Absent or reversed end-diastolic flow in the umbilical artery is known to be associated with an increased risk of intrauterine growth restriction and perinatal mortality.1 Data from evidence-based medicine suggest that the use of umbilical artery Doppler studies in high-risk pregnancy, especially those complicated by hypertension or presumed impaired fetal growth, is associated with a trend to a reduction in perinatal deaths, fewer inductions of labor and fewer admissions to hospital.2 The availability and use of umbilical artery Doppler studies for assessment of the fetal–placental circulation in pregnant women with suspected placental insufficiency has been now included in some standard obstetric guidelines.3 However, there is controversy as to clinical usefulness of abnormal Doppler findings and absent or reversed end-diastolic blood flow in the umbilical artery in unselected preterm deliveries. In term or moderately preterm high-risk pregnancies, umbilical artery Doppler studies are useful both in predicting an abnormal neonatal outcome and in monitoring the occurrence of fetal hypoxia thus reducing the incidence of cesarean section for fetal distress.4 However, several studies show that when the severity of intrauterine growth restriction is taken into account, the prediction of perinatal morbidity based on umbilical artery velocimetry is of limited value.5–6
The association of umbilical artery Doppler studies and long-term neurodevelopmental outcome of infants is even more controversial. Although earlier investigations suggested a strong relationship between absent or reversed end-diastolic flow in the umbilical artery and poor neurodevelopmental outcome,7 this association was less evident in more recent studies.8 None of the previous studies accounted for the potential confounding effect of the severity of fetal growth failure on subsequent infant neurodevelopmental outcome.
The main purpose of this study was to evaluate the association between umbilical artery Doppler velocimetry, short-term neonatal complications and the infants' neurodevelopmental outcome at 2 years in an unselected population of consecutive pregnancies delivered between 24 and 35 weeks of gestational age. A secondary objective of the study was to identify whether the presence of fetal growth restriction modifies the risk of a poor neonatal outcome associated with abnormal results of an umbilical artery Doppler study.
MATERIALS AND METHODS
This was an historical cohort study of all pregnant women admitted at our department and delivered of a singleton live fetus between 24 and 35 weeks of gestational age who had at least one Doppler velocimetry evaluation of umbilical artery blood flow immediately before birth in the period 1994–2000. The study was approved by the institutional review board of our institution. Of a total of 625 pregnancies delivered in the period examined, 22 severely malformed fetuses, 10 pregnancies with uncertain gestational age and 11 with incomplete umbilical artery Doppler data, were excluded from the study. The remaining 582 pregnancies with at least one umbilical artery Doppler study in the week before birth constituted the study group. Maternal data and clinical variables were collected prospectively and stored in a computer database. Preeclampsia was defined according to standard criteria. Fetal growth restriction was diagnosed when abdominal circumference ultrasonographic measurements decreased to below the 10th percentile of a standard growth curve for our population; the abnormal measurement of abdominal circumference was confirmed on at least 2 ultrasonographic examinations, 2–4 weeks apart, after the standard sonogram at 18–22 weeks. To evaluate the severity of growth restriction we also computed the birth weight ratio of each infant at any given gestational age. The birth weight ratio is the ratio between the observed and the expected birth weight, where expected birth weight is the 50th percentile of birth weight of our population corrected for the sex of the infant.9 Doppler waveform examinations were obtained using pulsed Doppler ultrasound with a 3.5-MHz transducer after identification of the appropriate vessel by color flow imaging. All the examinations were performed in the 7 days preceding the delivery and usually before administration of ritodrine or steroids. Among patients on expectant management for fetal growth restriction or for preterm premature rupture of membranes Doppler studies were repeated weekly and, in any case, at least 48 hours after stopping corticosteroids or ritodrine. The angle of insonication between the Doppler beam and the direction of flow was kept at less than 30 degrees for all measurements. We evaluated the ratio of peak-systolic to end-diastolic (S/D) blood flow velocities in the umbilical artery of all the patients. All Doppler waveforms were obtained in the absence of fetal body or breathing movements. The mean of three values was used for subsequent analyses. Both fetal biometric data and Doppler studies were available to attending obstetricians for clinical decision making. To evaluate the association between an increasing resistance to blood flow in the umbilical artery and the frequency of abnormal neonatal outcomes, patients were grouped into 3 categories according to the results of the Doppler studies. The first group comprised patients with a normal (< 95th percentile of local standards) S/D ratio of umbilical artery blood flow, the second group was formed of patients with an elevated S/D ratio (≥ 95th percentile of local standards) but with end-diastolic flow still present, and the third group comprised patients with either absent or reversed end-diastolic blood flow in the umbilical artery.
After delivery, all the neonates were admitted to the neonatal intensive care unit of our department. A standard echoencephalogram was obtained from all the infants within 24 hours of admission. Subsequently serial cranial ultrasounds were carried out every 48–72 hours during the first 2 weeks of life and weekly thereafter until the infants were discharged. Infant intracranial hemorrhage was graded into four categories according to Papile et al.10 Periventricular leukomalacia was defined as both frank cystic degeneration of periventricular white matter and persistently (more than a week) increased echodensity in the periventricular regions without cystic degeneration. The infants' neurodevelopment was evaluated by a child neuropsychiatrist not involved in the intensive care of the infants and unaware of the maternal history. Examinations were performed at discharge from hospital and at 3, 6, 12, and 24 months of corrected age. The neurologic evaluations were based on the methods of Amiel-Tison and Grenier.11 The Bayley scales of infant development were used to assess cognitive development (Mental Developmental Index at 12–24 months. Infants were grouped into 4 categories according to their final examination: 1) unimpaired (normal neurological status and Bayley Mental Developmental Index > 84); 2) minor impairment (abnormalities of tone or reflexes but functionally normal or borderline [71–84] Bayley Mental Developmental Index); 3) moderate impairment (spastic diplegia or hemiplegia with a Bayley Mental Developmental Index ≥ 71); and 4) severe impairment (spastic tetraplegia and/or severe mental retardation: Bayley Mental Developmental Index < 71). Infants with moderate or severe impairments were considered to have cerebral palsy. This classification is similar to the one proposed by De Vries.12 Of the 556 surviving infants, 506 (91%) were followed up throughout the study program. The 50 infants lost to follow-up had been judged neurologically normal at discharge from hospital.
Statistical analysis was carried out using the Mann-Whitney U test or the χ2 test to compare continuous and categorical variables, respectively. Extended Mantel-Haenszel test was used to evaluate linear trends across ordered categories.13 A χ2 for heterogeneity of trend was used to test the uniformity of linear trends between the groups of pregnancies with normal and impaired fetal growth.13 Logistic regression analysis was used to compute multivariable odds ratios of neonatal outcomes, to perform tests for linear trend, and to test for interactions adjusting for potential confounders.14 For the interaction study, the categories of S/D blood flow velocities in the umbilical artery were coded as a 3-level dummy variable (0 = S/D ratio < 95th percentile of local standards, 1 = S/D ratio ≥ 95th percentile but with forward end-diastolic velocities, 2 = absent/reversed end-diastolic blood flow). Dummy codes of umbilical artery S/D ratios together with crude values of gestational age and proportion of expected birth weight were entered into the logistic models as single continuous terms whereas fetal growth restriction was entered as a categorical variable. The interaction term between umbilical artery S/D ratio and fetal growth restriction for each neonatal outcome was considered significant if its inclusion improved the model fit (P < .01) based on a likelihood ratio test.15 Since this was not a randomized study we did not compute a preliminary sample size. However, given the overall prevalence of the neonatal outcomes studied, we had a 96.5% power (α = 0.05) to detect an odds ratio of 3 or more of either neonatal death or cerebral palsy among pregnancies complicated by absent or reversed end-diastolic blood flow compared with among controls with forward end-diastolic velocities.
The rate of fetal growth restriction as diagnosed by serial ultrasonograms was 45.7% (266/582). The main maternal characteristics and complications of pregnancy according to fetal growth are reported in Table 1. As expected, pregnancies complicated by fetal growth restriction had higher rates of abnormal umbilical artery Doppler studies, preeclampsia, nonreassuring cardiotocograms, steroid administration, and cesarean deliveries. However, pregnancies with normal fetal growth were more often complicated by premature rupture of fetal membranes, placenta previa and were more likely to have been subject to ritodrine tocolysis. Although the gestational age of the two groups was similar, neonatal weight and the proportion of expected birth weight were significantly lower among pregnancies complicated by fetal growth restriction.
Table 2 reports the main neonatal complications and the neurodevelopmental outcome at 2 years of the surviving infants according to the presence of growth restriction. Overall, among 536 surviving infants followed-up to 2 years of corrected age, there were 13 cases (2.4%) of cerebral palsy (8 cases of spastic quadriplegia, 3 cases of spastic diplegia, and 2 cases of spastic hemiplegia) and 74 cases (13.8%) of mild neurodevelopmental impairment. Fetal growth restriction was not associated with any of the short or long-term neonatal outcomes studied.
Considering the entire population under study, in univariate analysis, absent/reversed umbilical artery end-diastolic blood flow was associated with an increased risk of either cystic leukomalacia or severe (grade III–IV) intracranial hemorrhage (12/88 versus 31/494, odds ratio [OR[2.36, 95% confidence interval [CI] 1.12–4.72), neonatal death (8/88 versus 18/494, OR 2.64,95% CI 1.05–6.19), cerebral palsy (5/77 versus 8/429, OR 3.65, 95% CI 1.06–12.8), and either neonatal death or cerebral palsy (13/88 versus 26/494, OR 3.12, 95% CI 1.49–6.28). After adjustment in logistic analysis for the potential confounding effects of gestational age and proportion of expected birth weight, absent/reversed end-diastolic blood flow in the umbilical artery was still significantly associated with the occurrence of cerebral palsy (OR 3.47, 95% CI 1.09–11.01, P = .035) and of either neonatal death or cerebral palsy (OR 3.08, 95% CI 1.46–6.5, P = .003).
To evaluate the association of abnormal umbilical artery Doppler studies on the main neonatal outcomes according to the presence or absence of fetal growth restriction, we first performed univariate and multivariate analyses of linear trend. The first step of the analysis was to evaluate whether increasingly abnormal umbilical artery Doppler studies were linearly related with an increased rate of a poor short-term and long-term neonatal outcome. As reported in the methods, the results of umbilical artery Doppler studies were grouped into 3 categories (S/D ratio < 95th percentile, S/D ratio ≥ 95th percentile but end-diastolic flow present, absent/reversed end-diastolic blood flow) expressing a progressive increase in the resistance to blood flow in the umbilical artery, and then a progressive reduction in the end-diastolic blood flow velocity. Table 3 reports the results of trend analysis among pregnancies without fetal growth restriction. In this subgroup, we did not find any association between increasing resistance in the umbilical artery and variables describing neonatal outcome in either univariate or multivariate analysis of linear trend. The 13 pregnancies with absent or reversed end-diastolic blood flow in the umbilical artery and normal fetal growth included 4 cases of mild to moderate placental abruption, 7 cases of severe preeclampsia, and 2 cases of maternal diabetes.
Table 4 reports the same analysis in the subgroup of pregnancies complicated by fetal growth restriction. Increasing resistance in the umbilical artery was linearly associated with a lower gestational age and with an increasing severity of fetal growth restriction as expressed by a lower proportion of expected birth weight. In univariate analysis of linear trend, the rates of overall leukomalacia, severe intracranial hemorrhage and/or cystic leukomalacia, cerebral palsy, neonatal death and either cerebral palsy or neonatal death were significantly correlated with a progressive reduction in end-diastolic blood flow velocity of the umbilical artery. The tests of departure from the linear trend were not significant in any case.
After adjustment for the potential confounding effect of gestational age and proportion of expected birth weight, multivariate analysis of the trend indicated that increasing resistance in the umbilical artery was still associated with an increasing frequency of leukomalacia, cerebral palsy, and either neonatal death or cerebral palsy. In logistic regression, after adjustment for gestational age and proportion of expected birth weight, absent or reversed end-diastolic blood flow in the umbilical artery was still associated with an increased risk of cerebral palsy (OR 6.58, 95% CI 1.19–3.64, P = .03) and either neonatal death or cerebral palsy (OR 3.2, 95% CI 1.18–8.66, P = .02).
In the subsequent analysis we sought to evaluate whether the trends in the risks of several neonatal outcomes associated with increasing resistance in the umbilical artery, as reported in Table 4, were significantly different from those of pregnancies with normal growth. In univariate stratified analysis, the risks of neonatal death (heterogeneity χ2 = 4.03, P = .045), overall leukomalacia (heterogeneity χ2 = 8.3, P = .004) and either neonatal death or cerebral palsy (heterogeneity χ2 = 7.02, P = .008) associated with increasing resistance in the umbilical artery were significantly increased among fetuses with growth restriction but not among those with normal growth. In logistic models, after adjustment for the effect of gestational age and proportion of expected birth weight, the study of first-order interaction terms demonstrated that the risk of leukomalacia (P value for interaction = .012) and either neonatal death or cerebral palsy (P value for interaction = .02) associated with increasing resistance in the umbilical artery was significant only among growth-restricted fetuses.
In our study, we found that absent or reversed end-diastolic blood flow velocity in the umbilical artery is associated with an increased risk of neonatal death or cerebral palsy in an unselected population of preterm infants. This effect was independent of gestational age and was confined to pregnancies complicated by fetal growth restriction. In particular, among growth-restricted infants, 5 of the 7 infants with cerebral palsy and 13 out of the 21 cases of either neonatal death or cerebral palsy were born from pregnancies complicated by absent or reversed end-diastolic flow in the umbilical artery. These data confirm other investigations suggesting that absent or reversed end-diastolic flow velocities in the umbilical artery of growth-restricted fetuses is associated with an increased risk of perinatal mortality and subsequent mental and motor impairment in surviving infants.16–17 In addition, univariate and multivariate analyses of linear trend suggest that among growth-restricted fetuses delivered preterm, there is a linear increase in the likelihood of adverse neonatal events with increasing resistance in the umbilical artery.
The main limitations of our study could be related to both the limited number of cases with a poor neonatal outcome (death and cerebral palsy) or to a selection bias. To increase the power of the study, after a preliminary analysis, we decided to combine neonatal outcomes with a low prevalence but with a substantial clinical impact (death, cerebral palsy). Regarding a potential selection bias, infants with subtle forms of fetal growth restriction and normal or slightly abnormal umbilical artery Doppler studies who were delivered after the 35th week of pregnancy were not included in the study. For this reason, our results should be interpreted with caution and referred only to the more severe forms of fetal growth restriction.
One of the main criticisms of previous studies exploring the relationship between umbilical artery Doppler waveform analysis and neonatal outcome has been the lack of attention to the potential confounding effects of gestational age and severity of the growth restriction.5,6 According to McCowan et al,5 the progressive reduction of umbilical artery blood flow, as measured by Doppler analysis, reflects the severity of growth restriction and does not have an independent prognostic significance when birth weight and gestational age are included in logistic models. These authors drew their conclusions from a cohort study of small-for-gestational-age infants mostly delivered at term or moderately preterm. In our study, we confirm that increasing resistance in the umbilical artery was directly correlated with the severity of growth restriction and inversely with gestational age at delivery. However, the prognostic significance of absent or reversed end-diastolic blood flow in the umbilical artery remained significant in logistic models after the inclusion of gestational age and the proportion of expected birth weight, suggesting a possible independent effect. There are several methodological differences between our study and that of McCowan et al5 that could explain this discrepancy. We studied mainly growth-restricted infants delivered preterm because of a complication of pregnancy (eg, preeclampsia) or severe fetal or placental dysfunction. In the study by McCowan et al,5 a high percentage of infants were constitutionally small infants delivered at term without fetal growth restriction or placental dysfunction.
There are several biological reasons supporting an association between abnormal umbilical artery Doppler studies, both as an independent factor or as a proxy for severity of growth restriction, and a poor neonatal outcome. A recent study18 found that neonatal intracranial volume and cerebral cortical gray matter, as measured by magnetic resonance imaging, were significantly reduced in infants born from pregnancies complicated by fetal growth restriction and abnormal umbilical artery Doppler studies compared with matched infants of the same gestational age but with appropriate intrauterine growth. A reduction of cerebral gray matter volume is in turn associated with an increased risk of neurodevelopmental impairment in the infant. Growth-restricted fetuses with absent or reversed end-diastolic blood flow in the umbilical artery are not only severely growth restricted but also metabolically compromised.19 Blood-gas analyses of specimens obtained both at cordocentesis and at birth have found a direct correlation between the severity of umbilical artery blood flow compromise and fetal hypoxemia and acidemia.19 Studies of infants born after intrauterine growth restriction suggest a significant additional risk of neurodevelopmental compromise if they were also hypoxemic/acidemic.20 Ultrastructural placental studies suggest that absent or reversed umbilical artery end-diastolic blood flow is associated with uteroplacental ischemia and a severe defect in placental perfusion.21 In a study of isolated and perfused placental cotyledons, placental hypoperfusion resulted in increased production of both interleukin 6 and tumor necrosis factor alpha.22 Increased production of these cytokines at the uteroplacental junction has been repeatedly linked to the occurrence of cerebral white matter damage and infant cerebral palsy.23
In this study, we were not able to demonstrate a significant prognostic value of abnormal Doppler velocimetry of the umbilical artery in the absence of fetal growth restriction. In particular, in the 13 pregnancies complicated by absent or reversed end-diastolic blood flow but without fetal growth restriction there were no cases of neonatal death or neurodevelopmental impairment. These pregnancies included cases of a sharp increase in umbilical artery resistance because of placental abruption, or patients with severe preeclampsia or diabetes. Interestingly, other authors24 have stressed the limitations of umbilical artery Doppler studies in pregnancies complicated by diabetes or preeclampsia but without fetal growth restriction.
In conclusion, we found that umbilical artery Doppler waveform analysis is an independent predictor of either neonatal death or cerebral palsy in growth-restricted fetuses delivered between 24 and 35 weeks of pregnancy. The prognostic value of Doppler studies of the umbilical artery in the absence of fetal growth restriction is very limited.
1. Karsdorp VH, Van Vugt JMG, Van Geijn HP, Kostense PJ, Arduini D, Montenegro N, et al. Clinical significance of absent or reversed end diastolic velocity waveform in umbilical artery. Lancet 1994;344:1664–8.
2. Neilson JP, Alfirevic Z. Doppler ultrasound for fetal assessment in high risk pregnancies (Cochrane Review). In: The Cochrane Library, Issue 2, 2000. Oxford: Update Software.
3. Gagnon R, Van den Hof Michiel. SOGC Clinical Practice Guidelines. The use of fetal Doppler in obstetrics. J Obstet Gynaecol Can 2003;25:601–7.
4. Vergani P, Andreotti C, Roncaglia N, Zani G, Pozzi E, Pezzullo JC, Ghidini A. Doppler predictors of adverse neonatal outcome in the growth restricted fetus at 34 weeks' gestation or beyond. Am J Obstet Gynecol 2003;189:1007–11.
5. Chang TC, Robson S, Spencer JA, Gallivan S. Prediction of perinatal morbidity at term in small fetuses: comparison of fetal growth and Doppler ultrasound. Br J Obstet Gynaecol 1994;101:422–7.
6. McCowan LM, Harding JE, Stewart AW. Umbilical artery Doppler studies in small for gestational age babies reflect disease severity. BJOG 2000;107:916–25.
7. Valcamonico A, Danti L, Frusca T, Soregaroli M, Zucca S, Abrami F, Tiberti A. Absent end diastolic velocity in umbilical artery: risk of neonatal morbidity and brain damage. Am J Obstet Gynecol 1994;170:796–801.
8. Kirsten GF, Van Zyl JI, van Zijl F, Maritz JS, Odendaal HJ. Infants of women with severe early pre-eclampsia: the effect of absent end-diastolic umbilical artery doppler flow velocities on neurodevelopmental outcome. Acta Paediatr 2000;89:566–70.
9. Parazzini F, Cortinovis I, Bortolus R, Zanardo V. Birth weight of infants born between the 23rd and 42nd gestational week in Italy. Pediatr Med Chir 1998;20:93–7.
10. Papile L, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weight less than 1500 gm. J Pediatr 1978;92:529–34.
11. Amiel-Tison C, Grenier A. Evaluation neurologique du nouveau-ne et du nourisson. Paris, France: Masson; 1980.
12. De Vries LS, Dubowitz V, Dubowitz V, Kaiser A, Lary S, Silverman M, et al. Predictive value of cranial ultrasound in the newborn baby: a reappraisal. Lancet 1985;2:137–40.
13. Abramson JH, Gahlinger PM. Computer programs for epidemiologists: PEPI version 4. Salt Lake City (UT): Sagebrush Press; 2001.
14. Hosmer DW, Lemeshow S. Applied logistic regression. New York (NY): John Wiley and Sons; 1989. p. 720–4.
15. Greenland S. Modeling and variable selection in epidemiologic analysis. Am J Public Health 1989;79:340–9.
16. Vossbeck S, de Camargo OK, Grab D, Bode H, Pohlandt F. Neonatal and neurodevelopmental outcome in infants born before 30 weeks of gestation with absent or reversed end-diastolic flow velocities in the umbilical artery. Eur J Pediatr 2001;160:128–34.
17. Schreuder AM, McDonnell M, Gaffney G, Johnson A, Hope PL. Outcome at school age following antenatal detection of absent or reversed end diastolic flow velocity in the umbilical artery. Arch Dis Child Fetal Neonatal Ed 2002;86:F108–14.
18. Tolsa CB, Zimine S, Warfield SK, Freschi M, Sancho Rossignol A, Lazeyras F, et al. Early alteration of structural and functional brain development in premature infants born with intrauterine growth restriction. Pediatr Res 2004;56:132–8.
19. Steiner H, Staudach A, Spitzer D, Schaffer KH, Gregg A, Weiner CP. Growth deficient fetuses with absent or reversed umbilical artery end-diastolic flow are metabolically compromised. Early Hum Dev 1995;41:1–9.
20. Soothill PW, Ajuyi RA, Campbell S, Ross EM, Candy DCA, Snijders RM, et al. Relationship between fetal acidemia at cordocentesis and subsequent neurodevelopment. Ultrasound Obstet Gynecol 1992;2:80–3.
21. Pardi G, Cetin I, Marconi AM, Bozzetti P, Buscaglia M, Makowski EL, et al. Venous drainage of the human uterus: respiratory gas studies in normal and fetal growth retarded pregnancies. Am J Obstet Gynecol 1992;166:699–706.
22. Pierce BT, Pierce LM, Wagner RK, Apodaca CC, Hume RF Jr, Nielsen PE, et al. Hypoperfusion causes increased production of interleukin 6 and tumor necrosis factor alpha in the isolated, dually perfused placental cotyledon. Am J Obstet Gynecol 2000;183:863–7.
23. Yoon BH, Park C, Chaiworapongsa T. Intrauterine infection and the development of cerebral palsy. BJOG 2003;110(suppl):124–7.
24. Arduini D, Giacomello F. Doppler velocimetry in high risk pregnancy: other complications of pregnancy. In: Kurjak A, editor. Textbook of perinatal medicine: a comprehensive guide to modern clinical perinatology. Casterton Hall, Carnforth (UK): The Parthenon Group Ltd; 1998. p. 452–61.