This study investigated the perinatal outcome of 268 MC pregnancies undergoing 272 RFA procedures, and the overall survival rate was 73.9% in the co-twins. Postprocedural complications of RFA include thermal injuries5,8, IUFD and brain damage to the retained co-twin8,21,30 and preterm birth16,31. However, the effects of various indications on the adverse perinatal outcomes after RFA are still poorly understood. In this study, considering the large cohort size, we could analyze the pregnancy outcomes stratified by the indications of TTTS, sIUGR, TRAPs, discordant anomaly and EFR, and further investigate how the types of sIUGR, Quintero stages of TTTS, the position of foetal anomaly and the chorionicity of EFR contributed to the perinatal outcome after RFA.
The perinatal outcome was found to be significantly different with respect to various indications. Regarding the survival rate, the lowest was recorded in the TTTS group, whereas the highest was in the EFR and sIUGR groups, conforming to previous studies8,13,15,19,25. Kumar8 and Yinon19 reported a survival rate of 38/38 and 17/19 in the sIUGR group separately and that sIUGR as an indication for RFA had a more favourable perinatal outcome than other indications. In the current study, the poor perinatal outcome in the TTTS group and the unexpectedly lower survival rate in the MTDFA group could be explained by further subgroup analysis.
The perinatal outcomes were found to be significantly correlated with the Quintero stages of TTTS, which was worst in TTTS IV, having higher incidences of polyhydramnios and miscarriage, and selective termination for severe tricuspid insufficiency or oedema in the retained foetus. The sIUGR II subgroup was able to obtain a better survival rate than the sIUGR III subgroup, probably contributed by the higher incidence of IUFD in the subgroup sIUGR III, which involves larger artery–artery anastomosis1,25, allowing more blood exchange during ablation.
Additionally, two subgroups had high IUFD rates. As for the high rate of 40% in the subgroup of foetal anomaly in the anterior abdominal wall (gastroschisis or exomphalos), with the proximity of the target vessel and the heart of the targeted twin, we speculate that the disappearance of the cord blood ﬂow during RFA might be the consequence of the damage to the heart caused by ablation energy rather than a complete blockage of the blood flow, and subsequent exsanguination resulted in the co-twin demise. This finding provided an evidence to consider other UCO techniques in similar situations. Furthermore, an unexpected higher rate of 8/46 was found in the DCTA subgroup, with all IUFD cases occurred in those performed before 17 weeks, wherein technical difficulty was worsened by the small size of the targeted foetus. Meanwhile, IUFD did not occur in the MCTA subgroup with the same gestational age, probably because of the flexible option in any of the three foetuses, making the process technically simpler than in DCTA cases wherein only two of the three foetuses could be targeted. Technical difficulty and subsequent IUFD in the DCTA subgroup could be partially avoided by performing the procedure after 17 weeks.
The accurate mechanism of the co-twin demise remains unclear. Subsequent exsanguination caused by an incomplete blockage in the targeted vessel after ablation might be a reason4. In this study, no IUFD in the TRAPs group and the sIUGR II subgroup could be an indirect evidence for this presumption. The blood flow in the targeted vessels, completely from the donor twin or mainly from the retained larger co-twin, was not prone to be affected by the output from the reduced twin8. Blood flow disappearance indicated a complete blockage; hence, postprocedural exsanguination was avoided.
The incidence of brain damage and neurodevelopmental impairment in the reserved co-twin was reported to be 2/103 in a systematic review30 and 5/74 in an observational cohort study after selective foeticide in MC pregnancies32. In this cohort, no neurodevelopmental retardation was found, except for mild ventriculomegaly in one pregnancy, and all children were observed to be doing well. The reason could be that delivery before 28 weeks was considered as an abortion in local hospitals and most of those newborns delivered before 28 weeks died without active treatment. Preterm birth as a consequence of PROM was considered as one of the main risk factors for adverse perinatal outcomes after foetal intervention16,31.
The limitation of this study lies in its retrospective nature. The initial case selection such as selective avoidance of sIUGR III and TTTS IV, leading to the small sample size in certain subgroups, may result in statistical bias. Moreover, some surgical details, such as placental penetration, entry of a twin sac and whether the reduced twin was the presenting twin, were missed. Another shortcoming is owing to the distance from the patients’ home to our hospital, most patients were followed up and delivered at local hospitals where viable preterm newborns had to be given up and most of the neurodevelopmental follow-up was merely assessed via telephone interview. Long-term follow-up with a standard cognitive and developmental scale is necessary.
Compared to previous publications, this is the largest cohort study regarding the perinatal outcomes of RFA stratified by all possible indications. Furthermore, all of the procedures were conducted by one experienced operator and the same sonographer, thereby maximally decreasing the influence of the technique and measurement bias. In conclusion, the pregnancy outcome after RFA was found to be correlated to the indication, with the lowest survival rate in TTTS IV cases and the highest IUFD incidence in foetal anomaly cases affecting the anterior abdominal wall, followed by the sIUGR III cases. Selective RFA after 17 weeks may help prevent IUFD in DCTA pregnancies.