The interesting thing about a liver panel isn't any single value — it's that the same blood draw, read properly, catches fatty liver, hepatitis C, hereditary iron overload, an autoimmune bile-duct disease, and the harmless genetic quirk called Gilbert's syndrome that affects roughly one person in fifteen. None of those diseases announce themselves; most quietly fold the liver into scar tissue over decades. Learning the framework is an evening's work. After that every yearly check-up is a few seconds of triage instead of a vague thumbs-up.
Think of the six markers as three families.
ALT and AST are enzymes that live inside liver cells. When the cells get damaged or die, they leak into the blood, and the level on the printout is roughly how much leaked. ALT is the more liver-specific of the two — almost all of it lives in the liver. AST also lives in heart muscle, skeletal muscle, and red blood cells, so a high AST without a high ALT can be muscle (a hard workout, a recent fall) rather than liver Kwo et al. 2017.
ALP and GGT sit on the walls of the tiny bile ducts inside the liver and rise when bile can't flow out — blocked plumbing rather than damaged cells. ALP is sensitive but not specific; bone tissue makes the same enzyme, so a growing teenager, a pregnant woman, or someone with Paget's disease can have a high ALP that has nothing to do with the liver. GGT is the tiebreaker: if ALP is up and GGT is normal, the source is bone, not liver Whitfield 2001.
Bilirubin and albumin are the only two that measure what the liver is actually doing, in opposite directions. Bilirubin is a waste product the liver is supposed to clear; a high level means clearance has slowed somewhere along the way. Albumin is a protein the liver is supposed to make; a low level — in a stable person without protein-leaking kidneys or active inflammation — means the factory is failing. Albumin has a long fuse, around three weeks, so it shifts late; a chronically low albumin signals chronic disease, not yesterday's problem Lala et al. 2024.
The label LFTs — liver function tests — is misleading on its face. Five of the six numbers don't measure function. They measure injury or blockage. Keep that split clear: a panel with ALT 200 and a perfect albumin is a liver getting hit but still working; a panel with ALT 30 and albumin 2.5 is a liver that has stopped being hit because most of it is gone Kwo et al. 2017.
The three-pattern framework
Once you can read the markers individually, the actual interpretive skill is fast: sort the panel into one of three patterns, then size the magnitude, then check for the danger signals.
The R-ratio tells you which kind of problem you're looking at, which decides the rest of the workup. A hepatocellular panel goes one way (viral hepatitis testing, iron studies, autoimmune markers, ultrasound). A cholestatic panel goes another (ultrasound first to look for blocked ducts, then specific antibody tests). They don't share much downstream.
How big the numbers are matters too
Within the hepatocellular pattern, the size of the ALT rise narrows the cause sharply. ALT more than fifteen times the upper limit — so above roughly 500 IU/L in most labs — almost always means acute viral hepatitis, an acetaminophen overdose, a sudden drop in blood flow to the liver, or a fierce autoimmune flare. Between five and fifteen times normal opens up to chronic viral hepatitis on the boil, alcohol-related hepatitis, or drug reactions. Below five times normal is the everyday range, and the everyday cause is fatty liver — by a wide margin in the modern population Kwo et al. 2017Younossi et al. 2016.
The AST-to-ALT ratio — useful, often misread
The classic teaching is that AST more than twice ALT means alcohol. Half right. The ratio does rise in heavy drinking, but it also rises in any cirrhosis regardless of cause, because the scarred liver releases more of the mitochondrial form of AST and clears ALT differently. A ratio greater than 2 in a person who doesn't drink isn't vindication of the drinking question; it's a flag for advanced disease Nyblom et al. 2004Cohen and Kaplan 1979.
One danger signal that overrides everything
What missing the pattern costs you
The diseases this panel catches early don't announce themselves. They progress through years of feeling fine. Fatty liver — the modern epidemic, found in about a quarter of adults globally — sits silently in the liver of the slightly overweight person for a decade before it tips into the inflammatory version, and another decade before it tips into scarring serious enough to limit life Younossi et al. 2016Adams et al. 2005. The person living through that decade feels normal. The five-second look at the printout is the only signal.
Chronic hepatitis C is the same shape: years of quiet ALT bumps that the patient and the family doctor both shrug off, then a hospital admission for variceal bleeding or a cirrhosis diagnosis at sixty that an eight-week course of antiviral pills at forty would have cured outright. Hereditary iron overload, common in people of Northern European descent at roughly one in two hundred, the same shape again: silent ferritin and ALT creep across the thirties, joint pain and bronze-tinted skin in the fifties, diabetes and heart failure if it gets that far. Primary biliary cholangitis usually presents as nothing more than a quietly elevated ALP in a middle-aged woman; ten years of unmedicated progression takes the bile ducts apart in ways a single daily pill would have largely prevented Lavanchy 2009Lindor et al. 2009.
The story each of these tells the patient at sixty is the same: nobody ever told me my numbers meant anything. The numbers were on every annual blood draw for fifteen years. The panel was glanced at and filed.
The flip side cost is real but less catastrophic. Reading the panel in the wrong direction — panicking at a transient ALT of 60 — produces unnecessary ultrasounds and weeks of low-grade dread. The repeat draw, taken seriously, prevents most of that.
How to actually read your panel
The sequence is the same every time. A minute, once you've internalised it.
For a hepatocellular panel that doesn't clear on the repeat, the standard workup runs a hepatitis B and C panel, ferritin and transferrin saturation (for hereditary iron overload), autoimmune liver antibodies, and an ultrasound of the liver. Anyone under forty also gets a ceruloplasmin level, for the rare but treatable copper-overload condition called Wilson's disease.
For a cholestatic panel the ultrasound comes first — to see whether the bile ducts are dilated, which would mean a physical blockage somewhere — and if they aren't, an antimitochondrial antibody test (for primary biliary cholangitis, classically a middle-aged woman with an isolated ALP rise) and a more detailed scan called MRCP follow Kwo et al. 2017Newsome et al. 2018Lindor et al. 2009.
What most people get wrong
"My liver enzymes were high" doesn't mean your liver is failing. Enzymes leak when cells are getting damaged; they tell you about the assault, not the surviving capacity. A normal albumin and a normal INR mean the factory is still running, even if it's taking fire. The opposite is also true: enzymes can be near-normal in advanced cirrhosis, because so much of the liver has been replaced by scar tissue that there's little parenchyma left to leak Newsome et al. 2018.
An "elevated" lab marker isn't automatically elevated. The reference ranges most US labs print are based on healthy-donor cohorts from the 1990s, before fatty liver and hepatitis C were screened out of the donor pool. The modern evidence-based upper limit for ALT is closer to 30 in men and 19 in women, but most printouts still flag at 40 or 50 — meaning your "normal" 38 may genuinely be normal, or may be a real signal the lab range was too lenient to catch Prati et al. 2002Ruhl and Everhart 2012.
A high bilirubin alone is almost never a disease. Roughly one person in fifteen carries the genetic variant for Gilbert's syndrome, in which the liver's bilirubin-processing step works a bit more slowly than average. Bilirubin drifts up during a fast, a flu, or after a hard workout, and the printout flags it. There's no progression, no treatment, no "fix." If the rest of the panel is clean and the high fraction is the unconjugated kind, the worry stops here Strassburg 2008.
A high GGT on its own isn't a hidden alarm. GGT rises with alcohol, with extra weight, with several common medications (statins, anticonvulsants), and with fatty liver. It's sensitive to almost everything and specific to nothing. In a person with otherwise normal numbers, a high GGT rarely points at occult disease beyond what a five-minute lifestyle conversation already names Whitfield 2001.
Where reading goes wrong
Acting on a single set of labs. A 30% swing in ALT between two draws in the same healthy person is well within biological noise. A new medication, a viral cold, a weekend of heavy training can each bump the numbers transiently. The discipline for any mild, asymptomatic elevation is a repeat in two to four weeks before any imaging or further testing — most of the borderline cases resolve Lazo et al. 2008.
Calling everything "fatty liver" and stopping there. Fatty liver is the most common cause of a mild ALT bump in the modern adult — affecting roughly a quarter of the global adult population — but hepatitis C, hereditary iron overload, autoimmune hepatitis, and Wilson's disease all present identically on a routine panel. The workup matters even when the metabolic story seems to fit, because the metabolic story usually does fit (most people carry extra weight these days), and the other diagnoses get missed by default Younossi et al. 2016Lavanchy 2009.
Missing the bone source for an isolated ALP. ALP also comes from bone. A pregnant woman in her third trimester, a teenager in a growth spurt, an older adult with Paget's disease or a healing fracture, a person with vitamin D deficiency — all can show ALP elevation that has nothing to do with the liver. The GGT check is the cheap version of fractionating the enzyme; if GGT is normal alongside, the ALP is bone.
Trusting the printed reference range as the bar. Lab reference ranges are slow to update and vary by instrument. Treat them as a starting point, not the truth. The modern thresholds (ALT under about 30 in men, under about 19 in women) are the better evidence-based bar, and several major hepatology societies now teach them Prati et al. 2002.
Ignoring Hy's Law on a new medication. The combination of ALT more than three times normal plus bilirubin more than twice normal in someone on a new drug is the highest-stakes pattern on the panel. Acting fast (stopping the drug, calling the prescriber) is the default; "let's recheck in a month" is not Reuben 2004.
What's next if the panel points somewhere
If the panel triages toward a real diagnosis, the next layer of tools sits outside this entry: a FibroScan or transient elastography to estimate how much scarring is already there; MRCP imaging to look at the bile ducts in detail; the AST-to-platelet ratio as a rough fibrosis index; the MELD score for staging advanced disease. Each belongs to a follow-up conversation with a hepatologist, not to the panel itself.
If the panel points at fatty liver — the most likely outcome for the most common reader — the action question is its own subject (weight loss, alcohol, diet, the role of GLP-1 drugs, sleep). Same for hepatitis B vaccination, hemochromatosis phlebotomy, and the larger question of whether to drink at all.
- — If the panel points at fatty liver, GLP-1 drugs are one of the levers that actually move those numbers back down.
- — A quiet ALT bump with a thick waistline is most often fatty liver — the single most likely thing this panel is trying to flag.
- — The same blood draw can catch hereditary iron overload years before symptoms — one of the diagnoses hiding in an abnormal panel.
- — A high AST-to-ALT ratio with a raised GGT is the classic fingerprint of alcohol on the liver.
- — Supplements show up on liver tests too. Ashwagandha has caused real liver injury, so flag it if your ALT or AST comes back high.
- — Turmeric extracts are a documented cause of drug-induced liver injury; if your enzymes climb, this is on the suspect list.
- — A textbook-clean panel can sit on top of real disease — knowing how to read 'normal' versus meaningful is the same skill applied to liver numbers.
- — Red yeast rice is a statin in disguise, and like any statin it can nudge your liver enzymes — worth a panel if you take it.
- — The fatty-liver drug resmetirom is tracked by these exact enzymes — your liver panel is how you and your doctor watch it work.
- — Like the liver panel, the kidney's eGFR is an estimate worth learning to read for yourself.
Substance + claimed effects
"Reading a liver panel" is the interpretive skill — applied to the six serum markers that routine chemistry panels report under LFTs or hepatic panel: alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), total and direct bilirubin, and albumin. The claim is that competent pattern-recognition over these six values discriminates the three principal injury syndromes — hepatocellular (predominant aminotransferase elevation), cholestatic (predominant ALP/GGT elevation), and isolated hyperbilirubinemia (often benign Gilbert's, sometimes hemolysis or rare inherited transport defects) — and that this triage step determines an evidence-based downstream workup Kwo et al. 2017Newsome et al. 2018. Consequences scored holistically: longevity (early detection of NAFLD, chronic viral hepatitis, hemochromatosis, autoimmune hepatitis, PBC — all of which progress silently to cirrhosis if missed); short-term health (resolving a borderline ALT in 3-6 months avoids invasive workup for a transient drug effect); mood (interpretive literacy converts a panic-inducing flag into a graded plan); cost/effort burden (the panel is standard, free or near-free; the skill itself is one-time learned).
Evidence by addressing question
mechanism
ALT (alanine aminotransferase, formerly SGPT) is concentrated in hepatocyte cytosol with minor contributions from kidney and muscle; its release into serum reflects hepatocellular membrane injury — the cell either died (necrosis) or had a transient leak. ALT is the most liver-specific of the panel Kwo et al. 2017Lala et al. 2024. AST (aspartate aminotransferase, SGOT) exists in two isoforms — cytosolic and mitochondrial — and is abundant in liver, cardiac and skeletal muscle, kidney, brain, and red blood cells. The mitochondrial fraction predominates in heavy alcohol use (alcohol depletes pyridoxal-5'-phosphate, which ALT depends on more than AST, and induces mitochondrial AST release), producing the classic AST:ALT ratio >2 of alcoholic liver disease Cohen & Kaplan 1979Nyblom et al. 2004. ALP (alkaline phosphatase) is a family of isoenzymes on the canalicular membrane of hepatocytes, bile-duct epithelium, osteoblasts, placenta, and intestine; biliary obstruction or intrahepatic cholestasis upregulates its synthesis and releases it into serum, but bone turnover (growing children, pregnancy, Paget's, healing fracture, bone metastases) and placental production (third trimester) confound the liver signal Kwo et al. 2017. GGT (gamma-glutamyl transferase) is bound to hepatocyte canalicular and biliary epithelial membranes; it rises with any cholestatic process and is strongly inducible by alcohol, phenytoin, barbiturates, and other microsomal enzyme inducers. Its role in the panel is confirmatory: an isolated ALP elevation with a normal GGT is bone, not liver Whitfield 2001. Bilirubin partitions into unconjugated (indirect; bound to albumin in serum; produced from heme catabolism by macrophages) and conjugated (direct; glucuronidated in hepatocytes by UGT1A1; excreted in bile). Unconjugated hyperbilirubinemia points to overproduction (hemolysis) or impaired conjugation (Gilbert's, Crigler-Najjar); conjugated hyperbilirubinemia points to hepatocellular dysfunction, intrahepatic cholestasis, or biliary obstruction Strassburg 2008. Albumin is synthesized exclusively by hepatocytes with a serum half-life of ~21 days; a low albumin in a stable patient with no proteinuria or inflammatory state reflects synthetic dysfunction and signals chronic disease, not acute injury Lala et al. 2024.
evidence
The three-pattern framework (hepatocellular vs. cholestatic vs. mixed) is operationalised by the R-ratio: R = (ALT / ALT upper limit of normal) / (ALP / ALP upper limit of normal). R > 5 is hepatocellular; R < 2 is cholestatic; R 2–5 is mixed Kwo et al. 2017Newsome et al. 2018. The framework is endorsed by both the American College of Gastroenterology (Kwo et al. 2017) and the British Society of Gastroenterology (Newsome et al. 2018) and underlies regulatory definitions of drug-induced liver injury (DILI). Magnitude bands further constrain differential: ALT/AST >15× ULN narrows the differential to acute viral hepatitis, ischemic hepatitis ("shock liver"), acetaminophen toxicity, autoimmune hepatitis flare, and acute biliary obstruction; 5–15× ULN includes chronic viral hepatitis flare, drug-induced injury, alcoholic hepatitis; <5× ULN is dominated by NAFLD/MASLD, mild alcohol use, and medications Kwo et al. 2017Hultcrantz et al. 1986. Reference-range epidemiology is contested: the historical ULN for ALT (~40 IU/L) was derived from "healthy" blood-donor cohorts that included undiagnosed NAFLD and HCV. Prati et al. (2002), excluding subjects with risk factors, derived ULN of 30 IU/L for men and 19 IU/L for women Prati et al. 2002; the AASLD adopted similar lower thresholds, though many US labs still report the older ranges Ruhl & Everhart 2012. The clinical impact is real: a Korean prospective cohort of 142,055 workers found that liver-related mortality rises continuously with ALT above ~20 IU/L, with no inflection at the lab's stated ULN Kim et al. 2004. Short-term biological variability is also substantial: a 30% test-retest difference in ALT is within physiologic noise; transient elevations after intense exercise, viral infection, or new medication need repeat testing in 2–4 weeks before workup escalates Lazo et al. 2008. Hy's Law, a regulatory rule from drug-safety science, states that an ALT >3× ULN with concurrent total bilirubin >2× ULN in the absence of obstruction predicts severe drug-induced liver injury with ~10% mortality — a high-specificity stopping rule for any new medication Reuben 2004Lewis 2007.
protocol
Interpretive sequence supported by guidelines Kwo et al. 2017Newsome et al. 2018: (1) classify magnitude (mild <5×, moderate 5–15×, marked >15× ULN); (2) compute R-ratio to assign hepatocellular / mixed / cholestatic; (3) characterize bilirubin (isolated unconjugated vs. mixed/conjugated); (4) check synthetic markers (albumin, INR) — if abnormal alongside any of the above, this is decompensated disease and requires urgent referral. Workup branches: hepatocellular pattern triggers HBsAg, anti-HCV, ferritin/transferrin saturation (hemochromatosis), ANA/ASMA/IgG (autoimmune hepatitis), ceruloplasmin in patients <40 (Wilson's), alpha-1 antitrypsin level/phenotype, fasting lipids and glucose for NAFLD risk-staging, and ultrasound; cholestatic pattern triggers right-upper-quadrant ultrasound first to distinguish obstructive (dilated ducts → MRCP or ERCP) from non-obstructive (AMA for primary biliary cholangitis, MRCP for primary sclerosing cholangitis), drug/herbal review, and consideration of infiltrative disease; isolated unconjugated hyperbilirubinemia with normal hemoglobin, haptoglobin, LDH, reticulocyte count and peripheral smear is Gilbert's syndrome by exclusion and needs no further workup Strassburg 2008. The Kwo et al. (2017) ACG guideline formalises this into branching algorithms by R-ratio and magnitude band. Borderline isolated mild elevations (ALT 1–2× ULN) in an asymptomatic patient with no risk factors warrant repeat testing in 2–4 weeks before any imaging or serology — many resolve Lazo et al. 2008.
misconceptions
(a) "LFTs" is a misnomer. ALT, AST, ALP, GGT, and bilirubin measure liver injury or cholestasis, not function. The only true function markers on the panel are albumin and (indirectly, on a coagulation panel) INR/prothrombin time — both reflecting synthetic capacity. A patient with normal albumin and INR but ALT of 200 has injury without dysfunction; a patient with ALT 30 and albumin 2.5 and INR 1.8 has end-stage dysfunction with little ongoing injury Lala et al. 2024Kwo et al. 2017. (b) The enzyme magnitude does not correlate with the severity of underlying disease in chronic liver injury. Cirrhosis with portal hypertension can present with normal or near-normal ALT/AST because most hepatocytes have been replaced by fibrosis — there is little parenchyma left to leak. Conversely, a young patient with acute viral hepatitis A can have ALT of 3,000 IU/L and recover completely Newsome et al. 2018. (c) AST:ALT >2 is not pathognomonic for active drinking. The ratio rises in advanced (cirrhotic) liver disease of any etiology because the cirrhotic liver releases more mitochondrial AST and ALT clearance changes Nyblom et al. 2004. (d) Isolated unconjugated hyperbilirubinemia in a young, otherwise-well patient is almost always Gilbert's. Prevalence ~5–10% of the population; the underlying UGT1A1 promoter polymorphism is benign and the bilirubin fluctuates with fasting, illness, or strenuous exercise; "normalising" it is unnecessary and impossible Strassburg 2008. (e) GGT in isolation is not a disease marker. It is exquisitely sensitive but non-specific — alcohol, NAFLD, medications (statins, anticonvulsants), and obesity all elevate it. A 30-year-old with GGT 120 and otherwise normal panel almost never has occult disease beyond what lifestyle assessment already names Whitfield 2001.
failure-modes
(a) One-shot interpretation. Acting on a single set of labs without confirming with a repeat in 2–4 weeks for borderline elevations leads to over-investigation; biological variability in ALT alone is ~10–30% test-to-test in healthy controls Lazo et al. 2008. (b) Attributing borderline ALT to "fatty liver" without checking risk factors. NAFLD is the most common cause in metabolic-syndrome patients, but hepatitis C, hemochromatosis (1:200–1:300 in Northern European populations), and alcohol can present identically Younossi et al. 2016Lavanchy 2009. (c) Missing the bone source for an isolated ALP elevation. Without fractionating ALP into bone and liver isoenzymes (or using GGT as a surrogate), a Paget's disease or vitamin D deficiency picture is mistaken for cholestasis. (d) Trusting the lab's printed reference range as the bar. Reference ranges in most US labs are 1990s-era and 30–50% too high for ALT relative to modern healthy-donor cohorts Prati et al. 2002Ruhl & Everhart 2012. (e) Macroenzymes. Rarely, AST or ALP forms immunoglobulin complexes that delay clearance; serum levels are persistently elevated without disease. Suspected when one isolated marker stays elevated indefinitely without clinical correlate; confirmed by polyethylene glycol precipitation or electrophoresis Lala et al. 2024. (f) Failing to think about Hy's Law on a new medication. ALT >3× ULN with bilirubin >2× ULN on a new drug is a stopping signal — not a wait-and-watch Reuben 2004.
stakes
The stakes of mis-reading or ignoring the panel are not subtle. NAFLD now affects ~25% of adults globally Younossi et al. 2016, with progression to NASH in a substantial minority and cirrhosis/HCC in a smaller-still but real fraction; ALT is the cheapest early signal Adams et al. 2005. Chronic hepatitis C affects ~70 million people worldwide and is curable with direct-acting antivirals — but only if detected, which often starts with an unexplained ALT elevation Lavanchy 2009. Hereditary hemochromatosis (1:200 in Northern European populations) progresses silently to cirrhosis, cardiomyopathy, and diabetes if not caught at the ferritin/transferrin-saturation stage — which is typically triggered by an unexplained ALT bump. Primary biliary cholangitis presents as isolated ALP elevation in middle-aged women; missed early, it progresses to cirrhosis over years; treated with ursodeoxycholic acid, the trajectory is largely arrested Lindor et al. 2009. The asymmetric cost of mis-reading: a missed pattern on a single panel rarely changes a 3-month outcome but routinely changes a 10-year one. Conversely, anxious over-reading produces unnecessary CT scans, biopsies, and specialist referrals for transient ALT bumps that would have normalised on repeat testing.
payoff
The payoff of competent interpretation is graded. Most readers — those with normal panels — gain reassurance and a stable framework for future panels (knowing what "ALT 28, AST 24, ALP 80, bilirubin 0.6" actually says about the next decade of liver-related risk). Readers with mild ALT elevations and metabolic risk factors gain a 3–6 month window to act on NAFLD via weight loss and dietary change, demonstrably arresting and often reversing early steatosis EASL/EASD/EASO 2016. Readers with isolated bilirubin elevations gain freedom from years of intermittent worry. Readers in a Hy's-Law-triggering medication regimen avoid acute liver failure. The longest-tail payoff is the rare-but-real chronic disease (PBC, hemochromatosis, autoimmune hepatitis) detected a decade earlier than it would have been, when treatment still meaningfully changes outcome.
out-of-scope
Outside the scope of this entry: hepatic synthetic-function staging beyond what albumin and INR signal (MELD, Child-Pugh); platelet count and the AST-to-platelet ratio index (APRI) as fibrosis markers (related but a separate skill); imaging modalities (ultrasound, transient elastography / FibroScan, MRCP) — these are next-step tools, not panel-interpretation steps; specific treatment of any of the named diseases. NAFLD lifestyle interventions, viral hepatitis vaccination, alcohol modification, hemochromatosis phlebotomy each warrant their own entries.
The credibility range
The optimist case. Liver-panel interpretation is one of the most settled skills in primary care. The R-ratio framework is replicated across ACG, BSG, and EASL guidelines Kwo et al. 2017Newsome et al. 2018EASL/EASD/EASO 2016; magnitude-band differential is taught identically across hepatology fellowships globally; Hy's Law has shaped drug regulation for two decades Reuben 2004. A motivated reader can master pattern-recognition over an evening and apply it for life. The downstream prognostic value is well-documented: ALT predicts liver-related mortality continuously across the population Kim et al. 2004, and the conditions a competent reading catches early (NAFLD, viral hepatitis, PBC, hemochromatosis) have meaningful treatments. This is exactly the kind of literacy the catalogue exists to confer.
The skeptic case. Most "elevated" ALT/AST values in primary care are transient and resolve on repeat without intervention Lazo et al. 2008; teaching readers to scrutinise the panel risks producing health anxiety in a population whose pre-test probability of consequential liver disease is low. Reference ranges are unstable across labs and over time, so a "framework" applied to lab printouts is brittle. The R-ratio cleanly separates a textbook hepatocellular from a textbook cholestatic case but loses discriminatory power in mixed presentations and in early disease where all enzymes are mildly up. Pattern recognition without imaging and serology is incomplete: the panel triages, but it doesn't diagnose. Finally, the conditions caught early — NAFLD especially — often have no pharmacotherapy and resist behaviour change; detection doesn't equal benefit.
The author's call. Lands close to the optimist case with a calibration caveat. The interpretive skill is real, the framework is guideline-grade, and the downstream conditions matter. The two honest hedges: (a) the framework must be paired with the discipline to repeat-test mild elevations before escalating workup, and (b) the meta scores should reflect that this is knowledge, not action — the entry teaches the reader to read, not to treat. Evidence: 4 (multiple guideline bodies converge; the framework is decades-old and replicated; some unsettled questions on optimal ULN thresholds). Controversy: 2 (ULN debate is live; the framework itself is not).
Stakeholder + incentive map
- Hepatology societies (AASLD, EASL, BSG, ACG) — push for wider interpretive literacy and lower ALT thresholds to catch NAFLD/HCV earlier; aligned with the catalogue.
- Primary-care clinicians — own the first interpretation; under time pressure, often act on lab-printed reference ranges without R-ratio reasoning. The interpretive gap between guideline and bedside is the catalogue's opening.
- Diagnostic labs — set reference ranges per-instrument; have weak incentive to align with hepatology consensus and strong incentive to match historical ranges (continuity with past results). This is why printed reference ranges drift behind the literature.
- Pharma drug-safety / regulators — heavy users of Hy's Law and DILI definitions; pushed magnitude-based interpretation into routine practice.
- Patient communities (Gilbert's, PBC, NAFLD) — strong online presence; the Gilbert's community is generally well-informed; the NAFLD community is fragmented and target-rich for misinformation about supplements.
- Supplement industry — markets "liver detox" and "liver cleanse" products keyed to "elevated liver enzymes"; no evidence base; competing incentive to keep the interpretive skill mystified.
Population variability
- Sex. ALT ULN is meaningfully lower in women (~19 vs. ~30 IU/L by Prati et al. 2002 thresholds). Many labs still use unisex ranges, under-detecting injury in women Prati et al. 2002.
- Age. Children and adolescents have physiologically higher ALP from bone growth (can be 2–3× the adult ULN without disease). Older adults have lower albumin and modest ALP rises with age.
- Pregnancy. ALP doubles in the third trimester from placental isoenzyme; albumin drops from haemodilution; bilirubin should not rise — if it does, think HELLP, intrahepatic cholestasis of pregnancy, or acute fatty liver of pregnancy.
- BMI / metabolic. Higher BMI shifts the entire ALT distribution upward; the "normal" ALT for a BMI 32 patient may still represent steatosis.
- Athletes. AST and ALT rise after intense exercise (rhabdomyolysis at the extreme); a CK alongside settles the question — if CK is elevated, the AST is muscle-source.
- Genetic. ~5–10% of the population carries the UGT1A1*28 promoter variant (Gilbert's); higher prevalence in some West African and South Asian populations. Hereditary hemochromatosis HFE C282Y homozygosity ~1:200 in Northern European descent, vanishingly rare elsewhere.
- Race / lab calibration. Most reference-range studies are white-European cohorts; some evidence that baseline ALT differs by ancestry, though clinical impact is small relative to BMI.
Knowledge gaps
- Optimal ULN for ALT. Prati thresholds (30 M / 19 F) are arguably the most evidence-based but not universally adopted; the cost-benefit of lowering reference ranges to capture more NAFLD vs. generating false-positive workup has not been definitively settled Ruhl & Everhart 2012.
- Threshold for action on mild ALT elevation in metabolically healthy patients. Guidelines diverge on whether to image / serology-screen at ALT 1.5× ULN vs. 2× vs. repeat-only.
- GGT prognostic value. A growing literature ties GGT to cardiovascular and all-cause mortality independent of liver disease, but the clinical actionability remains unclear — should an isolated high GGT trigger metabolic intervention? Whitfield 2001
- Direct vs. indirect bilirubin assay precision at low values. Most labs report direct bilirubin to one decimal place, but the assay is imprecise below 0.3 mg/dL — a "direct fraction" calculation in mild hyperbilirubinemia can be misleading.
- Macroenzymes. Underrecognised prevalence; no routine screening protocol.
Scope coverage relative to the brief. All six markers named in the brief — ALT, AST, ALP, GGT, bilirubin, albumin — covered in the mechanism section. The three presentation patterns (hepatocellular, cholestatic, isolated bilirubin) are operationalised through the R-ratio framework in the evidence section. Workup branches per pattern are laid out in protocol. Nothing dropped from the brief.
Why category screening rather than medical. The interpretive skill is applied almost exclusively to results from standard primary-care chemistry — the panel itself is part of routine screening. Treating this as a screening-literacy entry, not a medical-decision entry, matches how the reader will use it.
Action know rather than test. The entry teaches interpretation of results the reader is already getting; it doesn't push them to order labs they weren't going to get. If we later add a sibling on "what labs to ask for at your annual physical," that entry takes test.
Rating call on longevity. Held at 3 rather than 4. The cited mortality and disease-progression reductions (Kim et al. 2004 for ALT; Younossi et al. 2016 for NAFLD; Lindor et al. 2009 for PBC) require the reader to act on findings, not merely read them. Scoring the interpretation skill at 4 would conflate it with the downstream treatments. The entry sits one step upstream — necessary but not sufficient.
Rating call on evidence. 4 not 5: the interpretive framework is settled across ACG, BSG, EASL, but the ULN debate (Prati et al. 2002 vs. legacy lab ranges) is genuinely unresolved, and the threshold for action on a 1.5×-ULN mild elevation varies by guideline.
Hard editorial choice on ULN. Chose to teach the Prati thresholds (30 M / 19 F) as the better evidence-based bar while acknowledging that printed lab ranges run higher. The alternative — defer to whatever the reader's lab prints — leaves a real signal on the table. Called out explicitly in the misconceptions and failure-modes sections so the reader isn't blindsided when a "normal" result is annotated against a stricter threshold.
Separate-entry candidates surfaced during writing
- Non-alcoholic / metabolic fatty liver disease (NAFLD/MASLD) — needs its own entry covering progression, lifestyle reversal, GLP-1 evidence, and FibroScan staging. Linked from out-of-scope.
- Hepatitis B and C screening / treatment — the curative DAA regimens for HCV deserve their own entry. Linked from stakes and out-of-scope.
- Hereditary hemochromatosis — screening (ferritin + transferrin saturation), HFE genotyping, phlebotomy treatment. Linked from stakes.
- FibroScan / transient elastography — the natural follow-on for a hepatocellular pattern that doesn't resolve. Linked from out-of-scope.
- Gilbert's syndrome — a small standalone entry would let us point Gilbert's carriers somewhere reassuring without bloating this one.
- Hy's Law / drug-induced liver injury — covered briefly here as a warning callout; could warrant its own entry tied to medication monitoring.
Excluded from the article. Detailed scoring systems (MELD, Child-Pugh, APRI, FIB-4) — these are clinician tools and would distract the reader. Mentioned by name in out-of-scope so the reader knows they exist.
Voice tension noted. The R-ratio and Hy's Law are inherently technical and resist full plain-Englishing. Both are wrapped in callouts so the surrounding prose stays felt; the callouts retain the math because the reader needs the exact rule, not a softened version of it.
Future cross-links. Once entries exist on annual blood panels, fasting before labs, NAFLD, alcohol cessation, and screening cadences, this entry should link to all of them.
Reading a Liver Panel
One evening to learn. After that it's reading numbers off a printout — minutes, only when labs come in.
The interpretation framework is endorsed by every major hepatology society and has been taught the same way for decades. The data is settled.
Fatty liver, hepatitis C, hemochromatosis, autoimmune liver disease — all silent for years, all visible on a standard blood draw if you know what you're looking at.
A bad reading triggers a CT scan you didn't need. A good one books a repeat blood draw and waits four weeks. The skill is the difference.
"ALT 47*" stops being a free-floating worry and becomes a number with a meaning and a plan. Especially big if you're one of the ~1 in 15 people with Gilbert's syndrome.