Hepatology-GI Calculators

What hepatology and gastroenterology cover, and the clinical questions these tools answer

Hepatology covers diseases of the liver, biliary system and the complications of portal hypertension: viral hepatitis, alcohol-related liver disease, non-alcoholic fatty liver disease, autoimmune liver disease, cirrhosis, ascites, variceal bleeding and hepatic encephalopathy. Gastroenterology covers the wider digestive tract, and in the emergency setting its sharpest questions concern acute appendicitis, acute pancreatitis and upper gastrointestinal bleeding. The tools in this library answer the recurring practical questions of both fields.

Prognosis in liver disease is dominated by two systems. The Child-Pugh score grades cirrhosis from A to C using bilirubin, albumin, INR, ascites and encephalopathy, and remains the everyday language for describing how sick a cirrhotic liver is. The MELD score, and its sodium-incorporating update MELD-Na, estimate 90-day mortality from bilirubin, creatinine, INR and sodium, and form the basis of liver transplant allocation in many countries: the higher the score, the more urgent the need. Together they answer the two questions every cirrhosis assessment asks: how advanced is this disease, and how urgently does this patient need a transplant?

Earlier in the disease course, the question is whether significant fibrosis is present at all. The FIB-4 index and the NAFLD fibrosis score combine routine blood results with age to estimate the likelihood of advanced fibrosis in patients with fatty liver disease, helping primary care clinicians decide who can be monitored locally and who needs hepatology referral. In the abdomen, the SAAG (serum-ascites albumin gradient) distinguishes portal-hypertensive ascites from other causes with a single threshold: a gradient of 1.1 g/dL or above points to portal hypertension. Light's criteria perform a parallel job for pleural fluid, classifying effusions as exudate or transudate to narrow the differential diagnosis.

The emergency tools address bleeding, inflammation and obstruction. The Glasgow-Blatchford bleeding score predicts the need for hospital-based intervention in upper gastrointestinal bleeding; a score of zero identifies very-low-risk patients who can often be managed as outpatients with urgent follow-up. The Alvarado score combines migratory pain, anorexia, nausea, tenderness, rebound, fever, leucocytosis and neutrophilia to estimate the likelihood of acute appendicitis. For acute pancreatitis, the BISAP score predicts mortality from five bedside variables available early in admission, while Ranson's criteria add a 48-hour assessment of severity.

When clinicians and students use these calculators

In hepatology clinics, FIB-4 and the NAFLD fibrosis score are calculated on patients with fatty liver or abnormal liver enzymes to triage referrals efficiently; repeating them over time tracks whether fibrosis is likely progressing. On the wards, every patient with decompensated cirrhosis gets a Child-Pugh grade and a MELD or MELD-Na score, which together frame discussions about prognosis, transplant referral and goals of care. The SAAG is calculated whenever diagnostic paracentesis is performed, because it immediately sorts the differential.

In the emergency department, suspected upper gastrointestinal bleeding is scored with Glasgow-Blatchford at triage to decide between admission and a safe discharge pathway. Suspected appendicitis is scored with Alvarado alongside clinical examination, particularly useful for junior doctors deciding who needs surgical review. Acute pancreatitis gets a BISAP score early and Ranson's criteria completed at 48 hours, structuring the decision about level of care. Light's criteria are applied whenever pleural fluid is sampled, including in patients with hepatic hydrothorax where the cause of the effusion is in question.

For students, these scores teach the logic of their specialities. Child-Pugh demonstrates how subjective bedside findings and objective labs combine into a single grade; MELD shows how a continuous score can drive a fair allocation system; and the Alvarado and BISAP scores illustrate how clinical prediction rules compress an examination into an actionable number.

Interpreting results: what the scores change in practice

Child-Pugh grades run from 5 to 15 points: 5 to 6 is class A, 7 to 9 class B, and 10 to 15 class C, with expected survival falling across the classes and surgical risk rising steeply in class C. MELD scores run from 6 to 40; higher scores mean higher 90-day mortality and, in transplant systems, higher priority. MELD-Na refines the estimate because hyponatraemia independently predicts mortality in cirrhosis. When the MELD-Na climbs, the conversation shifts from routine management toward transplant evaluation.

A FIB-4 below commonly used low cut-offs suggests advanced fibrosis is unlikely and the patient can usually stay in primary care with lifestyle management and periodic rechecking; values above the high cut-off suggest significant fibrosis and merit hepatology referral, with the intermediate zone calling for further assessment or repeat testing. A Glasgow-Blatchford score of zero is the best-known decision point in upper GI bleeding: these patients have a very low risk of needing intervention and can often be discharged with urgent outpatient endoscopy arranged. An Alvarado score of 7 or above indicates a high likelihood of appendicitis and usually prompts surgical review and imaging, while low scores support observation or alternative diagnoses.

In pancreatitis, higher BISAP scores predict higher mortality and help decide between ward care and higher-dependency monitoring; Ranson's criteria, with three or more positive at 48 hours signalling severe disease, guide ongoing management and counselling. A SAAG of 1.1 g/dL or above directs the workup toward portal hypertension and its causes, while a lower gradient points toward peritoneal disease such as infection or malignancy. Light's criteria classify an effusion as exudative if any one of the three protein and LDH conditions is met, which then focuses investigation on infection, malignancy or pulmonary embolism rather than systemic causes of transudates.

Limitations and pitfalls: what these tools cannot do

Every score here has blind spots. Child-Pugh includes two subjective components, the grading of ascites and encephalopathy, which introduces variability between observers. MELD can be distorted by factors that raise creatinine or INR for non-hepatic reasons, and transplant systems add standard exceptions for conditions such as hepatocellular carcinoma precisely because the raw score underestimates their urgency. FIB-4 is confounded by age and by acute illness that transiently alters its blood components; it is a triage tool, not a diagnosis, and indeterminate results need a next step rather than a shrug.

The Alvarado score performs less well in children, older adults and women of childbearing age, where the differential diagnosis is wider and the score's components less specific. Ranson's criteria require 48 hours of data, which limits their use for the earliest decisions; BISAP was designed to fill that gap but, like all early scores, cannot predict every complication. The Glasgow-Blatchford score predicts the need for intervention, not the diagnosis, and a low score never justifies discharging a patient who looks unwell for other reasons. Light's criteria can misclassify effusions in patients on diuretics, where a transudate may meet exudative thresholds.

Thresholds also shift with context. A MELD score that is routine in a transplant centre may be alarming in a general ward; a FIB-4 result means something different at 35 than at 75. Scores standardise the starting point of a decision, not its endpoint.

How to use this library

Match the tool to the question you are actually asking. Staging cirrhosis calls for Child-Pugh; transplant urgency calls for MELD or MELD-Na; fatty liver triage calls for FIB-4 or the NAFLD fibrosis score. Enter the values as measured, note the score alongside the clinical findings in the record, and act on the band it falls into while keeping the patient's overall picture in view. For scores with a time component, such as Ranson's, set a reminder to complete the calculation rather than relying on memory. And when a score and your clinical impression disagree, investigate the disagreement: either the inputs are wrong or the patient is telling you something the score cannot see.

Related specialities

These calculators are often used alongside tools from neighbouring fields:

• Gastroenterology
• Surgery
• Critical Care
• Infectious Disease

Frequently asked questions