Middle cerebral artery pulsatility index as possible predictive marker for neonatal death in fetuses with tricuspid valve malformation

Tricuspid valve malformations (TVMs) account for approximately 1% of all congenital heart defects (CHDs)1. Tricuspid valve leaflets are deformed in TVM, with apical displacement in Ebstein's anomaly (EA) and deficient coaptation in tricuspid valve dysplasia (TVD). Intrauterine/neonatal death is a possible consequence in the management of TVM and, thus, prediction of prognosis is crucial. Other than echocardiographic criteria2, no individual fetal Doppler parameter has been proposed as a possible predictor of neonatal death in affected fetuses. Cerebrovascular resistance may be affected in utero in the presence of CHD, which is evaluated easily by fetal middle cerebral artery (MCA) pulsatility index (PI). We aimed to investigate whether MCA-PI assessment has additional value in predicting neonatal death in fetuses with TVM. We analyzed retrospectively 15 fetuses diagnosed with TVM at the University of Maryland Fetal Heart Program, Baltimore, MD, USA from 2007 to 2018. All diagnoses were confirmed by an expert (S.T.) in fetal echocardiography and later by pediatric cardiologists. Fetal echocardiographic and Doppler variables obtained at both diagnosis and final examination were MCA, umbilical artery and ductus venosus Doppler, and Simpson–Andrews–Sharland (SAS) score parameters, including cardiothoracic ratio, Celermajer index, pulmonary valve flow, arterial duct flow and right–left ventricular diameter ratio (Table S1)2. Celermajer index was calculated as follows: (right atrial area + atrialized right ventricular area)/(functional right ventricular area + left atrial area + left ventricular area). Neonates were divided into survivors and non-survivors (at 2 months after birth). The Shapiro–Wilk test was used to assess the distribution of data for normality. Groups were compared using the Student's t-test, Mann–Whitney U-test, Fisher's exact test or chi-square test, as appropriate. The Kendall correlation test was used to determine the degree of correlation in SAS score between diagnosis and final examination. Statistical analyses were performed using SPSS software version 21.0 (IBM Corp., Armonk, NY, USA). P < 0.05 was considered statistically significant. Of the 15 cases, 10 had EA and five had TVD. Data from 14 cases were included, as one EA case was terminated at 23 weeks' gestation due to intrauterine fetal loss secondary to fetal hydrops. Survivors and non-survivors did not differ according to maternal age, birth weight or gestational age at diagnosis, final examination or delivery (Table 1). There were three fetuses with a chromosomal anomaly. One of the survivors had a gain of 5q21 and a loss of 22q12, with short femora being an additional finding. Of the non-survivors, one had 1p36 deletion with no additional findings, and the other had 18q21.1 terminal deletion with persistent right umbilical vein. Figure 1a and Table 2 show the data from fetal echocardiography measurements. Mean SAS score in survivors compared with non-survivors was 0.13 ± 0.35 vs 2.83 ± 1.94 (P = 0.002) at diagnosis and 1.38 ± 1.59 vs 5.67 ± 2.33 (P = 0.001) at final examination. The predictive threshold of SAS score for neonatal death was lower at diagnosis than at final examination (Kendall's correlation coefficient, 0.811; P = 0.001). Figure 1b and Table 3 show the data regarding MCA variables. Mean MCA-PI in survivors compared with non-survivors was 1.73 ± 0.34 vs 1.71 ± 0.30 (P = 0.930) at diagnosis and 1.48 ± 0.20 vs 2.09 ± 0.42 (P = 0.004) at final examination. Cerebroplacental ratio did not differ between the groups at diagnosis (P = 0.168) but there was a trend towards statistical significance at final examination (P = 0.053). This was likely due to the study's small sample size. The SAS scoring system aims to predict perinatal death, with a cut-off of ≥ 52. Based on our findings, this cut-off was more predictive with advancing gestation. Thus, the progression of the echocardiographic findings seems more significant in predicting neonatal death. Emerging studies have shown altered MCA Doppler variables in the context of CHD. Severe left-sided obstructive defects have been associated with reduced MCA-PI3, 4. In cases of right-sided obstructive defects, MCA-PI has been found to be similar to or greater than that in healthy controls5, 6. However, studies have not investigated the role of MCA-PI in the prediction of intrauterine/neonatal death in specific types of CHD. Hemodynamic deterioration in cases of complex CHD is usually a gradual process, making the prediction of intrauterine/neonatal death difficult. MCA assessment may uncover the progression of disease in fetuses with TVM if performed at regular intervals, as our study showed a noticeable difference in MCA-PI between survivors and non-survivors at the final examination but not at the time of diagnosis. This may be explained by the gradual progression of TVM in infants at risk for neonatal death7. Tricuspid regurgitation, right–left shunting at the foramen ovale and left-sided cardiac output may, in turn, increase as a result of decreasing right ventricle outflow. This gradual worsening may be attributable to reduced diastolic velocity in the MCA in non-survivors. This implies altered flow patterns that may include steal from the cerebral circulation in diastole when the ductal flow is reversed, and/or a centrally mediated response to reduce increased left ventricular output into the cerebral circulation. Other possible mechanisms that may explain changes in MCA hemodynamics in cases of right-sided dysfunction are development of diastolic dysfunction due to septal shift, or shift of blood flow from the ductus arteriosus towards the pulmonary artery8. These fetuses may not be able to compensate for the hemodynamic consequences of TVM after delivery. Fetuses that maintain consistent MCA-PI with advancing gestation are better able to tolerate the hemodynamic implications of TVM. Despite its retrospective nature and small sample size, our preliminary study suggests that fetuses with TVM at risk for neonatal death may show a progressive pattern in utero. With advancing gestation, MCA findings may reflect the progression of the disease, as demonstrated by a significant change in MCA-PI from initial diagnosis to final examination in non-survivors. Cases with TVM require close monitoring due to the progressive nature of the disease, and MCA assessment may add value in predicting neonatal death. However, it is difficult to propose a cut-off for MCA-PI due to our small sample size. Our findings may lead to the design of a larger prospective study evaluating the importance of MCA assessment in the prediction of neonatal death in cases of TVM. Table S1 Simpson–Andrews–Sharland scoring system Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.