What it is:This appendix isn't a disease chapter, it's a toolkit. It hands you two instruments: a scoring system (RUCAM) for deciding whether a drug actually caused a liver injury, and a lab pattern table for deciding what kind of liver injury you're looking at.
The core problem:Liver enzymes go up for dozens of reasons: viral hepatitis, alcohol, gallstones, fatty liver, autoimmune disease, sepsis, and drugs. If you don't have a structured way to separate "drug-induced" from "everything else," you either pull a perfectly good medication for no reason or you leave a hepatotoxic drug running while you chase the wrong workup.
What you do about it:Classify the injury pattern from the labs first (which enzymes are up, and by how much), then run the causality tool to score how likely the suspect drug actually is, while systematically ruling out the alternatives that could explain the same picture.
Everything in this appendix answers one of two separate questions, and mixing them up is the single biggest mistake. Question 1, "what kind of injury is this?"is answered by the enzyme pattern table: hepatocellular (necrotic), cholestatic, or mixed/chronic. Question 2, "did the drug actually do it?"is answered by RUCAM: a point-scored checklist covering timing, dechallenge and rechallenge, prior reports, competing causes, and host risk factors. You need both. The pattern tells you what to test for and how urgent it is. The causality score tells you whether you're even looking at the right suspect.
The liver has two functionally separate compartments that a toxic drug or metabolite can hit, and which one takes the damage determines which enzymes rise. Hit the hepatocyte itself(the cell doing metabolism and synthesis) and you get a necrotic, hepatocellular pattern. Hit the bile duct or canalicular transport machinery(the plumbing that moves bile out of the liver) and you get a cholestatic pattern. Hit both, or sustain lower-grade injury over a longer course, and you get a mixed or chronic pattern. The enzyme table below is really just a map of which compartment is under attack, read through five lab values.
Why the same six enzymes are always the ones you check:three of them (AST, ALT, LDH) live mostly insidethe hepatocyte and only spill into blood when the cell membrane is damaged or the cell dies, so they track hepatocellular injury. The other three (alkaline phosphatase, 5'-nucleotidase, GGT) are anchored to the canalicular membranethat lines the bile ducts, so they track cholestatic injury. A drug that mostly kills hepatocytes will light up the first group far more than the second. A drug that mostly blocks bile flow does the reverse.
Don't read "necrotic" as automatically more dangerous than "cholestatic." Necrotic/hepatocellular injury tends to move faster and carries a higher risk of acute liver failure per case, but cholestatic injury can smolder for months and cause its own complications (pruritus, malabsorption, fat-soluble vitamin deficiency). The pattern changes what you watch for and how urgently, not simply how worried you should be in absolute terms.
What the patient reports usually tracks the compartment involved, the same way the labs do.
| Pattern | What you'll hear and see |
|---|---|
| Hepatocellular (necrotic) | Fatigue, nausea, right upper quadrant discomfort, and (if severe) jaundice. Can progress quickly to signs of synthetic dysfunction: easy bruising, confusion, or worsening jaundice as the injury advances. |
| Cholestatic | Pruritus (often out of proportion to how sick the patient otherwise looks), jaundice, dark urine, and pale or clay-colored stools as conjugated bilirubin and bile acids back up into the bloodstream instead of draining into the gut. |
| Mixed / chronic | A blend of both, often lower-grade and slower to declare itself. Chronic low-level injury can present only as an incidental abnormal liver panel long before the patient notices anything. |
New confusion, asterixis, or a rapidly climbing bilirubin in a patient with a suspected drug-induced hepatocellular injury needs same-day escalation. That combination raises concern for evolving acute liver failure, which moves fast once synthetic function starts to fail.
This is the appendix's core reference table. Learn to read it left to right, not just memorize the arrows: it's telling you which compartment is hit and by how much, relative to the other pattern.
| Enzyme | Abbreviations | Necrotic (hepatocellular) | Cholestatic | Chronic |
|---|---|---|---|---|
| Alkaline phosphatase | Alk Phos, AP | ↑ | ↑↑↑ | ↑ |
| 5'-Nucleotidase | 5-NC, 5NC | ↑ | ↑↑↑ | ↑ |
| γ-Glutamyltransferase | GGT, GGTP | ↑ | ↑↑↑ | ↑↑ |
| Aspartate aminotransferase | AST, SGOT | ↑↑↑ | ↑ | ↑↑ |
| Alanine aminotransferase | ALT, SGPT | ↑↑↑ | ↑ | ↑↑ |
| Lactate dehydrogenase | LDH | ↑↑↑ | ↑ | ↑ |
Notice that everyenzyme goes up in all three columns, at minimum a single arrow. The diagnosis isn't "which labs are abnormal," almost all of them will be. It's "which groupof labs is abnormal out of proportion to the other group." A hepatocellular picture is AST/ALT/LDH climbing much harder than Alk Phos/5-NC/GGT. A cholestatic picture is the reverse. That's the whole logic of the table in one sentence.
GGT rises in both cholestatic and chronic injury, but notice it's the only one of the three canalicular enzymes that also climbs to double arrows in the chroniccolumn. A GGT that's persistently elevated more than Alk Phos and 5-NC, in a patient without a clear acute picture, points you toward a smoldering, chronic pattern rather than an acute cholestatic event.
Eyeballing "which column of arrows looks worse" works, but it's not reproducible between clinicians. The standard way to formalize exactly what the table above is showing is to convert the two lead enzymes, ALT and alkaline phosphatase, into a single ratio measured against their own upper limits of normal (ULN).
R = (ALT ÷ ALT ULN) ÷ (Alk Phos ÷ Alk Phos ULN)
| R value | Pattern | What it means |
|---|---|---|
| R ≥ 5 | Hepatocellular | ALT is rising far out of proportion to Alk Phos, the necrotic pattern from the table above, ALT/AST/LDH-dominant |
| R ≤ 2 | Cholestatic | Alk Phos is rising far out of proportion to ALT, the cholestatic pattern, Alk Phos/5-NC/GGT-dominant |
| 2 < R < 5 | Mixed | Both compartments are contributing meaningfully, matching the "chronic" column's more even spread of moderate elevations |
A patient's R value can migrate between categories over the course of an injury, since hepatocellular and cholestatic components don't always peak at the same time. Recheck the ratio with each new set of labs rather than anchoring on the number from the day of presentation.
Once you know the pattern, the next question is whether the suspect drug actually caused it. The Roussel Uclaf Causality Assessment Method (RUCAM)converts a clinical judgment call into a reproducible point total, so two different clinicians scoring the same case should land in the same range.
| Criterion | Result | Score |
|---|---|---|
| Timing | ||
| Days from startof therapy to the reaction | 5-90 days | +2 |
| <5 or >90 days | +1 | |
| Days from cessationof therapy to the reaction (except for slowly metabolized drugs) | ≥30 days | +1 |
| Rechallenge | ||
| Unintentional rechallenge with the suspect drug/herb alone | Alk Phos > 2x normal | +3 |
| Alk Phos ≤ normal | −2 | |
| Unintentional rechallenge with the suspect drug/herb plus other agents | Alk Phos > 2x normal | +1 |
| Alk Phos ≤ normal | 0 | |
| Prior evidence for this drug/reaction pair | ||
| This drug and this reaction have been reported before in… | Product labeling | +2 |
| Published literature | +1 | |
| Competing drugs/herbs | ||
| Concurrent hepatotoxic drug or herb present? | No | 0 |
| If yes, is that agent's timing… | Consistent with causing the injury | −2 |
| Inconsistent with causing the injury | 0 | |
| Is a concurrent drug/herb the more likely cause? | Positive rechallenge or a validated test points to it instead | −3 |
| Host risk factors | ||
| Alcohol use >2 drinks/day (female) or >3 drinks/day (male) | Yes | +1 |
| No | 0 | |
| Pregnant | Yes | +1 |
| No | 0 | |
| Age > 55 years | Yes | +1 |
| No | 0 | |
| Alternative causes ruled out | ||
| All alternatives systematically excluded | Groups I and II both fully ruled out | +2 |
| Group I only fully ruled out | +1 | |
| Alternatives partially excluded | 5-6 of Group I ruled out | 0 |
| Fewer than 5 of Group I ruled out | −2 | |
| Overall likelihood of a non-drug cause | Highly likely | −3 |
| Somewhat likely | 0 | |
Total score interpretation:
| Total score | Causality verdict |
|---|---|
| ≤ 0 | Drug/herb is not a likely cause |
| 1-2 | Unlikelyto be the cause |
| 3-5 | Possiblecause |
| 6-8 | Probablecause |
| > 8 | Highly probablecause |
Say a patient starts a new drug, develops jaundice on day 40(+2 for the 5-90 day window), has noconcurrent hepatotoxic drug (0), is 62 years old(+1 age risk factor), and every viral, biliary, and autoimmune alternative on the list comes back negative (+2 for both groups ruled out). No rechallenge was attempted (that criterion contributes 0 either way) and this drug has a prior case report in the literature (+1). Running total: +6, landing in the probablerange. Change nothing except add unintentional rechallenge with a rising Alk Phos, and the score jumps to +9, crossing into highly probable. That single data point, rechallenge, is worth more than almost anything else on the sheet, which is exactly why it's weighted at +3.
A positive unintentional rechallenge (Alk Phos rising again after re-exposure) is worth +3, the single largest positive value anywhere in RUCAM. It's also the one criterion you should never deliberately create. RUCAM scores unintentionalrechallenges, meaning the patient was accidentally re-exposed (a refill error, a duplicate therapy, a different brand of the same herb). Intentionally rechallenging a patient to confirm a causality score you could get another way is not how this tool is meant to be used.
RUCAM's "alternative causes" criterion is worth up to +2, but scoring it honestly means you actually have to work through two tiers of competing diagnoses before you can claim credit for ruling them out.
| Tier | What has to be checked |
|---|---|
| Group I | Hepatitis A (anti-HAV IgM), hepatitis B (anti-HBV IgM or HBV-DNA), hepatitis C (anti-HCV IgM or HCV-RNA), hepatitis E (anti-HEV IgM, anti-HEV IgG, or HEV-RNA), other structural/biliary disease on imaging (hepatobiliary sonography, color Doppler sonography, endosonography, liver CT, or liver MRI), alcoholic hepatitis (AST:ALT > 2 with a compatible history), and a recent hypotensive episode (particularly with underlying heart disease, which can cause ischemic hepatopathy). |
| Group II | Autoimmune hepatitis, chronic viral hepatitis, biliary or sclerosing cholangitis, complications of sepsis, metastatic malignancy, genetic liver disease, and infection with cytomegalovirus, Epstein-Barr virus, herpes simplex virus, or varicella-zoster virus. |
Fully excluding bothtiers earns +2. Excluding only Group I earns +1. This isn't busywork, it directly changes whether a borderline case lands in "possible" versus "probable." A causality assessment that skips the imaging or skips the viral serologies isn't just clinically incomplete, it's also mathematically under-scored, because you can't claim points for alternatives you never actually tested for.
Three host factors nudge the score, each worth +1 if present. None of them are diagnostic on their own, they just shift the probability slightly in the same direction the rest of the picture is already pointing.
| Risk factor | Threshold |
|---|---|
| Alcohol use | >2 drinks/day in a woman, >3 drinks/day in a man |
| Pregnancy | Currently pregnant |
| Age | Older than 55 years |
Each of these three is only worth a single point, but RUCAM's bands are narrow (0, 1-2, 3-5, 6-8, >8). A case sitting at a score of 5 (possible) with all three risk factors present but not yet counted could move to 6 (probable) once you actually ask about alcohol intake, pregnancy status, and age, three questions that take under a minute and are easy to forget to formally tally.
RUCAM scores rechallenge directly, but dechallenge (what happens after you stopthe drug) is the piece of the workup that happens in real time, at the bedside, before any score gets tallied.
Positive dechallengemeans the liver values improve after the suspect drug is stopped, supporting a drug-induced cause. Negative dechallenge, values keep climbing or fail to improve despite stopping the drug, should make you widen the differential back out, since it argues against that specific agent (or suggests injury severe enough to keep progressing on its own momentum for a period even after removal). Positive rechallenge, injury recurs on re-exposure, is the single strongest piece of evidence RUCAM can score, worth +3 when unintentional. Deliberately rechallenging a patient purely to confirm causality is generally not appropriate once a probable or highly probable picture already exists from the rest of the workup.
Stopping the suspect drug is the intervention, not a diagnostic test you order and wait on. You act first (stop the drug) and then use the trend in the labs over the following days to weeks as supporting evidence for your causality assessment, not as a prerequisite for stopping it.
Watching AST/ALT/Alk Phos tells you about ongoing cell injury, but it doesn't tell you whether the liver can still do its job. Worsening jaundice, a climbing INR, or new confusion in a patient with a hepatocellular pattern are signs the liver's synthetic and clearance functions are starting to fail, not just that cells are still being damaged, and that distinction is what should drive an urgent hepatology referral rather than routine follow-up labs.
| Parameter | When | Watching for |
|---|---|---|
| AST, ALT, LDH | Baseline at first suspicion, then trended on a defined interval until resolution or a clear alternative diagnosis is confirmed | Continued rise (negative dechallenge) versus a downward trend (positive dechallenge) after stopping the drug |
| Alkaline phosphatase, 5'-nucleotidase, GGT | Same schedule as the transaminases | Whether the cholestatic component is resolving in step with, ahead of, or behind the hepatocellular component; cholestatic recovery often lags |
| R value | Recalculated with every new lab set, not just once at presentation | Migration between hepatocellular, mixed, and cholestatic categories as the injury evolves |
| Bilirubin (total/direct) | Alongside every enzyme panel | Rising conjugated bilirubin, which tracks clinical jaundice and cholestasis severity more directly than the enzymes alone |
| INR / synthetic function markers | Any time the hepatocellular pattern is significant or worsening | Early evidence the liver's clotting-factor synthesis is failing, a marker of severity that outranks the enzyme trend itself |
| Mental status | Continuously in any patient with a significant hepatocellular pattern | New confusion or asterixis, which with a climbing bilirubin and INR signals possible evolving acute liver failure |