Interpretation
Our finding of a positive association between the COVID-19 lockdown and risk of PTB was consistent with some previous studies (8,12,28). Several reasons possibly explained the increased risk of PTB. First, the lack of medical resources during the COVID-19 pandemic and lockdown measures might interrupt the timely antenatal care for pregnant women (9,29). Secondly, fear and panic about the pandemic could make pregnant women reluctant to seek help from medical institutions, and further impacted the timely detection and diagnoses of pregnancy complications (9,30). For example, we observed a higher rate of GDM in the exposed group than the unexposed group. This suggested a potential mediation role of GDM, as GDM is a critical risk factor of PTB (10). In addition, pregnant women have always been considered a susceptible population to mental disorders (31). The lockdown and restriction measures could increase psychological problems in pregnant women through concomitant financial problems and increased stress (30,32), particularly if they were socioeconomically disadvantaged (33). The closure of entertainment venues also reduced the outlets for negative feelings (34). A previous study observed a more pronounced increase in depression and anxiety in pregnant women during the COVID-19 pandemic than in the general population (35). Lastly, the nutritional status of pregnant women was also of concern. During the lockdown period, the decreased supply of fresh foods could lead to inadequate intake of vegetables and high-fiber foods. Meanwhile, the intake of high-carbohydrate foods might have increased because they were relatively easier to obtain and store (9). It was reported that the overweight and obesity rates increased during the lockdown period due to unbalanced diets and less exercise (36). This suggested that the maternal stress and obesity during the lockdown might influence the risk of PTB (9,35).
We further observed that women in early and middle pregnancy during the Level I lockdown had a greater risk of PTB, which also contributed to the health effects of the COVID-19 lockdown. Zhang et al. reported that women in the first and second trimesters of pregnancy during the lockdown had more severe psychological disorders (16). A simple explanation could be that these mothers continued to experience Level II and III lockdown after the Level I lockdown, which may have led to more cumulative effects on their fetal health. This was supported by our observed positive association between PTB risk and cumulative exposure to lockdown of all levels in the first 24 GWs. An alternative explanation could be that early and middle pregnancy is a critical period for fetal development because the majority of fetal organ and tissues retain plasticity at that time (37). As a result, lockdown-induced poor diet, depression, and anxiety problems in early and middle pregnancy may substantially interrupt fetal development (38-40).
It should be noted that several other previous studies reported a reduction (rather than an increase as in our study) in rate of PTB during the COVID-19 lockdown (11,13). Although the mechanisms underlying these negative associations were unclear, several socio-environmental and behavioral modifiers were proposed (5,41). First, the lockdown measures increased company and support from partners and family, which could reduce the existing psychological stress in pregnant women. Second, working from home increased their rest time at home and decreased work-related stress. Third, the reduced anthropogenic emissions improved the air quality, which could benefit maternal and fetal health. Fourth, precautionary behavioral changes were promoted during the lockdown, including social distancing, enhanced hand hygiene, and use of face masks. These behavioral changes could potentially reduce the chances of other common viral infections in addition to COVID-19 during pregnancy. Finally, lockdown measures also reduced daily commuting, road traffic incidents, and consumption of cigarettes, coffee, alcohol, prescription drugs, and street drugs due to limited accessibility (5,41).
Previous studies reported inconsistent associations of lockdown exposure with maternal and fetal health (5,11,13). These inconsistencies may have a few explanations. First, some studies (11) had small sample sizes and potentially inadequate statistical power to detect an association between lockdown exposure and PTB. Second, the seemingly decreased risk of adverse pregnancy outcomes related to lockdown might be partially related to the reduced number of ultrasound scans and screening, which increased the possibility of under-diagnoses of early pregnancy loss, miscarriages, or stillbirths. Third, the health effects of lockdown may last for several months, but previous studies did not track participants long enough to assess the total effects of lockdown, which could have led to underestimations. In this study, we used the data of pregnant women who experienced the Level I lockdown until the end of 2020 and were able to obtain birth outcomes of all exposed women by covering the entire pregnancy. Fourth, air quality improvement during the lockdown was proposed as a major contributor to the reduced risk of PTB. In this study, we also found a substantial reduction in air pollution during the lockdown (Table S4), which was consistent with previous studies (2,42). Fifth, seasonal effects and pregnancy stages were not considered in most previous studies, which could lead to biased results. To evaluate this potential bias, we estimated the difference in PTB rates between new births during the Level I lockdown and all previous births during the entire years (rather than matching the calendar months) from 2015-2019. We did not find a significant association between lockdown and PTB risk (Table S5). Finally, although the lockdown measures may increase company and support from partners and family, the potential increase in family conflicts and domestic abuse should also be considered (43). These findings suggest that the health effects of COVID-19 lockdown were comprehensively affected by socio-environmental changes and behavioral modifications, and that improvement in one factor could not make up for the overall disadvantage (11,28).