Maternal and perinatal outcomes related to COVID-19 and pregnancy: An overview of systematic reviews

Abstract

Introduction: Evidence about coronavirus disease 2019 (COVID-19) and pregnancy has rapidly increased since December 2019, making it difficult to make rigorous evidence-based decisions. The objective of this overview of systematic reviews is to conduct a comprehensive analysis of the current evidence on prognosis of COVID-19 in pregnant women.

Material and methods: We used the Living OVerview of Evidence (L·OVE) platform for COVID-19, which continually retrieves studies from 46 data sources (including PubMed/MEDLINE, Embase, other electronic databases, clinical trials registries, and preprint repositories, among other sources relevant to COVID-19), mapping them into PICO (population, intervention, control, and outcomes) questions. The search covered the period from the inception date of each database to 13 September 2020. We included systematic reviews assessing outcomes of pregnant women with COVID-19 and/or their newborns. Two authors independently screened the titles and abstracts, assessed full texts to select the studies that met the inclusion criteria, extracted data, and appraised the risk of bias of each included systematic review. We measured the overlap of primary studies included among the selected systematic reviews by building a matrix of evidence, calculating the corrected covered area, and assessing the level of overlap for every pair of systematic reviews.

Results: Our search yielded 1132 references. 52 systematic reviews met inclusion criteria and were included in this overview. Only one review had a low risk of bias, three had an unclear risk of bias, and 48 had a high risk of bias. Most of the included reviews were highly overlapped among each other. In the included reviews, rates of maternal death varied from 0% to 11.1%, admission to intensive care from 2.1% to 28.5%, preterm deliveries before 37 weeks from 14.3% to 61.2%, and cesarean delivery from 48.3% to 100%. Regarding neonatal outcomes, neonatal death varied from 0% to 11.7% and the estimated infection status of the newborn varied between 0% and 11.5%.

Conclusions: Only one of 52 systematic reviews had a low risk of bias. Results were heterogeneous and the overlap of primary studies was frequently very high between pairs of systematic reviews. High-quality evidence syntheses of comparative studies are needed to guide future clinical decisions.

Keywords: coronavirus disease 2019; coronavirus infections; infant; newborn; overview; pregnant women; systematic reviews as topic.

FIGURE 1

PRISMA flowchart. SR, systematic review. *These two articles correspond to the same review (preprint version and journal article)

FIGURE 2

Overall risk of bias of the included systematic reviews

FIGURE 3

Detailed assessment of corrected covered area. Our overview includes several systematic reviews (SRs), and each SR includes primary studies. It is expected that some primary studies are included in two or more SRs, which is known as “overlap of primary studies”. To assess this overlap, there is a formula known as corrected covered area (CCA), where values below 5% are considered low overlap; between 5% and 10% are considered moderate; between 10% and 15% are considered high; and above 15% are considered very high. Usually overlap is presented as an overall assessment for the whole body of evidence, but this approach sometimes fails to identify which specific SRs are contributing to double‐counting of the same primary studies. In this figure, we present not an overall CCA, but a CCA for each pair of SRs. White boxes represent low overlap (CCA <5%), green boxes represent moderate overlap (CCA between >5% and <10%), yellow boxes represent high overlap (CCA between >10% and <15%), and red boxes represent very high overlap (CCA ≥ 15%). The interpretation of each one of these boxes or “nodes” involves two SRs: a white node means that there are none or a minimum proportion of primary studies shared by the two SRs assessed, whereas a red node means that there is a considerable amount of primary studies shared by the pair of SRs assessed