Blood gas bilirubin measurements in neonates must be adjusted for HbF to avoid misleading results

Abstract

Hyperbilirubinaemia is common in the first week of life with 60%–80% of babies developing jaundice. Though usually uncomplicated, untreated hyperbilirubinaemia, can lead to irreversible neurological injury. Prevention of bilirubininduced neurological dysfunction (BIND) includes universal risk factor and clinical assessment with additional transcutaneous bilirubinometer or Total serum bilirubin (TSB) assessment. Point of care testing (POCT) has the advantage of rapid bedside results, improving outcomes. Bilirubin measurement is available as a POCT via blood gas analysers in most neonatal intensive care units (NICU). Studies have shown good accuracy and precision of blood gas bilirubin compared with standard serum bilirubin assessment. 4 A blood gas machine (Radiometer ABL 800 flex) was installed by the company in a new level 3 NICU in Australia and calibrated by the laboratory to provide bilirubin assay as a POCT. Over the first 2 months of use, bilirubin levels did not correlate with the clinical condition, requiring repeat testing with TSB. Data where both TSB and gas bilirubin were done on the same day was analysed. A BlandAltman plot was constructed and accuracy and precision in terms of mean bias and limits of agreement were calculated. The mean bias between the gas and laboratory bilirubin was −76 μmol/L with the limits of agreement being −0.54 to −151.7 (figure 1), suggesting a significant difference. Investigation revealed that the fetal haemoglobin (HbF) correction had not been activated at installation. After this was activated, paired samples of gas and serum bilirubin done showed the mean bias was −17.3 with limits of agreement 3.9 to −35.9 (figure 1). The calibration of the POCT was done using adult blood. This did not factor in that 50%–95% of neonatal haemoglobin is HbF. The blood gas method measures bilirubin using multiwavelength spectrophotometry of haemolysed whole blood. As the absorbance spectra of adult and HbF are different, the results will differ based on the concentration of HbF. The blood gas machine has the option of HbF correction that needs to be enabled. This was not enabled by the company nor was this known to the laboratory that maintained the blood gas machine. The calibration used adult blood and hence passed quality control. The specimens used did not match patient specimen (noncommutability), exposing the weakness in the quality control procedures. BIND is a preventable serious complication of hyperbilirubinaemia with sick and preterm babies admitted in the NICU being at higher risk. The initial results of the blood gas machine were much lower than the laboratory bilirubin with wide limits of agreement (figure 1). This difference would miss babies with significant hyperbilirubinaemia with potential to cause significant harm if gas bilirubin was used to diagnose hyperbilirubinaemia. It is important when managing neonates to consider their unique biology and not to treat them as small children or tiny adults. Physiology even within species differs, varying with age, race, gender, etc. It is important to factor this in to provide safe and highquality healthcare. Our experience also highlights the importance of using commutable reference material for calibration.