We welcome Dr. Davenport's interest in our article reporting hepatic transcriptional signatures that differentiate two clinical phenotypes in infants with biliary atresia. His comments underscore two important issues that need to be carefully considered in any patient-based studies addressing the pathogenesis of biliary atresia: etiology and time of onset of disease. Both issues are interrelated and, when used in conjunction with the presence or absence of nonhepatic malformations, have been used in the literature to identify clinical forms. In regards to etiology, there are sufficient data to suggest that viral insults are a plausible etiology despite the variable identification of specific viruses in published studies from different populations of patients with biliary atresia. This is further supported by an experimental mouse model of rotavirus-induced biliary atresia, in which the virus is efficiently cleared from the liver after obstruction of the extrahepatic bile duct.1, 2 The inability to detect viral elements in these mice even when sensitive techniques are used (such as PCR) underscores the concept that a negative result does not rule out a previous infection by a virus, which may have triggered an inflammatory and obstructive injury to the bile ducts. Therefore, patient- and animal-based studies support the existence of a “perinatal” form of acquisition in a group of infants with biliary atresia, perhaps due to an infectious insult. The exact timing of the proposed viral insult is currently unknown.

Dr. Davenport's comments about the time of onset of disease and the use of the term “embryonic” to describe a group of patients included in our study highlight an important gap in our understanding of pathogenic mechanisms of disease and the need to develop a uniform system to classify clinical subtypes of biliary atresia. In regards to the time of onset of disease, although experimentally the administration of rotavirus to pregnant mice at term results in biliary obstruction in their offspring,3 an association between a viral insult to the developing human fetus and the postnatal diagnosis of biliary atresia has not been fully established. However, the early onset of jaundice and the coexistence of congenital nonhepatic malformations in a subset of infants with biliary atresia imply, at least in part, a prenatal onset of disease. As described in the literature cited by Dr. Davenport, the high prevalence of unique malformations in these infants, especially splenic malformations and laterality defects, allows for the grouping of these patients into the biliary atresia–splenic malformation or polysplenia syndrome. However, the nomenclature to describe infants with biliary atresia presenting with other types of nonhepatic malformations is far from clear or uniformly accepted. For example, two additional phenotypic groups have been proposed based on the presence of associated anomalies that do not follow any recognizable syndromic pattern/sequence or the presence of intestinal malrotation and atresia; notably some of the abnormalities included minor cardiovascular and urogenital malformations.4 These differences notwithstanding, common to all patients with nonhepatic malformations is the coexistence of one or more congenital malformations. In this context, we applied the term “embryonic” to all five infants with biliary atresia who also had nonhepatic malformations. Despite the phenotypic heterogeneity among these infants, they shared unique transcriptional profiles, as demonstrated by the coordinated expression of regulatory genes and the overexpression of imprinted genes when compared to infants with biliary atresia without nonhepatic malformations (termed “perinatal”).

Our experimental design and the limited number of subjects were not adequate to analyze the hepatic transcriptome in search of molecular signatures that are unique to subtypes of infants with biliary atresia and nonhepatic malformations, such as those with the biliary atresia–splenic malformation syndrome. These studies will need a population size that allows for much greater discriminatory (statistical) power of the gene expression profiling, and enable the testing of hypotheses relating to the expression of laterality genes in the subgroup of infants with laterality defects. We agree with Dr. Davenport that critical to these analyses is a rigorous phenotypic definition of clinical groups and a much higher degree of within-group homogeneity. When such a population is assembled, it will be equally important to use mathematical models to explore the existence of novel subtypes based on molecular signatures, to determine how they correlate with clinical phenotypes, and to explore their impact on long-term outcome.


  1. Top of page
  • 1
    Riepenhoff-Talty M, Schaekel K, Clark HF, Mueller W, Uhnoo I, Rossi T, Fisher J, et al. Group A rotaviruses produce extrahepatic biliary obstruction in orally inoculated newborn mice. Pediat Res 1993; 33: 394399.
  • 2
    Shivakumar P, Campbell KM, Sabla GE, Miethke A, Tiao G, McNeal MM, et al. Obstruction of extrahepatic bile ducts by lymphocytes is regulated by IFN-gamma in experimental biliary atresia. J Clin Invest 2004; 114: 322329.
  • 3
    Czech-Schmidt G, Verhagen W, Szavay P, Leonhardt J, Petersen C. Immunological gap in the infectious animal model for biliary atresia. J Surg Res 2001; 101: 6267.
  • 4
    Carmi R, Magee CA, Neill CA, Karrer FM. Extrahepatic biliary atresia and associated anomalies: etiologic heterogeneity suggested by distinctive patterns of associations. Am J Med Genet 1993; 45: 683693.

Jorge A. Bezerra M.D.*, Ronald J. Sokol M.D.†, * Cincinnati Children's Hospital Medical Center, Cincinnati, OH, † The Children's Hospital, Denver, CO.