007; Cutillas-Tolín et al., 2015). Briefly, blood samples were drawn from participants’ cubital veins on the same time of the day of semen sample collection and were stored and frozen. Serum levels of follicle-stimulating hormone (FSH), luteinizing hormone (LH), and sex hormone–binding globulin (SHBG) were determined using time-resolved immunofluorometric assays (DELFIA; PerkinElmer, Skovlund, Denmark). Intra- and inter-assay variations were < 5% in each of the three assays. Serum testosterone levels were determined using a timeresolved fluoroimmunoassay (DELFIA; PerkinElmer) with intra- and inter-assay variation of < 8%. Estradiol was measured by radioimmunoassay (Pantex, Santa Monica, CA) with an intra-assay variation of < 8% and an inter-assay variation of < 13%. Inhibin b levels were determined by a specific two-sided enzyme immunometric assay (Oxford Bio-Innovation Ltd, Bicester, UK) with intra- and inter-assay variation of 13% and 18%, respectively. Free testosterone (FT) was calculated using the equation of Vermeulen et al. (1999) assuming a fixed albumin of 43.8 g/L. Hormone ratios were also calculated in order to assess potential hormonal dysregulation.

2. Material and methods 2.1. Study population The Murcia Young Men’s Study (MYMS) is a cross-sectional study of university students 18–23 years old in the Murcia Region (Southern Spain), aimed at studying the influence of environmental and lifestyle factors on reproductive parameters. The study rationale and design have been previously described in detail (Cutillas-Tolín et al., 2015; Mendiola et al., 2013; Mínguez-Alarcón et al., 2017). Briefly, a total of 2

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

E. Adoamnei et al.

2.4. Urinary BP-type UV filters analyses

benzophenones amount (free plus conjugated) in urine, each sample was spiked with 50 μL of enzyme solution (β-glucuronidase/sulfatase) and incubated at 37 °C for 24 h. The treated urine was placed in a 15 mL screw-cap glass tube and spiked with 30 μL of the surrogate standard solution (1.25 mgL−1 of BP-D10). Urine was diluted to 10.0 mL with 5% NaCl aqueous solution (w/v) and the pH was adjusted to 2.0. Next, 0.75 mL of acetone and 0.75 mL of trichloromethane were mixed and injected rapidly into the aqueous sample with a syringe. After manual shaking, centrifugation and evaporation of the extract, the residue was dissolved with 100 μL of a mixture consisting of water (0.1% ammonia)/acetonitrile (0.1% ammonia), 70:30 (v/v), and finally 10 μL was injected in the LC system. The sum of BP-1, BP-2, BP-3, BP-8 and 4OH-BP concentrations (in nmol/mL) was also calculated (∑BP). Urinary creatinine concentration (mg/dL) was determined using an automated colorimetric determination based on the Jaffe assay, in the same urine samples in which environmental chemical was assessed.

Single spot urine samples were used to assess five BP-type UV filter concentrations. First-morning urine samples were collected in 100 mL polypropylene urine collection vessels pretested to ensure that they did not contain or leach any of the compounds under study. All urine samples were frozen at −80 °C in 4.5 mL polypropylene vials. Samples were sent to the University of Granada (Spain) on dry ice and stored at −20 °C until analyses were performed. These are the most common BPtype UV filters found in human populations and frequently analyzed and reported regarding reproductive health outcomes (e.g. semen quality, serum hormone levels, endometriosis, uterine leiomyoma, couple fecundity, birth outcomes) (Buck Louis et al., 2014, 2015; Krause et al., 2018; Kunisue et al., 2012; Pollack et al., 2015). In addition, these UV filters have the potential to interfere with essential sperm functions and may impair human fertilization (Rehfeld et al., 2016; Schiffer et al., 2014). Detection and quantification of 2,4-dihydroxybenzophenone (BP-1); 2,2′,4,4′-tetrahydroxybenzophenone (BP-2); 2-hydroxy-4-methoxybenzophenone (BP-3); 2,2′-dihydroxy-4-methoxybenzophenone (BP-8) and 4-hydroxybenzophenone (4OH-BP) were carried out by dispersive liquid–liquid microextraction (DLLME) and ultra-high performance liquid chromatography with tandem mass spectrometry detection (UHPLC–MS/MS) as previously described with minor modifications (Vela-Soria et al., 2014; Jiménez-Díaz et al., 2016). Briefly, urine samples were thawed completely at room temperature, centrifuged at 2600g for 10 min to sediment particulate matter and 0.75 mL were taken to carry out the analysis. In order to determine total

2.4.1. Calibration and quality control Matrix-matched calibration curves were constructed plotting the analyte/surrogate peak area ratio against the analyte concentration, in synthetic urine (Inn et al., 2001). BP-D10 was used as surrogate. Limit of detection (LOD) for all BP-type UV light filters obtained was 0.1 ng mL−1 and limit of quantification (LOQ) was 0.3 ng mL−1, but for BP-3 that was 0.2 ng mL−1 and limit of quantification (LOQ) was 0.6 ng mL−1. Urine samples were extracted in batches of 12, with each batch containing approximately 2 quality-control samples. These quality-control samples included 1 field blank (deionized water which is treated as a sample) and 1 blank urine sample spiked with five

Table 1 Descriptive characteristics of participants by quartile of sum of urinary concentrations of benzophenone (BP)-type ultraviolet (UV) light filters (nmol/mL). Variables

Sum of urinary concentrations of BP-type UV light filtersa Q1 (n = 54)

Q2 (n = 54)

p-valuesf

Q3 (n = 53)

Q4 (n = 54)

Total (n = 215)

Median (5–95) o n (%)20.5 (18.0-23.1) Age (years) BMI (kg/m2) Ejaculation abstinence time (h)b Current smokers, n (%) Presence of varicocele, n (%) Have had Prolonged diseasec, n (%) Taken any medicationd, n (%)

20.6 (18.1–22.6) 23.1 (20.3–29.6) 71.0 (35.0–155) 20 (29.4) 9 (27.3)

20.1 (18.2–22.6) 23.0 (19.0–29.9) 70.0 (38.0–127) 15 (22.1) 8 (24.2)

20.8 (18.5–23.0) 24.6 (19.2–31.5) 74.0 (38.1–133) 17 (25.0) 9 (27.3)

20.5 (18.1–23.1) 23.7 (19.1–30.9) 65.5 (34.0–143) 16 (23.5) 7 (21.2)

20.4 (18.1–22.8) 23.7 (19.4–30.0) 71.0 (39.0–136) 68 (31.6) 33 (15.0)

0.28 0.29 0.16 0.52 0.67

6 (35.3) 11 (21.2)

4 (23.5) 15 (28.8)

3 (17.7) 11 (21.2)

4 (23.5) 15 (28.8)

17 (8.0) 52 (24.3)

0.59 0.48

Semen parameters Seminal volume (mL) Sperm concentration (mill/mL) Total sperm count (Mill) % Motile sperm (PR + NP)e % Morphologically normal sperm

3.0 (1.2–7.2) 40.1 (6.3–125) 127 (12.5–369) 56.6 (38.3–74.8) 9.0 (2.0–22.3)

2.9 (1.3–6.5) 34.1 (12.4–138) 132 (25.4–445) 58.8 (40.1–75.1) 9.0 (2.8–25.8)

3.2 (0.94–7.1) 43.3 (4.5–121) 126 (8.4–448) 57.4 (34.0–71.1) 8.0 (2.7–25.4)

2.8 (0.62–6.0) 53.3 (8.3–147) 106 (16.6–398) 54.9 (35.6–75.3) 9.0 (1.8–21.5)

3.0 (1.0–6.4) 44.0 (8.9–129) 121 (17.8–400) 57.2 (38.9–74.0) 9.0 (2.8–23.0)

0.25 0.79 0.82 0.64 0.83

Reproductive hormones FSH (IU/L) Inhibin b (pg/mL) LH (IU/L) Testosterone (nmol/L) SHBG (nmol/L) Calculated FT (pmol//L) Estradiol (E2) (pmol/L) cFT/LH ratio T/E2 ratio T/LH ratio Inhibin b/FSH ratio

2.1 (0.97–5.4) 175 (80.8–332) 4.1 (2.2–7.1) 21.9 (9.7–38.6) 30.0 (15.0–55.0) 494 (243–832) 77.0 (51.8–125) 117 (60.4–228) 0.28 (0.17−0.42) 5.4 (2.5–10.3) 83.2 (14.8–275)

2.2 (0.80–6.1) 192 (109–361) 4.0 (1.9−7.2) 20.7 (11.8–32.7) 28.5 (15.0–55.0) 456 (255–868) 70.5 (40.0−111) 116 (62.6–261) 0.29 (0.15−0.41) 5.3 (2.8–10.2) 86.5 (23.2–289)

2.4 (0.90–6.5) 203 (100–348) 3.8 (1.8–7.5) 19.4 (11.1–32.2) 27.0 (14.7–50.6) 485 (241–757) 75.0 (46.2–120) 128 (45.3–319) 0.27 (0.17–0.41) 5.5 (2.3–13.8) 78.2 (17.1–318)

2.3 (0.90–4.9) 183 (100–339) 4.4 (1.7–6.8) 22.0 (12.9–37.0) 31.5 (16.0–60.3) 502 (300–989) 76.0 (53.3–135) 119 (52.6–273) 0.28 (0.18–0.47) 5.1 (2.4–11.3) 80.6 (21.3–327)

2.2 (0.93–5.4) 190 (101–336) 4.0 (1.9–7.2) 21.1 (11.5–34.0) 29.0 (16.0–54.2) 482 (276–845) 76.0 (48.0–117) 120 (57.0–260) 0.28 (0.17–0.43) 5.2 (2.6–10.9) 82.3 (21.5–270)

0.83 0.93 0.46 0.26 0.45 0.78 0.59 0.92 0.58 0.98 0.95

Q1: 1st quartile; Q2: 2nd quartile; Q3: 3rd quartile; Q4: 4th quartile. 5–95: 5th–95th percentile. (n): number and (%): percentage. a Sum of urinary concentrations of BP-type UV light filters: Sum of BP-1, BP-2, BP-3, BP-8 and 4OH-BP concentrations (in nmol/mL). b Ejaculation abstinence: period calculated as difference between time of current ejaculation and self-reported time of previous ejaculation. c Long-lasting disease (including diabetes/thyroid disease), sexually transmitted diseases (diagnosed with epididymitis, chlamydia or gonorrhea). d Taken any medication during 3 months prior to participation in study (mostly antibiotics or medication against allergy). e Percentage of motile sperm [progressive + non-progressive (PR + NP)]. f From Kruskal-Wallis test for continuous variables and Chi-squared test for categorical variables among quartiles of sum of urinary concentrations of BP-type UV light filters.

3

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

E. Adoamnei et al.

benzophenone-UV filters at a final concentration of 2.0 ng/mL. Recoveries in the quality control spiked samples were from 88% to 104%, with coefficients of variation (CV) less than 15%.

BMI, smoking status and time of blood sample collection to account for circadian variation in hormone levels (number of minutes after 8:00 am of blood collection) were included. Urinary samples with non-detectable BP-type UV filters concentrations were assigned the value Limit of Detection (LOD) divided by the square root of 2, which has been recommended when the data are not highly skewed (i.e. geometric standard deviation < 3) (Hornung and Reed, 1990). In order to assess the influence of several factors that might have affected results, we carried out three sensitivity analyses. First of all, we fit models adjusting for creatinine as a covariate (Barr et al., 2005). A second sensitivity analysis dichotomizing BP-2 and BP-8 at the median value, irrespective of LODs. Finally, to evaluate the limitations due to exposure detection limits (Schisterman et al., 2006), we fit models using instrument reported values for BP-type UV filters, irrespective of LODs (continuous exposure variables). All tests were two-tailed and the level of statistical significance was set at 0.05. Statistical analyses were performed with the statistical package IBM SPSS 19.0 (IBM Corporation, Armonk, NY, USA).

2.5. Statistical analyses Continuous variables were summarized by geometric mean, standard deviation (SD), median and 5th-95th percentiles, and categorical variables given as number (n) and percentage (%). Multiple linear regression was used to examine the associations between creatinine corrected urinary concentrations of BP-type UV filters (BP-1, BP-3 and 4OH-BP; continuous) and semen parameters and reproductive hormones. Due to low percentage of detection for BP-2 and BP-8, creatinine corrected values were dichotomized at below/above LOD values for these chemical compounds. Semen volume, sperm concentration, total sperm count (TSC), percentage of motile sperm (PR + NP), percentage of morphologically normal sperm, serum FSH and estradiol levels, and urinary concentrations of BP-type UV filters showed skewed (non-normal) distributions and were transformed using the natural log (ln) before analysis. We also examined linearity of associations between urinary concentrations of BP-type UV filters (BP-1, BP-3 and 4OH-BP) and semen parameters and reproductive hormone levels by regressing reproductive parameters on quartiles of creatinine-corrected urinary BP-type UV filters concentrations and testing for linearity of trend using median measure in each quartile as a continuous variable in the linear regression models (p-trend across quartiles). Several potential confounders [e.g. age (years), BMI (kg/m2), ejaculation abstinence time (hours), smoking (yes/no), presence of varicocele (yes/no), cryptorchidism (yes/no), prolonged disease (yes/no), taken any medication (yes/no), season (winter vs. spring, summer or fall), etc.] were assessed. The variables considered as potential confounders included factors previously associated with serum reproductive hormones or semen quality in this or other studies, and factors related to environmental exposures and reproductive outcomes in this study, regardless of whether they had been previously described as predictors of male reproductive health. When inclusion of a potential covariate resulted in a change in the ß-coefficient of < 10%, the variable was not retained in final models. The following variables were included in the final models for semen parameters: BMI, smoking status, presence of varicocele, ejaculation abstinence time, and time to start of semen analysis (for sperm motility only). For reproductive hormones,

3. Results Characteristics of the participants across quartiles of sum of BP-type UV filters are summarized in Table 1. Study participants were young college students [median (5th–95th)] 20.4 (18.1–22.8) years old, with BMI of 23.7 (19.4–30.0) kg/m2. Almost one-third of the subjects smoked (31.6%) and varicocele was detected in 15% of the participants. Median abstinence time was 71.0 (39.0–136) hours, median sperm concentration 44.0 (8.9–129) million/mL, total sperm count 121 (17.8–400) millions, motile sperm (PR + NP) 57.2% (38.9–74.0), morphologically normal sperm 9.0% (2.8–23.0) and semen volume 3.0 (1.0–6.4) mL. In general, all hormones showed serum levels within normal ranges, for example FSH [2.2 (0.93–5.4) IU/L], LH [4.0 (1.9–7.2) IU/L] or testosterone [21.1 (11.5–34.0) nmol/L]. Similar descriptive results were identified for the individual BP-type UV filters measured (data not shown). Overall, 97.7% of the urinary samples had measurable concentrations of at least one of the five BP-type UV filters quantified. Table 2 shows summary statistics for the urinary concentrations of BP-type UV filters. Detection frequencies of BP-1, BP-2, BP-3, BP-8 and 4OH-BP ranged between 19.5% and 97.2%. Urinary geometric mean concentrations of BP-1, BP-3 and 4OH-BP were 2.1, 1.3, and 0.21 ng/mL, respectively. The mean (SD) urinary creatinine concentration was 155

Table 2 Summary statistics for the unadjusted and creatinine corrected urinary concentrations of benzophenone (BP)-type ultraviolet (UV) light filters. BP-type UV light filter

Geometric Mean

Standard Deviation

Unadjusted (ng/mL) BP-1 BP-2 BP-3 BP-8 4OH-BP ∑BP (nmol/mL)b

2.1 0.12 1.3 0.19 0.21 0.11

5.4 2.8 22.5 6.5 0.57 0.12

Creatinine-corrected (μg/g) BP-1 BP-2 BP-3 BP-8 4OH-BP

1.6 0.09 0.97 0.14 0.16

4.2 2.1 16.2 4.8 0.41

% > LODa

97.2 19.5 65.6 27.9 70.7 –

Selected Percentiles 5th

25th

50th

75th

95th

0.21 < LOD < LOD < LOD < LOD 0.003

1.2 < LOD < LOD < LOD < LOD 0.012

2.3 < LOD 1.3 < LOD 0.21 0.026

4.3 < LOD 6.5 0.32 0.34 0.070

12.4 4.7 24.4 10.0 1.0 0.19

0.34 < LOD < LOD < LOD < LOD

0.86 < LOD 0.20 < LOD 0.10

1.6 < LOD 0.96 < LOD 0.15

3.1 0.16 4.6 0.36 0.29

9.9 2.7 16.3 8.9 0.95

Limit of Detection (LOD) for most BP-type UV light filters was 0.1 ng/mL, but for BP-3 was 0.2 ng/mL. BP-1: 2,4-dihydroxybenzophenone; BP-2: 2,2′,4,4′-tetrahydroxybenzophenone; BP-3: 2-hydroxy-4-methoxybenzophenone; BP-8: 2,2′-dihydroxy-4-methoxybenzophenone; 4OH-BP: 4hydroxybenzophenone. a Percentage of samples above the LOD. b ∑BP is the sum of BP-1, BP-2, BP-3, BP-8 and 4OH-BP concentrations (in nmol/mL).

4

5

Reference 1.2% −7.0% −17.9% 0.10 −0.04 0.8% Reference 0.1% 1.1% 2.1% 0.85 0.02 Reference 1.1% 4.9% −19.2% 0.07 −0.06

BP-3 1st quartile 2nd quartile (0.41) 3rd quartile (1.9) 4th quartile (8.9) p-trend across quartiles Continuous BP−8b

4OH-BP 1st quartile 2nd quartile (0.12) 3rd quartile (0.21) 4th quartile (0.54) p-trend across quartiles Continuous

∑BP 1st quartile 2nd quartile (0.017) 3rd quartile (0.037) 4th quartile (0.122) p-trend across quartiles Continuous (−0.13;0.02)

(−23.0;25.2%) (−20.2;30.0%) (−45.1;6.7%)

(−0.06;0.11)

(−23.9;24.1%) (−22.8;25.0%) (−22.6;26.8%)

(−0.09;0.005) (−18.0;19.6%)

(−22.5;24.9%) (−30.8;16.8%) (−41.7;5.9%)

(−0.14;0.03) (−31.7;10.8%)

Reference −0.10% −14.7% 12.7% 0.49 0.03

Reference −8.3% −17.9% 0.7% 0.88 0.01

Reference −17.9% −20.2% 9.4% 0.65 0.02 −1.8%

Reference −31.5% −34.0% 1.2% 0.96 −0.02 0.1%

(−0.09;0.14)

(−37.9;37.6%) (−53.8;24.5%) (−27.6;52.9%)

(−0.12;0.15)

(−45.0;28.4%) (−54.5;18.6%) (−37.1;38.5%)

(−0.05;0.09) (−31.2;27.5%)

(−54.5;18.7%) (−56.9;16.5%) (−27.2;45.9%)

(−0.15;0.11) (−33.0;33.1%)

(−68.0;5.1%) (−70.2;3.7%) (−35.0;37.5%)

Sperm concentration (million/mL)

Reference 2.3% −10.4% −5.2% 0.80 −0.03

Reference −8.5% −16.2% 3.7% 0.98 0.04

Reference −18.1% −26.0% −8.1% 0.61 −0.02 1.7%

Reference −10.6% −34.8% −4.3% 0.54 −0.07 −9.3%

(−0.15;0.09)

(−39.2;43.8%) (−53.5;32.6%) (−49.8;39.4%)

(−0.10;0.19)

(−49.1;32.1%) (−56.6;24.2%) (−38.1;45.5%)

(−0.10;0.06) (−30.5;33.9%)

(−58.6;22.3%) (−66.5;14.6%) (−48.7;32.4%)

(−0.21;0.07) (−45.5;26.8%)

(−51.2;30.0%) (−75.5;5.8%) (−44.7;36.1%)

Total sperm count (Millions)

Reference 0.97 −1.5 −1.4 0.31 −0.63

Reference −2.7 −2.2 −2.1 0.24 −0.63

Reference −2.2 −2.9 −2.2 0.25 −0.64 −2.2

Reference −0.16 −2.5 0.58 0.93 −0.61 −0.51

(−1.7;0.44)

(−3.1;5.0) (−5.7;2.7) (−5.8;2.9)

(−1.5;0.30)

(−6.8;1.2) (−6.2;1.8) (−6.1;1.7)

(−1.4;0.12) (−5.3;0.92)

(−6.2;1.8) (−6.9;1.1) (−6.1;1.8)

(−1.9;0.76) (−4.2;3.2)

(−4.1;3.8) (−6.5;1.5) (−3.4;4.5)

% Motile sperm (PR + NP)

Reference 0.3% −8.0% −6.1% 0.64 −0.03

Reference 3.4% 5.5% −9.4% 0.54 −0.04

Reference −34.3% −10.2% −13.9% 0.65 −0.02 −5.1%

Reference −0.2% −1.3% 5.7% 0.67 −0.01 5.8%

(−0.10;0.05)

(−24.6;25.2%) (−34.0;18.0%) (−32.9;20.7%)

(−0.12;0.05)

(−21.0;27.8%) (−18.9;29.8%) (−34.4;15.7%)

(−0.07;0.03) (−24.2;14.0%)

(−58.3;-10.4%) (−34.1;13.7%) (−38.0;10.2%)

(−0.09;0.07) (−16.2;27.8%)

(−24.8;24.3%) (−25.9;23.4%) (−18.7;30.1%)

Morphologically normal forms (%)

BP-1: 2,4-dihydroxybenzophenone; BP-2: 2,2′,4,4′-tetrahydroxybenzophenone; BP-3: 2-hydroxy-4-methoxybenzophenone; BP-8: 2,2′-dihydroxy-4-methoxybenzophenone; 4OH-BP: 4-hydroxybenzophenone. Percentage (%) change shows natural logarithm values back-transformed to improve interpretability, and untransformed model coefficients shows the mean difference in percentage points between the% of motile sperm (PR + NP) in a given paraben quartile (2nd, 3rd or 4th) and the reference group (1st quartile). a Controlled for BMI, smoking status, presence of varicocele, ejaculation abstinence time, and time to start of semen analysis (for sperm motility only). b Dichotomized at below/above LOD values. c ∑BP is the sum of BP-1, BP-2, BP-3, BP-8 and 4OH-BP concentrations (in nmol/mL), model further controlled for urinary creatinine concentration.

c

Reference 22.5% 4.1% −6.6% 0.29 −0.05 −10.4%

BP-1 1st quartile 2nd quartile (1.2) 3rd quartile (1.9) 4th quartile (5.2) p-trend across quartiles Continuous BP−2b (−1.2;46.2%) (−19.6;27.9%) (−30.2;17.0%)

Semen volume (mL)

Creatinine-corrected median for each quartile of BP-type UV light filter (μg/g)

Table 3 Associations between semen quality parameters and urinary concentrations of benzophenone (BP)-type ultraviolet (UV) light filters [reported as percentage (%) change or untransformed model coefficients, with 95%CI].a

E. Adoamnei et al.

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

6

Reference 0.1% (−20.7;20.8%) 11.1% (−9.5;31.8%) 3.4% (−18.3;25.0%) 0.52 0.02 (−0.06;0.09) Reference 6.4% (−14.7;27.5%) 15.7% (−6.3;37.7%) 18.3% (−4.3;40.9%) 0.14 0.07 (0.009;0.13)d

4OH-BP 1st quartile 2nd quartile (0.12) 3rd quartile (0.21) 4th quartile (0.54) p-trend across quartiles Continuous

∑BPb 1st quartile 2nd quartile (0.017) 3rd quartile (0.037) 4th quartile (0.122) p-trend across quartiles Continuous

Reference 2.4% −1.6% 0.7% 0.95 −0.01 −9.4% Reference −5.1% 2.7% −8.3% 0.56 −0.005 2.3%

BP-1 1st quartile 2nd quartile (1.2) 3rd quartile (1.9) 4th quartile (5.2) p-trend across quartiles Continuous BP−2b

BP-3 1st quartile 2nd quartile (0.41) 3rd quartile (1.9) 4th quartile (8.9) p-trend across quartiles Continuous BP−8b

cFT/LH ratio

Reference −0.5% (−21.0; 20.0%) 21.8% (1.2;42.4%) 11.6% (−9.2;32.4%) 0.10 0.04 (0.002;0.08)d 8.3% (−20.8;11.8%)

BP-3 1st quartile 2nd quartile (0.41) 3rd quartile (1.9) 4th quartile (8.9) p-trend across quartiles Continuous BP−8b

Creatinine corrected median for each quartile of BP-type UV light filter (μg/g)

Reference 9.7% (−10.9;30.4%) 21.8% (1.1;42.6%) 18.6% (−1.9;39.2%) 0.04 0.08 (0.009;0.15)d 7.9% (−10.6;26.4%)

FSH (IU/L)

BP-1 1st quartile 2nd quartile (1.2) 3rd quartile (1.9) 4th quartile (5.2) p-trend across quartiles Continuous BP−2b

Creatinine corrected median for each quartile of BP-type UV light filter (μg/g)

(−0.04;0.03) (−11.8;16.5%)

(−23.0;12.8%) (−15.3;20.7%) (−26.4;9.9%)

(−0.07;0.06) (−25.4;6.5%)

(−15.6;20.4%) (−19.7;16.6%) (−17.2;18.7%)

Reference 12.6 (−16.7;41.9) 2.5 (−28.0;33.1) −0.56 (−31.9;30.8) 0.51 −4.3 (−12.1;3.6)

Reference 7.9 (−20.6;36.4) 18.6 (−9.8;47.0) 12.5 (−17.2;42.2) 0.30 4.5 (−5.9;14.9)

Reference 8.5 (−20.0;36.9) −2.1 (−30.8;26.6) −4.6 (−33.5;24.3) 0.60 −2.1 (−7.7;3.6) 0.59 (−21.9;23.2)

Reference 6.7 (−21.9;35.3) −6.1 (−34.9;22.7) −6.6 (−35.1;21.9) 0.48 −6.9 (−16.8;2.9) 6.4 (−19.2;31.9)

Inhibin b (pg/mL)

Reference −9.9% −6.7% −1.0% 0.97 0.004 6.2%

Reference 14.9% 11.3% 15.2% 0.01 0.04 6.0%

T/E2 ratio

(−0.02;0.02) (−2.1;14.4%)

(−20.3;0.5%) (−17.2;3.7%) (−11.5;9.6%)

(0.002;0.07)d (−3.3;15.4%)

Reference −4.2% 1.0% −7.3% 0.53 −0.005 8.3%

Reference 4.6% 3.1% 2.6% 0.82 −0.001 −9.6%

(−0.04;0.03) (−5.2;21.7%)

(−21.3;12.9%) (−16.2;18.3%) (−24.6;10.1%)

(−0.06;0.06) (−24.9;5.7%)

Reference −10.4% −9.2% −6.9% 0.87 −0.01

Reference −3.7% −2.7% −1.3% 0.87 −0.002

Reference 1.1% 3.7% 0.7% 0.77 0.003 0.01%

Reference −17.4% −10.5% −6.6% 0.45 −0.01 −6.6%

Reference 8.8% −22.6% −10.2% 0.21 −0.05 6.2%

Reference −7.3% −26.7% −20.2% 0.09 −0.11 −3.5%

(continued on next page)

(-17.5;29.9%)

(−20.8;38.5%) (−52.5;7.3%) (−40.3;20.0%) (-0.11;0.01)

(−0.21;-0.006)d (-30.4;23.3%)

(−37.2;22.6%) (−56.8;3.4%) (−50.0;9.6%)

(−0.05;0.02)

(−21.0;0.1%) (−20.2;1.8%) (−18.2;4.4%)

(−0.04;0.04)

(−14.1;6.7%) (−13.1;7.7%) (−12.2;9.6%)

(−0.02;0.02) (−8.2;8.2%)

(−9.4;11.5%) (−6.8;14.2%) (−9.8;11.3%)

(−0.05;0.02) (−15.9;2.6%)

(−27.6;-7.2%) (−20.7;-0.2%) (−16.8;3.5%)

E2 (pmol/L)

Inhibin b/FSH ratio

Reference −53.9 (−129;21.0) −59.3 (−137;18.9) −1.7 (−81.9;78.5) 0.20 2.3 (−20.3;24.8)

Reference −33.1 (−107;40.6) 2.3 (−71.1;75.8) −18.8 (−95.6;58.0) 0.89 −4.3 (−31.3;22.7)

Reference −54.1 (−127;19.1) −10.6 (−84.4;63.2) −1.6 (−75.9;72.8) 0.74 4.9 (−9.6;19.3) 40.1 (−17.8;98.1)

Reference −27.2 (−101;46.2) −7.8 (−81.7;66.0) 47.1 (−25.9;120) 0.17 11.9 (−13.6;37.6) −3.0 (−68.9;62.9)

cFT (pmol/L)

(−12.6;21.9%) (−14.3;20.4%) (−14.5;19.8%)

Reference −1.4 (−5.6;2.6) −1.5 (−5.8;2.8) 1.3 (−3.1;5.6) 0.31 0.31 (−0.89;1.5)

Reference 0.42 (−3.6;4.4) 2.6 (−1.4;6.5) 0.62 (−3.5;4.8) 0.57 0.48 (−0.97;1.9)

Reference −0.32 (−4.3;3.7) −1.2 (−5.2;2.9) 0.76 (−3.3;4.8) 0.82 0.09 (−0.70;0.88) 3.0 (−1.3;6.8)

Reference 0.95 (−3.1;4.9) 2.9 (−1.2;6.9) 2.1 (−1.9;6.1) 0.21 0.59 (−0.79;2.0) −0.39 (−3.9;3.2)

SHBG (nmol/L)

T/LH ratio

Reference −1.9 (−4.5;0.73) −2.1 (−4.8;0.66) −0.06 (−2.8;2.7) 0.20 0.08 (−0.70;0.86)

Reference −1.1 (−3.7;1.4) 0.08 (−2.5;2.6) −0.65 (−3.3;2.0) 0.89 −0.15 (−1.1;0.79)

Reference −1.9 (−4.4;0.66) −0.37 (−2.9;2.2) −0.06 (−2.6;2.5) 0.74 0.17 (−0.33;0.67) 1.4 (−0.62;3.4)

Reference −0.94 (−3.5;1.6) −0.27 (−2.8;2.3) 1.6 (−0.90;4.2) 0.17 0.42 (−0.47;1.3) −0.10 (−2.4;2.2)

T (nmol/L)

(4.6;25.2%) (1.0;21.7%) (4.9;25.4%)

Reference −0.04 (−0.72;0.63) 0.01 (−0.69;0.72) 0.31 (−0.41;1.0) 0.39 0.07 (−0.11;0.25)

Reference 0.21 (−0.45;0.87) 0.09 (−0.57;0.74) 0.26 (−0.43;0.94) 0.57 0.04 (−0.20;0.28)

Reference −0.16 (−0.81;0.50) −0.06 (−0.72;0.60) 0.29 (−0.38;0.95) 0.40 0.05 (−0.08;0.18) −0.13 (−0.65;0.39)

Reference −0.20 (−0.86;0.46) 0.15 (−0.51;0.81) 0.22 (−0.44;0.87) 0.36 0.12 (−0.11;0.34) 0.21 (−0.38;0.80)

LH (IU/L)

Table 4 Associations between serum reproductive hormone concentrations and urinary concentrations of benzophenone (BP)-type ultraviolet (UV) light filters [reported as percentage (%) change or untransformed model coefficients, with 95%CI].a

E. Adoamnei et al.

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

(−0.07;0.03) (−0.02;0.04) (−0.08;0.03)

Bolded values are significant. BP-1: 2,4-dihydroxybenzophenone; BP-2: 2,2′,4,4′-tetrahydroxybenzophenone; BP-3: 2-hydroxy-4-methoxybenzophenone; BP-8: 2,2′-dihydroxy-4-methoxybenzophenone; 4OH-BP: 4-hydroxybenzophenone. Percentage (%) change shows natural logarithm values back-transformed to improve interpretability, and untransformed model coefficients shows the mean difference in percentage points between the hormone values in a given BP quartile (2nd, 3rd or 4th) and the reference group (1st quartile). a Controlled for body mass index (BMI), smoking status and time of blood sample collection. b Dichotomized at below/above LOD values. c ∑BP is the sum of BP-1, BP-2, BP-3, BP-8 and 4OH-BP concentrations (in nmol/mL), model further controlled for urinary creatinine concentration. d p-value < 0.05.

(-0.18;0.002)

(−30.7;30.7%) (−44.4;19.8%) (−49.0;16.9%)

Reference 0.01% −12.3% −16.1% 0.29 −0.09 (−22.0;13.1%) (−23.7;13.0%) (−25.7;11.9%)

Reference −4.5% −5.4% −6.9% 0.57 −0.02 (−7.0;14.5%) (−9.8;12.6%) (−3.1;19.9%)

Reference 3.7% 1.4% 8.4% 0.26 0.01 Reference −3.7% −4.3% −8.9% 0.42 −0.02

(−0.09;0.04)

∑BPc 1st quartile 2nd quartile (0.017) 3rd quartile (0.037) 4th quartile (0.122) p-trend across quartiles Continuous

(−22.0;14.6%) (−23.4;14.8%) (−28.5;10.7%)

(-0.09;0.13) (−0.04;0.04)

(−0.08;0.04)

Reference 7.2% −3.5% 5.7% 0.92 0.02 Reference −8.7% −0.1% −9.9% 0.49 −0.02 Reference −1.2% 4.5% −1.8% 0.95 −0.004 Reference −10.6% −4.6% −12.0% 0.33 −0.03 4OH-BP 1st quartile 2nd quartile (0.12) 3rd quartile (0.21) 4th quartile (0.54) p-trend across quartiles Continuous

Creatinine corrected median for each quartile of BP-type UV light filter (μg/g)

Table 4 (continued)

cFT/LH ratio

(−28.5;7.3%) (−22.5;13.2%) (−30.7;6.6%)

T/E2 ratio

(−11.7;9.3%) (−5.9;15.0%) (−12.7;9.1%)

(−25.8;8.4%) (−17.1;16.9%) (−27.7;7.9%)

Inhibin b/FSH ratio T/LH ratio

(−22.8;37.3%) (−33.4;26.4%) (−25.6;37.0%)

E. Adoamnei et al.

(78.1) mg/dL (median, 146 mg/dL). Creatinine corrected geometric mean (25th–75th) urinary BP-1, BP-3 and 4OH-BP concentrations were 1.6 (0.86–3.1) μg/g, 0.97 (0.20–4.6) μg/g and 0.16 (0.10–0.29) μg/g, respectively. Table 3 summarizes multivariate analysis for urinary concentrations of BP-type UV filters and men’s semen quality parameters. Neither crude (not shown) nor adjusted significant associations were found with semen parameters. Multivariate analysis for urinary concentrations of BP-type UV filters and men’s reproductive hormones are shown in Table 4. After adjustment for important covariates, there was a significant positive association between urinary BP-1, BP-3 and ∑BP concentrations and serum FSH levels (β = 0.08, 95%CI: 0.009;0.15, β = 0.04, 95%CI: 0.0002;0.08, β = 0.07, 95%CI: 0.009;0.13, respectively) (Fig. 1). Urinary BP-1 concentration was also significantly positively associated with T/E2 (β = 0.04, 95%CI: 0.002; 0.07) (Fig. 2) and negatively with inhibin b/FSH ratio (β = −0.11, 95%CI: −0.21; −0.006) (Fig. 3). No significant associations were found between other BP-type UV filters and other reproductive hormone levels or ratios or any semen parameters. Lastly, similar results were obtained after carrying out the three sensitivity analyses. 4. Discussion We observed no associations between urinary concentrations of BPtype UV filters and semen quality parameters. However, urinary concentrations of BP-1 and BP-3 were associated with altered FSH levels, T/E2 and inhibin b/FSH ratios in our study population. To our knowledge, this is the first study to evaluate the associations between urinary concentrations of BP-type UV filters and markers of reproductive function in young unselected men. The positive association between urinary BP-1 and T/E2 ratio supports the hypothesis that BP-type UV filters affect steroidogenesis through inhibition of the conversion of testosterone to estradiol by aromatase. However, we did not find any increase in LH levels, which would have been expected under this hypothesis. The positive association between BP-1 and serum FSH levels coupled with a negative association with inhibin b/FSH ratio might reflect a primary negative effect on spermatogenesis (Sertoli cell function), which, however, was compensated for by an increased FSH. To our knowledge, only two previous studies have investigated associations between BP-type UV filters exposure and serum reproductive hormones in men (Janjua et al., 2004; Scinicariello and Buser, 2016), with results that are inconsistent with ours. In an experimental human study by Janjua et al. (2004) slightly lower testosterone levels, but no changes in gonadotrophins (FSH or LH) were observed after wholebody topical application of selected sunscreens (e.g. BP-3) by healthy volunteers. Scinicariello and Buser (2016) reported that urinary BP-3 concentrations were associated with lower total testosterone in male adolescents (12–19 years) participating in NHANES 2011–2012. This study did not examine the associations of BP-3 with serum free testosterone or gonadotrophin levels. Our results are, however, consistent with experimental studies in fish demonstrating a negative effect of BP-3 exposure on steroidogenesis (Blüthgen et al., 2012) and an increase of the T/E2 ratio followed by a significant increase in plasma testosterone concentrations (Kim et al., 2014), as BP-3 can be transformed to BP-1 (Kim et al., 2014). Our results are consistent with a previous case-control study that reported no associations between urinary concentrations of BP-3 in 877 idiopathic infertile men (including a subset of men having abnormal semen parameters) versus 713 fertile controls (Chen et al., 2013). However, our findings differ from results from a recent study reporting significant inverse associations between urinary concentrations of BP-2 and sperm concentration, motility and tail morphology in 413 men participating in a cohort study of US couples attempting pregnancy (Buck Louis et al., 2015). In this case, differences between these results and those from our study could be related to differences in population, 7

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

E. Adoamnei et al.

et al., 2013). Lastly, in Danish (Frederiksen et al., 2014) and German (Moos et al., 2014) male volunteers from the general population unadjusted urinary concentrations of BP-3 at the 95th percentile were higher (80.4 ng/mL) and lower (8.5 ng/mL) respectively, compared to 24.4 ng/mL in our study population. This study is limited in a number of ways. Most importantly, our statistical power was relatively low. Nonetheless, our study had adequate power (80%) to detect, for example, differences of 40% decrease in sperm concentration or 42% decrease in total sperm count between men in the first and fourth quartile of sum of urinary concentrations of BP-type UV filters concentrations. Both, our exposure assessments and outcomes were based on a single sample. With respect to reproductive outcomes, it has been shown that one sample is enough to assess semen quality in epidemiological studies (Chiu et al., 2017; Stokes-Riner et al., 2007) and a single sample can adequately classify men’s reproductive hormones (Vermeulen and Verdonck, 1992). However, newer approaches (e.g., measures of DNA damage) may be more sensitive indicators of reproductive toxicity and a consideration for future work. Moreover, semen quality and reproductive hormone levels are proxies for male fecundity with limited utility for predicting actual couple fertility. Therefore, it is still entirely possible that UV filters could have an effect on male fertility potential outside of the mechanisms examined here as well. With respect to exposure, several studies reported that, despite variability in BP-type UV filters concentrations, a single spot urine measurement seems to correctly represent the long-term exposure of an individual, including relatively high intraclass correlation coefficient (ICC) ranging from 0.46–0.81 (Dewalque et al., 2015; Koch et al., 2014; Lassen et al., 2013; Meeker et al., 2013). Nevertheless, exposure measurement error or misclassification cannot be ruled out but, if non-differential, one may expect on average that this would likely bias effect estimates toward the null. Selection bias is unlikely, since our study subjects were unselected young men unaware of their fertility status (semen quality and reproductive hormone levels) or BP-type UV filters exposure. It is worth mentioning that our results do not exclude a potential negative impact of these exposures in fetal life, especially during the critical masculinization programming window (Hsieh et al., 2007; van den Driesche et al., 2017). Finally, as with all observational studies causal inference is limited, and chance and confounding by unmeasured factors, or by imperfectly measured factors (residual confounding) may remain as possible explanations for our findings.

Fig. 1. Adjusted percentage change (and 95% CI) in serum FSH levels using the 1st quartile of the urinary concentration of 2,4-dihydroxybenzophenone (BP-1) quartiles as the reference.

exposure levels, methods of assessing semen parameters, sample size and/or also chance. Similarly, in the same cohort study, male BP-2 concentrations were associated with reduced couple fecundity (FOR = 0.69, 95% CI: 0.49; 0.97), suggesting that male exposure to select UV filters may diminish couples' fecundity (Buck Louis et al., 2014). The body of animal evidence focusing on BP-type UV filters exposure and male reproduction is very limited (Ghazipura et al., 2017). Daston et al. (1993) demonstrated that no reproductive parameters, including reproductive organ weight or sperm parameters, were affected by BP-3 exposure in mice. However, other studies in mice and fish have shown the disruption of male reproductive system, including gonadal development or spermatogenesis, induced by BP-2 or BP-3 exposure (French, 1992; Weisbrod et al., 2007). Creatinine-corrected urinary concentrations of BP-3 in our subjects, geometric mean (GM) (95% CI) 0.97 (0.76–1.24) μg/g were dramatically lower compared to concentrations measured in US males from NHANES 2011–12 [16.1 (11.3–22.8) μg/g], and relatively lower compared to concentrations in male partners of US couples attempting pregnancy [4.13 (3.35–5.09) μg/g] (Buck Louis et al., 2015). It cannot be ruled out that this may account for the null results. We might speculate on the potential for a dose-response threshold, given what appear to be null results from studies with very low levels of BP-type UV filters and positive relationships from studies with higher levels of exposure. In a case-control study on Chinese men relatively lower unadjusted urinary concentrations of BP-3 [fertile controls GM:0.07 ng/ mL versus 1.3 ng/mL in our study population] were reported (Chen

5. Conclusions Our results add to previous studies that BP-type UV filters may act as endocrine disrupters. In men not selected because of testis function

Fig. 2. Adjusted percentage change (and 95% CI) in serum testosterone (T) and estradiol (E2) levels and T/E2 ratio using the 1st quartile of the urinary concentration of 2,4-dihydroxybenzophenone (BP-1) quartiles as the reference.

8

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

E. Adoamnei et al.

Fig. 3. Adjusted percentage change (and 95% CI) in serum inhibin b and FSH levels and inhibin b/FSH ratio using the 1st quartile of the urinary concentration of 2,4-dihydroxybenzophenone (BP-1) quartiles as the reference.

we detected that urinary concentrations of BP-type UV filters were associated with a modest alteration of serum reproductive hormones. Although the effects we report on reproductive function are small they should not be ignored. Nonetheless, we should be cautious about the interpretation of our findings, and whether these changes are sufficient to affect male reproductive function measured by serum reproductive hormones remain to established, as do primary mechanisms. Therefore, further research is warranted to confirm and expand these findings.

Buck Louis, G.M., Chen, Z., Kim, S., Sapra, K.J., Bae, J., Kannan, K., 2015. Urinary concentrations of benzophenone-type ultraviolet light filters and semen quality. Fertil. Steril. 104, 989–996. CDC (Centers for Disease Control and Prevention). 2015. Fourth National Report on Human Exposure to Environmental Chemicals. Updated Tables, February 2015. Atlanta, GA., CDC, National Center for Environmental Health; Division of Laboratory Sciences. https://www.cdc.gov/biomonitoring/pdf/fourthreport_updatedtables_ feb2015.pdf. (Accessed 15 Aug 2017). Calafat, A.M., Wong, L.Y., Ye, X., Reidy, J.A., Needham, L.L., 2008. Concentrations of the sunscreen agent benzophenone-3 in residents of the United States: national Health and Nutrition Examination Survey 2003–2004. Environ. Health Perspect. 116, 893–897. Chen, M., Tang, R., Fu, G., Xu, B., Zhu, P., Qiao, S., Chen, X., Xu, B., Qin, Y., Lu, C., Hang, B., Xia, Y., Wang, X., 2013. Association of exposure to phenols and idiopathic male infertility. J. Hazard. Mater. 250–251, 115–121. Chiu, Y.H., Edifor, R., Rosner, B.A., Nassan, F.L., Gaskins, A.J., Minguez-Alarcon, L., Williams, P.L., Tanrikut, C., Hauser, R., Chavarro, J.E., 2017. What does a single semen sample tell you?: Implications for male factor infertility research. Am. J. Epidemiol. 186, 918–926. Cutillas-Tolín, A., Mínguez-Alarcón, L., Mendiola, J., López-Espín, J.J., Jørgensen, N., Navarrete-Muñoz, E.M., Torres-Cantero, A.M., Chavarro, J.E., 2015. Mediterranean and western dietary patterns are related to markers of testicular function among healthy men. Hum. Reprod. 30, 2945–2955. Daston, G.P., Gettings, S.D., Carlton, B.D., Chudkowski, M., Davis, R.A., Kraus, A.L., Luke, C.F., Oellette, R.E., Re, T.A., Hoberman, A.M., 1993. Assessment of the reproductive toxic potential of dermally applied 2-hydroxy-4-methoxybenzophenone to male B6C3F1 mice. Fundam. Appl. Toxicol. 20, 120–124. Dewalque, L., Pirard, C., Vandepaer, S., Charlier, C., 2015. Temporal variability of urinary concentrations of phthalate metabolites, parabens and benzophenone-3 in a Belgian adult population. Environ. Res. 142, 414–423. Frederiksen, H., Jensen, T.K., Jørgensen, N., Boye Kyhl, H., Husby, S., Skakkebaek, N.E., Main, K.M., Juul, A., Andersson, A.M., 2014. Human urinary excretion ofnon-persistent environmental chemicals: an overview of Danish data collected 2006–2012. Reproduction 147, 555–565. French, J.E., 1992. NTP technical report on the toxicity studies of 2-hydroxy-4-methoxybenzophenone (CAS No. 131-57-7) adminstered topically and in dosed feed to F344/ N rats and B6C3F1 mice. Toxic. Rep. Ser. 21, 1–E14. Ghazipura, M., McGowan, R., Arslan, A., Hossain, T., 2017. Exposure to benzophenone-3 and reproductive toxicity: a systematic review of human and animal studies. Reprod. Toxicol. 73, 175–183. Hornung, R.W., Reed, L.D., 1990. Estimation of average concentration in the presence of non-detectable values. Appl. Occup. Environ. Hyg. 5, 46–51. Hsieh, M.H., Grantham, E.C., Liu, B., Macapagal, R., Willingham, E., Baskin, L.S., 2007. In utero exposure to benzophenone-2 causes hypospadias through an estrogen receptor dependent mechanism. J. Urol. 178, 1637–1642. Inn, K.G.W., McCurdy, D., Kuruvilla, L., Barss, N.M., Pietrzak, R., Kaplan, E., Inkret, W., Efurd, W., Rokop, D., Lewis, D., Gautier, P., Bell, R.T., 2001. Intercomparison study of inductively coupled plasma mass spectrometry, thermal ionization mass spectrometry and fission track analysis of mu Bq quantities of Pu-239 in synthetic urine. J. Radioanal. Nucl. Chem. 249, 121–131. Janjua, N.R., Mogensen, B., Andersson, A.M., Petersen, J.H., Henriksen, M., Skakkebaek, N.E., Wulf, H.C., 2004. Systemic absorption of the sunscreens benzophenone-3, octylmethoxycinnamate, and 3-(4-methyl-benzylidene) camphor after whole-body topical application and reproductive hormone levels in humans. J. Invest. Dermatol. 123, 57–61. Janjua, N.R., Kongshoj, B., Andersson, A.M., Wulf, H.C., 2008. Sunscreens in human plasma and urine after repeated whole-body topical application. J. Eur. Acad. Dermatol. Venereol. 22, 456–461. Jiang, R., Roberts, M.S., Collins, D.M., Benson, H.A., 1999. Absorption of sunscreens across human skin: an evaluation of commercial products for children and adults. Br. J. Clin. Pharmacol. 48, 635–637.

Conflicts of interest None. Funding statement This work was supported by Fundación Séneca, Agencia de Ciencia y Tecnología de la Región de Murcia [08808/PI/08, 19443/PI/14]; Consejería de Innovación, Junta de Andalucía [P09-CTS-5488] and Ministerio de Economía, Industria y Competitividad, Instituto de Salud Carlos III (Acción Estratégica en Salud, AES) [PI10/00985, PI13/01237, PI13/02406]. Ethics approval and consent to participate The Research Ethics Committee of the University of Murcia approved this study (no. 495/2010, approved May 14, 2010), and written informed consent was obtained from all subjects. Acknowledgments The authors gratefully acknowledge Drs Manuela Roca, Laura Sarabia-Cos, Guillermo Vivero-Salmerón and all the Quirónsalud Dexeus Murcia and Fertilidad Roca clinic staff for their assistance in data collection; and the young men of the study for their participation. References Asklund, C., Jørgensen, N., Skakkebaek, N.E., Jensen, T.K., 2007. Increased frequency of reproductive health problems among fathers of boys with hypospadias. Hum. Reprod. 22, 2639–2646. Barr, D.B., Wilder, L.C., Caudill, S.P., Gonzalez, A.J., Needham, L.L., Pirkle, J.L., 2005. Urinary creatinine concentrations in the U.S. population: implications for urinary biologic monitoring measurements. Environ. Health Perspect. 113, 192–200. Blüthgen, N., Zucchi, S., Fent, K., 2012. Effects of the UV filter benzophenone-3 (oxybenzone) at low concentrations in zebrafish (Danio rerio). Toxicol. Appl. Pharmacol. 263, 184–194. Buck Louis, G.M., Kannan, K., Sapra, K.J., Maisog, J., Sundaram, R., 2014. Urinary concentrations of benzophenone-type ultraviolet radiation filters and couples' fecundity. Am. J. Epidemiol. 180, 1168–1175.

9

International Journal of Hygiene and Environmental Health xxx (xxxx) xxx–xxx

E. Adoamnei et al.

Pillai, S., Oresajo, C., Hayward, J., 2005. Ultraviolet radiation and skin aging: roles of reactive oxygen species, inflammation and protease activation, and strategies for prevention of inflammation-induced matrix degradation – a review. Int. J. Cosmet. Sci. 27, 17–34. Pollack, A.Z., Buck Louis, G.M., Chen, Z., Sun, L., Trabert, B., Guo, Y., Kannan, K., 2015. Bisphenol A, benzophenone-type ultraviolet filters, and phthalates in relation to uterine leiomyoma. Environ. Res. 137, 101–107. Rehfeld, A., Dissing, S., Skakkebæk, N.E., 2016. Chemical UV filters mimic the effect of progesterone on Ca2+ signaling in human sperm cells. Endocrinology 157, 4297–4308. Schiffer, C., Müller, A., Egeberg, D.L., Alvarez, L., Brenker, C., Rehfeld, A., Frederiksen, H., Wäschle, B., Kaupp, U.B., Balbach, M., Wachten, D., Skakkebaek, N.E., Almstrup, K., Strünker, T., 2014. Direct action of endocrine disrupting chemicals on human sperm. EMBO Rep. 15, 758–765. Schisterman, E.F., Vexler, A., Whitcomb, B.W., Liu, A., 2006. The limitations due to exposure detection limits for regression models. Am. J. Epidemiol. 163, 374–383. Schlumpf, M., Cotton, B., Conscience, M., Haller, V., Steinmann, B., Lichtensteiger, W., 2001. In vitro and in vivo estrogenicity of UV screens. Environ. Health Perspect. 109, 239–244. Schlumpf, M., Schmid, P., Durrer, S., Conscience, M., Maerkel, K., Henseler, M., Gruetter, M., Herzog, I., Reolon, S., Ceccatelli, R., Faass, O., Stutz, E., Jarry, H., Wuttke, W., Lichtensteiger, W., 2004. Endocrine activity and developmental toxicity of cosmetic UV filters- an update. Toxicology 205, 113–122. Schlumpf, M., Kypke, K., Wittassek, M., Angerer, J., Mascher, H., Mascher, D., Vökt, C., Birchler, M., Lichtensteiger, W., 2010. Exposure patterns of UV filters, fragrances, parabens, phthalates, organochlor pesticides, PBDEs, and PCBs in human milk: correlation of UV filters with use of cosmetics. Chemosphere 81, 1171–1183. Schreurs, R.H., Sonneveld, E., Jansen, J.H., Seinen, W., van der Burg, B., 2005. Interaction of polycyclic musks and UV filters with the estrogen receptor (ER), androgen receptor (AR), and progesterone receptor (PR) in reporter gene bioassays. Toxicol. Sci. 83, 264–272. Scinicariello, F., Buser, M.C., 2016. Serum testosterone concentrations and urinary bisphenol A, benzophenone-3, triclosan, and paraben levels in male and female children and adolescents: NHANES 2011–2012. Environ. Health Perspect. 124, 1898–1904. Stokes-Riner, A., Thurston, S.W., Brazil, C., Guzick, D., Liu, F., Overstreet, J.W., Wang, C., Sparks, A., Redmon, J.B., Swan, S.H., 2007. One semen sample or 2? Insights from a study of fertile men. J. Androl. 28, 638–643. Suzuki, T., Kitamura, S., Khota, R., Sugihara, K., Fujimoto, N., Ohta, S., 2005. Estrogenic and antiandrogenic activities of 17 benzophenone derivatives used as UV stabilizers and sunscreens. Toxicol. Appl. Pharmacol. 203, 9–17. Vela-Soria, F., Ballesteros, O., Zafra-Gómez, A., Ballesteros, L., Navalón, A., 2014. UHPLCMS/MS method for the determination of bisphenol A and its chlorinated derivatives bisphenol S, parabens, and benzophenones in human urine samples. Anal. Bioanal. Chem. 406, 3773–3785. Vermeulen, A., Verdonck, G., 1992. Representativeness of a single point plasma testosterone level for the long term hormonal milieu in men. J. Clin. Endocrinol. Metab. 74, 939–942. Vermeulen, A., Verdonck, L., Kaufman, J.M., 1999. A critical evaluation of simple methods for the estimation of free testosterone in serum. J. Clin. Endocrinol. Metab. 84, 3666–3672. WHO (World Health Organization), 2010. WHO Laboratory Manual for the Examination and Processing of Human Semen, 5th ed. WHO Press, Geneva, Switzerland. Weisbrod, C.J., Kunz, P.Y., Zenker, A.K., Fent, K., 2007. Effects of the UV filter benzophenone-2 on reproduction in fish. Toxicol. Appl. Pharmacol. 225, 255–266. van den Driesche, S., Kilcoyne, K.R., Wagner, I., Rebourcet, D., Boyle, A., Mitchell, R., McKinnell, C., Macpherson, S., Donat, R., Shukla, C.J., Jorgensen, A., Meyts, E.R., Skakkebaek, N.E., Sharpe, R.M., 2017. Experimentally induced testicular dysgenesis syndrome originates in the masculinization programming window. JCI Insight 23 (2), e91204.

Jiménez-Díaz, I., Artacho-Cordón, F., Vela-Soria, F., Belhassen, H., Arrebola, J.P., Fernández, M.F., Ghali, R., Hedhili, A., Olea, N., 2016. Urinary levels of bisphenol A, benzophenones and parabens in Tunisian women: A pilot study. Sci. Total Environ. 562, 81–88. Kawamura, Y., Ogawa, Y., Nishimura, T., Kikuchi, Y., Nishkawa, J., Nishihara, T., Tanamoto, K., 2003. Estrogenic activities of UV stabilizers used in food contact plastics and benzophenone derivatives tested by the yeast two-hybrid assay. J. Health Sci. 49, 205–212. Kawamura, Y., Mutsuga, M., Kato, T., Iida, M., Tanamoto, K., 2005. Estrogenic and antiandrogenic activities of benzophenones in human estrogen and androgen receptor mediated mammalian reporter gene assays. J. Health Sci. 51, 48–54. Kim, S., Jung, D., Kho, Y., Choi, K., 2014. Effects of benzophenone-3 exposure on endocrine disruption and reproduction of Japanese medaka (Oryzias latipes)– a two generation exposure study. Aquat. Toxicol. 155, 244–252. Koch, H.M., Aylward, L.L., Hays, S.M., Smolders, R., Moos, R.K., Cocker, J., Jones, K., Warren, N., Levy, L., Bevan, R., 2014. Inter- and intra-individual variation in urinary biomarker concentrations over a 6-day sampling period. Part 2: personal care product ingredients. Toxicol. Lett. 231, 261–269. Krause, M., Frederiksen, H., Sundberg, K., Jørgensen, F.S., Jensen, L.N., Nørgaard, P., Jørgensen, C., Ertberg, P., Holm Petersen, J., Feldt-Rasmussen, U., Juul, A., Drzewiecki, K., Skakkebæk, N.E., Andersson, A.M., 2018. Maternal exposure to UV filters and associations to maternal thyroid hormones and IGF-I/IGFBP3 and birth outcomes. Endocr. Connect pii: EC-17-0375. Kunisue, T., Chen, Z., Buck Louis, G.M., Sundaram, R., Hediger, M.L., Sun, L., Kannan, K., 2012. Urinary concentrations of benzophenone-type UV filters in U: S. women and their association with endometriosis. Environ. Sci. Technol. 46, 4624–4632. Lassen, T.H., Frederiksen, H., Jensen, T.K., Petersen, J.H., Main, K.M., Skakkebæk, N.E., Jørgensen, N., Kranich, S.K., Andersson, A.M., 2013. Temporal variability in urinary excretion of bisphenol A and seven other phenols in spot morning, and 24-h urine samples. Environ. Res. 126, 164–170. Mínguez-Alarcón, L., Chavarro, J.E., Mendiola, J., Roca, M., Tanrikut, C., Vioque, J., Jørgensen, N., Torres-Cantero, A.M., 2017. Fatty acid intake in relation to reproductive hormones and testicular volume among young healthy men. Asian J. Androl. 19, 184–190. Ma, R., Cotton, B., Lichtensteiger, W., Shlumpf, M., 2003. UV filters with antagonistic action at androgen receptors in the MDA-kb2 cell transcriptional-activity assay. Toxicol. Sci. 74, 43–50. Meeker, J.D., Cantonwine, D.E., Rivera-González, L.O., Ferguson, K.K., Mukherjee, B., Calafat, A.M., Ye, X., Anzalota Del Toro, L.V., Crespo-Hernández, N., Jiménez-Vélez, B., Alshawabkeh, A.N., Cordero, J.F., 2013. Distribution, variability, and predictors of urinary concentrations of phenols and parabens among pregnant women in Puerto Rico. Environ. Sci. Technol. 47, 3439–3447. Mendiola, J., Jørgensen, N., Mínguez-Alarcón, L., Sarabia-Cos, L., López-Espín, J.J., Vivero-Salmerón, G., Ruiz-Ruiz, K.J., Fernández, M.F., Olea, N., Swan, S.H., TorresCantero, A.M., 2013. Sperm counts may have declined in young university students in Southern Spain. Andrology 1, 408–413. Menkveld, R., Stander, F.S., Kotze, T.J., Kruger, T.F., van Zyl, J.A., 1990. The evaluation of morphological characteristics of human spermatozoa according to stricter criteria. Hum. Reprod. 5, 586–592. Moos, R.K., Angerer, J., Wittsiepe, J., Wilhelm, M., Brüning, T., Koch, H.M., 2014. Rapid determination of nine parabens and seven other environmental phenols in urine samples of German children and adults. Int. J. Hyg. Environ. Health 217, 845–853. Muncke, J., 2011. Endocrine disrupting chemicals and other substances of concern in food contact materials: an updated review of exposure, effect and risk assessment. J. Steroid. Biochem. Mol. Biol. 127, 118–127. Nakagawa, Y., Tayama, K., 2011. Estrogenic potency of benzophenone and its metabolites in juvenile female rats. Arch. Toxicol. 75, 74–79. Philippat, C., Bennett, D., Calafat, A.M., Picciotto, I.H., 2015. Exposure to select phthalates and phenols through use of personal care products among Californian adults and their children. Environ. Res. 140, 369–376.

10