Portal Hypertension and Cirrhosis Complications: Pulmonary, Coagulopathy, and Vascular Complications

0:00Welcome to Board Pearls. This is episode three of three of the Portal Hypertension and Cirrhosis Complications chapter, in the Liver Disease module. This episode is the remaining decompensated complications: hepatopulmonary syndrome, portopulmonary hypertension, hepatic hydrothorax, the coagulopathy where the INR misleads about bleeding risk, and the vascular liver diseases.

0:23Start with the pulmonary side, because hepatopulmonary syndrome and portopulmonary hypertension look related but are physiologically opposite, and knowing which one you're looking at picks the therapy and changes the transplant calculus. Hepatopulmonary syndrome is a shunt problem at the capillary level: the cirrhotic liver fails to clear circulating vasoactive mediators, chiefly nitric oxide, which reach the pulmonary capillary bed and dilate the capillaries, so deoxygenated blood enters a bed too wide for full equilibration and traverses the lung without picking up the oxygen it should, and the patient becomes hypoxemic in a way that gives the diagnosis away.

1:00That behavior is platypnea and orthodeoxia: the patient feels more short of breath upright and improves lying down, with saturations falling on standing and rising when flat, and the mechanism is gravity, because the dilated capillaries are denser at the bases, so upright posture preferentially perfuses them and oxygenation falls while lying flat redistributes flow, which is the highest-yield recognition cue here because the directionality is the opposite of what a stem-reader expects.

1:26Diagnosis is contrast echocardiography, injecting agitated saline and timing the microbubble appearance in the left atrium, where a delayed appearance three to six cardiac cycles after the right atrium opacifies is an intrapulmonary shunt and thus this syndrome, while an early appearance within one to three cycles is an intracardiac shunt and something else, with confirmation by a shunt-quantifying scan or oxygenation criteria of a room-air arterial oxygen below eighty with an elevated alveolar-arterial gradient.

1:56Severity grades by arterial oxygen, and the threshold that matters most is sixty, because a value below sixty earns transplant-priority exception points, and the only definitive cure is transplant, since supplemental oxygen is supportive but doesn't address the shunt and no vasodilator helps because the capillaries are already dilated, so the lung doesn't need vasodilation, it needs the capillaries to constrict back, which only removing the failing liver accomplishes.

2:23Portopulmonary hypertension is the opposite physiology, where the pulmonary arterioles remodel and constrict in the setting of portal hypertension, so vascular resistance and artery pressure rise and the right ventricle works against an obstruction rather than a wide-open bed, making it a resistance problem at the arteriolar level while the other is a dilation problem at the capillary level, and that single distinction predicts everything.

2:46Diagnosis is by right heart catheterization, because echocardiography alone overestimates pressures and can't reliably distinguish it, and the three criteria run together: a mean pulmonary arterial pressure of at least twenty-five, a pulmonary vascular resistance over three Wood units, and a wedge pressure of at most fifteen, with the wedge being the discriminator that excludes left-heart-driven pulmonary hypertension, because a high wedge means the pressure comes from the left side rather than intrinsic arteriolar disease.

3:15The transplant calculus is where this gets tested: severe disease with a mean pressure at or above fifty is a transplant contraindication, because the right ventricle can't handle the preload swings against a fixed high afterload and right ventricular failure on the table is prohibitive, while a mean pressure in the thirty-five-to-forty-five range is a relative contraindication that becomes acceptable when pre-transplant vasodilator therapy lowers the gradient, and mild-to-moderate disease responds to pulmonary-hypertension-specific therapy, the PDE-5 inhibitors, endothelin antagonists, and prostacyclins, so lowering the pressure makes the patient transplantable. That's the asymmetry: vasodilators help portopulmonary hypertension because the lesion is constriction, and don't help hepatopulmonary syndrome because the lesion is already dilation.

4:00Hepatic hydrothorax is the third pulmonary-adjacent complication and is mechanically distinct, a transudative pleural effusion forming because microscopic diaphragmatic defects let ascitic fluid pass directly into the pleural space, drawn through by the negative intrathoracic pressure of inspiration, and most are right-sided because the right hemidiaphragm has more fenestrations and less muscular reinforcement. The clinically important point is that ascites doesn't need to be detectable on exam, because fluid can pass directly through the diaphragm without first accumulating in the peritoneum, so a cirrhotic with a new right-sided effusion and minimal ascites should still trigger this on the differential, with diagnostic thoracentesis showing fluid that mirrors ascites, low protein and transudative with a high albumin gradient, and a pleural neutrophil count over two hundred fifty defining spontaneous bacterial empyema treated like peritonitis plus prophylaxis.

4:48Initial management is sodium restriction, diuretic optimization, and therapeutic thoracentesis as needed, with TIPS controlling the effusion in most refractory cases by lowering the portal pressure that fuels the trans-diaphragmatic flow.

5:02The most-tested point is what not to do: chest tube placement and pleurodesis are contraindicated, because the tract becomes infected, the pleural space is continuously contaminated by ongoing transit of ascitic fluid, the resulting empyema can preclude transplant, and drainage produces enormous protein and fluid loss that worsens the cirrhotic state, so small-bore catheters are reserved for palliation in non-transplant candidates and surgical repair for highly selected refractory cases, and the question typically rewards the salt-and-diuretic-with-TIPS answer over any drainage-based option.

5:32Now coagulopathy, where the cirrhotic patient sets a trap. The instinct is to treat the INR, because an INR of two in a non-cirrhotic on warfarin predicts bleeding and pre-procedural correction makes sense, but the same INR of two in an advanced cirrhotic doesn't predict bleeding and correction may make things worse, because cirrhotic hemostasis is rebalanced rather than impaired.

5:56Rebalanced means parallel reductions: the liver makes less of the pro-coagulant factors, so the INR rises, but it also makes less of the anti-coagulant factors, protein C, protein S, and antithrombin, and the fibrinolytic proteins fall in parallel too, so the net state can be balanced, hypercoagulable, or hypocoagulable, and the INR and prothrombin time measure only the pro-coagulant arm, so they don't predict bleeding risk in cirrhosis. That's why portal vein thrombosis and venous thromboembolism rates run elevated despite the elevated INR, and why prophylaxis is standard during cirrhotic hospitalization, because the patient who looks anticoagulated on paper is paradoxically prothrombotic.

6:40The assay that captures the rebalanced state is viscoelastic testing, which integrates clot formation, propagation, and lysis on whole blood and returns a global picture rather than a single time, identifying the specific defect, low fibrinogen, low platelet contribution, or hyperfibrinolysis, so targeted correction by defect reduces transfusion without raising bleeding.

7:03Periprocedural correction is keyed to the procedure, not the lab value: low-risk procedures need no prophylactic transfusion regardless of the numbers, with paracentesis the canonical example, its bleeding rate under one percent even with a very high INR or very low platelets, so routine plasma for an elevated INR isn't evidence-based, and the same applies to thoracentesis and diagnostic endoscopy.

7:26The reason transfusion fails to help is mechanistic: correcting the INR doesn't change bleeding because the INR wasn't predicting it, and plasma adds volume that raises portal pressure and can worsen variceal bleeding, so the transfusion meant to reduce bleeding can worsen it, with additional transfusion risks and no offsetting benefit, so the right move for paracentesis or thoracentesis is to proceed without correction. So the common stem of an advanced cirrhotic with a high INR and low platelets scheduled for diagnostic paracentesis, where the hospitalist orders plasma and platelets, has the correct answer of holding both and tapping without correction, and vitamin K doesn't help either because the defect is reduced synthesis rather than vitamin K deficiency unless there's concurrent cholestasis.

8:10High-risk procedures shift the calculation, with fibrinogen replacement to keep above a floor and platelet transfusion to keep above fifty thousand reserved for documented bleeding or the highest-risk settings like neuraxial anesthesia, and the platelet threshold scaling to the bleeding consequence, but even then the INR alone is not a transfusion trigger.

8:30A few practical points: cirrhotics get standard prophylaxis when hospitalized because the rebalanced state runs prothrombotic often enough, anticoagulation decisions weigh the prothrombotic substrate against the bleeding risk, the direct oral anticoagulants are increasingly used in compensated cirrhosis but avoided in advanced disease, severe coagulopathy, or after recent variceal bleeding, and cirrhotic cardiomyopathy is screened by echocardiogram and ECG because it converts a marginal procedure into a hemodynamic crisis when not anticipated.

9:00Now vascular liver disease, where the level of vascular obstruction determines both presentation and treatment, so recognize the level and the answer follows. Budd-Chiari syndrome is hepatic vein outflow obstruction, with abdominal pain, hepatomegaly, ascites, and elevated transaminases, often acutely in a previously healthy patient, and imaging showing absent hepatic vein flow with caudate lobe enlargement in most cases, which is the most-tested anatomic clue because the caudate drains independently through small veins directly into the vena cava, so when the main veins occlude the rest of the liver congests while the caudate keeps draining and hypertrophies, with the injury congestive and the ascites classically showing a high albumin gradient and high protein because the leak is from congested sinusoids.

9:44Thrombophilia workup is mandatory, because the syndrome is almost always a clue to a systemic driver, and the single most common is a myeloproliferative neoplasm with the JAK2 mutation, so JAK2 testing has the highest yield because the mutation is often positive before overt erythrocytosis develops, with antiphospholipid syndrome, paroxysmal nocturnal hemoglobinuria, factor V Leiden, and oral contraceptives following, so a young woman on the pill with new Budd-Chiari still gets JAK2 first.

10:07Treatment is stepwise: anticoagulation first in every patient without contraindication, sufficient alone in some but most needing escalation, then angioplasty with or without stenting for focal webs or short stenoses, then TIPS to bypass the obstruction and decompress the liver, and transplant for fulminant disease or end-stage failure of the prior steps. So a patient on anticoagulation alone for a week or two who fails to improve with a focal short-segment stenosis has met the criteria to escalate to angioplasty, and systemic thrombolysis is not the standard because of high bleeding risk without proven benefit.

10:44Sinusoidal obstruction syndrome occupies a different level, with the injury at the sinusoid producing downstream central venule occlusion while the hepatic veins stay patent on Doppler, and diagnosis depends on exposure history, the classic trigger being myeloablative stem cell transplant conditioning, with pyrrolizidine-alkaloid herbal teas producing the same pathology.

11:04Presentation is painful hepatomegaly, weight gain, ascites, and rising bilirubin within the first three weeks after transplant, and the classical criteria are a bilirubin over two plus two of painful hepatomegaly, weight gain over five percent, or ascites within that window, with the distinction from Budd-Chiari sitting on imaging, since patent hepatic veins exclude Budd-Chiari and direct attention to the sinusoidal level.

11:28Defibrotide is approved for severe disease, a compound with antithrombotic and endothelial-protective effects, dosed at six and a quarter milligrams per kilogram every six hours for at least three weeks, reducing mortality compared with historical controls and the only therapy with prospective evidence, while heparin monotherapy is ineffective and adds bleeding risk in thrombocytopenic patients, the old thrombolytic-plus-heparin regimen was abandoned for fatal hemorrhage, and ursodiol prophylaxis started before conditioning reduces incidence but has no role as monotherapy in established disease.

11:59Portal vein thrombosis is the third level and the most common in cirrhotic practice, developing in up to a quarter of cirrhotics over time because portal flow slows and the procoagulant imbalance favors thrombosis, while in non-cirrhotics the drivers are thrombophilia or local inflammation like pancreatitis seeding the splanchnic system, and anticoagulation selection depends on extent, mesenteric involvement, transplant candidacy, and bleeding risk.

12:24The most-tested trap is excluding hepatocellular carcinoma with tumor thrombus before anticoagulating, because tumor thrombus is intravascular tumor extension, which is metastatic disease, and anticoagulating it doesn't help and may worsen bleeding, with the discriminator being arterial enhancement, since bland thrombus doesn't enhance while tumor thrombus enhances arterially because it carries the tumor's hepatic-artery feeding pattern, and cavernous transformation, where collaterals form around chronic thrombosis, generally doesn't warrant anticoagulation because the vein is functionally occluded.

12:54Idiopathic non-cirrhotic portal hypertension, or porto-sinusoidal vascular disease, occupies the fourth level, a heterogeneous group of obliterative portal venopathies producing presinusoidal portal hypertension without cirrhosis, and the classic discordance is the high-yield clue: the patient has clinically significant portal hypertension with splenomegaly, varices, and thrombocytopenia, but synthetic function is preserved, the liver stiffness may be only mildly elevated, and the pressure gradient can be normal or only modestly elevated, so the mismatch between clinical portal hypertension and a low or normal gradient distinguishes it from sinusoidal cirrhosis.

13:33Biopsy confirms it, showing nodular regenerative hyperplasia or obliterative venopathy without bridging fibrosis, and the associations matter for recognition, HIV with prior didanosine the most-tested, with oxaliplatin and azathioprine following, and management focuses on the portal-hypertensive complications, variceal screening, beta-blockers, and banding, because there's no parenchymal disease to treat since the lesion is vascular.

13:57And shock liver is the systemic equivalent, massive transaminases often over five thousand with an elevated LDH mirroring the AST, modest hyperbilirubinemia, and a mild alkaline phosphatase rise, triggered by documented hypotension or hypoxia often in an ICU patient, and the trajectory gives it away, because the transaminases halve every day or two once perfusion is restored and the picture resolves over a week or so, with the discriminators being an LDH-to-ALT ratio that often exceeds one and a half and the temporal association with the hemodynamic insult, and acetaminophen the major confounder when transaminases run very high, settled by exposure history.

14:29So the synthesis. The pulmonary vascular complications break into three mechanisms sharing only the cirrhotic substrate: hepatopulmonary syndrome is capillary dilation with shunt physiology, recognized by platypnea and orthodeoxia and confirmed by delayed contrast echocardiography, with an arterial oxygen below sixty earning exception points and transplant the only cure; portopulmonary hypertension is arteriolar constriction, diagnosed by right heart catheterization, treated with vasodilators for mild-to-moderate disease, and contraindicated for transplant above a mean pressure of fifty; and hepatic hydrothorax is mechanical trans-diaphragmatic flow, managed with salt restriction, diuretics, and TIPS, with chest tubes and pleurodesis explicitly off the menu.

15:06Coagulopathy looks like a bleeding tendency on paper and isn't, because the INR doesn't predict bleeding in cirrhosis, rebalanced hemostasis governs procedures, paracentesis and thoracentesis require no correction, and viscoelastic testing captures the real state.

15:19And the vascular liver diseases divide by anatomic level: the hepatic vein for Budd-Chiari with JAK2 as the highest-yield workup and a stepwise escalation, the sinusoid for sinusoidal obstruction syndrome with defibrotide the only therapy with mortality evidence, the portal vein for portal vein thrombosis with tumor thrombus the high-yield trap, the presinusoidal venule for porto-sinusoidal vascular disease with the discordance between clinical portal hypertension and a normal gradient, and systemic perfusion for shock liver, which is a trajectory diagnosis as much as a magnitude one.

15:45The next chapter moves into liver tumors and transplantation: benign lesions versus hepatocellular carcinoma with surveillance and the imaging-diagnosis framework, the treatment algorithm with immunotherapy first-line for advanced disease, cholangiocarcinoma with its own staging and the perihilar transplant protocol, and the transplant evaluation framework anchored by the allocation score.

16:08For the full chapter, the practice vignettes, and the topic-tagged question bank, head to board pearls dot com. You'll find the rest of the series on Apple Podcasts, Spotify, or wherever you listen to podcasts. That brings us to the end of episode three of three of chapter twenty-three, and I'll see you in the next one.