Maternal exposure to ambient temperature and risk of preterm birth in Chengdu, China, from 2017 to 2020: a cohort study

Abstract

Background: Due to climate change, the frequency and intensity of heat waves and other extreme weather events are rapidly increasing. Compared to the general population, pregnant women and fetuses are increasingly vulnerable to the effects of extreme temperatures and are associated with the occurrence of adverse birth outcomes, including preterm birth (PTB). However, its risk of preterm birth is currently uncertain. The objective of the research is to examine the effect of ambient temperature on PTB in pregnant women.

Methods: This study included 6,850 pregnant women from the Tongji-Shuangliu Birth Cohort. Meteorological data for Chengdu through the European Centre for Medium-Range Weather Forecasts. The main exposure assessment was conducted during eight different exposure windows, including the first three months of pregnancy, 7 weeks periods during the first two trimesters, throughout pregnancy, 1-week preceding delivery, and 4 weeks preceding delivery. The effect of environmental temperature on PTB during different exposure windows was assessed using the logistic regression based on the percentile of the mean temperature in different exposure cycles. Additionally, the lagged effect of environmental temperature on preterm births throughout the study period was analyzed using a distributed lag non-linear model.

Results: Among the 6850 pregnant women, 301 (4.4%) were diagnosed with PTB. Compared to mild temperature (10th to 90th percentile), exposure to extreme cold (< 10th percentile) temperature during the 4 weeks preceding delivery (RR = 2.45, 95% CI:1.11,5.40) and throughout pregnancy (RR = 3.85, 95% CI:1.56,9.53) increased the risk of PTB. In addition, hot temperature (> 90th percentile) at 4 weeks preceding delivery (RR = 0.33, 95% CI:0.13,0.86) and 22-28 weeks of pregnancy (RR = 0.25, 95% CI:0.11,0.59), and cold exposure at 1-week preceding delivery(RR = 0.51, 95% CI:0.27,0.96), reduced risk of PTB. In the lagged model, compared with 18° C (50th percentile), 7 °C (10th percentile) had the strongest effect on lag day 21 and lag 22 (RR = 1.20, 95% CI:1.03,1.40; RR = 1.20, 95% CI:1.03,1.39). A temperature of 27° C (90th percentile) was protective for PTB from the 22nd day of lag(RR = 0.86, 95% CI:0.75,0.99).

Conclusions: This study indicates that high temperature may be a protective factor for PTB, while low temperature may be a risk factor, with an obvious lag effect.

Keywords: Ambient temperature; Climate change; Distributed delay nonlinear model; Lag effect; Preterm birth.

Fig. 1

Plot of RR by lag at extreme temperature distribution. The model is adjusted according of relative humidity, air pressure, precipitation, ground wind speed and direct radiation (during the corresponding day), long-term trend and seasonality, DOW and PH. Lags represents single-day lags (lag 0, representing temperature on the same day of preterm birth, up to30-day lag.). All figures are referenced to 18°. RR, relative risk

Fig. 2

Contour plot for variation trend of RR with temperature and lag time. The model is adjusted according of relative humidity, air pressure, precipitation, ground wind speed and direct radiation (during the corresponding day), long-term trend and seasonality, DOW and PH. Lags represents single-day lags (lag 0, representing temperature on the same day of preterm birth, up to30-day lag.). All figures are referenced to 18°. RR, relative risk

Fig. 3

Three-D plot for variation trend of RR with temperature and lag time. The model is adjusted according of relative humidity, air pressure, precipitation, ground wind speed and direct radiation (during the corresponding day), long-term trend and seasonality, DOW and PH. Lags represents single-day lags (lag 0, representing temperature on the same day of preterm birth, up to30-day lag.). All figures are referenced to 18°. RR, relative risk