Dr. Clayton's letter raises pertinent points with regard to the interpretation of the finding of increased urinary excretion of Δ4-3- oxo bile acids. These metabolites can be detected in a variety of severe cholestatic liver disorders. The interpretation of their presence in liver failure is much more complex. Currently the finding of high proportions of Δ4-3-oxo bile acids and markedly reduced or absent primary bile acids is used as supportive evidence of a quantitative defect in Δ4-3-oxosteroid 5β-reductase enzyme activity. This pattern of urinary bile acid metabolites is not necessarily a simple reflection of liver dysfunction. Fast atom bombardment ionization-mass spectrometry analyses of urine from many patients with liver failure, including four with the clinical diagnosis of neonatal hemochromatosis, did not reveal this bile acid profile. An additional five patients have been identified who had a clinical diagnosis of neonatal hemochromatosis, and had a bile acid metabolite pattern similar to the two patients that we described.1Shneider BL Setchell KDR Whitington PF Neilson KA Suchy FJ. Δ4-3-Oxosteroid 5β-reductase deficiency causing neonatal liver failure and hemochromatosis.J PEDIATR. 1994; 124: 234-238Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar Definitive diagnosis of a primary genetic deficiency of the enzyme is hampered by the fact that it is not expressed in fibroblasts. When possible we have attempted to obtain additional evidence for an enzyme defect by immunoblotting for the 38 kd protein with monoclonal antibodies against the enzyme.2Russell DW Setchell KDR Bile acid biosynthesis.Biochemistry. 1992; 31: 4737-4749Crossref PubMed Scopus (644) Google Scholar As pointed out by Dr. Clayton, the finding of an absence of immunodetectable protein may be caused by a generalized destruction of hepatic proteins. The finding of a truncated protein in our first patient is intriguing.1Shneider BL Setchell KDR Whitington PF Neilson KA Suchy FJ. Δ4-3-Oxosteroid 5β-reductase deficiency causing neonatal liver failure and hemochromatosis.J PEDIATR. 1994; 124: 234-238Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar This feature is more consistent with a primary defect, rather than simple nonspecific loss of proteins as a result of hepatocellular injury. At present, confirmatory evidence of a primary enzyme defect can best be established from the observation that an interruption of primary bile acid therapy in these patients, after liver function has returned to normal, results in a reemergence of Δ4-3-oxo bile acids as major metabolites in urine. A human complementary DNA encoding 5β-reductase was recently sequenced, and molecular genetic techniques should soon be available to aid in clarifying the role of a primary genetic deficiency of this enzyme in the pathogenesis of liver failure.3Kondo K-H Kai M-H Setoguchi Y et al.Cloning and expression of cDNA of human Δ4-3-oxosteroid 5β-reductase and substrate specificity of the expressed enzyme.Eur J Biochem. 1994; 219: 357-363Crossref PubMed Scopus (121) Google Scholar The role of biliary excretion of iron has been persistently underestimated and poorly studied, especially in the newborn period. Examination of the studies of Hultcrantz et al.4Hultcrantz R Angelin B Bjorn-Rasmussen E Ewerth S Einarsson K Biliary excretion of iron and ferritin in idiopathic hemochromatosis.Gastroenterology. 1989; 96: 1539-1545PubMed Google Scholar indicates that normal biliary excretion amounts to between 0.4 and 2.3 mg of iron/day, in the range of normal uptake of iron. We agree that the finding of 5β-reductase deficiency should not be the end of the diagnostic evaluation in an infant with neonatal liver failure, but it should be part of the evaluation. Only through systematic evaluation of infants with neonatal liver failure will we determine the significance of this finding, in particular its relationship to the development of neonatal hemochromatosis. Unfortunately, at this time pediatric hepatologists are faced with very difficult therapeutic decisions. More exact molecular probes may simplify those decisions in the future. 9/35/58922