Balloon-Occluded Retrograde Transvenous Obliteration of Gastric Fundal Varices with Hemorrhage

Transcription

1 Mikiya Kitamoto 1,2 Michio Imamura 1 Koji Kamada 1 Hiroshi Aikata 1 Yoshiiku Kawakami 1 Akiko Matsumoto 1 Yoshika Kurihara 1 Hirotaka Kono 1 Hiroo Shirakawa 1 Toshio Nakanishi 1,3 Katsuhide Ito 4 Kazuaki Chayama 1 Received June 13, 2001; accepted after revision November 20, First Department of Internal Medicine, Hiroshima University School of Medicine, Kasumi, Minami-ku, Hiroshima , Japan. 2 Present address: Department of Gastroenterology, Hiroshima Prefectural Hospital Ujina-Kanda, Minami-ku, Hiroshima , Japan. Address correspondence to M. Kitamoto. 3 Department of Clinical Radiology, Hiroshima University School of Medicine, Hiroshima , Japan. 4 Department of Radiology, Hiroshima University School of Medicine, Hiroshima , Japan. AJR 2002;178: X/02/ American Roentgen Ray Society Balloon-Occluded Retrograde Transvenous Obliteration of Gastric Fundal Varices with Hemorrhage OBJECTIVE. This study was performed to evaluate the clinical efficacy, feasibility, and complications of balloon-occluded for patients with hemorrhage from gastric fundal varices. SUBJECTS AND METHODS. Between December 1994 and February 2001, 24 consecutive patients with hemorrhage from gastric fundal varices were enrolled in this study. Balloon-occluded consisted of injecting 5% ethanolamine oleate iopamidol through the outflow vessels during balloon occlusion. The treatment was performed during acute bleeding in 11 patients and electively in 13 patients. Among those patients with acute bleeding, six were treated for temporary hemostasis with balloon tamponade, and five were treated endoscopically. RESULTS. Cannulation into the outflow vessels was performed in 23 patients, but the balloon catheter could not be inserted in one patient who had inferior phrenic vein outflow. Complete success was obtained in 88% (21/24) of patients, and partial success was obtained in two patients. In nine of 11 patients with acute bleeding, complete success was achieved. Rebleeding from gastric varices was not observed in patients treated with complete success, whereas two patients treated partially rebled within 1 week of the treatment (rate of rebleeding, 9%). Eradication of gastric varices was obtained in all patients (n = 19) who were examined by endoscopy 3 months after the treatment. Eight patients experienced worsening of esophageal varices. These patients were treated endoscopically because of findings that suggested a risk of hemorrhage. The overall mortality rate was 4% (1/24). No damage to the kidney was observed, although 11 patients had macrohematuria. CONCLUSION. Balloon-occluded followed by any hemostatic procedure might be effective for both prophylaxis of rebleeding and eradication of gastric fundal varices, even in urgent cases. O nce gastric fundal varices bleed, the mortality rate ranges from 25% to 55% [1 4]. Because of poor underlying liver function and fast blood flow, effective treatment of fundal varices is difficult. Although various therapeutic approaches to gastric fundal varices are available, each treatment has its disadvantages. Among the interventional procedures, percutaneous transhepatic obliteration and transjugular intrahepatic portosystemic shunt do not always provide a favorable clinical outcome [5 7]. Balloon-occluded has been safely performed for gastric fundal varices or ectopic varices with almost complete eradication [8 13]. However, in most published series, patients were treated prophylactically rather than at the time of bleeding. To our knowledge, the usefulness of balloon-occluded retrograde transvenous obliteration for hemorrhage from gastric fundal varices has not yet been sufficiently shown [14]. The purpose of our study was to evaluate the clinical significance of balloon-occluded for patients with hemorrhage from gastric varices with special reference to the treatment s clinical efficacy, feasibility, and complications. Subjects and Methods Patients Twenty-four consecutive patients with gastric fundal varices were enrolled in this study at Hiroshima University Hospital between December 1994 and March 2001 (Fig. 1). In 11 patients, endoscopy revealed active bleeding (spurting or oozing) from gastric fundal varices or showed signs of recent bleeding, such as a clot or fibrin clot. In the remaining 13 patients, gastric fundal varices were considered to be the source of bleeding because no other source could be identified in the presence of distinct and large gastric varices (Fig. 2A). Informed consent was obtained from all patients at the time of enrollment. The characteristics of the patients are shown in Table 1. The study included 18 men and six women with AJR:178, May

2 1168 AJR:178, May 2002 TABLE 1 Patient Characteristics of Patients and Details of Therapeutic Results Sex (Male/ Female) Age Underlying Disease Child s Grade a Temporary Hemostasis Drainage Route Size of Balloon (mm) Note. EOI = ehanolamine oleate iopamidol, EV = esophageal varices, LC = liver cirrhosis, ET = endoscopic treatment, GR = gastorenal shunt, IPV = inferior phrenic vein, HCC = hepatocellular carcinoma, BT = balloon tamponade, EHO = extrahepatic portal vein obstruction, NE = not evaluated, PCPV = pericardiophrenic vein. a Child s classification system [15]. b Criteria for difficulty of according to retrograde venography under balloon occlusion [13]. c Considered risky varices, which were determined on endoscopy as large form (enlarge tortuous or large coil-shaped) or red spot. Hirota s Dose of 5% Grade b EOI (ml) Treatment of Collaterals Eradication of Gastric Varices Worsening of EV c (months) Result (months) Cause of Death 1 M 57 LC C ET GR Complete +3 Died (3) Gastric cancer 2 M 52 LC B Spontaneous IPV Complete +25 Living (64) 3 F 48 LC B Spontaneous GR Complete Living (33) 4 M 66 LC with HCC B BT GR Complete Died (3 HCC 5 F 53 LC B Spontaneous GR Complete +18 Living (33) 6 M 24 EHO A Spontaneous and IPV GR Balloons for GR and IPV Complete +3 Living (51) 7 M 67 LC B Spontaneous GR Complete Living (48) 8 F 79 LC B Spontaneous IPV NE Failure Died (12) Bleeding 9 M 45 LC with HCC B BT GR Complete +11 Died (14) Liver failure 10 F 67 LC B BT GR Coils for PCPV Complete Living (27) 11 M 60 LC A Spontaneous GR Complete Died (10) Pneumonia 12 M 51 LC A Spontaneous GR Coils for IPV Complete Living (18) 13 M 26 LC C BT GR Complete +6 Living (34) 14 M 72 LC B Spontaneous GR Complete Living (18) 15 M 41 LC B ET GR Complete Living (26) 16 M 68 LC A Spontaneous GR Complete Living (30) 17 M 69 LC with HCC B Spontaneous GR Coils for IPV, intercostal vein Complete Living (11) 18 M 66 LC A Spontaneous GR Complete +18 Living (18) 19 M 49 LC A ET GR Complete Living (12) 20 M 56 LC A Spontaneous GR Coils for IPV Complete +7 Living (7) 21 M 47 LC A BT GR Partial Living (7) 22 F 71 LC with HCC A ET GR Coils for IPV Complete Living (5) 23 F 80 LC C BT GR Coils for IPV, Partial NE Died (1 day) Liver failure intercostal vein 24 M 52 LC B ET GR Coils for intercostal vein Complete Living (3) Kitamoto et al.

3 Transvenous Obliteration of Gastric Fundal Varices a mean age of 57 years. Among 23 patients with liver cirrhosis, the cause was viral liver cirrhosis in 13 patients (two patients were positive for hepatitis B surface antigen, and 11 patients were positive for antihepatitis C virus antibody), alcoholic liver cirrhosis in eight, primary biliary cirrhosis in one, and unknown cause in one (negative for viral markers). Four patients had complicating hepatocellular carcinoma with no tumor thrombosis in the main portal vein. Using Child s classification of esophageal varices [15], we categorized eight patients as grade A, 13 as grade B, and three as grade C. Serum creatinine levels varied from 0.51 to 1.4 mg/dl (average, 0.78 mg/dl). Endoscopic findings for varices were evaluated according to the general rules proposed by the Japanese Research Society for Portal Hypertension [16]. The location of gastric varices (Lg) was classified as adjacent to the cardiac ring (Lg-c), separated from the cardiac ring (Lg-f), and continuing from the cardiac ring to the gastric fundus (Lg-cf). The location of gastric varices was Lg-f in 22 patients and Lg-cf in the remaining two patients. The form of the varices was classified as small straight, enlarged tortuous, or large coil-shaped. The size of gastric fundal varices was enlarged tortuous in 11 patients and large coil-shaped in the remaining 13 patients. The size of esophageal varices was small straight in 12 patients, enlarged tortuous in three, and large coil-shaped in one. In the remaining eight patients, no esophageal varices were noted. The clinical courses of patients from hospital admission to balloon-occluded retrograde transvenous obliteration and the details of the therapeutic results are shown in Figure 1. Seven patients had sustained hypovolemic shock on admission. They were initially treated with IV fluid administration, blood transfusion, and balloon tamponade and underwent endoscopic examination after improvement of their general condition. Eleven patients with bleeding signs were treated for temporary hemostasis with balloon tamponade (n = 6) and endoscopically (n = Fig. 1. Schematic diagram shows clinical courses and results of patients with bleeding gastric fundal varices. Hypovolemic shock, present at admission, was defined as systolic blood pressure less than 100 mm Hg and pulse rate greater than 100 beats per minute. Bleeding involved presence of active bleeding (spurting or oozing) from gastric varices or signs of recent bleeding such as clot or fibrin clot. (+) = positive, ( ) = negative. 5; two were treated with endoscopic injection sclerotherapy and three with endoscopic variceal ligation). Then all patients underwent contrast-enhanced CT to determine the presence of a gastrorenal shunt and, thereafter, to evaluate balloon-occluded retrograde transvenous obliteration under temporary hemostasis. In six patients, the presence of a gastrorenal shunt had already been confirmed by CT before hemorrhage from gastric varices. These patients were rapidly treated with balloon-occluded retrograde transvenous obliteration. Balloon-Occluded Retrograde Transvenous Obliteration Selective angiography of the celiac and superior mesenteric arteries was performed before balloon-occluded to evaluate portosystemic collaterals. Balloon-occluded retrograde transvenous obliteration was performed by injecting 5% ethanolamine oleate iopamidol (Oldamin; Grelan Pharmaceutical, Tokyo, Japan) through the outflow vessels during balloon occlusion [8 14]. A 6.5-French balloon catheter (Artec Balloon Catheter, B-RTO type I SML or B-RTO type II SML; Create Medic, Tokyo, Japan) was inserted into the inferior vena cava through the right femoral vein. The catheter was advanced into the outflow vessels, such as the gastrorenal shunt or inferior phrenic vein. On cannulation into the gastrorenal shunt, the balloon catheter was inserted into the right atrium over a guidewire (Fig. 3A). After pulling out the guidewire, we reshaped the catheter by pushing on the interatrial septum (Fig. 3B). Clockwise rotation of the balloon catheter was performed until the tip of the catheter was twisted around the catheter itself (Fig. 3C). Then the balloon catheter was placed in the inferior vena cava, and the tip of the catheter was released by the insertion of a guidewire (Fig. 3D). The balloon catheter was inserted into the left renal vein cannulating the gastrorenal shunt (Fig. 3E). Blood flow directed into the inferior vena cava was blocked by inflating the balloon, and retrograde venography was conducted to determine the hemodynamics of gastric fundal varices and collaterals (Fig. 2C). Thereafter, ml of 5% ethanolamine oleate iopamidol was intermittently injected into the gastric varices through the inflated balloon-occlusion catheter under fluoroscopy. When gastric varices were not visible or only partially opacified because blood flow was stealing through other collaterals, it was necessary to obliterate those vessels [9, 13, 14]. In patients in whom other collaterals were obviously present, the vessels should be obliterated with embolic coils. However, in patients with small vessels or in patients who could not be treated with coils, the use of 50% glucose or ethanol was indicated. These pretreatment occlusions of collaterals could limit the dosage of 5% ethanolamine oleate iopamidol. With or without these interventions, a microcatheter was introduced through a balloon catheter to the gastric varices, and selective administration of 5% ethanolamine oleate iopamidol was performed during balloon occlusion. When varices and inflow vessels could be shown in their entirety, injection was suspended (Fig. 2D). To avoid incomplete therapeutic efficacy and pulmonary infarction due to an unstable thrombus, we left the catheter in the vein with the balloon inflated for approximately 20 hr and removed it after retrograde venography (Fig. 2E). Human haptoglobin was prophylactically administered systemically IV before infusion of 5% ethanolamine oleate iopamidol. The therapeutic effects were evaluated by contrast-enhanced CT approximately 1 week after the treatment (Fig. 2F). Two radiologists interpreted the angiograms and retrograde venograms. Mortality was defined as any death occurring within 6 weeks after admission [17]. Statistical Methods The differences between the values of renal function and hepatic vein wedge pressure before and after balloon-occluded were analyzed using the Student s t test. The incidence of embolic coils and the dose of 5% ethanolamine oleate iopamidol in relation to Hirota s classification [13] were statistically analyzed. The statistical difference was evaluated as significant when the p value was less than 0.05 in all analyses. Results The outflow vessels of gastric fundal varices were gastrorenal shunts in 21 patients, inferior phrenic veins in two, and both collaterals in one. Although cannulation into the outflow vessels was performed in 23 patients, we could not insert the balloon catheter in one patient who had an inferior phrenic vein outflow. A representative clinical course of balloon-occluded is shown in Figure 2. According to criteria for the difficulty of retrograde transvenous obliteration proposed by Hirota et al. [13], in which the degree of progression of gastric varices and collaterals was classified into five grades on the basis of the results of retrograde AJR:178, May

4 Kitamoto et al. out the use of embolic coils, and the average dose of 5% ethanolamine oleate iopamidol was 24 ml. In six grade II varices, embolic coils were necessary in one patient, and the average dose of the sclerosant was 18 ml. In six grade III varices, embolic coils were needed in two patients, and the venography during balloon occlusion, four patients were classified as grade I, six as grade II, seven as grade III, and six as grade IV in the present study. With respect to patients who were treated completely, all four patients whose varices were considered to be grade I were treated with- A average dose of the sclerosant was 29 ml. In five grade IV varices, embolic coils were needed in three patients, and the average dose of the sclerosant was 36 ml. Other collaterals treated by embolic coils were the inferior phrenic vein (n = 4), the pericardio- B D E C F Fig year-old woman with alcohol-related liver cirrhosis and gastric fundal varices. Gastric fundal varices with acute bleeding were successfully treated with balloon-occluded. A, Endoscopic image at admission reveals huge gastric fundal varices with presence of fresh blood. B, Enhanced CT scan reveals large gastric fundal varices (arrow). Arrowheads indicate balloon tamponade inserted into gastric fundus. C, Retrograde venogram obtained during balloon occlusion reveals pericardiophrenic vein (small arrow), but gastric varices (large arrow) are only partially observed and disappeared soon. D, Venogram shows retrograde obliteration. Gastric varices were completely obliterated by 5% ethanolamine oleate iopamidol injected in retrograde manner during balloon occlusion. After insertion of embolic coils into pericardiophrenic vein (small arrow), gastric varices (large arrow) were delineated by 5% ethanolamine oleate iopamidol. Subsequently, short gastric vein (small arrowhead) and posterior gastric vein (large arrowhead) were retrogradely opacified. E, Retrograde venogram obtained 1 day after treatment reveals almost completely obliterated gastrorenal shunt (large arrow) and no evidence of gastric varices. Small arrow indicates embolic coils inserted into pericardiophrenic vein. F, Contrast-enhanced CT scan obtained 10 days after procedure reveals gastric varices with low attenuation (arrow), suggesting complete obliteration. G, Endoscopic image obtained 3 months after treatment reveals eradication of gastric varices. G 1170 AJR:178, May 2002

5 Transvenous Obliteration of Gastric Fundal Varices phrenic vein (n = 1), and the intercostal vein (n = 2). One patient was successfully treated with occlusion of both the gastrorenal shunt and the inferior phrenic vein by inflating two balloons. Contrast-enhanced CT that was performed 1 week after treatment revealed complete thrombosis of gastric varices in 21 patients (Fig. 2F). These results showed good feasibility of the procedure. We could not place the embolic coils into other collaterals in two patients classified as grade IV, and only partial thrombosis was obtained with the use of 50% glucose and ethanol delivered through a microcatheter. These two partially treated patients could not receive further treatment because of the retrograde appearance of the left gastric vein or subintimal injury of the gastrorenal shunt. The reason for partial treatment in these two patients might be insufficient examination of portal collaterals because angiography was not performed as a result of the patients urgent condition. In the completely treated patients, the average dose of 5% ethanolamine oleate iopamidol was 25.7 ml. Although the use of embolic coils (p = 0.03) was related to the criteria proposed by Hirota et al. [13], the dose of 5% ethanolamine oleate iopamidol administered was not associated with the criteria. In the two completely treated patients mentioned previously, we successfully used a A Fig. 3 Illustrations show reshaping of balloon catheter. A, Balloon catheter was inserted into right atrium over guidewire. SVC = superior vena cava, IVC = inferior vena cava, RA = right atrium, RV = right ventricle. B, After guidewire was pulled out, catheter was reshaped by pushing on interatrial septum. C, Clockwise rotation of balloon catheter was performed until tip of catheter was twisted around catheter itself. D, Balloon catheter was placed in inferior vena cava, and tip of catheter was released by insertion of guidewire. E, Balloon catheter was inserted into left renal vein, cannulating gastrorenal shunt. D B large amount of the sclerosant without severe complications. Pain in the epigastrium or in the back was observed in all patients, and fever was observed in six patients. These minor complications persisted for 2 days except in three patients who required an analgesic over 3 days. Macrohematuria was observed in 11 patients. One day after the procedure, a relatively high increase in serum lactic dehydrogenase occurred in all patients (average, 895 U/L), with a level greater than 1000 U/L of serum lactic dehydrogenase observed in one third of the patients, and a slight increase in serum aminotransferase level was observed in all the patients. C E AJR:178, May

6 Kitamoto et al. These results suggest that the occurrence of massive hemolysis was due to 5% ethanolamine oleate. However, the serum lactic dehydrogenase level returned to the baseline values within 2 weeks in almost all patients. The serum creatinine level varied from 0.47 to 1.6 mg/dl (average, 0.8 mg/dl) 1 day after the procedure and did not substantially change. Although a slight increase in the serum creatinine level was observed in three patients, measurements of renal function after the procedure returned to baseline values within a week in all patients. One patient developed ascites 3 days after the treatment, but the condition was successfully treated with furosemide. We measured hepatic vein wedge pressure as portal pressure in five patients before and after the procedure. The hepatic vein wedge pressure before the procedure varied from 260 to 400 mm H 2 O (average, 310 mm H 2 O), and the hepatic vein wedge pressure after the procedure varied from 240 to 470 mm H 2 O (average, 334 mm H 2 O). The differences between the values before and after the procedure were not significant. Rebleeding from gastric varices was not observed in patients who were treated with complete success, whereas two partially treated patients rebled within 1 week after the treatment (rate of rebleeding, 9%). Complete eradication of gastric varices was obtained in all the patients (n = 19) who were examined by endoscopy 3 months after the treatment (Fig. 2G). During the follow-up period (range, 3 64 months; mean, 21.2 months), no patients experienced bleeding from gastric varices, and reappearance of gastric varices was not observed by endoscopic examination in any patient. In patients who were treated with complete success, no deaths occured as a result of bleeding gastric varices, although four patients died of other diseases (Table 1). Among three patients who were treated unsuccessfully, one patient underwent Hassab s devascularization [18] 9 days after partial treatment, one patient with a Child s grade C classification and rebleeding shortly after the treatment died of liver failure, and one patient died of rebleeding from gastric fundal varices 1 year after the trial. The overall mortality rate was only 4% (1/24). Worsening of esophageal varices was determined by endoscopic examination when the form of varices became enlarged tortuous or large coil-shaped or when a red spot was observed. In eight of 17 patients who had esophageal varices before the treatment, worsening of esophageal varices was observed. Two patients were prophylactically treated by endoscopy. In four patients, the worsened esophageal varices ruptured and were treated endoscopically. New esophageal varices did not appear in any patient who did not have esophageal varices before the treatment. No patient died from worsened esophageal varices. Discussion The usefulness of balloon-occluded retrograde transvenous obliteration has been shown mainly by Japanese investigators [8 14]. In 1984, Olson et al. [19] first reported a case of gastric varices treated with retrograde obliteration through a gastrorenal shunt during balloon occlusion. However, those investigators did not show the disappearance of gastric varices, and the procedure could not reveal the obliterated vessels because of the use of ethanol as a sclerosant. On the other hand, Kanagawa et al. [8] successfully obliterated gastric varices using 5% ethanolamine oleate iopamidol in a similar manner. With respect to sclerosants, ethanol would be ideal because of its strong ability to damage venous endothelium, but its distribution cannot be monitored. On the other hand, ethanolamine oleate can be mixed with contrast medium, but it damages venous endothelium. Damage of venous endothelium caused by ethanolamine oleate is reported to induce platelet activation and thrombosis at least 6 hr after IV administration of a sclerosant. Thrombosis continues up to 1 month after treatment [20]. Fifty percent glucose or 1% polidocanol has only a weak effect on the endothelium. Thus, we used 5% ethanolamine oleate iopamidol for the procedure and obtained successful results [8 14]. Moreover, because ethanolamine oleate promptly binds with albumin in the blood with loss of activity, the sclerosant is ideal for local obliteration of vessels [21]. However, ethanolamine oleate related complications have been reported, including severe renal dysfunction [22], pulmonary edema [23], disseminated intravascular coagulation [24], and anaphylactic reaction [13, 25]. Among these complications, renal dysfunction is not rare. As a result of intravascular hemolysis due to ethanolamine oleate, free hemoglobin is circulated and excreted by glomeruli, and the reabsorption of an excess amount of free hemoglobin may cause renal tubular dysfunction [22]. To avoid renal dysfunction, administration of haptoglobin before the procedure and a limited dose of ethanolamine oleate are necessary. Haptoglobin conjugates hemoglobin and may prevent renal tubular dysfunction [22]. With endoscopic sclerotherapy, the safe dose of ethanolamine oleate is reported to be less than 0.5 ml/kg in animal models [26] and less than 0.6 ml/kg in humans [27]. In the present study, no damage to the kidney was observed after the administration of haptoglobin, although half the treated patients had macrohematuria. Safety, and the complications our patients experienced, were similar to those reported previously. In the treatment for bleeding gastric fundal varices, initial hemostasis and prevention of secondary bleeding are most important. Bleeding gastric varices are often treated with the administration of vasopressin and the use of balloon tamponade, but these treatments provide only a transient hemostatic effect. Although endoscopic sclerotherapy carries a considerable risk of prolonged bleeding from gastric ulceration at the injection site [1], endoscopic sclerotherapy using cyanoacrylate results in a more durable hemostasis [28, 29]. However, the use of cyanoacrylate is complicated. Moreover, endoscopic sclerotherapy using cyanoacrylate does not always successfully eradicate gastric varices, and it may cause unexpected complications such as cerebral infarction [30]. Surgical treatments such as Hassab s devascularization and transection have a limited role in the treatment of gastric varices because of hepatic dysfunction [18, 31]. Percutaneous transhepatic obliteration could obtain hemostasis, but rebleeding may occur [5]. It is likely that the antegrade procedure could obliterate inflow vessels but could not easily obliterate varices completely. The formation of new collateral inflow vessels may cause rebleeding from nonembolized varices. After the placement of a transjugular intrahepatic portosystemic shunt for gastric variceal hemorrhage, the rate of hemostasis is greater than 90%, and the incidence of rebleeding is approximately 30% [6, 7]. Transjugular intrahepatic portosystemic shunts may achieve hemostasis, but rebleeding rates are not satisfactory. Moreover, the procedure may induce hepatic encephalopathy or chronic hepatic failure [7]. On the other hand, balloon-occluded retrograde transvenous obliteration, as presented here, shows a high rate of hemostasis and a low rate of rebleeding. These findings would indicate that the retrograde procedure could easily obliterate whole varices in comparison with antegrade obliteration. Indeed, many reports, including our series, showed that the rate of eradication for gastric varices is approximately 90% [8 10, 12]. Subsequently, portal pressure would increase because outflow vessels such as a gastrorenal 1172 AJR:178, May 2002

7 Transvenous Obliteration of Gastric Fundal Varices shunt or an inferior phrenic vein were obliterated. Although balloon-occluded retrograde transvenous obliteration could worsen preexisting esophageal varices, patients who have no small varices have not developed worsened esophageal varices after the procedure. Thus, patients with small varices before retrograde obliteration must be carefully followed up with endoscopy. The disadvantages associated with retrograde transvenous obliteration include an increase in portal pressure and the difficulty performing the procedure. It is likely that retrograde transvenous obliteration increases portal pressure because of obliteration of collateral veins connecting the portal venous system and systemic circulation. Indeed, portal pressure was reported to rise in all (seven) patients at least 1 week after the treatment [32], although in the present study, no difference was seen in the value of wedged hepatic venous pressure immediately before and 1 day after the treatment. Many reports, including our series, showed a worsening of esophageal varices after the treatment [8 10, 12, 13]. However, the worsening of esophageal varices is offset by improvements in liver function and because portosystemic encephalopathy is well controlled after [13, 32, 33]. It is likely that the portal vein pressure of patients with worsened esophageal varices after the treatment may be high, because we studied patients with bleeding episodes from gastric varices. Moreover, the worsened esophageal varices could be successfully treated endoscopically. Hirota et al. [13] showed the criteria for the difficulty of and that it is necessary to occlude small collaterals with an injection of ethanol or with embolic coils in half of the patients. Although we could achieve successful results using 50% glucose, ethanol, and embolic coils delivered through a microcatheter, we also experienced two technical failures. Moreover, it is not easy to perform this procedure without the presence of a shunt. In patients with inferior phrenic veins acting as outflow vessels, it was difficult to perform. We obtained a successful result in one patient. Because the prevalence of a shunt is reported to be approximately 90% in gastric fundal varices, such cases may be rare [34, 35]. Thus, the results presented here indicate that balloon-occluded of gastric varices is feasible. Treatment of varices is focused on decreasing the risk of a first variceal hemorrhage, increasing survival from the acute bleeding, and preventing rebleeding. Balloon-occluded may perform all three functions. It remains unclear whether this procedure should be used prophylactically because of the high rate of worsened esophageal varices. In the present study, balloon-occluded retrograde transvenous obliteration proved feasible with low mortality and almost complete eradication of gastric fundal varices even in urgent conditions, although this aspect of the procedure was limited by the small number of patients. No panacea exists for portal hypertension except liver transplantation. For treatment of patients with hemorrhage from gastric varices, the usefulness of retrograde transvenous obliteration might overcome the risk of worsening of esophageal varices. Thus, balloon-occluded retrograde transvenous obliteration followed by any hemostatic procedure might be useful for gastric fundal varices even in urgent conditions. Acknowledgments We thank the radiographic staff for their cooperation and the nursing staff for their treatment of the patients. References 1. Trudeau W, Prindiville T. Endoscopic injection sclerosis in bleeding gastric varices. Gastrointest Endosc 1986;32: Sarin SK, Sachdev G, Nanda R, Misra SP, Broor SL. Endoscopic sclerotherapy in the treatment of gastric varices. Br J Surg 1988;75: Sarin SK, Lahoti D, Saxena SP, Murthy NS, Makwana UK. Prevalence, classification and natural history of gastric varices: a long-term follow-up study in 568 portal hypertension patients. Hepatology 1992;16: Kim T, Shijo H, Kokawa H, et al. Risk factors for hemorrhage from gastric fundal varices. Hepatology 1997;25: L Hermine C, Chastanet P, Delemazure O, Bonniere PL, Durieu JP, Paris JC. Percutaneous transhepatic embolization of gastroesophageal varices: results in 400 patients. AJR 1989;152: Chau TN, Patch D, Chan YW, Nagral A, Dick R, Burroughs AK. Salvage transjugular intrahepatic portosystemic shunts: gastric fundal compared with esophageal variceal bleeding. Gastroenterology 1998;114: Barange K, Peron JM, Imani K, et al. Transjugular intrahepatic portosystemic shunt in the treatment of refractory bleeding from ruptured gastric varices. Hepatology 1999;30: Kanagawa H, Mima S, Kouyama H, Gotoh K, Uchida T, Okuda K. Treatment of gastric fundal varices by balloon-occluded retrograde transvenous obliteration. J Gastroenterol Hepatol 1996;11: Chikamori F, Shibuya S, Takase Y, Ozaki A, Fukao K. Transjugular retrograde obliteration for gastric varices. Abdom Imaging 1996;21: Koito K, Namieno T, Nagakawa T, Morita K. Balloon-occluded for gastric varices with gastrorenal or gastrocaval collaterals. AJR 1996;167: Haruta I, Isobe Y, Ueno E, et al. Balloon-occluded (BRTO), a promising nonsurgical therapy for ectopic varices: a case report of successful treatment of duodenal varices by BRTO. Am J Gastroenterol 1996;91: Sonomura T, Sato M, Kishi K, et al. Balloon-occluded for gastric varices: a feasibility study. Cardiovasc Intervent Radiol 1998;21: Hirota S, Matsumoto S, Tomita M, Sako M, Kono M. Retrograde transvenous obliteration of gastric varices. Radiology 1999;211: Chikamori F, Kuniyoshi N, Shibuya S, Takase Y. Urgent transjugular retrograde obliteration for prophylaxis of rebleeding from gastric varices in patients with a spontaneous portosplenorenal shunt. Dig Surg 2000;17: Child CG, Turcotte JG. Surgery in portal hypertension. In: Child CG, ed. Major problems in clinical surgery: the liver and portal hypertension. Philadelphia: Saunders, 1964: Idezuki Y. General rules for recording endoscopic findings of esophagogastric varices. World J Surg 1995;19: De Franchis R, Rascal JP, Ancona E, et al. Definitions, methodology and therapeutic strategies in portal hypertension: a consensus development workshop. J Hepatol 1992;15: Hassab MA. Gastroesophageal decongestion and splenectomy in the treatment of esophageal varices in bilharzial cirrhosis: further studies with a report on 355 operations. Surgery 1967;61: Olson E, Yune HY, Klatte EC. Transrenal-vein reflux ethanol sclerosis of gastroesophageal varices. AJR 1984:143: Evans DMD, Jones DB, Cleary BK, Smith PM. Oesophageal varices treated by sclerotherapy: a histopathological study. Gut 1982;23: Hoak JC, Warner ED, Connor WE. Platelets, fatty acids and thrombosis. Circ Res 1967;20: Hashizume M, Kitano S, Yamaga H, Sugimachi K. Haptoglobin to protect against renal damage from ethanolamine oleate sclerosant. Lancet 1988;2: Lee JY, Moon SH, Lee SM, et al. A case of noncardiogenic pulmonary edema by ethanolamine oleate. Korean J Intern Med 1994;9: Bellary SV, Isaacs P. Disseminated intravascular coagulation (DIC) after endoscopic injection sclerotherapy with ethanolamine oleate. (letter) Endoscopy 1990;22: Bordas JM, Feu F, Vilella A, Rodes J. Anaphylactic reaction to ethanolamine oleate injection in sclerotherapy of esophageal varices. (letter) Endoscopy 1989;21: Iso Y, Kitano S, Iwanaga T, Hashizume M, Sugima- AJR:178, May

8 Kitamoto et al. chi K. Morphological and hemodynamic changes of the lung after injection of 5% ethanolamin oleate into dogs. Eur Surg Res 1989;21: Takase Y, Kikuchi M, Ozaki A, Shibuya S, Iwasaki Y. Pathological studies on esophageal varices treated with injection sclerotherapy. Jpn J Surg 1985;15: Soehendra N, Grimm H, Nam VCh, Berger B. N- butyl-2-cyanoacrylate: a supplement to endoscopic sclerotherapy. Endoscopy 1987;19: Oho K, Iwao T, Sumino M, Toyonaga A, Tanikawa K. Ethanolamine oleate versus butyl cyanoacrylate for bleeding gastric varices: a nonrandomized study. Endoscopy 1995;27: See A, Florent C, Lamy P, Levy VG, Bouvry M. Cerebrovascular accidents after endoscopic obturation of esophageal varices with isobutyl-2-cyanoacrylate: report of 2 patients [in French]. Gastroenterol Clin Biol 1986;10: George P, Brown C, Ridgway G, Crofts B, Sherlock S. Emergency oesophageal transection in uncontrolled variceal haemorrhage. Br J Surg 1973; 60: Akahane T, Iwasaki T, Kobayasi N, et al. Changes in liver function parameters after occlusion of gastrorenal shunts with balloon-occluded retrograde transvenous obliteration. Am J Gastroenterol 1997;92: Kawanaka H, Ohta M, Hashizume M, et al. Portosystemic encephalopathy treated with balloonoccluded. Am J Gastroenterol 1995;90: Watanabe K, Kimura K, Matsutani S, Ohto M, Okuda K. Portal hemodynamics in patients with gastric varices: a study in 230 patients with esophageal and/or gastric varices using portal vein catheterization. Gastroenterology 1988;95: Matsumoto A, Kitamoto M, Imamura M, et al. Three-dimensional portography using multislice helical CT is clinically useful for management of gastric fundic varices. AJR 2001;176: The full text and images from the American Journal of Roentgenology may also be viewed online at or AJR:178, May 2002