Hepatic synthesis of several procoagulant factors falls, prolonging PT/INR.
← LCCH Physiology
Module 06
How can the INR be prolonged while thrombosis remains possible?
Build the haemostatic system as simultaneous changes in procoagulants, anticoagulants, platelets, endothelium and fibrinolysis — then see why standard plasma tests capture only one slice.
Clinical frame
The organising problem.
Cirrhosis does not create simple auto-anticoagulation. Haemostasis is rebalanced but fragile: compensating changes can preserve thrombin generation while leaving patients vulnerable to either bleeding or thrombosis when the system is perturbed.
Mechanism · Procoagulants fall
Many clotting factors decrease
01↓ hepatic synthesis
02↓ procoagulant factors
03Prolonged PT / INR
Factors II, V, VII, IX, X and XI may be reduced; fibrinogen concentration and function vary with disease and inflammation.
PT/INR is sensitive to this arm of the system, which is why it looks dramatically abnormal.
Step 1 of 6
Change the patient. Then change what the test sees.
Compare the in-vivo system with the narrower view provided by routine plasma testing.
Add cirrhosis. Several hepatic proteins fall while endothelial biology also changes.
- 01
Procoagulants
ReferenceUsual factor availability
Selected scenarioSeveral factors decrease
- 02
Anticoagulants
ReferenceProtein C / S + antithrombin
Selected scenarioNatural restraint also decreases
- 03
Platelet–endothelial axis
ReferencePlatelets and VWF interact
Selected scenarioLow platelets; VWF compensation
- 04
What INR tells you
ReferencePT-based plasma endpoint
Selected scenarioCaptures only part of the change
Why this happens
Paired reductions can preserve net thrombin potential despite prolonged PT. Platelet–endothelial compensation adds another process outside the routine plasma assay.
What to look for
Bleeding and thrombosis remain possible in the same patient.
Limit of the model These effects are qualitative. They cannot be added to produce a validated net haemostatic score.
Sources and context: ISTH: rebalanced haemostasis (2021) ↗EASL bleeding and thrombosis guideline (2022) ↗
Go deeper
Mechanisms beneath the model.
Open the topics you need. Keep several explanations visible to compare mechanisms.
Evidence key: Established: broadly supported. Nuanced: context affects interpretation. Evolving: evidence or definitions remain unsettled. These labels describe physiological certainty, not a formal GRADE rating.
01INRPT/INR is a liver synthetic test, not a global bleeding-risk assayEstablished
PT-based assays measure time to fibrin formation in platelet-poor plasma and are insensitive to major compensating anticoagulant and endothelial changes.
- INR was standardised for vitamin-K antagonist therapy, not cirrhosis.
- It does not capture VWF, platelet function or protein C balance adequately.
- Correcting INR numerically does not guarantee clinical haemostasis.
02EndotheliumVWF and ADAMTS13 shift platelet biologyEstablished
Markedly increased VWF and reduced ADAMTS13 can enhance platelet adhesion and partly compensate for thrombocytopenia.
- Platelet number is only one determinant of primary haemostasis.
- Endothelial activation becomes more pronounced in severe illness.
- This compensation is itself unstable.
03ViscoelasticGlobal assays answer different questionsNuanced
TEG/ROTEM can describe clot kinetics and strength in whole blood, but they do not perfectly reproduce in-vivo endothelial biology or predict every procedural bleed.
- They can support targeted transfusion strategies in selected settings.
- A normal trace does not mean the haemostatic system is normal.
- Interpretation depends on assay and clinical context.
04AssaysThe result depends on the system placed in the tubeEstablished
PT measures time to an early fibrin endpoint after tissue-factor activation in platelet-poor plasma. It does not recreate intact endothelium or the full thrombomodulin–protein C pathway. Loss of anticoagulant restraint is therefore poorly reflected in a test sensitive to reduced procoagulant factors.
- Thrombin-generation experiments that include thrombomodulin ask a different question from routine PT.
- VWF-mediated adhesion under shear, platelet number and endothelial activation belong to primary haemostasis and are not captured by INR.
- An abnormal synthetic marker can remain prognostically useful while being a poor predictor of procedural bleeding.
05BleedingPressure, vessel injury and clot formation are different failure pointsNuanced
A varix can bleed because portal pressure and local vessel properties create a mechanical problem. A procedure can injure a vessel despite preserved clot formation. Haemostatic deficits can compound either event, but correcting a plasma test does not directly remove these causes.
- First establish the bleeding source and context, then assess whether a specific haemostatic deficit contributes.
- Whole-blood viscoelastic assays describe clot kinetics, strength and lysis under their test conditions; they do not reproduce every vascular or endothelial process.
- Rebalanced haemostasis is a description of competing changes, not a guarantee of stability or a numerical bleeding/thrombosis score.
Interventions
What are we trying to change?
Therapy makes more sense when its physiological target is explicit. These are mechanism summaries, not prescribing guidance.
TargetActual bleeding mechanism
Treat the source
Endoscopic, surgical or interventional control addresses portal-pressure or procedural bleeding directly.
TargetSpecific haemostatic deficit
Targeted blood products
Fibrinogen, platelets or other components can be replaced when a clinically relevant deficiency is present.
TargetVTE risk
Thromboprophylaxis
May still be appropriate when standard clinical indications exist.
Apply the mechanism.
What not to conclude.
- Equating INR with bleeding risk.
- Assuming thrombocytopenia means no thrombotic risk.
- Giving plasma solely to normalise a laboratory number.
- Calling cirrhotic fibrinolysis uniformly hyperfibrinolytic.
Four things to keep.
- Pro- and anticoagulant systems fall together.
- VWF/ADAMTS13 biology partly compensates for thrombocytopenia.
- Routine plasma tests show only one side of the balance.
- The system is rebalanced but fragile — bleeding and thrombosis are both possible.
Evidence & current guidance
Where the physiology meets current practice.
Clinical criteria and treatment recommendations change faster than core physiology. This layer keeps the module tied to current consensus and primary evidence.
Cirrhosis is not auto-anticoagulation — it is a fragile rebalance
Advanced liver disease reduces several procoagulant factors, but it also reduces natural anticoagulants such as protein C, protein S and antithrombin. At the same time, platelet number falls while endothelial von Willebrand factor rises and ADAMTS13 activity falls. The result is not normal haemostasis, but a new, unstable balance that can tip toward either bleeding or thrombosis.
- PT/INR captures reduction in selected procoagulant factors but largely ignores the parallel fall in anticoagulant pathways.
- Thrombin-generation assays that include thrombomodulin can show preserved or increased thrombin potential despite a prolonged INR.
- Rebalanced does not mean protected: infection, AKI, ACLF, endothelial injury, procedures and stasis can destabilise the system.
- Bleeding in cirrhosis is often driven by portal pressure or mechanical injury rather than failure of plasma coagulation alone.
Thrombocytopenia is only half of the platelet story
Platelet counts commonly fall because of splenic sequestration, reduced thrombopoietin and marrow effects, but primary haemostasis is partly supported by marked increases in von Willebrand factor and reduced ADAMTS13 activity. Platelet count therefore does not map directly onto bleeding tendency.
- High VWF supports platelet adhesion under shear and can partly compensate for low platelet number.
- Platelet function is heterogeneous and becomes more vulnerable during sepsis, renal failure and severe decompensation.
- A platelet count above 50 × 10⁹/L should not routinely trigger prophylactic correction before procedures solely because cirrhosis is present.
- For high-risk procedures with counts between 20 and 50 × 10⁹/L, EASL supports case-by-case rather than routine correction.
A prolonged INR does not predict procedural bleeding — and FFP is not an INR antidote
Routine PT/INR and APTT are poor predictors of post-procedural bleeding in cirrhosis because they interrogate platelet-poor plasma and omit important anticoagulant, platelet and endothelial biology. Contemporary guidance therefore discourages prophylactic correction of INR with fresh frozen plasma before invasive procedures.
- EASL explicitly states that INR and APTT do not predict post-procedural bleeding in cirrhosis.
- FFP frequently fails to normalise the INR and adds substantial volume, which can worsen portal pressure, ascites or pulmonary congestion.
- Routine correction of fibrinogen deficiency or thrombocytopenia before procedures is also not supported outside selected high-risk contexts.
- The procedural plan, local haemostatic control, disease severity, AKI, infection and technical factors often matter more than a single coagulation result.
TEG/ROTEM can guide transfusion — but they are not universal bleeding predictors
Viscoelastic assays provide a whole-blood view of clot initiation, propagation, strength and lysis. They can reduce unnecessary blood-product exposure in selected procedural, surgical and bleeding settings, but they do not fully reproduce endothelial biology and should not be treated as a definitive test of in-vivo bleeding risk.
- Randomised studies in cirrhosis show reduced prophylactic blood-product use with VET-guided strategies without a clear increase in bleeding.
- EASL 2022 did not recommend routine VET solely to predict procedural bleeding risk.
- EASL 2025 surgical guidance supports VET to guide blood-product administration during active bleeding and high-risk operative settings.
- A 'normal' TEG or ROTEM does not mean the haemostatic system is physiologically normal; it means the measured whole-blood clot properties are relatively preserved under that assay's conditions.
An abnormal INR does not protect against VTE or portal vein thrombosis
Patients with cirrhosis remain susceptible to venous thromboembolism and splanchnic thrombosis. The fall in anticoagulant proteins, endothelial activation, inflammation and stasis all contribute. Thrombocytopenia and prolonged INR should therefore not be used as automatic reasons to withhold pharmacological thromboprophylaxis.
- ISTH guidance recommends against using thrombocytopenia or prolonged PT/INR as absolute contraindications to thromboprophylaxis.
- Hospital thromboprophylaxis should generally follow local protocols, with individual assessment in critically ill patients and especially in ACLF.
- Portal vein thrombosis is a clinically important thrombotic phenotype, not evidence that the rebalanced model has 'failed'.
- Active bleeding, severe thrombocytopenia, procedures and organ failure still require individual risk–benefit assessment.
Cirrhosis is not uniformly hyperfibrinolytic
Both pro- and antifibrinolytic proteins change in cirrhosis. Plasminogen may fall, but inhibitors and regulatory proteins also change; endothelial release of tissue plasminogen activator and acute inflammatory states further modify the phenotype. The net fibrinolytic state therefore varies between patients and across phases of illness.
- Hyperfibrinolysis can occur, particularly in advanced disease or major bleeding, but it should not be assumed from cirrhosis alone.
- Conversely, hypofibrinolytic states may contribute to thrombosis in some patients.
- The direction of fibrinolytic imbalance can change during ACLF, sepsis, surgery and transplantation.
- Treatment should be driven by a compatible clinical and laboratory pattern rather than by the diagnostic label of cirrhosis itself.
Key sourcesPrimary guidance and landmark evidence
Reviewed 12 September 2026 · next scheduled review September 2027EASL Clinical Practice Guidelines on prevention and management of bleeding and thrombosis in patients with cirrhosisEuropean Association for the Study of the Liver. J Hepatol. 2022;76:1151–1184. doi:10.1016/j.jhep.2021.09.003.
Core contemporary European guidance on procedural bleeding, transfusion, thrombosis and the limitations of routine coagulation tests in cirrhosis.
View source ↗The concept of rebalanced hemostasis in patients with liver diseaseLisman T, et al. J Thromb Haemost. 2021;19:1116–1122. doi:10.1111/jth.15239.ISTH SSC communication describing the mechanistic basis of rebalanced haemostasis and why routine tests misrepresent net haemostatic capacity.
View source ↗Thromboprophylaxis for venous thromboembolism prevention in hospitalized patients with cirrhosisRoberts LN, et al. J Thromb Haemost. 2022;20:2237–2245. doi:10.1111/jth.15829.ISTH SSC guidance supporting pharmacological thromboprophylaxis according to clinical risk rather than excluding patients solely because of thrombocytopenia or prolonged INR.
View source ↗Vascular Liver Disorders, Portal Vein Thrombosis, and Procedural Bleeding in Patients With Liver DiseaseNorthup PG, et al. Hepatology. 2021;73:366–413. doi:10.1002/hep.31646.AASLD practice guidance on procedural bleeding, portal vein thrombosis and vascular complications in liver disease.
View source ↗EASL Clinical Practice Guidelines on extrahepatic abdominal surgery in patients with cirrhosis and advanced chronic liver diseaseEuropean Association for the Study of the Liver. J Hepatol. 2025;83:768–789. doi:10.1016/j.jhep.2025.04.008.Recent perioperative guidance reinforcing that abnormal coagulation tests should not be systematically corrected and supporting VET-guided transfusion during active bleeding/high-risk surgery.
View source ↗Review policy: each module is reviewed at least annually and sooner after a major guideline, consensus statement or practice-changing study. A review date indicates editorial review, not that every linked source is open access.