Address reprint requests to Sung-Gyu Lee, M.D., F.A.C.S., Division of Hepatobiliary Surgery and Liver Transplantation, Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Poongnap-Dong, Songpa-Gu, Seoul, Korea 138-736. Telephone: 82-2-3010-3485; FAX: 82-2-474-9027; E-mail: email@example.com
When living donor liver transplantation (LDLT) is performed for pediatric patients, a left lateral section (LLS) graft is often used after consideration of graft-recipient size matching. Although the anatomy of the left hepatic vein (LHV) is diverse, most LHVs are suitable for direct anastomosis to the recipient hepatic vein stumps or to the inferior vena cava (IVC). However, in rare instances, the LLS can be drained through both the LHV and the middle hepatic vein (MHV).[1, 3] In living donors with an unusual hepatic vein anatomy, the preservation of the MHV trunk may result in a sizable hepatic vein opening positioned relatively far from the main LHV. Because the hepatic vein orifice is located at the liver's cut surface in such cases, customized venoplasty techniques are necessary to make it suitable for graft implantation, especially for infant recipients with a very small IVC.
Such anatomical variations of LHV are rare and are thus not often clinically encountered. Therefore, data obtained from the donor pool of a transplantation center performing large numbers of LDLT procedures were used to determine the incidence and to develop technical methods for effective reconstruction. This study was intentionally focused on effective reconstruction of the hepatic vein outflow in infant recipients weighing approximately 10 kg or less.
PATIENTS AND METHODS
The ages of pediatric liver transplant candidates span a wide range (from a few months to adolescence), and their body weights can differ by 10 times. In practice, all living donors are adults, so LDLT is more demanding for young pediatric patients versus older pediatric patients because of size mismatching of the graft and vessels. Thus, this study was designed to establish surgical techniques for the effective reconstruction of the hepatic vein outflow of LLS grafts and was especially focused on infant recipients weighing approximately 10 kg or less.
This study comprised 3 parts: an LHV variation analysis, a simulation-based design for the technical modification of graft LHV venoplasty, and its clinical application with an outcome analysis.
A retrospective analysis of LHV variations in potential LLS graft donors was conducted with a special emphasis on multiple, separate graft hepatic vein openings. First, the database of a high-volume LDLT center was reviewed. The database contained computed tomography image data for 1172 consecutive LDLT procedures performed between January 2008 and December 2011. Left lobe and LLS graft donors were selected; most of these were donors for dual-donor adult LDLT or pediatric LDLT. Thereafter, 300 donors were consecutively selected after sex matching to avoid potential bias because males were dominant in the whole donor pool but females were definitely dominant in the donor pool for pediatric LDLT. For these 300 donors, a computational simulation of virtual LLS graft harvesting was performed with dynamic computed tomography images and commercial 3-dimensional reconstruction software (Lucion, Infinitt Co., Seoul, Korea). The surgical feasibility was simulated for a single LHV opening and for 2 adjacent LHV openings, which enabled surgical unification by LHV transection in the virtual 3-dimensional environment. The anatomical variations of LHV were classified into 4 types according to the number, size, and location of the graft hepatic vein openings (Fig. 1). Very small hepatic vein branches (2-3 mm in diameter) draining into the MHV or IVC were not considered for this classification because they could be sacrificed without the risk of significant hepatic vein congestion.
In the second part of the study, customized venoplasty techniques were developed to unify 2 separate graft hepatic vein openings for a 10-kg infant recipient with an IVC diameter measuring approximately 1 cm. When the space between 2 hepatic vein openings was wide, the use of an interposition vessel graft was required to compensate for the void. In the case of a hepatic vein opening positioned at the liver's cut surface, there were 2 options for reconstruction: an end-to-end anastomosis or an end-to-side anastomosis with an interposition vessel graft. Reconstruction of the interposed vessel was performed with the infant recipient's IVC and a separate anastomosis or a single anastomosis after parallel and wedged unifications. Morphometric data obtained from the surgical simulations, including various combinations of these surgical options, were analyzed, and then the most suitable method was chosen.
In the final part of the study, the customized venoplasty technique was clinically applied, and its applicability and usefulness were assessed through a retrospective review of pediatric LDLT series using LLS grafts. The previously described wedged unification venoplasty technique for LLS grafts was established in the middle of 2006 at our institution. Thus, we set the study period for recipients as July 2006 to December 2011, and we followed the patients until June 2012. During this study period, 49 pediatric patients with an age of 24 months or less (age range = 4-24 months) underwent primary LDLT with an LLS graft.
Numerical variables are presented as means and standard deviations with ranges. Incidences were analyzed with Fisher's exact test. Survival rates were determined with the Kaplan-Meier method. Statistical significance was set at P < 0.05. The study protocol was approved by the institutional review board of the Asan Medical Center.
Anatomical LHV Variations in Living Donors Enabling LLS Grafts
The LHV anatomy of 300 potential LLS graft donors was classified according to the patterns of the graft hepatic veins. In all, 106 LLS grafts and 194 left lobe grafts were actually harvested from these living donors. The mean age of the donors was 33.7 ± 6.9 years (range = 20-55 years), and 150 (50%) were male. The patterns of the virtual LLS graft hepatic veins were classified into types according to the number, size, and location: (1) a single opening (n = 218 or 72.7%); (2) 2 large adjacent openings (n = 29 or 9.7%); (3) 2 adjacent openings, 1 large and 1 small (n = 34 or 11.3%); and (4) 2 large and widely spaced openings (n = 19 or 6.3%; Fig. 1).
Type 1 LHV grafts were divided into 3 subtypes according to the presence and location of a fissural vein, and all the subtypes were suitable for use as a single hepatic vein orifice with or without venoplasty. Type 2 LHV grafts were divided into 3 subtypes according to the pattern of the union with the IVC: all appeared to require unification venoplasty with deep wedging (septoplasty with excavation of some intervening liver parenchyma). Type 3 LHV grafts were divided into 3 subtypes according to their location relative to the MHV trunk: all required unification venoplasty with or without deep wedging. Type 4 LHV grafts were also divided into 3 subtypes according to their location relative to the MHV trunk and the dominance of the LHV trunk. Nearly all type 4 subtypes appeared to require an interposition vessel for unification venoplasty. In some instances, type 3 graft anatomy closely resembled type 4 graft anatomy (Fig. 1). In this study population, the type 4 donors donated left lobe grafts (except for 1 LLS graft requiring unification venoplasty with a wide vein patch).
Simulation-Based Design of Surgical Techniques
LLS grafts with a type 1 LHV are ideal for implantation in infant patients. Other types of grafts require venoplasty to make the hepatic vein orifices suitable for reconstruction.
For LLS grafts with type 2 or 3 hepatic veins, simulation studies were unnecessary because routine wedged unification venoplasty had been performed successfully in more than 50 LLS grafts for pediatric or adult LDLT with dual grafts by our transplant team. In fact, the technique used in these cases is very similar to that used for the unification of the MHV and LHV trunks in left lobe grafts. Wedged unification is intended to create a wide common channel for the joining of 2 intrahepatic vein branches. Thus, its shape becomes hemodynamically compliant and tolerant of extrinsic compression. The septum between the hepatic vein branches is cut, and the underlying liver parenchyma is excavated; then, this portion is simply repaired by central approximation tying and bidirectional suturing with a monofilament (Fig. 2). The use of absorbable suture material for pediatric patients is recommended. Occasionally, conventional venoplasty has been used with simple septotomy (not excavating septoplasty) for the unification of small hepatic vein orifices (Fig. 3). However, the resulting configuration is not hemodynamically compliant, especially under the conditions of IVC compression in infant recipients receiving a large-for-size graft.[9, 10] Therefore, this method is usually not suggested for major hepatic veins in grafts. In contrast to older pediatric or adult patients, unification venoplasty should not make the graft hepatic vein orifice too large in an infant patient because the diameter of the recipient's IVC is usually as small as 1 cm or less.
In contrast, for LLS grafts with a type 4 LHV, it was determined that the reconstruction technique should be refined further. Two surgical methods for a graft hepatic vein opening at the liver's cut surface are illustrated in Fig. 4. According to a morphometric analysis, the end-to-side method (Fig. 4A) appears to be more hemodynamically beneficial and more tolerant of extrinsic compression than the end-to-end method (Fig. 4B). For reconstruction with the recipient's hepatic vein/IVC, the separate reconstruction of an interposed vein to the retrohepatic IVC (Fig. 4C) does not appear to be appropriate because of the consequent exposure of the hepatic vein reconstruction to excessive stretching and compression after abdomen closure and posttransplant graft remodeling. Furthermore, parallel unification venoplasty (Fig. 4D) is usually not recommended because of the relatively low drainage efficacy despite its enlarged size. Therefore, the aforementioned wedged unification venoplasty method (Fig. 4E) appears to be most suitable for type 4 LHV grafts because it enables size adjustment of the unified hepatic vein orifice and enhancement of the drainage efficiency.
Clinical Application of Customized Reconstruction Techniques and Outcome Analysis
Forty-nine cases of pediatric LDLT with a recipient age of 24 months or less (age range = 4-24 months) were reviewed (Table 1). Unification venoplasty was applied to 13 of the LLS grafts with LHV variations. Hepatic vein stenosis requiring stenting occurred in 1 of the 36 type 1 LHV grafts (2.8%) 2 weeks after LDLT and in 1 of the 13 type 2-4 LHV grafts (7.7%) 16 months after LDLT (P = 0.464). These 2 stenting cases underwent LDLT surgery in 2007; thereafter, no such hepatic vein complications occurred, probably because of the maturation of the surgical techniques.
In the aforementioned series, 1 case received an LLS graft with a type 4 LHV: because the graft was not excessively large (the recipient weighed 11.3 kg, and the graft weight was 217 g, so the graft-recipient weight ratio was 1.92), graft hepatic vein reconstruction was successfully performed after wedged unification venoplasty.
In addition, interposition-wedged unification venoplasty for an LLS graft with a type 4 LHV was recently applied to a 10-month-old infant patient weighing 9.6 kg with a diagnosis of hepatoblastoma. He received a 280-g LLS graft from his mother (Fig. 5). This patient was doing well 6 months after transplantation with a patent graft hepatic vein.
LLS grafts seldom present hepatic vein variations of sufficient significance to preclude direct reconstruction because there is a vascular reserve located at the left medial section. In fact, the left medial section is destined to be sacrificed during LLS graft procurement because of the interruption of inflow vessels.[11, 12] Therefore, the MHV/LHV branches within the left medial section can be harvested without additional risk to the donor, and in some cases, even the proximal portion of the MHV trunk is harvested to facilitate hepatic vein reconstruction of the LLS graft. In addition to the lower incidence, this surgical latitude may primarily explain why type 4 LHV anatomy has not been a matter of concern. In our early period of pediatric LDLT, we had a case with a type 4 LHV in which a segment III hepatic vein (V3) branch was ligated at the liver's cut surface and the ventral one-third of the LLS graft was discolored after portal reperfusion. Soon after that, V3 drainage was performed by the in situ placement of an interposition graft between the V3 stump and the recipient's retrohepatic IVC. Since this case, we have paid special attention to LHV variations when we are using LLS grafts. In practice, for living donors with a type 4 LHV, we prefer to harvest a whole left lobe graft rather than an LLS graft when such selection is possible.
However, MHV interruption induces unnecessary derangement of the perfusion of the remnant right liver, and subsequently, atrophy develops in the corresponding area.[13, 14] Thus, it is worthwhile to preserve the MHV flow when an LLS graft is being harvested from a donor liver with MHV dominance. It was of significant interest to determine the incidence of variant hepatic vein anatomy in LLS grafts requiring unification venoplasty. Our simulative analysis using high-volume donor data revealed that interposition-wedged unification venoplasty was necessary in 6.3% of living donors providing an LLS graft. These findings imply that such interposition-unification venoplasty is a surgical option applicable to a small proportion of LLS grafts.
To effectively cope with such variant anatomy in LLS grafts, we performed a series of simulative studies to optimize the surgical techniques for interposition-wedged unification venoplasty. Upon completing these studies, we recognized that the actual shape of the overall hepatic vein reconstruction for such LLS grafts was very similar to the shape of unification quilt venoplasty used to integrate the right hepatic vein orifice and the MHV trunk/segment VIII hepatic vein branch for right liver graft implantation. The reconstruction shapes appeared as mirror images across the IVC. It has been previously demonstrated that the placement of a wide common channel portion is beneficial for accommodating potential graft remodeling–associated distortion of the hepatic vein anastomosis. Therefore, when wedged unification is being performed, the depth of the wedging should be adjusted and optimized after the consideration of the relative sizes of the graft and the recipient's hepatic veins. At this point, we emphasize the importance of size matching between the graft hepatic vein orifice and the diameter of the recipient's IVC. For example, when the diameter of the recipient's IVC is 1.0 cm, a graft LHV orifice with a widest diameter of 2.0 cm can be readily reconstructed with the widely conjoined orifice of the right, middle, and LHV stumps. However, when the graft LHV orifice increases to 3.0 cm, its reconstruction often becomes difficult in a patient with a 1.0-cm IVC, but it is more readily applicable for a 1.5-cm IVC.
To avoid size mismatching and to prevent outflow complications, instead of a direct anastomosis to the conjoined hepatic vein orifice, a triangular anastomosis can be used after the creation of a wide triangular orifice in the recipient's IVC at the confluence of all the hepatic veins as well as a unique and wide longitudinal anastomosis. We think that our unified graft LHV orifice is also suitable for such triangular or diamond-shaped anastomoses.
To perform vascular interposition, the availability of high-quality vein grafts is essential. To our knowledge, femoral vein allografts are the best for medium-sized venous replacement because they have the thickest vein walls, which tolerate vigorous hip movement. In the presented case with a type 4 LLS graft, a cryopreserved iliofemoral vein allograft was selected from the tissue bank at our institution. Other possible materials include the thick-walled external iliac vein, the IVC, the internal jugular vein, and others of similar thicknesses. Very thin-walled vein allografts should be avoided because they are difficult to handle and carry the potential risks of aneurysmal dilatation and spontaneous shrinkage. Artery allografts should also be avoided because most are too stiff or fragile and are thus difficult to handle. The last possible source for a suitable vessel is the donor. Intra-abdominal vessels such as ovarian and gonadal veins are usually too small for this purpose. The use of the greater saphenous vein is a feasible option because parallel stitching or spiral winding can make a large-caliber vein graft comparable to the femoral vein, but its procurement requires an incision of more than 10 cm on the donor's inner thigh. During our experience with more than 200 pediatric LDLT cases, we have harvested a donor's saphenous vein for graft hepatic vein venoplasty in only 1 case so far. Synthetic vessels of any kind should not be used for pediatric patients.
This study was performed as a part of graft standardization efforts to minimize the risk of vascular complications. To date, we have 3 technical approaches to bench work for hepatic vein orifices of left liver grafts for both adult and pediatric patients, including extended left livers, left livers without the MHV trunk, extended LLSs, LLSs, reduced LLSs, and monosegments. The first approach is unification venoplasty with central wedging/septoplasty to make a wide common channel; the second, the placement of a side patch to widen the hepatic vein orifice,[9, 18] is a procedure that can be combined with the first approach; and the third is a very unusual combination of the first approach and vein graft interposition. According to our experience, these techniques cover nearly all hepatic vein variations in various types of left liver grafts. In contrast, we have rarely used side-patch placement for infant patients because such venoplasty often makes the LHV orifice too large in comparison with the size of the recipient's IVC. However, if unification venoplasty is performed without concurrent central wedging or septoplasty, a side patch may be necessary to facilitate an anastomosis.
In conclusion, we have presented the features of LLS hepatic vein variations. Most of these variations can be managed with unification venoplasty, but some rare types require customized interposition-wedged unification venoplasty. We suggest that these techniques represent beneficial surgical options for hepatic vein reconstruction in pediatric LDLT as a part of graft standardization.