Send letter to journal:
Re: Prolonged indirect hyperbilirubinemia

Dear Editor

Lee at al correctly conclude that in infants with prolonged indirect hyperbilirubinemia, bile pigment analysis should be delayed until the child is at least three months old. (Arch Dis Child 2000;85:38-42) Certainly, there is no rush to make a diagnosis in most infants, particularly those who are breast fed, and thriving. I note that with one exception, all of the infants in the study who subsequently turned out to be normal or ? Gilbert's were breast fed. Indirect hyperbilirubinemia lasting for 10-12 weeks is well described in breast fed infants and practitioners need to know that in a thriving, breast fed infant, nothing needs to be done other than to ensure that the infant's thyroid function is normal (usually achieved by confirming that the metabolic screen done in the nursery was normal) and to confirm that the direct bilirubin level is not elevated, thus ruling out cholestasis.

Over the last few years I have had many calls from pediatricians about breast fed infants who are 8-12 weeks old with indirect bilirubin levels in excess of 170 micromol/L. I always mention the possibility of Crigler-Najjar syndrome (CNS) and Gilbert's syndrome but also tell the caller that by far the most likely diagnosis is breast milk jaundice. None have had CNS and I know of one with Gilbert's although only a few have had the laboratory testing for Gilbert's.

It is also important to emphasize that what we are talking about is prolonged unconjugated hyperbilirubinemia. The first sentence of the paper states that "common causes" of unconjugated hyperbilirubinemia include "over-production (that is, polycythemia), systemic disease (congenital hypothyroidism, sepsis), and inherited disorders of bilirubin metabolism". Polycythemia does not belong here - do the authors mean hemolysis? Perhaps this is a typographical error. The remaining conditions listed are exceptionally rare causes of unconjugated hyperbilirubinemia although they should certainly be considered in infants with prolonged jaundice.

I am also quite puzzled by the figure and the recommendations provided by the authors. The figure suggests that common causes should be excluded in infants with bilirubin levels exceeding 200 micromol/L and that phenobarbitone treatment should be offered to these infants. I would argue that no investigations (other than those I mention above) are needed in a thriving six week old breast fed infant who has a bilirubin level of 200 micromol/L, and I cannot think of any reason why such an infant should receive phenobarbitone. Are the authors suggesting that this level represents a threat to the infant's brain? Certainly, if the infant is not breast fed, further investigation should be considered if the jaundice persists. Treatment might be considered if the level exceeds 290 micromol/L.

Finally, in the section headed "genetic studies", the statement is made that "neither TA6 nor TA6/TA7 are known to be associated with hyperbilirubinemia." This may not be correct. It is true that in the absence of any additional icterogenic factors, the observed clinical effect of the 6/7 heterozygosity is minimal. But this is not the case when other factors that increase the bilirubin load or impair hepatic conjugating capacity are present. Kaplan et al found that when G6PD deficiency and the 6/7 heterozygosity for the UGT promotor polymorphism were both present, 31.6% of infants developed serum bilirubin levels > 257 micromol/L, compared with 9.7% of those G6PD deficient but homozygous for the normal UGT promotor (6/6). (Lancet 2000;356:652-3)