Zhang et al.1 report an elegant study looking at the expression of a large number of genes and gene products in infants with biliary atresia, the main aim being to demonstrate gene or gene expression differences between two distinct clinical phenotypes. However, both the underlying general assumption and the specific nature of the patients in this study need to be challenged before one can accept the results as conclusive.
Use of the terms embryonic and perinatal in association with biliary atresia is widespread in the literature. Both terms imply an explicit assumption regarding the timing of the etiological cause; however, for a number of reasons this approach may be somewhat simplistic. The “perinatal” form of biliary atresia implies that there is destruction of an already fully formed biliary tree by a virus (presumably) at or around the time of birth. The two elements of this assumption are controversial. First, many viral studies in humans are not referenced in the article that are entirely negative but still perfectly valid.2, 3 Second, the assumption that the timing of an etiological insult is perinatal has little actual evidence to support it. Antenatally detected biliary atresia, although it represents a small proportion of most series (≈5%) and has an unusual biliary appearance (cystic), has implications on the timing of biliary atresia occurrence. In our recently reported series of 9 infants with biliary atresia, all occurrences were detected between 18 and 20 weeks' gestation, with 8 of 9 being nonsyndromic.4 Furthermore, the key studies of Francoise Muller, who measured various gastrointestinal enzymes (specifically γ-glutamyltranspeptidase) in serial samples of amniotic fluid, have also shown that in those cases of nonsyndromic biliary atresia detected “incidentally,” there was definite evidence of bile obstruction early in the second trimester.5–7
In the current study, the authors have chosen very unusual examples of biliary atresia and classified them as “embryonic.” Infants with the embryonic form of biliary atresia typically have a constellation of extrahepatic anomalies characterized by splenic anomalies (100%), situs inversus (50%), preduodenal portal vein (60%), absence of the inferior vena cava (40%), and cardiac anomalies (50%) (all percentages are based on the King's College series, currently n = 50). We have used the term biliary atresia splenic malformation syndrome,8 and others, polysplenia9 or polyasplenia syndrome when describing such infants.
So why were the infants in Zhang's study exceptional? Of the 5 “embryonic” infants, only one was a typical example, with polysplenia and a preduodenal portal vein (infant 2). Infant 3 did not have a splenic anomaly but had other typical features (preduodenal portal vein, annular pancreas, and malrotation). Infant 5 had congenital cardiac abnormalities but only an interrupted inferior vena cava to suggest syndromic biliary atresia. Infant 1 was extremely abnormal and very atypical with diaphragmatic hernia, vaginovesicular fistula, cleft lip and palate, and so forth. Finally, the preterm infant (infant 4) had absolutely no features to suggest an embryonic cause for its biliary atresia, certainly not with a patent ductus arteriosum or hydronephrosis.
Nevertheless, using these infants, the authors then extrapolate their molecular and genetic findings and conclusions based on the more usual syndromic variant described above. For example, they searched for and found abnormal expression of laterality genes (i.e., Sprouty-4 like, Zinc family member-3-heterotaxy-1), when in fact not one of the “embryonic” infants had any clinical evidence of axial determination defects.
Searching for the roots of biliary atresia lies in unraveling the key molecular differences between the syndromic and nonsyndromic forms. The groups to be discriminated, however, need far better definition and a much higher degree of within-group homogeneity before we can speculate on any difference in their genes.