Conjugated bilirubin as a reflex test for increased total bilirubin in apparently healthy population

Background Total, unconjugated and conjugated bilirubin levels are usually ordered together in health check‐up populations. The aim of this study was to investigate whether using increased total bilirubin ( TBIL ) as a reflex test can reduce conjugated bilirubin ( CBIL ) test. Methods Medical records of 8433 males and 4496 females who visited Shuyang People's Hospital for health check‐ups were retrospectively reviewed and the fasting serum TBIL , unconjugated bilirubin ( UBIL ) and CBIL of patients were extracted. Reference intervals for TBIL , UBIL , CBIL and C/ TBIL were established using Q 2.5 to Q 97.5 . The relationship between TBIL and CBIL was analyzed by Spearman's approach. Receiver operating characteristics ( ROC ) curve analysis was used to evaluate the predictive accuracy of TBIL for abnormal CBIL and UBIL . Results The reference intervals for TBIL in males and females were 6.9‐29.3 μmol/L and 6.1‐23.8 μmol/L, respectively. For CBIL , the reference intervals were 1.9‐10.4 μmol/L and 1.6‐8.8 μmol/L for males and females, respectively. CBIL was significantly positively correlated with TBIL , either in males ( r =.75) or females ( r =.73). Area under curve ( AUC ) of TBIL for predicting abnormal CBIL was 0.99 in both male and females. The optimal threshold of TBIL for predicting abnormal CBIL and UBIL were 21.0 μmol/L in males and 17.0 μmol/L in females. At these thresholds, <2% of subjects with abnormal CBIL or CBIL might be missed, but approximately 87% of the CBIL test could be eliminated. Conclusion Conjugated bilirubin measurement is not needed for the apparently healthy males with TBIL <21.0 μmol/L or females with TBIL <17.0 μmol/L.