The effect of repeated phlebotomy on bilirubin turnover, bilirubin clearance and unconjugated hyperbilirubinaemia in the Crigler-Najjar syndrome and the jaundiced Gunn rat: application of computers to experimental design.

1. A multicompartmental model of erythrokinetics and bilirubin production has been developed to predict the consequences of chronic phlebotomy on daily bilirubin turnover. 2. Control values for four physiological variables including bilirubin turnover were determined in a 20-year-old woman with type I congenital nonhaemolytic jaundice (Crigler-Najjar syndrome). With these base-line data, the model predicted the following changes during phlebotomy: a 34% fall in bilirubin turnover; a 240% increase in the haemoglobin content of bone-marrow erythroid precursors; a 25% fall in the half-life of 51Cr-labelled erythrocytes; a characteristic alteration of the erythrocyte survival curve after labelling with [2-14C]glycine. 3. On the assumption, previously validated in normal volunteer subjects and patients with Gilbert's syndrome, that hepatic bilirubin clearance was independent of turnover and would therefore remain unchanged, a fall in plasma unconjugated bilirubin concentration during phlebotomy from 436 to 282 mumol/1 was expected. 4. Accordingly, the patient underwent phlebotomy 350 ml/week for 2 months, and 500 ml/week during an additional 3 months. Appropriate studies during phlebotomy confirmed each of the predictions in paragraph 2 above. In particular, turnover fell by 31%. Unexpectedly, plasma unconjugated bilirubin remained essentially unchanged. Analogous results were observed in phlebotomized jaundiced Gunn rats. 5. Kinetic studies in both the patient and the rats demonstrated that the failure of plasma unconjugated bilirubin to fall in parallel with bilirubin turnover resulted from a prolongation of the terminal half-life of radioactively labelled bilirubin and a fall in bilirubin clearance in every instance. 6. These studies indicate that (a) in congenital non-haemolytic jaundice, bilirubin clearance is uniquely influenced by bilirubin turnover and (b) compartmental modelling and kinetic studies are useful for predicting and interpreting the results of both physiological experiments and experimental therapeutic regimens.