Liver function tests: patterns and interpretation

• "Liver function tests" (LFTs) are a panel, not a single test, and the name is misleading. Most of them tell you that liver cells or bile ducts have been injured; only albumin, the prothrombin time or international normalised ratio (INR), and bilirubin tell you much about how well the liver is working.
• Alanine aminotransferase (ALT) is the most liver-specific enzyme; aspartate aminotransferase (AST) is found in heart, muscle, kidney and brain as well.
• Alkaline phosphatase (ALP) comes from liver cells and bone (and placenta, gut and kidney). Gamma-glutamyl transferase (GGT) rises with hepatobiliary disease and is used to confirm that ALP is of liver rather than bone origin. ^[3]
• Work out the pattern from the ratio of ALT to ALP: R above 5 is hepatocellular, R below 2 is cholestatic, R of 2 to 5 is mixed (R = (ALT divided by its upper normal limit) divided by (ALP divided by its upper normal limit)). ^[2]
• An AST to ALT ratio of at least 2 to 1 points towards alcoholic liver disease, whereas a high AST with normal or low ALT is more often extrahepatic muscle damage. ^[2]
• Isolated unconjugated hyperbilirubinaemia with normal enzymes is commonly Gilbert syndrome and needs no treatment.
• A single borderline enzyme needs context, not panic: confirm it, look for risk factors, and image or refer when the pattern is cholestatic, the enzymes are high, or they persist.

By the end of this page you should be able to:

1. Name each component of a standard liver panel and say whether it mainly reflects injury or function.
2. Quote typical adult reference ranges in both SI and conventional units and remember that the laboratory's own range always wins.
3. Explain why AST is less liver-specific than ALT, why ALP can be high in bone disease, and how GGT helps.
4. Calculate and classify an R ratio as hepatocellular, cholestatic or mixed.
5. List the common causes of each pattern and give a sensible next investigation for each.
6. Recognise Gilbert syndrome, alcohol-related injury, and a normal panel, and interpret a short result sheet.

Liver panels are among the most commonly requested blood tests in medicine, and they are frequently ordered for people who have no symptoms at all. Mildly elevated aminotransferases are common in the general population, and many panels return to normal when the test is repeated after removing an obvious trigger, such as a medicine or alcohol. The British Society of Gastroenterology (BSG) guideline on abnormal liver blood tests states that the NICE guidelines were closely followed in its development, and it underpins much UK practice; in the US, ACG and AASLD teaching is the usual reference point. ^[1][2][3]

The panel has two repeating ideas. The first is injury: enzymes leak from damaged cells or are induced by disease. The second is function: albumin and the clotting factors reflect what the liver synthesises, and bilirubin reflects its capacity to take up, conjugate and excrete. A person can have a badly injured liver with a normal serum albumin (because the injury is acute) and a normal liver on imaging with a low albumin (because of inflammation, malnutrition or renal loss). Keeping injury and function separate is the most useful discipline in reading a liver panel. ^[2][3]

No single test diagnoses a given disease. The value comes from the pattern across the panel, the clinical context (alcohol, medicines, viral risks, metabolic disease, age, sex, pregnancy), and the trajectory on repeat testing. Ultrasound imaging and fibrosis assessment then add structure, and referral or biopsy follows when the diagnosis or stage remains uncertain. Standard clinical examination texts describe the same panel in the same way. ^[1][2][8]

Alanine aminotransferase (ALT) is the most liver-specific routinely available enzyme; it is found principally in hepatocytes, with only small amounts elsewhere, and it rises when hepatocytes are injured or die. Its elevation is specific but not sensitive in the sense that it tells you hepatocytes are hurt, not how widespread or reversible the hurt is. ^[2][3]

Aspartate aminotransferase (AST) is also a useful marker of hepatocellular injury, but it is far less specific: it occurs in cardiac muscle, skeletal muscle, kidney and brain as well as in the liver. A rise driven mainly by AST, especially with a normal ALT, should prompt a search for muscle, cardiac or haematological causes rather than assuming liver disease. ^[2][3]

Alkaline phosphatase (ALP) is a membrane-bound enzyme found on the canalicular membrane of hepatocytes. It rises with hepatobiliary disease, especially cholestasis, because of increased synthesis and release into the sinusoids; however, it is also found in bone, placenta, intestine and kidney, and roughly 80% of circulating ALP is of hepatic or bone origin. In children and elderly women (over 50 years of age), ALP rises in part because of bone turnover. ALP may also rise after a fatty meal in people with blood groups O and B because of intestinal ALP, and it rises in pregnancy because of the placental isoform. ^[2][3]

Gamma-glutamyl transferase (GGT) is a canalicular enzyme that rises with hepatobiliary disease. It is useful for confirming that an isolated ALP elevation is of hepatic rather than bone origin, but it is not specific for liver disease: it can rise with alcohol use, pancreatic disease, myocardial infarction, renal failure, emphysema, diabetes and drugs such as phenytoin and barbiturates, and in many people with alcohol misuse but no clear liver disease. For this reason, GGT should not be used as a screening test for liver disease when the rest of the panel is normal. ^[2]

Bilirubin is the yellow end-product of haemoglobin breakdown. Unconjugated (indirect) bilirubin is produced in the reticuloendothelial system, is tightly bound to albumin and is insoluble in water, so it does not appear in urine. The liver takes it up and conjugates it with glucuronic acid, making it water-soluble (conjugated or direct bilirubin), then excretes it into bile; bacteria in the colon convert it to urobilinogen, which is excreted in stool and urine. Normal total bilirubin is mostly (about 70%) unconjugated. An elevated unconjugated fraction usually points to overproduction (for example, haemolysis), reduced hepatic uptake, or impaired conjugation such as Gilbert syndrome. An elevated conjugated fraction points to hepatocellular damage or impaired bile flow. Unconjugated bilirubin is not excreted in urine; clay-coloured stools and dark urine arise once conjugated bilirubin is elevated. ^[2][3]

Albumin is the most abundant plasma protein and is synthesised exclusively by hepatocytes. Because it has a half-life of about three weeks, a low serum level usually indicates liver disease lasting more than three weeks, although it also falls in systemic inflammation, malnutrition, nephrotic syndrome, protein-losing enteropathy and fluid overload. Its fall is therefore a marker of chronic liver dysfunction rather than of acute injury. ^[2][3]

Prothrombin time and INR measure the extrinsic coagulation pathway and depend on factors synthesised by the liver (all except factor VIII). Because factor VII has a short half-life, the INR prolongs within about 24 hours of severe hepatocellular dysfunction, so it is a more time-sensitive marker of function than albumin. Prolongation can also result from vitamin K deficiency, cholestasis (vitamin K is not absorbed), warfarin, heparin bolus, disseminated intravascular coagulation and hypothermia. In chronic liver disease the INR is not a reliable predictor of bleeding risk even when it is high. ^[2][6]

Total protein, platelets and INR are often reported alongside the panel. A pattern such as a low platelet count with a prolonged INR and a low albumin suggests chronic portal hypertension and advanced fibrosis, even before imaging is performed.

Reference intervals differ between laboratories, assays, populations, ages and sexes. They are usually constructed to include most of a healthy population, and a small fraction of healthy people fall outside the printed range by design. A normal result also does not exclude significant liver fibrosis. The values below are those given by the MedlinePlus/NLM encyclopedia for adults. Use your own laboratory's printed range as the final arbiter: each printed range is appropriate for the assay and population that the laboratory validated. ^[4][5][6][1]

Notes on interpretation:

• Sex and age. ALP is higher in adolescents and in women over 50 years of age, and rises in children and in pregnancy. AST and ALT ULNs also vary by sex; MedlinePlus quotes single ranges, but many laboratories print different male and female intervals. ^[2][4]
• "True" healthy ALT. The ACG guideline argues that population ranges are inflated by undiagnosed fatty liver. In a truly healthy population it proposes upper limits of ALT of 29 to 33 U/L in males and 19 to 25 U/L in females, and levels above these should be assessed. ^[2]
• Fractionation. Total bilirubin is the sum of conjugated and unconjugated portions. Fractionation is most useful when ALT, AST and ALP are normal or near normal. ^[2]
• Disease without laboratory abnormality. Both AST and ALT can be normal even in the setting of cirrhosis, and the utility of the AST to ALT ratio persists even when both values are within the normal reference interval. ^[1]
• SI units. The MedlinePlus ranges above are already paired in SI and conventional units; your own laboratory report lists both. StatPearls teaching emphasises reading the laboratory's own range before interpreting any of these numbers. ^[7]

The table below lists the causes most commonly associated with abnormalities of each component. It is not exhaustive, and many conditions produce more than one abnormality because hepatocytes and bile ducts inflame or fail together. ^[2][3]

Specific clinical rules worth remembering, with their sources:

• AST to ALT ratio. In alcoholic liver disease an AST to ALT ratio of at least 2 to 1, with AST or ALT rarely exceeding 300 U/L, is classical; a ratio exceeding 3 to 1 strengthens the probability. The ratio works because of pyridoxine depletion (reducing ALT synthesis) and the release of mitochondrial AST. In advanced fibrosis or cirrhosis from any cause the ratio also tends to rise above 1. ^{[2] [1]}
• Degree of ALT elevation. The ACG guideline classes borderline elevation as under 2 times the upper limit of normal (ULN), mild as 2 to 5 times, moderate as 5 to 15 times, severe as over 15 times, and massive as over 10,000 U/L. ^[2]
• GGT and alcohol. GGT can be raised in over 50% of patients with alcohol misuse without obvious liver disease, but it is a useful behaviour-change marker rather than a diagnostic one. ^[1][2]
• Bilirubin and history. In intrahepatic cholestasis, light-coloured stools, pruritus and dark urine may be elicited; in haemolytic presentations, a history of blood disorder is relevant. ^[2]
• Medications. Statins can raise ALT but are safe in chronic liver disease; hepatotoxicity is rare. An ALT above 1,000 U/L requires an explicit history of acetaminophen use. ^[2]

The key step is deciding whether the injury is hepatocellular, cholestatic, mixed or isolated hyperbilirubinaemia. The tool is the R ratio:

R = (ALT divided by its upper normal limit) divided by (ALP divided by its upper normal limit).

• R above 5: hepatocellular. The aminotransferases predominate; ALP may be normal or modestly raised. Typical causes: viral hepatitis, drug-induced or toxin-induced injury, alcohol (often with AST predominance), non-alcoholic fatty liver disease or steatohepatitis, autoimmune hepatitis, ischaemic or hypoxic hepatitis ("shock liver"), Wilson disease, haemochromatosis. ^[2][3]
• R below 2: cholestatic. ALP predominate (and GGT supports hepatic origin). Typical causes: choledocholithiasis and biliary obstruction, primary biliary cholangitis, primary sclerosing cholangitis, drug-induced cholestasis, sepsis, total parenteral nutrition, vanishing bile duct syndrome, sarcoidosis, malignancy. Cholestatic disease often presents with elevated ALP, with or without bilirubin elevation, and imaging is needed to separate intrahepatic from extrahepatic disease. ^[2][3]
• R of 2 to 5: mixed. Features of both; drug-induced liver injury can look hepatocellular, cholestatic or mixed depending on the agent. ^[2][3]
• Isolated hyperbilirubinaemia rises bilirubin with normal ALT, AST and ALP. If the rise is predominantly unconjugated, think haemolysis, Gilbert syndrome, fasting and the rare Crigler-Najjar syndrome; if conjugated (or mixed), think sepsis, cirrhosis, alcohol-related hepatitis, conjugated inherited syndromes (Dubin-Johnson and Rotor) and biliary disease. ^[2][3]

Practical patterns that sharpen diagnosis:

• Gilbert syndrome. A benign, inherited reduction in UDP-glucuronosyltransferase activity producing mild unconjugated hyperbilirubinaemia, with normal aminotransferases and ALP. It worsens with fasting, illness, sleep deprivation, strenuous exercise and alcohol, and improves with feeding. It needs no treatment and no investigation once the rest of the panel is normal. ^[7]
• Alcohol. AST above ALT, often with raised GGT and an AST to ALT ratio of at least 2 to 1. ^[2]
• Advanced fibrosis. The AST to ALT ratio rises above 1 in cirrhosis from any cause, although it may be less pronounced than in alcohol-related disease; even normal enzymes do not exclude cirrhosis, and in late disease aminotransferases can fall as hepatocyte mass falls. ^[1][2]
• Hepatocellular versus cholestatic. The degree of elevation, the pattern and the clinical scenario guide the work-up: consider viral and metabolic serology, iron studies, autoimmune markers and an abdominal ultrasound at baseline for any abnormality. ^[2]

All values below are illustrative samples invented for teaching. They are not real patient data. Reference ranges are the adult MedlinePlus ranges from the table above; every threshold and rule used below is cited to its source. ^[4][2]

Interpretation:

1. Compare each value with the laboratory's printed reference range. All seven components fall inside their adult MedlinePlus ranges. ^[4][6]
2. Because ALT and AST predominate in magnitude relative to ALP only trivially, there is no hepatocellular or cholestatic injury pattern. ^[2]
3. Albumin and INR are normal, so there is no evidence of impaired synthetic function on this sheet. ^[2]
4. No fractionation or imaging is needed when the whole panel is normal and the patient has no risk factors. ^[2]

Clinical conclusion: this is a normal liver panel, and as long as there are no red flags or clinical symptoms it should be managed with routine primary-care follow-up rather than further investigation. Next step: address risk factors and repeat the panel only if symptoms, new drugs or abnormal findings appear.

Interpretation:

1. ALT at 312 U/L is about 8.7 times the upper limit of a typical adult range, AST is about 6 times, while ALP at 96 U/L is within its range. These are hepatocellular values. ^[2][4]
2. The R ratio is (312 divided by 36) divided by (96 divided by 130), which is about 11.7, so this is a hepatocellular pattern (R above 5). ^[2]
3. The AST to ALT ratio is 198 to 312, about 0.6, which is below 2. A ratio of at least 2 to 1 would have pointed towards alcoholic liver disease, so alcohol is not the likely dominant driver on this sheet. ^[2]
4. Albumin and INR are normal, indicating preserved synthetic function; bilirubin is only mildly raised, so decompensation is not evident on this sheet. ^[2]
5. A moderate-to-severe hepatocellular rise (5 to 15 times ULN) warrants testing for viral hepatitis A, B and C, iron studies, immune and metabolic markers, and a careful medicine and toxin history. ^[2]

Clinical conclusion: this is an acute hepatocellular injury with preserved function and no alcohol pattern, and the next step is a same-week hepatology-oriented work-up with hepatitis serology, iron studies, autoimmune and metabolic tests, an abdominal ultrasound, and urgent review if jaundice, confusion or INR rise develops.

Interpretation:

1. ALP at 340 U/L is about 2.6 times its upper limit and GGT at 210 U/L is about 5 times its upper limit, both clearly abnormal; ALT and AST are only mildly raised. This is a cholestatic profile. ^[2][4][5]
2. The R ratio is (62 divided by 36) divided by (340 divided by 130), which is about 0.66, so R is below 2, confirming a cholestatic pattern. ^[2]
3. GGT is raised, so the elevated ALP is of hepatic rather than bone origin. ^[2]
4. Conjugated hyperbilirubinaemia with a cholestatic enzyme picture implies hepatocellular dysfunction or impaired bile flow, and the slightly low albumin is in keeping with prolonged cholestasis. ^[2]
5. Primary biliary cholangitis and primary sclerosing cholangitis are the autoimmune cholestatic diseases to consider, and extrahepatic obstruction must be excluded by imaging. ^[2]

Clinical conclusion: a cholestatic injury pattern with a raised conjugated bilirubin must be imaged before or alongside autoantibody testing, so the next step is an abdominal ultrasound (and MRCP or ERCP if the ducts are dilated) together with AMA and p-ANCA assessment, with early referral if obstruction or jaundice is confirmed.

• "Liver function tests" is a misnomer. ALT, AST, ALP and GGT are tests of injury or enzyme induction; only albumin, INR and bilirubin say much about function. Calling them all "function tests" leads to over-investigation of an injured but functioning liver and under-investigation of a failing one. ^[2][3]
• AST is not liver-specific. Always glance at ALT and the clinical picture before attributing an AST rise to the liver. ^[2][3]
• ALP is not liver-specific either. In children, in women over 50, in pregnancy and after a fatty meal it can be raised of bone, placental or intestinal origin. Confirm hepatic origin with GGT, or fractionate ALP, before treating it as cholestasis. ^[2][3]
• A normal test does not exclude disease. Aminotransferases and ALP can all be normal in cirrhosis, and the AST to ALT ratio is still informative even when both are normal. ^[1]
• The opposite also holds: abnormal need not mean disease. By definition 5% of healthy people fall outside the 95% reference interval, so a single borderline value should be repeated and put in context before labelling anyone. ^[4]
• ALT is not stable. Effective hepatitis C treatment can lower ALT further even when it was already within the "normal" range, and muscle, exercise or specimen problems move AST and ALT around. ^[2]
• INR is not a bleeding predictor in cirrhosis. It measures synthetic function, not haemostasis, in chronic liver disease. ^[3]
• Gilbert syndrome fluctuates. Values rise with fasting, illness, exercise and sleep loss and fall after feeding, which can make a patient look as if they have a labile liver disease. ^[7]
• GGT is a poor screening test. It lacks specificity and should not be ordered to screen for liver disease in an otherwise normal panel. ^[2]
• Panels vary. The values of ALP in late pregnancy, of ALP in adolescents, or of bilirubin in a jaundiced neonate carry meanings that adult ranges do not, and paediatric ranges must be used for paediatric patients. ^[4]
• Composition of the sample matters. Differences in assay method and reference population, especially for ALT, mean that each laboratory's printed range deserves respect. ^[2]

• "Liver function tests" are mostly injury enzymes; albumin, INR and bilirubin give the functional information.
• ALT is the most liver-specific enzyme; AST is present in several non-hepatic tissues.
• ALP rises in liver and bone disease; confirm hepatic origin with GGT, remembering GGT's poor specificity.
• R above 5 is hepatocellular, R below 2 is cholestatic, and R of 2 to 5 is mixed, using R = (ALT divided by its ULN) divided by (ALP divided by its ULN).
• Hepatocellular injury: viral, drug, alcohol, NAFLD, autoimmune, metabolic and vascular causes; cholestatic: obstruction and autoimmune bile-duct disease, drugs, sepsis and TPN.
• An AST to ALT ratio of at least 2 to 1 points towards alcohol; ratio above 1 in cirrhosis of any cause.
• Gilbert syndrome causes mild unconjugated hyperbilirubinaemia with otherwise normal results and needs no treatment.
• Conjugated bilirubin rises in hepatocellular and cholestatic disease and its presence darkens urine and lightens stool.
• Albumin falls only after weeks of liver disease but can fall sooner in systemic illness, malnutrition or renal or gut loss.
• INR changes within about a day of severe liver injury but does not predict bleeding in cirrhosis.
• Normal liver enzymes do not exclude advanced fibrosis, and a single abnormal value need not indicate disease.
• Statins are safe in chronic liver disease; an ALT above 1,000 U/L should trigger an acetaminophen history.
• The laboratory's own printed range, clinical context and repeat testing matter more than any single published cut-off.