Is Extrahepatic Biliary Atresia an HLA-Associated Disease?

To the Editor: Extrahepatic biliary atresia (EHBA) is an inflammatory obliteration of the extrahepatic biliary system of unknown origin, affecting 1 of 10,000 live births (1). In a subgroup of patients (10-15%), the disease is caused by congenital developmental abnormalities (the embryonic or fetal type). However, in most instances, the disease occurs as an isolated anomaly and is considered to be an acquired and progressive disease (of the perinatal type) (2), suggesting a different etiology or a pathogenesis that remains undefined. Environmental agents such as toxins, ischemic insults, and infectious agents have been proposed as etiological candidates (3). Nevertheless, even with an infectious agent as the pathogen, a minority of infected people develop EHBA, and it is tempting to speculate whether host factors as well as extrinsic factors may be important (2,3). The possibility of a genetic predisposition to biliary atresia has been raised. In a previous report in this journal, on the basis of HLA typing, Silveira et al. (4) determined the HLA phenotypes in 55 patients (8 embryonic and 47 perinatal types) with EHBA, using conventional serologic techniques, and described an association with HLA-B12 in those with the perinatal type. Using serologic and biologic molecular techniques, we have studied the polymorphism of HLA molecules in a group of 48 patients with EHBA (6 embryonic and 42 perinatal types) attended in the Pediatric Hepatology Department of La Paz Hospital, Madrid, to test whether HLA-B12 or other antigens are increased (4) and to study the possible association with other class II antigens. A total of 80 unrelated, healthy Spanish subjects were used as a control group. HLA class I phenotypes were determined in 26 patients, using a standard complement-dependent cytotoxicity assay (One Lambda, Canoga Park, CA, U.S.A.) to determine 16 A and 23 B antigens. Class II typing was determined in 46 patients, using a reverse hybridization-based test (Amplicor HLA DRB test; Roche Diagnostic Systems, Branchburg, NJ, U.S.A.) performed according to manufacturer's instructions. Currently, these DNA techniques are recommended for HLA typing. Statistical significance of differences in antigen frequencies between patients and controls was measured by Fisher's exact test, and the probability was corrected (Pc) for the number of antigens tested, a correction that is necessary when multiple tests are performed (5). Our results in class I typing show a frequency of HLA-B12 (B44 and B45) antigen that is even lower than that in the control population. No significant increase in the other HLA-I antigens was detected; consequently, no more patients were studied (Table 1). Class II typing shows an increase of the DR6 antigen in the total sample, but the increase is only slightly significant (41.6% versus 25%; relative risk (RR) = 2; p = 0.05; Table 2). Moreover, for an accurate interpretation of the results, it is mandatory to apply Sveejgard's recommendations (5). In accordance, when the p value was corrected, the association with DR6 was not significant. The DR6 antigen has a great variety of alleles, but the HLA typing system used in this work cannot discriminate between them, and thus an association with DR6 cannot be completely excluded. Our data contrast with the association with the B12 antigen reported by Silveira et al. (4) One possible explanation is a different genetic background in the two populations. We thought, however, that this discrepancy could be caused by the difference in the calculation of the p value: In the work of Silveira et al. the probability is obtained by comparing the percentages of patients and controls with the B12 antigen, whereas we compared the absolute number instead of the percentage. Therefore, according to our results, EHBA does not appear to be HLA-mediated, and further research is necessary to clarify other factors, especially known or unknown infectious agents, responsible for this devastating disease. Aurora Jurado; *Paloma Jara; *Carmen Camarena; *Loreto Hierro; Carlos Lahoz; Pilar Palomino Immunology Department, Fundación Jiménez Díaz;*Pediatric Hepatology Department, Hospital La Paz; Madrid, Spain