ALBI Grade Calculator (Albumin-Bilirubin Score)
Calculate the albumin-bilirubin (ALBI) score and grade 1, 2 or 3 for liver function in hepatocellular carcinoma and chronic liver disease, from serum albumin and bilirubin in US or SI units.
How it works
ALBI = 0.66 × log₁₀(bilirubin, µmol/L) − 0.085 × albumin (g/L). Grade 1: ≤ −2.60 · Grade 2: > −2.60 to ≤ −1.39 · Grade 3: > −1.39. (Bilirubin mg/dL × 17.1 = µmol/L; albumin g/dL × 10 = g/L.)
The ALBI grade (Johnson et al., Journal of Clinical Oncology 2015) was developed in hepatocellular carcinoma as an objective alternative to Child-Pugh, using only two lab values and no subjective items such as ascites or encephalopathy. Grade 1 has the best survival and grade 3 the worst, and ALBI separates patients within Child-Pugh class A into groups with different outcomes. It is now used in HCC staging discussions, treatment selection (resection, TACE, systemic therapy) and liver disease research.
Worked example
Bilirubin 1.2 mg/dL (20.5 µmol/L) and albumin 3.5 g/dL (35 g/L): ALBI = 0.66 × log₁₀(20.52) − 0.085 × 35 = 0.866 − 2.975 = -2.11 → grade 2 (intermediate liver function).
Frequently asked questions
What do ALBI grades 1, 2 and 3 mean?
Grade 1 (score ≤ −2.60) is the best-preserved liver function, grade 2 (−2.60 to −1.39) intermediate and grade 3 (above −1.39) the most impaired, with the shortest survival in the original cohorts.
Why is the ALBI score negative?
Because the albumin term (−0.085 × g/L) is larger than the bilirubin term in most patients. More negative is better.
ALBI vs Child-Pugh — which is better?
ALBI is fully objective and splits Child-Pugh A patients into different risk groups, but Child-Pugh still captures ascites and encephalopathy. Many clinicians report both.
Can ALBI be used outside liver cancer?
It has been studied in cirrhosis, primary biliary cholangitis and drug-induced liver injury, but it was validated first in hepatocellular carcinoma.
Related calculators
Clinical decision tool for education and for use by trained clinicians alongside clinical judgement — not a diagnosis and not a substitute for medical assessment. Scores validated in specific populations may not apply to every patient.