Functional Shunt with Small-for-Size Graft in Auxiliary Liver Transplantation for Portal Hypertension

Introduction

PH is a life-threatening complication of end-stage liver disease, primarily driven by increased intrahepatic resistance and hyperdynamic circulation [1]. It carries a significant risk of morbidity and mortality, most notably from acute variceal hemorrhage. The management of PH remains a critical clinical challenge. While multidisciplinary approaches such as pharmacotherapy (eg, non-selective β-blockers) and endoscopic variceal ligation, can alleviate symptoms and prevent bleeding to some extent, they are largely palliative and do not address the underlying disease progression [2,3]. Transjugular intrahepatic portosystemic shunt (TIPS) is a less-invasive technique that creates a direct intrahepatic portosystemic connection and effectively reduces PV pressure (PVP) and PH complications [4]. However, the risk of postoperative hepatic encephalopathy increases significantly. For patients with severe PH and recurrent gastrointestinal bleeding, LT remains the only definitive curative treatment. However, traditional LT requires a stringent GRWR criterion exceeding 0.8% to mitigate the risk of small-for-size syndrome (SFSS) [5]. Due to the hyperperfusion state of the PV, they are at high risk of developing SFSS after LT [6]. This requirement significantly limits the donor pool. ALT presents a potential solution to overcome graft size limitations, which involves implanting a partial graft while retaining part of the native liver. It has been successfully used for acute liver failure and metabolic diseases [7,8]. There are also special forms of ALT, such as domino cross-auxiliary liver transplantation for metabolic diseases, which have been pioneered and applied in our center [9].

Based on our center’s experience, we propose ALT with SFSGs to achieve functional shunt for the treatment of PH (Figures 1, 2) [10], which may expand the donor pool for PH patients who were excluded from traditional LT due to graft size. We have 2 considerations on the indication and therapeutic aims of functional shunt with SFSG in ALT: (1) For patients with pathological PH as the main clinical manifestation, the synthetic and metabolic functions of the diseased liver are in the compensatory stages. The physiological functions of the remnant liver may be utilized to reduce the need for graft size. For instance, the GRWR may be adjusted to 0.5% instead of the commonly recognized 0.8%. (2) The remnant liver has irreversible pathological changes in its structure and functions, whereas the structure of the graft is normal. When there are no immune-mediated injury or vascular/biliary complications, adequate perfusion can be achieved, which enables programmed liver regeneration and eventually achieves a radical cure for PH.

A major safety concern of the functional shunt is whether hyperperfusion injury or fatal SFSS will occur after ALT due to the SFSG, which can be affected by the pathophysiological changes of both the remnant liver and the graft [11]. The relevant variables include liver hemodynamics, synthetic and metabolic functions, and PH stages. However, limited research has been conducted in this area. Thus, this study aims to evaluate the hemodynamic changes and clinical outcomes of functional shunt with SFSG in ALT for PH.

Material and Methods

STUDY DESIGN AND PATIENT SELECTION:

A retrospective, single-center study was conducted on patients with PH who underwent functional shunt with SFSG in ALT at Beijing Friendship Hospital from 2014 to 2018, approved by the Ethics Committee of Beijing Friendship Hospital after rigorous ethics review (2025-P2-439-01). Each case of organ donation from a living donor complied with the relevant laws and regulations on organ transplantation and donation in China, and was approved by the Ethics Committee of Beijing Friendship Hospital after rigorous ethical review. All patients and donors signed the relevant informed consent forms.

ALT SURGICAL PROCEDURES:

In this study, all grafts were obtained from living donors. After partial liver resection, the grafts were perfused with UW (University of Wisconsin) solution or HTK (Histidine-Tryptophan-Ketoglutarate) solution and preserved at 0–4°C. The warm ischemia time was controlled at approximately 2–5 minutes, and the cold ischemia time was controlled at approximately 1–3 hours.

Key technical points of SFSG in ALT: When choosing to preserve either the left or right hepatic lobe, the lobe exhibiting compensatory hyperplasia should be preserved as much as possible, while the severely atrophied lobe should be resected (Figure 2A). During the native hepatectomy, the arterial, portal venous, and hepatic venous stumps should be kept as long as possible. End-to-end anastomosis was used for reconstruction. The outflow tract reconstruction technique was as follows. The orifices of the recipient’s left and middle hepatic veins were enlarged before reconstruction to increase the diameter of the outflow tract (Figure 2B). Anticipating the change in the angle of the outflow tract anastomosis after graft regeneration, the reconstruction angle of the hepatic venous outflow tract was adjusted accordingly, extending the anastomosis towards the right. Similarly, considering the anticipated change in the portal venous anastomosis angle after regeneration, a vascular graft (such as the PV sagittal portion from the resected diseased left lobe) can be used to connect the graft PV to the anterior wall of the recipient’s PV, adjusting the orientation of the graft’s PV after reconstruction. The arterial splitting plan is determined preoperatively, including the number of arterial anastomoses required, and a matching anastomotic site is created on the recipient side to correspond with the graft side [12]. If an arterial vascular graft is needed, the right gastroepiploic artery from either the recipient or donor can be selected for reconstruction [13]. During the operation, we monitored the PVP. After completing the anastomosis, a portal venous pressure catheter was used for direct puncture measurement. If the PVP exceeded 15 mmHg, additional PV modulation was required. The PVP was ultimately controlled to below 15 mmHg. The appropriate method (end-to-end biliary anastomosis or Roux-en-Y hepaticojejunostomy) is selected based on the biliary anatomy (Figure 2C). For more surgical details, refer to our team’s previous reports [10,14].

Immunosuppression was induced with basiliximab. The standard triple therapy including methylprednisolone, calcineurin inhibitor, and mycophenolate was used for postoperative immunosuppression.

DATA COLLECTION:

Data were retrospectively collected from our prospectively maintained institutional database, the China Liver Transplant Registry at Beijing Friendship Hospital. The collected data included:

Baseline characteristics: age, sex, primary disease etiology, Child-Pugh score, Model for End-Stage Liver Disease (MELD) score, history of gastrointestinal bleeding, and presence of ascites.

Intraoperative parameters were: graft type (left/right lobe), graft weight, GRWR, PVP (measured directly after graft reperfusion), operative time, and intraoperative transfusion volume of red blood cells and fresh frozen plasma. Postoperative outcomes were: graft volume (measured by computed tomography at 3 months after LT), major postoperative complications (defined as Clavien-Dindo classification ≥grade III), patient and graft survival, and occurrence of SFSS.

Patients were strictly followed up after surgery. During the follow-up period, vascular Doppler ultrasound was performed 1, 2, 3, 4, 5, 6, 7, 10, 14, 30, 90, 180, 360, and 720 days after functional shunt. We assessed PV velocity, PV sagittal diameter, PV sagittal velocity, and hepatic artery systolic/diastolic velocity. Serum levels of total bilirubin (TB) and albumin (ALB) were assessed routinely. Abdominal drainage volume was recorded daily.

STATISTICAL ANALYSIS:

In this study, we mainly observed the changes in PV perfusion and hepatic artery perfusion in both remnant native livers and graft livers. Since we did not establish an accurate mathematical model of liver fluid mechanics, the Newtonian fluid model was used for trend analysis, which assumes constant viscosity and is standard for clinical volumetric flow estimation in large vessels. The flow velocity and blood vessel diameter were recorded on ultrasound, and the blood flow was calculated using the following formula:

Statistical analysis was performed with Statistical Product and Service Solutions software 27.0 (SPSS, China). Continuous data were tested for normality using the Shapiro-Wilk test. Data conforming to a normal distribution were described as mean±standard deviation (χ̄±SD), while non-normally distributed data were presented as median and interquartile range [M (Q1, Q3)]. Categorical variables were described as frequencies (n) and percentages (%). Given the small sample size (n=6) and the descriptive nature of this study, the primary analysis focused on trend observation and descriptive statistics. Multivariable adjustment was not feasible due to the limited sample size.

Results

HEMODYNAMIC MONITORING IN THE REMNANT NATIVE LIVER AND GRAFT LIVER:

Data collection of the changes of PV flow and hepatic artery velocity in the remnant native liver and graft liver from the patients who underwent functional shunt are shown in Table 3.

In the remnant native liver, the PV blood flow gradually decreased from 7.1 mL/s (IQR: 0–26.0 mL/s) to nearly no perfusion (0 mL/s (IQR: 0–4.9 mL/s) by postoperative day 7, which means that the blood flow buffering in the PV system could only be maintained for about 7 days after surgery. On the contrary, the hepatic PV buffer bed was able to tolerate a certain degree of pressure of PV perfusion about 7 days after surgery and no longer relied solely on the vascular bed of the remnant native liver. After the loss of blood perfusion in the PV of native liver, the effect of hepatic artery buffering response (HABR) was markedly enhanced. The blood flow in the PV of native liver dramatically dropped, along with significantly increased hepatic artery velocity (Table 3, Figure 3A).

The PV blood flow in the remnant native liver disappeared 1 week after functional shunt and the PV blood flow gradually increased in the graft liver. About 1 month after functional shunt, the PV blood flow of the graft liver reached its peak at 47.7 mL/s (IQR: 29.7–53.9 mL/s) and began to decline to hemostasis (Table 3, Figure 3B).

TREND MODEL OF ASCITES DRAINAGE:

The volume of abdominal drainage served as an indirect indicator of PVP and overall fluid shift. Beginning 1 week after functional shunt, the blood flow and pressure in PV of the grafts increased significantly, along with increased abdominal drainage volume. About 2 weeks after the operation, the abdominal drainage volume gradually reached its peak. By postoperative day 30, abdominal drainage had resolved in all 6 patients, which aligned with the normalization of graft PV flow and the alleviation of systemic hyperdynamic circulation, as shown in Figure 4.

CHANGES OF LIVER ANABOLIC AND METABOLIC FUNCTIONS:

Bilirubin metabolism and albumin synthesis are 2 key hepatic biochemical and metabolic markers, which were collected and analyzed (Table 4). During the first week after surgery, TB levels continued to fluctuate at 2 to 5 times the upper limit of normal due to the possible presence of SFSS and other factors, including ischemia-reperfusion injury and immune-mediated injury. About 2 weeks after functional shunt, the TB level dropped to normal, indicating that the metabolic function reserve of the graft was near-normal, as shown in Figure 5A. The recipients’ serum albumin levels began to increase steadily 2 weeks after functional shunt and finally became normal, reflecting the normalization of the overall hepatic synthetic function, as shown in Figure 5B. No recipients had serious irreversible metabolic disorders due to SFSS throughout the postoperative surveillance, which also proved the safety of functional shunt.

Discussion

Patients with PH are at high risk of developing SFSS after LT, which imposes strict requirements on graft size and greatly limits the donor pool. Based on our center’s experience in LT, we propose ALT with SFSGs to achieve functional shunt for the treatment of PH. Our study innovatively applied ALT using SFSG to the treatment of PH, which can expand the donor pool and provide treatment opportunities for more patients with this condition. Clinical observation revealed that functional shunt may be safe and effective in treating end-stage liver disease with PH, although the potential pathophysiological mechanisms and hemodynamics need to be further investigated.

For patients with cirrhosis and severe PH, the most life-threatening complication is variceal bleeding from ruptured esophagogastric varices. Although TIPS can reduce PVP to some extent, the latest meta-analysis has shown that the efficacy of TIPS in patients with recurrent gastrointestinal bleeding is unsatisfactory, with one-third of patients dying within 6 weeks after the procedure and one-fourth developing hepatic encephalopathy postoperatively [15], which also limits its use to some extent.

Unlike TIPS, our surgical approach offers the potential for definitive treatment through the compensatory hypertrophy of the transplanted graft and the atrophy of the diseased native liver, ultimately achieving a curative outcome. In this study, none of the patients showed manifestations of PH after functional shunt.

In traditional LT, postoperative SFSS is a highly dangerous complication that affects both patient and graft survival. Since the graft volume from split liver transplantation or living donor liver transplantation may be insufficient, many centers refuse to perform the surgery. These patients can only continue to wait, during which they may die from recurrent bleeding, thereby losing the opportunity for treatment. Our approach can expand the donor pool by allowing the use of split grafts or SFSGs from living donors, increasing the opportunity for patients to receive lifesaving treatment. However, this procedure also has limitations: it requires extremely precise and high-level technical expertise, which means that the surgeon must have substantial experience in LDLT, as the learning curve is relatively long.

In 2015, the Oslo group proposed the concept of Resection and Partial Liver Segment 2/3 Transplantation with Delayed Total Hepatectomy (RAPID) [16]. This technique uses the regenerative capacity of the graft liver while implementing PV flow modulation to the remnant native liver, effectively integrating and applying the theoretical principles of both auxiliary partial

orthotopic liver transplantation (APOLT), and associating liver partition and portal vein ligation for staged hepatectomy (ALPPS). Compared with the RAPID technique, our technique similarly utilizes graft regeneration and PV flow modulation to achieve a functional shunt, thereby reducing the demand for graft volume and decreasing the risk of postoperative SFSS. In fact, our team has been implementing this technique since 2013.

Functional shunt has 2 aims: (1) to resolve the PH using a SFSG with normal structure and to ultimately cure PH through the regeneration of graft; and (2) the remnant native liver temporarily uses the biochemical synthetic and metabolic functions of hepatocytes, which supplements the functions of SFSG [17]. The remnant native liver can be used as a required supplement to the liver’s synthetic and metabolic functions within a certain period of time [18].

The liver has a dual system of blood supply: the hepatic artery and the PV. About 75% of hepatic blood flow comes via the PV, whereas 25% is supplied by the hepatic artery. When the PV flow decreases, the flow in the hepatic artery increases rapidly, while an increase in PV flow leads to decreased arterial flow. This phenomenon is known as HABR, which reflects the ability of the hepatic artery to produce compensatory flow changes in response to changes in portal venous flow. Since PV perfusion decreases rapidly in the early postoperative period, the physiological functions of the hepatocytes mainly depend on HABR [19,20]. Therefore, after the loss of PV blood flow in the remnant native liver, the hepatic artery perfusion can still maintain the synthetic and metabolic functions of hepatocytes, which serves as a transitional supplement to the biochemical functions of the graft.

The functional reserve of the remnant liver can be assessed by Child-Pugh score, monitoring of indocyanine green clearance (ICG), hepatobiliary iminodiacetic acid (HIDA) scan, and other methods. Most of our patients had liver function of Child-Pugh score A or B, and there was no postoperative synthetic/metabolic decompensation. However, due to the small sample size, more practice will be needed in the future, and at the same time, it is still necessary to rely on precise liver function assessment methods or mathematical models.

The remnant hepatic PV shunt is based on the hemodynamic regulation of the HVPG pressure gradient of the liver on both sides, with an attempt to avoid absolute hyperperfusion injury in the graft. In our study, due to the dominating perfusion in the PV of the graft, the PV flow in the native liver had decreased velocity. Thus, the recipients will still have a considerable degree of PH and visceral hyperdynamic circulation. About 1 month after functional shunt, the PV flow in the graft began to decline and then reached a steady state, indicating that the visceral hyperdynamic circulation throughout the body was relieved. There is a correlation between the amount of ascites produced and the PVP. There are 2 prerequisites for PVP to return to normal: (1) the size of the graft after regeneration is large enough to provide buffer effect for the PV; and (2) the portal resistance is normal. A certain level of PVP helps to promote liver regeneration.

Although the management of ascites after functional shunt is complicated, monitoring of the changes in PV blood flow can be helpful. Malnutrition, surgical trauma, lymphatic vessel breakage, and many other factors in the early postoperative stage led to a rapid increase in ascites production, which continued until the PVP returned to physiological state. Immunosuppressive intervention was applied and the ischemia-reperfusion injury of the liver was repaired, and the portal resistance of the grafts gradually decreased. Meanwhile, graft regeneration increased the vascular bed area in the PV system. As a result, the PV blood flow gradually increased, and the PVP gradually decreased, along with a slow reduction in volume of drainage. However, the systemic visceral hyperdynamic circulation persisted. At 1 month after surgery, the PV blood flow of grafts began to decline, suggesting the systemic visceral hyperdynamic circulation had been alleviated, and the clinical cure of PH was achieved.

In this study, the blood flow was measured using the Newtonian fluid model, in which the blood flows through a straight circular pipe with a constant cross-sectional area. However, the actual hemodynamics of PV contains multiple variables, including density, viscosity, shear stress, Reynolds number, and changing flow patterns. Unlike a Newtonian fluid, the flow patterns change with the vascular geometry. Thus, pilot measurements or computational fluid dynamics (CFD) simulations are required to accurately analyze the impacts of complex circular pipes and flow patterns.

In liver hemodynamic models treated with functional shunt, the biofluid mechanics theories concerning the cardiovascular system and central nervous system have been established, whereas the biofluid mechanics model of the liver has not been identified [21,22]. In this study, ultrasound and radiographic scans in functional shunt cases showed typical features. The PVP of the grafts was maintained at < 15 mmHg during intraoperative monitoring. Continuous ultrasound scans in the early postoperative period showed that the PV blood flow of the grafts increased within 1 week after the operation, along with the speculated PVP >15 mmHg; however, no typical SFSS was clinically observed in these patients. Therefore, the risk factors of traditionally-defined SFSS during a specific functional shunt procedure can be discretely analyzed: (1) the variables of the synthetic/metabolic functions of hepatocytes; and (2) the hydrodynamic variables of blood flow in hepatic PV system [23]. The traditionally-defined SFSS has a higher risk weight as there is an absolute deficiency in the synthetic/metabolic functions of hepatocytes, and the upper limit of the perfusion pressure tolerance of the PV system is relatively broad. The functional shunt model can discretely analyze these 2 risk factors and deduce the definition of SFSS.

Our study has several limitations. First, its retrospective design and very small sample size limit the generalizability of the findings. Second, the calculation of portal flow using the Newtonian fluid model and Doppler ultrasound, while clinically standard, is an estimation and may be influenced by factors such as assumed vessel geometry and operator technique. Third, the study failed to exclude the influence of some confounding factors – for example, the patients’ status and the treatment applied may have an impact on the outcomes. Finally, the single-center nature of the study may have introduced selection bias. Despite these limitations, this study provides unique and detailed serial hemodynamic observations and clinical outcomes on functional shunt. Relevant multicenter, prospective studies are needed to further clarify the efficacy and mechanisms of functional shunt.

Conclusions

Our study provides initial clinical evidence that ALT with SFSGs to create a functional shunt is a viable and safe therapeutic strategy for patients with PH. While our observations suggest that the retained native liver can provide transitional flow buffering and that the HABR might help maintain synthetic/metabolic functions, these interpretations should be considered preliminary due to the study’s limitations, including the small sample size and the lack of a control group. The generalizability of our findings requires validation through larger, controlled studies.

In summary, the functional shunt using SFSGs in ALT may safely expand the donor pool for PH patients who would otherwise be excluded from transplantation due to graft size constraints. Future studies should focus on validating these findings in larger cohorts, establishing optimal patient selection criteria, and developing precise hemodynamic models to predict individual outcomes.