Perinatal outcome after selective fetal reduction in monochorionic twin pregnancies: A comparison of techniques over a 20‐year period

To assess the perinatal outcome after fetal reduction in complicated monochorionic (MC) twin pregnancies by comparing different techniques.

� To date, there is no international consensus on which technique for which the indication is ideal and has the best perinatal outcome.

What does this study add?
� Selective fetal reductions in monochorionic twins have the lowest mortality rate of the cotwin after bipolar cord coagulation and the highest after interstitial laser coagulation.
� High preterm pre-labor rupture of membranes rates before 32 weeks gestational age was seen irrespective of the technique.
� The selection of the most ideal technique should be made per pregnancy, considering gestational age, indication and feasibility.

| INTRODUCTION
Monochorionic (MC) twin pregnancies carry a higher risk of perinatal morbidity and mortality than dichorionic pregnancies due to placental sharing and vascular anastomoses. 1,2 Unique complications such as twin-to-twin transfusion syndrome (TTTS), twin anemia polycythemia sequence (TAPS), selective fetal growth restriction (sFGR), twin reversed arterial perfusion (TRAP) and discordance for structural anomalies can occur. 3,4 These conditions may lead to fetal demise (FD) in one of the twins. Acute exsanguination can lead to fetal demise or severe cerebral injury in the co-twin in 15% and 26%-34% of the cases, respectively. 5,6 To optimize outcomes for the co-twins, selective fetal reduction can be offered. 7 In MC twins, selective reductions are carried out using several techniques that aim to immediately stop the intertwin exchange of blood. The occlusion of the umbilical cord by intrafetal procedures (interstitial laser coagulation [ILC] and radiofrequency ablation [RFA]) or coagulation of the umbilical cord (fetoscopic laser coagulation [FLC] and bipolar cord coagulation [BCC]) can prevent acute exsanguination. BCC and FLC often require the insertion of a relatively large instrument into the amniotic sac, which poses a significant risk of complications, such as premature pre-labor rupture of membranes (PPROM), hemorrhage, infections and preterm labor. 8 However, BCC or FLC can be the preferred option due to, for example, technical predilection, gestational age and size of the fetus, or the need for transection in case of mono-amniotic twins. Various studies have suggested that RFA has better outcomes due to the smaller diameter of the instrument, resulting in fewer complications. 8,9 It has been speculated that a more advanced gestational age (GA) with a larger fetus in RFA has unfavorable outcomes due to larger intrahepatic vessels that impede the ablation process, resulting in a longer period of hemodynamic instability. 10 Thus far, there is no international consensus on which technique for which the indication is superior and has the best perinatal outcome. 7,11 The primary objective of this study was to compare the perinatal outcome in complicated MC twin pregnancies treated with selective reduction using ILC, RFA, FLC, or BCC.

| METHODS
In this 20-year retrospective cohort study, obstetric, perinatal and neonatal data between June 2000 and December 2020 from the Leiden University Medical Center, the national referral center for fetal therapy, were evaluated. All consecutive MC twin pregnancies that underwent selective fetal reduction were included. Monoamniotic twin pregnancies were also included in this analysis. Triplets and higher order pregnancies were excluded. Selective fetal reductions are performed before 24 weeks' GA at our center abiding by the Dutch Abortion Law. Data including obstetric history, treatment details, delivery information and neonatal outcomes were extracted from the medical records. We contacted the referring clinicians to collect the relevant data for patients who did not deliver at our center. In this study, we compared four techniques for selective reduction: ILC, RFA, FLC and BCC. Some of the cases included in this study were previously reported in other publications from our group. 12,13 This study was approved by the ethics committee of our center (G21.018).

| Indications for selective fetal reduction
The following complications in MC twin pregnancies were amenable for selective fetal reduction: TRAP, TTTS, sFGR, TAPS and discordant congenital anomalies. TRAP was defined as the partial or complete absence of development of the heart in one twin, and the presence of a pump-twin that provides circulation through vascular anastomoses for the acardiac co-twin. 14 TTTS was defined as oligohydramnios in the donor sac with a maximum vertical pocket of ≤2 cm, and ≥8 cm in the recipient sac or ≥10 cm after 20 weeks gestation and was staged according to Quintero. 15 Selective fetal reductions in this group were performed when laser coagulation of vascular anastomoses was technically not feasible or in the presence of severe congenital anomalies. sFGR was defined as a growth restriction in one fetus with an estimated fetal weight <10th centile and intertwin estimated fetal weight discordance of ≥25%, and was categorized according to Gratacos. 16 Discordant congenital anomalies include malformations of the circulatory system, nervous system, respiratory system, van HOEK ET AL. digestive system, urinary system, musculoskeletal system, and face and neck observed on ultrasound examination. 17 TAPS was defined as discordant middle cerebral artery peak systolic velocity (MCA-PSV) (>1.5 multiples of the median (MoM) in donors and <1.0 in recipients) without signs of oligo-and polyhydramnios and was classified into stage I-V. 18 In 2018, delta MCA-PSV >0.5 MoM was implemented as a new diagnostic criterion for TAPS. 19

| Techniques
All patients were counseled about possible risks of the procedure, such as fetal demise, PPROM, preterm delivery and intrauterine infection, against expectant management or other interventions.
The procedures were performed by experienced fetal surgeons.
Patients received intravenous antibiotics prior to the procedure (cefazolin) and a tocolytic agent (indomethacin) before and after the procedure. Local anesthetics and sedation were chosen appropriately. In the case of oligohydramnios and BCC or FLC, amnioinfusion was performed prior to the procedure to expand the amniotic sac.

| Intrafetal techniques
ILC was performed using an 18-or 19-G needle that was inserted into the fetal abdomen aiming for the umbilical vein. Vessels were ablated under ultrasound guidance using a Diode or Nd:YAG laserfiber in short bursts until the cessation of blood flow was confirmed with color Doppler.
In RFA, a 17G needle tip was positioned in the fetal abdomen aiming for the umbilical vein in the liver. The applied energy generated changes in the alternating current between 200 and 1200 kHz until the target temperature of 100°C was reached. At least two cycles were performed, and additional cycles if necessary, until the cessation of blood flow was confirmed on color Doppler.

| Umbilical cord coagulation
In FLC, the fetoscope or embryoscope was inserted through an introduction sheath. The sizes of the instruments were a 1.0 mm embryoscope through a 5.6 Fr sheath, a 1.3 mm scope through an 8 Fr sheath, or a 2.0 or 3.0 mm scope through a 10 Fr sheath. Using a Diode or Nd:YAG laserfiber, the umbilical cord was ablated either close to the fetal abdomen or at the cord insertion at the placenta until cessation of blood flow was confirmed with color Doppler. Cord transection was performed using the laserfiber after cord coagulation in mono-amniotic twins.
In BCC, bipolar forceps were inserted through a 10 Fr sheath into the amniotic cavity of the targeted fetus. The cord was grasped and occluded at multiple sites under continuous ultrasound guidance.
Complete cessation of flow was confirmed afterward with color Doppler. In case of mono-twins, cord transection was performed with a fetoscopic laser using the previously inserted sheath.
The following day, ultrasound examination was performed and the patients were discharged home. When fetal anemia was suspected, Doppler interrogation of the MCA was performed. After 2 and 4 weeks after the procedure, another ultrasound examination was performed to assess the fetal brain. In case of suspected cerebral injury, fetal MRI was followed. Eventually, the patients were referred back to their referring clinician.

| Primary and secondary outcome
The primary outcome was perinatal mortality, defined as fetal demise, termination of pregnancy, immature delivery and neonatal death

| Statistical analyses
Statistical analyses were performed using SPSS version 25.0 (IBM).
Data are presented as mean with standard deviation for normally distributed variables, frequency or percentage for categorical data, median with interquartile range or odds ratio with 95% confidence interval. Chi-square test was used with post hoc testing using the Bonferroni correction for categorical data. For nominal data, the analyses were performed using ANOVA or ANOVA Welch test, as appropriate, to compare multiple groups, and with the independent t-test to compare two groups. Multivariable logistic regression was performed to investigate the association between multiple variables and perinatal mortality. A p value of <0.05 was considered to indicate statistical significance. (25%), 85 with FLC (33%) and 81 with BCC (31%). An outline of the indications for selective reduction per technique is given in Figure 1. Table 1

F I G U R E 1
Indications per technique of 259 monochorionic twin pregnancies. Data shown as n. BCC, bipolar cord coagulation; FLC, fetoscopic laser coagulation; ILC, interstitial laser coagulation; RFA, radiofrequency ablation; sFGR, selective fetal growth restriction; TAPS, twin anemia polycythemia sequence; TRAP, twin reversed arterial perfusion; TTTS, twin-to-twin transfusion syndrome. Indications for selective reduction in the 76 pregnancies with TTTS were persistence or recurrence of TTTS after laser coagulation of the vascular anastomoses (n = 10), laser being technically not feasible (n = 23) (decided before laser treatment in 9 or during treatment in 14 pregnancies), coexistence of severe congenital anomalies (n = 22), termination on the request of the parents (n = 8), and are unreported in 13 cases. In 9 of the 39 sFGR cases, selective reduction was performed because of coexistent congenital anomalies. 9 cases with uncommon indications were placed in the "Other" group that included 7 elective reductions (on grounds of psychosocial factors, mono-amniocity or uterus didelphidus) and 2 of umbilical cord strangulation in mono-amniotic twin pregnancies. In logistic regression analysis, the odds of mortality decreased by 62% for patients who underwent FLC compared to ILC (95% CI 0.14-0.99, p = 0.048), and by 78% for BCC compared to ILC (95% CI 0.07-0.72, p = 0.012), controlling for GA at therapy and indication ( Table 2). The odds of mortality increased 5.14 times for TTTS compared with sFGR (95% CI 1.38-19.18, p = 0.015), and 7.49 times for congenital anomalies compared with sFGR (95% CI 1.93-29.05, p = 0.004), controlling for GA at therapy and technique. GA at therapy was not significantly associated with a higher mortality rate.

| Mono-amniotic twins
Of the 259 pregnancies, 38 were spontaneous mono-amniotic. In 31 cases after cord occlusion, transection of the umbilical cord was also performed. In 9 of these cases, cord occlusion was performed using BCC and a fetoscopic laser was used for transection. In the other 22 cases, the cord was coagulated and dissected with laser. Sub-analyses of the 9 cases treated with BCC followed by transection with fetoscopic laser showed no differences in perinatal mortality; hence, they are included in the BCC group. Incomplete transections (n = 4, all following FLC) were due to bleeding of the umbilical cord (n = 1), amnio-chorion dehiscence (n = 1), or complex umbilical cord entanglement (n = 2). In 7 mono-amniotic pregnancies no transection was T A B L E 1 Perinatal outcome after selective reduction per technique for 259 monochorionic twin pregnancies. Note: Data shown as mean � SD, n (%) or n/n (%).

| Incomplete procedures
In 2% (6/259) of the MC pregnancies post-operative cessation of blood flow in the umbilical cord was not achieved. Causes of incomplete procedures were complex cord entanglement (FLC), double fetal demise peri-operative (FLC), scope dislocation due to maternal vomiting (FLC), deep panniculus (ILC) and two cases treated with RFA with unknown cause. In three additional cases, reintervention was necessary to completely coagulate the umbilical cord, whereafter delivery at term was followed in all cases. The interval between the first intervention and re-intervention ranged from 1 to 7 days, and the first attempts were with FLC, ILC and RFA.

| DISCUSSION
This study describes a large cohort of complicated MC twin pregnancies that underwent selective fetal reduction at a national referral center for fetal therapy over a 20-year period. Selective fetal reductions are precarious procedures with a high risk of perinatal mortality. Our results show the lowest mortality rates after the use of BCC. The poorest outcomes were seen after ILC.
The survival rates found in our cohort correspond to Bebbington et al., 8 reporting survival rates for RFA and BCC of 71% and 85%, respectively. Similarly, Dadhwal et al. 24 reported survival rates after RFA, BCC and ILC in 71%, 75% and 50%, respectively. In our cohort, MC pregnancies complicated by congenital anomalies had lower survival after RFA (33%) than after BCC (83%). This result may be due to the small number of cases that underwent RFA, but it is also observed in the study of Bebbington et al., 8 who reported survival rates of 33% after RFA and 85% after BCC. Another possible explanation is that RFA is technically less effective in fetuses with congenital abnormalities owing to the need for intra-abdominal insertion. Overall, it seems that the outcomes in sFGR pregnancies are relatively favorable, with a survival rate of 92%. This is supported by the findings of Parra et al. 25 who only examined sFGR cases treated with BCC or laser coagulation and found a 93% survival rate.
Our hypothesis is that the favorable outcomes in sFGR may be due to the absence of disbalance in intertwin transfusion, resulting in a better condition of the remaining fetus.
PPROM is considered to be one of the most frequent complications after selective reduction. Other studies have reported an overall PPROM rate between 20% and 41%. 7,8,11,12,26 In accordance with our study, previous studies reported lower PPROM rates after RFA compared to BCC or FLC, with differences ranging from 11% to 14%. 8,11,26,27 The most reasonable explanation for this is the use of a smaller instrument in RFA, resulting in a smaller defect of the  26 reported similar rates of neurological injury in 5% of their neonates, and Lanna et al. 29 found severe neurologic morbidity after BCC in 2% of patients at follow-up between 1 and 9 years. The study by van Klink et al. 13 on long-term outcomes after selective reduction reported neurodevelopmental impairment in 7% of survivors at a median age of 54.5 months using standardized psychometric tests in all survivors, which is higher than in the general population. Standardized follow-up studies beyond the age of 5 years are needed to provide clinicians and future parents with reliable information on the long-term neurodevelopmental status of co-twin survivors.
The limitations of our study are its retrospective and observational nature. It is difficult to compare groups in this large cohort because for each patient, the technique used is carefully selected T A B L E 2 Multivariate analysis of the effect of technique on perinatal mortality, controlled for indication and GA at therapy. Abbreviations: BCC, bipolar cord coagulation; FLC, fetoscopic laser coagulation; GA, gestational age; ILC, interstitial laser coagulation; RFA, radiofrequency ablation; sFGR, selective fetal growth restriction; TAPS, twin anemia polycythemia sequence; TRAP, twin reversed arterial perfusion; TTTS, twin-to-twin transfusion syndrome.

Independent variables OR (95% CI) p-value
van HOEK ET AL.
-1033 based on indication, gestational age and feasibility. The study period of 20 years could have affected outcomes due to improvements over time in for example, neonatal care, anesthesia, instruments and a learning curve in surgeons. Furthermore, we did not have access to data on several peri-operative characteristics (e.g. presentation of the fetus) from which we know they could affect outcome. 30 Important strengths are the large cohort size and the fact that all procedures are being performed at a single national referral center by a small group of experienced surgeons. The procedures have been performed by a total of 6 surgeons, rotating over a period of 20 years.
In conclusion, our findings confirm that selective fetal reduction is a precarious procedure that has a high perinatal mortality rate with differences between techniques and indications. Further studies are required to evaluate the long-term neurodevelopmental outcome in survivors. Since selective fetal reduction can be an emotional experience for parents, research into the psychological sequelae is warranted.

ACKNOWLEDGMENTS
We thank our Fetal Therapy nurses for their valuable support in conducting this research and collecting the data. We received no financial support for the research, authorship, and/or publication of this article.