Glasgow Alcoholic Hepatitis Score Calculator
Glasgow Alcoholic Hepatitis Score Calculator
Score the GAHS from five variables, with Forrest’s day-28 and day-84 survival beside the total and the 9-point threshold declared inclusive.
Glasgow alcoholic hepatitis score
5 variables → 5–12 pointsAge 54, white cells 16.8 ×10⁹/L, urea 7.2 mmol/L, INR 1.8, bilirubin 310 µmol/L
Scoring
Each variable scores 1, 2 or 3; total 5 to 12; threshold 9 or more
- Age
- < 50 → 1 · ≥ 50 → 2
- White cell count
- < 15 ×10⁹/L → 1 · ≥ 15 → 2
- Urea
- < 5 mmol/L → 1 · ≥ 5 → 2
- PT ratio / INR
- < 1.5 → 1 · 1.5–2.0 → 2 · > 2.0 → 3
- Bilirubin
- < 125 µmol/L → 1 · 125–250 → 2 · > 250 → 3
- Threshold
- 9 or more identifies the patients most at risk of death, scored on admission or at days 6 to 9
Worked example
Age 54, white cells 16.8 ×10⁹/L, urea 7.2 mmol/L, INR 1.8, bilirubin 310 µmol/L
Age 54 ≥ 50 → 2 · white cells 16.8 ≥ 15 → 2 · urea 7.2 ≥ 5 → 2
INR 1.8 is 1.5–2.0 → 2 · bilirubin 310 > 250 → 3
2 + 2 + 2 + 2 + 3 = 11 points
11 is 9 or more → day-28 survival 46%, day-84 survival 40% in Forrest's series
The scoring table
| Variable | 1 point | 2 points | 3 points |
|---|---|---|---|
| Age (years) | < 50 | ≥ 50 | — |
| White cell count (×10⁹/L) | < 15 | ≥ 15 | — |
| Urea (mmol/L) | < 5 | ≥ 5 | — |
| PT ratio or INR | < 1.5 | 1.5–2.0 | > 2.0 |
| Bilirubin (µmol/L) | < 125 | 125–250 | > 250 |
GAHS against Maddrey’s discriminant function
| Maddrey DF | GAHS | |
|---|---|---|
| Published | 1978 | 2005 |
| Inputs | PT in seconds against a laboratory control, bilirubin in mg/dL | Age, white cells, urea, PT ratio, bilirubin |
| Takes an INR? | No — the formula is not defined for it | Yes, the PT ratio is one of its bands |
| Treatment threshold | 32 or more | 9 or more |
| Accuracy for 28-day outcome in Forrest’s series | 49% | 81% |
| What it is used for | Entry criterion for corticosteroids in most guidelines | Identifying, within that group, who benefits |
What the GAHS adds to Maddrey
The Glasgow alcoholic hepatitis score was derived on 241 patients and validated on a further 195, with the explicit aim of improving on Maddrey’s discriminant function. In the derivation cohort the GAHS predicted 28-day outcome with an overall accuracy of 81% against the modified discriminant function’s 49% — which is the single figure that explains why the score exists. Five variables, each banded 1 to 3, total 5 to 12, and a threshold of 9 or more.
The discriminant function has not been displaced and this page is not arguing that it should be. A DF of 32 or more remains the entry criterion for corticosteroids in most guidelines, because the trials that established steroid benefit were built on it. What the GAHS contributes is specificity within that group. The two scores disagree in a defined population: patients with a discriminant function of 32 or more — severe by the conventional criterion — but a GAHS below 9. That discordant group is the one in which the case for corticosteroids is weakest, and Forrest’s 2007 follow-up paper, which reported steroid outcomes stratified by GAHS within the DF ≥ 32 population, is the reference for that claim. The survival percentages from that stratification are deliberately not quoted here: the paper’s abstract could not be read directly and no reachable source reproduces its figures, so the page states the finding and names the reference rather than printing numbers it cannot stand behind.
There is a practical reason the GAHS is easier to score. Maddrey’s formula requires a prothrombin time in seconds against the reporting laboratory’s own control, and many laboratories now report only an INR — substituting it gives a number that is not the discriminant function. The GAHS takes a prothrombin time ratio, and an INR is a ratio, so the score can be computed from what a modern report actually contains. It also takes bilirubin in µmol/L, which is how UK laboratories report it, rather than requiring a conversion first.
Neither score is the response measure. Once corticosteroids are started, the Lille model at day 7 decides whether to continue: a Lille above 0.45 indicates non-response and steroids should be stopped rather than completed. The sequence in current practice is therefore three scores doing three different jobs — a severity score to decide whether to treat, the GAHS to sharpen that decision, and Lille to decide whether to carry on. Running any one of them in isolation misses the point of the other two.
Two operational details from the derivation are worth keeping. First, the GAHS discriminates as well when calculated at days 6 to 9 as on the day of admission, so a score of 8 on arrival is worth recomputing rather than treated as settled. Second, the threshold is published as “9 or more”, not “above 9” — a score of exactly 9 is in the high-risk group, and the bands above are declared inclusive so that a 9 cannot fall through into the wrong side. As with any severity score in this disease, sepsis, gastrointestinal bleeding, acute kidney injury and uncontrolled hyperglycaemia all need addressing before corticosteroids are considered, and active infection remains a contraindication whatever the score says.
Frequently asked questions
What GAHS score is severe?
9 or more. Survival in Forrest’s series was 46% at day 28 and 40% at day 84 at or above 9, against 87% and 79% below it. The criterion is “9 or more”, so a score of exactly 9 is in the high-risk group.
How does the GAHS differ from Maddrey’s discriminant function?
It uses five banded variables rather than two continuous ones, takes a prothrombin time ratio or INR rather than a PT in seconds against a laboratory control, and predicted 28-day outcome with 81% accuracy against the modified discriminant function’s 49% in the derivation cohort. The discriminant function remains the entry criterion for corticosteroids in most guidelines; the GAHS sharpens the decision within that group.
What about a patient with a discriminant function over 32 but a GAHS under 9?
That is the group in which the two scores disagree, and it is the group in which the case for corticosteroids is least clear. Forrest’s 2007 paper in Gut, which stratified steroid outcomes by GAHS among patients with a discriminant function of 32 or more, is the reference for that. The survival figures from that stratification are not quoted on this page because they could not be verified from a reachable source.
Can I use an INR in the GAHS?
Yes. The prothrombin time variable is a ratio, and an INR is a ratio — under 1.5 scores 1, 1.5 to 2.0 scores 2, above 2.0 scores 3. This is one of the practical advantages over Maddrey’s discriminant function, which requires a prothrombin time in seconds against a laboratory control and is not defined for an INR.
When should the GAHS be scored?
On admission, or at days 6 to 9 — the derivation found comparable discrimination at both. Because the score changes as bilirubin and the prothrombin time move, a score just below the threshold is worth recalculating rather than treated as final.
Related calculators
References
- Forrest EH, Evans CDJ, Stewart S, et al. Analysis of factors predictive of mortality in alcoholic hepatitis and derivation and validation of the Glasgow alcoholic hepatitis score. Gut. 2005;54(8):1174–9.
- Forrest EH, Morris AJ, Stewart S, et al. The Glasgow alcoholic hepatitis score identifies patients who may benefit from corticosteroids. Gut. 2007;56(12):1743–6.
- Maddrey WC, Boitnott JK, Bedine MS, et al. Corticosteroid therapy of alcoholic hepatitis. Gastroenterology. 1978;75(2):193–9.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.