Patients with chronic liver disease variably exhibit abnormalities of coagulation related to decreased synthesis of coagulation proteins. This results in prolongation of routine coagulation tests such as prothrombin time (PT) and activated partial thromboplastin time (APTT). In addition, patients suffer from thombocytopenia of variable severity and the mechanisms underlying this decreased platelet count are multifactorial, including decreased production of thrombopoietin (TPO). These abnormalities of coagulation are often equated with an excess risk of bleeding and patients are often treated in the context of active bleeding, or peri-procedurally in the presence of abnormalities. Commonly used haemostatic replacement therapies include fresh frozen plasma (FFP), prothrombin complex concentrates (PCC), cryoprecipitate, and/or platelets. The antifibrinolytic agent tranexamic acid along with TPO receptor agonists have been used in restricted clinical circumstances, the latter to increase platelet counts.
Practicing clinicians are faced with two questions in the management of this coagulopathy: the first is when to treat, and the second is what to use. The evidence base for products that are available for use is less contentious. FFP has been used for many decades in this context and multiple studies have demonstrated that the response to treatment is related to volume, i.e. a higher dose of FFP results in a better correction of PT with higher factor levels.1 Similarly, PCC have been used with complete correction of the international normalised ratio seen in the majority of the patients, unlike the situation seen with the use of FFP.2 PCC have been demonstrated to be efficacious and there does not appear to be an obvious excess of thrombotic events. Platelet transfusions are commonly prescribed for bleeding patients and peri-procedurally in the context of moderate-to-severe thombocytopenia. There is ongoing debate regarding transfusion thresholds and values used in liver disease, as they are based on evidence gathered from other areas, principally oncology patients undergoing intense chemotherapy and stem cell transplantation. TPO agonists have also been used peri-procedurally in the context of clinical trials and one of them (eltrombopag) is licensed for use in the context of interferon- based treatment for hepatitis C infection, where thrombocytopenia limits treatment options. TPO agonists have been associated with an unexpectedly high risk of thrombosis, in particular portal vein thrombosis, and caution needs to be exercised when used in routine clinical practice.3
Whilst the effects of the various products are well documented, there is much debate concerning the appropriate transfusion thresholds. Laboratory and clinical data show dysregulated coagulation with an increased risk of thrombosis; this is concurrently associated with a diminished reserve and a potential for excess risk of bleeding.4-6 Conventionally, the PT, APTT, fibrinogen, and platelet count have been used to define transfusion thresholds. There is increasing evidence that use of viscoelastic tests of whole blood (TEG®, ROTEM®) may decrease the number of patients requiring treatment.7,8 A recent study evaluating two different tests for transfusion demonstrated markedly decreased requirements for transfusion when a TEG-based strategy was instituted. Prospective studies with PT as an endpoint demonstrate that PCC are more effective than FFP, and with the latter the effectiveness is related to the volume infused with sub- therapeutic doses being in common usage.1,2 Thresholds for platelet transfusion are similarly variable and in a prospective study of eltrombopag and platelet transfusion that aimed for a target of 50,000 per mm3, bleeding complications were seen in approximately one-fifth of the patients.3
The dilemma relates to appropriate transfusion thresholds. Pragmatically, this is dictated by individual clinicians’ perception of risk versus benefit in a given patient for a given procedure, and indeed a national audit of transfusion practices in cirrhosis in the UK reflects this anxiety and ambiguity.9 There is an urgent need for observational studies correlating abnormal coagulation parameters to bleeding outcomes and interventional studies to redefine the transfusion thresholds for procedures and haemostatic products.