Albumin, bilirubin, or ALBI score for predicting drug clearance in hepatic impairment

Introduction: Hepatic impairment is classified as mild, moderate, and severe based on either the Child–Pugh or the National Cancer Institute (NCI) classification system. Another classification system based on bilirubin and albumin levels, known as albumin–bilirubin (ALBI) scoring system, has been proposed mainly for patients with cancer with hepatic impairment. The objective of this study was to predict drug clearance in subjects with varying degrees of hepatic impairment using albumin or bilirubin or albumin–bilirubin (ALBI) score.

Materials and methods: There were thirty-five drugs with 64 observations (data points with different degrees of hepatic impairment) in this study. The methods used to predict drug clearance (CL) in subjects with hepatic impairment were bilirubin or albumin or ALBI score. Prediction fold-errors of 0.5–2, 0.5–1.5, and 0.7–1.3 and several statistical parameters such as average fold error (AFE), average absolute fold error (AAFE), bias, and root mean square error (RMSE) were used for comparison purposes among the three methods for the prediction of CL in subjects with different degrees of hepatic impairment.

Results: More than 90% of observations were within 0.5–2-fold prediction error by all three methods. Based on albumin, bilirubin, and ALBI, 75.0%, 78.1%, and 89.1% of observations were within 0.5–1.5-fold prediction error, respectively. ALBI in particular provided a good prediction of CL in subjects with severe liver impairment (100% within 0.5–2, 84.6% within 0.5–1.5, and 61.5% within 0.7–1.3-fold prediction error).

Conclusions: All three approaches are simple. ALBI is suitable for the prediction of drug clearance in patients with hepatic impairment and is simple and fairly accurate in its predictive performance.