Importance: Using provisional or opportunistic data, three nationwide studies (The Netherlands, USA, and Denmark) have identified a reduction in preterm or extremely preterm births during periods of COVID-19 restrictions. However, these findings have been challenged as none of the studies accounted for perinatal deaths.
Objective: To determine whether the reduction in extremely preterm births, observed in Denmark during the COVID-19 lockdown, could be a result of an increase in number of perinatal deaths; and to assess the impact of extended COVID-19 restrictions on the prevalence of preterm birth and perinatal death.
Design: The study is a nationwide Danish register-based prevalence proportion study using detailed data to capture all births (induced abortions were excluded) throughout Denmark. We assessed the prevalence of stillbirth, preterm birth, and early neonatal death.
Setting: Population-based study
Participants: All singleton pregnancies delivered in Denmark, between February 27, and September 30, 2015-2020. COVID-19 lockdown was studied in 31,164 births and the extended period of COVID-19 restrictions in 214,862 births.
Exposure: COVID-19 restrictions broadly (February 27, to September 30, 2020) or COVID-19 lockdown specifically (March 12, to April 14, 2020).
Main Outcome: Prevalence of preterm births, stillbirths and early neonatal deaths across the periods under study.
Results: The extremely preterm birth rate was reduced (OR 0.27, 95% CI 0.07 to 0.86) during the strict lockdown period, while perinatal mortality was not significantly different. During the extended period of restrictions, the extremely preterm birth rate was marginally reduced, and a significant reduction in the stillbirth rate (OR 0.69, 0.50 to 0.95) was observed. No changes in early neonatal mortality rates were found.
Conclusion and Relevance: Stillbirth and extremely preterm birth rates were reduced in Denmark during the period of COVID-19 restrictions and lockdown, respectively, suggesting that aspects of these containment and control measures confer an element of protection. The present observational study does not allow for causal inference; however, the results support design of studies to ascertain whether behavioural or social changes for pregnant women may improve pregnancy outcomes.