Anticoagulation for portal vein thrombosis in cirrhosis

Watch a video presentation of this article Thrombosis of the splanchnic venous system, especially the portal vein (PV), is an increasingly recognized phenomenon in patients with cirrhosis that has garnered substantial clinical interest. With improvement in diagnostic radiologic technology and increasing awareness among clinicians, this diagnosis is routinely encountered. Substantial evidence now shows that alterations in procoagulant and anticoagulant properties of the hemostatic system in cirrhosis are rebalanced and can be perturbed to produce either hemorrhage or thrombosis. The portal venous system in cirrhosis represents a local environment particularly prone to thrombosis formation by virtue of reduced blood flow from portal hypertension and the inflammatory milieu secondary to hepatic injury and possibly even to gut translocation of bacteria or their by-products. The decision to treat portal vein thrombosis (PVT) is often determined by the extent of thrombosis (partial or complete), the presence of symptoms, the patient's transplant status, and the timing of the event (acute versus chronic)1 (Table 1). A prospective randomized, but non–placebo-controlled trial demonstrated that anticoagulation therapy (AT) with enoxaparin prevents development of PVT and reduces hepatic decompensation and mortality.2 However, studies examining the treatment of PVT vary tremendously in design and patient population. Although guidelines3 support AT for progressive PVT in transplant-eligible patients, a large study from France indicates that PVT may not represent a risk itself, but instead may only signify advancing portal hypertension and liver disease.4 In this article, we will review AT for PVT in cirrhosis with a particular focus on safety and efficacy. Low molecular weight heparin (LMWH) and vitamin K antagonist (VKA) are the most extensively studied agents in the cirrhosis population5-11 (Table 2). Drawbacks include parenteral administration for LMWH and challenges in monitoring and dietary adherence for VKA. Consequently, interest in direct oral anticoagulants (DOAC) is increasing, but clinical data in cirrhosis are limited to case reports12, 13 and one larger retrospective cohort study.14 LMWH VKA 13.2% (22/167) Major bleeding events Complete/partial response (n = 51) No response (n = 14) 23.5% in daily (1.5 mg/kg) versus 6.4% in bid (1 mg/kg) No variceal bleeding Complete response (n = 11) Partial response (n = 12) No response (n = 5) LMWH VKA Complete response (n = 25) Partial response (n = 8) No response (n = 22) 9% (5/55) Non-EV (2 gastrointestinal, 3 other) 6 EV Complete response (n = 12) Partial response (n = 9) No response (n = 12) 8.6% (3/35) non-EV (1 ICH) EV: 1 exposed, 5 nonexposed Complete response (n = 21) Partial response (n = 2) No response (n = 5) 7.1% (2/28) 2 UGI LMWH VKA Complete/partial response (n = 8) No response (n = 11) 5.2% (1/19) 1 UGI A large retrospective study from Spain examined selected patients with cirrhosis with PVT treated with LMWH or VKA and found a 60% partial or complete recanalization rate.8 This study and others have demonstrated that early initiation of AT (before 6 months) results in a greater likelihood of success.8, 10 Both lower and higher recanalization rates have been reported in other studies, underscoring the wide variation in patient populations, study endpoints, and response definitions between studies. The most feared consequence of AT is bleeding. In a recent prospective, multicenter study from Italy examining both cirrhotic and noncirrhotic PVT, the cirrhosis cohort (n = 167) was the most likely to bleed; however, the majority of patients who bled did not receive AT.5 In a study from China examining LMWH (enoxaparin) in once-daily (1.5 mg/kg) versus twice-daily (1 mg/kg) dosing, investigators demonstrated that patients receiving higher once-daily dosing demonstrated more risk for bleeding compared with standard twice-daily dosing.7 Delgado et al.8 noted that all “anticoagulation-related bleeding” events were treated with VKA as compared with patients treated with LMWH. Although it is challenging to compare studies in terms of bleeding risk, the majority of studies indicate that AT for PVT appears relatively safe. A recent study examined upper gastrointestinal (UGI) bleeding in patients with cirrhosis receiving AT compared with a similar cohort who did not receive AT.15 Five-day endoscopic treatment failure was the same between these two groups, indicating that AT per se does not change the outcome of UGI bleeding, at least in the early timeframe. In carefully selected patients, AT appears to be generally safe and effective in this population (Table 3). Certain categories of patients with cirrhosis and PVT that likely benefit from AT include patients with acute PVT who are candidates for liver transplantation (LT) or patients with progressive PVT extending into superior mesenteric vein (SMV). In patients listed for transplantation, AT in patients with cirrhosis and PVT is given with the aim of achieving at least partial recanalization of the PV before surgery to restore portal flow to the graft by conventional end-to end anastomosis. Englesbe et al.16 analyzed 22,291 LT recipients wherein the prevalence rate of PVT was 4.02%. Patients with PVT had significantly higher post-LT mortality (hazard ratio 1.32, P = 0.02). PVT at LT is significantly associated with higher transfusion requirements and postoperative complications.1 Notably, only PVT involving more than 50% of the lumen is associated with higher long-term mortality. Similar findings were found in a systematic review analyzing 25,753 LTs in patients with cirrhosis in which 2004 LTs were performed in patients with PVT.17 There are no definitive markers identifying which patients with cirrhosis with SMV thrombosis develop bowel infarction. Mortality in acute SMV thrombosis ranges from 20% to 50% depending on age, comorbidities, timing of diagnosis, and surgical interventions. In our experience, it is a very uncommon complication of cirrhotic PVT, but imaging should include this vascular territory at diagnosis. Because it can influence the outcome of therapeutic intervention, it is important to try to distinguish between acute and chronic PVT. The clinical context, diagnosis coincidence with symptoms such as abdominal pain, ascites, hydrothorax, or worsening portal hypertension, and certain imaging characteristics can help to determine the chronicity of PVT in many circumstances.18 The presence of cavernous transformation (chronic PVT) is usually characterized by numerous small collateral vessels in the place of a normal PV. Most studies examining the effect of AT include patients with acute or subacute PVT and demonstrate that successful recanalization is highly dependent on time from diagnosis to therapy (less than 6 months). The presence of a cavernoma indicates a longstanding PVT that is unlikely to recanalize completely with AT. In certain situations with chronic PVT and portal hypertension, transjugular intrahepatic portosystemic shunt (TIPS) or other vascular approaches may be more useful. Whether anticoagulant therapy has a net benefit even in the chronic setting remains to be determined in prospective trials. Based on existing data we suggest considering AT in patients with occlusive PVT involving the trunk or both branches of PV in candidates for LT and in patients with nonocclusive or occlusive SMV thrombosis without direct contraindication (Fig. 1). Because data are heterogeneous and randomized prospective controlled trials are lacking, individualization of therapy is a necessity and other factors such as presence of hypercoagulable states or symptoms related to PVT in patients may make treatment more desirable in certain cases. We recommend endoscopic screening for varices and starting or optimizing prophylaxis of bleeding with nonselective beta-blockers or endoscopic band ligation in patients with large esophageal and/or gastric varices before starting AT.19, 20 TIPS may be indicated in patients developing thrombosis progression on AT or complications related to AT such as major bleeding events. The duration of therapy for PVT in cirrhosis is currently undefined, but we recommend at least 6 months of therapy. We consider indefinite anticoagulation in patients listed for transplant or who display other concurrent indications for AT. Is the risk for extension of cirrhotic PVT into the SMV with subsequent mesenteric ischemia and possible bowel infarction an established indication for anticoagulation? This is not resolved. In our collective experience this appears to be relatively uncommon in cirrhosis-related PVT (perhaps because of the antecedent development of the collateral bed), but it is a potentially serious and possibly devastating complication. In general, this would favor more aggressive anticoagulant therapy. However, as with much in this field, the relative risk versus benefit remains untested in prospective and randomized trials. Therapy must be decided by experience and clinical findings on a case-by-case basis.