Methods
The ESPRiT study was a large multicenter prospective cohort study that was carried out by the Perinatal Ireland Research Consortium at 8 tertiary-level obstetric units in Ireland from May 2007 through to October 2009. This study represents a secondary analysis of the ESPRiT data. Institutional review board approval was obtained at each of the study sites, and all patients who were included gave informed consent. The inclusion criteria for this study were twin pregnancies enrolled at <22 weeks’ gestation, with both twins alive at the time of enrolment and intact membranes. Confirmed or suspected major fetal structural abnormality or fetal aneuploidy led to exclusion from the study. Participants underwent ultrasound determination of fetal growth and interrogation of umbilical and middle cerebral artery Doppler scans that were performed every 2 weeks from 24 weeks’ gestation until delivery. All scans were carried out with GE Voluson Expert 730 equipment (General Electric Company, Fairfield, CT). Delivery outcome data included mode of delivery, gestational age, birthweight, and perinatal morbidity. A quality assurance system was in place whereby the sonographers were obliged to submit ultrasound images and Doppler traces to a central quality assurance committee. The ultrasound and obstetric information that related to each pregnancy that was enrolled were collected contemporaneously on an ultrasound software system (Viewpoint; General Electric Company) and uploaded to a live web-based central consolidated database.
Chorionicity was assigned at the first visit with the use of standard ultrasonographic markers such as the placental number, visualization of a lambda or T sign, thickness of the intertwin membrane, and/or fetal sex to support the diagnosis. Postnatal placental histologic examination confirmed the chorionicity to be correct in all but 1.7% of the pregnancies that were enrolled; the data were analyzed with the use of the corrected chorionicity. The mean gestational age of recruitment was 15.8 ± 3.9 weeks’ gestation, and all enrolled pregnancies were recruited at <22 weeks’ gestation. At each scheduled ultrasound visit, the study team recorded clinical obstetric events such as hypertension, preeclampsia, rupture of the membranes, gestational diabetes mellitus, antepartum hemorrhage, threatened preterm labor, admission to hospital, or the administration of antenatal corticosteroids for fetal lung maturity. Decision-making regarding the management of each pregnancy was at the discretion of the individual clinician who was charged with the care of the patient. In each of the 8 centers, the peripartum care protocol for twin pregnancies involved the routine use of regional anesthesia, continuous intrapartum fetal heart rate monitoring, immediate access to operating facilities for emergent cesarean delivery, and availability of clinicians who were skilled in operative vaginal delivery, internal podalic version, and breech extraction. Each center had tertiary level neonatal intensive care unit facilities.
Complete delivery and perinatal outcome data were recorded for all pregnancies. Where delivery was prompted solely by FGR, abnormal UA Doppler examination, or oligohydramnios (and in the absence of any other clinical complications such as ruptured membranes, onset of labor, preeclampsia, or antepartum hemorrhage), the delivery was considered “ultrasound indicated.” FGR was defined as estimated fetal weight <10th percentile. An abnormal UA Doppler scan was classified as a pulsatility index >95th percentile or absent or reversed end diastolic flow. Oligohydramnios was defined as a maximal vertical pocket <2 cm.
For these solely “ultrasound-indicated” deliveries, detection probabilities for FGR/abnormal UA Doppler scans/oligohydramnios were determined according to the interval between examinations by suppression of the alternate examination data. We examined whether a 4-week ultrasound schedule or a schedule that comprised scans every 4 week at <32 weeks’ gestation and every 2 weeks thereafter would impact the occurrence or timing of an “ultrasound-indicated” delivery.
To reduce sampling bias, 1000 random samples of every 4 weeks within-pregnancy visits were generated that yielded detection probabilities that were defined as the proportion of pregnancies with FGR, abnormal UA Doppler, or oligohydramnios that was detected on ultrasound scanning. Mean detection probabilities and associated margins of error were estimated from the sampling distribution of detection probabilities. The χ 2 test was used to determine whether mean detection probabilities (and corresponding estimated n) differed from the every 2 weeks schedule of assessments. SAS software (version 9.2; SAS Institute Inc, Cary, NC) was used for data management and statistical analysis. The endpoint of this study therefore was these detection probabilities rather than clinical pregnancy outcomes.
Results
Maternal demographics of the dichorionic pregnancies are shown in Table 1 . Overall there were 789 dichorionic twin pregnancies that completed the study protocol. The proportion of pregnancies that arose from assisted reproductive therapies was 33% (227/789 pregnancies).
| Variable | Measure |
|---|---|
| Mean maternal age, y ± SD | 33 ± 5 |
| Parity (≥1 births), n (%) | 493 (52) |
| Nonwhite ethnicity, n (%) | 49 (7) |
| Assisted reproduction, n (%) | 227 (33) |
| Previous cesarean delivery, n (%) | 76 (10) |
| Smoking, n (%) | 92 (12) |
| Body mass index, kg/m 2 ± SD | 25 ± 5 |
| Spontaneous labor, n (%) | 242 (31) |
| Preterm prelabor rupture of the membranes, n (%) | 20 (3) |
| Hypertension, n (%) | 104 (13) |
| Gestational diabetes mellitus, n (%) | 14 (2) |
| Antepartum hemorrhage, placental abruption, placenta previa, n (%) | 54 (7) |
| Gestational age at delivery, wk ± SD | 36.3 ± 2.4 |
Table 2 outlines the indications for delivery that were recorded for each dichorionic twin pregnancy. Seventy-three percent of dichorionic twin pregnancies (577/789) either labored spontaneously or underwent uncomplicated term induction of labor or a prelabor cesarean delivery; 97 women (12%) had delivery for which “obstetric complications” was the indication. This included obstetric cholestasis in 11 cases, antepartum hemorrhage in 11 cases, preeclampsia in 58 cases, and gestational hypertension in 17 cases. Sixty-six women had an “ultrasound indicated delivery.” It is important to underline that these 66 deliveries did not have any other complicating clinical issues such as preeclampsia, ruptured membranes, or antepartum hemorrhage.
| Indication for delivery | Measure (n = 789), n (%) |
|---|---|
| Fetal growth restriction b | 62 |
| Abnormal umbilical artery Doppler scan b | 8 |
| Oligohydramnios b | 22 |
| Term uncomplicated induction of labor or prelabor cesarean delivery | 335 (43) |
| Spontaneous onset of labor | 242 (31) |
| Obstetric complications | 97 (12) |
| Maternal request | 24 (3) |
| Nonreassuring cardiotocograph | 21 (3) |
| Other/indication not listed | 4 (0.005) |
a Some cases had >1 indication for delivery stated
b Ultrasound indicated delivery; in total 66 cases (12.3%) had ≥1 of the indications listed.
Table 3 describes the number of scheduled ultrasound examinations (including fetal biometry scan, Doppler scans (UA/multiple congenital anomalies and ductus venosus if either was abnormal or amniotic fluid assessments) for all dichorionic twin pairs in the study after 24 weeks’ gestation. The median number of scheduled visits for which this complete assessment was carried out was 6 per pregnancy (range, 1–9 visits).
| Variable | Measure |
|---|---|
| Pregnancies, n | 789 |
| Study visits (2-wk schedule), n | 5304 |
| Median scheduled visits per pregnancy, n (interquartile range) | 6 (5–7) |
| Scheduled visits per pregnancy, n (%) | |
| 1-3 | 41 (5) |
| 4 | 64 (8) |
| 5 | 171 (22) |
| 6 | 258 (33) |
| 7 | 173 (22) |
| 8-9 | 55 (7) |
Detection of FGR was reduced from 88-69%, and detection of abnormal UA Doppler was reduced from 82-62% when an every 4 week ultrasound schedule was simulated. Both of these reductions reached statistical significance ( Table 4 ). There was a nonsignificant trend toward a reduction in the recording of oligohydramnios with a 4-week interval between examinations. A program of 4-week examinations between 24 and 32 weeks’ gestation and scans every 2 weeks thereafter did not result in a significant reduction in the detection of ultrasound-derived triggers for delivery. Because of the sampling of visits at 4-week intervals, the detection probabilities (%) had margins of error of at most 2% for FGR, abnormal UA Doppler, and oligohydramnios. For the combined 4-week (<32 weeks’ gestation) plus 2-week (>32 weeks’ gestation) sampling schedule, the margin of error for detection probabilities was <1%.
| Frequency of scan | Ultrasound-indicated delivery (n = 66) | ||||||
|---|---|---|---|---|---|---|---|
| Fetal growth restriction or abnormal umbilical artery or oligohydramnios | Fetal growth restriction b | Abnormal umbilical artery Doppler scan c | Oligohydramnios d | ||||
| n (%) | P value | n (%) | P value | n (%) | P value | ||
| Every 2 wks | 66 (100) | 58 (88) | 54 (82) | 21 (32) | |||
| Every 4 wks (<32 wks’ gestation) + every 2 wk (>32 wks’ gestation) | 48 (73) | 54 (83) | .332 | 48 (73) | .213 | 20 (31) | .851 |
| Every 4 wks | 31 (47) | 45 (69) | .006 | 41 (62) | .011 | 15 (23) | .241 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
