Portal Hypertension
Clinical Snapshot
Portal hypertension is elevated pressure within the portal venous system, measured as the hepatic venous pressure gradient (HVPG) — the difference between the wedged (occluded) and free hepatic venous pressures, a close surrogate for portal pressure itself in sinusoidal disease. Normal HVPG is 1-5 mmHg. Portal hypertension exists once HVPG exceeds 5 mmHg, but nothing clinically important happens until HVPG reaches 10 mmHg — the threshold defining clinically significant portal hypertension (CSPH), above which gastroesophageal varices form, ascites and other decompensating events become possible, and postoperative/post-TIPS risk rises sharply.[1] Variceal bleeding risk climbs further still above roughly 12 mmHg — varices essentially do not bleed below this level regardless of their endoscopic appearance.[2]
See this platform's Cirrhosis chapter for general cirrhosis staging (Child-Pugh, MELD) and the longitudinal sodium-restriction/diuretic-titration protocol for ascites, and the Hepatic Encephalopathy chapter for staging and the lactulose/rifaximin protocol for the third major decompensating event — both are assumed background here, not re-explained. This chapter is the dedicated deep-dive on portal hypertension itself: the HVPG/CSPH framework, non-invasive variceal screening, primary and secondary prophylaxis of variceal hemorrhage, acute variceal hemorrhage as an in-hospital emergency, and the diagnosis/active treatment of spontaneous bacterial peritonitis (SBP) — the last of which the Cirrhosis chapter covers only as far as prophylaxis, not treating a confirmed episode.
Two mechanisms raise portal pressure
Portal pressure rises from the combination of (1) increased intrahepatic vascular resistance — a fixed, structural component from fibrosis and regenerative nodules distorting the sinusoidal architecture, plus a genuinely reversible, dynamic component from reduced intrahepatic nitric-oxide bioavailability causing excessive sinusoidal vasoconstriction — and (2) increased portal venous inflow, driven by splanchnic arterial vasodilation and the resulting hyperdynamic circulation that develops once CSPH is established.[1] This two-mechanism model is not academic: it is exactly why nonselective beta-blockers (NSBBs) work (reducing cardiac output and splanchnic inflow, the second mechanism) and why carvedilol outperforms propranolol/nadolol (its added alpha-1 blockade also relaxes intrahepatic vascular tone, addressing the dynamic component of the first mechanism directly) — see Choose Initial Treatment.
Recognize the Condition
The clinical consequences of portal hypertension
Everything in this chapter follows from one hemodynamic abnormality producing portosystemic collateral formation once CSPH is reached. The collaterals that matter clinically:
- Gastroesophageal varices — the dominant collateral pathway and the source of the single most lethal complication of portal hypertension, variceal hemorrhage (see Red Flags & Refer and Choose Initial Treatment).
- Ascites — driven by splanchnic vasodilation, effective arterial underfilling, and renal sodium/water retention; the Cirrhosis chapter owns the longitudinal diuretic-titration protocol, while this chapter adds the SAAG diagnostic link (Assessment & Diagnosis) and active SBP treatment.
- Splenomegaly and hypersplenism — thrombocytopenia and, less prominently, leukopenia from splenic sequestration; thrombocytopenia is often the earliest laboratory clue to portal hypertension before any other finding appears.
- Portosystemic (caput medusae, abdominal-wall) collaterals — dilated periumbilical veins from recanalized umbilical vein flow; a visible, non-bleeding marker of long-standing portal hypertension.
- Portal hypertensive gastropathy (PHG) — a distinct mucosal finding (a "snakeskin" or mosaic mucosal pattern on endoscopy) from congested gastric mucosal capillaries, causing chronic, low-grade, often occult GI blood loss and iron-deficiency anemia rather than the dramatic hemorrhage of a ruptured varix; PHG is managed with an NSBB to lower portal pressure, not with endoscopic band ligation or clipping, since it is a mucosal perfusion problem rather than a discrete bleeding vessel.[2]
- Ectopic varices (rectal, duodenal, peristomal) and portal hypertensive colopathy — less common, but a real source of unexplained lower GI bleeding in a patient with known portal hypertension.
- Hepatic encephalopathy — from portosystemic shunting of gut-derived neurotoxins bypassing hepatic clearance; see the Hepatic Encephalopathy chapter.
Non-cirrhotic portal hypertension exists, and changes management
Cirrhosis (sinusoidal portal hypertension) is overwhelmingly the cause encountered in Canadian practice, but portal hypertension can arise anywhere along the portal venous circuit — see Risk Factors for the full classification. The distinction matters clinically: a patient with non-cirrhotic portal hypertension can have life-threatening variceal bleeding with entirely preserved hepatic synthetic function (normal albumin, INR, bilirubin), so Child-Pugh/MELD-based risk stratification from the Cirrhosis chapter does not apply, and the underlying vascular lesion itself (e.g., portal vein thrombosis) often needs its own separate, specialist-directed treatment.
Risk Factors
Classifying portal hypertension by anatomic site
| Site | Examples |
|---|---|
| Prehepatic | Portal vein thrombosis; splenic vein thrombosis specifically causes isolated gastric varices without esophageal varices or ascites ("sinistral"/left-sided portal hypertension), classically from pancreatic disease compressing or invading the splenic vein |
| Intrahepatic — presinusoidal | Schistosomiasis (a leading global cause, though rare in Canadian-born patients — consider in patients with relevant travel/residence history); congenital hepatic fibrosis; nodular regenerative hyperplasia; early-stage primary biliary cholangitis or primary sclerosing cholangitis before cirrhosis develops; idiopathic noncirrhotic portal hypertension |
| Intrahepatic — sinusoidal | Cirrhosis of any cause — by far the dominant category in Canadian practice; see the Cirrhosis chapter for the full list of underlying etiologies (alcohol, MASLD/MASH, viral hepatitis, autoimmune/cholestatic liver disease, hereditary metabolic disease) |
| Intrahepatic — postsinusoidal | Sinusoidal obstruction syndrome (veno-occlusive disease), classically post-hematopoietic stem cell transplant |
| Posthepatic | Budd-Chiari syndrome (hepatic venous outflow obstruction); chronic right heart failure, constrictive pericarditis, or severe tricuspid regurgitation ("cardiac" portal hypertension) |
Risk factors for progression once CSPH is established
- Ongoing liver injury — continued alcohol use, untreated viral hepatitis, or uncontrolled metabolic risk factors accelerate the rise in HVPG and shorten screening/surveillance intervals (see Assessment & Diagnosis).
- Variceal size and endoscopic stigmata — medium/large caliber and red wale marks/red spots are the two endoscopic features that most directly predict bleeding, independent of HVPG measurement.[1]
- Child-Pugh class — any size of varix in a Child-Pugh C patient is classified as "high-risk," reflecting the added risk conferred by advanced synthetic failure itself, not just the varix's appearance.[1]
- HVPG level itself — bleeding risk rises roughly in proportion to HVPG once it exceeds about 12 mmHg; a rise in HVPG on treatment (or failure to fall) also predicts treatment failure, though routine repeat HVPG measurement is not practical outside specialized centres.[2]
- A prior variceal bleed is the single strongest predictor of a future bleed absent secondary prophylaxis — see Treatment Targets and Choose Initial Treatment.
Red Flags & Refer
- Hematemesis or melena in any patient with known or suspected portal hypertension — treat as active variceal hemorrhage until proven otherwise. This is a medical emergency requiring immediate emergency department assessment and, unlike the Cirrhosis chapter's ambulatory scope, is managed in full in this chapter — see Choose Initial Treatment and Clinical Decision Algorithm for the complete in-hospital protocol.
- New fever, abdominal pain/tenderness, or otherwise-unexplained worsening encephalopathy or renal function in a patient with ascites — suspect spontaneous bacterial peritonitis; keep a low threshold for diagnostic paracentesis even without classic severe abdominal pain, and note that routine diagnostic paracentesis is recommended for any hospital admission with ascites regardless of symptoms (see Assessment & Diagnosis).[2]
- Hemodynamic instability, ongoing hematemesis, agitation, or altered mental status in a patient with suspected variceal hemorrhage — assess airway protection needs and consider elective intubation before endoscopy rather than after a witnessed aspiration event.
- New or worsening ascites, rapid abdominal distension, or symptoms after large-volume paracentesis (hypotension, worsening renal function, new confusion) — evaluate for paracentesis-induced circulatory dysfunction (see Special Populations) or an inadequate albumin dose.
- Rapidly rising creatinine in a patient with ascites — evaluate urgently for hepatorenal syndrome; this is covered in the Cirrhosis chapter's Special Populations section and remains a specialist-directed, typically inpatient diagnosis of exclusion.
- New acute abdominal pain with a new diagnosis of portal or splenic vein thrombosis — refer urgently; bowel ischemia from mesenteric venous extension is a genuine, time-critical risk.
Assessment & Diagnosis
Confirming portal hypertension and estimating its severity
HVPG measurement (hepatic vein catheterization comparing wedged and free pressures) is the reference standard but is invasive, technically demanding, and available only at specialized centres — it is not something a community pharmacist needs to calculate, and (matching the Cirrhosis chapter's own precedent for Child-Pugh/MELD) no calculator for it is embedded on this platform. In practice, portal hypertension and its severity are inferred non-invasively.
Non-invasive diagnosis of clinically significant portal hypertension (CSPH)
In a patient with compensated advanced chronic liver disease (cACLD — the modern term for compensated cirrhosis or advanced fibrosis), Baveno VII combines liver stiffness measurement (LSM, by elastography) with platelet count:[1]
| Finding | Interpretation |
|---|---|
| LSM ≤15 kPa and platelet count ≥150 × 10⁵/L | CSPH ruled out |
| LSM ≥25 kPa | CSPH ruled in (probable), regardless of platelet count |
| Everything in between | Indeterminate ("grey zone") — additional non-invasive testing (e.g., spleen stiffness) or HVPG measurement at a specialized centre is needed |
Deciding who needs screening endoscopy at all
The same LSM/platelet combination, using slightly different thresholds validated specifically for ruling out varices needing treatment rather than CSPH broadly, identifies patients who can safely skip screening endoscopy altogether:[1]
- Baveno VI criteria — LSM <20 kPa and platelet count >150 × 10⁵/L: very low probability of high-risk varices; endoscopy can be deferred, with LSM and platelet count reassessed annually.
- Expanded Baveno VI criteria — LSM <25 kPa and platelet count >110 × 10⁵/L: spares more patients an endoscopy while keeping the proportion of missed high-risk varices low, at the cost of slightly more false negatives than the original criteria.
- Any patient not meeting these criteria needs screening upper endoscopy.
- Once a patient is started on an NSBB — whether for CSPH-driven decompensation prevention or for confirmed high-risk varices — routine surveillance endoscopy is no longer needed, since progression on endoscopy would not change the management already in place; this is a genuine, current Baveno VII practice shift away from the older "repeat endoscopy on a fixed schedule regardless of therapy" approach.[1]
What counts as a "high-risk" varix
Per the current Baveno VII definition, a varix is high-risk (and therefore warrants primary prophylaxis) if it is medium/large (≥5 mm) in size, regardless of size if red wale marks/red spots are present, or any size in a Child-Pugh C patient.[1]
Ascites: confirming portal hypertension is the cause
The serum-ascites albumin gradient (SAAG) — serum albumin minus ascitic fluid albumin, drawn on the same day — is the key diagnostic test distinguishing portal-hypertensive ascites from other causes. A SAAG ≥11 g/L (≥1.1 g/dL) indicates portal hypertension as the underlying mechanism with approximately 97% accuracy, far outperforming the older exudate/transudate total-protein classification it replaced.[3] A SAAG below this threshold should prompt a search for a non-portal-hypertensive cause (peritoneal carcinomatosis, tuberculous peritonitis, pancreatic ascites, nephrotic syndrome) rather than assuming cirrhosis explains the fluid.
Diagnostic paracentesis — routine on any hospital admission with ascites
Send ascitic fluid for cell count with differential, albumin (for SAAG), total protein, and culture — ideally inoculated directly into blood culture bottles at the bedside, which meaningfully improves culture yield over sending fluid to the lab in a plain container. A polymorphonuclear (PMN/neutrophil) count ≥250 cells/mm³ confirms spontaneous bacterial peritonitis (SBP) and is treated as such regardless of whether culture ultimately grows an organism — culture-negative neutrocytic ascites is managed identically to culture-positive SBP.[2] Because SBP is frequently asymptomatic or subtle, diagnostic paracentesis is recommended routinely on any hospital admission with ascites, not only when fever or abdominal pain is present.[2] A markedly elevated total protein, multiple organisms on culture (polymicrobial), or a low ascitic glucose should raise suspicion for secondary peritonitis (e.g., a perforated viscus) rather than SBP, and warrants urgent surgical evaluation rather than antibiotics alone.
Imaging
Doppler ultrasound of the portal vein assesses patency (excluding portal vein thrombosis) and flow direction — reversed (hepatofugal) flow is itself a sign of advanced portal hypertension. Cross-sectional imaging obtained for other indications often incidentally shows varices or collaterals and should prompt a screening-endoscopy discussion if one has not already occurred.
Treatment Targets
- Prevent a first variceal bleed in a patient found to have high-risk varices, and prevent further decompensation in a patient with confirmed CSPH even before varices form.
- Control active variceal hemorrhage rapidly and safely when it occurs — this is a genuine medical emergency with substantial short-term mortality if management is delayed.
- Prevent rebleeding after a first variceal hemorrhage through appropriate combination secondary prophylaxis.
- Diagnose and treat spontaneous bacterial peritonitis promptly, not only prevent it.
- Avoid iatrogenic harm from over-transfusion, over-resuscitation, or unmonitored NSBB use in a patient whose circulatory reserve can no longer tolerate it.
- Coordinate with, rather than duplicate, the Cirrhosis chapter's ascites diuretic-titration protocol and the Hepatic Encephalopathy chapter's lactulose/rifaximin protocol.
Choose Initial Treatment
Four scenarios, four different urgency levels
Primary prophylaxis
Compensated CSPH or high-risk varices found on screening: NSBB (carvedilol preferred) or endoscopic band ligation.
Acute variceal hemorrhage
A life-threatening emergency: resuscitation, a vasoactive drug, antibiotics, urgent endoscopy, and TIPS criteria — all started together, not sequentially.
Secondary prophylaxis
After any bleed: combination NSBB plus endoscopic band ligation, not either alone.
Spontaneous bacterial peritonitis
Diagnose by ascitic PMN count; treat confirmed episodes with antibiotics plus albumin, not prophylaxis alone.
Once screening endoscopy identifies high-risk varices (see Assessment & Diagnosis for the definition), start either an NSBB or endoscopic variceal ligation (EVL), with a preference for an NSBB where tolerated — NSBBs lower portal pressure globally and reduce the risk of decompensation events beyond bleeding, a benefit banding alone does not confer.[1]
Carvedilol is the preferred first NSBB in current consensus: its added alpha-1 blocking activity produces a significantly greater HVPG reduction than propranolol alone by relaxing intrahepatic vascular tone directly, alongside the splanchnic-inflow reduction shared by all NSBBs.[1] Start 6.25 mg once daily; if tolerated (systolic BP remaining above roughly 90-95 mmHg), increase after several days to 6.25 mg twice daily, to a practical maximum of 12.5 mg/day — higher doses add hypotension risk without a demonstrated further portal-pressure benefit for this indication.[1]
Propranolol or nadolol remain effective alternatives, titrated to a resting heart rate of 55-60 beats/min (see Drug Tables for exact titration steps — full detail is also carried in the Cirrhosis chapter).
Endoscopic band ligation is the appropriate primary-prophylaxis choice when an NSBB is contraindicated or not tolerated.
Do not routinely combine an NSBB with EVL for primary prophylaxis — current consensus reserves combination therapy for secondary prophylaxis, after a bleed has already occurred, not first-time prevention.[1]
A genuine, current practice shift: start NSBB therapy (carvedilol preferred) in compensated cACLD with non-invasively confirmed CSPH even before any varices are found, specifically to prevent the first decompensation event (predominantly ascites) rather than only to prevent bleeding. This is supported directly by the PREDESCI trial, a randomized, double-blind, placebo-controlled study in compensated patients with CSPH but without high-risk varices, in which NSBB therapy significantly reduced the risk of a first decompensation event compared with placebo over a median follow-up of roughly three years.[4][1]
Airway and resuscitation — assess airway protection before endoscopy in any patient with ongoing hematemesis, altered mental status, or agitation; consider elective intubation rather than risking aspiration during the procedure. Resuscitate with a restrictive transfusion strategy: target hemoglobin 7-8 g/dL, not a "normal" value — a landmark randomized trial found restrictive transfusion significantly reduced mortality and rebleeding compared with a liberal strategy, and specifically benefited the cirrhotic/Child-Pugh A-B subgroup; over-transfusion raises portal pressure by restoring intravascular volume and worsens bleeding.[5] Avoid over-aggressive crystalloid resuscitation for the same portal-pressure reason. Higher transfusion targets are reasonable for ongoing massive hemorrhage, hemodynamic instability, or active cardiovascular disease.[5]
Vasoactive drug — start immediately, before endoscopy, and continue for 2-5 days once bleeding is controlled: octreotide, 50 mcg IV bolus, then 50 mcg/hour continuous infusion, is the practical first-line choice in Canadian practice; a second 50 mcg bolus may be given in the first hour if hemostasis is not achieved. Terlipressin (a vasopressin analogue with the strongest mortality-reduction trial evidence internationally, dosed 2 mg IV every 4 hours initially, stepped down to 1 mg every 4 hours once bleeding is controlled) has no Drug Database entry as of this writing — verify current local formulary/marketing status before assuming availability. Legacy agent vasopressin requires concomitant nitroglycerin to offset its systemic (including coronary) vasoconstrictive effects and has been largely displaced by octreotide/terlipressin given its narrower safety margin.
Prophylactic antibiotics — core therapy, not merely infection prevention: give to every patient with cirrhosis and suspected or confirmed variceal hemorrhage, regardless of whether infection is clinically apparent — antibiotic prophylaxis in this setting measurably reduces mortality, rebleeding, and infectious complications (including SBP), on top of the vasoactive drug and endoscopic therapy.[2] Ceftriaxone 1 g IV once daily is preferred in advanced cirrhosis, in patients already on quinolone prophylaxis, or in settings with high fluoroquinolone resistance; norfloxacin 400 mg PO twice daily is a reasonable alternative in less advanced disease. Continue for up to 7 days.
Endoscopy, with a pre-procedure prokinetic — perform upper endoscopy within 12 hours of presentation once the patient is hemodynamically resuscitated (not the instant they arrive, and not delayed past this window). Consider erythromycin (or metoclopramide as an alternative) IV 30-120 minutes beforehand to clear retained gastric blood and improve visualization. Endoscopic band ligation is first-line for esophageal varices; isolated gastric varices (a different endoscopic problem) are generally treated with cyanoacrylate glue injection rather than banding, given a higher rebleeding risk with banding at that site.
Rescue if endoscopic control fails — balloon tamponade (Sengstaken-Blakemore/Minnesota tube) or a self-expanding covered esophageal metal stent is a temporizing bridge only, not definitive therapy; limit balloon-tamponade dwell time (generally under 24 hours) given ischemia/perforation risk, and proceed to TIPS.
Early ("pre-emptive") TIPS — place within 24-72 hours of presentation in patients at genuinely high risk of treatment failure: Child-Pugh class C (score 10-13), or Child-Pugh B with active bleeding seen at endoscopy. A landmark randomized trial in exactly this population found early TIPS (placed within 72 hours) significantly reduced treatment failure and improved survival compared with the standard drug-plus-endoscopic-therapy approach, a finding subsequently confirmed by later trials and meta-analyses.[6]
Do not routinely correct an elevated INR with fresh frozen plasma or transfuse platelets based on lab values alone before the endoscopic procedure itself — the Canadian Association for the Study of the Liver's own Choosing Wisely Canada recommendation specifically advises against this before minor invasive procedures including endoscopic variceal band ligation, reflecting the same "rebalanced hemostasis" concept the Cirrhosis chapter describes for elective procedures.[7] Do not give tranexamic acid or another antifibrinolytic — a large randomized trial in GI bleeding broadly found no mortality benefit and a signal of harm from venous thromboembolism, and it has no established role in variceal hemorrhage specifically.
Combination therapy — an NSBB (carvedilol preferred, or propranolol/nadolol titrated to resting HR 55-60 bpm) plus a program of endoscopic band ligation sessions repeated roughly every 1-4 weeks until variceal obliteration — is standard secondary prophylaxis, and is superior to either modality alone.[1] Start the NSBB once the acute vasoactive-drug infusion has been discontinued (overlapping the two adds unnecessary hypotension risk) and the patient is otherwise hemodynamically stable, typically before hospital discharge.
After obliteration, surveillance endoscopy is generally repeated at 3-6 months and then every 6-12 months, or per the treating gastroenterology/hepatology team's specific interval.
Rebleeding despite combination NSBB+EVL secondary prophylaxis is an indication for TIPS — see Treatment Failure / Next Step.
Confirmed SBP (ascitic PMN ≥250 cells/mm³, with or without a positive culture) is treated empirically with an IV third-generation cephalosporin — cefotaxime 2 g IV every 8 hours (the original pivotal-trial regimen) or ceftriaxone 2 g IV once daily (a practical, equally effective, once-daily alternative) — for approximately 5 days.[8][2] Repeat paracentesis at 48 hours is reasonable if the clinical response is inadequate, to confirm the PMN count has fallen and guide any antibiotic change.
Add IV albumin: 1.5 g/kg within 6 hours of diagnosis, then 1 g/kg on day 3. A landmark randomized trial found this regimen significantly reduced both renal impairment (33% to 10%) and in-hospital mortality (29% to 10%) compared with antibiotics alone, and it is particularly important in patients with the highest baseline risk (elevated bilirubin or creatinine at diagnosis), though it is widely given for all SBP episodes.[8]
Primary prophylaxis — daily antibiotic prophylaxis for ascitic fluid total protein <15 g/L plus impaired renal function or advanced liver disease (Child-Pugh ≥9 with bilirubin ≥3 mg/dL), matching the Cirrhosis chapter's criteria; secondary prophylaxis — indefinite prophylaxis for every patient with a prior SBP episode. Norfloxacin 400 mg PO once daily is the best-studied first choice for both; sulfamethoxazole-trimethoprim is a reasonable alternative. Plain oral ciprofloxacin has no standalone Drug Database entry (see this file's header) and is referenced by name only where used as an alternative.
Non-Pharmacologic Treatment
- Complete alcohol cessation and control of other underlying liver-injury drivers — see the Cirrhosis chapter; the single biggest lever for slowing progression of portal hypertension itself, not only the underlying liver disease.
- Sodium-restricted diet (2 g/day) once ascites is present — full protocol owned by the Cirrhosis chapter; not repeated here.
- Avoid NSAIDs — beyond the renal-perfusion risk already covered in the Cirrhosis chapter, NSAIDs also blunt the prostaglandin-mediated compensatory vasodilation that partially offsets portal hypertension's splanchnic effects.
- Endoscopic surveillance executed on the schedule determined by baseline findings and Baveno-criteria eligibility (see Assessment & Diagnosis and Monitoring & Follow-Up) — this is itself a non-pharmacologic intervention, not merely a monitoring afterthought, since it is what triggers primary prophylaxis in the first place.
- Transjugular intrahepatic portosystemic shunt (TIPS) — a radiologically placed intrahepatic stent connecting the portal and hepatic venous systems, directly and substantially lowering portal pressure; used as rescue therapy for uncontrolled/refractory acute bleeding, pre-emptively in the specific high-risk early-TIPS population, for rebleeding despite combination secondary prophylaxis, and for refractory ascites (see Special Populations for contraindications).
- Patient and family education: teach the specific warning signs — vomiting blood or "coffee-ground" material, and black, tarry, or foul-smelling stool — as reasons to seek emergency care immediately rather than waiting to see a regular provider.
- Nutrition — see the Cirrhosis chapter; never restrict protein.
Special Populations
Refractory ascites and circulatory dysfunction — the NSBB "window hypothesis"
NSBBs are genuinely beneficial across most of the disease course, but current consensus recognizes a real, specific exception: in a patient with refractory ascites (defined in the Cirrhosis chapter), NSBBs blunt the compensatory rise in cardiac output needed to tolerate large-volume paracentesis and can precipitate or worsen paracentesis-induced circulatory dysfunction (PICD) and hepatorenal syndrome — the mechanistic explanation for why continuing an NSBB unmonitored through this stage can increase mortality. Reduce or discontinue the NSBB if a patient with refractory ascites develops systolic blood pressure <90 mmHg, serum sodium <130 mmol/L, or acute kidney injury; discontinuation should be regarded as temporary, with cautious reinstatement once the precipitating event resolves rather than a permanent stop.[1] PICD itself is the direct physiologic consequence of the pathophysiology in Clinical Snapshot: removing a large volume of ascitic fluid acutely worsens the already-present splanchnic vasodilation, and without adequate albumin replacement (see Choose Initial Treatment and the Cirrhosis chapter), effective arterial volume falls further, activating the renin-angiotensin-aldosterone system and worsening sodium/water retention and renal perfusion.
TIPS candidacy and contraindications
Beyond the acute-hemorrhage and refractory-ascites indications above, contraindications to TIPS that a pharmacist should recognize when a patient is being considered for referral include severe heart failure or significant pulmonary hypertension (the shunt acutely increases venous return to the heart), uncontrolled or severe hepatic encephalopathy (TIPS bypasses hepatic ammonia clearance and can precipitate or worsen HE — see the Hepatic Encephalopathy chapter's Special Populations section for post-TIPS HE management), and severe hepatic synthetic failure where the procedure itself carries prohibitive risk.
Cardiac and peripheral vascular disease
Terlipressin and vasopressin are vasoconstrictors and carry a genuine risk of splanchnic, coronary, and peripheral ischemia; use with caution — or avoid — in significant ischemic heart disease or peripheral vascular disease, and pair vasopressin with concomitant nitroglycerin when it is used. Octreotide's hemodynamic effects are comparatively modest and it does not carry the same ischemic-risk profile.
Pregnancy
Physiologic plasma-volume expansion in pregnancy increases variceal bleeding risk, so pre-conception variceal screening and prophylaxis are ideal when a pregnancy is planned (see the Cirrhosis chapter). Propranolol is the preferred NSBB in pregnancy based on broader obstetric safety experience. If acute variceal hemorrhage occurs during pregnancy, octreotide is generally preferred over terlipressin/vasopressin, whose vasoconstrictive effects carry theoretical risk to uteroplacental blood flow; manage jointly with maternal-fetal medicine and hepatology.
Renal impairment
Terlipressin's hyponatremia risk is more pronounced with reduced renal reserve; norfloxacin and ciprofloxacin need renal dose adjustment; octreotide requires no specific renal adjustment. See the Cirrhosis chapter for diuretic-related renal monitoring.
Non-cirrhotic portal hypertension
Patients with portal vein thrombosis or other non-cirrhotic causes (see Risk Factors) still need variceal screening and prophylaxis using the same principles above, generally with preserved hepatic synthetic function. Anticoagulation for the underlying thrombosis itself is frequently appropriate and is not categorically withheld because of coexisting varices — this parallels the "rebalanced hemostasis" reasoning the Cirrhosis chapter applies to portal vein thrombosis in cirrhotic patients, and should be directed by hepatology/specialist input.
Monitoring & Follow-Up
- Resting heart rate and blood pressure at every visit for a patient on an NSBB, confirming the target (55-60 bpm for propranolol/nadolol; systolic BP tolerance for carvedilol) is met without symptomatic hypotension — full detail in the Cirrhosis chapter.
- Endoscopic surveillance interval determined by baseline findings and Baveno-criteria eligibility (see Assessment & Diagnosis): annually if Baveno VI criteria met and endoscopy deferred; 1-2 years if small varices found and no NSBB started; no further surveillance needed once an NSBB is started for CSPH or confirmed varices, since progression would not change management.[1]
- After endoscopic band ligation for primary or secondary prophylaxis, repeat sessions roughly every 1-4 weeks until variceal obliteration is achieved, then surveillance endoscopy at 3-6 months and every 6-12 months thereafter.
- A short course of a proton pump inhibitor (e.g., pantoprazole) after band ligation reduces the size and bleeding risk of post-banding mucosal ulcers as the bands slough, typically continued for about 2 weeks.
- Reassess SBP-prophylaxis criteria at every ascites-related visit — see the Cirrhosis chapter's ascites monitoring schedule, which this chapter's SBP diagnosis/treatment content complements.
- Renal function and electrolytes with every diuretic dose change and with any NSBB dose change in a patient with refractory ascites, given the window-hypothesis stopping criteria in Special Populations.
- See the Cirrhosis chapter for the broader routine surveillance schedule (Child-Pugh/MELD recalculation, hepatocellular carcinoma surveillance) that continues in parallel.
Common Errors to Avoid
- Delaying antibiotic prophylaxis in acute variceal hemorrhage until infection is clinically apparent — antibiotics are core therapy in this setting, proven to reduce mortality, not merely an infection-prevention afterthought.[2]
- Over-transfusing to a "normal" hemoglobin during acute variceal hemorrhage — a restrictive strategy (target 7-8 g/dL) improves outcomes, while over-transfusion raises portal pressure and worsens bleeding/rebleeding.[5]
- Giving fresh frozen plasma, vitamin K, or platelets to correct lab-value coagulopathy before paracentesis or endoscopic band ligation based on the numbers alone — explicitly flagged against by the Canadian Association for the Study of the Liver's Choosing Wisely Canada recommendation.[7]
- Using tranexamic acid or another antifibrinolytic for variceal (or GI bleeding generally) — no established benefit and a signal of thrombotic harm in the largest trial.
- Combining an NSBB with endoscopic band ligation for primary prophylaxis — current consensus reserves that combination for secondary prophylaxis only.[1]
- Continuing full-dose NSBB unmonitored once a patient develops refractory ascites, spontaneous bacterial peritonitis, acute kidney injury, or systolic BP <90 mmHg — apply the window-hypothesis stopping criteria rather than assuming the drug remains purely beneficial at every disease stage.[1]
- Treating isolated gastric varices with the same banding technique used for esophageal varices — cyanoacrylate glue injection is the appropriate endoscopic approach for isolated gastric varices.
- Missing routine diagnostic paracentesis on hospital admission with ascites because the patient has no fever or abdominal pain — SBP is frequently subtle or asymptomatic.[2]
- Managing confirmed SBP with antibiotics alone and omitting IV albumin, particularly in a patient with an elevated bilirubin or creatinine at diagnosis — the combination, not antibiotics alone, is what reduces renal impairment and mortality.[8]
- Attributing all chronic, low-grade GI blood loss in a cirrhotic patient to varices without considering portal hypertensive gastropathy, which needs an NSBB rather than endoscopic band ligation.
- Delaying endoscopy well past the 12-hour window once a patient is hemodynamically resuscitated, or performing it before adequate resuscitation has occurred — both extremes carry real risk.
Clinical Decision Algorithm
Drug Tables
How to use this table: Doses are adult doses and require patient-specific confirmation (renal/hepatic function, weight, allergy history). Pediatric dosing is weight-based per the current product monograph unless stated. Drug pricing is intentionally omitted. Select any drug name to open its complete Pharmasist Drug Database page.
Nonselective beta-blockers — primary/secondary variceal-bleed prophylaxis and compensated-CSPH decompensation prevention
Titrate to a resting heart rate of 55-60 beats/min (propranolol/nadolol) or the stated maximum dose (carvedilol), whichever comes first; hold/reduce for symptomatic hypotension, and apply the window-hypothesis stopping criteria in Special Populations once refractory ascites develops.
| Drug | Dosage | Adverse effects | Drug interactions | Comments |
|---|---|---|---|---|
| Carvedilol | Starting dose: 6.25 mg PO once daily Titration: Increase to 6.25 mg twice daily as tolerated Maximum: 12.5 mg PO daily | Hypotension, dizziness, fatigue, bradycardia | Additive hypotension with other antihypertensives; avoid combining with another nonselective beta-blocker | Preferred first-choice NSBB given its added alpha-1 blockade and greater HVPG reduction; also the preferred agent for starting therapy in compensated CSPH before high-risk varices develop.[1] |
| Propranolol | Starting dose: 20 mg PO twice daily Titration: Increase by 10 mg increments every 2-3 days to resting HR 55-60 bpm Maximum: 160 mg/day (up to 320 mg/day if ascites present) | Fatigue, bradycardia, bronchospasm, masked hypoglycemia symptoms in diabetes | Additive bradycardia with other rate-limiting agents; extensively hepatically metabolized | Preferred NSBB in pregnancy based on broader obstetric safety experience. Full titration steps are also carried in the Cirrhosis chapter's drug table.[2] |
| Nadolol | Dosing: Titrated once daily to resting HR 55-60 bpm | Fatigue, bradycardia, bronchospasm | Additive bradycardia with other rate-limiting agents | Renally rather than hepatically cleared, giving more predictable pharmacokinetics as liver disease advances; requires dose reduction in renal impairment.[2] |
Vasoactive drugs for acute variceal hemorrhage
Start immediately, before endoscopy, and continue for 2-5 days once bleeding is controlled.
| Drug | Dosage | Adverse effects | Drug interactions | Comments |
|---|---|---|---|---|
| Octreotide | Bolus: 50 mcg IV Infusion: 50 mcg/hour continuous IV, for 2-5 days | Hyperglycemia, bradycardia, abdominal cramping/nausea | None clinically significant at standard use | Practical first-line vasoactive agent in Canadian practice. A second 50 mcg bolus may be given within the first hour if hemostasis is not achieved.[2] |
| Terlipressin | Initial: 2 mg IV every 4 hours Step-down: 1 mg IV every 4 hours once bleeding controlled | Hyponatremia, splanchnic/coronary/peripheral ischemia, abdominal cramping | Additive ischemic risk with other vasoconstrictors | No Drug Database entry as of this writing — verify current local formulary/marketing status before assuming a slug exists. Internationally, the vasoactive agent with the strongest mortality-reduction trial evidence.[2] |
| Vasopressin | —: Specialist-directed dosing with concomitant nitroglycerin | Systemic, coronary, and peripheral vasoconstriction; arrhythmia; bowel ischemia | Requires concomitant nitroglycerin to offset ischemic risk | Legacy agent, largely displaced by octreotide/terlipressin given a narrower safety margin; avoid in significant coronary or peripheral vascular disease even with nitroglycerin co-administration.[2] |
Antibiotics — acute bleed prophylaxis and SBP prophylaxis/treatment
Antibiotic prophylaxis is core therapy in acute variceal hemorrhage, given to every patient regardless of apparent infection.
| Drug | Dosage | Adverse effects | Drug interactions | Comments |
|---|---|---|---|---|
| Ceftriaxone | Bleed prophylaxis: 1 g IV once daily, up to 7 days SBP treatment: 2 g IV once daily, approximately 5 days | Hypersensitivity reactions, biliary sludging with prolonged use | None clinically significant at short-course dosing | Preferred acute-bleed prophylaxis choice in advanced cirrhosis, prior quinolone exposure, or high local fluoroquinolone resistance. Also used for confirmed SBP treatment at the higher dose.[2] |
| Norfloxacin | SBP prophylaxis: 400 mg PO once daily, indefinitely for secondary prophylaxis Acute bleed prophylaxis: 400 mg PO twice daily, up to 7 days | GI upset, rare tendinopathy, QT prolongation | Chelation with divalent cations reduces absorption | Best-studied first agent for both primary and secondary SBP prophylaxis; alternative to ceftriaxone for acute-bleed prophylaxis in less advanced disease. Requires renal dose adjustment.[2] |
| Cefotaxime | SBP treatment: 2 g IV every 8 hours, approximately 5 days | Hypersensitivity reactions, GI upset | None clinically significant at standard use | The original pivotal-trial SBP treatment regimen; give together with IV albumin, not alone.[8] |
| Sulfamethoxazole-trimethoprim | Dosing: 800/160 mg (1 double-strength tablet) PO once daily, indefinitely | GI upset, hyperkalemia, rash, hematologic effects | Additive hyperkalemia with spironolactone; potentiates warfarin | Alternative to norfloxacin for SBP prophylaxis.[2] |
Albumin and paracentesis/SBP-treatment adjuncts
| Drug | Dosage | Adverse effects | Drug interactions | Comments |
|---|---|---|---|---|
| Albumin, IV (human) | SBP treatment: 1.5 g/kg IV within 6 hours of diagnosis, then 1 g/kg IV on day 3 Large-volume paracentesis: 6-8 g IV per litre of ascitic fluid removed beyond a 5 L threshold | Fluid overload if over-dosed; rare hypersensitivity | None clinically significant | Two distinct, non-interchangeable indications: reduces renal impairment and mortality when given with antibiotics for confirmed SBP, and reduces post-paracentesis circulatory dysfunction when given for large-volume paracentesis.[8] |
Pre-endoscopy prokinetic and post-banding adjuncts
| Drug | Dosage | Adverse effects | Drug interactions | Comments |
|---|---|---|---|---|
| Erythromycin | Dosing: IV, single dose 30-120 minutes pre-endoscopy | QT prolongation, GI upset, injection-site irritation | Numerous CYP3A4 interactions; additive QT effect with other QT-prolonging agents | Given IV 30-120 minutes before endoscopy in suspected acute variceal hemorrhage to clear retained gastric blood and improve visualization.[2] |
| Metoclopramide | Dosing: IV, single dose pre-endoscopy | Extrapyramidal symptoms, sedation | Additive sedation with CNS depressants; additive extrapyramidal risk with antipsychotics | Alternative prokinetic to erythromycin before endoscopy.[2] |
| Pantoprazole | Dosing: Standard once-daily PPI dose for approximately 2 weeks post-banding | Headache, GI upset; long-term use risks (osteoporosis, B12 deficiency, C. difficile) not relevant to this short course | None clinically significant at short-course dosing | Short course after endoscopic band ligation reduces the size and bleeding risk of post-banding mucosal ulcers as the bands slough, typically continued about 2 weeks. Esomeprazole or omeprazole are equally reasonable alternatives. |
Agents to avoid in acute variceal hemorrhage
| Drug | Dosage | Adverse effects | Drug interactions | Comments |
|---|---|---|---|---|
| Tranexamic acid | —: Avoid for variceal hemorrhage | Venous thromboembolism risk | Additive thrombotic risk with other prothrombotic states | No established mortality benefit in GI bleeding and a signal of thrombotic harm in the largest trial; not part of standard variceal hemorrhage management.[5] |
References
- Baveno VII – Renewing consensus in portal hypertension — Baveno VII Faculty (de Franchis R, Bornschein J, Ripoll C, Abraldes JG, Reiberger T, Vinel JP, et al.). J Hepatol 2022;76(4):959-974
- AASLD Practice Guidance on risk stratification and management of portal hypertension and varices in cirrhosis — American Association for the Study of Liver Diseases (Kaplan DE, Bosch J, Ripoll C, Thiele M, Fortune BE, Simonetto DA, Garcia-Tsao G). Hepatology 2024;79(5):1180-1211
- The Serum-Ascites Albumin Gradient Is Superior to the Exudate-Transudate Concept in the Differential Diagnosis of Ascites — Annals of Internal Medicine (Runyon BA, Montano AA, Akriviadis EA, Antillon MR, Irving MA, McHutchison JG). Ann Intern Med 1992;117(3):215-220
- Beta blockers to prevent decompensation of cirrhosis in patients with clinically significant portal hypertension (PREDESCI): a randomised, double-blind, placebo-controlled trial — The Lancet (Villanueva C, Albillos A, Genesca J, Garcia-Pagan JC, Calleja JL, Aracil C, et al.). Lancet 2019;393(10181):1597-1608
- Transfusion Strategies for Acute Upper Gastrointestinal Bleeding — New England Journal of Medicine (Villanueva C, Colomo A, Bosch A, Concepcion M, Hernandez-Gea V, Aracil C, et al.). N Engl J Med 2013;368(1):11-21
- Early Use of TIPS in Patients with Cirrhosis and Variceal Bleeding — New England Journal of Medicine (Garcia-Pagan JC, Caca K, Bureau C, Laleman W, Appenrodt B, Luca A, et al.). N Engl J Med 2010;362(25):2370-2379
- Choosing Wisely Canada-Top Five List in Hepatology: Official Position Statement of the Canadian Association for the Study of the Liver (CASL) and Choosing Wisely Canada (CWC) — Canadian Association for the Study of the Liver and Choosing Wisely Canada (Brahmania M, Renner EL, Coffin CS, Yoshida EM, Wong P, Zeman M, Shah H). Ann Hepatol 2019;18(1):165-171
- Effect of Intravenous Albumin on Renal Impairment and Mortality in Patients with Cirrhosis and Spontaneous Bacterial Peritonitis — New England Journal of Medicine (Sort P, Navasa M, Arroyo V, Aldeguer X, Planas R, Ruiz-del-Arbol L, et al.). N Engl J Med 1999;341(6):403-409
Professional reference only. Apply clinical judgment and verify patient-specific decisions against current authoritative guidance and product information.