What it is:The end result of chronic liver injury. Normal liver tissue gets replaced by regenerative nodules wrapped in dense scar (fibrosis). That's the histologic hallmark, and it's irreversible.
The core problem:Scarred, nodular architecture does two things at once. It physically blocks blood flowthrough the liver (portal hypertension, which drives varices, variceal bleeding, and ascites), and it kills the liver's factory function(loss of albumin synthesis, loss of clotting factor synthesis, loss of ammonia clearance, which drives coagulopathy and hepatic encephalopathy). Almost every complication in this chapter traces back to one of those two mechanisms.
What you do about it:Treat the cause if you can (stop the alcohol, treat the hepatitis C). Then it's complication management for life: screen for varices and prophylax, manage ascites, watch for spontaneous bacterial peritonitis (SBP), and treat hepatic encephalopathy (HE) when it flares.
Think of cirrhosis as "broken plumbing plus broken factory."The plumbing problem (portal hypertension) is what you treat with beta blockers and band ligation. The factory problem (lost synthetic and detox capacity) is what shows up as low albumin, a high INR that isn'ta reason to give FFP prophylactically, and ammonia that isn't being cleared. Every drug in this chapter is aimed at one side or the other, never at "fixing" the liver itself, because you can't reverse fibrosis with medication.
Two systems, two different jobs. Child-Pughis bedside and drives treatment decisions in this chapter (who gets prophylactic beta blockers, who's high risk for TIPS). MELD-Nais the transplant-allocation score, calculated from labs alone.
| Criterion | 1 point | 2 points | 3 points |
|---|---|---|---|
| Total bilirubin | <2 mg/dL | 2–3 mg/dL | >3 mg/dL |
| Albumin | >3.5 g/dL | 2.8–3.5 g/dL | <2.8 g/dL |
| Ascites | None | Mild | Moderate |
| Encephalopathy grade | None | 1–2 | 3–4 |
| PT (seconds prolonged) | <4 | 4–6 | >6 |
Grade A<7 points · Grade B7–9 points · Grade C10–15 points. Two labs (bilirubin, albumin), one coag value (PT), and two clinical findings you assess on exam (ascites, HE grade) - that mix of objective and subjective is exactly why it's a bedside score and not a pure lab calculation.
Grade C(or Grade B with active bleeding at endoscopy) is the threshold for considering early TIPSinstead of standard EVL plus octreotide, because these patients are high-risk for failing standard therapy. Grade C alone (even without bleeding, if small varices are present) is also a trigger for starting primary prophylaxiswith a beta blocker.
MELD score = 9.57 × ln(creatinine mg/dL) + 3.78 × ln(bilirubin mg/dL) + 11.20 × ln(INR) + 6.43. Any lab value below 1 gets rounded up to 1 before plugging in. The final score is multiplied by 10 and rounded to the nearest whole number.
MELD-Na = MELD − sodium(mEq/L) − [0.025 × MELD × (140 − sodium)] + 140. Adding sodium matters because hyponatremia independently predicts mortality in cirrhosis, on top of whatever the raw MELD says.
Chronic injury (Table below) triggers fibrosis and nodule formation. Once the architecture is distorted, two downstream cascades explain nearly everything that follows.
Distorted architecture raises intrahepatic resistance, which raises the pressure gradient between the portal vein and the central venous system. Portal hypertension is defined as a gradient >5 mm Hgbetween portal and central venous pressure. That pressure has to go somewhere, so the body opens collateral vessels, most importantly gastroesophageal varices, which are thin-walled and not built to carry high-pressure blood. They rupture. First variceal hemorrhage happens at an annual rate of roughly 15%and carries a mortality of 7–15%per bleed, which is why prophylaxis is not optional once varices are found.
Separately, portal hypertension drives splanchnic vasodilationand new vessel formation. More blood pools in the splanchnic bed, which (paradoxically, since the patient is fluid-overloaded) makes the kidneys think the body is underfilled. That renal hypoperfusion activates the renin-angiotensin-aldosterone system, and the resulting sodium and water retention is what produces ascites.
A cirrhotic liver can't make what it used to make. Albuminsynthesis drops (low oncotic pressure contributes to ascites too, on top of the RAAS-driven volume overload). Clotting factorsynthesis drops, but so does synthesis of the liver's own anticoagulants- antithrombin, protein C, and protein S. Meanwhile factor VIII and von Willebrand factor are actually elevated. Net effect is unpredictable: some patients bleed, some clot, and a prolonged INR in a cirrhotic patient does not mean they're "auto-anticoagulated" the way it would in a warfarin patient.
Read the two cascades and the drug list writes itself. Nonselective beta blockerslower portal pressure by cutting cardiac output andsplanchnic blood flow, directly countering cascade 1. Spironolactone and furosemidecounter the RAAS-driven sodium retention from cascade 1's renal hypoperfusion loop. Lactulose and rifaximinexist because the factory can no longer clear gut-derived ammonia. Nothing here regenerates hepatocytes, because nothing can.
| Category | Examples |
|---|---|
| Alcohol / metabolic | Alcohol use disorder, nonalcoholic fatty liver disease |
| Viral | Chronic hepatitis C |
| Metabolic liver disease | Hemochromatosis, Wilson's disease |
| Immunologic | Autoimmune hepatitis, primary biliary cholangitis |
| Vascular | Budd-Chiari syndrome |
| Drug-induced | Isoniazid, macrolides, amoxicillin-clavulanate, nitrofurantoin, fluoroquinolones, amiodarone, NSAIDs, allopurinol, sulfasalazine, methotrexate, interferons, anti-TNF agents, valproate, lamotrigine, phenytoin, carbamazepine |
HE results from gut-derived nitrogenous substances, mainly ammonia, that would normally be cleared by the liver. With a fibrotic liver and portosystemic shunting, those substances bypass hepatic clearance, reach the CNS, and disrupt neurotransmission. Glutamate, benzodiazepine receptor agonists, and aromatic amino acids are also implicated. There is no established correlation between the blood ammonia level and how altered the patient actually is- don't chase the number, treat the exam.
Type Ais from acute liver failure, Type Bis portosystemic shunting without underlying liver disease, and Type Cis the one you'll see constantly: HE in a patient with cirrhosis. HE itself is further described as episodic, persistent, or minimal.
Early cirrhosis is often silent or nonspecifically tired-and-hungry-less. What actually brings patients in is usually a decompensation event: new ascites (abdominal distention) or new confusion (HE).
| Category | Findings |
|---|---|
| Nonspecific | Fatigue, anorexia, weight loss, malaise |
| Portal HTN / decompensation | Ascites and edema, splenomegaly, encephalopathy (confusion, lethargy) |
| Cutaneous / endocrine stigmata | Jaundice, palmar erythema, spider angiomas (spider nevi), gynecomastia, testicular atrophy, axillary hair loss, parotid gland enlargement, white nails, Dupuytren contracture |
| Neuro | Asterixis (the flapping tremor) |
| Labs | Hypoalbuminemia, elevated PT, thrombocytopenia, elevated alkaline phosphatase, elevated AST/ALT/GGT |
Wasting happens because a failing liver shifts to catabolizing muscle for energy and gluconeogenic substrate once glycogen stores and albumin synthesis fall off. It's part of the same synthetic-failure picture as hypoalbuminemia, not a separate process.
There is no single lab or imaging test that diagnoses cirrhosis."Liver function tests" are really two different categories answering two different questions, and mixing them up is a classic trap.
| Category | Tests | What it tells you |
|---|---|---|
| Markers of liver injury | AST, ALT, GGT, alkaline phosphatase | Hepatocytes are being damaged right now (or were recently) |
| Markers of hepatocellular function | PT, bilirubin, albumin | How much working liver tissue is left to synthesize things |
Albuminhas a long half-life, so a low albumin implies the disease process has been going on for more than 3 weeks. PTreflects clotting factors with much shorter half-lives, so severe liver injury can push the PT up in under 24 hours. If a patient's PT is climbing fast but their albumin is still normal, think acute-on-something, not necessarily long-standing cirrhosis.
Once you suspect cirrhosis, Child-Pugh(see above) grades severity and predicts survival, surgical risk, and variceal bleeding risk. MELD-Nais the newer score used mainly for transplant prioritization.
Goals of treatment:resolve acute complications when they happen (tamponade a bleed, stabilize hemodynamics), and prevent complications the rest of the time by lowering portal pressure and supporting abstinence from the causative agent (usually alcohol).
The general approach, straight from the guideline logic: identify and eliminate the cause · assess variceal bleeding risk and start pharmacologic prophylaxis where indicated · evaluate for ascites and manage with diuretics plus paracentesis as needed, watching closely for SBP · treat HE with dietary changes, removing CNS depressants, and ammonia-lowering therapy · monitor for hepatorenal syndrome, pulmonary insufficiency, and endocrine dysfunction.
Screen at diagnosis. NSBB or EVL for prevention; combo therapy for an active bleed.
Sodium restriction plus diuretics; paracentesis if tense or refractory.
Any patient with ascites who acutely deteriorates gets a tap.
Lower ammonia production and increase its removal.
Three distinct clinical scenarios, three distinct strategies: (1) primary prophylaxisbefore any bleed has happened, (2) treatmentof an active variceal hemorrhage, and (3) secondary prophylaxisafter a patient has already bled once.
Screen every cirrhotic patient with portal hypertension for varices at diagnosis. The mainstay is a nonselective beta blocker (NSBB): propranolol, nadolol, or carvedilol. These lower portal pressure through two separate mechanisms, decreased cardiac output and decreased splanchnic blood flow, which is exactly the "cascade 1" fix from the pathophysiology above.
Carvedilol adds alpha-1 blockade on top of nonselective beta blockade, so it's perceived as better tolerated than pure propranolol or nadolol in patients who develop fatigue, weakness, or shortness of breath. It's a reasonable swap, not a downgrade.
This is a medical emergency. Initial goals, in order: adequate volume resuscitation, airway protection from aspiration, prophylaxis against SBP and other infection, control of the bleed, prevention of rebleeding, preservation of liver function (avoiding an HE trigger), and prevention of acute kidney injury.
Ceftriaxoneis genuinely there to prevent infection (SBP and other bacterial complications are common after a GI bleed in cirrhosis). Erythromycinis dosed once, right before the scope, purely for its prokinetic GI motility effect so the endoscopist can actually see the bleeding source. If a question implies erythromycin is treating an infection here, that's the trap.
Acute bleed → resuscitation (ABCs, sedation as needed) plus octreotide and prophylactic antibiotics started immediately → diagnostic and therapeutic endoscopy.
NSBB plus EVL together is the best optionfor a patient who has already had a variceal bleed, better than either alone.
Goals:control the ascites, relieve symptoms (dyspnea, abdominal pain, distention), and prevent SBP and hepatorenal syndrome.
Send a serum-ascites albumin gradient (SAAG)on any new ascites workup. A gradient ≥1.1 g/dLmeans the patient almost certainly has portal hypertension driving the fluid, as opposed to a non-portal-hypertensive cause (malignancy, infection, pancreatitis).
Spironolactone is potassium-sparing and furosemide is potassium-wasting. Keeping them locked at that ratio as you titrate up maximizes natriuresis (the actual therapeutic goal) while keeping serum potassium roughly stable, instead of chasing hyperkalemia with one drug and hypokalemia with the other.
Uncontrolled or recurrent encephalopathy develops, severe hyponatremia (Na <120 mEq/L) persists despite fluid restriction, or renal insufficiency develops (SCr >2 mg/dL). These are all signs the diuretics are doing more harm than good.
Suspect SBP in any cirrhotic patient with ascites who has an acute deterioration. It's caused by translocation of gut bacteria into ascitic fluid, mainly E. coli, Klebsiella pneumoniae, and pneumococci.
If PMN ≥250 cells/mm³, clinical suspicion is high, andthe patient has SCr >1 mg/dL, BUN >30 mg/dL, or total bilirubin >4 mg/dL, give albumin 1.5 g/kg within 6 hours of detectionplus 1 g/kg on day 3. This combination reduces the risk of hepatorenal syndrome developing on top of the infection - it's not given to every SBP patient, only the ones with these markers of higher renal risk.
| Scenario | Regimen | Duration |
|---|---|---|
| Cirrhosis with acute GI hemorrhage | Antibiotic prophylaxis (ceftriaxone during the bleed, see above) | Short-term, 7 days |
| Survived a prior SBP episode | Daily ciprofloxacin or trimethoprim-sulfamethoxazole | Long-term / indefinite |
| Low-protein ascites (<1.5 g/dL) plus SCr ≥1.2, BUN ≥25, Na ≤130, or Child-Pugh ≥9 with bilirubin ≥3 | Daily ciprofloxacin or TMP-SMX | Long-term / indefinite |
All three scenarios use antibiotics to prevent SBP, but only the GI-hemorrhage indication is time-limited. The other two are ongoing because the underlying risk (low-protein ascites, or having already survived one episode) doesn't go away.
Management is four-pronged: supportive care for the patient's altered consciousness, rule out other causes of altered mental status, identify and treat any precipitating factor, and start empiric HE-specific treatment.
| Grade | Consciousness | Personality / intellect | Neuro findings |
|---|---|---|---|
| Unimpaired | Normal | Normal | Normal |
| Minimal | No clinical change | No clinical change (only on psychometric testing) | No clinical change |
| I | Trivial lack of awareness, shortened attention span | Euphoria or anxiety; impaired addition/subtraction | Altered sleep rhythm |
| II | Lethargic | Obvious personality change, inappropriate behavior, apathy | Asterixis, dyspraxia, disoriented to time |
| III | Somnolent but arousable | Bizarre behavior | Confused, grossly disoriented to time and space |
| IV | Coma, unarousable | None | Does not respond to stimuli |
Rifaximin dosing is indication-specific and it's easy to mix up. For hepatic encephalopathy, it's 550 mg twice daily, continued indefinitely, on top of lactulose. For IBS-D, it's an entirely different regimen: 550 mg three times daily for 14 days only, with retreatment allowed up to twice if symptoms recur. Same drug, same rough total daily exposure ballpark, completely different duration and dosing frequency depending on why you're using it.
Lactulose isn't a laxative here in the usual sense. As an unabsorbed disaccharide it's fermented by colonic bacteria into acidic byproducts, which converts ammonia (NH₃) into ammonium (NH₄⁺) that can't cross back into the bloodstream, and the resulting catharsis physically clears it out. Rifaximin instead reduces the population of ammonia-producing gut bacteria in the first place. Together they attack production and removal.
| Nonselective beta blockers - primary & secondary prophylaxis of variceal bleeding | ||
| Propranolol | 20 mg BID | Titrate q2–3 days (primary) or weekly (secondary) to HR 55–60 or max tolerated. Max 320 mg/day (160 mg/day if ascites) |
| Nadolol | 20–40 mg daily | Same titration goal. Max 160 mg/day (80 mg/day if ascites) |
| Carvedilol | 3.125 mg BID | Increase to 6.25 mg BID after 3 days; better tolerated alternative |
| Acute variceal hemorrhage | ||
| Octreotide | 50 mcg IV bolus, then 50 mcg/hr infusion | Continue 2–5 days; may stop after 24h bleed-free |
| Ceftriaxone | 1 g IV every 24 hours | Infection prophylaxis for all cirrhotic patients with variceal bleed |
| Erythromycin | 250 mg IV once, 30–120 min pre-endoscopy | Prokinetic, improves visualization, not for infection |
| Ascites - diuretics (100:40 ratio) | ||
| Spironolactone | 100 mg every morning | Titrate q3–5 days; max 400 mg/day |
| Furosemide | 40 mg every morning | Titrate q3–5 days maintaining ratio; max 160 mg/day |
| Ascites - refractory / tense | ||
| Albumin (postparacentesis) | 6–8 g per liter removed | Only if >5 L removed at once |
| SBP treatment | ||
| Cefotaxime | 2 g IV every 8 hours | 5 days; ceftriaxone or piperacillin/tazobactam are alternatives |
| Albumin (SBP-specific) | 1.5 g/kg within 6h of detection, then 1 g/kg on day 3 | Only if PMN ≥250 andSCr >1, BUN >30, or bilirubin >4 |
| SBP prophylaxis | ||
| Ciprofloxacin | 500 mg daily | Long-term, after surviving SBP or in high-risk low-protein ascites |
| Trimethoprim-sulfamethoxazole | Daily | Alternative long-term prophylaxis |
| Hepatic encephalopathy | ||
| Lactulose | 30 mL (20 g) PO q1–2h until catharsis | Then titrate to 2–3 soft stools/day maintenance |
| Rifaximin | 550 mg BID (with lactulose) | Indefinite for HE - do not confuse with the IBS-D regimen |
Propranolol and nadolol block both beta-1 (cardiac) and beta-2 (splanchnic vasculature) receptors. Blocking beta-1 drops cardiac output; blocking beta-2 leaves alpha-mediated splanchnic vasoconstriction unopposed, reducing splanchnic blood flow. Both effects lower portal venous inflow, which is the whole point. Carvedilol adds alpha-1 blockade on top, giving additional vasodilation and often better tolerability.
Monitor:blood pressure and heart rate at every titration. Goal HR 55–60 bpm, goal systolic BP generally >90 mmHg. Watch for bradycardia, bronchospasm, and (especially in insulin-dependent diabetics) masked hypoglycemia symptoms.
Important operational point:as cirrhosis progresses, patients can develop new contraindications to NSBBs - renal impairment and hypotension both become more common in decompensated disease. A drug that was appropriate at diagnosis may need to be stopped later; this isn't a failure of adherence, it's disease progression outpacing the drug's tolerability.
Spironolactone is the backbone here, not furosemide, because ascites in cirrhosis is fundamentally a hyperaldosteronism problem (RAAS activation from perceived renal underfilling, as described in the pathophysiology section). Furosemide is added for extra natriuretic power once spironolactone alone isn't cutting it, and the 100:40 mg ratio keeps potassium roughly balanced while both drugs are titrated upward together.
Watch for electrolyte disturbances (especially hyperkalemia from spironolactone, or hypokalemia if the ratio drifts), dehydration, renal insufficiency, and hypotension. Weight loss faster than 0.5 kg/day is itself a warning sign of over-diuresis, not a treatment success.
Empiric coverage targets the usual gut-translocation organisms: E. coli, Klebsiella pneumoniae,and pneumococci. Cefotaxime (or another third-generation cephalosporin) covers this reliably; ceftriaxone and piperacillin/tazobactam are reasonable substitutes, particularly for community-acquired presentations.
The albumin adjunct in SBP isn't about the infection itself - it's about preventing the renal hypoperfusion cascade (hepatorenal syndrome) that SBP can trigger in a patient whose kidneys are already borderline. That's why it's reserved for patients with markers of renal risk (elevated SCr, BUN, or bilirubin) rather than given to every SBP patient automatically.
Lactulose's therapeutic endpoint during an acute episode is catharsis (getting the patient moving), and then the maintenance target flips to a specific stool frequency (2–3 soft stools/day) rather than "as much as possible." Overshooting causes dehydration, hypernatremia, and can paradoxically worsen HE by causing volume depletion. Monitor electrolytes and mental status together, not just one or the other.
Rifaximin's adverse effect profile is favorable because it's minimally absorbed, but the dosing regimen for HE (550 mg BID, indefinite) is structurally different from IBS-D dosing (550 mg TID, 14 days, with limited retreatment). Confirm the indication before assuming a "550 mg" order is correct.
Cirrhosis lowers both procoagulant factor synthesis andanticoagulant synthesis (antithrombin, protein C, protein S), while factor VIII and von Willebrand factor actually rise. Thrombocytopenia (from splenic sequestration and reduced thrombopoietin) adds another layer. The net effect is a fragile, rebalanced hemostatic system that can tip either direction - clinically significant bleeding orthrombosis - rather than a system that's simply "anticoagulated." An elevated PT/INR in cirrhosis does not reliably predict bleeding risk the way it does on warfarin, and prophylactic correction with FFP is not automatically indicated just because the number is high.
| Drug / Therapy | Adverse effect to watch | Monitor |
|---|---|---|
| Propranolol / nadolol | Heart failure, bronchospasm, glucose intolerance | BP, HR - goal HR 55–60 or max tolerated dose |
| Carvedilol | Similar to NSBB, but potentially better tolerated | BP - goal systolic >90 mmHg |
| Octreotide | Bradycardia, hypertension, arrhythmia, abdominal pain | BP, HR, EKG, abdominal exam |
| Spironolactone / furosemide | Electrolyte disturbance, dehydration, renal insufficiency, hypotension | Serum electrolytes (especially K⁺), SCr, BUN, BP, daily weight (goal ≤0.5 kg/day loss); spot urine Na > K correlates with adequate natriuresis (>78 mmol/day) |
| Lactulose | Electrolyte disturbance from over-catharsis | Serum electrolytes, stool frequency (goal 2–3 soft stools/day), mental status |
| Rifaximin | Nausea, diarrhea | Clinical response, HE recurrence |
| Cefotaxime / SBP antibiotics | Usual antibiotic AEs | Repeat paracentesis at 48h if not improving; renal function |