Population-related differences in fetal aortic arch dimensions

Abstract

Diagnosing fetal aortic arch anomalies, such as coarctation of the aorta (COA), is difficult. Hornberger et al. [1] constructed nomograms of the normal development of the aortic arch to help detect anomalies by comparing sonographic images with the nomograms for the same gestational ages. However, these graphs are not always reliable across populations. For instance, our measurements of the distance between the mitral and tricuspid valves in growing Thai fetuses differed from those published in a Western study [2]. We are reporting on how we constructed nomograms of the fetal aortic arch in growing Thai fetuses using the transverse aortic arch diameter (TAD) and the distal aortic isthmus diameter (DAID), and evaluated the nomograms' helpfulness in detecting fetal aortic arch anomalies in the Thai and other populations. We also report on the differences between these Thai nomograms and those constructed by Hornberger et al. in the United States [1]. We conducted this work from January to September 2007 with 294 women whose pregnancy duration was accurately known. None of the 20to 40-week fetuses had structural anomalies. Ultrasound exploration was done once in each participant using an Aloka model SSD Prosound 5000 (Aloka, Tokyo, Japan) or a Voluson 730 Expert (GE Medical Systems, Waukesha, Wisconsin, USA). With the arch viewed in the long axis, TAD was measured between the common carotid and the left subclavian arteries and DAID was measured distal to the origin of the left subclavian artery, both during the ventricular systole. The mean values of 3 measurements were calculated and plotted against gestational age. Correlation coefficients were determined using the Pearson correlation. The TAD and DAID measurements presented are the 5th, 50th, and 95th percentile (Table 1). Using the Mann-Whitney U test, we compared echocardiographic findings for 5 fetuses with COA obtained between the 22nd and 38th weeks of gestation to our nomograms. We also compared our nomograms to those published in the report by Hornberger et al. [1] using the Z test. The following regression equations were found to have the best fit: