The review protocol was registered on PROSPERO (CRD42020161980) and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist (See Appendix S1).(23) As a systematic review of published studies, ethical approval was not required.
Eligibility criteria
Eligible studies were primary research studies that reported the prevalence of AEDF and/or REDF amongst low-risk or unselected-risk pregnant women with singleton pregnancies. Randomized or non-randomized study designs (including cohort or cross-sectional studies, and randomized or non-randomized interventional studies) were eligible for inclusion, though case reports, case series and commentaries were not eligible. Conference abstracts were eligible provided that sufficient information was available for data extraction and quality assessment. The primary outcomes of interest were the prevalence of AEDF or REDF. Secondary outcomes included mortality (stillbirth, perinatal death, early neonatal death, neonatal death) and mode of birth (percentage of caesarean births).
Studies were eligible if they included either “unselected-risk” pregnant women (defined as women attending non-specialist antenatal care clinics, but have not been classified as being low- or high-risk based on local clinical guidelines, or as defined by study authors) or low-risk women (defined as women attending non-specialist antenatal care clinics, who have been classified as “low-risk” according to local clinical guidelines, or defined as low-risk by the study authors). Studies involving only high-risk women (regardless of how they were defined) were not eligible. If studies included mixed populations, they were included only if data could be extracted separately for low- and/or unselected-risk populations. Studies were eligible if umbilical arterial flow was assessed by Doppler ultrasound, regardless of whether this was continuous or pulsed-wave Doppler. We excluded studies where Doppler assessment was performed only in the first 20 weeks of pregnancy – this threshold was used as it reflects a when a consistently positive umbilical arterial flow can be expected. We also excluded studies of multiple pregnancies, post-term pregnancies, or where Doppler assessment was done during labour or immediately prior to caesarean section or labour induction.
Literature searching, data collection and analysis
We searched PubMed, Embase, CINAHL, Cochrane CENTRAL and Global Index Medicus for relevant studies with no date, setting or language restrictions (initial search 19 December 2017, updated search 10 January 2020, see Appendix S2 for search strategy). At least 2 review authors independently screened all titles and abstracts, assessed full texts of potentially eligible studies and extracted data (disagreements were resolved by discussion). Covidence software was used for title, abstract and full text screening.(24)
Extracted data from each study included characteristics of the study (design, year, country, sample size, study eligibility criteria), participants (including maternal risk profile) and Doppler assessment. Where necessary, we contacted authors of the included studies for additional information on the primary outcomes. Characteristics of studies and populations and prevalence of review outcomes were reported descriptively (SPSS version 26).(25) We planned but could not perform a meta-analysis due to significant methodological and reporting heterogeneity and sparseness of primary outcome events. Countries were classified into low, low-middle, upper-middle and high-income countries based on the current World Bank classification.(26)
As this review include different study designs, we used different tools to assess methodological quality. Observational studies were assessed using the relevant iteration of the Newcastle-Ottawa scale for case-control, cohort or cross-sectional studies.(27, 28) Randomised controlled trials were assessed using the Cochrane risk of bias tool.