Prognosis of isolated congenital diaphragmatic hernia using lung-area-to-head-circumference ratio: variability across...

Congenital diaphragmatic hernia (CDH) is a severe anomaly. The observed-to-expected lung-area-to-head-circumference ratio (o/e-LHR) has been shown to provide a useful prediction of subsequent survival of fetuses with CDH in referral centers with expertise and a large caseload. However, the accuracy of o/e-LHR measurements in general fetal medicine units with less expertise is not well known. The aim of this study was to evaluate the capacity of o/e-LHR to provide a useful prediction of mortality of fetuses with CDH when the measurement is performed in fetal medicine units with different levels of expertise.

Between January 2008 and November 2013, 305 live births with expectantly managed left-sided isolated CDH were recorded in the database of the French National Center for Rare Diseases (31 centers) and followed up after birth. Centers were grouped into two categories according to their mean annual CDH caseload over the study period: large centers with an average of ≥ 14 cases and smaller centers with < 14 cases per year. The relationship between o/e-LHR and 28-day and 6-month mortality was modeled using fractional polynomials and the predictive value of o/e-LHR was quantified using the area under the receiver-operating characteristics curve. Comparisons between the two center categories were carried out. Analyses were adjusted for potential confounders such as thoracic herniation of the liver and gestational age at birth and at diagnosis.

During the study period, two large centers managed a total of 82 CDH cases and 29 smaller centers a total of 223 CDH cases. Overall, there was a significant inverse relationship between 28-day mortality rate and o/e-LHR, which decreased from 54% when o/e-LHR was 20% to 6% when o/e-LHR was 75% (P < 0.01). When the category of center was considered, adjusted associations between o/e-LHR and 28-day mortality were significantly different (P = 0.032) between large and smaller centers. The ability to predict survival at 28 days postpartum based on o/e-LHR was better in large centers; for a specificity of 0.30, the sensitivity was 0.71 in large centers and 0.55 in smaller ones. The results were similar for 6-month mortality.

Our results show that o/e-LHR measured on two-dimensional ultrasound is a good indicator of neonatal prognosis in cases of CDH that may be used even in fetal medicine centers with a small caseload. However, our results also suggest that LHR measurement may be difficult to perform correctly. Therefore, appropriate training should be offered to professionals. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.