Editor,—We write in support of Tripp and McNinch’s paper1 recommending daily oral administration of 25 μg phytomenadione to all breast fed infants up to the age of six months. However, we remain concerned that high risk infants with cholestatic jaundice (up to 1 in 500 births2) are still liable to develop bleeding secondary to vitamin K deficiency.

An audit of all infants younger than 3 months admitted between 1 January 1998 and 30 June 1998 was carried out by retrospective review of case notes. The aim of the review was to evaluate vitamin K prophylaxis and the prevention of coagulopathy in jaundiced neonates. The setting was a supraregional centre for paediatric liver disease in Birmingham, UK.

Twenty seven jaundiced infants were admitted during the study period with the following diagnoses: extrahepatic biliary atresia (7), acute liver failure (4), α₁ antitrypsin deficiency (3), neonatal hepatitis (4), Alagille syndrome (3), Niemann pick disease (2), other (4). The mean age at presentation was 6 weeks (range 4–10). Three children with extrahepatic biliary atresia presented with gastrointestinal bleeding. All three children had rectal bleeding and one of them had haematemesis while another had extensive bruising, which his parents were initially accused of inflicting. All three had prothrombin times in excess of 100 seconds, which was corrected after a single 1 mg dose of intravenous vitamin K. All three babies had been exclusively breast fed before presentation and had received two to three doses of oral vitamin K. Despite this, these high risk infants developed a life threatening coagulopathy and it is fortunate that none had an intracranial haemorrhage.

It is now recommended that all babies with jaundice persisting beyond 14 days of age be evaluated medically.3 We would like to add to Tripp and McNinch’s recommendation about oral vitamin K prophylaxis, that breast fed infants who are referred for evaluation of conjugated hyperbilirubinaemia have coagulation measured. Intramuscular vitamin K (1 mg) should be given without delay if coagulation studies cannot be easily obtained. In this group of children with liver disease, the risk of vitamin K deficiency bleeding was 11%, which exceeds the putative risk of childhood leukaemia.4