Safety and Efficacy of Anticoagulation in Patients with Cirrhosis: A Meta-Analysis

Background and Aims Portal vein thrombosis is a serious adverse event that occurs during liver cirrhosis. We performed a meta-analysis to evaluate the safety and efficacy of anticoagulant therapy and prophylactic anticoagulant therapy in cirrhosis patients with (/without) portal vein thrombosis. Methods Eligible comparative studies were identified by searching the following electronic databases: PubMed, Embase, Cochrane Library, Web of Science, and CNKI. A meta-analysis was performed to calculate odds ratios and 95% confidence intervals using fixed-effects models. Recanalization and thrombus progression were defined as the primary outcomes. Secondary outcomes included adverse events and death mortality. Results A total of 3479 patients were included in this analysis. Compared with the control group, the recanalization rate in the anticoagulant therapy group was increased (P < 0.00001) in patients with cirrhosis and portal vein thrombosis without increasing adverse events. Multiple use of enoxaparin in small doses is safer than single large doses (P=0.004). Direct oral anticoagulants are more effective (P < 0.00001) and safer than traditional anticoagulants. Prophylactic anticoagulant therapy can effectively prevent portal vein thrombosis formation (P < 0.00001). Conclusions Anticoagulation therapy can treat or prevent portal vein thrombosis in patients with liver cirrhosis and is a relatively safe treatment.


Introduction
Portal vein thrombosis (PVT) is a common adverse event of liver cirrhosis, and its incidence increases as liver disease progresses and is even higher in patients with various portal hypertension procedures [1][2][3]. Patients with acute and severe PVT may experience symptoms, such as fever, abdominal pain, ascites, and splenomegaly, but many patients do not exhibit symptoms in the early stage of onset [4]. e hidden onset of PVT can cause significant harms to patients, including intestinal congestion and necrosis, secondary serious infections, increased risk of bleeding from esophageal varices rupture of the stomach, increased decompensation of the liver, more intraoperative and postoperative adverse events, and higher mortality [5,6]. erefore, to improve patient prognosis, timely and effective treatments of portal PVT are very important. As one of the main treatments of PVT, anticoagulation has received increasing attention in recent years, and prophylactic anticoagulation has even been proposed for patients at high risk of PVT. However, no definitive conclusion on anticoagulation effectiveness and safety has been reported. Some studies found that the recanalization rate of PVT after anticoagulation treatment is greater than 80% [7,8]. However, other studies showed that anticoagulation treatment might be ineffective for PVT [9,10]. erefore, it is necessary to analyze relevant previous studies.
is article is divided into two parts, namely, anticoagulation and prophylactic anticoagulation therapy, and both topics are analyzed using and metaanalysis to provide a reference for clinicians to treat or prevent PVT in patients with cirrhosis.
is study included papers published up to December 2019.

Inclusion
Criteria. ① RCT or nRCT; ② study subjects were patients older than 18 years of age with liver cirrhosis at any stage attributed to various etiologies, and there were no restrictions on the race, nationality, or region; ③ the observation group was administered anticoagulants for anticoagulation, and the control group was treated with placebo or blank control, different anticoagulants, or different doses or treatment times with the same anticoagulant; ④ data reported should include these outcome indicators: portal vein recanalization or new onset, bleeding events, death, and other adverse events, including the new onset of decompensation of liver function, ascites, spontaneous peritonitis, sepsis, hepatorenal syndrome, or hepatic encephalopathy.

Exclusion Criteria. ① Nonclinical research;
② studies for which a full text is not available; ③ republished literatures; ④ studies that do not provide complete data; ⑤ research subjects are noncirrhotic patients; ⑥ subjects have an underlying primary blood disease, membranous obstruction of the inferior vena cava, or preexisting extrahepatic thrombosis; ⑦ interventions other than anticoagulation; ⑧ research that is not germane to our subject.

Screening and Quality Evaluation.
After reading the titles and abstracts of all the retrieved studies, preliminary screening was performed. e full text of the documents that passed the preliminary screening was read to exclude documents that clearly do not meet the requirements or are duplicate studies. e Cochrane bias risk assessment tool was used to assess the bias risk of included RCTs, and the Newcastle-Ottawa Scale (NOS) was used to assess the quality of included nRCTs.

Data Extraction and Statistical
Analysis. Data extracted from each study included the following: first author, year of publication, country of publication, number of patients, liver function score, specific interventions, overall follow-up time, portal vein recanalization or new occurrence, bleeding events, other adverse events, and death.

Results
A total of 403 articles passed the preliminary screening, and 302 were excluded due to noncompliance of the study subjects or the use of intervention methods other than anticoagulation. In addition, 29 were nonclinical studies, and the full text of 16 articles could not be obtained. Moreover, 20 articles did not meet the requirements. us, thirty-six papers [2,9, were ultimately selected to complete this meta-analysis ( Figure 1).

Basic Characteristics of Included Literatures.
Of the 36 selected papers, 21 were reported in English, and 15 were in Chinese. Of the selected papers, 11 studies reported RCTs, and 25 reported nRCTs. e study sites included China, the United States, Europe, Japan, and other places. e dates of publication ranged from 2005 to 2019, and a total of 3479 patients were included. e basic characteristics of the included studies are provided in Table 1.

Bias Risk Assessment.
Cochrane bias risk assessment tool and NOS scale were selected for evaluation, as shown in Figures 2(a) and 2(b) and Table 2. Figure 3(a) shows that the PVT recanalization rate in the observation group (anticoagulation) is increased compared with the control group, and the results are statistically significant (OR � 5.10, 95% CI: 3.93∼6.61, P < 0.00001). Subgroup analysis based on different drugs (other represents other anticoagulants, heparin, and/or warfarin combined with others) (Figure 3(b)) more specifically shows that different anticoagulants have therapeutic effects on PVT. Figure 3 C shows that the thrombus progression or new thrombus formation in the observation group was reduced compared with the control group (OR � 0.22, 95% CI: 0.13∼0.37, P < 0.00001). Compared with the control group, anticoagulation did not increase the incidence of bleeding events (OR � 0.70, 95% CI: 0.49∼1.02, P � 0.06) or the incidence of other adverse events (OR � 0.62, 95% CI: 0.37∼1.02, P � 0.06), but the mortality rate was reduced (OR � 0.25, 95% CI: 0.08 ∼ 0.81, P � 0.02) (Figures 2(d)-2(f )) (see Supplementary Figure 1 for histogram).

Effect of Anticoagulant erapy with Different Enoxaparin Doses.
When different doses of enoxaparin were used for anticoagulation, the same effects were noted in the observation group (1.0 mg/kg q 12 h) and the control group (1.5 mg/kg q 24 h) of patients with liver cirrhosis and PVT (OR � 1.03, 95% CI: 0.47∼2.27, P � 0.94) (Figure 4(a)), but the incidence of bleeding events was reduced in the former (OR � 0.24, 95% CI: 0.09∼0.62, P � 0.004) (Figure 4(b)). No significant difference in the incidence of other adverse events was between the two groups (OR � 1.43, 95% CI: 0.67∼3.08, P � 0.36) (Figure 4(c)) (see Supplementary Figure 2 for histogram).

e Effect of Prophylactic
Anticoagulation with Different Drugs. During preventive anticoagulation, the rate of thrombosis formation did not differ in the observation group (warfarin) and the control group (aspirin) (OR � 0.33, 95% CI: 0.03∼3.76, P � 0.37) (Figure 7) (see Supplementary Figure 5 for histogram).

Integration of Traditional Chinese and Western
Medicine to Prevent PVT Formation. In anticoagulation therapies, the addition of drugs to promote blood circulation and prevent blood stasis can reduce the incidence of portal vein thrombosis (OR � 0.24, 95% CI: 0.17∼0.34, P < 0.00001) (Figure 8

Discussion
e liver is an important organ that maintains the balance of the hemostatic system. As cirrhosis progresses, disorders of the coagulation and fibrinolytic system may occur, which can easily lead to bleeding and thromboembolism in   Canadian Journal of Gastroenterology and Hepatology 5  [45][46][47][48]. However, the use of anticoagulants may cause some side effects, such as elevated liver enzymes, thrombocytopenia, prolonged prothrombin time, and even life-threatening cases [21,49]. erefore, the effectiveness and safety of anticoagulation therapy were further discussed in this article.
In terms of therapeutic anticoagulation, the results showed that anticoagulation prevents thrombus progression and increases the thrombosis recanalization rate. It is worth mentioning that in a study [50], the portal cavernomas were disappeared in two patients after anticoagulation. In addition, compared with the control group, bleeding events and other adverse events did not increase, and the mortality rate was decreased in the observation group. ese results show that anticoagulation can treat cirrhosis PVT and improve patient survival without increasing side effects. Studies indicated that microthrombosis in the liver sinus exists in patients with cirrhosis [51,52]. Microthrombi can increase portal pressure and cause intimal fibrosis and venous occlusion, eventually causing adjacent liver cells to be lost and replaced by fibrous tissue. Anticoagulation can improve liver fibrosis by combating microthrombosis, further improving liver function and reducing portal hypertension. Francoz et al. [12] found that liver function and renal function were improved in patients treated with enoxaparin. He also noted that enoxaparin could reduce intestinal cell damage by improving intestinal microcirculation, thereby reducing bacterial translocation. e rombosis Canada and 7th International Coagulation in Liver Disease Conference recommended liver transplant candidates with PVT for anticoagulation therapy and pointed out nontransplant candidates with acute PVT may also benefit [53]. erefore, anticoagulation represents a safe, effective, and reliable option for patients with cirrhosis PVT, even those with poor liver function. e 2016 Consensus of the Italian Society of Hepatology and the Italian Medical Association: Hemostasis Balance of Cirrhosis reported that thromboprophylaxis is not absolutely contraindicated in patients with cirrhosis [54]. However, through repeated searches of these literature libraries, only one controlled study [29] on preventive anticoagulation in nonsurgical cirrhosis patients was identified. Villa et al. found that enoxaparin was safe in preventing PVT in cirrhosis patients and delayed the occurrence of hepatic decompensation. However, related studies remain scarce. e possible reasons are as follows [8,21,55]: anticoagulation has serious side effects; PVTdoes not occur in all patients with cirrhosis; some PVT has a very high rate of spontaneous recanalization; and even if PVT is resolved with the use of anticoagulants, it may recur after stopping treatment. Many scholars have employed preventive anticoagulation after splenectomy or cancer resection in patients with liver cirrhosis. e surgical process and postoperative recovery may lead to a persistent hypercoagulable state, hemodynamic changes of the portal vein system, and local vascular disease, further promoting the occurrence of PVT [56,57]. Our data shows that compared with the control group, PVT risk in the observation group does not decrease in patients with liver cirrhosis after cancer resection, but the risk did increase in patients after splenectomy, which is consistent with previous studies [58,59]. However, whether preventive anticoagulation should be a routine treatment for patients with liver cirrhosis remains unclear because the study included in this article assessed patients after surgery for cirrhosis. Our data shows that the incidence of bleeding events in the observation group is higher than that in the control group. We believe that preventive anticoagulation is worth considering in those patients at high risk of PVT, such as those undergoing splenectomy.
Given that common anticoagulants have advantages and disadvantages, they should be used with the principle of "individualization." Our results show that the effect of direct oral anticoagulants is improved compared with traditional anticoagulants, and warfarin and aspirin exhibit no significant differences when used in prophylactic anticoagulant therapy. In addition, the combination of traditional Chinese and Western medicine can also achieve good results without increasing the risk of abnormal blood clotting. Intagliata et al. [27] reported that dabigatran or rivaroxaban combined with antiplatelet agents is safer compared with warfarin. Despite these findings, we still need to choose the ideal drug based on the actual situation of the patient. e first factor to consider is pharmacokinetics, especially the functional state of the liver and kidney, which are involved in drug metabolism and clearance. A reduced glomerular filtration rate (GFR) will affect the pharmacokinetics of low molecular  weight heparin (LMWH), and the low density of antithrombin-III in patients with liver cirrhosis may lead to heparin resistance [21]. Patients with renal insufficiency should avoid using dabigatran. e pharmacodynamics of rivaroxaban may be enhanced in patients with liver cirrhosis with poor liver function, while edoxaban, a new oral anticoagulant, is not metabolized by the liver [28,60]. e interaction of drugs with food and other drugs cannot be ignored. For example, some foods rich in vitamin K and antibiotics and other drugs can affect the activity of CYP2C9 enzymes and potentially interfere with the efficacy of warfarin [61]. Economic capacity and compliance should also be taken into account. From our results, it seems that heparin is safer than vitamin K antagonists during the treatment of PVT. However, the high cost, preservation conditions, and daily injection of LMWH cause medical centers to prefer          vitamin K antagonists [62]. For emergency operations, the effect of LMWH exhibits a shorter duration, and the dosage can be adjusted easily and accurately. us, LMWH is better than VKA [12]. e Consensus Statement of the 7th Meeting on Coagulation of Liver Disease suggests that it is important to use direct oral anticoagulants (DOACs) as a treatment option for compensatory liver cirrhosis. LMWH is preferred in an emergency, and treatment should continue until hepatic decompensation is stable. In addition, long-term anticoagulation DOACs can be considered as a safe alternative. DOACs are an effective choice for anticoagulant therapy for patients with heparin-induced thrombocytopenia [15]. Next, we should clarify specific treatment dosages and anticoagulant regimens. In the studies included in this article, the dose and timing of anticoagulant drugs are subjective, and currently, no international standard exists for these parameters. Only two articles discussed the use of enoxaparin and found that it is safer to use it in small doses and at multiple times. e anticoagulant time suggested in each guideline or consensus also varies.
e American Association for the Study of Liver Diseases (AASLD) recommends anticoagulant therapy for at least 3 months to recanalize the PVT in cases with the deterioration of intestinal infarction and portal hypertension [63]. In 2018, the National Comprehensive Cancer Network (NCCN) recommended anticoagulation for at least 6 months without contraindications [64]. In patients with superior mesenteric vein thrombosis, with a past history suggestive of intestinal ischemia or liver transplant candidates, the European Association for the Study of Liver recommended lifelong anticoagulation [65]. e clinical evidence for these problems is inadequate, and data from more clinical trials are needed to support these findings.
In addition, the effects of anticoagulant therapy are affected by many factors, such as age, liver function score, thrombus condition, platelet count, time of thrombosis, hepatic encephalopathy, and hereditary thrombotic disease [13,16]. Delgado et al. [55] proposed that anticoagulant therapy should begin as early as 2 weeks before the discovery of thrombosis because the processes of fibrosis in chronic PVT are irreversible. One study reported that SMV thrombus is an important parameter related to the continuous recanalization of the portal vein. When the PVT extends out of the SMV and the flow rate is reduced by 50%, the anticoagulant effect may be offset by a reduced flow rate [66]. Varicose veins rupture, so bleeding is also associated with PVT recanalization [13]. e 2015 European Guidelines for Hepatic Vascular Disease state that it is important to fully assess the risk of acute bleeding or esophageal and gastric variceal rupture bleeding prior to anticoagulant therapy and to prepare methods to prevent bleeding [65]. It should be noted that approximately 70% to 75% of PVTs occur in malignant tumors [67]. e prognosis of patients with tumor thrombus infiltration is extremely poor, so the use of anticoagulants is not recommended. erefore, attention should also be paid to distinguish a cancer thrombus from a benign thrombus by the combined judgment of imaging features and alpha-fetoprotein levels before anticoagulant treatment [66]. In summary, the clinical decision-making process for anticoagulant therapy requires many comprehensive considerations.  A major limitation of this study is that some articles are nonrandomized controlled trials. ese studies carry a certain level of bias, such as patient selection, drug dosage and course, treatment evaluation, and follow-up. In addition, the lack of patients stratification according to the severity of cirrhosis (compensated/decopensated, CP class A/ B/C, MELD, etc.. . .) in the evaluation of treatment effects prevents us from determining whether all patients with cirrhosis should be treated with anticoagulation. Anticoagulant therapy based on combined traditional Chinese and Western medicine seeks to promote blood circulation by preventing blood stasis during PVT treatment. Preventive anticoagulation also requires comparative clinical trials between the anticoagulant with and without traditional Chinese medicine to further confirm the effect on promoting blood circulation and preventing blood stasis. e longest median follow-up time in the study in this paper is 5 years, and the effect of anticoagulants on long-term prognosis requires further study.

Summary
PVT is a serious adverse event in patients with cirrhosis.
e results show that anticoagulant therapy can effectively and safely treat PVT in patients with cirrhosis and effectively reduce the mortality rate. In addition, this paper also demonstrates that prophylactic anticoagulant therapy can prevent PVT after splenectomy. e necessity of prophylactic anticoagulant therapy requires further discussion. In cases without contraindications, anticoagulants are recommended for liver cirrhosis patients with PVT. e selection of anticoagulant drugs and the dosage and course of drugs should be considered based on the patient's conditions.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare no conflicts of interest.