Cochrane Reviews' summaries and their relevance for imaging: Doppler in obstetrics

The use of Doppler in obstetrics has spread progressively with the ever advancing technology and better understanding of the pathophysiology of the fetoplacental circulation. Although the time scale over which the Doppler changes develop in impaired placentation can vary, the cascade of events is mostly reproducible. This constitutes the rationale for the use of Doppler ultrasound in high-risk pregnancies as one of the main tools for the assessment of fetal wellbeing and leads to interventions such as increased fetal surveillance, labor induction and administration of corticosteroids in anticipation of preterm birth. Undoubtedly, false-positive findings may lead to unnecessary interventions and iatrogenic prematurity that could contribute to neonatal morbidity and increased maternal anxiety. Doppler ultrasound of the uterine and umbilical arteries has also been proposed as a screening test in low-risk pregnancies in order to identify ‘at-risk’ pregnancies and fetuses that may benefit from preventative strategies such as low-dose aspirin. Again, there is a real risk that any potential benefit may be overshadowed by unnecessary interventions and increased maternal anxiety. Here we summarize the findings from three recently updated Cochrane reviews1-3 that replace older versions4-6. Two focused on the fetoplacental (umbilical) circulation, one in high-risk1 and the other in low-risk2 pregnancies. The third review included trials of uterine artery Doppler in both high- and low-risk groups3. Standard Cochrane methodology was used, the details of which can be found in The Cochrane Library, available online7. The first systematic review1 addressed the comparison of fetal and umbilical Doppler ultrasound versus no Doppler ultrasound in high-risk pregnancies. Eighteen studies were included, involving 10 156 women. Fifty-eight meta-analyses were performed, looking separately at various sub-populations (singleton or multiple pregnancies [both primary and secondary outcomes]; small-for-gestational age, pre-eclampsia and hypertension, diabetes, prolonged pregnancy and previous pregnancy loss [primary outcomes only]). There was a statistically significant reduction in perinatal mortality in the Doppler group compared with the no-Doppler group (risk ratio (RR), 0.71; 95% CI, 0.52–0.98; 16 studies; 10 225 babies, perinatal mortality, 1.2% vs. 1.7%; number needed to treat, 203; 95% CI, 103–4352). The difference in perinatal death remained significant when the analysis focused just on potentially preventable perinatal deaths (excluding cases of termination of pregnancy and lethal chromosomal or other fetal abnormalities) (RR, 0.67; 95% CI, 0.46–0.98; 16 studies; 10 225 babies). The data for stillbirth, neonatal deaths and low Apgar score fitted into the overall picture, showing fewer adverse outcomes in the Doppler group, but without reaching statistical significance. The use of Doppler ultrasound in high-risk pregnancies was associated with a reduction in Cesarean sections (RR, 0.90; 95% CI, 0.84–0.97; 14 studies; 7918 women). When separating the Cesarean sections into elective and emergency, the reduction in the number of procedures appeared to be confined to the latter group. No significant differences were found for spontaneous vaginal births and operative vaginal births between the Doppler and no-Doppler groups. There was, however, a significant average reduction in induction of labor for women with Doppler intervention (RR, 0.89; 95% CI, 0.80–0.99; 10 studies; 5633 women). The subgroup analyses suggested that the benefits may be confined to singleton pregnancies, but these data should be interpreted cautiously, because there was just one study involving twins. As yet, there are no published clinical trials focusing on the possible benefit of adding Doppler studies of the fetal circulation (e.g. ductus venosus, middle cerebral arteries), but at least one large trial is currently ongoing8. The second systematic review2 addressed the comparison of routine fetal and umbilical Doppler ultrasound versus no-Doppler ultrasound in low-risk pregnancies. Twenty publications were identified, of which five studies (14 185 women) were included. There was no statistically significant reduction identified in the risk of perinatal death when Doppler ultrasound was used (average RR across studies, 0.85; 95% CI, 0.47–1.54; four studies; 11 190 women). Just one study assessed serious neonatal morbidity and showed no significant difference (RR, 0.99; 95% CI, 0.06–15.75; one study; 2016 women). There were no significant differences identified for any of the secondary outcomes, including stillbirth, neonatal death, Apgar scores < 7 at 5 min, Cesarean section, operative vaginal birth, spontaneous vaginal birth, induction of labor, neonatal resuscitation and preterm birth. The third systematic review3 addressed the issue of whether the use of uteroplacental Doppler ultrasound (uterine arteries and placental vessels) improves important perinatal outcomes in low- and high-risk pregnancies. Only two relevant trials were included, both assessing low-risk women in the second trimester, with low-dose aspirin administered in the presence of abnormal uterine artery findings. The studies involved a combined total of 4993 women and were of good quality, but the review remained underpowered to detect clinically important differences in serious maternal and neonatal morbidity/mortality. There was no significant difference identified in the pooled estimate of the intervention effect for the primary outcomes (‘any perinatal death after randomization’ and ‘hypertensive disorders’) or for the range of secondary outcomes. As yet, there are no published randomized studies assessing Doppler ultrasound of uterine arteries in the first trimester or in high-risk pregnancies. Updated systematic reviews1-3 using more stringent methodology and quality assessment confirmed previous results4, 5. Use of umbilical artery Doppler in singleton high-risk pregnancies showed a reduction in perinatal death without an increase in obstetric intervention. Nevertheless, the optimum frequency of Doppler assessments and the best timing of delivery in the presence of abnormal Doppler findings remain elusive. The role of umbilical Doppler assessment in other high-risk groups such as post-term, diabetic and dichorionc twin pregnancies remain debatable. On the other hand, there is, at present, no good-quality evidence to support routine umbilical artery Doppler ultrasound, or a combination of umbilical and uterine artery Doppler ultrasound in low-risk or unselected populations2. Similarly, there is insufficient evidence to support a screening package that comprises ‘routine’ uterine artery Doppler followed by low-dose aspirin given to women who have abnormal Doppler findings (i.e. screen-positive women)3. The main difficulty for three of these reviews is the lack of power to test the hypotheses related to rare events such as severe perinatal and/or maternal morbidity. Clearly, setting up large and well-designed, multicenter clinical trials is a difficult task. Data from some observational studies suggest that the assessment of other fetal vessels, such as the ductus venosus, could be more useful for clinical decision-making. Similarly, increasing numbers of observational studies are reporting possible benefits of first-trimester uterine artery Doppler as a potential predictor of pregnancy complications, particularly when combined with maternal biochemistry. Again, rigorous, large clinical trials are needed to show the benefits of this test when managing pregnancy. It is important to explore new avenues for preventative strategies in obstetric care, and Doppler ultrasound will remain the focus of interest for years to come. However, we should not forget that any intervention may bring not just benefit but also harm. Making sure that all future prospective studies collect data not only on effectiveness, but also on safety, cost and women's views would be both a significant contribution to the quality of evidence and of great help to the women and clinicians who have to rely on them.