Does balanced haemostasis equate to normal coagulation in patients with acute liver failure?
2014; Wiley; Volume: 34; Issue: 5 Linguagem: Inglês
10.1111/liv.12477
ISSN1478-3231
Autores Tópico(s)Blood Coagulation and Thrombosis Mechanisms
ResumoPatients with liver disease whether acute or chronic suffer from a coagulopathy and this, in the past, has been associated with a bleeding phenotype. Laboratory investigations typically reveal a high Prothrombin Time (PT)/International Normalized Ratio (INR) and Activated Partial Thromboplastin Time (APTT) and sometimes a low fibrinogen level. The high PT was noted since the inception of the test by Armand Quick 1. It was later realized that this test was exquisitely sensitive to low factor VII levels. The PT/INR became established tests to monitor coumarin anticoagulant therapy and as high PT/INR values correlate with an increased risk of bleeding in patients on similar drugs, the raised PT in patients with liver disease was considered to indicate a state of ‘autoanticoagulation’. This notion has been challenged over the past decade. Interest in the coagulopathy of liver disease started accruing momentum in 2005 when Tripodi et al. showed that using a novel assay of coagulation, the thrombin generation test (TGT), people with cirrhotic liver disease generated similar amounts of thrombin as a normal control population 2. One has to point out that the normal values came after the thrombin receptor, thrombomodulin, was added to the assay which makes the system sensitive to some of the natural anticoagulants. In vivo, thrombin binds to the endothelial cell surface receptor thrombomodulin. This interaction neutralises thrombin's procoagulant activity and allows its active site to target Protein C (PC) and activate the latter to Activated Protein C (APC). APC is a very potent anticoagulant which targets the clotting cascade at least in two crucial points, by inactivating both the propagation and amplification loops through its activity on activated Factor V (FVa) and activated Factor VIII (FVIIIa) respectively. Tripodi's work which has since been replicated by other groups 3, demonstrated that as the levels of PC and Protein S (PS) are reduced in liver disease, there is less APC production and less inactivation of coagulation. This has been interpreted as ‘balanced coagulation’ as we have a reduction in both the pro-coagulants such as FI, FII, FV, FVII, FIX, FX and FXI and the anticoagulants namely PC, PS and also Protein Z and antithrombin (AT). Not only that, but Factor VIII and Von Willebrand Factor are almost invariably increased in these patients. As a natural progression in research, groups started looking at a different cohort of patients, those with acute liver failure (ALF). Lisman et al. and Agarwal et al. both showed that the coagulopathy in these cases is similar to that in patients with cirrhosis with some differences 4, 5. Patients with ALF tend to have higher INR levels. In these patients, the unmodified TG assay shows that the patient group does not generate as much thrombin as the normal cohorts. This has been replicated by Habib et al. in this edition of the Journal 6. What has not been replicated is the increased velocity of thrombin generation in the patients with ALF for unclear reasons. Again, it has been shown that the patients are highly resistant to APC after an empiric concentration of soluble thrombomodulin is added to the test. The patients also generate as much thrombin as their matched normal counterparts after the inclusion of this thrombin receptor. Does this mean that patients with ALF have a normal coagulation potential resulting in a normal bleeding risk when exposed to haemostatic challenges? Does this mean that clinicians do not need to try to correct the ‘coagulopathy’ of liver disease? We could try to answer these questions by drawing comparison with coumarin anticoagulation. The INR in patients on warfarin correlates strongly with TGT 7. It shows a parabolic curve with increasing INRs with a plateau being reached at INR of around 4.0 7, 8. The maximum amount of thrombin generated in the TGT or the endogenous thrombin potential (ETP) goes down to 28.2% of normal in patients with INRs in the therapeutic range of 2.4–2.6 and goes down to 22% in those with INRs of 2.9–3.0. In the ALF cohort presented in the Journal, the median INR was 3.36 with an interquartile range of 2.67–7.01. So, allowing for differences in the PT reagent (STA-Neoplastine) used in this study which has an ISI of ~1.3 and would be expected to yield higher INR values than the human recombinant TF reagents such as Innovin and Recombiplastin (ISI ~1.0), Habib et al.'s ALF patients should have shown an ETP of less than 22% of normal. Instead, dividing the mean patient ETP by the mean normal ETP gives a value of 55.5% normal. This is high, despite these patients having FII, FVII and FV levels <23% as one would expect at such high INRs similar to patients on warfarin 9. Intriguingly, it seems that ALF patients can still generate significant amounts of thrombin possibly allowing for efficient secondary haemostasis to take place. Why should this happen as clotting factor levels are actually similarly low in patients with ALF as in patients with similar INRs on warfarin? The authors and other groups give us one possibility i.e that these patients have FVIII levels which are much higher than normal and are probably contributing to the generation of greater amounts of thrombin through the amplification loop of the coagulation cascade. However, again turning our attention to the warfarin model, patients on warfarin commonly also have elevated FVIII levels 7 albeit lower than the levels reported in this study. This suggests that there are other stronger factors at play including a low AT, procoagulant microparticles that are clearly increased in ALF 5 and associated with clinical outcomes 10 and possibly tissue factor produced by damaged hepatocytes 11. Both tissue factor and phospholipids greatly affect the thrombin generation assay while at low levels do not impact on the routine tests such as the PT or the APTT 12. In addition, as expected, this group demonstrated resistance to APC. However, this should be interpreted with caution and certainly not in isolation as even patients on warfarin show a similar APC resistance pattern because of a decrease in γ-carboxylation of PC and PS 13 but this does not lead to thrombosis as the effect on FII, FVII, FIX and FX is much greater and instead results in anticoagulation. The activation of PC by this mechanism is a regulatory phenomenon mainly aimed at preventing clot propagation on healthy endothelium rather than an immediate effect to dampen initial thrombin formation. What this could mean is that these patients like those with chronic cirrhosis could paradoxically also be at relatively high risk of thrombotic complications 14. Another interesting finding in Habib's study is the fact that the group showed weak to moderate negative correlations of thrombin generation parameters with the INR. This is in contrast to the study by Agarwal et al. where no correlation between the two tests was detected 5. This analysis is not documented in the other study in ALF 4. These differences could be explained by the small numbers of patients in these studies which are logistically very difficult to conduct. However, it could also be the result of the different thromboplastin reagents used in the two studies as explained before. It is well known that different PT reagents have different sensitivity to plasma from patients with liver disease 15, 16 and that those reagents with a higher ISI lead to relatively higher INR values. This is the reason why the Liver ISI was proposed, in order to try to harmonize similar differences 17. While acknowledging that this proposal has been verified only in patients with chronic cirrhosis and not patients with ALF, it could be the case that some thromboplastins might more accurately reflect coagulation in patients with liver disease. This hypothesis still needs to be confirmed. However, the negative correlation of the INR with thrombin generation parameters such as the ETP demonstrated in Habib et al.'s study could mean that the higher the INR the more hypocoagulable these ALF patients become. This would mean that certain patients would actually require correction of their coagulopathy prior to certain invasive procedures. This paper coming from the King's College Hospital, London group is very important in that it not only provides us with more insight on a complex issue, but also raises more questions, thus opening more avenues for further research in this field. Is the time ripe for clinical trials using thrombin generation as a diagnostic tool for identifying patients at risk of bleeding and to monitor coagulation factor replacement therapy to cover invasive procedures in patients with ALF, instead of the routine coagulation tests currently performed? Certainly, it is still early days before we can make major changes to our guidelines on this topic. Conflict of interest: The authors do not have any disclosures to report.
Referência(s)