Association of sulfur dioxide exposure with sperm quality in reproductive-aged men: the mediating and moderating roles of essential elements in seminal plasma

Abstract

Currently, there are limited data to support the association between sulfur dioxide (SO₂) exposure and semen quality. Here, we analyzed nine elements including vanadium (V), chromium (Cr), iron (Fe), copper (Cu), cobalt (Co), nickel (Ni), manganese (Mn), selenium (Se), and strontium (Sr) in seminal plasma of 738 participants using inductively coupled plasma mass spectrometry (ICP–MS). Each participant was assigned an average SO₂ exposure of 0–90 days preceding semen collection. Associations were determined using multivariable linear models, susceptible exposure windows were identified via distributed lag nonlinear models (DLNM), and indirect effects and effect modification were examined based on mediation and interaction analyses. Notably, SO₂ exposure was negatively associated with sperm concentration (β: −0.07; 95% confidence interval [CI]: −0.13 to −0.01; P = 0.02), total sperm count (β: −0.09; 95% CI: −0.15 to −0.03; P < 0.01), progressive motility (β: −0.04; 95% CI: −0.07 to −0.01; P = 0.02), and total motility (β: −0.03; 95% CI: −0.07 to −0.01; P < 0.01), but positively associated with sperm abnormal morphology (β: 0.27; 95% CI: 0.12 to 0.42; P < 0.01). DLNM identified a susceptibility window during spermiogenesis (lags 37–69 days). Se was positively associated with sperm concentration (β: 0.34; 95% CI: 0.23 to 0.45; P < 0.01) and total sperm count (β: 0.28, 95% CI: 0.16 to 0.40; P < 0.01), mediating 23.6% and 14.6% of the effects of SO₂, respectively. A significant interaction between SO₂ and Ni was detected in sperm abnormal morphology (β: 0.28; 95% CI: 0.05 to 0.51; P = 0.02). These results indicate that SO₂ exposure may reduce semen quality, and this effect may be mediated by Se and modified by Ni.

INTRODUCTION

Ambient air pollution, which includes particulate matter (PM_2.5 and PM₁₀), carbon monoxide (CO), sulfur dioxide (SO₂), nitrogen oxides (NO₂), and ozone (O₃), is recognized to exert detrimental effects on human health.1,2 Epidemiological studies have reported that ambient air pollution can induce respiratory, cardiovascular, and reproductive diseases, potentially increasing the risk of mortality. Hwang et al.3 found that a 24-h geometric mean PM_2.5 exposure level of 14.12 μg m⁻³, though below the 35 μg m⁻³ standard limit of United States Environmental Protection Agency (USA EPA), can increase health risks in vulnerable populations. A cohort study comprising 2.1 million Canadian adults demonstrated that even at relatively low exposure levels, with a mean PM_2.5 concentration of 8.7 μg m⁻³, each 10 μg m⁻³ increase in PM_2.5 contributed to a 49% increase in the risk of diabetes-related mortality (hazard ratio [HR]: 1.49, 95% confidence interval [CI]: 1.37 to 1.62). Moreover, this elevated risk persisted even at concentrations below the World Health Organization (WHO) guideline of 5 μg m⁻³.4 A longitudinal study by Ma et al.5 showed that SO₂ levels below 40 μg m⁻³ still exhibited a negative association with sperm quality.

Male infertility affects nearly 15% of couples worldwide and is largely driven by poor semen quality.6,7 There has been a global decline in sperm concentration and progressive motility, with data from 1973 to 2018 showing that sperm quality has decreased by more than 62%.8 However, the factors driving this reduction in sperm quality remain unclear. Researchers have found that exposure to ambient air pollution may be associated with poor sperm quality.9

In our previous study, we found that SO₂, a major atmospheric pollutant, adversely affected sperm quality in Anhui Province even at concentrations lower than the WHO-recommended 24-h average limit of 40 μg m⁻³.5 Findings from a meta-analysis also showed a negative association between SO₂ exposure and sperm concentration.10 In addition, exposure to SO₂ is negatively associated with sperm count and concentration.11 However, there are inconsistencies in the association between SO₂ exposure and sperm quality across studies. For instance, Huang et al.12 found that SO₂ exposure may have no significant effect on sperm concentration or sperm count. Zhou et al.13 found no pronounced association between SO₂ exposure and sperm concentration or motility.

Essential elements have been reported to be associated with sperm quality.14 The main routes of exposure to essential elements include air inhalation, food consumption, and water intake.15 Essential elements are crucial for various physiological functions and are the important biochemical components of semen. Human semen contains several essential elements, including copper (Cu), zinc (Zn), calcium (Ca), magnesium (Mg), manganese (Mn), and selenium (Se), which play critical roles in spermatogenesis, sperm maturation, and capacitation.16 Any deficiency or excess of specific essential elements can compromise male fertility or even induce infertility.17 Numerous population-based studies have reported strong associations between essential element levels in blood, urine, or seminal plasma and sperm quality. Liu et al.18 reported that Cu, Se, nickel (Ni), iron (Fe), and other trace elements may affect sperm function and quality. However, these associations have varied across population studies.

Recently, Hassan et al.19 showed that ambient air pollution levels might affect metal levels in the body. This finding reinforces the biological link between air pollution exposure and altered body metal levels.

In this study, we hypothesized that the association between ambient SO₂ exposure and sperm quality might be mediated and moderated by essential element concentrations in seminal plasma. To confirm this hypothesis, we collected 738 seminal plasma samples and their corresponding residential locations. Initially, multiple linear regression models were adopted to determine the associations between the 0–90-day (human sperm development takes approximately 90 days, including spermatogenesis and spermiogenesis) average SO₂ concentrations20 and sperm quality parameters. Next, we tested the day-specific effects of SO₂ exposure during the entire spermatogenesis period on sperm quality using a distributed lag nonlinear model (DLNM), which captures the critical exposure window of SO₂ exposure on sperm quality. Moreover, the mediating and moderating roles of essential elements in the relationship between SO₂ exposure and sperm quality parameters were explored. Finally, a two-pollutant model was developed to evaluate the robustness of our findings.

To our knowledge, limited studies have examined the association between SO₂ exposure during the entire spermatogenesis period (0–90 days) and sperm quality, as well as the impact of essential element concentrations in seminal plasma on this association. Our findings provide evidence that SO₂ exposure impairs sperm quality and reveal how essential elements in seminal plasma mediate and moderate this association.

PARTICIPANTS AND METHODS

Participants

A retrospective cohort study involving 738 male participants was conducted at the Department of Obstetrics and Gynecology, The First Affiliated Hospital of Anhui Medical University (Hefei, China) from March 2, 2021, to November 23, 2022. All male participants were assessed based on semen analyses following the 5^th edition of the WHO Laboratory Manual for the Examination and Processing of Human Semen.21 Participants were excluded if they had genetic abnormalities (e.g., chromosomal anomalies or gene mutations), metabolic disorders (e.g., thyroid dysfunction, diabetes, or adrenal disorders), a history of sexually transmitted infections, occupational exposure to pesticides, heavy metals, or industrial chemicals with known reproductive toxicity, cryptorchidism, orchitis, testicular torsion, mumps, or long-term consumption of unrefined cottonseed oil. Demographic and clinical information of all eligible participants was collected from the hospital’s electronic medical records. This study was approved by the Ethics and Research Committee of Anhui Medical University (Approval No. 20160270). Written informed consent was obtained from all the participants prior to enrollment and follow-up. This study was performed in accordance with the Declaration of Helsinki for human medical research.

SO₂ exposure assessment

Daily SO₂, NO₂, O₃, CO, PM_2.5, and PM₁₀ data with a spatial resolution of 1 km × 1 km were collected from the China High Air Pollutants (CHAP) database, which provides full-coverage and high-quality datasets on ground-level air pollutants (AP). Considering the latitude and longitude of each participant’s residence, the daily average SO₂ exposure for the 90 days preceding semen sample collection was assigned to each participant. SO₂ exposure during the 0–90 days before semen donation, which represents the entire human spermatogenesis period and includes four key stages: days 0–9 (epididymal storage), days 10–14 (development of sperm motility), days 15–69 (spermatogenesis stage II), and days 70–90 (spermatogenesis stage I), was determined for each participant and then analyzed in relation to sperm parameters.22

Covariates

Potential covariates included age at semen collection, education level, smoking status, body mass index (BMI), month of semen sample collection, relative humidity (RH), temperature, and other air pollutants (NO₂, O₃, CO, PM_2.5, and PM₁₀). Temperature and RH data were obtained from the fifth-generation European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis dataset (ERA5).23

Sample collection and detection of essential elements

Seminal plasma samples donated by participants were collected during sperm routine analysis and stored at −80°C until further analysis of essential element concentrations. The concentrations of V, Cr, Fe, Cu, Co, Ni, Mn, Se, and Sr in seminal plasma were measured using inductively coupled plasma mass spectrometry (ICP–MS; NexION 350X; PerkinElmer, Shelton, CT, USA). The detailed experimental procedure is described below: seminal plasma (200 μl) was mixed with 200 μl of nitric acid (HNO₃) and digested at 105°C for 2 h. The sample was diluted to a final volume of 5 ml with a solution containing 0.05% Triton X-100 and 1% HNO₃. Instrument ionization drift was corrected using an internal standard solution. Next, a mixed-element standard stock solution (10 μg ml⁻¹; Catalog No. N9300233; PerkinElmer) and a mixed-internal element standard solution (10 μg ml⁻¹; Catalog No. N9303832; PerkinElmer) were prepared, and the samples were analyzed using a previously described method.24

Statistical analyses

The demographic characteristics of the male participants were expressed as mean ± standard deviation (s.d.) for continuous variables and as percentages for categorical variables. The concentrations of the nine essential elements were expressed using the geometric mean (GM) and five selected percentiles (5^th, 25^th, 50^th, 75^th, and 95^th). The concentrations of the six outdoor air pollutants, temperature, and relative humidity were presented as mean ± s.d. and median (interquartile range [IQR]). Sperm quality parameters, including sperm concentration (×10⁶ ml⁻¹), total sperm count (×10⁶), progressive motility rate (PR; %), total motility (%), and sperm abnormal morphology, were expressed as median (IQR).

Multivariate linear regression models were used to explore the associations between average SO₂ exposure during 0–90 days, essential element concentrations in seminal plasma, and sperm parameters.

To further investigate the day-specific effects of SO₂ exposure on sperm quality, we utilized a DLNM using a cross-basis function to simultaneously estimate the exposure–response and lag–response relationships. Specifically, a generalized linear model (GLM) and DLNM were applied to clarify the association between SO₂ exposure and sperm quality. In this model, the exposure–response and lag–response relationships were simultaneously modeled. The following GLM with DLNM was employed: Log (Yt) = α + Cb (SO₂, lag) + natural spline (ns; temperature) + ns (RH) + other covariates, where t was the day of investigation; Yt represents the sperm quality parameters on the day t; Log (Yt) follows a normal distribution, and GLM with a Gaussian link was used; α was the intercept; Cb (SO₂, lag) was a cross-basis matrix, in which a nonlinear exposure–response function was applied to determine the lag–response function of SO₂ exposure; and ns denotes a natural spline function used to model the nonlinear effects of temperature and relative humidity. This function was modeled using a natural cubic spline function with three degrees of freedom and lags 0–90 as the exposure variable. Meteorological factors, including daily mean temperature and relative humidity, were included in the model to control for potential confounding effects. The model also included covariates such as age at semen collection, BMI, education level, smoking status (yes or no), and month of semen sample collection. The lag period was set to 0–90 days with reference to the entire human spermatogenesis period.

The mechanism by which SO₂ exposure (independent variable) affects sperm parameters (dependent variable) was determined through mediation analysis, with essential element concentrations in seminal plasma as the mediators. This approach breaks down the total effect (TE) of exposure into direct and indirect components, revealing the pathways through which exposure alters sperm quality. TE represents the sum of the direct and indirect effects of SO₂ exposure on sperm parameters. The direct effect (DE) was defined as the impact of SO₂ exposure on sperm parameters via pathways other than the mediators. If the TE, DE, and indirect effects operated in the same direction, the “proportion mediated” was calculated to determine how much of the total effect was explained by the mediators. Mediation analysis can reveal complex causal relationships, distinguish between direct and indirect effects, and quantify the role of mediators in environmental exposure.25

To clarify the interaction effects between SO₂ exposure and essential trace elements on sperm quality, we performed multiplicative interaction analyses. In this test, exposure to SO₂ was calculated as the average concentration during the 90 days prior to semen collection. Interaction terms between SO₂ and each essential trace element were included in multivariable linear regression models to determine whether the association between SO₂ exposure and sperm quality parameters could differ according to the element levels. Significant interactions were defined as those with P for interaction < 0.05, and further stratified analyses were conducted to explore these modifying effects.

In the developed model, we adjusted for age at semen collection, BMI, education level, smoking status (yes or no), month of semen sample collection, RH, and temperature.26

To further determine the robustness of the association between average SO₂ exposure during 0–90 days and sperm quality, sensitivity analyses were performed using a two-pollutant model to control for potential confounding by NO₂, O₃, CO, PM_2.5, and PM₁₀.27

All statistical analyses were conducted using R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria) and SPSS (IBM Corp., Armonk, NY, USA). P < 0.05 was considered statistically significant.

RESULTS

Characteristics of male participants

Overall, 738 male participants were enrolled in this study. The basic demographic characteristics of the participants are presented in Table 1. The age (mean ± s.d.) was 31.7 ± 3.8 years and the BMI (mean ± s.d.) was 25.0 ± 3.6 kg m⁻². Of the participants, 35.3% had a university-level education or higher, 67.1% were non-smokers, and 60.7% had secondary infertility. The median period of abstinence was 4 (IQR: 3–5) days. The median values for semen volume, PR, total motility, sperm concentration, total sperm count, and sperm abnormal morphology were 3.9 ml, 44.6%, 68.4%, 62.5×10⁶ ml⁻¹, 234.2×10⁶, and 95.1%, respectively.

Table 1.

The characteristics of included study population (total=738)

Characteristic	Value
Age (year), mean±s.d.	31.7±3.8
BMI (kg m−2), mean±s.d.	25.0±3.6
Smoke, n (%)
No	495 (67.1)
Yes	243 (32.9)
Education, n (%)
High school or below (≤12 years)	321 (43.5)
Junior college (13–15 years)	156 (21.1)
University or above (≥16 years)	261 (35.4)
Sampling month, n (%)
January	28 (3.8)
February	28 (3.8)
March	143 (19.4)
April	98 (13.3)
May	40 (5.4)
June	105 (14.2)
July	5 (0.7)
August	121 (16.4)
September	97 (13.1)
October	57 (7.7)
November	11 (1.5)
December	5 (0.7)
Infertile type, n (%)
Primary infertility	290 (39.3)
Secondary infertility	448 (60.7)
Sperm parameters, median (IQR)
Abstinence period (day)	4 (3–5)
Semen volume (ml)	3.9 (2.9–4.9)
Progressive motility (%)	44.6 (32.8–57.1)
Total motility (%)	68.4 (52.7–80.1)
Sperm concentration (×106 ml−1)	62.5 (34.1–104.0)
Total sperm count (×106)	234.2 (126.0–393.1)
Sperm abnormal morphology (%)	95.1 (94.0–97.5)

IQR: interquartile range; s.d.: standard deviation; BMI: body mass index

Assessment of ambient SO₂ exposure and concentrations of nine essential elements in seminal plasma

The data shown in Supplementary Table 1 reflect the distribution of air pollutant concentrations, temperature, and relative humidity for 0–90 days before semen sample collection. In the 90-day exposure window, the SO₂ concentration (mean ± s.d.) was 8.50 ± 1.09 μg m⁻³, the median value was 8.49 (IQR: 7.95–8.97) μg m⁻³, and the overall range across the study population was 6.00 μg m⁻³ to 20.72 μg m⁻³. The median concentrations of V, Cr, Fe, Cu, Co, Ni, Mn, Se, and Sr in seminal plasma were 1.19 μg l⁻¹, 4.55 μg l⁻¹, 379.04 μg l⁻¹, 131.78 μg l⁻¹, 0.28 μg l⁻¹, 1.33 μg l⁻¹, 10.08 μg l⁻¹, 93.02 μg l⁻¹, and 106.10 μg l⁻¹, respectively. Moreover, the limits of detection (LODs) for the nine essential elements ranged from 0 to 0.24 μg l⁻¹. The spiked recoveries for these essential elements ranged from 95.7% to 118.3% (Supplementary Table 2).

Supplementary Table 1.

Distributions of average ambient temperature, relative humidity, and six air pollutants’ exposure during 0–90 days before semen test of the study population

Variables	Mean±s.d.	Median (IQR)	Range
Temperature (°C)	12.28±11.19	14.01 (1.09–23.56)	−13.05–27.52
Relative humidity (%)	86.6±3.6	85.4 (83.7–89.6)	79.4–96.9
PM2.5 (µg m−3)	34.71±16.11	28.66 (19.91–48.83)	11.19–74.98
PM10 (µg m−3)	63.66±24.23	59.65 (40.87–85.36)	24.44–123.56
SO2 (µg m−3)	8.50±1.09	8.49 (7.95–8.97)	6.00–20.72
NO2 (µg m−3)	26.62±11.36	24.08 (17.17–35.05)	833–61.65
CO (mg m−3)	0.69±0.11	0.66 (0.61–0.78)	0.36–1.05
O3 (µg m−3)	110.19±26.13	117.81 (85.23–132.53)	55.88–158.15

IQR: interquartile range; s.d.: standard deviation; PM: particulate matter; SO₂: sulfur dioxide; NO₂: nitrogen oxides; CO: carbon monoxide; O₃: ozone

Supplementary Table 2.

The distributions of nine metals in the seminal plasma of all male participants

Metals	Detection rates (%)	LODs	Spike recovery rates (%)	Geometric mean	Percentile
					P 5	P 25	P 50	P 75	P 95
V (mg/L)	100	0.00	102.1	1.24	0.66	0.92	1.19	1.60	2.76
Cr (mg/L)	100	0.01	112.1	5.00	2.49	3.41	4.55	6.88	12.44
Mn (mg/L)	100	0.00	95.7	10.20	5.16	7.76	10.08	13.08	20.53
Fe (mg/L)	100	0.07	97.1	407.06	199.66	295.80	379.04	521.86	1094.31
Co (mg/L)	100	0.00	100.1	0.28	0.14	0.21	0.28	0.36	0.56
Ni (mg/L)	100	0.01	118.3	1.41	0.68	1.02	1.33	1.87	3.29
Cu (mg/L)	100	0.01	111.9	139.46	59.39	93.67	131.78	187.34	413.28
Se (mg/L)	100	0.24	106.0	92.70	39.20	66.77	93.02	125.89	238.80
Sr (mg/L)	100	0.00	113.4	109.43	51.75	79.95	106.10	139.81	274.64

The concentrations of nine essential trace elements in seminal plasma are presented as the 5^th percentile (P₅), 25^th percentile (P₂₅), 50^th percentile (P₅₀), 75^th percentile (P₇₅), and 95^th percentile (P₉₅), respectively. Detection rates refer to the proportion of samples with concentrations above the LOD. LOD: limit of detection; Cr: chromium; Mn: manganese; V: vanadium; Se: selenium; Cu: copper; Co: cobalt; Ni: nickel; Sr: strontium; Fe: iron

Association between SO₂ exposure and sperm quality parameters using linear regression and DLNM model

Linear regression analysis results for the association between SO₂ exposure during 0–90 days and sperm quality parameters are shown in Figure 1 and Supplementary Table 3. SO₂ exposure was negatively associated with total sperm count (β: −0.09; 95% CI: −0.15 to −0.03; P < 0.01), sperm concentration (β: −0.07; 95% CI: −0.13 to −0.01; P = 0.02), PR (β: −0.04; 95% CI: −0.07 to −0.01; P = 0.02), and total motility (β: −0.03; 95% CI: −0.07 to −0.01; P < 0.01) but positively associated with sperm abnormal morphology (β: 0.27; 95% CI: 0.12 to 0.42; P < 0.01). Moreover, DLNM results indicate that SO₂ had a significant lagged effect on sperm quality parameters (Figure 2). Specifically, exposure to SO₂ at lags 47–69 was negatively associated with total sperm count; exposure to SO₂ at lags 37–58 was negatively associated with sperm concentration; exposure to SO₂ at lags 41–58 was negatively associated with progressive motility (PR); exposure to SO₂ at lags 40–62 was negatively associated with total motility; and exposure to SO₂ at lags 44–54 was positively associated with sperm abnormal morphology.

Figure 1.

Association of SO₂ exposure during 0–90 days and sperm quality using multivariate linear regression models. ^aThese models were adjusted for age, BMI, education level, smoking status, month of semen sample collection, RH, and temperature. SO₂: sulfur dioxide; CI: confidence interval; BMI: body mass index; RH: relative humidity.

Supplementary Table 3.

The associations between sulfur dioxide exposure during entire spermatogenesis (0–90 days) and sperm quality parameters

Air pollutants	Lags (days)	Outcomes	Adjusted β (95% CI) a	P
SO2	0–90	Total sperm count	−0.09 (−0.15 to−0.03)	<0.01
SO2	0–90	Sperm concentration	−0.07 (−0.13 to−0.01)	0.02
SO2	0–90	progressive motility	−0.04 (−0.07 to−0.01)	0.02
SO2	0–90	Total motility	−0.03 (−0.07 to−0.01)	<0.01
SO2	0–90	Sperm abnormal morphology	0.27 (0.12 to 0.42)	<0.01

^aModel adjusted for age, BMI, smoking status, educational level, month of semen collection, average temperature, and relative humidity during the 0–90 days. CI: confidence interval; SO₂: sulfur dioxide; BMI: body mass index

Figure 2.

The association between average SO₂ exposure during 0–90 days and sperm quality using DLNM. These models were adjusted for age, BMI, education level, smoking status, month of semen sample collection, RH, and temperature. Lag–response curve showing the associations between SO₂ exposure and (a) sperm concentration, (b) total sperm count, (c) progressive motility, (d) total motility, and (e) sperm abnormal morphology rate. The curves present the estimated adjusted β values and 95% CI at lags 0–90. SO₂: sulfur dioxide; CI: confidence interval; DLNM: distributed lag nonlinear model; BMI: body mass index; RH: relative humidity.

The associations between SO₂ exposure and nine essential elements using linear regression model

The results for the association between SO₂ exposure during 0–90 days and nine essential element concentrations in seminal plasma are shown in Figure 3 and Supplementary Table 4. Notably, SO₂ exposure was positively associated with V (β: 0.07; 95% CI: 0.04 to 0.09; P < 0.01) and Ni (β: 0.04; 95% CI: 0.00 to 0.08; P = 0.04) concentrations but negatively associated with Fe (β: −0.04; 95% CI: −0.08 to 0; P = 0.04), Se (β: −0.06; 95% CI: −0.10 to −0.02; P < 0.01), and Sr (β: −0.06; 95% CI: −0.10 to −0.03; P < 0.01) concentrations in seminal plasma. Further linear regression analyses demonstrated that Mn, Cu, Se, and Fe were positively associated with sperm concentration (β_Mn: 0.19, 95% CI: 0.06 to 0.32, P < 0.01; β_Cu: 0.21, 95% CI: 0.10 to 0.31, P < 0.01; β_Se: 0.34, 95% CI: 0.23 to 0.45, P < 0.01; β_Fe: 0.38, 95% CI: 0.26 to 0.49, P < 0.01) and total sperm count (β_Mn: 0.15, 95% CI: 0.02 to 0.29, P = 0.03; β_Cu: 0.15, 95% CI: 0.04 to 0.26, P < 0.01; β_Se: 0.28, 95% CI: 0.16 to 0.40, P < 0.01; β_Fe: 0.30, 95% CI: 0.17 to 0.42, P < 0.01), whereas Co and Sr were positively associated with sperm concentration (β_Co: 0.18, 95% CI: 0.06 to 0.31, P < 0.01; β_Sr: 0.13, 95% CI: 0 to 0.26, P = 0.05), as shown in Figure 4 and Supplementary Table 5.

Figure 3.

The association between SO₂ exposure during 0–90 days with nine essential trace elements concentrations in seminal plasma using multivariate linear regression models. ^aThese models were adjusted for age, BMI, education level, smoking status, month of semen sample collection, RH, and temperature. SO₂: sulfur dioxide; CI: confidence interval; BMI: body mass index; RH: relative humidity.

Supplementary Table 4.

Regression coefficients and percentage changes (95% confidence interval) in essential elements associated with average exposure to sulfur dioxide during 0–90 days prior to semen testing based on multiple linear regression

Air pollutants	Lags (days)	Outcomes	Adjusted β (95% CI) a	P
SO2	0–90	V	0.07 (0.04 to 0.09)	<0.01
SO2	0–90	Cr	0.02 (−0.01 to 0.06)	0.22
SO2	0–90	Mn	−0.03 (−0.06 to 0.01)	0.11
SO2	0–90	Fe	−0.04 (−0.08 to 0.00)	0.04
SO2	0–90	Co	−0.02 (−0.05 to 0.02)	0.35
SO2	0–90	Ni	0.04 (0.00 to 0.08)	0.04
SO2	0–90	Cu	−0.04 (−0.09 to 0.00)	0.06
SO2	0–90	Se	−0.06 (−0.10 to−0.02)	<0.01
SO2	0–90	Sr	−0.06 (−0.10 to−0.03)	<0.01

^aModel adjusted for age, BMI, smoking status, educational level, month of semen collection, average temperature, and relative humidity during the 0–90 days. SO₂: sulfur dioxide; CI: confidence interval; BMI: body mass index; Cr: chromium; Mn: manganese; V: vanadium; Se: selenium; Cu: copper; Co: cobalt; Ni: nickel; Sr: strontium; Fe: iron

Figure 4.

The association between essential trace elements concentrations in seminal plasma and sperm quality using multivariate linear regression models. ^aThese models were adjusted for age, BMI, education level, and smoking status. CI: confidence interval; BMI: body mass index.

Supplementary Table 5.

The association between the levels of nine metals in seminal plasma and sperm quality parameters analyzed using multiple linear regression

Metals	Outcomes	Adjusted β (95% CI) a	P
V	Total sperm count	−0.08 (−0.24 to 0.07)	0.29
	Sperm concentration	0.01 (−0.14 to 0.16)	0.93
	progressive motility	0.02 (−0.06 to 0.10)	0.62
	Total motility	0.03 (−0.03 to 0.09)	0.37
	Sperm abnormal morphology	−0.01 (−0.38 to 0.36)	0.95
Cr	Total sperm count	−0.06 (−0.19 to 0.06)	0.34
	Sperm concentration	−0.03 (−0.15 to 0.09)	0.58
	progressive motility	−0.06 (−0.12 to 0.01)	0.10
	Total motility	−0.04 (−0.09 to 0.00)	0.07
	Sperm abnormal morphology	0.29 (−0.01 to 0.59)	0.06
Mn	Total sperm count	0.15 (0.02 to 0.29)	0.03
	Sperm concentration	0.19 (0.06 to 0.32)	<0.01
	progressive motility	0.00 (−0.07 to 0.07)	0.94
	Total motility	0.02 (−0.03 to 0.08)	0.40
	Sperm abnormal morphology	−0.16 (−0.49 to 0.17)	0.35
Fe	Total sperm count	0.30 (0.17 to 0.42)	<0.01
	Sperm concentration	0.38 (0.26 to 0.49)	<0.01
	progressive motility	−0.07 (−0.13 to 0.00)	0.05
	Total motility	−0.02 (−0.07 to 0.03)	0.36
	Sperm abnormal morphology	−0.17 (−0.50 to 0.14)	0.28
Co	Total sperm count	0.10 (−0.04 to 0.23)	0.17
	Sperm concentration	0.18 (0.06 to 0.31)	<0.01
	progressive motility	−0.01 (−0.07 to 0.07)	0.89
	Total motility	0.01 (−0.05 to 0.06)	0.81
	Sperm abnormal morphology	−0.09 (−0.42 to 0.23)	0.57
Ni	Total sperm count	−0.04 (−0.16 to 0.09)	0.56
	Sperm concentration	0.11 (−0.01 to 0.23)	0.08
	progressive motility	−0.04 (−0.11 to 0.02)	0.19
	Total motility	−0.02 (−0.07 to 0.03)	0.47
	Sperm abnormal morphology	0.27 (−0.03 to 0.57)	0.08
Cu	Total sperm count	0.15 (0.04 to 0.26)	<0.01
	Sperm concentration	0.21 (0.10 to 0.31)	<0.01
	progressive motility	−0.05 (−0.10 to 0.01)	0.12
	Total motility	−0.02 (−0.06 to 0.03)	0.47
	Sperm abnormal morphology	−0.32 (−0.59 to−0.06)	0.02
Se	Total sperm count	0.28 (0.16 to 0.40)	<0.01
	Sperm concentration	0.34 (0.23 to 0.45)	<0.01
	progressive motility	−0.03 (−0.09 to 0.03)	0.31
	Total motility	0.00 (−0.05 to 0.05)	1.00
	Sperm abnormal morphology	−0.23 (−0.52 to 0.06)	0.12
Sr	Total sperm count	0.13 (0.00 to 0.27)	0.06
	Sperm concentration	0.13 (0.00 to 0.26)	0.05
	progressive motility	−0.05 (−0.12 to 0.02)	0.18
	Total motility	−0.02 (−0.07 to 0.03)	0.44
	Sperm abnormal morphology	−0.06 (−0.38 to 0.27)	0.73

^aModel adjusted for age, BMI, smoking status, education level. BMI: body mass index; CI: confidence interval; Cr: chromium; Mn: manganese; V: vanadium; Se: selenium; Cu: copper; Co: cobalt; Ni: nickel; Sr: strontium; Fe: iron

Mediation analyses

Considering that a mediator correlates with both the exposure and the outcome, a mediation analysis was performed to determine whether the concentrations of essential elements could mediate the association between SO₂ exposure during 0–90 days and sperm quality. Initially, we assessed the correlations between average SO₂ exposure during 0–90 days, the concentrations of essential elements in seminal plasma, and sperm quality parameters. Results indicated that SO₂ exposure was significantly and negatively associated with sperm concentration, total sperm count, progressive motility, and total motility. Furthermore, a positive association was observed between SO₂ exposure and sperm abnormal morphology. During the 0–90-day window, SO₂ exposure exhibited a significant positive association with V and Ni concentrations in seminal plasma. In contrast, it was significantly negatively associated with Se, Fe, and Sr concentrations. Moreover, positive associations were observed between the concentrations of Mn, Fe, Cu, and Se in seminal plasma and total sperm count. Similarly, the concentrations of Mn, Fe, Cu, and Se in seminal plasma were positively associated with sperm concentration. The concentration of Cu in seminal plasma was negatively associated with sperm abnormal morphology. With reference to these findings, we conducted mediation analyses to investigate the mediating role of Fe and Se concentrations in the association between average SO₂ exposure during 0–90 days and sperm concentration and total sperm count. The mediating role of Sr concentration in the association between average SO₂ exposure during 0–90 days and total sperm count was also investigated. The results indicated that Se concentrations in seminal plasma mediated the association between average SO₂ exposure during 0–90 days and total sperm count (mediation proportion: 14.6%) as well as the association between average SO₂ exposure during 0–90 days and sperm concentration (mediation proportion: 23.6%), as shown in Figure 5. With respect to Fe and Sr, no mediation effect was detected for the associations between average SO₂ exposure during 0–90 days and total sperm count or sperm concentration.

Figure 5.

Mediation analysis of the associations between SO₂ exposure and semen quality parameters by essential element concentrations in seminal plasma. (a) Se mediating the association between SO₂ exposure and sperm concentration. (b) Se mediating the association between SO₂ exposure and total sperm count. (c) Sr mediating the association between SO₂ exposure and sperm concentration. (d) Fe mediating the association between SO₂ exposure and sperm concentration. (e) Fe mediating the association between SO₂ exposure and total sperm count. Se: selenium; Sr: strontium; Fe: iron; SO₂: sulfur dioxide; CI: confidence interval.

Interaction analyses

Among all tested interactions, a significant interaction (P = 0.02) was observed only between ambient SO₂ exposure and Ni levels in seminal plasma in relation to sperm abnormal morphology (Supplementary Table 6). The interaction term between seminal plasma Ni and SO₂ was positively associated with sperm abnormal morphology (β: 0.28; 95% CI: 0.05 to 0.51; P = 0.02), suggesting that the adverse effect of SO₂ on sperm morphology was more pronounced at higher levels of seminal plasma Ni. Further stratified analysis confirmed that high seminal plasma Ni levels were significantly associated with an increased rate of sperm abnormal morphology (β: 0.36; 95% CI: 0.15 to 0.57; P < 0.01); however, no significant association was observed in the low Ni group (β: 0.12; 95% CI: −0.10 to 0.35; P = 0.27), as shown in Supplementary Table 7.

Supplementary Table 6.

Interaction effects between ambient sulfur dioxide exposure during 0–90 day and essential trace element levels in seminal plasma on sperm quality parameters

Exposure	Outcome	Metals	β interaction	95% CI a	P interaction
SO2	Total sperm count	V	−0.04	−0.23 to 0.14	0.66
		Cr	0.03	−0.10 to 0.16	0.63
		Mn	0.11	−0.01 to 0.23	0.07
		Fe	0.06	−0.04 to 0.16	0.21
		Co	0.04	−0.09 to 0.17	0.53
		Ni	−0.03	−0.13 to 0.07	0.54
		Cu	0.03	−0.07 to 0.13	0.62
		Se	0.01	−0.09 to 0.11	0.84
		Sr	0.02	−0.10 to 0.13	0.74
	Sperm concentration	V	0.03	−0.15 to 0.21	0.75
		Cr	0.04	−0.08 to 0.17	0.50
		Mn	0.09	−0.03 to 0.20	0.13
		Fe	0.04	−0.05 to 0.13	0.36
		Co	0.03	−0.09 to 0.13	0.66
		Ni	−0.01	−0.10 to 0.08	0.85
		Cu	−0.00	−0.10 to 0.09	0.97
		Se	−0.01	−0.01 to 0.08	0.79
		Sr	−0.01	−0.12 to 0.10	0.92
	Progressive motility	V	0.03	−0.07 to 0.12	0.59
		Cr	−0.05	−0.11 to 0.02	0.14
		Mn	0.04	−0.02 to 0.10	0.16
		Fe	−0.04	−0.09 to 0.01	0.12
		Co	−0.01	−0.07 to 0.05	0.76
		Ni	−0.02	−0.07 to 0.03	0.43
		Cu	−0.01	−0.06 to 0.04	0.79
		Se	−0.03	−0.08 to 0.02	0.21
		Sr	−0.05	−0.11 to 0.00	0.06
	Total motility	V	0.02	−0.05 to 0.09	0.58
		Cr	−0.03	−0.08 to 0.02	0.26
		Mn	0.04	−0.00 to 0.08	0.07
		Fe	−0.02	−0.06 to 0.02	0.24
		Co	−0.00	−0.05 to 0.05	0.92
		Ni	−0.01	−0.05 to 0.02	0.48
		Cu	−0.00	−0.04 to 0.04	0.97
		Se	−0.02	−0.06 to 0.02	0.32
		Sr	−0.04	−0.08 to 0.01	0.09
	Sperm abnormal morphology	V	−0.17	−0.61 to 0.26	0.43
		Cr	0.14	−0.16 to 0.45	0.36
		Mn	−0.16	−0.43 to 0.12	0.27
		Fe	0.09	−0.14 to 0.32	0.43
		Co	0.09	−0.21 to 0.39	0.54
		Ni	0.28	0.05 to 0.51	0.02
		Cu	0.02	−0.21 to 0.26	0.84
		Se	0.10	−0.13 to 0.33	0.41
		Sr	0.16	−0.11 to 0.43	0.25

^aModel adjusted for age, BMI, smoking status, educational level, month of semen collection, average temperature, and relative humidity during the 0–90 days. SO₂: sulfur dioxide; Cr: chromium; Mn: manganese; V: vanadium; Se: selenium; Cu: copper; Co: cobalt; Ni: nickel; Sr: strontium; Fe: iron; BMI: body mass index

Supplementary Table 7.

Stratified analysis of the association between sulfur dioxide exposure during 0–90 day and abnormal sperm morphology, stratified by median Ni level in seminal plasma

Variable	Interaction term β interaction	P interaction	Ni exposure status	β (95% CI)	P
SO2	0.28	0.02	Lower	0.12 (−0.10 to 0.35)	0.27
			Higher	0.36 (0.15 to 0.57)	<0.01

Stratified analysis was performed to evaluate whether the association between SO₂ exposure and abnormal sperm morphology differs by Ni levels in seminal plasma. Ni exposure levels in seminal plasma was divided at the median into “lower” and “higher” groups. Model adjusted for age, BMI, smoking status, educational level, month of semen collection, average temperature, and relative humidity during the 0–90 days. BMI: body mass index; SO₂: sulfur dioxide; CI: confidence interval

Sensitivity analyses

To test the robustness of the findings, sensitivity analyses using two-pollutant models adjusted for PM_2.5, PM₁₀, CO, NO₂, and O₃ were performed to determine the relationship between average SO₂ exposure during 0–90 days and sperm quality (Supplementary Table 8–12). No significant differences were observed between the single-pollutant and two-pollutant models (both P < 0.05).

Supplementary Table 8.

The associations between sulfur dioxide exposure during entire spermatogenesis (0–90 days) and sperm quality parameters

Air pollutants	Lags (days)	Outcomes	Adjusted β (95% CI) a	P
SO2	0–90	Total sperm count	−0.09 (−0.16 to−0.03)	<0.01
SO2	0–90	Sperm concentration	−0.08 (−0.14 to−0.01)	0.02
SO2	0–90	progressive motility	−0.04 (−0.07 to−0.01)	0.01
SO2	0–90	Total motility	−0.04 (−0.06 to−0.01)	<0.01
SO2	0–90	Sperm abnormal morphology	0.28 (0.12 to 0.43)	<0.01

^aModel adjusted for age, BMI, smoking status, education level, month of semen collection, average temperature, relative humidity, and NO₂ during entire spermatogenesis. SO₂: sulfur dioxide; CI: confidence interval; NO₂: Nitrogen oxides; BMI: body mass index

Supplement Table 12.

The associations between sulfur dioxide exposure during entire spermatogenesis (0–90 days) and sperm quality parameters

Air pollutants	Lags (days)	Outcomes	Adjusted β (95% CI) a	P
SO2	0–90	Total sperm count	−0.09 (−0.16 to−0.03)	<0.01
SO2	0–90	Sperm concentration	−0.08 (−0.14 to−0.01)	0.02
SO2	0–90	Progressive motility	−0.04 (−0.07 to−0.01)	0.02
SO2	0–90	Total motility	−0.03 (−0.06 to−0.01)	<0.01
SO2	0–90	Sperm abnormal morphology	0.29 (0.14 to 0.45)	<0.01

^aModel adjusted for age, BMI, smoking status, education level, month of semen collection, average temperature, relative humidity, and O₃ during entire spermatogenesis. SO₂: sulfur dioxide; BMI: body mass index; CI: confidence interval; O₃: ozone

Supplementary Table 9.

The associations between sulfur dioxide exposure during entire spermatogenesis (0–90 days) and sperm quality parameters

Air pollutants	Lags (days)	Outcomes	Adjusted β (95% CI) a	P
SO2	0–90	Total sperm count	−0.09 (−0.15 to−0.02)	<0.01
SO2	0–90	Sperm concentration	−0.07 (−0.13 to−0.01)	0.02
SO2	0–90	progressive motility	−0.04 (−0.07 to−0.01)	0.02
SO2	0–90	Total motility	−0.03 (−0.06 to−0.01)	<0.01
SO2	0–90	Sperm abnormal morphology	0.27 (0.12 to 0.43)	<0.01

^aModel adjusted for age, BMI, smoking status, education level, month of semen collection, average temperature, relative humidity, and CO during entire spermatogenesis. SO₂: sulfur dioxide; BMI: body mass index; CI: Confidence interval; CO: carbon monoxide

Supplementary Table 10.

The associations between sulfur dioxide exposure during entire spermatogenesis (0–90 days) and sperm quality parameters

Air pollutants	Lags (days)	Outcomes	Adjusted β (95% CI) a	P
SO2	0–90	Total sperm count	−0.09 (−0.15 to−0.02)	<0.01
SO2	0–90	Sperm concentration	−0.07 (−0.14 to−0.01)	0.02
SO2	0–90	progressive motility	−0.04 (−0.07 to−0.01)	0.01
SO2	0–90	Total motility	−0.03 (−0.06 to−0.01)	<0.01
SO2	0–90	Sperm abnormal morphology	0.27 (0.12 to 0.43)	<0.01

^aModel adjusted for age, BMI, smoking status, education level, month of semen collection, average temperature, relative humidity, and PM_2.5 during entire spermatogenesis. SO₂: sulfur dioxide; BMI: body mass index; PM_2.5: particulate matter; CI: confidence interval

Supplementary Table 11.

The associations between sulfur dioxide exposure during entire spermatogenesis (0–90 days) and sperm quality parameters

Air pollutants	Lags (days)	Outcomes	Adjusted β (95% CI) a	P
SO2	0–90	Total sperm count	−0.09(−0.15 to−0.02)	0.01
SO2	0–90	Sperm concentration	−0.07 (−0.14 to−0.01)	0.02
SO2	0–90	progressive motility	−0.04 (−0.07 to−0.01)	0.02
SO2	0–90	Total motility	−0.03 (−0.06 to−0.01)	<0.01
SO2	0–90	Sperm abnormal morphology	0.26 (0.10 to 0.41)	<0.01

^aModel adjusted for age, BMI, smoking status, education level, month of semen collection, average temperature, relative humidity, and PM₁₀ during entire spermatogenesis. SO₂: sulfur dioxide; CI: confidence interval; BMI: body mass index; PM_2.5: particulate matter

DISCUSSION

Research has shown that global fertility levels have been decreasing owing to decreased sperm quality, which calls for further studies to identify the risk factors influencing sperm quality. Here, we investigated the associations between SO₂ exposure during 0–90 days, the concentrations of nine essential elements in seminal plasma, and sperm quality.

Results of linear regression analysis indicated that SO₂ exposure during 0–90 days was negatively associated with sperm concentration, total sperm count, and progressive motility and positively associated with sperm abnormal morphology. These results are consistent with the findings of Guo et al.,28 who reported negative associations between SO₂ exposure during 0–90 days and sperm normal morphology, total sperm motility, and progressive motility. In this study, we further confirmed the association between SO₂ exposure and sperm quality in Anhui Province (China), consistent with the findings of our previous study.5

DLNM revealed that SO₂ exposure during 0–90 days was significantly and negatively associated with total sperm count (lags 47–69), sperm concentration (lags 37–58), sperm progressive motility (lags 41–58), and sperm total motility (lags 40–62). However, it exhibited a positive association with sperm abnormal morphology (at lags 44–54). These findings are congruent with the results of Guo et al.,28 who observed that SO₂ exposure was negatively associated with sperm normal form, vitality, total motility, and progressive motility.

Human spermatogenesis takes approximately 90 days and is divided into four biologically relevant stages: epididymal storage (days 0–9), acquisition of motility (days 10–14), spermatogenesis phase II (days 15–69), and spermatogenesis phase I (days 70–90).29 Spermatogenesis phase II, which corresponds to meiotic and early spermiogenic processes, is characterized by multiple DNA replication cycles that can be affected by external stimuli.30 In our study, we used the DLNM model to identify sensitive exposure windows in the association between SO₂ exposure during the 90 days preceding semen examination and sperm quality. Notably, the most vulnerable period was spermatogenesis phase II (days 15–69), which encompasses the meiotic and early spermiogenic processes. Considering the cellular activity and chromatin remodeling during these phases, they are likely to be affected by external environmental factors such as air pollutants, heavy metals, and oxidative stress. Disruptions during this period can compromise the fidelity of DNA transmission and affect sperm production and quality.31

In this study, we found that exposure to SO₂ at lags 47–69 was negatively associated with total sperm count, while exposure to SO₂ at lags 37–58 was negatively associated with sperm concentration These results indicated that spermatogenesis in phase II represents a key susceptibility window for these parameters. Consistent with our results, other studies have also revealed that exposure to ambient air pollutants during this period can impair sperm quality. For example, Lu et al.31 reported a negative association between O₃ exposure and sperm concentration (β: −1.85; 95% CI: −2.87 to −0.83) and total sperm count (β: −6.65; 95% CI: −10.62 to −2.69) during spermatogenesis stage II, which implies that higher exposure during this developmental window can reduce spermatogenic output. This is likely because spermatogenesis phase II involves a series of DNA replication cycles that regulate the total sperm count and are highly susceptible to external stimuli.30 Notably, if meiosis is disrupted to affect sperm count, a large number of spermatocytes may be detectable in semen.13 Considering the predictive value of immature germ cells in semen for assessing testicular function, quantifying their abundance is likely to be a non-invasive alternative to testicular biopsy for evaluating spermatogenesis. However, our study did not include direct measurements of spermatocytes in semen. Nevertheless, other studies have reported its utility and may provide indirect evidence supporting the hypothesis that air pollution exposure compromises spermatogenesis.32 In the future, it will be important to combine quantification of spermatocytes in semen with DLNM modeling to precisely identify the critical exposure windows during which air pollutants affect semen quality.

With respect to sperm motility, we found that exposure to SO₂ at lags 41–58 and 40–62, corresponding to the meiotic phase of spermatogenesis, was negatively associated with progressive and total sperm motility. This phenomenon can be ascribed to the possibility that SO₂ exposure disrupts meiosis, increases spermatocyte apoptosis, and impairs sperm motility.33 Zhao et al.34 observed a significant negative association between PM exposure during the entire period of sperm development and sperm motility. Analysis of the critical exposure windows showed that adverse effects of PM exposure during the period of spermatogenesis on sperm motility were markedly larger than those in the sperm motility development period and epididymal storage, suggesting that PM exposure during spermatogenesis may affect sperm motility more profoundly than the other two exposure windows. An explanation for this phenomenon is that PM exposure during this key period of sperm development may alter the expression of genes involved in transcription and translation.34 Another possible mechanism is that SO₂ exposure during meiotic arrest may alter mitochondrial integrity and disrupt energy metabolism, thereby reducing sperm motility.35

In contrast, the epididymal storage phase, corresponding to the 0–9-day lag window, represents a crucial period for sperm motility. For example, previous studies have reported that exposure to SO₂ at lags 0–9 was negatively associated with sperm motility.5,11 However, no significant association was found between O₃ exposure at lags 0–9 and sperm progressive motility or total motility.31 Wu et al.29 also found that PM_2.5 exposure at lags 0–9 was not strongly associated with sperm progressive motility or total motility. Although the epididymal phase (0–9 days before ejaculation) influences the acquisition of sperm motility, we found that exposure to SO₂ at lags 41–62 may be associated with reduced sperm motility. This indicates that testicular events during spermatogenesis may compromise the structural integrity or functional capacity of sperm prior to epididymal maturation. However, we did not evaluate sperm tail abnormalities in this study, and thus, we cannot rule out the possibility that some of the observed motility impairment may be attributable to epididymal dysfunction. Future studies should perform detailed morphological examinations of sperm tails to clarify the relative contributions of testicular and epididymal factors to impaired motility.

A study has shown that the epididymis is involved in post-testicular sperm maturation and that air pollution exposure during the immediate pre-ejaculation period (0–9 days) may modify sperm morphology, possibly by affecting epididymal transit or membrane remodeling.36 In addition, Wang et al.37 found that a 1°C increase in temperature above the threshold correlated with a 2.04% decrease in sperm normal morphology at lags 0–9. However, we observed a significant association between exposure to SO₂ at lags 44–54 and sperm abnormal morphology. Consistently, Henry et al.38 demonstrated a strong association between PM_2.5 exposure during the early spermatogenic window and increased rates of sperm head abnormalities, although this effect was observed during the later windows. Mitochondrial dysfunction may lead to meiotic arrest and thereby induce morphological defects in sperm, providing a plausible mechanistic explanation consistent with the timing observed in our study.39 Altogether, these findings suggest that sperm morphology may be influenced not only by exposure during the epididymal storage phase but also by earlier disruptions during spermatogenesis phase II, specifically through a mechanism involving mitochondrial damage that interferes with meiotic progression and triggers structural abnormalities.

Notably, DLNM does not directly uncover biological mechanisms but instead identifies time windows during which exposures are more strongly associated with outcomes. Researchers can interpret these windows alongside established stages of human spermatogenesis to infer biologically sensitive periods. Moreover, spermatogenesis in humans is a relatively synchronous process, which contrasts with that observed in many animal models such as rodents, where spermatogenesis is often cyclic and asynchronous. These interspecific differences may limit the direct extrapolation of mechanistic findings from animal studies to humans. This implies that our results should be interpreted with caution when comparing across species. Above all, these findings suggest that minimizing SO₂ exposure at least 90 days prior to semen collection may improve sperm quality in infertile men.

Mediation analysis indicated that Se levels in seminal plasma significantly mediated the association between average SO₂ exposure during the 0–90 days prior to semen collection and total sperm count (mediation proportion: 14.6%), as well as the association with sperm concentration (mediation proportion: 23.6%). Collectively, the present findings indicate that Se levels in seminal plasma may mediate the association between SO₂ exposure and impaired spermatogenesis. Our results suggest that the effects of SO₂ may be mediated, at least in part, through trace metals in seminal plasma. Currently, there are no standardized data on atmospheric metal concentrations, which further complicates the investigation of their potential role in the relationship between SO₂ exposure and sperm quality. A review of the current literature revealed that few investigations have explored the mechanisms by which SO₂ exposure damages sperm quality. Similarly, the mechanisms by which trace metals contribute to impaired sperm quality, particularly the role of oxidative stress (OS), are unclear. OS can damage spermatogenic cells and reduce semen quality.40 Se contributes to the biosynthesis of selenoproteins, thereby acting as an antioxidant.41 Nelson et al.42 reported that Se deficiency increased oxidative stress levels in macrophages. Taken together, these findings suggest that Se may play a mediating role in the association between SO₂ exposure and sperm quality. Therefore, we hypothesize that SO₂ may deplete Se-dependent antioxidant defenses (e.g., by reducing glutathione peroxidase activity), which could exacerbate oxidative stress in testicular or epididymal tissues and ultimately impair sperm quality.

Interaction analyses revealed a potential synergistic effect between SO₂ and Ni on sperm abnormal morphology, which is biologically plausible considering that both pollutants can induce oxidative stress and impair testicular function via complementary mechanisms.43,44 This analysis was motivated by the concern that certain metals (e.g., Ni and V) may act not only as independent mediators but also act as co-pollutants stemming from the same anthropogenic sources as SO₂. For example, combustion processes in industrial activities and diesel engines emit particulate matter containing V and Ni.45 Peralta et al.46 identified V and Ni as constituents of PM_2.5. In such cases, these trace elements may exert direct toxicological effects or modify the association between SO₂ exposure and sperm quality. In this study, we incorporated the interaction terms into models to clarify whether the combined exposure to SO₂ and specific essential elements produced additive or synergistic effects on sperm quality. Interestingly, we found a significant interaction only for SO₂ exposure and Ni, suggesting a positive association between their combined exposure and sperm abnormal morphology. This finding implies a potential synergistic effect between SO₂ and Ni, which is biologically plausible considering that both pollutants can induce oxidative stress and disrupt testicular function through complementary mechanisms.47,48

To date, few studies have investigated the association between SO₂ exposure and the concentrations of essential elements in seminal plasma. There is some evidence regarding the impact of SO₂ exposure on metal levels in the human body. Hassan and colleagues identified positive associations between SO₂ exposure and blood concentrations of arsenic (As) and lead (Pb).19 In our study, we found that SO₂ exposure during 0–90 days was negatively associated with the concentrations of Fe, Se, and Sr and positively associated with the concentrations of V and Ni in seminal plasma. Mediation analysis indicated that Se levels in seminal plasma mediated the association between SO₂ exposure during 0–90 days and sperm concentration and total sperm count. Interaction analyses further revealed a potential synergistic effect of SO₂ and Ni on sperm abnormal morphology.

This study has several strengths. First, our study provides novel insights into the association between male reproductive health and SO₂ exposure. Considering the declining trend in sperm quality from 1973 to 2018 and the significantly negative association between SO₂ and sperm quality, even at concentrations below the WHO criterion of 40 μg m⁻³ for a 24-h average, these findings warrant further investigation. Second, this study provides novel evidence whether the concentrations of essential elements in seminal plasma mediate or moderate the association between SO₂ exposure and sperm quality. Third, similar results were observed in both single-pollutant and two-pollutant models, which support the robustness of our findings.

Nevertheless, this study has several limitations that should be acknowledged. First, although we adjusted for several important confounding factors known to affect semen quality, such as male age, days of abstinence, and the month of semen collection, the existence of unmeasured confounders, including dietary habits and other lifestyle factors, cannot be ruled out. Moreover, considering resource constraints, we could not explore the markers of subclinical inflammation, such as leukocyte counts in semen, cytokine levels, or reactive oxygen species (ROS), which limits our ability to fully account for the potential influence of subclinical inflammatory processes on semen quality. In addition, although participants with overt genital tract infections were excluded, undetected subclinical inflammation could not be ruled out and may increase individual susceptibility to SO₂ exposure. Such conditions act as residual confounders or effect modifiers, which need to be clarified further. Second, while we estimated the ambient temperature exposure of the study subjects, other environmental factors, such as sleep and work environment temperatures, could not be assessed, and their potential effects need to be considered. Third, we did not investigate the impact of individual exposure to residential outdoor pollutants, indoor air quality, or activity patterns that could potentially increase the daily exposure to gaseous air pollution. Fourth, although the DLNM estimates time windows of effect, interpretation regarding specific stages should be conducted with caution and grounded in biological plausibility. In future studies, additional data should be collected to account for these factors. Fifth, the inclusion of subjects from a reproductive medicine center may have introduced a collider bias. Sixth, owing to the retrospective observational nature of our study, we could not establish a causal relationship between SO₂ exposure and sperm quality. Therefore, future prospective or mechanistic studies are required to validate and further elucidate our findings. Finally, these findings suggest potential mediating and moderating roles. Further animal studies are required to clarify these underlying mechanisms.

CONCLUSION

Here, we found that SO₂ exposure during 0–90 days was negatively associated with sperm concentration, total sperm count, PR, and total motility but positively associated with sperm abnormal morphology. DLNM results also revealed significant lagged effects of SO₂ exposure at lags 37–69 on sperm parameters. Moreover, SO₂ exposure during 0–90 days exhibited a negative association with the concentrations of Fe, Se, and Sr and a positive association with the concentrations of V and Ni in seminal plasma. Importantly, we observed that Se concentrations in seminal plasma, rather than Fe or Sr, mediated the association between SO₂ exposure during 0–90 days and sperm concentration and total sperm count. This suggests that Se levels in seminal plasma may serve as a surrogate indicator of the influence of SO₂ exposure on these sperm parameters. An interaction between SO₂ exposure and Ni levels in seminal plasma was detected in relation to sperm abnormal morphology, indicating a potential synergistic effect between SO₂ and Ni. Finally, this study provides new findings regarding the effect of SO₂ exposure on male reproductive function. Further prospective and mechanistic studies are required to validate these findings.

AUTHOR CONTRIBUTIONS

TY carried out the investigation, designed the methodology, curated the data, performed the formal analysis, and drafted the original manuscript. XYZ conducted the visualization and participated in the formal analysis. XQL contributed to the methodology design and assisted with the formal analysis. RD participated in the methodology design and supported the investigation. QL aided in the methodology design and assisted with data curation. XYY contributed to the methodology design and supported data curation. DYZ participated in the methodology design. FBT assisted with the methodology design, acquired the funding for the study, and conceived of the study. GYL participated in conceiving the study. DMJ contributed to the methodology design and assisted in conceiving the study. CML reviewed and edited the manuscript, supervised the research process, and participated in funding acquisition. YXC reviewed and edited the manuscript, supervised the research process, participated in funding acquisition, conceived of the study, and administered the project. All authors read and approved the final manuscript.

COMPETING INTERESTS

All authors declared no competing interests.