CLINICAL STUDY
Emergent Salvage Direct Intrahepatic Portocaval Shunt Procedure for Acute Variceal Hemorrhage Thomas J. Ward, MD, Tust Techasith, MD, John D. Louie, MD, Gloria L. Hwang, MD, Lawrence V. Hofmann, MD, and Daniel Y. Sze, MD, PhD
ABSTRACT Purpose: To review the safety and effectiveness of direct intrahepatic portocaval shunt (DIPS) creation with variceal embolization for acute variceal hemorrhage after a failed transjugular intrahepatic portosystemic shunt (TIPS) creation attempt or in patients with prohibitive anatomy. Materials and Methods: Transjugular intrahepatic portosystemic shunt and DIPS procedures performed for variceal hemorrhage between January 2008 and July 2014 were reviewed. The default procedure was TIPS creation, with DIPS creation reserved for patients with unfavorable anatomy or who had technically unsuccessful TIPS creation. Thirteen patients underwent DIPS creation (mean age, 60 y ⫾ 12; Child–Pugh class A/B/C, 8%/62%/30%; Model for End-stage Liver Disease score, 15 ⫾ 5; range, 8–26) and 117 underwent TIPS creation. Four patients underwent a TIPS attempt and were converted to DIPS creation upon technical failure; 9 were treated primarily with DIPS creation because of preprocedural imaging revealing unfavorable anatomy (intrahepatic portal thrombosis, n ¼ 2; venous distortion from prior hepatic resections, n ¼ 2; severely angulated hepatic veins, n ¼ 5). Results: Direct intrahepatic portocaval shunt creation with variceal embolization (six gastric or esophageal; seven stomal, duodenal, or rectal) was successful in all patients; 11 also had concomitant variceal sclerotherapy. Mean DIPS procedure time was less than 2 hours. There was 1 major procedural complication. During a mean follow-up of 13.0 months ⫾ 15.5, 1 patient developed DIPS thrombosis and recurrent hemorrhage; 1 patient underwent successful transplantation. Two deaths were observed within 30 days, neither associated with recurrent hemorrhage. Conclusions: Direct intrahepatic portocaval shunt creation appears to be a safe, expedient, and effective treatment for patients with acute variceal hemorrhage who are poor anatomic candidates for TIPS creation or who have undergone unsuccessful TIPS creation attempts.
ABBREVIATIONS DIPS = direct intrahepatic portocaval shunt, MELD = Model for End-stage Liver Disease, TIPS = transjugular intrahepatic portosystemic shunt
From the Division of Interventional Radiology, Stanford University Medical Center, 300 Pasteur Dr., Stanford, CA 95305. Received December 15, 2014; final revision received March 9, 2015; accepted March 11, 2015. Address correspondence to T.J.W.; E-mail: [email protected] From the SIR 2015 Annual Meeting. D.Y.S. is a paid consultant for Amgen (Thousand Oaks, California), BTG (West Conshohocken, Pennsylvania), Sirtex Medical (North Sydney, Australia), W.L. Gore & Associates (Flagstaff, Arizona), Covidien (Mansfield, Massachusetts), Guerbet (Villepinte, France), Cook (Bloomington, Indiana), Boston Scientific, Inc. (Marlboro, Massachusetts), and Codman (Raynham, Massachusetts), and serves on the advisory boards of SureFire Medical (Westminster, Colorado), Koli Medical (Fremont, California), Northwind Medical (San Jose, California), Treus Medical (Redwood City, California), RadiAction Medical (Tel Aviv, Israel), EmbolX (Los Altos, California), and Lunar Design (Palo Alto, California). None of the other authors have identified a conflict of interest. & SIR, 2015 J Vasc Interv Radiol 2015; 26:829–834 http://dx.doi.org/10.1016/j.jvir.2015.03.004
Variceal hemorrhage represents a major cause of morbidity and mortality in patients with portal hypertension and cirrhosis (1–4). The creation of a transjugular intrahepatic portosystemic shunt (TIPS) has been shown to decrease recurrent hemorrhage and mortality compared with endoscopic band ligation and vasoactive drugs (5). Creation of a TIPS, however, may be technically difficult or impossible in patients with hepatic venous anomalies, intra- or extrahepatic portal venous obstruction, or distorted postsurgical anatomy. The creation of an intravascular ultrasound (US)– guided direct intrahepatic portocaval shunt (DIPS) is an alternative to TIPS creation (6,7) with high rates of technical and clinical success. DIPS creation has been primarily reported in the elective setting for portal hypertension complicated by ascites. There is less
830
’
Emergent Salvage DIPS Procedure for Acute Variceal Hemorrhage
experience with the emergent creation of DIPSs for acute variceal hemorrhage, with DIPSs likely associated with additional setup time at less experienced centers. In the largest series of 40 patients treated with the DIPS procedure for the sequela of portal hypertension (8), only five patients were treated for bleeding. In these patients, there was one episode of recurrent bleeding at 1 year. The technical success rate for the entire cohort was 100%, with the clinical success rate being 74% for patients treated for ascites. The purpose of the present study was to investigate the application of the DIPS procedure for acute variceal hemorrhage, reserved for use after a failed TIPS attempt or in patients with prohibitive anatomy for TIPS.
MATERIALS AND METHODS This study was an institutional review board–approved, single-center, retrospective review of all portosystemic shunt procedures performed between January 2008 and July 2014. Data were handled in accordance with the Health Insurance Portability and Accountability Act. At the performing institution, TIPS creation is the default emergent procedure performed for acute variceal hemorrhage. DIPS creation is reserved for use after a failed TIPS attempt or in patients with prohibitive anatomy, eg, severe cranial angulation and other hepatic venous anomalies, intra- or extrahepatic portal venous thrombosis or obstruction, or distorted postsurgical vascular anatomy such as cranial displacement of intrahepatic portal veins. During the study period, 130 shunt procedures were performed for variceal hemorrhage. Of these, 13 were DIPSs and 117 were conventional TIPSs. All patients first underwent diagnostic endoscopy with attempted band ligation and/or sclerotherapy and were referred for chronic or acute refractory hemorrhage. No patients refused TIPS or DIPS creation, and no surgical shunts were performed for variceal hemorrhage during the study period. Of the patients who received a DIPS, the mean age was 60 years ⫾ 12, 46% were male, Child–Pugh– Turcotte class A, B, and C disease was seen in 8%, 62%, and 30%, respectively, and mean Model for End-stage Liver Disease (MELD) score was 15 ⫾ 5 (standard deviation; range, 8–26; Table 1). Four patients underwent a TIPS attempt and were converted to DIPS upon technical failure. Nine were treated primarily with DIPS creation because of cross-sectional imaging revealing unfavorable anatomy: two with intrahepatic portal vein thrombosis, two with distortion of venous anatomy from earlier hepatic resections, and five with severely cranially angulated hepatic veins (Fig 1). In patients with intrahepatic portal vein occlusion, any patency, even partial, of the main portal vein allowed US-guided puncture from the inferior vena cava and creation of a
Ward et al
’
JVIR
Table 1 . Patient Demographics (N ¼ 13) Characteristic
Value
Age (y) Mean ⫾ SD Range INR Mean ⫾ SD Range Creatinine (mg/dL) Mean ⫾ SD Range Total bilirubin (mg/dL)
59.9 ⫾ 12.3 35–80 1.5 ⫾ 0.3 1.1–2.1 1.3 ⫾ 0.7 0.7–2.9
Mean ⫾ SD
2.0 ⫾ 1.5
Range Albumin (mg/dL)
0.5–6.2
Mean ⫾ SD
2.3 ⫾ 0.5
Range MELD score Mean ⫾ SD Range Sex (male:female)
1.4–3.5 15.2 ⫾ 5.3 8–26 6:7 (46:54)
Child–Pugh–Turcotte class A B
1 (8) 8 (62)
C
4 (30)
Note–Values in parentheses are percentages. INR ¼ International Normalized Ratio, MELD ¼ Model for Endstage Liver Disease, SD ¼ standard deviation.
shunt. All patients presented with variceal bleeding within 72 hours of the DIPS procedure, including seven within 24 hours. Three patients required inotropic support, and two underwent the procedure with a Sengstaken–Blakemore gastroesophageal tamponade tube in place. DIPS creations were performed by using intravascular US guidance with the AcuNav intracardiac echocardiography probe (Acuson/Siemens, Mountain View, California). The technique used is a modification of the procedure described in detail by Petersen and Clark (6). Right jugular 10-F and right femoral 12-F sheaths were telescoped coaxially, and, through a side slit in the femoral sheath, a 65.5-cm, 21-gauge Chiba needle sheathed in a shortened 4-F CXI catheter (Cook, Bloomington, Indiana) was advanced through a Rosch–Uchida cannula (Cook), through the caudate lobe, into the main portal vein. Stent length was prescribed by measuring the tract length by intravascular US and simultaneous portal and caval venography using a calibrated pigtail catheter. A VIATORR device (W. L. Gore & Associates, Flagstaff, Arizona) was used in all patients. Because of acute hemorrhage, coil or plug embolization was performed with the goal of elimination of all variceal blood flow and pressure in all patients. Concomitant variceal sclerotherapy was performed with variceal embolization at the discretion of the operator with the use of a slurry
Volume 26
’
Number 6
’
June
’
2015
831
Figure 1. (a) Coronal reformat of an intravenous contrast-enhanced CT scan in an 80-year-old woman with nonalcoholic steatohepatitis and acute duodenal variceal hemorrhage shows an elevated right hemidiaphragm causing severe cranial angulation of the right hepatic vein (arrow), an anatomic configuration that precluded a conventional TIPS pathway from the right hepatic vein to the right portal vein from a transjugular approach. (b) Coronal reformat of an intravenous contrast-enhanced CT scan in a 64-year-old woman with primary biliary cirrhosis and acute rectal variceal hemorrhage reveals thrombosed intra- and extrahepatic portal veins (arrow) causing difficulty in targeting and in confirmation of a successful needle pass. (c) Right oblique maximum-intensity projection of an intravenous contrast-enhanced CT scan in a 72-year-old man with a history of hepatitis B and hepatocellular carcinoma that had been resected shows severe distortion of portal venous, hepatic venous, and vena caval anatomy after right hepatectomy, with the venous drainage coursing predominantly through a very medial and anterior left hepatic vein (arrow).
of 3% sodium tetradecyl sulfate (Sotradecol; Mylan/ AngioDynamics, Latham, New York), ethiodized oil (Lipiodol; Guerbet, Villepinte, France), and gelatin sponge (Surgifoam; Ethicon/Johnson & Johnson, Somerville, New Jersey). Sclerotherapy and variceal embolization were performed after DIPS creation in an antegrade fashion in all patients (9). The time required to complete different stages of the procedure (ie, US setup, portal vein access, stent graft deployment, sheath removal) was extracted from the electronic medical record. Time to DIPS conversion was defined for the four patients who had an unsuccessful TIPS attempt converted to a DIPS procedure as the time from right internal jugular vein puncture for the initial TIPS attempt to right common femoral vein puncture. Procedural complications were graded according to the Society of Interventional Radiology clinical guidelines consensus document (10). Clinical outcomes, including 30-day mortality, recurrence of hemorrhage, occurrence of encephalopathy or hepatic failure, and “bridge” to transplantation were reviewed. Continuous variables are presented as mean and standard deviation; categorical variables are presented as percentage and count.
RESULTS Of the 130 patients treated with portosystemic shunt creation for variceal hemorrhage, 13 underwent DIPS creation because of a failed TIPS attempt or unfavorable anatomy for TIPS creation. DIPS creation was successful at stopping variceal bleeding (six gastric or esophageal; seven stomal, duodenal, or rectal) in all 13
patients. Portosystemic gradient was reduced from 14.0 mm Hg ⫾ 3.2 (range, 10–21 mm Hg) to 5.7 mm Hg ⫾ 2.1 (range, 2–9 mm Hg). Four of 13 patients had DIPS creation after a failed TIPS attempt, with a mean time to DIPS conversion of 124.0 minutes ⫾ 34.4 (range, 86.0– 164.0 min). The mean time for DIPS creation in all 13 patients was 110.8 minutes ⫾ 36.2 (range, 71.0–203.0 min), not including time for embolization and/or sclerotherapy. A 10-mm-diameter stent graft was used in nine patients; an 8-mm-diameter stent graft was used in four patients who were believed to be at high risk of hepatic failure and/or encephalopathy because of a history of encephalopathy and/or high Child–Pugh– Turcotte score. Devices used were relatively short, reflecting the generally short tract lengths and lack of necessity to span an outflow hepatic vein (4 cm covered plus 2 cm bare, n ¼ 4; 5 cm covered plus 2 cm bare, n ¼ 3; 6 cm covered plus 2 cm bare, n ¼ 6). All patients were treated for variceal bleeding with adjunctive variceal embolization in addition to the DIPS procedure. Eleven patients also underwent concomitant variceal sclerotherapy for chemical obliteration of the varices (9). No residual flow into any of the varices was visible at completion venography. The geometry of a DIPS did not impact the ability to perform variceal embolization and sclerotherapy. The times required for the different procedural components are provided in Table 2. One patient had a severe intraprocedural complication when dilation of the DIPS tract caused a portal vein wall laceration, resulting in tachycardia and hypotension. Portal venography confirmed extravasation, which was successfully treated with balloon tamponade before
832
’
Emergent Salvage DIPS Procedure for Acute Variceal Hemorrhage
being sealed with two VIATORR stent-grafts. The patient required fluid resuscitation and initiation of a massive transfusion protocol. A perihepatic peritoneal drainage catheter was inserted, from which 750 mL of bright red blood was aspirated. The patient was discharged after 6 days of hospitalization and did not have recurrent hemorrhage in 3 months of follow-up. There was one recurrent variceal hemorrhage during a mean follow-up of 13.0 months ⫾ 15.5. A patient with inflammatory bowel disease, idiopathic hypercoagulability with branch portal venous thrombosis, and peristomal variceal hemorrhage refractory to embolization and sclerotherapy was treated with a 10-mm stent graft constrained by a balloon-expandable stent dilated to 8 mm (Fig 2). This technique was employed to minimize the risk of encephalopathy. The hourglass-shaped DIPS thrombosed within 48 hours, resulting in repeat hospitalization at another institution, failed recanalization, and Table 2 . Procedural Time Requirements Time (min) Mean ⫾ SD
Range
124.0 ⫾ 34.4 110.8 ⫾ 36.2
86.0–164.0 71.0–203.0
Variceal embolization/sclerotherapy
89.1 ⫾ 30.4
42.0–141. 0
Total DIPS procedure Pre-DIPS gradient
199.8 ⫾ 52.0 132.0–257.0 14.0 ⫾ 3.2 10.0-21.0
Procedural Component TIPS-to-DIPS conversion (n ¼ 4) DIPS creation
Post-DIPS gradient
5.7 ⫾ 2.1
2.0-9.0
DIPS ¼ direct intrahepatic portocaval shunt, SD ¼ standard deviation, TIPS ¼ transjugular intrahepatic portosystemic shunt.
Ward et al
’
JVIR
creation of a parallel 10-mm DIPS, which remained patent. No definite cause for DIPS thrombosis was identified in this patient, but the underlying hypercoagulability may have been a factor. The DIPS tract was completely lined by the covered portion of the graft from the portal vein to the vena cava. In the study cohort of 13 patients, 11 were treated with adjunctive variceal embolization and sclerotherapy and two were treated with embolization alone. The one patient who had recurrent hemorrhage had undergone variceal embolization with sclerotherapy, but the small sample sizes preclude drawing conclusions about comparative efficacy. Because of referral patterns, only seven of 13 patients underwent follow-up imaging or endoscopy at our institution (mean follow-up, 649 d ⫾ 708; range, 28–1,889 d), and no varices were detectable in these seven patients. Two deaths were observed within 30 days, both in patients with massive variceal bleeding and hemodynamic collapse requiring placement of a Sengstaken–Blakemore tube and initiation of a massive transfusion protocol. The first patient had Child–Pugh class C disease (Child-Pugh score: 10) with a MELD score of 26 and an Acute Physiology and Chronic Health Evaluation II score of 21 (expected 30-d mortality rate of 38.9%). The patient had uncorrectable coagulopathy and was bleeding uncontrollably from the site of recent tooth extractions and site of traumatic nasogastric tube insertion. The family decided to withdraw care. The second patient had Child–Pugh class C disease (Child-Pugh score: 10) with a MELD score of 20 and an Acute Physiology and Chronic Health Evaluation II score of 28 (expected 30-d mortality rate of 63.9%). The patient required high positive pressure ventilation to maintain adequate
Figure 2. (a) Radiograph of the course of the selective catheterization of the right hepatic vein in a 52-year-old man with acute stomal variceal hemorrhage shows a severely angulated hepatic vein (approximately 1201). Note the dome of the diaphragm (arrow) cranial to the hepatic vein. (b) Right hepatic venography after forced straightening of the curve with an Amplatz super-stiff guide wire (Boston Scientific, Natick, Massachusetts) and a 10-F braided sheath (Flexor; Cook) showed bird-beak–shaped stenosis pointing caudally into the vena cava (arrow), preventing positioning of a metal cannula or needle to enable TIPS creation. (c) Final venography after creation of a DIPS shows the caval end of the stent at approximately the level of hepatic venous confluence, but unaffected by the unfavorable angulation. A 10-mm stent graft was deployed inside an 8-mm constraining balloon-expandable stent.
Volume 26
’
Number 6
’
June
’
2015
oxygenation and developed a pneumothorax and pulmonary infection complicated by sepsis. These deaths were not attributed to the DIPS procedure; neither patient had recurrence of variceal hemorrhage before death. Four patients developed new or worsened encephalopathy and underwent medical management with lactulose or rifaximin. One patient underwent an uncomplicated liver transplantation involving bicaval anastomosis and resection of the retrohepatic inferior vena cava and DIPS stent.
DISCUSSION Despite extensive literature describing techniques to enable portal vein imaging and access, TIPS creation may be technically difficult or impossible in certain patients for a variety of anatomic reasons. DIPS creation is an alternative technique that may successfully circumvent these anatomic constraints (6,7), and may be a valuable contingency plan after an unsuccessful TIPS creation attempt. Advocates cite short procedural times, low complication rates, and high patency rates. However, there is a learning curve and a requirement for investment in intravascular US hardware. Increased setup time and additional equipment needed in a crowded procedure room may also discourage routine use. Still, in the present series of emergent procedures, DIPS creation was successful at reducing the portosystemic gradient and stopping variceal hemorrhage in all 13 patients. Our procedural time was substantially longer than the reported time of 1 hour from institutions in which DIPSs are more routinely created (6). Achieving improved procedural times and efficiency, especially for emergency procedures, would require a concerted effort to train all practitioners and support staff, and to make equipment setup programmed and routine. Although the benefits of variceal embolization or sclerotherapy are still controversial (11), the geometry of the DIPS did not impact the ability to perform these adjunctive procedures. However, none of the patients underwent periumbilical varix embolization, which could be hindered by a caudally directed extrahepatic portal vein puncture site. Variceal embolization with (n ¼ 11) or without sclerotherapy (n ¼ 2) was successfully performed in all patients in a mean time of 89.1 minutes ⫾ 30.4. The treatment of variceal bleeding is varied, with treatment options dependent on liver function, patient anatomy, and local expertise, with the TIPS procedure shown to be associated with superior recurrent bleeding rates compared with endoscopic methods (12). Recommended societal guidelines cite a 95% technical and 90% clinical success rate for TIPS creation when performed in patients with patent hepatic and portal veins (13). Addition of the DIPS technique for salvage cases allowed us to achieve a 100% technical success rate, regardless of venous patency.
833
In the present patient cohort, only one patient underwent an uncomplicated liver transplantation with a bicaval anastomosis and resection of the retrohepatic inferior vena cava and DIPS stent, and other successful transplantations have been previously reported (7). For bicaval and piggyback anastomosis transplantations, pedicle lengths, stent resection, cavotomy, and venous anastomoses may require additional considerations and surgical planning. The present report is limited by its retrospective nature and lack of blinding or randomization. West Haven stages of encephalopathy were not recorded for the four patients in whom encephalopathy developed after the DIPS procedure. Nine of the patients were judged to have unfavorable anatomy for TIPS creation and no attempt was made to perform TIPS creation before proceeding to a DIPS procedure. The mean follow-up period was 12 months, and, as such, the long-term durability of the salvage DIPS is unknown. In addition, many patients returned to the referring institutions for follow-up, from which clinical documents were available, but postprocedural imaging or endoscopy may not have been routinely performed, so assessment for complete variceal obliteration was limited. Our short-term results are equivalent to those of other DIPS reports (8), which are at least equivalent to published results for TIPS creation. The performing institution is a center where DIPS creation is not the default primary intervention for portal hypertension–related variceal hemorrhage, making the observed results more likely to be achievable in “real-world” practice in which TIPS creation is likely to remain the default intervention. In this setting, DIPS creation appears to be a safe, expedient, and effective treatment for patients with acute variceal hemorrhage who are poor anatomic candidates for conventional TIPS creation or who have undergone an unsuccessful TIPS creation attempt. Despite salvage circumstances, outcomes are comparable to expected outcomes after conventional TIPS creation.
REFERENCES 1. Turon F, Casu S, Hernandez-Gea V, Garcia-Pagan JC. Variceal and other portal hypertension related bleeding. Best Pract Res Clin Gastroenterol 2013; 27:649–664. 2. D’Amico G, De Franchis R. Cooperative Study G. Upper digestive bleeding in cirrhosis. Post-therapeutic outcome and prognostic indicators. Hepatology 2003; 38:599–612. 3. Fidelman N, Kwan SW, LaBerge JM, Gordon RL, Ring EJ, Kerlan RK Jr. The transjugular intrahepatic portosystemic shunt: an update. AJR Am J Roentgenol 2012; 199:746–755. 4. Garcia-Tsao G, Bosch J. Management of varices and variceal hemorrhage in cirrhosis. N Engl J Med 2010; 362:823–832. 5. Garcia-Pagan JC, Caca K, Bureau C, et al. Early use of TIPS in patients with cirrhosis and variceal bleeding. N Engl J Med 2010; 362:2370–2379. 6. Petersen BD, Clark TW. Direct intrahepatic portocaval shunt. Tech Vasc Interv Radiol 2008; 11:230–234. 7. Hoppe H, Wang SL, Petersen BD. Intravascular US-guided direct intrahepatic portocaval shunt with an expanded polytetrafluoroethylenecovered stent-graft. Radiology 2008; 246:306–314.
834
’
Emergent Salvage DIPS Procedure for Acute Variceal Hemorrhage
8. Petersen B, Binkert C. Intravascular ultrasound-guided direct intrahepatic portacaval shunt: midterm follow-up. J Vasc Interv Radiol 2004; 15: 927–938. 9. Saad WE, Sze DY. Variations of Balloon-occluded Retrograde Transvenous Obliteration (BRTO): Balloon-occluded Antegrade Transvenous Obliteration (BATO) and Alternative/Adjunctive Routes for BRTO. Semin in Interv Radiol 2011; 28:314–324. 10. Sacks D, McClenny TE, Cardella JF, Lewis CA. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003; 14: S199–S202.
Ward et al
’
JVIR
11. Chen S, Li X, Wei B, et al. Recurrent variceal bleeding and shunt patency: prospective randomized controlled trial of transjugular intrahepatic portosystemic shunt alone or combined with coronary vein embolization. Radiology 2013; 268:900–906. 12. Zheng M, Chen Y, Bai J, et al. Transjugular intrahepatic portosystemic shunt versus endoscopic therapy in the secondary prophylaxis of variceal rebleeding in cirrhotic patients: meta-analysis update. J Clin Gastroenterol 2008; 42:507–516. 13. Haskal ZJ, Martin L, Cardella JF, et al. Quality improvement guidelines for transjugular intrahepatic portosystemic shunts. SCVIR Standards of Practice Committee. J Vasc Interv Radiol 2001; 12:131–136.
Emergent Salvage Direct Intrahepatic Portocaval Shunt Procedure for Acute Variceal Hemorrhage Thomas J. Ward, MD, Tust Techasith, MD, John D. Louie, MD, Gloria L. Hwang, MD, Lawrence V. Hofmann, MD, and Daniel Y. Sze, MD, PhD
ABSTRACT Purpose: To review the safety and effectiveness of direct intrahepatic portocaval shunt (DIPS) creation with variceal embolization for acute variceal hemorrhage after a failed transjugular intrahepatic portosystemic shunt (TIPS) creation attempt or in patients with prohibitive anatomy. Materials and Methods: Transjugular intrahepatic portosystemic shunt and DIPS procedures performed for variceal hemorrhage between January 2008 and July 2014 were reviewed. The default procedure was TIPS creation, with DIPS creation reserved for patients with unfavorable anatomy or who had technically unsuccessful TIPS creation. Thirteen patients underwent DIPS creation (mean age, 60 y ⫾ 12; Child–Pugh class A/B/C, 8%/62%/30%; Model for End-stage Liver Disease score, 15 ⫾ 5; range, 8–26) and 117 underwent TIPS creation. Four patients underwent a TIPS attempt and were converted to DIPS creation upon technical failure; 9 were treated primarily with DIPS creation because of preprocedural imaging revealing unfavorable anatomy (intrahepatic portal thrombosis, n ¼ 2; venous distortion from prior hepatic resections, n ¼ 2; severely angulated hepatic veins, n ¼ 5). Results: Direct intrahepatic portocaval shunt creation with variceal embolization (six gastric or esophageal; seven stomal, duodenal, or rectal) was successful in all patients; 11 also had concomitant variceal sclerotherapy. Mean DIPS procedure time was less than 2 hours. There was 1 major procedural complication. During a mean follow-up of 13.0 months ⫾ 15.5, 1 patient developed DIPS thrombosis and recurrent hemorrhage; 1 patient underwent successful transplantation. Two deaths were observed within 30 days, neither associated with recurrent hemorrhage. Conclusions: Direct intrahepatic portocaval shunt creation appears to be a safe, expedient, and effective treatment for patients with acute variceal hemorrhage who are poor anatomic candidates for TIPS creation or who have undergone unsuccessful TIPS creation attempts.
ABBREVIATIONS DIPS = direct intrahepatic portocaval shunt, MELD = Model for End-stage Liver Disease, TIPS = transjugular intrahepatic portosystemic shunt
From the Division of Interventional Radiology, Stanford University Medical Center, 300 Pasteur Dr., Stanford, CA 95305. Received December 15, 2014; final revision received March 9, 2015; accepted March 11, 2015. Address correspondence to T.J.W.; E-mail: [email protected] From the SIR 2015 Annual Meeting. D.Y.S. is a paid consultant for Amgen (Thousand Oaks, California), BTG (West Conshohocken, Pennsylvania), Sirtex Medical (North Sydney, Australia), W.L. Gore & Associates (Flagstaff, Arizona), Covidien (Mansfield, Massachusetts), Guerbet (Villepinte, France), Cook (Bloomington, Indiana), Boston Scientific, Inc. (Marlboro, Massachusetts), and Codman (Raynham, Massachusetts), and serves on the advisory boards of SureFire Medical (Westminster, Colorado), Koli Medical (Fremont, California), Northwind Medical (San Jose, California), Treus Medical (Redwood City, California), RadiAction Medical (Tel Aviv, Israel), EmbolX (Los Altos, California), and Lunar Design (Palo Alto, California). None of the other authors have identified a conflict of interest. & SIR, 2015 J Vasc Interv Radiol 2015; 26:829–834 http://dx.doi.org/10.1016/j.jvir.2015.03.004
Variceal hemorrhage represents a major cause of morbidity and mortality in patients with portal hypertension and cirrhosis (1–4). The creation of a transjugular intrahepatic portosystemic shunt (TIPS) has been shown to decrease recurrent hemorrhage and mortality compared with endoscopic band ligation and vasoactive drugs (5). Creation of a TIPS, however, may be technically difficult or impossible in patients with hepatic venous anomalies, intra- or extrahepatic portal venous obstruction, or distorted postsurgical anatomy. The creation of an intravascular ultrasound (US)– guided direct intrahepatic portocaval shunt (DIPS) is an alternative to TIPS creation (6,7) with high rates of technical and clinical success. DIPS creation has been primarily reported in the elective setting for portal hypertension complicated by ascites. There is less
830
’
Emergent Salvage DIPS Procedure for Acute Variceal Hemorrhage
experience with the emergent creation of DIPSs for acute variceal hemorrhage, with DIPSs likely associated with additional setup time at less experienced centers. In the largest series of 40 patients treated with the DIPS procedure for the sequela of portal hypertension (8), only five patients were treated for bleeding. In these patients, there was one episode of recurrent bleeding at 1 year. The technical success rate for the entire cohort was 100%, with the clinical success rate being 74% for patients treated for ascites. The purpose of the present study was to investigate the application of the DIPS procedure for acute variceal hemorrhage, reserved for use after a failed TIPS attempt or in patients with prohibitive anatomy for TIPS.
MATERIALS AND METHODS This study was an institutional review board–approved, single-center, retrospective review of all portosystemic shunt procedures performed between January 2008 and July 2014. Data were handled in accordance with the Health Insurance Portability and Accountability Act. At the performing institution, TIPS creation is the default emergent procedure performed for acute variceal hemorrhage. DIPS creation is reserved for use after a failed TIPS attempt or in patients with prohibitive anatomy, eg, severe cranial angulation and other hepatic venous anomalies, intra- or extrahepatic portal venous thrombosis or obstruction, or distorted postsurgical vascular anatomy such as cranial displacement of intrahepatic portal veins. During the study period, 130 shunt procedures were performed for variceal hemorrhage. Of these, 13 were DIPSs and 117 were conventional TIPSs. All patients first underwent diagnostic endoscopy with attempted band ligation and/or sclerotherapy and were referred for chronic or acute refractory hemorrhage. No patients refused TIPS or DIPS creation, and no surgical shunts were performed for variceal hemorrhage during the study period. Of the patients who received a DIPS, the mean age was 60 years ⫾ 12, 46% were male, Child–Pugh– Turcotte class A, B, and C disease was seen in 8%, 62%, and 30%, respectively, and mean Model for End-stage Liver Disease (MELD) score was 15 ⫾ 5 (standard deviation; range, 8–26; Table 1). Four patients underwent a TIPS attempt and were converted to DIPS upon technical failure. Nine were treated primarily with DIPS creation because of cross-sectional imaging revealing unfavorable anatomy: two with intrahepatic portal vein thrombosis, two with distortion of venous anatomy from earlier hepatic resections, and five with severely cranially angulated hepatic veins (Fig 1). In patients with intrahepatic portal vein occlusion, any patency, even partial, of the main portal vein allowed US-guided puncture from the inferior vena cava and creation of a
Ward et al
’
JVIR
Table 1 . Patient Demographics (N ¼ 13) Characteristic
Value
Age (y) Mean ⫾ SD Range INR Mean ⫾ SD Range Creatinine (mg/dL) Mean ⫾ SD Range Total bilirubin (mg/dL)
59.9 ⫾ 12.3 35–80 1.5 ⫾ 0.3 1.1–2.1 1.3 ⫾ 0.7 0.7–2.9
Mean ⫾ SD
2.0 ⫾ 1.5
Range Albumin (mg/dL)
0.5–6.2
Mean ⫾ SD
2.3 ⫾ 0.5
Range MELD score Mean ⫾ SD Range Sex (male:female)
1.4–3.5 15.2 ⫾ 5.3 8–26 6:7 (46:54)
Child–Pugh–Turcotte class A B
1 (8) 8 (62)
C
4 (30)
Note–Values in parentheses are percentages. INR ¼ International Normalized Ratio, MELD ¼ Model for Endstage Liver Disease, SD ¼ standard deviation.
shunt. All patients presented with variceal bleeding within 72 hours of the DIPS procedure, including seven within 24 hours. Three patients required inotropic support, and two underwent the procedure with a Sengstaken–Blakemore gastroesophageal tamponade tube in place. DIPS creations were performed by using intravascular US guidance with the AcuNav intracardiac echocardiography probe (Acuson/Siemens, Mountain View, California). The technique used is a modification of the procedure described in detail by Petersen and Clark (6). Right jugular 10-F and right femoral 12-F sheaths were telescoped coaxially, and, through a side slit in the femoral sheath, a 65.5-cm, 21-gauge Chiba needle sheathed in a shortened 4-F CXI catheter (Cook, Bloomington, Indiana) was advanced through a Rosch–Uchida cannula (Cook), through the caudate lobe, into the main portal vein. Stent length was prescribed by measuring the tract length by intravascular US and simultaneous portal and caval venography using a calibrated pigtail catheter. A VIATORR device (W. L. Gore & Associates, Flagstaff, Arizona) was used in all patients. Because of acute hemorrhage, coil or plug embolization was performed with the goal of elimination of all variceal blood flow and pressure in all patients. Concomitant variceal sclerotherapy was performed with variceal embolization at the discretion of the operator with the use of a slurry
Volume 26
’
Number 6
’
June
’
2015
831
Figure 1. (a) Coronal reformat of an intravenous contrast-enhanced CT scan in an 80-year-old woman with nonalcoholic steatohepatitis and acute duodenal variceal hemorrhage shows an elevated right hemidiaphragm causing severe cranial angulation of the right hepatic vein (arrow), an anatomic configuration that precluded a conventional TIPS pathway from the right hepatic vein to the right portal vein from a transjugular approach. (b) Coronal reformat of an intravenous contrast-enhanced CT scan in a 64-year-old woman with primary biliary cirrhosis and acute rectal variceal hemorrhage reveals thrombosed intra- and extrahepatic portal veins (arrow) causing difficulty in targeting and in confirmation of a successful needle pass. (c) Right oblique maximum-intensity projection of an intravenous contrast-enhanced CT scan in a 72-year-old man with a history of hepatitis B and hepatocellular carcinoma that had been resected shows severe distortion of portal venous, hepatic venous, and vena caval anatomy after right hepatectomy, with the venous drainage coursing predominantly through a very medial and anterior left hepatic vein (arrow).
of 3% sodium tetradecyl sulfate (Sotradecol; Mylan/ AngioDynamics, Latham, New York), ethiodized oil (Lipiodol; Guerbet, Villepinte, France), and gelatin sponge (Surgifoam; Ethicon/Johnson & Johnson, Somerville, New Jersey). Sclerotherapy and variceal embolization were performed after DIPS creation in an antegrade fashion in all patients (9). The time required to complete different stages of the procedure (ie, US setup, portal vein access, stent graft deployment, sheath removal) was extracted from the electronic medical record. Time to DIPS conversion was defined for the four patients who had an unsuccessful TIPS attempt converted to a DIPS procedure as the time from right internal jugular vein puncture for the initial TIPS attempt to right common femoral vein puncture. Procedural complications were graded according to the Society of Interventional Radiology clinical guidelines consensus document (10). Clinical outcomes, including 30-day mortality, recurrence of hemorrhage, occurrence of encephalopathy or hepatic failure, and “bridge” to transplantation were reviewed. Continuous variables are presented as mean and standard deviation; categorical variables are presented as percentage and count.
RESULTS Of the 130 patients treated with portosystemic shunt creation for variceal hemorrhage, 13 underwent DIPS creation because of a failed TIPS attempt or unfavorable anatomy for TIPS creation. DIPS creation was successful at stopping variceal bleeding (six gastric or esophageal; seven stomal, duodenal, or rectal) in all 13
patients. Portosystemic gradient was reduced from 14.0 mm Hg ⫾ 3.2 (range, 10–21 mm Hg) to 5.7 mm Hg ⫾ 2.1 (range, 2–9 mm Hg). Four of 13 patients had DIPS creation after a failed TIPS attempt, with a mean time to DIPS conversion of 124.0 minutes ⫾ 34.4 (range, 86.0– 164.0 min). The mean time for DIPS creation in all 13 patients was 110.8 minutes ⫾ 36.2 (range, 71.0–203.0 min), not including time for embolization and/or sclerotherapy. A 10-mm-diameter stent graft was used in nine patients; an 8-mm-diameter stent graft was used in four patients who were believed to be at high risk of hepatic failure and/or encephalopathy because of a history of encephalopathy and/or high Child–Pugh– Turcotte score. Devices used were relatively short, reflecting the generally short tract lengths and lack of necessity to span an outflow hepatic vein (4 cm covered plus 2 cm bare, n ¼ 4; 5 cm covered plus 2 cm bare, n ¼ 3; 6 cm covered plus 2 cm bare, n ¼ 6). All patients were treated for variceal bleeding with adjunctive variceal embolization in addition to the DIPS procedure. Eleven patients also underwent concomitant variceal sclerotherapy for chemical obliteration of the varices (9). No residual flow into any of the varices was visible at completion venography. The geometry of a DIPS did not impact the ability to perform variceal embolization and sclerotherapy. The times required for the different procedural components are provided in Table 2. One patient had a severe intraprocedural complication when dilation of the DIPS tract caused a portal vein wall laceration, resulting in tachycardia and hypotension. Portal venography confirmed extravasation, which was successfully treated with balloon tamponade before
832
’
Emergent Salvage DIPS Procedure for Acute Variceal Hemorrhage
being sealed with two VIATORR stent-grafts. The patient required fluid resuscitation and initiation of a massive transfusion protocol. A perihepatic peritoneal drainage catheter was inserted, from which 750 mL of bright red blood was aspirated. The patient was discharged after 6 days of hospitalization and did not have recurrent hemorrhage in 3 months of follow-up. There was one recurrent variceal hemorrhage during a mean follow-up of 13.0 months ⫾ 15.5. A patient with inflammatory bowel disease, idiopathic hypercoagulability with branch portal venous thrombosis, and peristomal variceal hemorrhage refractory to embolization and sclerotherapy was treated with a 10-mm stent graft constrained by a balloon-expandable stent dilated to 8 mm (Fig 2). This technique was employed to minimize the risk of encephalopathy. The hourglass-shaped DIPS thrombosed within 48 hours, resulting in repeat hospitalization at another institution, failed recanalization, and Table 2 . Procedural Time Requirements Time (min) Mean ⫾ SD
Range
124.0 ⫾ 34.4 110.8 ⫾ 36.2
86.0–164.0 71.0–203.0
Variceal embolization/sclerotherapy
89.1 ⫾ 30.4
42.0–141. 0
Total DIPS procedure Pre-DIPS gradient
199.8 ⫾ 52.0 132.0–257.0 14.0 ⫾ 3.2 10.0-21.0
Procedural Component TIPS-to-DIPS conversion (n ¼ 4) DIPS creation
Post-DIPS gradient
5.7 ⫾ 2.1
2.0-9.0
DIPS ¼ direct intrahepatic portocaval shunt, SD ¼ standard deviation, TIPS ¼ transjugular intrahepatic portosystemic shunt.
Ward et al
’
JVIR
creation of a parallel 10-mm DIPS, which remained patent. No definite cause for DIPS thrombosis was identified in this patient, but the underlying hypercoagulability may have been a factor. The DIPS tract was completely lined by the covered portion of the graft from the portal vein to the vena cava. In the study cohort of 13 patients, 11 were treated with adjunctive variceal embolization and sclerotherapy and two were treated with embolization alone. The one patient who had recurrent hemorrhage had undergone variceal embolization with sclerotherapy, but the small sample sizes preclude drawing conclusions about comparative efficacy. Because of referral patterns, only seven of 13 patients underwent follow-up imaging or endoscopy at our institution (mean follow-up, 649 d ⫾ 708; range, 28–1,889 d), and no varices were detectable in these seven patients. Two deaths were observed within 30 days, both in patients with massive variceal bleeding and hemodynamic collapse requiring placement of a Sengstaken–Blakemore tube and initiation of a massive transfusion protocol. The first patient had Child–Pugh class C disease (Child-Pugh score: 10) with a MELD score of 26 and an Acute Physiology and Chronic Health Evaluation II score of 21 (expected 30-d mortality rate of 38.9%). The patient had uncorrectable coagulopathy and was bleeding uncontrollably from the site of recent tooth extractions and site of traumatic nasogastric tube insertion. The family decided to withdraw care. The second patient had Child–Pugh class C disease (Child-Pugh score: 10) with a MELD score of 20 and an Acute Physiology and Chronic Health Evaluation II score of 28 (expected 30-d mortality rate of 63.9%). The patient required high positive pressure ventilation to maintain adequate
Figure 2. (a) Radiograph of the course of the selective catheterization of the right hepatic vein in a 52-year-old man with acute stomal variceal hemorrhage shows a severely angulated hepatic vein (approximately 1201). Note the dome of the diaphragm (arrow) cranial to the hepatic vein. (b) Right hepatic venography after forced straightening of the curve with an Amplatz super-stiff guide wire (Boston Scientific, Natick, Massachusetts) and a 10-F braided sheath (Flexor; Cook) showed bird-beak–shaped stenosis pointing caudally into the vena cava (arrow), preventing positioning of a metal cannula or needle to enable TIPS creation. (c) Final venography after creation of a DIPS shows the caval end of the stent at approximately the level of hepatic venous confluence, but unaffected by the unfavorable angulation. A 10-mm stent graft was deployed inside an 8-mm constraining balloon-expandable stent.
Volume 26
’
Number 6
’
June
’
2015
oxygenation and developed a pneumothorax and pulmonary infection complicated by sepsis. These deaths were not attributed to the DIPS procedure; neither patient had recurrence of variceal hemorrhage before death. Four patients developed new or worsened encephalopathy and underwent medical management with lactulose or rifaximin. One patient underwent an uncomplicated liver transplantation involving bicaval anastomosis and resection of the retrohepatic inferior vena cava and DIPS stent.
DISCUSSION Despite extensive literature describing techniques to enable portal vein imaging and access, TIPS creation may be technically difficult or impossible in certain patients for a variety of anatomic reasons. DIPS creation is an alternative technique that may successfully circumvent these anatomic constraints (6,7), and may be a valuable contingency plan after an unsuccessful TIPS creation attempt. Advocates cite short procedural times, low complication rates, and high patency rates. However, there is a learning curve and a requirement for investment in intravascular US hardware. Increased setup time and additional equipment needed in a crowded procedure room may also discourage routine use. Still, in the present series of emergent procedures, DIPS creation was successful at reducing the portosystemic gradient and stopping variceal hemorrhage in all 13 patients. Our procedural time was substantially longer than the reported time of 1 hour from institutions in which DIPSs are more routinely created (6). Achieving improved procedural times and efficiency, especially for emergency procedures, would require a concerted effort to train all practitioners and support staff, and to make equipment setup programmed and routine. Although the benefits of variceal embolization or sclerotherapy are still controversial (11), the geometry of the DIPS did not impact the ability to perform these adjunctive procedures. However, none of the patients underwent periumbilical varix embolization, which could be hindered by a caudally directed extrahepatic portal vein puncture site. Variceal embolization with (n ¼ 11) or without sclerotherapy (n ¼ 2) was successfully performed in all patients in a mean time of 89.1 minutes ⫾ 30.4. The treatment of variceal bleeding is varied, with treatment options dependent on liver function, patient anatomy, and local expertise, with the TIPS procedure shown to be associated with superior recurrent bleeding rates compared with endoscopic methods (12). Recommended societal guidelines cite a 95% technical and 90% clinical success rate for TIPS creation when performed in patients with patent hepatic and portal veins (13). Addition of the DIPS technique for salvage cases allowed us to achieve a 100% technical success rate, regardless of venous patency.
833
In the present patient cohort, only one patient underwent an uncomplicated liver transplantation with a bicaval anastomosis and resection of the retrohepatic inferior vena cava and DIPS stent, and other successful transplantations have been previously reported (7). For bicaval and piggyback anastomosis transplantations, pedicle lengths, stent resection, cavotomy, and venous anastomoses may require additional considerations and surgical planning. The present report is limited by its retrospective nature and lack of blinding or randomization. West Haven stages of encephalopathy were not recorded for the four patients in whom encephalopathy developed after the DIPS procedure. Nine of the patients were judged to have unfavorable anatomy for TIPS creation and no attempt was made to perform TIPS creation before proceeding to a DIPS procedure. The mean follow-up period was 12 months, and, as such, the long-term durability of the salvage DIPS is unknown. In addition, many patients returned to the referring institutions for follow-up, from which clinical documents were available, but postprocedural imaging or endoscopy may not have been routinely performed, so assessment for complete variceal obliteration was limited. Our short-term results are equivalent to those of other DIPS reports (8), which are at least equivalent to published results for TIPS creation. The performing institution is a center where DIPS creation is not the default primary intervention for portal hypertension–related variceal hemorrhage, making the observed results more likely to be achievable in “real-world” practice in which TIPS creation is likely to remain the default intervention. In this setting, DIPS creation appears to be a safe, expedient, and effective treatment for patients with acute variceal hemorrhage who are poor anatomic candidates for conventional TIPS creation or who have undergone an unsuccessful TIPS creation attempt. Despite salvage circumstances, outcomes are comparable to expected outcomes after conventional TIPS creation.
REFERENCES 1. Turon F, Casu S, Hernandez-Gea V, Garcia-Pagan JC. Variceal and other portal hypertension related bleeding. Best Pract Res Clin Gastroenterol 2013; 27:649–664. 2. D’Amico G, De Franchis R. Cooperative Study G. Upper digestive bleeding in cirrhosis. Post-therapeutic outcome and prognostic indicators. Hepatology 2003; 38:599–612. 3. Fidelman N, Kwan SW, LaBerge JM, Gordon RL, Ring EJ, Kerlan RK Jr. The transjugular intrahepatic portosystemic shunt: an update. AJR Am J Roentgenol 2012; 199:746–755. 4. Garcia-Tsao G, Bosch J. Management of varices and variceal hemorrhage in cirrhosis. N Engl J Med 2010; 362:823–832. 5. Garcia-Pagan JC, Caca K, Bureau C, et al. Early use of TIPS in patients with cirrhosis and variceal bleeding. N Engl J Med 2010; 362:2370–2379. 6. Petersen BD, Clark TW. Direct intrahepatic portocaval shunt. Tech Vasc Interv Radiol 2008; 11:230–234. 7. Hoppe H, Wang SL, Petersen BD. Intravascular US-guided direct intrahepatic portocaval shunt with an expanded polytetrafluoroethylenecovered stent-graft. Radiology 2008; 246:306–314.
834
’
Emergent Salvage DIPS Procedure for Acute Variceal Hemorrhage
8. Petersen B, Binkert C. Intravascular ultrasound-guided direct intrahepatic portacaval shunt: midterm follow-up. J Vasc Interv Radiol 2004; 15: 927–938. 9. Saad WE, Sze DY. Variations of Balloon-occluded Retrograde Transvenous Obliteration (BRTO): Balloon-occluded Antegrade Transvenous Obliteration (BATO) and Alternative/Adjunctive Routes for BRTO. Semin in Interv Radiol 2011; 28:314–324. 10. Sacks D, McClenny TE, Cardella JF, Lewis CA. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003; 14: S199–S202.
Ward et al
’
JVIR
11. Chen S, Li X, Wei B, et al. Recurrent variceal bleeding and shunt patency: prospective randomized controlled trial of transjugular intrahepatic portosystemic shunt alone or combined with coronary vein embolization. Radiology 2013; 268:900–906. 12. Zheng M, Chen Y, Bai J, et al. Transjugular intrahepatic portosystemic shunt versus endoscopic therapy in the secondary prophylaxis of variceal rebleeding in cirrhotic patients: meta-analysis update. J Clin Gastroenterol 2008; 42:507–516. 13. Haskal ZJ, Martin L, Cardella JF, et al. Quality improvement guidelines for transjugular intrahepatic portosystemic shunts. SCVIR Standards of Practice Committee. J Vasc Interv Radiol 2001; 12:131–136.
