Reconstruction of Inferior Right Hepatic Veins in Living Donor Liver Transplantation Using Right Liver Grafts

Transcription

1 LIVER TRANSPLANTATION 18: , 2012 ORIGINAL ARTICLE Reconstruction of Inferior Right Hepatic Veins in Living Donor Liver Transplantation Using Right Liver Grafts Shin Hwang, 1 * Tae-Yong Ha, 1 * Chul-Soo Ahn, 1 Deok-Bog Moon, 1 Ki-Hun Kim, 1 Gi-Won Song, 1 Dong-Hwan Jung, 1 Gil-Chun Park, 1 Jung-Man Namgoong, 1 Sung-Won Jung, 1 Sam-Youl Yoon, 1 Kyu-Bo Sung, 2 Gi-Young Ko, 2 Byungchul Cho, 3 Kyoung Won Kim, 2 and Sung-Gyu Lee 1 1 Division of Hepatobiliary Surgery and Liver Transplantation, Department of Surgery, 2 Department of Medical Imaging, and 3 Division of Medical Physics, Department of Radiation Oncology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea Because revascularization of the inferior right hepatic vein (IRHV) is a major component of right liver graft (RLG) reconstruction, we assessed the surgical techniques and clinical outcomes of IRHV reconstruction so that we could formulate practical guidelines for standardized procedures. From July 2004 to February 2010, we performed separate IRHV reconstructions in 487 of 1142 adult RLG recipients (42.7%). These recipients included 364 patients with a natural single IRHV and 123 patients with multiple IRHVs; in the latter group, the IRHVs were unified by venoplasty, which enabled a single anastomosis. The 1- year stenosis rates for the single-vein and venoplasty groups were 23% and 18.9%, respectively, and the early stent insertion rates were 7.1% and 9.8%, respectively (P ¼ 0.09). Late IRHV occlusion did not lead to graft dysfunction, and all large major IRHVs were patent. A morphometric analysis showed that IRHV stenosis was associated with IRHV stretching and an anastomotic level discrepancy. This led to refinements of the surgical techniques: IRHV orifices were shaped into funnels, and the IRHV anastomosis was accurately placed at the recipient inferior vena cava (IVC). In an ongoing prospective study of 35 patients, our funneling unification venoplasty resulted in only 1 episode (2.9%) of early IRHV stenosis requiring stenting at a median follow-up of 8 months. The final configurations of the reconstructed IRHVs after funneling unification venoplasty and extensive IVC dissection were very similar to those of the native donor liver. In conclusion, we suggest that in combination with extensive recipient IVC dissection, funneling and unification venoplasty techniques are useful for securely reconstructing single or multiple IRHVs during the implantation of RLGs. Liver Transpl 18: , VC 2011 AASLD. Received July 13, 2011; accepted October 24, The hepatic venous drainage pathways of a right liver graft (RLG) consist of the right hepatic vein (RHV), the middle hepatic vein (MHV), and the short hepatic vein (SHV). The inferior right hepatic vein (IRHV) is the most common type of major SHV and is present in approximately 40% of donor livers. 1,2 The complete reconstruction of these venous outflow pathways during living donor liver transplantation (LDLT) with RLGs is essential for preventing hepatic venous congestion and for making these RLGs fully functional. 3,4 Although there have been many studies about the reconstruction of the RHV and the MHV, the revascularization of the IRHV has been studied much less frequently, likely because many transplant surgeons Abbreviations: CT, computed tomography; IRHV, inferior right hepatic vein; IVC, inferior vena cava; LDLT, living donor liver transplantation; MHV, middle hepatic vein; RHV, right hepatic vein; RLG, right liver graft; SHV, short hepatic vein. This study was supported by the Asan Medical Center Organ Transplantation Center Research Fund. *These authors contributed equally to this work. Address reprint requests to Sung-Gyu Lee, M.D., Division of Hepatobiliary Surgery and Liver Transplantation, Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, Poongnap-Dong, Songpa-Gu, Seoul, Korea Telephone: ; FAX: ; sglee2@amc.seoul.kr DOI /lt View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI /lt. Published on behalf of the American Association for the Study of Liver Diseases VC 2011 American Association for the Study of Liver Diseases.

2 LIVER TRANSPLANTATION, Vol. 18, No. 2, 2012 HWANG ET AL. 239 regard a direct anastomosis to the inferior vena cava (IVC) as more feasible. 5,6 IRHV reconstruction enhances the functional recovery of implanted RLGs. During the performance of more than 1800 adult LDLT procedures with RLGs, we have found that it is often more time-consuming and complex to perform IRHV reconstruction in addition to orthodox RHV revascularization. Secure reconstruction becomes more difficult when there are 2 or more major SHVs, so various types of venoplasty techniques have been used to facilitate the revascularization of multiple SHVs. 5 To date, few analytic studies have assessed IRHV reconstruction. Therefore, we investigated the surgical techniques and clinical outcomes of IRHV reconstruction in 487 patients who underwent LDLT with RLGs so that we could formulate practical guidelines for standardized surgical procedures. PATIENTS AND METHODS Patient Selection We have previously described several innovative surgical techniques for securely reconstructing multiple SHVs. 5 Since then, these techniques have become the standards for RLG implantation at our institution. Because serious vascular anastomotic complications rarely occur more than 1 year after LDLT, a minimum follow-up of 12 months was set for this study. Patients were, therefore, recruited from July 2004 to February 2010 and were followed up until the end of February During the study period, 1142 adult patients at our institution received a single RLG for first LDLT. Patients undergoing dual graft implantation with an RLG and a left liver graft were excluded from this study. Five hundred three of these 1142 patients (44.1%) underwent IRHV reconstruction. Because this study was focused on separate single reconstructions of IRHVs, we also excluded patients for whom unusual reconstruction methods (n ¼ 16) were used [ie, complex quilt patching (RHV and IRHVs), the double IVC technique, separate triple anastomoses (orthodox RHV, middle RHV, and IRHV), and IVC replacement with Dacron grafts]. 5-7 Thus, our patient population consisted of 487 patients (42.7%) with a single IRHV anastomosis; these patients included 364 with a natural single IRHV and 123 with multiple IRHVs who underwent unification venoplasty. The outcomes of these 2 subgroups were compared retrospectively. We also performed a prospective study of refined surgical techniques for IRHV reconstruction from June 2010 to March 2011, and the patients were followed up until June These 2 associated studies were approved by the institutional review board of the Asan Medical Center. Patient Groups According to the Number of IRHVs For donor livers containing sizable single IRHVs (usually 5 mm), these veins were harvested with a slight rim of the donor IVC wall. A vessel patch from the recipient greater saphenous vein or a homologous vein patch was attached to the periphery of the vein orifice if the IVC rim appeared to be too small or too weak for direct suturing. This patch plasty was usually not necessary for reconstructing most of the medium to large IRHVs. These patients were placed in the singlevein group, regardless of patch attachment. We were reluctant to perform a direct anastomosis of a sizable IRHV to the IVC only after the placement of a niche at the IRHV stump because it often induced unwanted tearing of the thin side wall of the IRHV stump. For donor livers containing 2 or 3 IRHVs (each 4 mm in size), each IRHV was harvested separately, and unification quilt venoplasty was performed to make a single, large orifice. Detailed surgical techniques for this procedure have been described before. 5 In some donors with 2 IRHVs located closely together, a concurrent wide excision was performed, and this usually resulted in a snout-shaped common orifice and required additional circumferential attachment of a narrow vein patch. These patients were placed in the venoplasty group, regardless of the venoplasty types, because the final shapes were quite similar. Each single IRHV orifice, whether it was natural or artificial, was directly anastomosed to the recipient IVC. According to the size and location of the vein orifices, the corresponding part of the recipient IVC was extensively dissected to make deep side clamping safe and feasible. Some of the recipient IVC wall was excised to widen the anastomotic orifice. An anastomosis was performed with a 5-0 monofilament after the placement of 2 corner stay sutures. Evaluation of IRHV Patency and Treatment for Stenosis For the graft inflow and outflow pathways (including the RHV, MHV and IRHV), Doppler ultrasonography was performed intraoperatively, once per day during the first week, and once per week until the patient was discharged from the hospital. Patients were assessed by dynamic computed tomography (CT) every week during their hospital admission and after 1 month, 3 months, 1 year, 3 years, and 5 years. CT scans were also performed for patients with Doppler abnormalities or high serum concentrations of liver enzymes (> IU/L) during the immediate posttransplant period. CT scans of patients with hepatocellular carcinoma were performed more frequently for cancer surveillance during the first 5 years. 8 Significant hepatic vein stenosis can be diagnosed by a significant weakness of pulsatility or wave changes (from triphasic patterns to biphasic or monophasic patterns) on Doppler ultrasonography and by concurrent significant narrowing at the anastomosis on a CT scan. Changes in parenchymal perfusion on CT scans are also indicative of hepatic venous congestion. 3 Especially for relatively small IRHVs (4-5 mm in size), significant stenosis was defined as a loss of

3 240 HWANG ET AL. LIVER TRANSPLANTATION, February 2012 visualization or continuity at the anastomosis site on venous phase CT scans. Stents were inserted into most patients with apparent IRHV stenosis during the early posttransplant period. Stents were also inserted prophylactically into patients with excessive stretching or extrinsic compression due to an anastomotic level discrepancy between the graft IRHV and the recipient IVC (ventrodorsal or cephalocaudal), graft size mismatching, or disproportionate graft regeneration, which suggested ongoing IRHV stenosis. Most preventive IRHV stenting was performed concurrently with other hepatic vein access procedures such as RHV/MHV stenting and transjugular liver biopsy. A metallic stent was inserted across the IRHV anastomosis through an internal jugular vein approach. Detailed procedures for hepatic vein stenting have been described before. 9 No specific anticoagulation therapy was administered after IRHV stent insertion except for a routine treatment with antiplatelet agents for the first 3 to 6 months after LDLT. Morphometric Assessment and Computational Simulation for the Technical Evolution of Unification Venoplasty Just after the patient recruitment period, a cross-sectional study was performed to investigate the time-dependent morphometric changes occurring at the IRHV reconstruction site, especially in the venoplasty group. Stereoscopic structures of IRHV anastomoses were visualized from dynamic CT images with commercial 3-dimensional reconstruction software (Lucion, Infinitt Co., Seoul, Korea). 10 This study showed that some narrow, long-necked IRHVs had become elongated because of the mismatching of their level of insertion into the recipient IVC (an anastomotic level discrepancy) and gradual graft regeneration, which led to stenosis or occlusion. To understand the underlying mechanisms of this outflow disturbance, we tested several computational simulation models with computational fluid dynamics software (fluid-structure interaction model, ADINA R&D, Inc., Watertown, MA). We finally devised a computational simulation model that showed a wide range of hemodynamic tolerance against anastomotic level discrepancies and stretching from gradual graft regeneration. These findings were used to formulate a detailed technique for newly revised unification venoplasty. Since June 2010, this technical refinement for shortening and widening IRHV stumps has been used in practice as part of a prospective study. Statistics All numerical data are reported as means and standard deviations or medians and ranges. Late IRHV occlusion due to massive hepatocellular carcinoma recurrence was censored from the analysis. Incidence rates were compared with the v 2 test or Fisher s exact test. IRHV patency and patient survival rates were estimated with the Kaplan-Meier method and were compared with the log-rank test. A P value < 0.05 was considered statistically significant. RESULTS Patient Profiles and Survival Outcomes During the study period, 1142 patients each received a single RLG, 11 and 487 of these patients (42.7%) underwent separate reconstructions of major IRHVs. These patients were divided into 2 groups: patients with a natural single IRHV (n ¼ 364) and patients with 2 (n ¼ 118) or 3 IRHVs (n ¼ 5) who required unification venoplasty. The clinicopathological profiles of the recipients and donors in these 2 groups are summarized in Table 1. The graft MHV was concurrently reconstructed in 89.1% of these patients; interposition vessel grafts were used, or the MHV trunk was concurrently harvested (modified, modified extended, and extended RLGs). There was no difference in the overall survival rates of the 2 groups (P ¼ 0.83; Fig. 1). Hepatocellular carcinoma recurrence was the most frequent cause of late patient death. IRHV Patency in the Single-Vein Group A separate anastomosis of a natural single IRHV was performed for 364 patients. Twenty-six patients (7.1%) required stenting for IRHV stenosis 0 to 50 days after LDLT: 12 on days 0 to 1, 8 on days 2 to 7, 3 on days 8 to 14, 1 on day 19, 1 on day 31, and 1 on day 50 (Fig. 2). Only 4 of these 26 patients (15.4%) showed noticeable hepatic venous congestion of the corresponding IRHV territory on CT scans; all underwent stenting on day 0 or 1, and this resulted in gradual improvements in perfusion abnormalities. Two of these 4 patients underwent concurrent RHV stenting. The remaining 22 patients underwent preventive stenting, although there was no definite hepatic venous congestion on CT scans. All 22 patients underwent hepatic vein access procedures [concurrent RHV/MHV stenting (n ¼ 18) or transjugular liver biopsy (n ¼ 4)]; during these procedures, a significant pressure gradient (usually >5 cmh 2 O) along with excessive stretching or extrinsic compression was incidentally found, and this led to preventive IRHV stenting. Delayed IRHV stenting after 1 month was due to prolonged graft dysfunction with IRHV stenosis. Eighteen of 26 IRHV stents (69.2%) appeared patent on CT scans at a median follow-up of 36.8 months (range ¼ 1-75 months), and the 1-year patency rate was 80.8%. Among these 26 patients, there was only 1 peritransplant death due to pneumonia on day 29. Abrupt cutting or nonvisualization of the IRHV on follow-up CT scans was observed in 72 of the 338 patients not requiring stenting, but none of them showed hepatic perfusion defects despite persistent stenotic configuration. No large IRHV (usually 1 cm) was occluded.

4 LIVER TRANSPLANTATION, Vol. 18, No. 2, 2012 HWANG ET AL. 241 TABLE 1. Clinicopathological Profiles of RLG Donors and Recipients Requiring Separate Reconstruction of the IRHVs During Implantation Single-Vein Group Venoplasty Group Parameter (n ¼ 364) (n ¼ 123) P Value Age (years)* (18-69) (25-65) 0.08 Sex: male/female (n/n) 267/97 93/ Model for End-Stage Liver Disease score* (6-50) (6-45) 0.20 Hepatitis B virus associated liver cirrhosis [n (%)] 271 (74.5) 102 (82.9) 0.06 Hepatocellular carcinoma [n (%)] 148 (40.7) 56 (45.5) 0.34 Graft type [n (%)] 0.62 Conventional right lobe 38 (10.4) 15 (12.2) Modified right lobe 316 (86.8) 105 (85.4) Extended right lobe 10 (2.7) 3 (2.4) Graft-to-recipient weight ratio Overall 5-year survival rate (%) Perioperative mortality [n (%)] 15 (4.1) 2 (1.6) 0.26 Causes of overall patient death (n) Graft dysfunction 6 2 Infection 18 2 Hepatocellular carcinoma recurrence 16 9 Other 8 2 *The data are presented as means and standard deviations (ranges are shown within parentheses). Modified extended right lobes are included. The data are presented as means and standard deviations. Less than 3 months. The overall 1-month, 1-year, and 3-year stenosis rates (including stent insertion) after the anastomosis of a natural single IRHV were 7.1%, 23%, and 28.9%, respectively (Fig. 3). IRHV Patency in the Unification Venoplasty Group Figure 1. Overall survival curves for RLG recipients requiring IRHV reconstruction. The open circles and the solid triangles indicate the single-vein group and the venoplasty group, respectively. A single anastomosis of unified IRHVs was performed for 123 patients. Twelve of these patients (9.8%) required stenting 1 to 69 days after LDLT: 7 patients on day 1, 2 patients on days 2 to 7, 1 patient on day 20, 1 patient on day 43, and 1 patient on day 69. All 12 patients underwent preventive stenting despite the absence of noticeable hepatic venous congestion of the corresponding IRHV territory; thus, the larger branch was preferentially selected for stenting if double stenting was not feasible. Eight of these 12 patients also underwent concurrent RHV/MHV stenting. Eight of the 12 IRHV stents (66.7%) appeared patent on followup CT scans at a median of 39.4 months (range ¼ 1-74 months) after stenting; the 1-year patency rate was 75%. There was only 1 peritransplant death, which was due to severe graft dysfunction on day 16; this was not associated with a hepatic outflow problem. Abrupt cutting or nonvisualization of the IRHV on follow-up CT scans was observed in 11 of 111 patients (9.9%) not requiring stenting, but none showed any abnormalities in hepatic parenchymal perfusion despite persistent stenotic configurations. Nearly all large IRHVs (usually 1 cm) remained patent during follow-up. The overall 1-month, 1-year, and 3-year occlusion rates (including stenting) after the anastomosis of unified IRHVs were 8.1%, 18.9%, and 18.9%, respectively (Fig. 3). The patency rates did not differ significantly between the single-vein and venoplasty groups (P ¼ 0.09). Morphometric Analysis of IRHV Reconstruction and Subsequent Technical Refinement Abnormal IRHV reconstruction configurations were divided into 2 main types: lengthy stretching due to excessive length and stretching of the extrahepatic IRHV

5 242 HWANG ET AL. LIVER TRANSPLANTATION, February 2012 Figure 2. CT images showing anastomotic stenosis after IRHV reconstruction in the single-vein group. All images were taken 1 day after liver transplantation. None of these patients had noticeable hepatic venous congestion corresponding to the IRHV territory, but they all underwent IRHV stenting to treat or prevent anastomotic stenosis. neck (Fig. 2A) and an anastomotic level discrepancy (Fig. 2C). Gradual graft regeneration was not proven to be one of the primary causes of IRHV anastomotic stenosis. A simulation analysis indicated that IRHV stenosis due to lengthy stretching could be overcome by the transformation of the extrahepatic IRHV into a short and wide funnel and by meticulous level matching of the IRHV anastomosis after extensive recipient IVC dissection (Fig. 4). Therefore, we revised our surgical techniques so that we could design short, wide, funnel-shaped IRHV orifices, especially for relatively small single IRHVs (Fig. 5) and unified IRHVs (Fig. 6). An additional small incision to the IRHV wall made the IRHV short and wide. When the walls of 2 IRHVs appeared long and thick, the simple placement of a central patch could complete the unification venoplasty, which required only a small vein patch, and shortened the bench-work time (Fig. 6). IRHV stenosis due to an anastomotic level discrepancy arose from the inadequate location of the IRHV with respect to the recipient IVC. An anastomotic level discrepancy occurred in the ventrodorsal axis, the cephalocaudal axis, or the combined axes. A ventrodorsal discrepancy was observed more frequently and was usually associated with shallow side clamping of the recipient IVC without extensive dissection of the right adrenal area. This led to an IRHV anastomosis in the axial direction of 9 to 10 o clock (Fig. 2A,C), even though the native graft IRHV anatomy required an anastomosis to the recipient IVC at 7 to 8 o clock (Fig. 2B,D). The cephalocaudal discrepancy was often associated with a shortened distance between the anastomosis sites of the orthodox RHV and the IRHV. Figure 3. Occlusion rates during IRHV reconstruction. The open circles and the solid triangles represent the single-vein group and the venoplasty group, respectively. These findings indicate that accurate localization of the IRHV anastomosis is essential after extensive mobilization of the recipient IVC, especially around the right adrenal gland. We describe our knack for preventing anastomotic level discrepancies in Fig. 7. Results of the Clinical Application of New Technical Refinements During the 10 months from June 2010 to March 2011, the refined funneling technique was applied to 35 RLGs for a prospective study.

6 LIVER TRANSPLANTATION, Vol. 18, No. 2, 2012 HWANG ET AL. 243 Figure 4. Computational simulation models for assessing the morphometric changes in 2 unified IRHVs. (A) After conventional unification venoplasty, the IRHVs appear patent in the normal position, but at least 1 IRHV becomes stenotic in the stretched state (indicated by an asterisk). (B) Funneling unification venoplasty makes the shapes of the IRHVs resemble the situation in the donor liver, in which the IRHVs appear patent in both normal and stretched positions. This small difference in the anastomotic configuration seems to provide patency tolerance against various extrinsic factors. Different colors are used for the liver capsule (brown), the extrahepatic IRHV branches (black), the vein patches (red), and the donor/recipient IVC (blue). Figure 5. Refined surgical techniques for the reconstruction of a single small (4-mm) IRHV. (A) A narrow vein neck was opened (B) by a small incision (indicated by an arrow) into the liver parenchyma. (C) A small vein patch was prepared, and (D) it was sutured. (E) The final shape was similar to a funnel. (F,G) This incisional patch venoplasty converted the shape of the vein branch. This small vein was not indicated for reconstruction in principle, but it had to be reconstructed because of the low graft-to-recipient ratio of Funneling incision patch plasty (Fig. 5) was performed for a small or medium IRHV in 19 patients. During a median follow-up of 7 months, only 1 patient (5.3%) underwent IRHV stenting because of anastomotic stenosis. We also performed funneling unification venoplasty (Fig. 6) on 16 RLGs containing 2 IRHVs. Its indication was the presence of long, narrow IRHVs with or without a lack of an adequate saphenous vein patch. During a median follow-up of 8 months, none of these patients underwent IRHV stenting. In addition to graft IRHV venoplasty, the recipient IVC was dissected more deeply than before to minimize the anastomotic level discrepancy. The right half of the recipient suprarenal IVC was extensively dissected, and the site of the IRHV anastomosis was prudently determined after the consideration of the IRHV direction (according to donor liver CT scans), the actual distance between the RHV and IRHV orifices on the back table, and the relative location of the graft IRHV in the right subphrenic fossa after the completion of the RHV anastomosis (Fig. 7). The early stenting rate after the application of the newly devised funneling technique [2.9% (1 of 35 patients)] was more favorable than the rate with conventional methods [7.8% (38 of 487 patients)], but there was no statistical difference (P ¼ 0.46). Proposed Guidelines for Major SHV Reconstruction Our current guidelines for reconstructing single or multiple major SHVs with respect to the orthodox RHV are summarized in Fig ,12

7 244 HWANG ET AL. LIVER TRANSPLANTATION, February 2012 Figure 6. Simplified surgical techniques for the reconstruction of 2 adjacent small IRHVs. (A) A narrow vein neck was opened (B) by a small incision (indicated by arrows) into the liver parenchyma. (C) A 3-cmlong vein patch was placed between the incised orifices and then sutured. (D,E) The final shape of the unified IRHVs was a short funnel, so circumferential fence attachment was not required. Small incisions combined with a small central patch converted (F) a conventional unification venoplasty into (G) a simplified unification funneling venoplasty. DISCUSSION It is generally accepted that major SHVs larger than 5 mm in diameter indicate revascularization. Major SHVs were identified in 44.1% of the RLGs in this study, and this indicates that IRHV reconstruction is one of the primary surgical components of RLG implantation. Unlike orthodox RHV reconstruction, IRHV revascularization usually does not use the recipient IRHV stump for reconstruction because of the very low probability of adequately matching the sizes and locations of the graft and recipient IRHVs. The harvesting of a sizable IRHV with a slight rim of the donor IVC wall is often sufficient, and then usual IRHV reconstruction follows the usual order: extensive dissection of the recipient IVC, deep side clamping of the IVC, a corresponding IVC orifice incision/excision, and running sutures. This straightforward sequence of IRHV reconstruction may be the primary reason that IRHV reconstruction has been considered beyond the focus of main surgical concerns. Although the wide range of anatomical variations in donor SHVs is already known and their reconstruction often makes LDLT surgery more difficult, 1,2,13 the need for effective IRHV reconstruction might have been underestimated. The presence of a large IRHV has been associated with a reciprocal reduction in the size of the orthodox RHV orifice, 13 and this makes IRHV reconstruction more important for successful graft implantation. As our experience with RLG implantation increased, we realized the need for formulating reliable guidelines for secure IRHV reconstruction with full coverage of wide anatomical variations. Therefore, we investigated our institutional experience with IRHV reconstruction. Technically, our development of unification venoplasty in 2004 was essential for coping with Figure 7. Knack for matching the cephalocaudal and ventrodorsal levels for IRHV reconstruction. (A,B) Two IRHVs are unified, and the center corresponds to the direction of 8 o clock. (C) After orthodox RHV reconstruction, the right lobe graft is placed into the right subphrenic fossa; 2 stay sutures (indicated by arrows) are placed in the direction of 9 o clock. (D) This area is pulled upward and side-clamped deeply. A longitudinal incision is made in the direction of 8 o clock (indicated by a green bidirectional arrow). The ventral part of this incision is excised (indicated by a dotted curve), and its dorsal part (indicated by an asterisk) is also plicated deeply during posterior wall suturing; this results in an oval orifice at the recipient IVC wall. We suggest not resecting the dorsal part of this incision because it is difficult to control bleeding in this area.

8 LIVER TRANSPLANTATION, Vol. 18, No. 2, 2012 HWANG ET AL. 245 Figure 8. Current institutional guidelines for reconstructing a major SHV with respect to the orthodox RHV. Quilt venoplasty is a complex patch plasty procedure used to unify multiple short SHVs without the application of a niche to the SHV stump. 5 anatomical variations in donor SHVs because its widespread adoption led to significant improvements in IRHV patency rates. 5 After performing more than 500 IRHV reconstructions, we felt that this high volume required analysis so that we could further improve the patency rates. For single-vein IRHV anastomoses, our surgical technique resulted in a stenting rate of 7.1%, but only 1.1% of the single-vein patients experienced significant hepatic venous congestion. Thus, we think that our current technique can be applied to routine clinical practice. Morphometric analyses indicated the importance of adequately matching the IRHV insertion level to the recipient IVC because a noticeable anastomotic level discrepancy was present in approximately half of the patients who required early stenting. 10 We previously thought that RLG grafts would regenerate and that IRHV insertion would, therefore, move ventrally and caudally, but we did not observe such a regeneration-associated effect in this study. The distance between the RHV and IRHV orifices was also not noticeably elongated according to graft regeneration. These observations indicate that it is reasonable to place the IRHV anastomosis exactly on the corresponding side wall of the recipient IVC without consideration of later regeneration. IRHV insertion level matching and wide anastomoses for IRHVs resulted in nearly perfect long-term patency. However, it may be difficult to exactly locate the IRHV insertion if the extrahepatic portion of the graft IRHV is narrow and lengthy because this type of graft is vulnerable to unwanted elongation or extrinsic compression. Our computational simulation has provided reasonable answers to some important questions related to IRHV reconstruction. The first question is how to exactly match IRHV insertion levels. Unlike RHV reconstruction, there is no fixed landmark for attaching the donor IRHV to the recipient IVC. Therefore, we place the RLG at its natural position in the right subphrenic fossa just after the RHV anastomosis, and we mark the corresponding center of the IRHV orifice on the lateral wall of the recipient IVC. This is followed by deep clamping at this point under upward traction with 1 or 2 small stay sutures. It is helpful to check the exact axial direction of the IRHV insertion in the native donor liver by a review of the donor CT scans. For deep and secure side clamping of the recipient IVC, it is usually necessary to extensively dissect the right half of the suprarenal IVC, and some branches of the right adrenal veins often need to be cut. Another question is how to overcome the drawbacks of long, narrow IRHV necks. The solution is to shorten and widen the IRHV neck and make a funnel shape after a deep 1-sided incision and patch application. Funneling venoplasty can achieve secure venous drainage, especially in patients with a small-for-size RLG. In addition, we emphasize that a small longitudinal incision (1.5-2 cm long) at the recipient IVC occasionally makes the effective diameter of the IVC opening insufficiently small because of a narrow slit effect due to the thickened IVC wall. Thus, we suggest the excision of a small part of the IVC wall after the IVC incision to prevent this stiff slit effect. We observed noticeable stenosis in 21.8% of the single IRHV anastomoses during the first posttransplant year in patients who did not undergo early stenting. There was no significant graft dysfunction due to this gradual occlusion, however, because nearly all of the occluded IRHVs were small or medium (not large). Had we made the indication for IRHV revascularization slightly narrower (eg, strictly excluding IRHVs < 5 mm), many of these occluded IRHVs would not have been reconstructed at all. We previously reported that the patency rates after stent insertion for hepatic venous outflow obstruction

9 246 HWANG ET AL. LIVER TRANSPLANTATION, February 2012 were 82.3% at 1 year and 75.0% at 3 years. The diameter of the stents was an independent factor associated with the patency of stents. 9 The results of this study showed a 1-year patency rate of 78.9% for all IRHV stents. Not a negligible proportion of the patients with IRHV stenting showed gradual diminution of intrahepatic IRHV blood flow over a few years, which finally led to stent occlusion. Because of its high technical success rate and acceptable long-term patency, early IRHV stenting must be an effective treatment for IRHV stenosis regardless of the underlying causes (eg, mechanical stretching, compression, twisting, or thrombosis). In contrast, balloon angioplasty may be applicable to IRHV stenosis, but it is not currently advisable because unacceptably high rates of elastic restenosis have been reported. 14,15 Interestingly, we found very high rates of simultaneous stenting for both orthodox RHV and IRHV stenting: 53.9% for the single-vein group and 41.7% for the venoplasty group. We previously reported a 2-week RHV stenting rate of 4% (9 of 225 patients) in a 1- year study. 10 These high rates may be closely related to the fact that many of the IRHV stenting procedures were also performed as a preventive measure in patients undergoing RHV or MHV stenting. In contrast, a few patients required multiple concurrent stents for the RHV and the IRHV and even for the MHV; in these patients, severe collapse of the recipient IVC due to extrinsic compression was often observed. Preventive IRHV stenting was also performed just after the measurement of the hepatic vein pressure gradient during transjugular liver biopsy. If these hepatic vein access procedures had not been performed, many of the preventive IRHV stents would not have been inserted. Unification venoplasty can convert 2 adjacent IRHVs into a large IRHV. This was proven to be beneficial, in that the patency rates were comparable to those observed for the single-vein group. We have previously described 2 types of unification venoplasty for 2 IRHVs. 5 The first type involves the wrapping of 2 long patches around the 2 IRHVs and their placement in a line, but this method became obsolete after its performance in only a few patients. The second technique consists of the placement of a central patch and additional circumferential fences; this was our preferred unification technique for 2 or more IRHVs. However, this technique has 2 drawbacks, which were masked by its excellent outcomes. First, it requires a relatively long vein patch (6-8 cm) and thus a longer incision for harvesting the recipient s saphenous vein segment. A very short vessel segment from the recipient s portal vein or other available sources is usually not adequate patch material. This technique also requires a relatively long time for bench work (approximately 15 minutes) because of the long or complex suture lines. The second drawback is the vulnerability of the unified IRHV to elongation or extrinsic compression. We have developed a third type of unification venoplasty (Fig. 6), which can convert 2 adjacent IRHVs into a short, funnel-shaped vein with only central placement of a small vein patch. This type seems to be time- and resource-effective (10 minutes or less for bench work with a 2- to 3-cm-long vein patch) and hemodynamically compliant. The unified vein stump is rather short and prevents excessive redundancy. Thus, deeper IVC side clamping after more extensive IVC mobilization is required to prevent unnecessary tension during the anastomosis procedure. If the unified IRHV stump appears too short for a direct anastomosis, a semicircular or circumferential vein fence can be attached to facilitate the anastomosis to the recipient IVC. Before this study, our main unification technique for 2 IRHVs involved the creation of a circumferential fence with or without a central patch. In contrast, since this study, our preferred method has been the creation of a tight central patch with or without an outer fence; this depends on the length of the available vessel patch and the condition of the native IRHV stumps. During the ongoing prospective study, we were very concerned about the final configuration of the reconstructed IRHVs. Unexpectedly, we recognized that the final configuration of 2 reconstructed IRHVs became very similar to the configuration of the native donor liver after the application of funneling unification venoplasty and meticulous anastomosis level matching; in these patients, no noticeable IRHV stenosis occurred. Although the effect of funneling venoplasty was not statistically proven in this study, primarily because of the small sample size of the prospective study group, we think that it is beneficial to make the IRHV reconstruction similar to the native IRHV anatomy of the donor liver to minimize anastomotic stenosis and to confer resistance to extrinsic stretching or compression. For any type of patch plasty, it is important to consider the quality and consistency of the vessel patch. For IRHV reconstruction, the recipient s autologous saphenous vein appears to be optimal, but it is occasionally too weak or too small to use for this purpose. Alternatives include short segments from other resources, including a cryopreserved femoral vein or vena cava, a fresh iliofemoral artery, or an autologous portal vein patch. We emphasize that cryopreserved iliac artery patches should not be used because their intima and media are often too fragile to endure tension during the anastomosis procedure. 16 Most homograft patches are the remnants of segments after RHV venoplasty or MHV interposition reconstruction. The concept of maximal hepatic outflow reconstruction, which originated from MHV reconstruction, has been applied to IRHV reconstruction because every hepatic vein has its own drainage territory. The anatomy of SHVs is as complex and variable as that of the MHV. Therefore, to effectively cope with a wide range of variations in SHV anatomy, we thought it worthwhile to establish experience-based practical guidelines for reconstructing every variant type of SHV. We have summarized our experience with more than 2000 adult LDLT procedures, and we have proposed our current guidelines for the reconstruction of single

10 LIVER TRANSPLANTATION, Vol. 18, No. 2, 2012 HWANG ET AL. 247 or multiple major SHVs with respect to the orthodox RHV. The existence of multiple SHVs often makes hepatic outflow reconstruction difficult and unreliable. If a single IRHV anastomosis is not feasible, we suggest clustered venoplasty because like a side-to-side cavocaval anastomosis during deceased donor liver transplantation, the reconstruction of such a large vein orifice is highly reliable. 5,12 In conclusion, we suggest that funneling and unification venoplasty methods in combination with extensive recipient IVC dissection are useful for the secure reconstruction of single or multiple IRHVs during the implantation of RLGs. ACKNOWLEDGMENTS The authors thank the following registered nurses for their contributions to data collection: Hea-Seon Ha, Jeong-Ja Hong, In-Ok Kim, Mi-Kyeong Jeon, Ji-Seon Yun, Seon-Young Choi, Yu-Jin Kang, and Eun-Joo Sim. REFERENCES 1. Uchida K, Taniguchi M, Shimamura T, Suzuki T, Yamashita K, Ota M, et al. Three-dimensional computed tomography scan analysis of hepatic vasculatures in the donor liver for living donor liver transplantation. Liver Transpl 2010;16: Varotti G, Gondolesi GE, Goldman J, Wayne M, Florman SS, Schwartz ME, et al. Anatomic variations in right liver living donors. J Am Coll Surg 2004;198: 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. Techniques of reconstruction of hepatic veins in living-donor liver transplantation, especially for right hepatic vein and major short hepatic veins of right-lobe graft. J Hepatobiliary Pancreat Surg 2006;13: Hwang S, Lee SG, Park KM, Kim KH, Ahn CS, Moon DB, Ha TY. Quilt venoplasty using recipient saphenous vein graft for reconstruction of multiple short hepatic veins in right liver grafts. Liver Transpl 2005;11: 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: Matsuda H, Sadamori H, Shinoura S, Umeda Y, Yoshida R, Satoh D, et al. Aggressive combined resection of hepatic inferior vena cava, with replacement by a ringed expanded polytetrafluoroethylene graft, in living-donor liver transplantation for hepatocellular carcinoma beyond the Milan criteria. J Hepatobiliary Pancreat Sci 2010;17: Hwang S, Ahn CS, Kim KH, Moon DB, Ha TY, Song GW, et al. Super-selection of a subgroup of hepatocellular carcinoma patients at minimal risk of recurrence for liver transplantation. J Gastrointest Surg 2011;15: Ko GY, Sung KB, Yoon HK, Kim KR, Kim JH, Gwon DI, Lee SG. Early posttransplant hepatic venous outflow obstruction: long-term efficacy of primary stent placement. Liver Transpl 2008;14: Hwang S, Lee SG, Ahn CS, Moon DB, Kim KH, Sung KB, et al. Morphometric and simulation analyses of right hepatic vein reconstruction in adult living donor liver transplantation using right lobe grafts. Liver Transpl 2010;16: Lee SG, Hwang S, Kim KH, Ahn CS, Moon DB, Ha TY, et al. Toward 300 liver transplants a year. Surg Today 2009;39: Kishi Y, Sugawara Y, Matsui Y, Akamatsu N, Motomura N, Takamoto S, Makuuchi M. Alternatives to the double vena cava method in partial liver transplantation. Liver Transpl 2005;11: Radtke A, Sotiropoulos GC, Sgourakis G, Molmenti EP, Schroeder T, Saner FH, et al. Hepatic venous drainage: how much can we learn from imaging studies? Anatomic-functional classification derived from three-dimensional computed tomography reconstructions. Transplantation 2010;89: Ko GY, Sung KB, Yoon HK, Kim JH, Song HY, Seo TS, Lee SG. Endovascular treatment of hepatic venous outflow obstruction after living-donor liver transplantation. J Vasc Interv Radiol 2002;13: Huang TL, Chen TY, Tsang LL, Sun PL, Chen YS, Wang CC, et al. Hepatic venous stenosis in partial liver graft transplantation detected by color Doppler ultrasound before and after radiological interventional management. Transplant Proc 2004;36: Pascual G, Jurado F, Rodríguez M, Corrales C, López- Hervás P, Bellón JM, Buján J. The use of ischaemic vessels as prostheses or tissue engineering scaffolds after cryopreservation. Eur J Vasc Endovasc Surg 2002;24: