Endovascular Stent Placement for Interposed Middle Hepatic Vein Graft Occlusion after Living-Donor Liver Transplantation Using Right- Lobe Graft

Transcription

1 LIVER TRANSPLANTATION 12: , 2006 ORIGINAL ARTICLE Endovascular Stent Placement for Interposed Middle Hepatic Vein Graft Occlusion after Living-Donor Liver Transplantation Using Right- Lobe Graft Ji Hoon Shin, 1 Kyu-Bo Sung, 1 Hyun-Ki Yoon, 1 Gi-Young Ko, 1 Kyoung Won Kim, 1 Sung-Gyu Lee, 2 Shin Hwang, 2 Chul-Soo Ahn, 2 Ki-Hun Kim, 2 Deok-Bog Moon, 2 Ho-Young Song, 1 and Tae-Yong Ha 2 Departments of 1 Radiology and Research Institute of Radiology and 2 Surgery, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea Middle hepatic vein (MHV) reconstruction is performed to drain the right paramedian sector to prevent hepatic venous congestion (HVC). The aim of the present study was to evaluate endovascular stent placement in patients with stenosed and/or occluded interposition vein graft (IVG) to segment V hepatic vein (V 5 ) and segment VIII hepatic vein (V 8 ) after living-donor liver transplantation (LDLT). The procedure was performed in 11 recipients; 7 underwent it within 24 hours of LDLT. The following parameters, including technical success, clinical success, complications, patient survival data, and serial computed tomography (CT) findings during follow-up, were documented retrospectively. Technical success was defined as both successful stent placement and resolution of stenosis or occlusion with copious flow of contrast medium through the stent, while clinical success was defined as both improvement of liver function tests (LFTs) and reduction or disappearance of hepatic low-attenuation areas on follow-up CT scans taken within 1 week of stent placement. Technical success was achieved in 10 of 11 patients (91%), and clinical success was achieved in 9 of 11 patients (82%). Acute thrombotic occlusion of the stent-inserted hepatic vein occurred in 1 patient 1 day following stent placement. During the mean follow-up period of 468 days (range, days), 9 patients survived and 2 patients died. No death was directly related to stent placement or its related complications. The low-attenuation area in the involved hepatic segment V (S 5 ) and/or VIII (S 8 ) area prior to stent placement disappeared completely on follow-up CT scans performed at 3-12 days (mean, 5.4 days) after stent placement in all 9 patients with clinical success. No attenuation change occurred even in cases with chronic occlusion of the stent-inserted hepatic veins. In conclusion, though IVG to V 5 and V 8 remains controversial, the treatment of their stenosis or occlusion is safe and effective, even during their immediate postoperative period. Liver Transpl 12: , AASLD. Recieved March 20, 2005; accepted August 11, Hepatic venous complications are not common following whole liver transplantation because direct anastomosis of the hepatic vein is usually not involved. On the other hand, it is mandatory in living-donor liver transplantation (LDLT) to perform hepatic vein anastomosis. In many of the right-lobe recipients, hepatic venous congestion (HVC), which was caused by deprivation of middle hepatic vein (MHV) drainage, has been tolerated, but there has been much debate on the natural resolution of HVC as well as the need for MHV reconstruction. 1-3 MHV reconstruction is performed by connecting the segment V hepatic vein (V 5 ) and/or segment VIII hepatic vein (V 8 ) to the recipient s hepatic vein stump or inferior vena cava (IVC) with the interposition vein graft (IVG). Recently, many surgeons have agreed on the need for MHV reconstruction in order to prevent HVC. 1,4,5 The adverse aspect of MHV reconstruction using IVG Abbreviations: MHV, middle hepatic vein; HVC, hepatic venous congestion; IVG, interposition vein graft; V 5, segment V hepatic vein; V 8, segment VIII hepatic vein; LDLT, living-donor liver transplantation; CT, computed tomography; IVC, inferior vena cava; US, ultrasonography; LFT, liver function test; S 5, hepatic segment V; S 8, hepatic segment VIII; HU, Hounsfield unit. Address reprint requests to Kyu-Bo Sung, MD, Department of Radiology, Asan Medical Center, University of Ulsan College of Medicine, 388-1, Pungnap-2dong, Songpa-gu, Seoul , Republic of Korea. Telephone: ; FAX: ; kbsung@amc.seoul.kr DOI /lt Published online in Wiley InterScience ( American Association for the Study of Liver Diseases.

2 270 SHIN ET AL. is IVG stenosis or occlusion during the follow-up. 6 In liver transplants, percutaneous balloon angioplasty or stent placement has been advocated widely as the initial treatment for many transplant-related anastomotic abnormalities The purpose of this study was to evaluate the effectiveness and safety of endovascular stent placement for the treatment of stenosed and/or occluded IVG to V 5 and V 8 after LDLT. PATIENTS AND METHODS Patient Population Written informed consent was obtained for endovascular stent placement from all patients. From August 2002 through July 2004, 370 patients who had undergone liver transplantation were initially included (31 patients with cadaveric donor liver transplantation and 339 patients with LDLT). Of 339 patients with LDLT, 305 patients (90%) underwent LDLT with MHV reconstruction. Of the 305 patients with LDLT with MHV reconstruction, HVC was suggested on the follow-up Doppler ultrasonography (US) and computed tomography (CT) scans in 30 patients. Twenty-two patients did not undergo hepatic venogram/stent placement, and they were followed up with liver function tests (LFTs) and imaging studies (Doppler US and CT scans) because their graft-to-recipient weight ratio was more than 1, and their congestion volume was estimated as less than 30% of the graft volume. Among them, 19 patients recovered uneventfully, whereas 3 patients underwent stent placement due to an increase of congestion volume. The remaining 8 patients with small-for-size grafts (graft-to-recipient weight ratio 1) or congestion volume more than 30% of the graft volume underwent primary stent placement for the correction of IVG stenosis and/or occlusion. There was no patient with IVG stenosis and/or occlusion who received other treatments, such as percutaneous balloon angioplasty or surgical revision during the study period. There were 8 male and 3 female patients; age ranged from 30 to 61 yr (median, 49 yr). Ten of these patients had received right-lobe graft, and the remaining 1 patient received dual grafts (1 right lobe and 1 left lobe). IVGs using cryopreserved iliac veins, the recipient s saphenous veins, or portal vein were made to bridge the distance between the recipient s hepatic vein stump or IVC and the venous ends (V 5 and V 8 ) at the cut surface of the liver graft. IVG stenosis or occlusion was suggested when Doppler US showed a persistent monophasic wave pattern or slow Doppler flow less than 10 cm/second, or when 2-phase CT scan showed nonopacification of the IVGs and their tributaries, with low-attenuation areas in the right paramedian sector in the clinical setting of HVC. 7,14 Of 11 subjects, 5 patients showed persistent monophasic wave pattern and the remaining 6 patients, slow Doppler flow less than 10 cm/second on Doppler US. A low-attenuation lesion in the right paramedian sector was defined as an area with the relative Hounsfield unit (HU) less than 1.00 on both noncontrast phase and hepatic arterial phase of the 2-phase CT scans. Because low attenuation on hepatic arterial phase is highly reliable, rather than that on portal venous phase, in predicting HVC on 2-phase CT scans, we measured HU on both noncontrast phase and hepatic arterial phase. 15 Relative HU was measured to avoid inaccuracy of the absolute HU value, depending on individuals different contrast dynamics. We determined the HU value by calculating the mean of 3 regions of interest (1 cm 2 ) measurements. The relative HU was calculated by dividing the mean HU of the lowattenuation lesion by the mean HU of the normal liver parenchyma. All 11 patients had a low-attenuation lesion on CT. Liver biopsy was performed on 1 patient (patient 2) and confirmed the centrilobular hepatic congestion without evidence of rejection, compatible with outflow occlusion. Stent Placement The mean ( standard deviation) interval between LDLT and stent placement was 14.9 days 26.9 (range, 0-75 days). Seven patients underwent stent placement within 24 hours after LDLT; the remaining 4 patients, 7 to 75 days after LDLT because of delayed decision following increase of congestion volume during follow-up (n 3) or delayed onset of HVC (n 1). The mean ( standard deviation) interval between the identification of the abnormal CT findings and stent placement was 4.2 days 10.1 (range, 0-34 days). The detailed data are shown in Table 1. In all patients, the right internal jugular vein was punctured under US guidance, and a selective hepatic venogram was performed using a 5-F Cobra catheter to evaluate the abnormality of the IVGs and their anastomoses. A inch angled hydrophilic guide wire (Terumo, Tokyo, Japan) was used to traverse the stenosis or occlusion. Since kinking of redundant IVG or extrinsic compression of the IVG from slight movement of the graft or adjacent structures cannot be resolved by balloon angioplasty in the early posttransplantation periods, primary stent placement was performed using self-expandable metallic stents. Wallstents (Boston Scientific, Galway, Ireland) or Zilver stents (Cook, Bloomington, IN) were used. The stent s size for each case was chosen on the basis of the measured normal hepatic vein diameter. Stents were intentionally oversized by approximately 1-2 mm to minimize the risk of migration. The pressure gradient between hepatic vein and IVC across the IVG was obtained in only 3 patients. Definitions and Study Endpoints All patients were followed up utilizing laboratory data, Doppler US, and 2-phase CT. Doppler US was performed in all patients 1 day after stent placement, weekly during hospital stay, and then 1, 6, and 12 months after discharge from the hospital. Two-phase CT scans were routinely taken once a week during the

3 STENT PLACEMENT FOR MHV IVG OCCLUSION POST-LDLT 271 TABLE 1. Patient Characteristics and Stent Placement for Interposition Vein Graft Stenosis or Occlusion Involved Main Interval- Patency of occluded H.V. Survival data No./Gender/ Age segment on CT stenotic area on venogram (LT-SP) (days) Tech. success Clinical success Patent or not F.U. days on CT Alive or not Days 1/M/47 S 5 and S 8 V 5 -IVG 1 Yes Yes Patent 210 Dead 271 2/M/40 S 5 and S 8 V 5 IVG-IVC 60 Yes Yes Occluded 262 Alive 891 S 6 and S 7 3/F/43 S 5 V 5 -IVG 75 Yes Yes Occluded 355 Alive 877 4/M/47 S 5 and S 8 V 5 -IVG and V 8 -IVG 0 Yes Yes Patent 583 Alive 771 5/M/46 S 5 and S 8 V 5 -IVG and V 8 -IVG 7 Yes Yes Occluded 72 Alive 413 6/F/61 S 5 and S 8 V 5 -IVG and V 8 -IVG 0 Yes No N.A. N.A. Alive 344 7/M/50 S 5 and S 8 V 5 -IVG and V 8 -IVG 1 Yes Yes Patent 220 Alive 265 8/M/55 S 5 and S 8 RHV-IVC, V 5 -IVG, 0 No Yes Patent 207 Alive 308 and V 8 -IVG 9/M/49 S 5 and S 8 V 5 -IVG and V 8 -IVG 0 Yes Yes V 5 -patent V Alive 274 V 8 -occluded V /F/56 S 5 V 5 -IVG 20 Yes No N.A., N.A. Dead 13 11/M/51 S 5 IVG-IVC 0 Yes Yes Occluded 110 Alive 721 Abbreviations: IVG, interposition vein graft; IVC, inferior vena cava; RHV, right hepatic vein; Tech. Success, technical success; H.V., hepatic vein; N.A., not applicable; F.U., follow-up. admission period. However, the schedule and interval of Doppler US and CT scans varied depending upon the patients clinical situations. The following parameters were documented retrospectively: technical success, clinical success, complications, patient survival data, and serial CT findings during follow-up. The technical success was defined as the successful stent placement and resolution of the hepatic venous stenosis or occlusion with copious flow of contrast medium through the stent. The clinical success was defined as when the improvement of LFT (including aspartate aminotransferase, alanine aminotransferase, and total bilirubin) and reduction or disappearance of hepatic low-attenuation areas on follow-up 2-phase CT scans taken within 1 week following stent placement were achieved. The reduction or disappearance of hepatic low-attenuation areas was confirmed by the increased relative HU on both noncontrast phase and hepatic arterial phase. The Wilcoxon signed rank test was used to evaluate the statistical significance of the differences between preand postprocedural aspartate aminotransferase, alanine aminotransferase, and the total bilirubin. The analysis was conducted using SPSS software (version 10, SPSS, Chicago, IL), and a P value lower than 0.05 was considered to be statistically significant. According to the Society of Interventional Radiology reporting standards, we defined major complications as those that required further treatment or more hospitalization, and minor complications as those that resolved spontaneously or with conservative treatment. 16 Survival data and length of follow-up were obtained for each patient through review of the medical records. Serial CT findings were focused on analysis of the change of low-attenuation areas and maintenance or loss of the opacified hepatic veins on follow-up CT scans. RESULTS Technical success was achieved in 10 of 11 patients (91%) (Fig. 1). A total of 20 stents (9 Wallstents and 11 Zilver stents) were used in 11 patients, with use of 1-4 stents for each patient. These stents were 6-10 mm in diameter and 4-8 cm in length. A total of 19/20 stents were used for stenosis and/or occlusion of V 5 or V 8 anastomoses or IVG itself, and the remaining stent was used for the right hepatic vein-ivc anastomosis. The pressure gradients across the IVG before the stent placement were 26, 10, and 6 mmhg, respectively, in the first 3 patients, and they became 6, 3, and 6 mmhg, respectively, after stent placement. The pressure measurement was not obtained in the other patients. A technical failure occurred in 1 patient (patient 8) with right-lobe graft with MHV reconstruction (Fig. 2), and the nonvisualization of the V 5 and V 8 branch and stenosis of the right hepatic vein-ivc, V 5 -IVG, and V 8 - IVG was verified on hepatic venogram. In this patient, stent placement for the right hepatic vein-ivc and V 5 - IVG was successful; however, after passage of the guide wire, the advancement of the stent introducer system into the V 8 failed because of severe twisting and kinking of the IVG. Interestingly, the follow-up CT obtained 5 days later showed resolution of the low-attenuation area with clear visualization of the V 8. We presumed that passage of the guide wire through the kinked stenotic portion caused recanalization of the obstructed venous flow. The clinical success was achieved in 9 of 11 study patients (82%). The mean relative HU of the low-attenuation areas changed from 0.73 (range, ) to

4 272 SHIN ET AL. Figure 1. Images of a 49-yr-old man with IVG occlusion (patient 9). (A and B) Contrast-enhanced CT images obtained on the same day as LDLT show low-attenuation areas (arrowheads) predominantly in the S 5 and S 8 areas. Note a small amount of bilateral pleural effusion (asterisks in A) and hematoma (asterisks in B) in the posterior portion of the graft. (C and D) Hepatic venogram images obtained on the same day as LDLT show kinking and narrowing of the V 5 -IVG (arrows) and V 8 -IVG (arrowheads). Note complete resolution of the hepatic venous outflow occlusion after stent placement. (E and F) Contrast-enhanced CT images obtained 6 days after stent placement (one stent at V 8 -IVG, arrow in E; another stent at V 5 -IVG, arrow in F) show complete resolution of the low-attenuation areas (range, ) on noncontrast phase and from 0.61 (range, ) to 0.95 (range, ) on hepatic arterial phase CT scans. In the remaining 2 patients (patients 6 and 10), despite the improved laboratory data, a decrease of the low-attenuation areas was not observed on follow-up CT scans taken within 1 week of stent placement. There were 2 clinical failures. In 1 patient (patient 6) with probable acute stent thrombosis, stent placement through narrowing of V 5 -IVG and V 8 -IVG was technically successful on the same day as LDLT. However, follow-up Doppler US taken 1, 2, 3, and 6 days following stent placement and CT scans taken 1 and 2 weeks following stent placement showed no blood flow through the stent, with interval atrophy of the involved hepatic segment V (S 5 ) and hepatic segment VIII (S 8 ). In the other patient (patient 10) with probable underlying frank hepatic necrosis, CT scans taken at the time of stent placement showed definite low attenuation, suggesting infarct on S 5. In this patient, the low-attenuation area did not decrease on the CT scans despite the successful stent placement for V 5 -IVG. Laboratory data showed significant improvement 1 week after stent placement (P 0.05) (Fig. 3). The mean aspartate aminotransferase, alanine aminotransferase, and total bilirubin declined from IU/L (range, IU/L) to 71.5 IU/L (range, IU/L), from IU/L (range, IU/L) to IU/L (range, IU/L), and from 5.7 mg/dl (range, mg/dl) to 4.2 mg/dl (range, mg/dl), respectively. In 9 clinically successful patients, 7 with abnormally high LFT before stent placement showed a decrease of all 3 values of LFT after stent placement; the remaining 2 patients with normal LFT before stent placement still showed normal LFT after stent placement. In 2 clinically unsuccessful patients, abnormally high LFTs before stent placement also improved after stent placement. There were no notable complications other than 1 case of acute thrombosis during or immediately after the procedure. Acute thrombotic occlusion of the stentinserted hepatic vein was detected on 1-day follow-up Doppler US in 1 patient (patient 6), which finally caused interval atrophy of the involved liver segments. During the mean follow-up period of 468 days (range, days) following stent placement, 9 patients survived and 2 patients died. In the 9 living patients, the mean follow-up period was 540 days (range, days) to date. The cause of death of the other 2 patients was concurrent leukemia (patient 1) and hepatic artery thrombosis (patient 10). No death was directly related to stent placement or its related complications. Patient 10 proved to have concurrent hepatic artery thrombosis on follow-up CT scans obtained 5 days after stent placement and died 13 days following stent placement and 33 days after LDLT operation.

5 STENT PLACEMENT FOR MHV IVG OCCLUSION POST-LDLT 273 Figure 2. Images of a 55-yr-old man with IVG and right hepatic vein occlusion (patient 8). (A) Contrast-enhanced CT image obtained on the same day as LDLT shows low-attenuation areas in the S 8 area. The V 8 (arrows) is indistinct because of nonopacification. (B and C) Hepatic venogram images obtained on the same day as LDLT show marked kinking of V 8 -IVG and the IVG itself (arrows in B). Traversal of the guide wire through the V 8 was possible (arrows in C); however, passage of the stent introducer system failed due to severe twisting and kinking of the IVG. Instead, stent placement was performed in the V 5 -IVG and the right hepatic vein-ivc (not shown). (D) Contrast-enhanced CT scan obtained 5 days after the hepatic venogram shows enhancement of the V 8 (arrows) and complete resolution of the low-attenuation area. Patency of the V 8 was achieved after traversing of the guide wire despite failure of the stent placement. As for the serial CT findings in the 9 clinically successful patients, the low-attenuation area in the involved S 5 and/or S 8 area before stent placement disappeared completely on the follow-up CT scans performed 3-12 days (mean, 5.4 days) after stent placement (Figs. 1 and 2). The opacification of the nonopacified V 5 and/or V 8 before stent placement was achieved in 8 patients while the opacification of the V 5 even before stent placement was maintained in 1 patient (patient 11). The patency of the opacified V 5 and/or V 8 was maintained until the last CT follow-up, with the mean follow-up period of 305 days (range, days) in 4 patients, while that of the opacified V 5 and/or V 8 was lost on the follow-up CT scans, with the mean follow-up period of 200 days (range, days) in another 4 patients. In the remaining patient, the patency of the opacified V 5 was maintained while that of the opacified V 8 was lost on the follow-up CT scans. As for the serial CT findings in the 2 clinically unsuccessful patients, the low-attenuation area in the involved S 5 and/or S 8 area did not disappear after stent placement, and opacification of the nonopacified V 5

6 274 SHIN ET AL. and/or V 8 before stent placement was not achieved. Interval atrophy of the involved S 5 and S 8 areas occurred in 1 patient (patient 6). Figure 3. Changes of aspartate aminotransferase (A), alanine aminotransferase (B), and total bilirubin (C) after stent placement. PRE and POST indicate values before and 1 week after stent placement, respectively. DISCUSSION LDLT with a right-lobe graft has been advocated for adult recipients with large metabolic demand, as it provides grafts of good quality and size. 3,15 To prevent HVC after LDLT, the large hepatic veins of V 5 and V 8 of the graft are reconstructed using the IVG. However, these veins are occasionally stenosed or occluded because of vessel size mismatching, kinking of the IVG due to its long course, and venous twisting from displacement of the regenerating graft. 6,7 When this complication occurs in the early posttransplantation period, it can cause deterioration of liver functions and can even result in loss of the patient s life. Therefore, early diagnosis and adequate treatment of hepatic venous outflow abnormalities ensure good graft function. Endovascular balloon angioplasty and stent placement have been advocated as the safe and effective initial treatment to manage the hepatic venous outflow abnormalities following LDLT, 7-11 and stent placement has been reported in the treatment for stenosis and/or occlusion of IVG between the recipient s right hepatic vein and the hepatic venous end of the liver graft in cases of dual LDLT. 7 However, there have been no reports on the endovascular treatment for the stenosis and/or occlusion of IVG to V 5 and V 8. To treat IVG stenosis and/or occlusion, we considered primary stent placement instead of balloon angioplasty for several reasons. First, the majority (9/11) of our subjects underwent the procedure within 3 weeks following LDLT, and there were concerns of possible disruption of anastomoses during balloon angioplasty. Second, considering the fact that frequent presumptive etiologies of IVG stenosis and/or occlusion are kinking of a redundant IVG and extrinsic compression from a rotated graft or edema, we assumed that these conditions could not be corrected effectively with balloon angioplasty alone. Third, we believed that primary stent placement could diminish the need for repeated interventional management to preserve fluent hepatic venous outflow. The need for repeat balloon angioplasty or further stent placement was not uncommon in patients who underwent balloon angioplasty. 7,11,13 We think that passage of the guide wire and stent introducer system does not damage or disrupt the IVGs or their anastomoses, even in the immediate posttransplantation period. In a report by Ko et al. 7 regarding 27 patients with LDLT, it was confirmed that hepatic vein stent placement (n 22) or balloon angioplasty (n 5) was safe without any procedure-related complications. Considering the mean interval between LDLT and the procedure was 9.4 weeks (range, 1-66 weeks) in that study, safety related to stent placement less than 1 day after LDLT should be emphasized in this study. Our technical success was achieved in 10 of 11 patients (91%). Considering acute angle or tortuosity of the IVG, we think that this success rate was favorably accept-

7 STENT PLACEMENT FOR MHV IVG OCCLUSION POST-LDLT 275 able. Ko et al. 7 reported 100% of the technical success rate for balloon angioplasty or stent placement in 27 patients with hepatic venous outflow stenosis and/or occlusion in LDLT 7 ; however, only 1/3 of the lesions had IVG connected to the recipient s right hepatic vein stump. In a report by Lee et al., 5 1 patient died of HVC and resultant ischemic necrosis of the right paramedian sector among the 5 patients who underwent right liver grafts without MHV reconstruction. In another report by Moon and Lee, 17 both 6-month graft and patient survival rates have markedly increased from around 70% to around 95% after the adoption of MHV reconstruction in the right-lobe graft and liberal use of largevolume graft. We herein advocate that salvage of IVG to V 5 and V 8 is clinically important, especially in patients with the marginal graft volume and during the immediate posttransplantation period. In this study, after patency of IVG resumed, all 9 patients with clinical success showed eventual recovery from HVC on CT scans and LFT follow-up. Although the 2 patients with the clinical failure showed some improvement of LFT, probably due to collateral circulation through the sinusoids, or short hepatic or portal veins, they did not show improvement of HVC on follow-up CT scans. In comparison with the previous report, reporting 73% of the clinical success for both balloon angioplasty and stent placement, 7 our clinical success rate (82%, 9/11) was also acceptably high, and there was no patient requiring a further intervention. As for the serial CT findings following stent placement, all of the clinically successful patients showed the disappearance of the low-attenuation areas in the involved area within 2 weeks of stent placement; the response on the follow-up CT scans seems to be reliable in reflecting the presence or absence of HVC. During the follow-up of these 9 clinically successful patients, the chronic occlusion with thrombosis of the stent-inserted hepatic veins occurred in 5 patients (1/5 patients showed patency of the V 5 and the occlusion of the V 8 ) more than 200 days following stent placement. Fortunately, deterioration of LFT or attenuation changes on the CT scans did not develop in these patients; this might be attributable to the development of sufficient intrahepatic collateral pathway during the chronic course of occlusion. Acute thrombotic occlusion of the stent-inserted hepatic veins occurred in only 1 patient, and this patient also showed improved liver function without additional interventional management, although there remained some atrophy of the involved area. Frazer and Gupta 13 stated that as thrombosis is most likely to occur during the first 3 weeks prior to endothelization, they administered aspirin at 50 mg daily for 1 month after stent placement. However, we did not administer anticoagulants following stent placement because of the patients coagulopathic state following liver transplantation. From the results of follow-up CT scans in this series, we emphasize the necessity of IVG for V 5 and V 8 because all 9 patients with IVG patency showed the normalization of hepatic parenchymal attenuation eventually without atrophy. On the other hand, 1 patient without IVG patency showed atrophy of the S 5 and S 8. Especially in vulnerable patients, functional loss secondary to ischemia or atrophy in a part of the graft can result in graft failure in early posttransplantation periods. We believe that stent placement for IVG to V 5 and V 8 deserves to be performed despite the risk of stent occlusion because there remains a great chance for development of collaterals. This study had limitations. First, the CT follow-up interval was not constant due to the retrospective case analysis. Therefore, there was a chance for error in accurately determining the follow-up interval or duration data. Second, the pressure gradients across stenosis or occlusion before and after stent placement were not routinely measured in this study. The criteria for the abnormal pressure gradient have not been established, and some patients showed normal pressure gradient despite abnormal venogram findings, as in 5 of 27 patients in the report by Ko et al. 7 In conclusion, though IVG to V 5 and V 8 remains controversial, the treatment of their stenosis or occlusion is safe and effective, even during their immediate postoperative period. REFERENCES 1. Sano K, Makuuchi M, Miki K, Maema A, Sugawara Y, Imamura H, et al. Evaluation of hepatic venous congestion: proposed indication criteria for hepatic vein reconstruction. Ann Surg 2002;236: Lee SG, Park KM, Hwang S, Kim KH, Choi DN, Joo HS, et al. Modified right liver graft from a living donor to prevent congestion. Transplantation 2002;74: Hwang S, Lee SG, Park KM, Kim KH, Ahn CS, Lee YJ, et al. Hepatic venous congestion in living donor liver transplantation: preoperative quantitative prediction and follow-up using computed tomography. Liver Transpl 2004;10: Lee SG, Park KM, Hwang S, Lee YJ, Kim KH, Ahn CS, et al. Adult-to-adult living donor liver transplantation at the Asan Medical Center, Korea. Asian J Surg 2002;25: Lee S, Park K, Hwang S, Lee Y, Choi D, Kim K, et al. Congestion of right liver graft in living donor liver transplantation. Transplantation 2001;71: Sugawara Y, Makuuchi M, Akamatsu N, Kishi Y, Niiya T, Kaneko J, et al. Refinement of venous reconstruction using cryopreserved veins in right liver grafts. Liver Transpl 2004;10: Ko GY, Sung KB, Yoon HK, Kim JH, Song HY, Seo TS, et al. Endovascular treatment of hepatic venous outflow obstruction after living-donor liver transplantation. J Vasc Interv Radiol 2002;13: Yamagiwa K, Yokoi H, Isaji S, Tabata M, Mizuno S, Hori T, et al. Intrahepatic hepatic vein stenosis after living-related liver transplantation treated by insertion of an expandable metallic stent. Am J Transplant 2004;4: Mazariegos GV, Garrido V, Jaskowski-Phillips S, Towbin R, Pigula F, Reyes J. Management of hepatic venous obstruction after split-liver transplantation. Pediatr Transplant 2000;4: Cheng YF, Chen YS, Huang TL, de Villa V, Chen TY, Lee TY, et al. Interventional radiologic procedures in liver transplantation. Transpl Int 2001;14: Huang TL, Chen TY, Tsang LL, Sun PL, Chen YS, Wang CC, et al. Hepatic venous stenosis in partial liver graft

8 276 SHIN ET AL. transplantation detected by color Doppler ultrasound before and after radiological interventional management. Transplant Proc 2004;36: Sze DY, Semba CP, Razavi MK, Kee ST, Dake MD. Endovascular treatment of hepatic venous outflow obstruction after piggyback technique liver transplantation. Transplantation 1999;68: Frazer CK, Gupta A. Stenosis of the hepatic vein anastomosis after liver transplantation: treatment with a heparin-coated metal stent. Australas Radiol 2002;45: Ko EY, Kim TK, Kim PN, Kim AY, Ha HK, Lee MG. Hepatic vein stenosis after living donor liver transplantation: evaluation with Doppler US. Radiology 2003; 229: Kim BS, Kim TK, Kim JS, Lee MG, Kim JH, Kim KW, et al. Hepatic venous congestion after living donor liver transplantation with right lobe graft: two-phase CT findings. Radiology 2004;232: Sacks D, McClenny TE, Cardella JF, Lewis CA. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003;14:S199-S Moon DB, Lee SG. Adult-to-adult living donor liver transplantation at the Asan Medical Center. Yonsei Med J 2004;45: