Beyond the mean: quantile regression to differentiate the distributional effects of ambient PM2.5 constituents on sperm quality among men

Abstract

Ambient PM2.5 (particulate matter ≤ 2.5 μm in aerodynamic diameter) constituents have been related to mean changes in semen quality, but focusing on the mean response may not well capture distributional and heterogeneous effects of PM2.5 constituents on semen quality. In this study, 2314 semen samples of 622 men between Jan 1, 2019 and Dec 31, 2019 from Guangdong Human Sperm Bank were subjected to semen quality analysis. Daily average concentrations of PM2.5 constituents including 4 water-soluble ions and 15 metals/metalloid were measured for 7 days per month at 3 fixed atmospheric pollutant monitoring stations. We used quantile regression for longitudinal data to examine whether the associations between PM2.5 constituents and quality indicators of semen varied across quantiles of outcome distribution. Heterogeneous associations were found between PM2.5 constituents and sperm quality across different quantiles. An interquartile range (14.0 μg/m3) increase in PM2.5 mass was negatively associated with lower tails of sperm concentration and upper tails of sperm count distribution. PM2.5 vanadium exposure was significantly related to the 90th percentile of sperm count distribution, but not to the lower quantiles. In addition, those subjects with relatively high sperm motility were more susceptible to sulfate, chromium, and manganese constituents in PM2.5. Our results indicate that PM2.5 and certain constituents were associated with sperm quality, especially sperm motility, and the associations are more pronounced in men with relatively high or low sperm motility.