Comparison of five routine methods with the candidate reference method for the determination of bilirubin in neonatal...

Using five routine methods and the candidate reference method of Doumas (Clin. Chem. 31, 1779-1789 (1985)), total bilirubin was determined in 77 neonatal serum samples (concentration range 63-444 mumol/l, average value 227 mumol/l). Four of the routine methods (Jendrassik & Grof's (Biochem. Z. 297, 81-89 (1938)) method, Hertz's (Scand. J. Clin. Lab. Invest. 33, 215-230 (1974)) method, the bilirubinometer procedure, and the method employing 2,5-dichlorophenyldiazonium (Scand. J. Clin. Lab. Invest. 29, Suppl. 126, Abstr. 11. 12. (1972]) gave values that were generally higher than those of the reference method. In contrast, the results from Vink's (Clin. Chem. 34, 67-70 (1988] direct spectrophotometric method differed only negligibly from those of the reference method. The accuracy of Jendrassik & Grof's method, and to a limited extent that of the 2,5-dichlorophenyldiazonium method, can be improved by redetermination of the molar absorption coefficient, or by using a standard containing a matrix of human albumin, with an assigned value determined by the reference method. It was found that Hertz's direct spectrophotometric method can be replaced by that of Vink. The accuracy of the bilirubinometer results could be improved only by using calibrators with assigned values specific for the bilimeter, or by calibration with a serum pool.