Methods:
This retrospective cohort study included women with a monochorionic diamniotic twin pregnancy who underwent laser surgery for the management of TTTS, between 2012 and 2018 at Baylor College of Medicine in Houston, Texas and between 2002 and 2017 at the University Of Maryland School Of Medicine in Baltimore, Maryland. Permission was obtained from Institutional Review Boards of both institutions (H-19834, H-36458 and HP-00040715-12). Laser surgery was performed when Quintero stage was II to IV, or in Quintero stage I associated with symptomatic polyhydramnios, cervical shortening, or preterm labor.
Laser photocoagulation of placental anastomoses was performed using a single trocar under either local anesthetic with intravenous sedation, regional, or general anesthesia. The trocar was introduced percutaneously using the Seldinger method at one participating institution, whereas direct trocar entry was used at the other institution. The placental vascular anastomoses were identified, and laser was used for photocoagulation under direct fetoscopic visualization. Amnioinfusion (using Ringer’s lactate or normal saline solutions) and/or amnioreduction were used at the discretion of the surgeon(s). After laser surgery, most patients returned to their referring institutions to continue with their antenatal care.
Absolute intertwin differences in umbilical artery PI was estimated prior to laser surgery by subtracting the lower pulsatility index (PI) of the UA of either twin from the PI of the UA of the co-twin, as previously reported9. TTTS cases were stratified using a DUAPI cutoff value of 0.4 derived from a receiving characteristic curve analysis to evaluate double infant survival to 30 days of life. In order to determine which component of DUAPI is associated with infant survival, separate ROC curve analyses were done for the UA PI of the donor and UA PI of the recipient to predict double survival to 30 days of life in the whole study cohort. Since UA PI changes with gestational age, correlation analyses were done using spearman’s rho tests to evaluate the relationship between DUAPI and gestational age at ultrasonography as well as between UA PI of the donor or the recipient twin with gestational age at ultrasonography.
Double infant survival and survival of at least one twin to one month of age was compared between women with a DUAPI ≥0.4 and those with a DUAPI <0.4 in the whole cohort and then separately in women with TTTS Quintero stage I or II combined as well as in women with TTTS Quintero stage III or IV combined. The rationale for this is that Quintero stages I and II differ only on the visualization of the fetal bladder of the donor twin; whereas fetal hydrops is the main feature differentiating stage III and IV. Moreover, besides fetal echocardiography, there is currently no other method to further stratify TTTS cases with Quintero stages I or II in order to predict infant survival prior to laser surgery. We previously reported that among TTTS with absent or reversed UA EDF, intermittent UA Doppler abnormalities are associated with higher infant survival when compared to persistent UA Doppler abnormalities9.
Stepwise backwards conditional regression analyses were performed to evaluate the association of DUAPI <0.4 and double infant survival using a DUAPI of ≥0.4 as a reference, adjusted for gestational age (GA) at surgery, GA at delivery, Quintero stage, selective fetal growth restriction (sFGR) (estimated fetal weight or birthweight discordance ≥25% and <10th percentile of one twin)18, maternal age ≥35 years old, BMI>35, placental location (anterior, posterior or lateral with an anterior or posterior component), use of Seldinger method to place the operative trocar, size of the vascular trocar, participating center, use of Solomon technique, laparoscopic-assisted procedure, cerclage and preterm prelabor rupture of membranes (PPROM). These regression analyses were done in the whole cohort and then separately in women with TTTS Quintero stage I or II combined as well as in women with TTTS Quintero stage III or IV combined. To evaluate if the sonographic components used in Quintero staging are independently associated with infant survival, a separate regression analysis was performed including the above-mentioned co-variables in addition to the following ones: maximum vertical pocket (MVP) >8 cm in the recipient twin, MVP <2 cm in the donor twin, absent or reversed UA end diastolic flow (EDF) in either twin, absent or reversed flow during atrial contraction in the ductus venosus of either fetus, and fetal hydrops in either fetus defined as the sonographic finding of two or more of the following: skin edema, ascites, pleural or pericardial effusion. In the regression analyses evaluating the sonographic criteria used in the Quintero classification, Quintero stage was excluded to avoid multicollinearity. Since the MVP cut-off used in the Quintero classification was not derived from an ROC curve analysis,7 additional regression analyses were done to evaluate the association of the MVP of both donor and recipient twins (as continuous variables) with infant survival adjusting for the above-mentioned confounders.
Comparisons were done using Pearson Chi-square, Fisher Exact, and Mann-Whitney U tests. P<0.05 was considered significant. All statistical analysis was performed using IBM Statistical Package for Social Sciences (SPSS version 24).