GAHS was developed for adults admitted with suspected or biopsy-confirmed alcoholic hepatitis, usually after recent heavy alcohol use and with marked jaundice. It is intended for use in hospital once baseline blood tests are available and has not been validated in other causes of acute hepatitis, isolated cirrhosis decompensation without alcoholic hepatitis, or stable chronic liver disease.

Evidence & context

GAHS was derived in 241 inpatients with clinical alcoholic hepatitis and serum bilirubin ≥80 µmol/L at two Glasgow hospitals, then validated in a separate UK cohort of 195 patients from eight centres. No patients in these cohorts received corticosteroids or pentoxifylline at the time of scoring; outcomes were 28- and 84-day mortality.
Age, blood urea, serum bilirubin, prothrombin time/INR and peripheral white cell count were selected by logistic regression as independent predictors of short-term mortality. GAHS showed higher overall accuracy than the modified Maddrey discriminant function for predicting 28- and 84-day outcome in both derivation and validation datasets.
Performance was similar whether prothrombin time was expressed as INR or as a ratio, and whether the diagnosis of alcoholic hepatitis was based on clinical criteria alone or confirmed on biopsy.

Risk bands and short-term outcomes (derivation + validation cohorts)

Timing	GAHS band	Day 28 survival	Day 84 survival
Day 1	<9	87%	79%
Day 1	≥9	46%	40%
Days 6–9	<9	93%	86%
Days 6–9	≥9	47%	37%

Subsequent work showed that among patients with severe alcoholic hepatitis by modified discriminant function (mDF ≥32), clinically important corticosteroid benefit was confined to those with GAHS ≥9, with little benefit seen when GAHS <9. Major international guidelines, including NICE and EASL guidance on alcohol-related liver disease, recognise GAHS ≥9 as one way to identify high-risk patients when considering corticosteroids, alongside other tools such as mDF, MELD, ABIC and Lille.

Glasgow Alcoholic Hepatitis Score (GAHS)