Portal Hypertension

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.

Endoscopic view of esophageal varices with prominent red wale marks — longitudinal red streaks on the variceal surface that, per Baveno VII, define a "high-risk" varix warranting prophylaxis regardless of its measured size. (Samir, public domain, via Wikimedia Commons)

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.

Classifying portal hypertension by anatomic site

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.

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]

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.

• 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.

• 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.

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.

• 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.

1. Failure to control acute variceal hemorrhage despite vasoactive drug plus endoscopic band ligation (ongoing bleeding, an unstable transfusion requirement, or rebleeding within 5 days) — proceed to balloon tamponade or a covered esophageal stent as a bridge, then TIPS; do not simply repeat endoscopic therapy indefinitely while the patient remains unstable.^[6][1]
2. A patient meeting early-TIPS criteria (Child-Pugh C 10-13, or Child-Pugh B with active bleeding at endoscopy) not yet referred — refer for TIPS placement within 24-72 hours rather than waiting to see whether standard therapy fails first; this is a pre-emptive strategy, not a rescue-only one.^[6]
3. Rebleeding despite combination NSBB + EVL secondary prophylaxis — refer for TIPS evaluation.^[1]
4. NSBB not tolerated because of hypotension in a patient with refractory ascites — apply the window-hypothesis stopping criteria in Special Populations; if the NSBB must be stopped, secondary prophylaxis then relies on the endoscopic band ligation program alone until/unless the NSBB can be safely reinstated.
5. Recurrent SBP despite adequate prophylaxis — reassess for antibiotic resistance (culture and sensitivity from the prior episode, if available) and consider a different prophylactic agent class; a first or recurrent SBP episode is itself a transplant-referral trigger per the Cirrhosis chapter's criteria.
6. Inadequate clinical response at 48 hours during SBP treatment — repeat paracentesis to confirm the PMN count is falling; if not, broaden or change antibiotic coverage and reconsider secondary peritonitis (see Assessment & Diagnosis).

• 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.

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.

Vasoactive drugs for acute variceal hemorrhage

Start immediately, before endoscopy, and continue for 2-5 days once bleeding is controlled.

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.

Albumin and paracentesis/SBP-treatment adjuncts

Pre-endoscopy prokinetic and post-banding adjuncts

Agents to avoid in acute variceal hemorrhage