Abstract: Objective To understand the exposure-lag-response associations between exposure to PM_2.5 and urinary leukocyte and white blood cell count in late pregnancy in Guangzhou, and identify the susceptible exposure windows. Methods According to a birth cohort study on Prenatal Environments and Offspring Health （PEOH） established in Guangzhou in the early stage, a spatiotemporal land-use-regression model was used to estimate a pregnant woman's weekly PM_2.5 exposure from 13 weeks before pregnancy to the sampling week in the late pregnancy. A distributed lag non-linear model incorporated with a Cox proportional hazard model was applied to assess the effect of weekly-specific maternal PM_2.5 exposure on the risk of urinary leukocyte positive, and a distributed lag non-linear model was applied to assess the exposure-lag-response associations between weekly PM_2.5 exposure and white blood cell count, respectively. Results The sample included 3 753 pregnant women. The risk of urinary leukocyte positive was positively associated with maternal PM_2.5 exposure during the 20th to 27th gestational weeks, and the strongest effect was observed in 24th gestational weeks （HR=1.030, 95% CI： 1.007-1.054, per 10 μg/m³ increment in PM_2.5）. Maternal exposure to PM_2.5 during the 5th week before pregnancy to 6th gestational week was positively associated with white blood cell count, and the strongest effect was observed in 0 gestational week, with a 10 μg/m³ increment of PM_2.5, white blood cell count increased 0.22% （95% CI： 0.10%-0.35%）. Conclusion Maternal exposure to PM_2.5 may increase the risk of urinary leukocyte positive and increase the white blood cell count in late pregnancy, the 20th to 27th gestational week and the 5th week before pregnancy to 6th gestational week may be the susceptible exposure windows, respectively.

Key words: PM_2.5, White blood cell, Urinary leukocyte, Susceptible exposure window, Birth cohort study