Splanchnic Vein Thrombosis In China: Etiology And Treatment With Transjugular Intrahepatic Portosystemic Shunt

【Background】Splanchnic vein thrombosis (SVT) includes portal venous system thrombosis (PVST)and Budd-Chiari syndrome (BCS). In spite of its rarity, SVT often results in lethalcomplications. In Western countries, more than80%of SVT patients have at least onethrombotic risk factor. Current practice guidelines recommend that the SVT patientsshould routinely screen for the thrombotic risk factors to identify the etiology and durationof anticoagulation. However, no study has systematically the prevalence of variousthrombotic risk factors in Chinese SVT patients.Transjugular intrahepatic portosystemic shunt (TIPS) is the mainstay treatment optionfor the complications of portal hypertension in liver cirrhosis. However, in the setting ofSVT, the technical difficulty and risk of TIPS is remarkably increased. According to the current practice guideline, portal vein thrombosis (PVT) is considered as a relativecontraindication of TIPS. Several small case series have explored the feasibility and safetyof TIPS in patients with PVT. But the information regarding the predictors of TIPSsuccess and prognostic factors are greatly lacking. On the other hand, pure hepatic vein(HV) thrombosis is the most common type of BCS in Western patients, and TIPS is oftenemployed in such patients. Contrarily, Chinese BCS patients often present with pureinferior vena cava (IVC) obstruction or combined HV/IVC obstruction, and percutaneousrecanalization is more frequently used, but TIPS is rarely attempted. The outcome ofChinese BCS patients treated with TIPS remains unknown.【Aims】The thesis primarily aims to explore the etiology and treatment of BCS in China, andis divided into three main parts. In the first part, we summarized the quantity and type ofscientific publications regarding SVT worldwide, and analyzed the distribution oftreatment modalities for BCS in China. In the second part, we systematically explored thesignificance of screening for thrombotic risk factors in SVT patients, and analyzed theprevalence of various thrombotic risk factors in Chinese SVT patients admitted to ourhospital. In the third part, we evaluated the outcome of TIPS in non-cirrhotic patients withPVT, cirrhotic patients with PVT and BCS patients. Additionally, we reported the studyprotocol and preliminary results of a randomized controlled trial (RCT) conducted by ourteam comparing the safety and efficacy of TIPS versus endoscopic therapy combined withdrug therapy for the prevention of variceal rebleeding in cirrhotic patients with PVT.【Methods】1. All papers regarding SVT were identified via the PubMed, EMBASE, andCochrane library databases. The publication year, country, type of paper, and study designwere summarized. Good quality clinical papers were defined as those in which a highproportion of homogeneous patients with BCS and/or PVT were observed and the samplesize was more than10. The number of citations of original articles regarding PVT and BCS was searched by using Google Scholar.2. All papers regarding the treatment of BCS in Chinese patients were identified viathe PubMed, Chinese Scientific and Technological Journal, and China NationalKnowledge Infrastructure databases. Data regarding the number of BCS patients treatedwith different treatment modalities were collected.3. First, all observational studies regarding the prevalence of JAK2V617F mutationin SVT patients were identified via PubMed database. Meta-analyses were performed tocalculate the pooled prevalence of JAK2V617F mutation or MPN in SVT patients, tocompare the prevalence of JAK2V617F mutation between patients with SVT and thosewith venous thrombosis at other sites, and to compare the prevalence of MPN in patientswith and without JAK2V617F mutation. Second, all consecutive patients with BCS andPVT diagnosed between September2009and May2011were prospectively enrolled inthe observational study and underwent the JAK2V617F mutation detection. Theprevalence of JAK2V617F mutation and its related predictors were calculated.4. PubMed, EMBASE, and Cochrane Library databases were employed to identify allstudies in which inherited antithrombin (AT), protein C (PC), and protein S (PS)deficiencies in PVST and/or BCS were evaluated by family study or gene analysis.Meta-analyses were performed by using StatsDirect software to calculate the prevalenceof these inherited deficiencies in PVST or BCS patients and to compare the prevalence ofthese inherited deficiencies between PVST or BCS patients and healthy individuals.5. PubMed database was employed to identify all studies in which AT, PC, and PSconcentrations were measured in both cirrhotic patients with and without PVT. Using theRevMan software, a standardized mean difference (SMD) with95%confidence interval(CI) was calculated to evaluate the effect of AT, PC, and PS on PVT.6. PubMed, EMBASE, Cochrane Library, and ScienceDirect databases were searched.Meta-analyses were performed by using RevMan software to compare the prevalence ofthe FVL and prothrombin G20210A mutations between patients with BCS or PVT withoutcirrhosis and healthy individuals and between patients with cirrhosis, with and withoutPVT. Odds ratios (OR) were expressed. 7. PubMed, EMBASE, Cochrane Library, and ScienceDirect databases were searched.Eligible studies should compare the prevalence of the5,10-methylenetetrahydrofolatereductase (MTHFR) C677T mutation or hyperhomocysteinemia (HH) or the homocysteinelevels between BCS or non-cirrhotic PVT patients and healthy controls or betweencirrhotic patients with and without PVT. OR or weighted mean difference (WMD) with95%CI was calculated.8. Patients with SVT who were consecutively admitted to our department or regularlyfollowed up between September2009and December2011were eligible for the study anddetected the expression of CD55and CD59on erythrocytes and granulocytes. PNH wasdiagnosed by both CD55and CD59deficient clone at flow cytometry of peripheral bloodcells. The prevalence of PNH was calculated.9. First,169BCS patients who were consecutively admitted to our departmentbetween July1999and December2011underwent the screening tests for thrombotic riskfactors. Second,141non-malignant and non-cirrhotic patients with PVT who wereconsecutively admitted to our department between September2009and August2012underwent the screening tests for the thrombotic risk factors. The prevalence of variousthrombotic risk factors was calculated.10. First, all studies in which TIPS were attempted in PVT patients were identifiedvia PubMed database. We analyzed the current status regarding this topic, TIPStherapeutic strategy, TIPS success rate, and complications. Second,20consecutivenon-cirrhotic patients with portal cavernoma who were admitted to out center betweenJuly2002and December2009and underwent TIPS were included. The TIPS success rate,complications, and prognosis were retrospectively evaluated. The incidence of varicealrebleeding and death were compared between TIPS success and failure groups. Third,57consecutive patients receiving TIPS between December2001and September2008wereincluded. TIPS therapeutic strategy was described. Kaplan-Meier curve was used toevaluate the cumulative risk of shunt dysfunction, post-TIPS hepatic encephalopathy, andsurvival. Logistic regression model was used to calculate the independent predictors ofTIPS success, and Cox regression model was used to calculate the prognostic factors. Finally, a RCT is being conducted by our team to compare the safety and efficacy of TIPSversus endoscopic therapy combined with drug therapy for the prevention of varicealrebleeding in cirrhotic patients with PVT. The study protocol and preliminary results werereported.11. First, all literatures in which TIPS were attempted in BCS patients were identifiedvia the PubMed, EMBASE, and Cochrane library databases. TIPS success rate,complications, efficacy, and prognosis were summarized. Second, all consecutive BCSpatients treated with TIPS between December2004and June2012were included.Indications, TIPS-related complications, post-TIPS hepatic encephalopathy, shuntdysfunction, and death were reported. Kaplan-Meier curve was used to describe thecumulative risks. As for the time-dependent variables, Cox regression model was used tocalculate the independent predictors. The outcome between the early and converted TIPSgroups was compared by using log-rank test.【Results】1. Overall,10,936and11,402papers regarding PVT and BCS were searched,respectively. After excluding the duplicates,6691and4325papers regarding PVT andBCS were identified, respectively. The number of papers gradually increased over time.Researchers from the USA published the greatest number of papers (PVT: n=1418; BCS:n=888). After2001, the number of papers from China was significantly elevated. Clinicalstudy was the most common type of paper (PVT: n=5395; BCS n=3171), but fewer thanhalf of the clinical studies had greater than10patients (PVT: n=2667/5395; BCS:n=1092/3171). Furthermore, fewer than half of the clinical studies with greater than10patients were of good quality (PVT:976/2667; BCS:466/1092). According to the studydesign, the good quality papers were classified as cohort studies (PVT: n=865; BCS:n=421), case-control studies (PVT: n=98; BCS: n=45), and randomized controlled trials(PVT: n=13; BCS: n=0). The5most frequently cited original articles andguidelines/consensuses were also listed.2. A total of3005papers about BCS in China were initially retrieved. Among them, 300papers were included in the systematic review. These papers included23352BCSpatients treated with different treatment modalities. The treatment modalities includedsurgery (n=8625), interventional treatment (n=13940), surgery combined withinterventional treatment (n=363), medical therapy (n=277), other treatment (n=91), and notreatment (n=56). After2005, the number of BCS patients treated with surgery wasdrastically decreased, but the number of BCS patients treated with interventional treatmentwas almost maintained. Shunt surgery was the most common type of surgery (n=3610).Liver transplantation was rarely employed (n=2). Balloon angioplasty with or withoutstenting was the most common type of interventional treatment (n=13747). TIPS wasrarely employed (n=81).3. First,23studies were included to explore the association between SVT and JAK2V617F mutation. Regardless of underlying aetiological factors, the pooled prevalence ofJAK2V617F mutation was37%and24%in patients with BCS and PVST, respectively.After pre-existing MPN was excluded, the pooled prevalence was decreased to26%and19%. Heterogeneity among studies was significant for the prevalence of JAK2V617Fmutation. Compared with healthy subjects and patients with thrombosis in other sites, theprevalence of JAK2V617F mutation was significantly higher in patients with BCS andPVST. The prevalence of MPN was significantly higher in patients with JAK2V617Fmutation than those without. Second, the prevalence of the JAK2V617F mutation was4.3%(4/92) in Chinese patients with primary BCS. All BCS patients with the JAK2V617F mutation had both platelet count (PLT) of above100*109/L (range,107-188*109/L)and splenomegaly. Prevalence of the JAK2V617F mutation was26.6%(17/64) innon-malignant and non-cirrhotic patients with PVT. Higher PLT (P=0.005, OR=1.006,95%CI:1.002-1.011) and older ages (P=0.001, OR=1.117,95%CI:1.064-1.302) were theindependent predictors of the JAK2V617F mutation. Further, the difference in PLTbetween the patients with and without the mutation displayed greater significance in thesubgroup of patients with splenomegaly (P<0.0001), but the statistical significancedisappeared in the subgroup of patients with splenectomy (P=0.1312). Prevalence of theJAK2V617F mutation was1.4%(1/71) in cirrhotic patients with PVT. 4. Nine studies were included to explore the prevalence inherited AT, PC, and PSdeficiencies of in SVT. The pooled prevalence of inherited AT, PC, and PS deficiencieswere3.9%,5.6%, and2.6%in PVST, and2.3%,3.8%, and3.0%in BCS, respectively.Heterogeneity among studies was not significant except for the analysis of inherited PCdeficiency in BCS. Three studies compared the prevalence of these inherited deficienciesbetween PVST patients and healthy subjects. The pooled odds ratios of inherited AT, PC,and PS deficiencies for PVST patients were8.89(95%CI:2.34-33.72, P=0.0011),17.63(95%CI:1.97-158.21, P=0.0032), and8.00(95%CI:1.61-39.86, P=0.011), respectively.Only one study demonstrated that no inherited deficiency was found in both BCS patientsand healthy subjects.5. Seven studies were included to explore the association between AT, PC, and PSconcentrations and PVT in liver cirrhosis. AT and PC concentrations were similar betweenPVT and non-PVT group (AT: SMD=-0.21,95%CI=-0.56to0.14, P=0.24; PC:SMD=-0.23,95%CI=-0.55to0.09, P=0.16). But PS concentration was significantly lowerin PVT than non-PVT group (SMD=-0.29,95%CI=-0.49to-0.08, P=0.006). Subgroupanalyses were further conducted in four studies in which baseline liver function wassimilar between cirrhotic patients with and without PVT, showing similar AT, PC, and PSconcentrations between the two groups (AT: SMD=-0.10,95%CI=-0.36to0.16, P=0.57;PC: SMD=-0.18,95%CI=-0.62to0.25, P=0.41; PS: SMD=-0.10,95%CI=-0.59to0.39,P=0.69).6. Twenty-seven papers were included to explore the association between FVL orfactor II G20210A mutation and SVT. Compared with healthy individuals, patients withBCS had a significantly higher prevalence of the FVL mutation (OR=6.21;95%CI,3.93-9.79) and a similar prevalence of the factor II G20210A mutation (OR=1.90;95%CI,0.69-5.23); patients with PVT without cirrhosis also had a significantly higher prevalenceof the FVL mutation (OR=6.21;95%CI,3.93-9.79) or the factor II G20210A mutation(OR=5.01;95%CI,3.03-8.30). Compared to patients with cirrhosis without PVT, patientswith cirrhosis and PVT had a significantly higher prevalence of the FVL mutation(OR=2.55;95%CI,1.29-5.07). We observed a trend toward a higher prevalence of the factor II G20210A mutation in patients with cirrhosis and PVT, but the difference was notstatistically significant (OR=2.93;95%CI,0.94-9.07).7. Twenty papers were included to explore the association between MTHFR mutationor HH and SVT. BCS and non-cirrhotic PVT patients had a higher prevalence ofhomozygous MTHFR mutation than healthy controls. The difference was statisticallysignificant in BCS patients (OR=2.01,95%CI:1.12-3.61), but not in non-cirrhotic PVTpatients (OR=1.72,95%CI:0.90-3.29). BCS and non-cirrhotic PVT patients had asignificantly higher prevalence of HH (BCS: OR=2.57,95%CI:1.19-5.51; PVT: OR=4.21,95%CI:1.01-17.54) and homocysteine level (BCS: WMD=3.30,95%CI:0.94-5.66; PVT:WMD=2.40,95%CI:0.17-4.63) than healthy controls. Cirrhotic patients with PVT had asignificantly higher prevalence of homozygous MTHFR mutation (OR=2.44,95%CI:1.58-3.76) than those without PVT. But the association between homocysteinelevel and PVT in cirrhotic patients was inconsistent among three studies.8. CD55and/or CD59deficiencies were found in1.6%(2/127) of patients withprimary BCS,1.0%(1/100) of non-malignant and non-cirrhotic patients with PVT, and4.7%(4/85) of cirrhotic patients with PVT. Only one patient had both CD55and CD59deficiencies on granulocytes. But he had been diagnosed with PNH before BCS.9. JAK2V617F mutation was positive in4of169Chinese BCS patients tested.Neither MPL W515L/K nor JAK2exon12mutation was found in any of135patientstested. Overt MPN were diagnosed in5patients (polycythemia vera, n=3; essentialthrombocythemia, n=1; idiopathic myelofibrosis, n=1). Two of them had positive JAK2V617F mutation. Anticardiolipin IgG antibodies were positive or weekly positive in6of166patients tested. Both CD55and CD59deficiencies were found in one of166patientstested. Neither FVL nor factor II G20210A mutation was found in any of136patientstested. HH was found in64of128patients tested. MTHFR C677T mutation was found in96of135patients tested.JAK2V617F mutation was positive in35of141Chinese non-malignant andnon-cirrhotic PVT patients tested. Neither JAK2exon12nor MPL W515L/K mutationwas found in any of50patients tested. Overt MPNs were diagnosed in13patients (polycythemia vera, n=1; essential thrombocythemia, n=9; idiopathic myelofibrosis, n=3).Latent MPNs were considered in23patients with JAK2V617F mutation but without anysignificant abnormalities in regular blood tests. Anticardiolipin IgG antibodies werepositive in none of136patients tested. PNH was not found in any of141patients tested.Neither FVL nor factor II G20210A mutation was found in any of72patients tested.MTHFR C677T mutation was found in29of38patients tested. HH was found in8of39patients tested.10. First,424patients with PVT undergoing TIPS were reported in54articles. TIPSinsertion success rate was67-100%in19case series. In these studies,85patients withportal cavernoma underwent successful TIPS insertions. Three therapeutic strategies ofTIPS were:1) creation of a TIPS followed by portal vein recanalization via the shunt;2)portal vein recanalization via percutaneous approaches followed by TIPS placement; and3)TIPS insertion between a hepatic vein and a large collateral vessel without portal veinrecanalization. Four approaches to access the portal vein included transjugular,transhepatic, transsplenic, and transmesenteric approach. Intra-abdominal hemorrhagesecondary to hepatic capsule perforation was lethal in only three patients. No episode ofpulmonary embolism was reported. Other procedure-related complications were reversible.The overall incidence of shunt dysfunction and hepatic encephalopathy was8-33%and0-50%, respectively.Second, TIPS were successfully placed in35%(7/20) of non-cirrhotic patients withportal cavernoma via a transjugular approach alone (n=1), a combinedtransjugular/transhepatic approach (n=4), and a combined transjugular/transsplenicapproach (n=2). TIPS were inserted in a large collateral vein in two patients in whomrecanalization of the occluded main portal vein (MPV) was impossible. Procedure-relatedcomplication was hepatic capsule perforation in one patient who was cured by medicaltherapy alone. Shunt dysfunction occurred in two patients, but TIPS revision was failed inone of them. Portosystemic pressure gradient was significantly reduced in TIPS successgroup (26.3±1.1versus12.4±1.1mmHg, p<0.001). The incidence of variceal bleeding inTIPS success group is lower than that in TIPS failure group (14%versus69%, p=0.057). In TIPS success group, two patients died of systemic infection and accident, respectively.In TIPS failure group, two patients died of liver failure.Third, TIPS were successfully placed in75%of cirrhotic patients with PVT (43/57).The independent predictors of technical success included portal cavernoma, and the degreeof thrombosis within the MPV, the portal vein branches and the superior mesenteric vein.Only one patient died of severe procedure-related complication. The cumulative one-yearshunt dysfunction and hepatic encephalopathy rates were21%and25%, respectively. Thecumulative one-and five-year variceal rebleeding rates differed significantly between theTIPS success and failure groups (10%and28%v.s.43%and100%, respectively;P=0.0004), while the cumulative one-and five-year survival rates were similar betweenthe two groups (86%and77%v.s.78%and62%, respectively; P=0.34). The independentpredictor of survival in PVT patients with decompensated cirrhosis was the degree ofMPV occlusion (hazard ratio=0.189,95%CI:0.042-0.848).Fourth, this RCT regarding TIPS for prevention of variceal rebleeding in cirrhoticpatients with PVT was approved by the ethics committee of Xijing hospital (No.20110224-5) on February24,2011, and was registered at ClinicalTrials.gov(NCT01326949) on March29,2011. The first patient was recruited into our study on4June2011. Until January15,2014, all patients were enrolled, a total of52patientsparticipated in this trial. Among them,38.5%were Child class A, and71.2%had partialPVT. The follow-up length is not enough, so we cannot compare the efficacy and safety oftwo groups.11. First,160papers from29countries reported the application of TIPS for BCS. Thenumber of publications was increased over time, but the level of evidence in this field waslow. Common indications for TIPS in BCS patients included refractory ascites, recurrentvariceal bleeding, diffuse HV thrombosis, and progressive liver failure. Successful TIPSinsertion could improve the hemodynamic and clinical parameters. TIPS procedure-relatedcomplications were not infrequent (range:0%-56%), but procedure-related death was rare.Shunt dysfunction rate appeared to be higher (range:18%-100%). Compared with barestents, covered stents could significantly decrease the rate of shunt dysfunction. Hepatic encephalopathy rate after TIPS was relatively low (range:0%-25%). Short and long-termprognosis of BCS-TIPS patients was excellent with1-year cumulative survival rate of80%-100%and5-year cumulative survival rate of74%-78%.Second,39of51BCS patients underwent percutaneous recanalization1024days(0-4574) before TIPS. Early TIPS group (n=19) has a shorter history of BCS and a lowerproportion of prior percutaneous recanalization than converted TIPS group (n=32). Mainindications were diffuse obstruction of three HVs (n=12), liver failure (n=2), liver functiondeterioration (n=8), refractory ascites (n=10), and variceal bleeding (n=19).Procedure-related intraperitoneal bleeding was reversible in3patients. The cumulative1-year rate of being free of first episode of post-TIPS hepatic encephalopathy and shuntdysfunction was78.38%and61.69%, respectively. The cumulative1-,2-, and3-yearsurvival rates were83.82%,81.20%, and76.93%, respectively. BCS-TIPS score, but notChild-Pugh, MELD, Clichy, or Rotterdam score, could predict the survival. Age, totalbilirubin, and IVC thrombosis were also significantly associated with overall survival.Survival was similar between early and converted TIPS groups.【Conclusions】1. There is a gradually increasing trend in the number of papers regarding PVT andBCS worldwide, but the grade of evidence is still low. The selection of treatmentmodalities is greatly different between China and Western countries. In China, drugtherapy alone, TIPS and liver transplantation are rarely used, but percutaneousrecanalization is the most common treatment modality used.2. Routine screening for JAK2V617F mutation in Western patients with SVT isnecessary. However, the screening is just suitable for non-cirrhotic patients with PVT inChina, but not for patients with BCS.3. In Western patients, the prevalence of inherited AT, PC, and PS deficiencies arevery low. Its association with non-cirrhotic PVT is remarkable, but the current evidencefails to support its association with BCS. Additionally, AT, PC and PS may not beassociated with the development of PVT in liver cirrhosis. 4. In Western patients, the FVL mutation increases the risk of BCS and PVT withoutcirrhosis. The factor II G20210A mutation increases the risk of PVT without cirrhosis, butnot BCS. Additionally, the FVL and factor II G20210A mutation may contribute to thepathogenesis of PVT in patients with cirrhosis. However, none of the Chinese patientswith SVT has FVL or factor II G20210A mutation.5. Homozygous MTHFR mutation and HH might be associated with the occurrenceof BCS and non-cirrhotic PVT. In addition, homozygous MTHFR mutation might increasethe risk of PVT in cirrhotic patients. However, the current evidence failed to support theassociation of HH with PVT in cirrhotic patients.6. In Chinese patients with SVT, PNH is rare. So the routine screening for CD55andCD59expression is unnecessary.7. TIPS is technically feasible in patients with PVT. Percutaneous trans-hepatic andtrans-splenic approaches could facilitate the TIPS procedure, but the procedure-relatedcomplication should be fully evaluated before TIPS. Although TIPS can decrease theincidence of variceal rebleeding, its effect on the survival should be further explored. Theindependent predictor of survival is the degree of MPV thrombosis in the cirrhotic patientswith PVT. Our team is conducting a RCT comparing the efficacy and safety of TIPSversus conventional first-line therapy for the prevention of variceal rebleeding in cirrhoticpatients. Unitil now, the patients have been enrolled and are being followed.8. TIPS can achieve an excellent survival in Chinese patients in whom percutaneousrecanalization is ineffective or inappropriate, but results in a high incidence of shuntdysfunction. BCS-TIPS score can effectively predict the candidates for TIPS. Additionally,early TIPS may not influence the patients’ survival.