Characteristics and outcomes of monochorionic diamniotic twin pregnancies complicated by selective fetal growth restriction with intermittent absent end diastolic flow Authors: Juliana S. Gebb, Alekhya Jampa, Violetta Bakunina, Yunliang Zhao, Kiersten Barr, Federica Bellussi, Desiree Fiorentino, Suzanne DeBari, Edward R. Oliver, Nahla Khalek, Shelly Soni Published: 2026 ABSTRACT AbstractIntroductionMonochorionic diamniotic twins with selective fetal growth restriction (sFGR) are categorized based on the umbilical artery (UA) end diastolic flow (EDF) of the growth‐restricted fetus. While persistent flow abnormalities are attributed to placental insufficiency, intermittent flow changes are often attributed to anastomoses. We hypothesized that not all intermittent flow is due to anastomotic fluctuation and sought to determine if pregnancies with intermittent forward and absent EDF may represent a unique cohort of sFGR pregnancy.MethodsRetrospective analysis of prospectively collected data from a monochorionic pregnancy registry. Records of patients with monochorionic diamniotic twins with sFGR evaluated from January 2013 to December 2024 were reviewed. sFGR was defined by the Delphi‐based consensus criteria. Pregnancies were grouped by the umbilical artery Doppler of the smaller fetus: (1) Type I: persistent forward EDF (+EDF); (2) Type II: no forward EDF (A/REDF); (3) Type IIIa: intermittent forward and absent EDF (iAEDF); or (4) Type IIIb: intermittent forward, absent, and reversed EDF (iAREDF). Pregnancy characteristics were compared between groups and outcomes were then compared in expectantly managed patients.ResultsOf 330 patients with sFGR, 114 (34.5%) had +EDF (Type I), 70 (21.2%) had A/REDF (Type II), 90 (27.3%) had iAEDF (Type IIIa), and 56 (17.0%) had iAREDF (Type IIIb). Gestational age at first evaluation was lower in those with Type II compared to Type IIIb (p = 0.003) and fetal intervention was higher in patients with any absent or reversed EDF compared to those with +EDF (p < 0.001). In expectantly managed patients (n = 189), median gestational age at delivery in weeks was higher in patients with Type I (32.5) and Type IIIa (32.4) compared to those with Type II (27.4) and Type IIIb (29.1) (Type I vs. Type II, p < 0.001; Type I vs. Type IIIa, p = 0.319; Type I vs. Type IIIb, p < 0.001; Type II vs. Type IIIa, p = 0.003; Type II vs. Type IIIb, p = 0.598; Type IIIa vs. Type IIIb, p < 0.001). Additionally, dual survival to discharge was higher in those with Type I compared to Type II (p < 0.001) but not compared to Type IIIa (p = 0.292) or Type IIIb (p = 0.262).ConclusionPatients with sFGR and intermittent absent EDF may represent a unique cohort with outcomes more similar to those with +EDF than those with intermittent forward, absent, and reversed EDF. INTRODUCTION Selective fetal growth restriction (sFGR) complicates up to 25% of monochorionic (MC) pregnancies and is characterized by unequal placental share as well as differences in vascular communications on the surface of the placenta [1, 2]. Unlike twin–twin transfusion syndrome (TTTS), there is no fetal surgical management for sFGR that optimizes the survival of both fetuses. In severe cases, patients face the devastating decision of whether to undergo surgical intervention to optimize the outcome of the larger fetus (often while losing the smaller one), or to continue expectant management that carries a high risk of significant morbidity and mortality to both fetuses [3]. The two most utilized definitions for sFGR are those proposed by the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) and those proposed by a Delphi‐based consensus criteria [4, 5]. ISUOG defines sFGR as one fetus less than the 10th percentile with a ≥25% discordance between the two fetuses. While this is a standard approach that will identify most at risk pregnancies, a guideline developed by Delphi‐based consensus expanded the criteria to be one fetus less than the 3rd percentile or at least two of four criteria including (1) estimated fetal weight <10th percentile, (2) abdominal circumference <10th percentile, (3) umbilical artery pulsatility index > 95th percentile in the smaller fetus, and (4) discordance in estimated fetal weights ≥25%. The most pressing concern in sFGR pregnancies is that death of one fetus in a monochorionic pair can lead to death or neurologic injury of the co‐twin in up to 75% of cases [6, 7, 8, 9]. This risk can be mitigated by fetal surgical intervention including selective cord occlusion or fetoscopic laser ablation of the placental anastomoses, but cord occlusion results in definite loss of the smaller fetus and laser similarly often results in growth‐restricted fetal demise [8, 10, 11, 12, 13, 14, 15]. Because of the risks of both expectant management and surgical intervention, numerous studies have sought to determine which sFGR pregnancies are at highest risk for deterioration. In 2007, Gratacos et al. proposed the classification system that divides sFGR pregnancies based on the umbilical artery Doppler waveform of the growth‐restricted fetus [16]. Pregnancies are characterized as follows: (1) Type I: persistent forward end diastolic flow; (2) Type II: persistent absent or reversed end diastolic flow; and (3) Type III: intermittent forward, absent, and/or reversed end diastolic flow. Type I pregnancies have a good prognosis with low risk of fetal loss and neurologic injury while Type II and III pregnancies have a higher risk profile [8, 17, 18, 19]. Type II pregnancies, in particular, carry a high risk of fetal loss or very early preterm delivery and have been further risk‐stratified by Chmait et al. into Type IIa and Type IIb based on abnormal middle cerebral artery or ductus venosus Dopplers in the smaller fetus (Type IIb) [20, 21]. Fetal surgical intervention is generally reserved for Types II and III. While pregnancies with intermittent forward and absent end diastolic flow in the smaller fetus have typically been grouped with those with intermittent forward, absent and reversed flow, we hypothesized that these pregnancies may have a different pathophysiology and different outcomes. The objective of this study, therefore, was to describe the characteristics of sFGR pregnancies with intermittent forward and absent EDF and compare outcomes in expectantly managed patients with intermittent forward and absent EDF versus those with persistent forward EDF, no forward EDF, and intermittent forward, absent, and reversed EDF to determine if this group may represent a unique sFGR cohort. METHODS After institutional review board approval, we conducted a retrospective analysis of prospectively collected data from a monochorionic (MC) pregnancy registry at our Level III fetal care center [22]. Written informed consent was obtained from the participants prior to enrollment in the registry. Records of patients with MC diamniotic twins diagnosed with sFGR were reviewed and those with sFGR by Delphi consensus criteria between January 2013 and December 2024 were included [4]. Patients with TTTS, twin anemia polycythemia sequence (TAPS), and structural malformations were excluded. Data was extracted from the registry and included maternal demographic characteristics, pregnancy details, fetal intervention details, and pregnancy outcome. Pregnancies were divided into groups based on the umbilical artery Doppler of the smaller fetus: (1) Type I: persistent forward EDF (+EDF); (2) Type II: no forward EDF (A/REDF); (3) Type IIIa: intermittent forward and absent EDF (iAEDF); or (4) Type IIIb: intermittent forward, absent, and reversed EDF (iAREDF). The Doppler grouping was based on initial evaluation at our center. Patients with abnormal umbilical artery Dopplers consisting of persistent or intermittent absent or reversed flow patterns and those with other evidence of complicated sFGR (oligohydramnios or markedly increased middle cerebral artery [MCA] peak systolic velocity Doppler of the growth‐restricted fetus) were offered selective cord occlusion of the FGR fetus or in some cases, laser photocoagulation of the vessels on the surface of the placenta [23]. Those that opted for expectant management generally followed with our center until around 24–26 weeks gestational age and then returned for local care with at least weekly ultrasound assessments. Delivery planning was determined by the local team in most cases and records were reviewed to determine the indication for delivery. Delivery indications were grouped as (1) planned (based solely on stable sFGR), (2) deterioration (worsening Dopplers or fetal heart rate abnormalities prompting delivery prior to planned date), or (3) other (preterm premature rupture of membranes, preterm labor, preeclampsia warranting delivery prior to planned date). Demographic characteristics included maternal age and body mass index (BMI). Pregnancy details included gestational age (GA) at consultation, intertwin estimated fetal weight (EFW) discordance, fetal Children's Hospital of Philadelphia (CHOP) cardiovascular score, deepest vertical pocket (DVP) of the appropriately grown fetus, DVP of the FGR fetus, placental cord insertion distance, presence of proximate placental cord insertions defined as less than 4 cm apart, MCA peak systolic velocity of the smaller fetus ≥1.5 multiples of the median, and absent or reversed a‐wave on ductus venosus Doppler of the smaller fetus [24]. The CHOP cardiovascular score incorporates recipient cardiovascular findings and donor umbilical artery Dopplers to quantify the cardiovascular burden of monochorionic twin pathologies, specifically TTTS [24]. A higher score reveals a higher cardiovascular burden and a score greater than or equal to 5 has been associated with an increased risk of progression of TTTS but there is also evidence that similar cardiovascular burden exists in some sFGR pregnancies even without TTTS [25]. Pregnancy management details included whether the patients opted for expectant management or had fetal intervention (SCO, laser, or termination of pregnancy) as well as gestational age at fetal intervention in patients that underwent therapy. Pregnancy outcome data included gestational age at delivery, delivery indication, birthweight, birthweight discordance, number liveborn, and number with survival to hospital discharge. Statistical analysis Continuous variables expressed as median [interquartile range, IQR] and categorical variables as n/N(%). Demographic and pregnancy characteristics were compared between groups and outcomes were then compared in expectantly managed patients using Kruskal‐Wallis, Mann‐Whitney U, chi‐square, or Fisher's exact test as appropriate. Statistical analysis was performed using Stata version 16. To account for multiple testing, a significance threshold of p < 0.008 was used for all comparisons. RESULTS Of 330 patients with sFGR, 114 (34.5%) had Type I, 70 (21.2%) had Type II, 90 (27.3%) had Type IIIa, and 56 (16.7%) had Type IIIb. All patients were seen for their initial consultation between 15 weeks 1 day and 26 weeks 3 days gestational age. In the entire cohort, 189 (57.3%) underwent expectant management, 132 (40.0%) had selective cord occlusion, 8 (2.4%) underwent laser photocoagulation of the vessels on the surface of the placenta and one (0.3%) opted for termination of pregnancy (Table 1). TABLE 1Pregnancy characteristics in all patients diagnosed with sFGR based on the umbilical artery Doppler of the FGR fetus. Entire cohort n = 330 Type I: All forward EDF (+EDF) n = 114 Type II: No forward EDF (A/REDF) n = 70 Type IIIa: Forward with intermittent absent EDF (iAEDF) n = 90 Type IIIb: Forward with intermittent absent & reversed EDF (iAREDF) n = 56 p valueGA consult (weeks)19.4 [17.5–21.3]17.8 [17.0–20.1]18.2 [17.2–20.6]20.2 [17.6–21.6]<0.001* Maternal age32 [28–35]30 [26–32]31 [27–35]30 [25–34]0.102BMI26.6 [23.7–32.1]28.0 [25.9–31.7]26.8 [23.8–31.5]27.0 [24.7–30.8]0.363EFW Discordance29 [24–35]33 [28–39]30 [25–35]33 [24.5–38.5]0.013PCI distance9.4 [6.6–12.9]8.5 [5.9–10.3]9.4 [7.0–11.9]8.5 [4.8–10.9]0.096Proximal PCI (<4 cm)10/96 (10.4%)9/62 (14.5%)6/78 (7.7%)7/40 (17.5%)0.367Cardiac score ≥56/111 (5.4%)6/68 (8.8%)6/90 (6.7%)8/55 (14.5%)0.214DVP (AGA)6.0 [4.9–6.9]6.3 [5.2–7.6]6.5 [5.1–7.8]6.7 [5.7–8.0]0.058DVP (FGR)3.1 [2.3–3.9]2.7 [2.0–3.4]2.9 [2.3–3.6]3.2 [1.9–4.2]0.026Ductus venosus a‐wave (FGR)0.003* Forward99 (86.8%)45 (64.3%)73 (81.1%)46 (82.1%)Absent/reversed15 (13.2%)23 (32.9%)13 (14.4%)8 (14.3%)Agenesis of DV02 (2.9%)4 (4.4%)2 (3.6%)MCA PSV ≥ 1.5 MoM (FGR)15 (13.2%)18 (25.7%)12 (13.3%)7 (12.5%)0.084Fetal intervention<0.001* None98 (86.0%)24 (34.3%)48 (53.3%)20 (35.7%)SCO16 (14.0%)43 (61.4%)41 (45.6%)32 (57.1%)Laser02 (2.9%)1 (1.1%)4 (7.1%)TOP01 (1.4%)00GA at fetal intervention (weeks)19.4 [18.4–21.4]19.2 [17.6–21.1]19.6 [18.0–22.2]20.4 [18.4–22.6]0.378 Note: Results are median [IQR] or proportions.Abbreviations: AGA, appropriate for gestational age; BMI, body mass index; DVP, deepest vertical pocket; EDF, end diastolic flow; EFW, estimated fetal weight; FGR, fetal growth restriction; GA, gestational age; MCA PSV, middle cerebral artery peak systolic velocity; PCI, placental cord insertion; SCO, selective cord occlusion; TOP, termination of pregnancy.*Statistically significant value based on p < 0.008 significance level.John Wiley & Sons, Ltd. Pregnancy characteristics in all patients diagnosed with sFGR based on the umbilical artery Doppler of the FGR fetus. Entire cohort n = 330 Type I: All forward EDF (+EDF) n = 114 Type II: No forward EDF (A/REDF) n = 70 Type IIIa: Forward with intermittent absent EDF (iAEDF) n = 90 Type IIIb: Forward with intermittent absent & reversed EDF (iAREDF) n = 56 Note: Results are median [IQR] or proportions. Abbreviations: AGA, appropriate for gestational age; BMI, body mass index; DVP, deepest vertical pocket; EDF, end diastolic flow; EFW, estimated fetal weight; FGR, fetal growth restriction; GA, gestational age; MCA PSV, middle cerebral artery peak systolic velocity; PCI, placental cord insertion; SCO, selective cord occlusion; TOP, termination of pregnancy. *Statistically significant value based on p < 0.008 significance level. Demographic and pregnancy characteristics Complete cohort Table 1 details the demographic and pregnancy characteristics of the complete cohort. Median maternal age, BMI, intertwin EFW discordance, elevated CHOP cardiovascular score greater than or equal to 5, placental cord insertion distance, incidence of proximal placental cord insertions, appropriate for gestational age fetal DVP, growth‐restricted fetal DVP, and proportion with elevated middle cerebral artery peak systolic velocity (MCA PSV) in the smaller fetus were not statistically different between the four groups (all p > 0.008). The incidence of absent or reversed a‐wave on ductus venosus Doppler of the smaller fetus was significantly higher in pregnancies with Type II compared to those with Type I (p = 0.001) but not those with Type IIIa (p = 0.008) and Type IIIb (p = 0.021). Additionally, the gestational age at first consultation was lower in those with Type II compared to those with Type I (p = 0.002) and Type IIIb (p = 0.001) and fetal intervention was higher in patients with Type II, Type IIIa, and Type IIIb compared to those with Type I (all p < 0.001). Expectantly managed cohort Table 2 details the demographic and pregnancy characteristics of the expectantly managed cohort. Median maternal age, BMI, intertwin EFW discordance, elevated cardiovascular score greater than or equal to 5, placental cord insertion distance, incidence of proximal placental cord insertions, appropriate for gestational age fetal DVP, growth‐restricted fetal DVP, incidence of elevated MCA PSV in the smaller fetus and incidence of absent or reversed a‐wave on ductus venosus Doppler in the smaller fetus were not statistically different between the four groups (all p > 0.008). The gestational age at first consultation was lower in those with Type II compared to those with Type IIIb (p = 0.003) but not compared to those with Type I (p = 0.033) or Type IIIa (p = 0.256). TABLE 2Pregnancy characteristics in expectantly managed patients diagnosed with sFGR based on the umbilical artery Doppler of the FGR fetus. Expectantly managed cohort n = 189 Type I: All forward EDF (+EDF) n = 98 Type II: No forward EDF (A/REDF) n = 24 Type IIIa: Forward with intermittent absent EDF (iAEDF) n = 47 Type IIIb: Forward with intermittent absent & reversed EDF (iAREDF) n = 20 p valueGA consult (weeks)19.6 [17.6–21.4]17.8 [16.5–20.6]18.6 [17.3–21.2]21.3 [20.1–23.2]<0.008* Maternal age32 [27–35]30 [27–32]30 [25–35]31 [27–34]0.477BMI26.5 [23.6–31.0]28.3 [25.9–30.0]27.6 [24.7–33.6]28.2 [25.1–31.8]0.322EFW Discordance28 [23–34]31 [23–37]26 [21–32]30 [23–36]0.427PCI distance9.4 [6.6–12.9]8.5 [5.9–10.3]9.4 [7–11.9]8.5 [4.8–10.9]0.179Proximal PCI (<4 cm)9/82 (11.0%)4/20 (20.0%)3/40 (7.5%)3/16 (18.8%)0.402Cardiac score ≥52/96 (2.0%)2/23 (8.7%)1/47 (2.1%)0/190.281DVP (AGA)6.0 [5.0–6.8]6.1 [4.7–7.8]6.0 [5.0–8.0]6.5 [6.1–8.1]0.194DVP (FGR)3.3 [2.5–4.3]2.9 [2.4–3.5]3.0 [2.4–3.8]3.9 [3.3–5.3]0.027Ductus venosus a‐wave (FGR)0.329Forward88 (89.8%)18 (75%)39 (83.0%)18 (90.0%)Absent/reversed10 (10.2%)5 (20.8%)7 (14.9%)1 (5.0%)Agenesis of DV01 (4.2%)1 (2.1%)1 (5.0%)MCA PSV ≥ 1.5 MoM (FGR)11 (11.2%)4 (16.7%)3 (6.4%)3 (15%)0.482 Note: Results are median [IQR] or proportions.Abbreviations: AGA, appropriate for gestational age; BMI, body mass index; DVP, deepest vertical pocket; EDF, end diastolic flow; EFW, estimated fetal weight; FGR, fetal growth restriction; GA, gestational age; MCA PSV, middle cerebral artery peak systolic velocity; PCI, placental cord insertion.*Statistically significant value based on p < 0.008 significance level.John Wiley & Sons, Ltd. Pregnancy characteristics in expectantly managed patients diagnosed with sFGR based on the umbilical artery Doppler of the FGR fetus. Expectantly managed cohort n = 189 Type I: All forward EDF (+EDF) n = 98 Type II: No forward EDF (A/REDF) n = 24 Type IIIa: Forward with intermittent absent EDF (iAEDF) n = 47 Type IIIb: Forward with intermittent absent & reversed EDF (iAREDF) n = 20 Note: Results are median [IQR] or proportions. Abbreviations: AGA, appropriate for gestational age; BMI, body mass index; DVP, deepest vertical pocket; EDF, end diastolic flow; EFW, estimated fetal weight; FGR, fetal growth restriction; GA, gestational age; MCA PSV, middle cerebral artery peak systolic velocity; PCI, placental cord insertion. *Statistically significant value based on p < 0.008 significance level. Outcome characteristics in expectantly managed patients A total of 189 patients opted for expectant management including 98 (51.9%) with Type I, 24 (12.7%) with Type II, 47 (24.9%) with Type IIIa, and 20 (10.6%) with Type IIIb. Their outcome characteristics and survival information are summarized in Tables 3 and 4. TABLE 3Outcome data in expectantly managed patients diagnosed with sFGR based on the umbilical artery Doppler of the FGR fetus. Expectantly managed cohort n = 189 Type I: All forward EDF (+EDF) n = 98 Type II: No forward EDF (A/REDF) n = 24 Type IIIa: Forward with intermittent absent EDF (iAEDF) n = 47 Type IIIb: Forward with intermittent absent & reversed EDF (iAREDF) n = 20 p valueGA delivery (weeks)32.5 [30.5–34.2]27.4 [23.7–33.2]32.4 [30.6–34.0]29.1 [27.0–32.0]<0.001* AGA birthweight (g)1755 [1519–2120]1510 [800–1980]1814 [1290–2070]1216 [1010–1550]<0.001* FGR birthweight (g)1265 [1131–1605]1060 [675–1353]1272 [940–1510]810 [737–1160]<0.001* Birthweight discordance (%)26.2 [16.5–33.6]31.8 [20.6–42.9]27.4 [22.1–37.6]25.5 [18.6–29.3]0.381Delivery indication0.018Planned41 (41.8%)6 (25.0%)14 (29.8%)4 (20.0%)Deterioration32 (32.7%)15 (62.5%)25 (53.2%)14 (70.0%)Other20 (20.4%)3 (12.5%)5 (10.6%)1 (5.0%)Unknown5 (5.1%)03 (6.4%)1 (5.0%)Livebirth0.15105 (5.1%)5 (20.8%)3 (6.4%)2 (10.0%)13 (3.1%)2 (8.3%)2 (4.3%)1 (5.0%)288 (89.8%)17 (70.8%)42 (89.4%)17 (85.0%)LTFU2 (2.0%)000Survivors to discharge0.006* 05 (5.1%)6 (25.0%)5 (10.6%)2 (10.0%)15 (5.1%)5 (20.8%)2 (4.3%)2 (10.0%)285 (86.7%)13 (54.2%)40 (85.1%)16 (80.0%)LTFU3 (3.1%)000 Note: Results are median [IQR] or proportions.Abbreviations: AGA, appropriate for gestational age; EDF, end diastolic flow; FGR, fetal growth restriction; GA, gestational age; IQR, interquartile range; LTFU, lost to follow‐up.*Statistically significant value based on p < 0.008 significance level.John Wiley & Sons, Ltd. Outcome data in expectantly managed patients diagnosed with sFGR based on the umbilical artery Doppler of the FGR fetus. Expectantly managed cohort n = 189 Type I: All forward EDF (+EDF) n = 98 Type II: No forward EDF (A/REDF) n = 24 Type IIIa: Forward with intermittent absent EDF (iAEDF) n = 47 Type IIIb: Forward with intermittent absent & reversed EDF (iAREDF) n = 20 Note: Results are median [IQR] or proportions. Abbreviations: AGA, appropriate for gestational age; EDF, end diastolic flow; FGR, fetal growth restriction; GA, gestational age; IQR, interquartile range; LTFU, lost to follow‐up. *Statistically significant value based on p < 0.008 significance level. TABLE 4Outcome data in expectantly managed patients diagnosed with sFGR with either intermittent forward and absent flow (iAEDF) or forward, absent and reversed end diastolic flow (iAREDF). Expectantly managed type III cohort n = 67/189 Type IIIa: Forward with intermittent absent EDF (iAEDF) n = 47 Type IIIb: Forward with intermittent absent & reversed EDF (iAREDF) n = 20 p valueGA delivery (weeks)32.4 [30.6–34.0]29.1 [27.0–32.0]<0.001* AGA birthweight (grams)1814 [1290–2070]1216 [1010–1550]<0.001* FGR birthweight (grams)1272 [940–1510]810 [737–1160]0.006* Birthweight discordance (%)27.4 [22.1–37.6]25.5 [18.6–29.3]0.235Delivery indication0.723Planned14 (29.8%)4 (20.0%)Deterioration25 (53.2%)14 (70.0%)Other5 (10.6%)1 (5.0%)Unknown3 (6.4%)1 (5.0%)Livebirth0.83203 (6.4%)2 (10.0%)12 (4.3%)1 (5.0%)242 (89.4%)17 (85.0%)LTFU00Survivors to discharge0.65505 (10.6%)2 (10.0%)12 (4.3%)2 (10.0%)240 (85.1%)16 (80.0%)LTFU00 Note: Results are median [IQR] or proportions.Abbreviations: AGA, appropriate for gestational age; EDF, end diastolic flow; FGR, fetal growth restriction; GA, gestational age; IQR, interquartile range; LTFU, lost to follow‐up.*Statistically significant value based on p < 0.008 significance level.John Wiley & Sons, Ltd. Outcome data in expectantly managed patients diagnosed with sFGR with either intermittent forward and absent flow (iAEDF) or forward, absent and reversed end diastolic flow (iAREDF). Expectantly managed type III cohort n = 67/189 Type IIIa: Forward with intermittent absent EDF (iAEDF) n = 47 Type IIIb: Forward with intermittent absent & reversed EDF (iAREDF) n = 20 Note: Results are median [IQR] or proportions. Abbreviations: AGA, appropriate for gestational age; EDF, end diastolic flow; FGR, fetal growth restriction; GA, gestational age; IQR, interquartile range; LTFU, lost to follow‐up. *Statistically significant value based on p < 0.008 significance level. Median gestational age at delivery in weeks was higher in patients with Type I (32.5) and Type IIIa (32.4) compared to those with Type II (27.4) and Type IIIb (29.1) (Type I vs. Type II, p < 0.001; Type I vs. Type IIIa, p = 0.317; Type I vs. Type IIIb, p < 0.001; Type II vs. Type IIIa, p = 0.003; Type II vs. Type IIIb, p = 0.590; Type IIIa vs. Type IIIb, < 0.001). Delivery for fetal deterioration prior to planned delivery date occurred in 32.7% of patients with Type I, 62.5% of patients with Type II, 53.3% of patients with Type IIIa, and 70.0% of patients with Type IIIb. While it appeared that pregnancies complicated by Doppler abnormalities were more likely to be delivered prior to planned for evidence of fetal deterioration, the finding did not reach statistical significance based on the multiple comparisons (p = 0.018). Birthweight data were available for 161 (85.2%) appropriately grown fetuses and 154 (81.5%) growth‐restricted fetuses. Birthweight of the larger fetus was higher in patients with Type I and Type IIIa compared to those with Type IIIb (Type I vs. Type II, p = 0.055; Type I vs. Type IIIa, p = 0.823; Type I vs. Type IIIb, p < 0.001; Type II vs. Type IIIa, p = 0.120; Type II vs. Type IIIb, p = 0.468; Type IIIa vs. Type IIIb, < 0.001). Similarly, birthweight of the smaller fetus was larger in pregnancies with Type I and Type IIIa compared to those with Type IIIb (Type I vs. Type II, p = 0.020; Type I vs. Type IIIa, p = 0.238; Type I vs. Type IIIb, p < 0.001; Type II vs. Type IIIa, p = 0.107; Type II vs. Type IIIb, p = 0.770; Type IIIa vs. Type IIIb, p = 0.007). The discordance in birthweights did not differ among groups (p = 0.381). While there was no difference in proportion of livebirths among groups (p = 0.151), dual survival to hospital discharge was significantly higher in pregnancies with Type I sFGR compared to those with Type II (86.7% vs. 54.2%; p < 0.001) but not compared to those with Type IIIa (86.7% vs. 85.1%; p = 0.492) and those with Type IIIb (86.7% vs. 80.0%; p = 0.310). Direct comparisons between patients with Type IIIa versus those with Type IIIb are shown in Table 4. Median gestational age at delivery, birthweight of the appropriately grown fetus, and birthweight of the growth‐restricted fetus were significantly higher in patients with Type IIIa compared to those with Type IIIb. Nonetheless, there was no difference in birthweight discordance, delivery indication, proportion with livebirth, and proportion with survival to hospital discharge. DISCUSSION Pregnancies with intermittent forward and absent EDF have traditionally been grouped with those with intermittent forward, absent and reversed EDF, despite the possibility that these findings may indicate different pathophysiologies [16]. In this study, we report that pregnancies with sFGR and intermittent forward and absent EDF in the smaller fetus (Type IIIa) deliver at a gestational age that is similar to those with persistent forward EDF (Type I) and significantly later than those with no forward EDF (Type II) and those with intermittent forward, absent and reversed EDF (Type IIIb). Placental pathology provides insight into the pathophysiological mechanisms of the types of sFGR. The placentas of Gratacos Type II pregnancies have few or very small anastomoses and the deterioration of the growth‐restricted fetus is similar to the deterioration in singleton pregnancies with placental insufficiency [2]. The placentas in pregnancies with Gratacos Type III sFGR, on the other hand, show a larger discordance but also have large artery–artery anastomoses that allow the fetus with the larger placental share to support the one with the smaller placental share but can also lead to large volume shifts that place the fetuses at risk for death or neurologic injury [2]. The large artery–artery anastomosis is also thought the be the mechanism behind the intermittent forward, absent, and reversed flow pattern in Gratacos Type III sFGR as the umbilical artery Doppler waveform of the larger fetus is transmitted to the waveform of the smaller fetus [26]. In singleton pregnancies, patients with umbilical artery intermittent forward and absent EDF have been shown to have a lower risk of progression and longer latency to delivery compared to those with persistent absent or persistent reversed end diastolic flow [27, 28, 29]. This is likely due to a less severe form of placental insufficiency. Similarly, some cases of monochorionic twins with sFGR and intermittent forward and absent EDF may have abnormal Dopplers due to placental insufficiency of the smaller fetus that is more severe than cases with increased pulsatility index but persistent forward EDF and less severe than cases with no forward EDF at all. Alternatively, cases with intermittent forward and absent flow may have similar pathophysiology to those with intermittent forward, absent and reversed flow, but may have a smaller artery‐artery connection leading to a lower risk of severe hemodynamic consequences. For the purposes of comparison, we separated these patients into a Type IIIa. In this study, while the gestational age at delivery was later in pregnancies with Type IIIa compared to those with Type II and Type IIIb, dual survival to hospital discharge was not different among the groups. Only pregnancies with Type I showed significantly higher dual survival to discharge versus those with Type II (86.7% vs. 54.2%; p < 0.001). Nonetheless, dual survival to hospital discharge in pregnancies with Type IIIa (85.1%) and Type IIIb (80.0%) did appear to be higher compared to pregnancies with Type II, and the lack of significant findings may be due to an inadequate sample size. Given the rare cohorts of sFGR pregnancies in the literature with strictly expectant management, we determined that attempt at power analysis would be flawed. The strength of this study is that it provides information from a relatively large cohort of expectantly managed sFGR patients at an established fetal therapy center. In addition to reporting outcomes in expectantly managed cases, we chose to include baseline characteristics in the entire sFGR cohort for two main reasons. First, we wanted to describe the unique sFGR cohort with intermittent forward and absent EDF (Type IIIa) and allow baseline comparisons with sFGR pregnancies with other Doppler findings in the largest cohort possible. Second, we wanted to point out that a large proportion of cases seen with Type IIIa sFGR (45.6% in our cohort) undergo cord occlusion procedures when those that are expectantly managed appear to have outcomes quite similar to those with Type I sFGR. Despite the strengths, the study has several limitations. First, although the cohort is relatively large, the number of expectantly managed patients with Type II (n = 24), Type IIIa (n = 47), and Type IIIb (n = 20) remains small thus limiting comparisons. Second, although all Doppler studies were performed and interpreted by fetal imaging experts based on the analysis of multiple waveforms in free loop as well as placental cord insertion sites, aberrant appearance of intermittent absent flow can sometimes be due to issues with a wall filter setting that is too high. While we believe that this would not be an issue in the vast majority of our images, we cannot completely exclude this factor particularly in earlier years of imaging. Additionally, while Gratacos originally reported that Doppler findings do not change throughout gestation, Noll et al. reported changes in 59% of pregnancies thus limiting the prognostic ability of one scan during initial consultation [8]. Unfortunately, we are not able to reliably report the final Doppler findings given most patients were ultimately followed and delivered locally. Changes in Doppler findings would have changed the grouping and may have impacted timing of delivery. Third, since delivery was at the referring hospital in most cases and delivery decisions were made by referring providers, the study lacks standardization that could be accomplished with strict delivery recommendation guidelines. Also potentially limiting standardization was the decision to use the Delphi‐based consensus definition for sFGR instead of the ISUOG definition. While this was done in an effort to use guidelines recommended by experts to promote research that uses an agreed upon definition, there continue to be many different ways to define sFGR that limit the comparability of different studies. Finally, placental pathology was not available therefore we are unable to show any differences between placental findings in each of the subsets of sFGR pregnancies. Notwithstanding, we believe that our findings are useful for counseling patients and should be the basis of future studies evaluating intermittent forward and absent EDF as a unique sFGR subset (Type IIIa). Ideally, future studies may provide further insight into the placental pathological changes seen with all the varying Doppler findings. Additionally, future research will ideally lead to therapies that can help stabilize or reverse placental insufficiency to allow optimization for both fetuses in the sFGR pair. CONCLUSION In our cohort, patients with intermittent forward and absent EDF (Type IIIa) deliver later in gestation than those with Type II and Type IIIb at a time that is similar to those with Type I sFGR. Despite this finding, dual survival to hospital discharge was not significantly higher in pregnancies with Type IIIa compared to those with Type II and those with Type IIIb sFGR. AUTHOR CONTRIBUTIONS Juliana Gebb: Writing—original draft and editing; data curation; investigation. Alekhya Jampa: Data curation. Suzanne DeBari: Data curation; investigation. Nahla Khalek: Writing—review and editing. Desiree Fiorentino: Investigation. Federica Bellussi: Writing—review and editing; investigation. Kiersten Barr: Data curation. Yunliang Zhao: Data curation. Violetta Bakunina: Data curation. Edward Oliver: Investigation; writing—review and editing. Shelly Soni: Writing—review and editing; validation. CONFLICT OF INTEREST STATEMENT The authors declare no conflicts of interest. FUNDING INFORMATION The authors received no specific funding for this work. ETHICS STATEMENT The study was reviewed and approved by the Children's Hospital of Philadelphia Institutional Review Board, IRB#09‐007085. Patient consent to have data collected and compiled in the registry was obtained. The research complies with the guidelines for human studies and the research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.