A total of 3,355 unique citations were screened, of which 2,796 were excluded (Fig. 1). Of the 559 citations included for full text review, 518 were excluded. A total of 42 studies were included, from which data on 18,282 women were extracted (Appendix S3). Included studies were mostly observational designs (38 studies), though four randomized controlled trials were included. A total of 37 studies were conducted in high income or upper-middle income countries, 4 studies were conducted in low-middle income countries and one was a multi-country study was conducted in Brazil, Kenya and the UK (zero studies in low-income countries).(28) Data collection periods ranged from 1987 to 2017, however 20 studies did not specify the year data were collected.
In total, 36 studies included low-risk pregnant women only, while 6 studies included unselected-risk women (though these definitions varied across studies) (Table 2). Sample sizes from which data were extracted ranged from 14 to 2,868 women; 24 studies had a sample of less than 200 women. The mean maternal age (reported in 27 studies) ranged from 19.9 to 32 years. The mean gestational age of the study population was only specified in 22 studies, ranging from 20 to 40 weeks (in 8 studies, the mean gestational age was 38 to 40 weeks). Eight studies did not specify the type of Doppler used, though 6 used continuous-wave, 6 used pulsed-wave and 1 study used continuous or pulsed wave (practice varied between participating centres (29)). Overall, 6 of 7 case-control studies, 9 of 12 cross-sectional studies and all cohort studies were assessed as satisfactory, good or very good quality (Appendix S4). The four trials were assessed as low or unclear risk of bias across all domains, except for lack of blinding of participants, personnel and outcome assessment in three of four trials.(29–31)
Table 1
Characteristics of included studies
Study design
|
N
|
%
|
Case-control study
|
7
|
16.7
|
Cross-sectional study
|
12
|
26.2
|
Cohort study
|
19
|
45.2
|
Randomized controlled trial
|
4
|
9.5
|
Country income level
|
|
|
High-income countries
|
24
|
57.1
|
Upper-middle income countries
|
13
|
31.0
|
Low-middle income countries
|
4
|
9.5
|
Low-income countries
|
0
|
0.0
|
Multiple*
|
1
|
2.1
|
* one study was a multi-country study conducted in Brazil (upper-middle income country), Kenya (low-middle income country) and the UK (high-income country)
|
Table 2
Characteristics of women and Doppler assessments in included studies
Study population
|
N (studies)
|
%
|
Maternal risk
Low-risk women only
Unselected women
|
36
6
|
85.7
14.3
|
Doppler assessments
Doppler ultrasound of umbilical artery only reported
Doppler ultrasound of umbilical artery plus other Doppler investigations reported*
|
15
27
|
35.7
64.2
|
Type of Doppler
Colour Doppler
Continuous-wave Doppler
Pulsed-wave Doppler
Pulsed-wave Doppler; colour Doppler
Pulsed-wave and continuous-wave Doppler
Doppler (not otherwise specified)
|
12
6
6
9
1
8
|
28.6
14.3
14.3
21.4
2.1
19.0
|
*Fetal cerebral circulation, fetal aorta, fetal renal arteries, umbilical vein, ventricular outlets, femoral vessels, uterine/placental vessels
|
Table 3
Studies with a prevalence of absent or reversed end diastolic flow greater than zero*
Study
|
Design
|
Study population
|
AEDF prevalence
|
Low risk women
|
Souka 2012 (35)
|
Cross-sectional study
|
2189 low-risk women in Greece
|
1/1289 (0.05%)
|
Mason 1993 (31)
|
RCT
|
863 low-risk, nulliparous women in the UK (Doppler arm only)
|
2/863 (0.23%)
|
Unselected risk women
|
Davies 1992 (30)
|
RCT
|
1246 unselected-risk women in the UK (Doppler arm only)
|
1/1246 (0.08%)
|
Beattie 1989 (33)
|
Cohort
|
2097 unselected-risk women in USA
|
6/2097
(0.29%)
|
Nkosi 2019 (36)
|
Cohort
|
2868 unselected-risk women in South Africa
|
38/2868 (1.32%)
|
Yoon 1993 (37)
|
Cohort
|
328 unselected women in South Korea
|
7/328 (2.13%)**
|
*Results of all included studies are in Appendix S3
|
**Defined as AEDF or REDF
|
Across all 42 studies (18,282 women), 36 studies had zero AEDF events, while 6 studies reported a total of 55 cases of AEDF or REDF. Forty-eight of these cases were AEDF only and seven were AEDF or REDF, reported in a 1993 study in South Korea by Yoon et al.(32) These six studies reported a prevalence ranging from 0.08–2.13% (Table 4).(30–36) They were generally larger samples (ranging from 328 to 2,868 women) – four studies were in unselected-risk women and five were conducted in high-income countries. The seventh study by Nkosi et al was conducted in South Africa (an upper-middle income country) and contributed 38 of the AEDF cases we identified.(36) They used a low-cost, handheld, continuous flow Doppler apparatus as a routine screening tool for assessing umbilical vessel flow in 2,868 unselected-risk women attending two community health centres for antenatal care. The highest prevalence was reported by Yoon et al, who reported a prevalence of AEDF or REDF of 2.13% amongst 328 unselected-risk women in South Korea.(37) The four studies (898 women) in lower middle-income countries (Bangladesh, Papua New Guinea, Nigeria and Tunisia) which did not report any cases of AEDF or REDF.(38–41) Similarly, a three-country prospective cohort study of 431 healthy, low-risk women by Drukker et al in Brazil, Kenya and the UK identified no AEDF cases in any country.(28)
The CS rate for the population of interest was generally not reported (25 studies). Where it was reported, it ranged from 2.9–57.1% (Appendix S3). Similarly, in most studies, stillbirth (30 studies), early neonatal death (36 studies), perinatal death (31 studies) and neonatal death (34 studies) were not reported for the population of interest. Where these rates were reported, they were very low, except for the study by Yoon et al that reported 7 cases of AEDF or REDF, 10 stillbirths and 18 perinatal deaths among 328 unselected-risk women.(37)