Thromboprophylaxis with heparin in hospitalized patients with cirrhosis: friend or foe

The paradigm that the haemostatic disturbances observed in patients with severe chronic liver disorders protect them from thrombotic events has completely changed in the last years 1, 2. In fact, despite bleeding is the most common clinical manifestation of advanced cirrhosis, thrombotic complications can be also observed. It has been recently demonstrated that patients with cirrhosis are at risk of developing not only thrombotic complications in the portal venous axis 3 but also in other territories 4, 5. In particular, the risk of developing a DVT/PE has been shown to range from 0.5 to 1.87% 4, 6-9, prevalence is not negligible but lower than that observed in patients with other comorbidities such as cancer or cardiorespiratory disease 10.

The increased rate of thrombotic events can be owing to the observed hypercoagulative state that is present in patients with cirrhosis. Indeed, patients with cirrhosis have a new unstable haemostatic balance characterized by a reduction in the hepatic procoagulant factors with a concomitant reduction in the hepatic anticoagulant factors together with an increase in endothelial procoagulant factors such as Factor VIII or vW Factor 2, 11. This weak new balance may be easily broken by infections, renal failure among other frequent situations observed in hospitalized patients with cirrhosis. In addition, patients with cirrhosis may also have inherited prothrombotic conditions such as factor V Leiden deficiency or prothrombin G20210A mutations that may further contribute to develop thrombotic events 12, 13.

This prothrombotic background associated to the observed increased rate of thrombotic events has lead to consider the prophylaxis of thrombosis in patients with cirrhosis when they are hospitalized. However, all studies in hospitalized patients evaluating the benefit of thromboprophylaxis in preventing thrombotic events systemically excluded patients with severe liver disorders. Therefore, it is unknown whether medical thromboprophylaxis is able to effectively prevent thrombosis in this specific population of patients and therefore if the benefit/risk ratio (prevention of thrombosis/bleeding risk) of medical thromboprophylaxis, which has been clearly demonstrated in other conditions, is also true for patients with cirrhosis. As a consequence, and despite hospitalized patients with cirrhosis frequently remain immobile because of the presence of severe ascites, infections or hepatic encephalopathy, current guidelines do not make formal recommendations on thromboprophylaxis in these patients 14-16. Even if several retrospective studies described thrombotic complications in cirrhotic patients and encouraged the use of anticoagulation to treat and/or prevent thrombotic complication 9, 17-19, up to now there are no prospective studies evaluating thromboprophylaxis and bleeding risk in these patients and so it is still an important matter of discussion. An additional concern that has risen when considering the use of heparin in patients with cirrhosis is the efficacy of this treatment. It has been suggested that because heparin require binding to endogenous antithrombin to exert their anticoagulant action and reduced levels of antithrombin is a typical feature of patients with cirrhosis 20, it may be less active in these patients. There are several in vitro studies giving conflicting results about the efficacy 21, 22 of LMWH in patients with cirrhosis. In fact, Bechmann et al. reported that in their series of cirrhotic patients that underwent thromboprophylaxis, only 14% of them achieved sufficient anti-Xa levels; interestingly any patient developed DVT. On the other hand, recent studies have shown that plasma from cirrhotic patients is more responsive to the anticoagulant effect of LMWH 23. Anyhow, a recent RCT has demonstrated that fixed doses of the LMWH enoxaparin is able to prevent PVT in cirrhosis 19 strongly suggesting that LMWHs are indeed effective in patients with cirrhosis.

The article in this issue of Liver international by Intagliata et al. 24 retrospectively evaluated the incidence of complications in a large series of hospitalized patients with cirrhosis (235 patients accounting 355 episodes of hospitalization) submitted for at least 2 days to thromboprophylaxis with heparin (mean duration of treatment was almost 5 days). Other 903 patients with cirrhosis hospitalized during the same period of time did not receive thromboprophylaxis. In the study, 1.4% of enrolled patients experienced an episode of VTE while on thromboprophylaxis. Unfortunately, the authors of the study do not provide the incidence of thrombosis of those patients with cirrhosis hospitalized during the same period of time and not receiving prophylactic anticoagulation to estimate the actual effect of heparins in preventing thrombotic events.

The study also evaluates the other side of the balance, the bleeding risk associated with the use of anticoagulation. Almost 4% of patients and 2.5% of episodes of treatments had an episode of gastrointestinal bleeding (GI) during hospitalization. It is important to recognize that spontaneous GI bleeding is a frequent event in hospitalized patients with cirrhosis. Unfortunately, the authors do not evaluate the incidence of GI bleeding in those hospitalized patients in their unit not receiving anticoagulation during the same period of time. Thus, it is difficult to estimate whether the risk was increased or not by treatment. Interestingly, a RCT in a more stable population of patients with cirrhosis treated with enoxaparin to prevent PVT development, the bleeding risk was similar in patients receiving or not receiving enoxaparin 19. A meta-analysis performed in non surgical patients has demonstrated that LMWH treatment is associated with less bleeding risk and need for transfusion compared to UFH-treated patients 25 In the study by Intagliata et al. 24, most patients received UFH and this may explain, at least in part, the relatively high incidence of GI bleeding, in this cohort of patients

In our view, the present study highlights the problem, until now almost neglected, of the risk of thrombosis in hospitalized patients with cirrhosis. Unfortunately, the study does not answer the open questions about the potential benefit/risk ratio of thromboprophylaxis in hospitalized patients with cirrhosis. Randomized double-blind placebo studies are needed to answer this question and to establish medical pharmacological strategies in this complex population of patients.