Ambient PM2.5 and clinically recognized early pregnancy loss: A case-control study with spatiotemporal exposure predictions

Abstract Background Experimental research suggests that fine particulate matter (PM 2.5 ) exposure might affect embryonic development. However, only few population-based studies have investigated the impact of maternal exposure to PM 2.5 on the early pregnancy loss. Objectives To estimate associations between clinically recognized early pregnancy loss (CREPL) and exposure to ambient PM 2.5 at individual residences during peri-conception periods, with the aim to identify susceptible exposure time windows. Methods CREPL cases and normal early pregnancy controls (of similar age and gravidity presenting within one week, a total of 364 pairs) were recruited between July 2017 and July 2018 among women residing in Tianjin, China. Average ambient PM 2.5 concentrations of ten exposure windows (4 weeks, 2 weeks and 1 week before conception; the first, second, third and fourth single week, the first and second 2-week periods, and the entire 4-week period after conception) at the women's residential addresses were estimated using temporally-adjusted land use regression models. Associations between PM 2.5 exposures at specific peri-conception time windows and CREPL were examined using conditional logistic regression models, adjusted for covariates. Results Based on adjusted models, CREPL was significantly associated with a 10 μg/m 3 increase in PM 2.5 exposure during the second week after conception (OR = 1.15; 95% CI: 1.04, 1.27; p  = 0.005), independent of effects at other time windows. There was also an association of CREPL with PM 2.5 during the entire 4-week period after conception (OR = 1.22; 95% CI: 1.02, 1.46; p  = 0.027). There was little evidence for associations with exposure during pre-conception exposure windows. Conclusions Maternal exposures to ambient PM 2.5 during a critical time window following conception are associated with CREPL, with the second week after conception possibly being the exposure window of most vulnerability. Future studies should focus on replicating these findings and on pathogenic mechanisms.