Background
Umbilical artery absent end-diastolic velocity indicates increased placental resistance and is associated with increased risk of perinatal demise and neonatal morbidity in fetal growth restriction. However, the clinical implications of intermittent vs persistent absent end-diastolic velocity are unclear.
Objective
We compared umbilical artery Doppler velocimetry changes during pregnancy and neonatal outcomes between pregnancies with fetal growth restriction and intermittent absent end-diastolic velocity and those with persistent absent end-diastolic velocity.
Study Design
In this retrospective study of singletons with fetal growth restriction and absent end-diastolic velocity, umbilical artery Doppler abnormalities were classified as follows: intermittent absent end-diastolic velocity (<50% of cardiac cycles with absent end-diastolic velocity) and persistent absent end-diastolic velocity (≥50% of cardiac cycles with absent end-diastolic velocity). The primary outcome was umbilical artery Doppler progression to reversed end-diastolic velocity. Secondary outcomes included sustained umbilical artery Doppler improvement, latency to delivery, gestational age at delivery, neonatal morbidity composite, rates of neonatal intensive care unit admission, and length of neonatal intensive care unit stay. Outcomes were compared between intermittent absent end-diastolic velocity and persistent absent end-diastolic velocity. Multivariate logistic regression was used to adjust for confounders. A receiver operating characteristic curve was generated to assess the sensitivity and specificity of the percentage of waveforms with absent end-diastolic velocity in predicting the neonatal composite. The Youden index was used to calculate the optimal absent end-diastolic velocity percentage cut-point for predicting the neonatal composite.
Results
Of the 77 patients included, 38 had intermittent absent end-diastolic velocity and 39 had persistent absent end-diastolic velocity. Maternal characteristics, including age, parity, and preexisting conditions did not differ significantly between the 2 groups. Progression to reversed end-diastolic velocity was less common in intermittent absent end-diastolic velocity than in persistent absent end-diastolic velocity (7.9% vs 25.6%; odds ratio, 0.25; 95% confidence interval, 0.06–0.99). Sustained umbilical artery Doppler improvement was more common in intermittent absent end-diastolic velocity than in persistent absent end-diastolic velocity (50.0% vs 10.3%; odds ratio, 8.75; 95% confidence interval, 2.60–29.5). Pregnancies with intermittent absent end-diastolic velocity had longer latency to delivery than those with persistent absent end-diastolic velocity (11 vs 3 days; P <.01), and later gestational age at delivery (33.9 vs 28.7 weeks; P <.01). Composite neonatal morbidity was less common in the intermittent absent end-diastolic velocity group (55.3% vs 92.3%; P <.01). Neonatal death occurred in 7.9% of intermittent absent end-diastolic velocity cases and 33.3% of persistent absent end-diastolic velocity cases ( P <.01). The differences in neonatal outcomes were no longer significant when controlling for gestational age at delivery. The percentage of cardiac cycles with absent end-diastolic velocity was a modest predictor of neonatal morbidity, with an area under the receiver operating characteristic curve of 0.71 (95% confidence interval, 0.58–0.84). The optimal percentage cut-point for fetal cardiac cycles with absent end-diastolic velocity observed at the sentinel ultrasound for predicting neonatal morbidity was calculated to be 47.7%, with a sensitivity of 65% and specificity of 85%.
Conclusions
Compared with persistent absent end-diastolic velocity, diagnosis of intermittent absent end-diastolic velocity in the setting of fetal growth restriction is associated with lower rates of progression to reversed end-diastolic velocity, higher likelihood of umbilical artery Doppler improvement, longer latency to delivery, and higher gestational age at delivery, leading to lower rates of neonatal morbidity and death. Our data support using an absent end-diastolic velocity percentage cut-point in 50% of cardiac cycles to differentiate intermittent absent end-diastolic velocity from persistent absent end-diastolic velocity. This differentiation in growth-restricted fetuses with absent end-diastolic velocity may allow further risk stratification.
Introduction
Fetal growth restriction (FGR) is a leading cause of infant morbidity, mortality, and adverse long-term health outcomes. , FGR is complicated by significant variation in diagnostic criteria and lack of effective strategies in disease prevention and management. Management of FGR is targeted at early diagnosis, fetal surveillance, and optimizing timing of delivery to minimize perinatal mortality and morbidity.
Why was this study conducted?
To determine the rates of progression to reversed end-diastolic velocity in growth-restricted fetuses with intermittent vs persistent absent end-diastolic velocity and compare the obstetrical and neonatal outcomes between growth-restricted fetuses with intermittent and those with persistent absent end-diastolic velocity.
Key findings
Growth-restricted fetuses with intermittent absent end-diastolic velocity have lower rates of progression to reversed end-diastolic velocity than those with persistent absent end-diastolic velocity. In addition, these fetuses have longer latency to delivery and decreased neonatal morbidity and mortality because of higher gestational age at delivery.
What does this add to what is known?
The percentage of cardiac cycles with absent end-diastolic velocity may allow further risk stratification in fetal growth restriction.
Umbilical artery Doppler (UAD) velocimetry measures the resistance to fetal blood flow within the placenta. Beginning around the 14th week of pregnancy, low placental resistance allows continuous forward flow in the umbilical arteries (UAs) throughout the cardiac cycle. However, myriad conditions can compromise the placental vasculature, leading to decreased, absent, or reversed end-diastolic velocity (EDV) in the UA.
In pregnancies complicated by FGR, abnormal UAD results are associated with increased risk of perinatal demise and neonatal morbidity. , UAD surveillance following FGR diagnosis is used to guide management and timing of delivery and has been shown to significantly reduce risks of perinatal death and adverse obstetrical outcomes. Because UAD results can progress over the course of a pregnancy, serial UAD assessments are recommended. , Frequency of UAD assessments, management algorithms, and delivery timing recommendations vary between major national societies in obstetrics. Absent EDV (AEDV) in the UA can be either persistent (pAEDV), occurring in most or all fetal cardiac cycles, or intermittent (iAEDV), only occurring in some of the cardiac cycles; however, there are no standardized definitions of these terms. One previous study has demonstrated that compared with fetuses with pAEDV, those with iAEDV are diagnosed with UAD abnormalities later in pregnancy and delived at later gestational age (GA) with greater birthweights. However, the natural history and clinical implications of iAEDV vs pAEDV remain unclear. Our study sought to compare changes in UAD velocimetry and pregnancy outcomes between cases of FGR with iAEDV and those with pAEDV.
Materials and Methods
We performed a retrospective cohort study of pregnancies with FGR and UAD studies demonstrating AEDV observed at the Washington University Medical Center from January 2017 to December 2020. This study was approved by the Washington University School of Medicine Human Research Protection Office (IRB #202012120).
We included fetuses with FGR, defined during the study period as an ultrasound-derived estimated fetal weight (EFW) <10th percentile on the basis of the Hadlock fetal growth curves. We included singleton fetuses with a GA of ≥23 weeks by best obstetrical estimate who underwent UAD studies that demonstrated AEDV. Cases were excluded if there were major fetal anomalies or genetic abnormalities diagnosed prenatally, if delivery or neonatal outcome data were unavailable, or if UAD studies demonstrated reversed EDV (REDV) before AEDV diagnosis.
All ultrasounds were performed by sonographers certified in obstetrics and gynecology by the American Registry for Diagnostic Medical Sonography on GE Voluson machines (GE Healthcare, Milwaukee, WI) and interpreted by a maternal-fetal medicine or genetics fellowship-trained obstetrician. UAD waveforms were obtained via transabdominal imaging from a free-floating loop of umbilical cord in the absence of fetal breathing. At least 3 separate UAD assessments were performed in each fetus. AEDV was defined as absence of blood flow within the UA at end-diastole in at least 1 fetal cardiac cycle. REDV was defined as reversal of flow in the UA in at least 1 fetal cardiac cycle. In the absence of AEDV or REDV, UA pulsatility index (PI) was calculated and reported as normal if the UA PI was ≤95th percentile for GA and as elevated if the UA PI was >95th percentile.
During the study period, the institutional management of FGR with AEDV included admission for antenatal corticosteroid administration and intensive inpatient monitoring, including UAD assessment twice weekly and antenatal testing twice daily. Delivery was recommended by GA of 34 weeks per American College of Obstetricians and Gynecologists (ACOG) recommendations or if another indication arose. If a diagnosis of AEDV or REDV was made after the recommended GA for delivery, delivery was recommended at the time of diagnosis. Maternal demographics, medical complications, antenatal history including additional ultrasound studies, delivery, and neonatal outcomes were extracted from the medical records by research staff.
To investigate our question regarding the significance of iAEDV vs pAEDV, all recorded UA waveforms from the first ultrasound for detecting AEDV were independently reviewed. The AEDV percentage was calculated by dividing the number of cardiac cycles with AEDV by the total number of fetal cardiac cycles assessed ( Figure 1 , A). Cases of iAEDV were defined as cases with AEDV percentage <50%, and pAEDV cases were defined as those with AEDV percentage ≥50%.
Our primary outcome was the progression to REDV before 34 weeks of gestation. This diagnosis was based on subsequent ultrasound reports. If no subsequent ultrasound was performed, the fetus was defined as having no progression of UAD. Secondary outcomes included improvement of UAD, defined by the presence of forward EDV in all waveforms on at least 2 subsequent ultrasounds; evidence of placental malperfusion on placental histopathology; GA at delivery; latency from time of AEDV diagnosis to delivery; indication for delivery; and a neonatal composite that included at least 1 of the following: neonatal death, hypoxic-ischemic encephalopathy (HIE), need for chest tube, need for intubation, total parenteral nutrition administration, seizures, need for ionotropic support, necrotizing enterocolitis, and intraventricular hemorrhage. Placental malperfusion was defined as extensive infarcts, intraplacental thrombi, massive perivillous or maternal floor fibrin deposition, extensive syncytial knotting, villous dysmaturity, distal villous hypoplasia, avascular villi, fetal vascular thrombi, or villous stromal-vascular karyorrhexis on placental histopathologic examination. These outcomes were compared between patients with iAEDV and those with pAEDV.
Baseline characteristics and ultrasound features were compared between groups using the chi square test, Fisher exact test, Mann–Whitney U test, or the Student t -test, as appropriate. Logistic regression was used to compare the odds of UAD progression and odds of neonatal composite between those with iAEDV and those with pAEDV and then adjusted for confounding factors, selected on the basis of biological plausibility and statistical significance in univariable analyses. Backward stepwise elimination was used to avoid overfitting. The final models included confounding variables with at least a 10% effect on the magnitude of the odds ratio (OR) associated with the exposure. Fit for the final model was assessed using the Hosmer–Lemeshow goodness-of-fit test. Kaplan–Meier curves and log-rank tests were used to compare latency to delivery and GA at delivery between groups. A receiver operating characteristic (ROC) curve was generated to assess the sensitivity and specificity of AEDV percentage at the sentinel ultrasound in predicting neonatal composite morbidity. The Youden index was used to calculate the optimal AEDV percentage cut-point for predicting the neonatal composite. Sensitivity, specificity, and the area under the ROC curve (AUC) were calculated using this cut-point.
Because we included all patients meeting the inclusion criteria during our study period, no a priori sample size estimation was performed. All tests were 2-tailed with P values <.05 considered statistically significant. Stata, version 12.1 (StataCorp, College Station, TX) was used to perform all statistical analyses.
Results
Our cohort included 77 pregnancies with FGR and AEDV. A median of 43 cardiac cycles per fetus were assessed (interquartile range [IQR], 33–66). The median percentage of cardiac cycles with AEDV observed in our cohort was 50.7% (IQR, 28.3–92.9) ( Figure 1 , B). In 38 fetuses, iAEDV was observed (median AEDV percentage, 28.0%; IQR, 17.5–34.4), and pAEDV was observed in the other 39 fetuses (median AEDV percentage, 92.9%; IQR, 56.0–100.0).
All maternal characteristics were similar between the 2 groups ( Table 1 ). The GA at diagnosis of AEDV was similar in the 2 groups, but diagnosis after 34 weeks of gestation was more common in the iAEDV group. More fetuses in the iAEDV group had subsequent UAD assessment (73.7% vs 48.7%; P =.03). All pregnancies diagnosed with AEDV before 34 weeks of gestation received at least 1 dose of betamethasone for fetal lung maturity, and the time between betamethasone administration and the subsequent UAD assessment was similar between the groups.
| Population characteristics | iAEDV (n=38) | pAEDV (n=39) | P value |
|---|---|---|---|
| Maternal age (y) | 28.9±5.0 | 29.5±6.4 | .62 |
| AMA a | 2 (5.3) | 7 (18.0) | .08 |
| BMI at delivery (kg/m 2 ) | 30.5 (26.2–37.6) | 30.6 (27.8–40.5) | .49 |
| BMI ≥30 kg/m 2 at delivery | 19 (51.4) | 23 (62.2) | .35 |
| Self-reported race/ethnicity | .76 | ||
| Black | 15 (40.5) | 17 (44.7) | |
| White | 21 (56.8) | 19 (50.0) | |
| Other | 1 (2.7) | 2 (5.3) | |
| Nulliparity | 19 (50.0) | 21 (53.9) | .74 |
| Tobacco use | 7 (18.4) | 6 (15.4) | .58 |
| Illicit drug use | 7 (18.4) | 4 (10.3) | .38 |
| Preexisting conditions | |||
| Chronic hypertension | 13 (34.2) | 13 (33.3) | .94 |
| Diabetes mellitus | 1 (2.6) | 2 (5.1) | .57 |
| Asthma | 7 (18.4) | 7 (18.0) | .96 |
| Gestational age at diagnosis (wk) | 30.1 (24.6–33.0) | 27.1 (25.0–29.4) | .09 |
| Diagnosis at >34 wk of gestation | 6 (15.8) | 0 | .01 |
| EFW <5th percentile | 20 (52.6) | 26 (66.7) | .21 |
| Oligohydramnios | 1 (3.7) | 2 (5.9) | .70 |
| Subsequent Doppler assessment before delivery | 28 (73.7) | 19 (48.7) | .03 |
| Betamethasone for fetal lung maturity | 36 (94.7) | 39 (100.0) | .15 |
| Days from betamethasone to subsequent Doppler assessment | 3 (2–4) | 3 (1–3) | .06 |
| AEDV percentage at diagnosis | 28.0 (17.5–34.4) | 92.9 (56.0–100) | <.01 |
| Total cardiac cycles analyzed at diagnosis | 52.5 (39–67) | 34 (25–54) | <.01 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
