Trends in ventricle size during pregnancy and its use for prediction of ventriculoperitoneal shunt in fetal myelomeningocele.

Objectives Fetal surgery for repair of open neural tube defect (ONTD) typically results in decreased need for a ventriculoperitoneal shunt (VPS). Our objectives were to determine the trend in ventricle size (VS) during pregnancy and whether VS and change in VS, as assessed by ultrasound, were predictive of the need for VPS in pregnancy with ONTD. Methods This was a retrospective analysis of prospectively collected data of consecutive pregnancies with ONTD, evaluated in a single center from January 2012 to May 2018. Two groups were identified: the first consisted of pregnancies that underwent in-utero repair (IUR) and the second those that had postnatal repair (PNR). Penalized B splines were used to determine the trend in VS, across 2-week gestational-age (GA) epochs, between 24 and 36 weeks of gestation. VS at each GA epoch and the change in VS between each GA epoch were compared between the IUR and PNR groups. To determine whether VS at any GA was predictive of VPS, receiver-operating-characteristics (ROC) curves were used and the optimal cut-off at each GA epoch was identified. Univariate analysis and multiple logistic regression were used for further analysis. Results ONTD was diagnosed in 110 fetuses, of whom 69 underwent IUR and 41 had PNR. Fetuses in the IUR group were more likely to have Chiari II malformation (100.0% vs 82.9%; P < 0.01), lower GA at delivery (34.9 +/- 3.2 vs 37.1 +/- 2.1weeks; P < 0.01) and lower rates of VPS within the first year postpartum (36.2% vs 61.0%; P= 0.02) compared with the PNR group. In both groups, VS increased steadily with GA from the initial evaluation to delivery. In the IUR group, there was a significant change in VS between the 24+ 0 to 25+ 6-week and the 26+ 0 to 27+ 6-week epochs (2.3 (95% CI, 0.4-4.1) mm; P= 0.02). There was a positive trend in the change in VS at later GAs, but this was not significant. Although there was no significant change in VS in the PNR group before 30 weeks, there was a positive trend after that time. On multivariate analysis, each week of advancing GA was associated with a mean increase of 0.74mm in VS (P< 0.0001) in both groups. VS was not associated with the level or type of lesion, but presence of Chiari II malformation was associated with a mean increase of 5.88mm (P < 0.0001) in VS in both the IUR and PNR groups. VS was modestly predictive of need for VPS in both groups, with area under ROC curves between 0.68 and 0.76 at the different GA epochs. Change in VS between the first and last measurements was also modestly predictive of the need for VPS, with better performance in the PNR group. Conclusions VS increased with advancing GA in all fetuses with ONTD, although in the IUR group this increase occurred immediately after fetal surgery and in the PNR group it occurred after 30 weeks of gestation. In-utero surgery was associated with a decreased rate of VPS and was more predictive of need for VPS than was VS. Postnatal factors resulting in increased need for VPS in the PNR group need to be assessed further. Copyright (C) 2019 ISUOG. Published by John Wiley & Sons Ltd.