Ductus venosus Doppler and postnatal outcomes in fetuses with absent or reversed end-diastolic flow in the umbilical arteries

1. Introduction 

The use and advances of Doppler velocimetry have revolutionized obstetrical practice and have permitted the in vivo study of the placental–fetal circulation, with assessment of the degree of blood flow resistance and, indirectly, of the level of fetal oxygenation [1]. Pregnancies with absent or reversed end-diastolic (ARED) flow have been associated with high perinatal mortality rates ranging from 16% to 60% [2], [3].

The high rates of prematurity and perinatal mortality observed in cases of ARED demonstrate the need for studies that assist in the decision of the most opportune time for interruption of these pregnancies, particularly between 26 and 32 weeks when the risks of complications resulting from extreme prematurity should be weighed against the risks of intrauterine hypoxia [3], [4], [5], [6]. The association between abnormal ductus venosus Doppler velocimetry and fetal acid-base status has been well-established in the literature, but studies demonstrating an association between ductus venosus flow and postnatal outcomes are still scarce, particularly those focusing on pregnancies with ARED [7], [8], [9], [10]. Maulik [11] suggested that pregnancies with absent end-diastolic flow and fetal growth restriction (FGR) at a gestational age of less than 34 weeks can be monitored as long as fetal well-being is assessed daily. Nevertheless, the presence of absent or reversed flow in the ductus venosus indicates immediate delivery.

The objective of the present study was to investigate the relationship between ductus venosus Doppler findings on the day of delivery and postnatal outcomes in pregnancies with ARED flow in the umbilical arteries.

2. Materials and methods 

A retrospective study was conducted on 103 newborns of pregnancies with ARED evaluated at the Fetal Surveillance Unit between January 1997 and December 2004. Included in the study were newborns of singleton pregnancies in which no fetal anomalies had been detected by ultrasonography or postnatal examination, who presented no premature membrane rupture, and on whom it was possible to perform fetal well-being assessment tests on the day of delivery including photodocumentation. The study was approved by the Ethics Committee for the Analysis of Research Projects of Hospital das Clínicas, University of São Paulo.

Pregnant patients at high risk for placental insufficiency were referred by the Prenatal Care Units for the assessment of fetal well-being. If ARED was diagnosed, the patient was hospitalized for bed rest and clinical monitoring until delivery. Color/pulsed Doppler studies were performed by one of the doctors of the team of the Fetal Surveillance Unit with a 3.5-MHz curved-array probe (ATL Ultramark 9 HDI, Advanced Technology Laboratories, Dee Why, Australia, and Ecocee, Toshiba, Tokyo, Japan). The high-pass filter was set at 50–100Hz. All recordings used for analysis were obtained in the absence of fetal breathing movements.

Placental and fetal evaluation consisted of arterial Doppler velocimetry (umbilical arteries and middle cerebral artery) and venous Doppler velocimetry (ductus venosus) performed at maximum intervals of 72h in all cases. In the umbilical arteries, the sample volume was positioned 3cm from the placental insertion of the umbilical cord, and in the ductus venosus at its origin from the umbilical vein. The amniotic fluid index (AFI) was assessed every 72h between the 20th and 27th week of gestation. After the 27th week, the following exams were performed daily: AFI measurement, cardiotocography and fetal biophysical profile scores. Each patient was thus examined until a reason to interrupt the pregnancy was observed. On the day of delivery, all exams were repeated and photodocumented regardless of the time that had elapsed since the previous assessment. The presence of at least one of the following indications was required to interrupt pregnancy: gestational age of 34 weeks or more or fetal maturity evaluated as described by Clements et al. [12], severe maternal disease which would be life-threatening to the patient if the pregnancy continued, late fetal heart decelerations, AFI less than 3.0, and fetal biophysical profile scores below 6. No prenatal corticosteroids were used in any case. The sonograms were analyzed by study of the A-wave and ductus venosus flow was classified as positive, absent or reversed.

All newborns were delivered by cesarean section. After delivery, the following immediate neonatal outcomes were obtained: gestational age at the time of delivery, 1- and 5-min Apgar scores, umbilical artery pH and base excess, the need for orotracheal intubation, newborn weight, and frequency of FGR. FGR was defined as a birthweight less than 10th percentile for gestational age of the curve of Alexander et al. [13]. Newborns were followed up during their stay at the hospital nursery, and the following postnatal outcomes were evaluated: hyaline membrane disease, pneumothorax, pulmonary hemorrhage, bronchopulmonary dysplasia, arterial canal persistence, septicemia, necrotizing enterocolitis, retinopathy of prematurity, thrombocytopenia, hypoglycemia, hyperglycemia, intracranial hemorrhage and the occurrence of death and its causes. The cases were classified into two groups according to the a-wave flow in the ductus venosus: group A, absent or reversed flow (20 cases); group B, positive flow (83 cases).

All variables mentioned above were submitted to comparative studies. The chi-square test, Fisher's exact tests, the Mann–Whitney U-test and regression logistic model were used to estimate the association of these variables with ductus venosus flow during the a-wave. The level of significance was set at 0.05.

3. Results 

The initial sample consisted of 106 patients. Subsequently, three patients were excluded due to the impossibility of obtaining postnatal follow-up data, two cases from group A and one case from group B. The final sample thus included 103 newborns. Mean maternal age at enrollment in the study was 30.08±6.7 years (range: 16–45) and 39/103 (37.8%) patients were nulliparous. Hypertensive syndromes were observed in 68/103 (66.0%) patients. All newborns were delivered by cesarean section. The indication for delivery was fetal biophysical profile scores less than 6 in 44 (42.7%) cases, late fetal heart decelerations in 27 (26.2%), an AFI less than 3.0 in 14 (13.6%), severe preeclampsia in 10 (9.7%), and the presence of fetal maturity in eight (7.8%).

The cases were classified into two groups according to ductus venosus Doppler flow (a-wave) on the day of delivery: group A, 20 cases (four with absent flow and 16 with reversed flow); group B, 83 cases with positive flow.

Analysis of blood flow in the umbilical arteries on the day of delivery showed a larger number of cases with reversed end-diastolic flow (15/20, 75%) in group A compared to group B (27/83, 32.5%) (P=0.001). The other Doppler velocimetry results are shown in Table 1.

Table 1. Doppler velocimetry results obtained for groups A and B

	Doppler results	Group A (absent/reversed flow)	Group B (positive flow)	P
	PI of the umbilical arteries	4.81 (1.98)	3.12 (1.26)	0.001
	PI of the MCA	1.08 (0.18)	1.18 (0.24)	0.07
	PIV of the ductus venosus	2.44 (0.55)	0.86 (0.31)	<0.001

Data are reported as mean (standard deviation). PI=pulsatility index; PIV=pulsatility index for veins; MCA=middle cerebral artery.

The mean gestational age at birth was 30 weeks in group A and 30.9 weeks in group B (P=0.23) (Table 2). The mean gestational age at the time of delivery of the 16 cases with reversed ductus venosus flow was 30.5 weeks and birthweight was 864g. Fetuses of group A presented a higher frequency of adverse outcomes (Table 2).

Table 2. Comparison of short-term postnatal outcomes between fetuses with absent or reversed end-diastolic flow and fetuses with positive flow waveforms in the ductus venosus

		Group A (absent/reversed flow)	Group B (positive flow)	p
	Gestational age at delivery, weeks (range)	30.0 (25.8–35.8)	30.9 (25–36)	.23 NS
	Birthweight, g (range)	831.2 (530–1375)	1105 (470–2350)	<.001
	FGR, % (n)	85 (17/20)	72.2 (60/83)	.14 NS
	1′ Apgar score<3, % (n)	85 (17/20)	20 (20/83)	.001
	5′ Apgar score < 7, % (n)	45 (9/20)	10.8 (9/83)	.001
	Orotracheal intubation, % (n)	100 (20/20)	48 (40/83)	.001
	NICU, % (n)	95 (19/20)	78 (65/93)	.11 NS
	Umbilical artery pH <7.20, % (n)	93.7 (15/16)	36 (22/61)	<.001
	Umbilical artery base excess <−12.0, % (n)	62 (10/16)	21.6 (13/60)	<.001
	Pulmonary hemorrhage, % (n)	40 (8/20)	16.9 (14/83)	.03
	Bronchopulmonary dysplasia, % (n)	0 (0/20)	6.9 (12/59)	.29 NS
	Hyaline membrane disease, % (n)	65 (13/20)	46 (38/83)	.12 NS
	Pneumothorax, % (n)	10 (2/20)	8.4 (7/83)	1.00 NS
	Necrotizing enterocolitis, % (n)	5.0 (1/20)	9.6 (8/83)	1.00 NS
	Sepsis, % (n)	40 (8/20)	37 (31/83)	.82 NS
	Intraventricular hemorrhage, % (n)	52.6 (10/19)	26.4 (19/72)	.02
	Hyperglycemia, % (n)	25 (5/20)	9.6 (8/83)	.12 NS
	Hypoglycemia, % (n)	85 (17/20)	56.6 (47/83)	.01
	Thrombocytopenia, % (n)	65 (13/20)	37.4 (31/83)	.02
	Retinopathy of prematurity, % (n)	42.8 (7/8)	40.0 (20/50)	.59 NS
	PAC, % (n)	60.0 (9/15)	56.6 (30/53)	.81 NS
	Postnatal death, % (n)	65 (13/20)	26.5 (22/83)	.001

FGR, fetal growth restriction; NICU, neonatal intensive care unit; ACP, arterial canal persistence; NS, not significant.

The mean frequency of postnatal death was 34% (35/103), and was significantly higher in group A (65%, 13/20) than in group B (26.5%, 22/83) (P=0.001). Analysis of the 35 cases of death showed a higher mortality rate until the seventh day of life in group A (12/13, 92.3%), with no case of death after day 28. In group B, deaths were more frequent after 28 days of life (10/22, 45.4%), with this difference being statistically significant (P=0.007).

Analysis of the autopsy reports showed a high frequency of respiratory insufficiency (30/35, 85.7%) as a consequence of pulmonary complications such as hyaline membrane disease, pulmonary condensation, shock lung and pulmonary hemorrhage. Septic shock was observed in 14.2% (5/35) of cases, and an association with cerebral edema and intracranial hemorrhage was present in 20% (7/35).

4. Discussion 

Severe placental insufficiency has been associated with a large number of adverse neonatal outcomes, which are mainly a consequence of FGR, low oxygen supply and prematurity. The decision of the ideal time to interrupt pregnancy in cases of fetuses with growth restriction and severe placental insufficiency continues to be a complex subject and observation and intervention studies are necessary to fill the gaps and to provide approaches to the follow-up of these pregnancies [8], [9], [11]. Little information exists regarding the neonatal complications that might be specific in these cases. However, preliminary evidence suggests that the study of the venous territory might improve the prediction of neonatal outcomes.

The present study evaluated the relationship between ductus venosus flow on the day of delivery and postnatal outcomes in order to obtain additional information that might assist in the decision of the best time to interrupt pregnancies with ARED. The groups were as homogenous as possible in terms of the degree of placental insufficiency, with only cases of ARED flow being included as suggested by Müller et al. [3], [9]. The similar mean gestational age of the two groups studied permitted the analysis of the association between ductus venosus Doppler findings and postnatal outcomes without interference from this parameter.

Mean newborn weight was 1052g, a value similar to those reported in the literature ranging from 908 to 1581g for cases of ARED [9], [14], [15]. A significant difference in mean newborn weight was observed between groups A and B. The lower weight observed in group A was possibly due to greater fetal clinical impairment as a result of longer exposure to or even more severe placental insufficiency and, consequently, deficient fetal nutrition. Similarly, Müller et al. [9], Baschat et al. [10] and Schwarze et al. [16] reported a lower birthweight in cases of abnormal ductus venosus flow.

The proportion of cases with a first-minute Apgar score lower than 3- and a 5-min Apgar score lower than 7 was significantly higher in group A, in agreement with the results of other studies [9], [10], [17], [18], [19], [20]. The umbilical cord blood pH was obtained in 77 cases and mean pH and base excess were significantly lower in group A. The frequency of the need for orotracheal intubation was 100% in group A and 48.1% in group B, a finding that can be explained by the lower Apgar scores and pH values observed in group A indicating more severe asphyxia. Similarly, Figueras et al. [6] reported a greater need for intubation in cases of abnormal ductus venosus flow. Müller et al. [9] also observed a 70% frequency of intubation in the group with absent or reversed ductus venosus flow.

The proportion of cases with hyaline membrane disease was similar in the two groups, in contrast to the study of Müller et al. [9] who reported a frequency of 48.4%, with 80% of cases in the group with absent or reversed ductus venosus flow. Necrotizing enterocolitis was observed in 8.7% of neonates, with no significant difference between groups A and B, in agreement with the results of other studies [6], [9], [18]. In contrast, Figueras et al. [6] and Baschat et al. [10] reported a higher risk of enterocolitis for cases with abnormal ductus venosus flow. Neonatal hypoglycemia was more frequent in group A (85%), probably because of a higher incidence of asphyxia which is an important cause of hypoglycemia as a result of anaerobic glycolysis inherent to hypoxia. Thrombocytopenia was present in 42.7% of cases. Hackett et al. [21] demonstrated a 31% incidence of thrombocytopenia among 26 newborns with FGR and ARED flow. Similarly Baschat et al. [10] reported a significant increase in the risk of thrombocytopenia among newborns of pregnancies with these complications, suggesting a complex vicious cycle of placental insufficiency, FGR, vascular injuries and abnormal blood flow. The presence of intracranial hemorrhage could be evaluated in 91 cases by transfontanellar ultrasound or autopsy. The high frequency of intracranial hemorrhage observed in the present study was expected since various reports have shown an elevated incidence of this event in newborns of pregnancies with ARED flow, with prematurity, low weight and hypoxia being considered important causes of this condition [22], [23]. This condition has been estimated to occur in 40–45% of neonates of less than 35 weeks of age, with the highest frequency being observed between 26 and 32 weeks [24], [25], [26]. Müller et al. [9] observed a frequency of intracranial hemorrhage of 50% among cases with absent or reversed ductus venosus flow and of 13% among cases with positive flow. Bilardo et al. [4], Figueras et al. [6] and Hartung et al. [17] also demonstrated an association between abnormal ductus venosus flow and a higher frequency of intracranial hemorrhage. Schwarze et al. [16] reported a frequency of intracranial hemorrhage of 24.2% but did not observe an association between this complication and absent or reversed flow in the ductus venosus.

In the present study, neonatal mortality was 34%, a rate similar to the 34.4% reported by Cosmi et al. [26] and to the 34.6% reported by Ferrazzi et al. [19]. However, mortality was significantly higher (68%) in the group with absent or reversed ductus venosus flow. In the present study, neonatal mortality was higher than that reported in most studies, possibly due to the presence of more severe cases, the low gestational age, the exclusive inclusion of pregnancies with ARED, and the lack of use of prenatal corticosteroids for fetal pulmonary maturation. No prenatal corticosteroids were used, because no consensus exists at our service regarding their routine use due to doubts about the possibility of adverse effects in relation to their true beneficial effects, especially in small-for-gestational age newborns at that time [27], [28], [29]. Thus, we did not interfere with the conduct established at our service. The lack of application of steroids by the service was not imposed by the research protocol but provided a unique opportunity to obtain the true correlation between ductus venosus flow and postnatal outcomes. Indeed, the use of steroids might have altered the results of Doppler velocimetry in the umbilical arteries and ductus venosus [30].

The high mortality rates in group A might be reversed, at least partially, by the administration of steroids since hyaline membrane disease and pulmonary hemorrhage were common findings in those newborns who died. Further studies will be necessary to investigate the importance of the finding of absent or reversed diastolic flow in the ductus venosus within the setting of steroid treatment.

Analysis of the present results provided an understanding of the importance of the relationship between alterations in fetal venous flow and adverse perinatal outcomes. We believe that the poorer results of group A might be due to the high frequency of absent or reversed ductus venosus flow. However, further randomized studies are necessary in order to answer this question.

Future studies analyzing only cases with positive flow in the ductus venosus may provide ductus venosus pulsatility indices and prediction curves for the main adverse outcomes, indicating the most appropriate time for the interruption of pregnancies with absent end-diastolic flow and FGR, especially between 26 and 32 weeks. It should thus be possible to reduce the high incidence of mortality and severe sequelae of intrauterine hypoxia, considerably improving the long-term prognosis of these newborns.

In conclusion, the presence of absent or reversed ductus venosus Doppler flow on the day of delivery in pregnancies with ARED in the umbilical arteries is related to a higher frequency of adverse perinatal outcomes.