Spontaneous Splenorenal Shunt in Liver Transplantation

OriginalClinicalScience

Spontaneous Splenorenal Shunt in LiverTransplantation: Results of Left Renal VeinLigation Versus Renoportal AnastomosisNicolas Golse, MD,1 Petru Octav Bucur, MD,1 François Faitot, MD,1 Mohamed Bekheit, MD,1,2,3

Gabriella Pittau, MD,1 Oriana Ciacio, MD,1 Antonio Sa Cunha, MD, PhD,1,4,5,6 René Adam, MD, PhD,1,4,5,6

Denis Castaing, MD, PhD,1,2,3,6 Didier Samuel, MD, PhD,1,2,3,6 Daniel Cherqui, MD, PhD,1,2,3,6

and Eric Vibert, MD, PhD1,2,3,6

Background. Management of portal inflow to the graft in patients with spontaneous splenorenal shunts (SRS) is a matter ofconcern especially in case of large varices (more than 1 cm). In case of portal vein (PV) thrombosis (PVT), renoportal anastomosis(RPA) directly diverts the splanchnic and renal venous blood assuring a good portal inflow to the graft. Disconnection of theportacaval shunt by left renal vein ligation (LRVL) is another option but requires a patent PV. The indication of primary RPA ratherthan LRVL in patients with small native PV, especially in case of large graft, should be questioned in these complex cases of livertransplantation.Methods. From 1998 to 2012, 17 patients with RPA and 15 patients with LRVL were transplanted in our center.We compared these 2 techniques for short- and long-term results. Results. The rate of preliver transplantation PVT (76% vs27%) and graft weight (1538 ± 383 g vs 1293 ± 216 g) was significantly higher in the RPA group. Renoportal anastomosis wasperformed in 4 cases of small but patent PV. Three-month mortality, morbidity, and massive ascitis were similar. No patient wasretransplanted. One year after transplantation, PV diameter was still larger in RPA group. Three-year survival was similar (RPA:79% vs LRVL: 53%, P = 0.1). Conclusions. In cirrhotic patients transplanted with large splenorenal shunts, RPA and LRVLreach similar survivals. In case of complete PVTand failure of thrombectomy, the RPA offers satisfactory long-term results.

(Transplantation 2015;00: 00–00)

Whole liver transplantation (LT) requires an hepa-topetal portal inflow of about 100 mL/100 g per

minute to ensure proper liver function.1,2 Portal inflowto the graft partly depends on the patency of the porto-mesenteric venous system, the diameter of the native portal

Received 14 November 2014. Revision requested 3 February 2015.

Accepted 18 March 2015.1 AP-HP Hôpital Paul-Brousse, Centre Hépato-Biliaire, Villejuif, F-94800, France.2 Inserm, Unité 1193, Villejuif, F-94800, France.3 Univ Paris-Sud, UMR-S 1193, Villejuif, F-94800, France.4 Inserm, Unité 776, Villejuif, F-94800, France.5 Univ Paris-Sud, UMR-S 776, Villejuif, F-94800, France.6 DHU Hepatinov, Villejuif, F-94800, France.

The authors declare no funding or conflicts of interest.

N.G. participated in data collection and writing the paper. P.O.B. participated in datacollection. F.F. participated in manuscript rewriting and statistical analysis. M.B.participated in data collection. G.P. participated in data collection and translation.O.C. participated in data collection and statistical analysis. A.S.C. participated inresearch design and manuscript correction. R.A. participated in research designand manuscript correction. D.C. participated in the supervision of the project. D.S.participated in research design and article correction. D.C. participated in researchdesign and manuscript correction. E.V., chief of project, participated in manuscriptwriting and correction.

Correspondence: Eric Vibert, MD, PhD, 12 Avenue Paul Vaillant Couturier, 94804Villejuif Cedex, France. ([email protected]).

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

ISSN: 0041-1337

DOI: 10.1097/TP.0000000000000766

Transplantation ■ Month 2015 ■ Volume 00 ■ Number 00

Copyright © 2015 Wolters Kluwer Health, Inc. Unaut

vein (PV), and the presence of portosystemic shunts. Amongthese shunts, spontaneous splenorenal shunts (SRS, Figure 1)are present in 20% to 35% of LT candidates.3–6 Theseportosystemic shunts are infrequently clinically significant.However, their diameter is associated to postoperative liver fail-ure secondary to portal steal syndrome, PV thrombosis (PVT)or encephalopathy due to liver graft portal hypoperfusion7–9

with an empirical threshold of 1 cm. Prevention of these compli-cations is usually achievedby large shunts division or disconnec-tion by left renal vein ligation (LRVL) in the case of SRS.10

In case of completely thrombosed11 or small diameter PV,the feasibility and efficiency of portoportal anastomosis (PPA)is uncertain, and large portosystemic shuntsmay have to be pre-served to perform a renoportal anastomosis (RPA).12–14Mesen-teric, splenic, and renal flows ensure adequate portal inflowafter realization of an end-to-end anastomosis between the re-cipient left renal vein (LRV) and the graft PV.

Pre-LT radiological shunt embolization has been re-ported representing an alternative to intraoperative com-plex procedures.3,15

In the presence of a patent recipient PV and large SRS,LRVL with end-to-end PPA is the usual approach. However,when recipient PV is small and/or with poor wall quality, aprimary RPA rather than LRVL might be a better strategyto prevent graft hypoperfusion especially in case of a largegraft. Nonetheless, this strategy will only be justified if theRPA results in patients without another alternative becauseof a thrombosed portal trunk are at least similar to thoseof patients with a large SRS, PPA, and LRVL. The aim of

www.transplantjournal.com 1

horized reproduction of this article is prohibited.

mailto:[email protected]

FIGURE 1. Left renal vein ligation and renoportal anastomosis. IVC indicates inferior vena cava; K, kidney; S, spleen.

2 Transplantation ■ Month 2015 ■ Volume 00 ■ Number 00 www.transplantjournal.com

this study was to compare immediate and long-term out-comes of RPA and LRVL, 2 strategies that are implementedin patients with large spontaneous SRS. The feasibility ofthe flowmetry will also be discussed in this particular situa-tionwhere SRS are present andwhere a reliablemeasurementwould only be feasible after completion of a temporaryportacaval anastomosis.

MATERIALS AND METHODS

Patient Selection

Between January 1, 1998, and December 31, 2012, 1574LT were performed at the Centre Hépato-Biliaire (HospitalPaul Brousse, Assistance Publique des Hôpitaux de Paris[APHP], Université Paris 11, Villejuif, France). During thisperiod, 17 RPA (1.1%) and 15 PPA with LRVL (PPA-LRVL) (0.95%) were performed accounting for 2% of theentire transplanted cohort. These 32 patients were retrospec-tively analyzed. They all displayed a spontaneous SRS greaterthan 1 cm on the pre-LT computed tomography (CT) scan.The temporal distribution of RPA was homogeneous (1 ±1.24 procedures/year) during the study period, whereas93% of PPA-LRVL were performed after 2009.

Demographic Data on Donors and Recipients

Donor's data included morphological data (body mass in-dex in kg/m2), sex, age, and liver laboratory tests (aspartateaminotransferase, alanine aminotransferase, γ glutamyltransferase, total and conjugated bilirubin levels and pro-thrombin time (PT), blood group, and cytomegalovirusstatus).

Recipient's preoperative data included demographic dataat the time of transplantation (sex, age, body mass index,blood group, and cytomegalovirus status), model of endstage liver disease (MELD),16 and age of donor � MELDscores,17 liver disease etiology, oncological data (size andnumber of lesions, treatments received), and a preoperativeCT scan analysis.

Intraoperative data included type of LT (iterativeLT, liver +kidney transplantation, 2-adult split, domino LT, living do-nor transplant), portal thrombectomy attempt and result, type

Copyright © 2015 Wolters Kluwer Health, Inc. Unau

of portal revascularization (RPAvs LRVL), graft weight, graft-to-recipient weight ratio, operative time, cold ischemia time,and transfusion rate. Postoperatively, hospital and intensivecare unit (ICU) length of stay, type, and grade of surgicalcomplications according to Clavien-Dindo classification,18

short- and long-term medical complications, including acuterejection, infections and PVT, duration of follow-up and sta-tus at the last follow-up.

Biological data included pre-LT assessments and post-operative days (PODs) 1, 3, 5, 7, 10, 30, 60, and 90 (aspar-tate aminotransferase, alanine aminotransferase, bilirubin,PT, creatinine). Pathological data on the graft determinedon a systematic liver biopsy performed at the end of LTwere analyzed to evaluate the rate of steatosis andischemia-reperfusion (I/R) injury. Biopsies were formalin-fixed paraffin-embedded and stained with hematoxylin-eosin and picrosirius. All the biopsies were reviewed bythe same (blinded) pathologist. Steatosis was defined asmicrovesicular or macrovesicular, and classified as absent,mild (<30%), moderate (30-60%), or severe (≥60%).Fibrosis of portal tracts and centrilobular veins as well asperisinusoidal fibrosis were recorded. Grafts were also evalu-ated for the presence of I/R injury defined by the combinationof apoptotic hepatocytes and inflammatory neutrophilicpolynuclear's infiltrate throughout the lobule and aroundcentrilobular veins. An I/R injury was graded as absent,mild, moderate, or severe. Severe I/R injury requiredcentrocentrilobular necrosis.

When digital imaging was available in the medical records(RPA n = 9 and LRVL n = 14), an exhaustive morphometricanalysis was performed using Myriam XP-Liver 1.14.1 soft-ware (Intrasense, Montpellier, France). On the preoperativeCT scan, performed less than 6 months before LT, weassessed PV patency (PVT was graded according to Yerdel'sclassification),19 spleen volume and splenic artery/PV/LRV/splenic vein/suprarenal and infrarenal vena cava diameters.

On repeated postoperative CT scans performed from thefirst week until POD 365, we assessed suprarenal andinfrarenal vena cava diameters and the venous diameter up-stream of the portal anastomosis (LRV in the case of RPAand native PV in the case of LRVL). Changes of (1) liver

thorized reproduction of this article is prohibited.

© 2015 Wolters Kluwer Golse et al 3

and spleen volumes and (2) PV diameters were evaluated upto 1 year.

Surgical Procedures

Deceased donor LT,20 living donor LT,21 split liver LT,22 ordomino LT23 were performed using previously describedtechniques. Left renal vein was controlled before total hepa-tectomy but ligated and/or divided only after completion ofcaval reconstruction to prevent any splanchnic congestion.13

The effect of the LRVL on the portal flow was visuallyassessed before performing the portal anastomosis. The por-tal clampwas briefly released before and after LRV clamping,and the blood flowwas qualitatively evaluated as stronger ornot. The LRV clamping was considered efficient if the bloodflow was stronger after clamping. In this latter case, portalanastomosis was performed with a posterior running sutureand anterior interrupted suture (5-0 or 6-0 Prolene), andLRV was secondarily ligated with 3-0 Mersuture. TheLRV was not sectioned to leave the possibility for switchingto an RPA if necessary. An RPA was performed if:(1) thrombectomy of the native PV was not feasible or(2) the portal flow remained low after LRV clamping andthe ligation of other non-SRS portosystemic shunts. The fail-ure of LRV clamping was interpreted as an insufficient diver-sion of renal flow through the SRS, thus requiring completeand direct diversion of it through an RPA. The end-to-endRPAwas thus achieved with a posterior running suture andanterior interrupted suture (5-0 or 6-0 Prolene) between thegraft PV and the native LRV (Figure 1). The distal part ofthe LRV was sectioned just upstream of its insertion on thevena cava, thus enabling a thicker vena cava ring to facilitateanastomosis with the PV. On the vena cava, the ostium of theLRV was closed with a running suture of 5-0 Prolene. Thistechnique does not represent any technical difficulties anddoes not need specific technical skill. An interposed vein graftwas used (autologous jugular) only in 1 case because the graftPV was too small.

At the end of the intervention, an intraoperative Dopplerultrasonography (DUS) was systematically performed tocheck for adequate arterial and portal flow spectrum. Portalflowmetry was not performed on native PVas it was not rel-evant in the post-LT setting given the intrahepatic resistanceof the native liver. As temporary portocaval shunt was notroutinely performed—and potentially not necessary in caseof spontaneous portosytemic shunts—the impact of LRVLwas not assessable before portal anastomosis thus reducingthe interest of flowmetry as a decision tool.

Short-Term Postoperative Management

After LT, all patients were systematically transferred to theICU. Until POD 7 daily DUS, blood tests and close clinicalsurveillance were conducted. Doppler ultrasonography in-cluded an assessment of the arterial resistive index and thedirection of portal flow. All patients received similar postop-erative intensive care with a routine triple immunosuppres-sive regimen that included corticosteroids, mycophenolatemofetil, and tacrolimus or cyclosporine, with basiliximab(day 1 and day 4) in the event of rising or initially high serumcreatinine levels. Massive ascites was defined as drain output> 1 L/24 hours for more than 2 weeks.

Prophylaxis against deep venous thrombosis included com-pression stockings, early mobilization and administration


of anticoagulation. Anticoagulation was initiated in the ICUas soon as possible depending on the patient's clinical status(no evidence of bleeding), coagulation tests (PT > 30%),and platelet counts (>50,000/mL), usually within 24 hoursof transplantation. The rate of the heparin infusion was ad-justed to obtain an activated partial thromboplastin time 1.5to 2 times higher than the reference level. This was replacedby a daily dose of low-molecular weight heparin once thepatient had moved to the ward and was administrated untildischarge. Heparinization also reduced the risk of earlyrethrombosis, especially after thrombectomy associated to en-dothelial lesions.

After their discharge, prophylactic long-term aspirin ther-apy (250 mg/day) was given to all patients to reduce the riskof arterial thrombosis.

Before 2007, CT-guided angiography was only performedin the event of DUS abnormalities and/or liver cytolysis. After2007, a CT-guided angiography was performed systemati-cally between POD 7 and POD 10 unless there were contra-indications to the injection of iodinated contrast medium.

Long-Term Postoperative Management

After their discharge, all patients visited our outpatientclinic regularly: every fortnight for the first 2 months, andthen once every 3 months. Follow-up consisted of a physicalexamination, liver function tests and DUS. The CT-guidedangiography was performed in the event of lower arterial in-dex resistances and/or biological abnormalities to assess vas-cular patency.

Data Analysis

All quantitative data are expressed as mean ± SD. TheMann-Whitney U test for continuous variables and the χ2

or Fisher exact test for categorical variables were used whenapplicable. Survival rates were calculated using the Kaplan-Meier method, and groups were compared by means of thelog-rank test. Statistical analyses were performed using SPSS20.0 (SPSS Inc, Chicago, IL). A P value less than 0.05 wasconsidered to be statistically significant.

RESULTS

Donor and Recipient Features

Demographic, biological, and morphological data (in do-nors and recipients) did not reveal any clinically relevant dif-ferences. Only donor γ glutamyl transferase and bilirubinlevels were statistically significantly different between the2 groups. Indication and type of LT (split, domino, living)were similar. One partial LT was performed in each group,and 2 patients were transplanted under emergency conditionsfor liver insufficiency: one for posthepatectomy liver failure(LT at POD 13 posthepatectomy for hepatocellular carci-noma [HCC] in the RPA group), and the second for decom-pensation of Wilson disease (PPA-LRVL group) (Table 1).

Morphometric Data Before Transplantation

Before LT, PVT rates were significantly higher in the RPAgroup (RPA vs PPA-LRVL: 13 [76%] vs 4 [27%] patients,P = 0.01). Three patients in each group had PVT graded IIIaccording to Yerdel classification, whereas 10 patients inthe RPA and 1 patient in the PPA-LRVL had grade II PVT.In patients without a complete PVT (including partiallythrombosed PV) on the preoperative CT scan, the PV


TABLE 1.

Preoperative donor and recipient features

Variable RPA (n = 17) PPA-LRVL (n = 15) P

Donors Age, (years) 41.5 ± 16.3 47.5 ± 20.3 0.37BMI, kg/m2 26.3 ± 6.4 24.5 ± 3.9 0.74Sex ratio (M/F) 13/4 7/8 0.26Deceased/domino/split/LDLT 15/1/0/1 13/1/1/0 0.87Biological parametersAST, UI/L 99 ± 146 45 ± 18.5 0.47ALT, UI/L 46.3 ± 36.9 56 ± 82.5 0.42GGT, UI/L 32 ± 32 61 ± 43 0.03Bilirubin, μmol/L 17 ± 9 11 ± 6 0.04PT, % 71 ± 18 67 ± 20 0.75

Recipients Age, y 45.9 ± 13.8 41.2 ± 18.8 0.41BMI, kg/m2 25.5 ± 5.2 26 ± 3.3 0.73Sex ratio (M/F) 14/3 11/4 0.67MELD 20.8 ± 7.1 18.3 ± 8 0.4D-MELD 856 ± 460 882 ± 550 0.89Biological parametersCreatinine, μmol/L 104 ± 84 119 ± 1 59 0.1Platelets, 103/mL 147 ± 200 74 ± 34 0.07Indication for transplantationCancer 7 (41%) 6 (40%) 1Cirrhosis 9 (53%) 8 (53%) 1Other 1 (6%) 1 (7%) 1

BMI indicates body mass index (kg/m2); M, male; F, female; LDLT, living donor liver transplantation; AST, aspartate aminotransferase; D-MELD, age of donor � MELD score; GGT, γ glutamyl transferase.Continuous variables are expressed as mean ± SD.


was larger in the PPA-LRVL group (0.99 cm vs 0.56 cm,P = 0.005). All other measures were comparable, particu-larly, the volume of the spleen (Table 2).

Surgical Procedures

Portal thrombectomywas successfully performed in all pa-tients with PVT in the PPA-LRVL group (n = 4), whereas nocases were achieved in the RPA group. One patient in thePPA-LRVL group required intraoperative ligation of amesentericocaval shunt.

A comparison of the intraoperative data revealed larger(1538 ± 383 g vs. 1293 ± 216 g, P = 0.04) and lower-quality liver grafts in the RPA group, in which severe I/R in-jury and macrosteatosis were more common. Only 2 grafts

TABLE 2.

Pre and postoperative morphometric data

Variable RPA

Pre-LT morphometric analysisPortal vein thrombosis (% of whole cohort) 7Spleen volume, cm3 835Portal vein diameter, cm* 0.56Renal vein diameter, cm 2.09Splenic vein diameter, cm 1.21Splenic artery diameter, cm 0.81Suprarenal/infrarenal vena cava ratio diameter 1.40Post-LT morphometric analysisVenous diameter upstream of portal anastomosis, cm 1.55Suprarenal/infrarenal vena cava ratio diameter 1.22Liver volume (CT scan < POD 10) 1959

Continuous variable are expressed as mean ± standard deviation (DS).*In patients with patent portal vein.


had macrosteatosis greater than 30%, and these livers werein the RPA group. Operating and cold ischemia times weresimilar in the 2 groups.

In the 2 cases of split LT (one in each group), the partial liverweights were 945 g and 1300 g, respectively (with respectivegraft to recipient body weight ratios of 1.3 and 1.8) (Table 3).

Graft and Patient Survivals

Mean survival reached 126 ± 19 months and 69 ± 18 afterRPA and PPA-LRVL, respectively. Median survival was133 months in the RPA group and not achieved in the PPA-LRVL group. There was no difference in the overall survivalbetween the 2 groups (P = 0.1). Survival rates at 3 months, 1,2, and 3 years were, respectively, 100%, 88%, 88%, and

(n = 9) PPA-LRVL (n = 14) P

6.5 26.7 0.01± 508 1261 ± 1017 0.26± 0.2 0.99 ± 0.27 0.005± 0.49 2.07 ± 0.5 0.92± 0.25 1.22 ± 0.24 0.93± 0.21 0.81 ± 0.25 0.76± 0.15 1.47 ± 0.27 0.65

± 0.39 1.18 ± 0.29 0.02± 0.16 1.15 ± 0.25 0.3± 460 1687 ± 444 0.19


TABLE 3.

Intraoperative data

RPA (n = 17) PPA-LRVL (n = 15) P

Iterative LT, % 23.6 0 0.1LT + KT, % 23.5 13.3 0.66No. transfused blood units 7.3 ± 6.9 4.1 ± 4.2 0.13No. transfused fresh frozen plasma units 11.1 ± 9.7 5.4 ± 7.3 0.038No. transfused platelet units 1 ± 2.6 0.13 ± 0.13 0.14Operating time, min 594 ± 180 619 ± 145 0.69Cold ischemia time, min 462 ± 140 450 ± 147 0.83Native liver weight, g 1374 ± 874 1257 ± 506 1Graft weight, g 1538 ± 383 1293 ± 216 0.04Graft/recipient weight, % 2.09 ± 0.54 1.75 ± 0.32 0.045Graft weight/splenic volume 2.81 ± 1.58 1.99 ± 1.5 0.23Microsteatosis (day 0), % 17.3 5.3 0.04Macrosteatosis (day 0), % 15.3 0.7 0.0002Severe I/R injury, % 35.3 6.7 0.09

Continuous variable are expressed as mean ± SD.


79% in the RPA group, and 93%, 67%, 53%, and 53% inthe PPA-LRVL group (Figure 2).

Morphometric Data After Transplantation

At the first postoperative CTscan, the diameter of the veininfusing the graft (LRV in the case of RPA and PV in the caseof PPA-LRVL) was larger in the RPA group (1.55 cm vs. 1.18cm, P = 0.02). This difference persisted with statistical signif-icance until the 12th month (Figure 3).

We also noted a regression of the suprarenal/infrarenal in-ferior vena cava ratio, whatever the group.

The evolution of graft volumes was similar in the2 groups during the first year and the kinetic of regression

FIGURE 2. Long-term survival after portoportal anastomosis with left rencumulative survival.


of splenomegaly was the same in both groups during the firstyear post-LT (data not shown).

Immediate Postoperative Biological Findings

An analysis of renal function showed, after exclusion ofliver + kidney recipients, a marked (but temporary) rise in se-rum creatinine levels in the RPA group, despite identicalpre-LT renal function. The long-term renal failure rate wasnot significantly different (P = 0.16) (Figure 4).

Peak cytolysis was higher in the RPA group (P = 0.01),although levels were only statistically different on PODs 1and 3 (Bonferroni post test). The findings were subsequentlycomparable.

al vein ligation versus renoportal anastomosis. Cum survival indicates


FIGURE 3. Left: Evolution of recipient portal vein and left renal vein diameters after liver transplantation. Right: Evolution of graft volumes untilthe 12th month. M indicates month.


The evolution of PT rates and thrombocytopenia andbilirubin level course were strictly identical in both groups(P = 0.5, P = 0.24, and P = 0.6, respectively).

Postoperative Mortality

In-hospital mortality (Clavien V) was 1 in the RPA group(5.9%) and 1 in the PPA-LRVL group (6.7%). In the RPAgroup, a 50-year-old patient died after transplantation forhepatitis B cirrhosis. Outcomes were complicated by perito-nitis due to ileal perforation, right hepatic artery thrombosisnecessitating repeat surgery (fungal thrombus), and respira-tory distress syndrome requiring prolonged mechanicalventilation. Death occurred on POD 101, secondary to amassive periventricular brain haemorrhage. In the PPA-LRVLgroup, a 52-year-old patient died after liver and kidney

FIGURE 4. Postoperative evolution of biological parameters. ALT indica


transplantation for hepatitis C cirrhosis with cryoglobulinemia.His immediate outcome was complicated by hemorrhagicshock secondary to a retropancreatic artery rupture (poten-tially secondary to intraoperative excessive mobilization ofthe duodenopancreatic block), and toxic shock with bacter-emia (Escherichia coli), leading to terminal renal failure.Death occurred on POD 36, secondary tomultiorgan failure.Neither of these 2 deaths was directly attributable to the por-tal revascularization technique.

Nonspecific Postoperative Morbidity

The incidence of severe complications (Clavien≥ III) wassimilar in the 2 groups (RPA: 35% vs PPA-LRVL: 53%;P = 0.47), as were the length of ICU stay (14 ± 16 days vs

tes alanine aminotransferase; D, day; KT, kidney transplantation.



16 ± 18 days; P = 0.9) and hospital stay (34 ± 19 vs 35 ±23 days; P = 0.69).

The frequency of rejections was not statistically different,despite a trend toward more episodes in the RPA group (re-jections before POD 90: 23% vs 0%; P = 0.1, and rejectionsafter POD 90: 35% vs 7%; P = 0.08). The rates of infections(all types) during the initial hospitalization were identical inthe 2 groups (23% vs. 40%; P = 0.45). The incidence of he-patic arterial stenosis requiring invasive management beforePOD 90 was 6% (surgical management in the RPA group)versus 13% (radiological procedures in the LRVL group;P = 0.59). One patient in the PPA-LRVL group presentedwith splenic artery steal syndrome complicated by hepatic ar-tery stenosis. He required radiological splenic embolizationon POD 60, which was successful.

Specific Postoperative Morbidity

There was no significant difference in post-LT PV thrombo-sis (RPA: n = 2 [12%] vs PPA-LRVL: n = 1 [7%], P = 1), andthere was no correlation between the presence of a pre-LTPVT and the occurrence of post-LT PVT. In the RPA group,PVT concerned: (1) a 39-year-old woman transplanted fornodular regenerative hyperplasia in a context of pre-LT PVTand HIV infection. She presented an early portal rethrombo-sis (not treated) associated with both acute and chronicrejections. Clinical features associated massive portal hyper-tension (PHT) with iterative bleeding episodes leading tomultiorgan failure and death on POD 300 (after dischargefrom hospital), and (2) a 31-year-old male patient with HCC(on degenerated polyadenomatosis) and portal cavernoma.Renoportal anastomosis was carried out using an interposedjugular graft (no autologous femoral vein was available be-cause of length and quality insufficiencies). Portal vein throm-bosis was diagnosed on POD 3, and portal angioplasty withstenting was performed on POD 6. His surgical outcomewas satisfactory but a recurrence of HCC occurred duringthe 19th month.

Note should also be made of the case of a 63-year-old pa-tient in the LRVL group (with no pre-LT PVT) who wastransplanted for alcoholic and post-hepatitis C cirrhosis. Hepresented with an early PVT requiring a rescue RPA (withthe interposition of an iliac venous graft) on POD 9. Thepostoperative course was uneventful. This patient continuedto be included in the PPA-LRVL group (analysis underintention-to-treat).

The incidence of postoperative ascites was similar in the 2groups (12% vs 13%; P = 0.7).

Long-Term Clinical and Pathological Results

During the first year, systematic histopathological analysesfound 2 cases of nodular regenerative hyperplasia (1 case ineach group) without any other vascular disease or architec-tural disorder potentially related to portal flow disturbance.

The etiologies of death (after hospital discharge) were:

– RPA group: multiorgan failure in a context of early RPAthrombosis (not treated) and chronic rejection leading tosevere PHT (n = 1, POD 300), multifocal metastatic HCC(n = 1, POD 900) and decompensated cirrhosis of unknownetiology (n = 1, POD 4000).– PPA-LRVL group: pulmonary sepsis on POD 120 (n = 2,Pneumocystis carinii and Legionella pneumophila), HCC


recurrence (n = 1, POD 180), and metastatic bladder cancer(n = 1, POD 560).

Preoperative and Intraoperative Prognostic Factors forSurvival

In the global cohort, negative factors for overall survivalwere a pre-LT small LRV diameter (P = 0.029) and operativetime (P = 0.01).

Compared to the PPA-LRVL group, the RPA group tendedto achieve better survival in the event of pre-LT PVT (P =0.08). There were no differences in survival in the case of pat-ent pre-LT PV (P = 0.2).

DISCUSSION

In this study, we showed that RPA and PPA-LRVL are ef-fective in achieving suitable portal inflow to the graft in thepresence of a spontaneous large SRS. We also demonstratedthat RPA could be a safe procedure in case of incompletePVT, small diameter PV, and large graft.

Preoperative PVT is a common situation in the setting ofLT, affecting between 10% and 20% of patients.19,24 Its inci-dence is closely correlated with the MELD score,25–28 PVToften occurring in the context of very advanced cirrhosis.With increasingly high MELD scores at listing,29 the inci-dence of SRS has reached approximately 30%.4,5Despite thishigh incidence, clinical significance is rare (2% in our prac-tice). The management of SRS remains a source of contro-versy because no consensus has yet been reached as to theindications and modalities of care.30–32 In this context, wecompared 2 cohorts of transplanted patients with spontane-ous SRS according to portal inflow management. The studywas designed after the occurrence of early PVT after LRVLthat were subsequently converted to RPA, also consideringthe high graft loss rate after post-LT PVT. Indeed, we won-dered if primary RPA, in the presence of a patent but smallPVand a large graft, might decrease the risk of post-LT PVT.

Renoportal anastomosis is classically reserved to patientswith complete PVT11 with SRS, using the LRVas a splanch-nic venous tributary. The RPA is technically easy and feasibleby any liver transplant surgeon particularly in case of largegrafts when the PV can be more easily brought to the levelof the LRV.

We herein report the largest series of patients undergoingRPA and PPA-LRVL for spontaneous SRS in deceased donorLT, thus excluding series evaluating only living donor LT.33

One of the most interesting findings of this work was thepost-LT vascular morphometric analysis, which showed thatthe diameter of the inflow vessel to the graft (LRV in case ofRPA) was larger than the native thrombectomized PV (PPA-LRVL group) theoretically allowing a higher portal flow.This is of major hemodynamic importance since it is knownthat resistance to a fluid is inversely proportional to the ra-dius of a vessel. Therefore, with a given differential pressure,flow would be higher in a larger vessel. It should also benoted that the liver volumes were equivalent after POD 30,and the regression of splenomegaly (up to 1 year) was similarin both groups, indicating a similar reduction in left PHTbetween the groups.

No primary nonfunction or emergency retransplantationoccurred among the patients.



Given the retrospective aspect of the study, the 2 groupsdiffered for the anatomy and physiology of PHT. Indeedpre-LT partial PVT was significantly more frequent in RPAgroup (76% vs. 27%; P = 0.01) and PV diameter was largerin the PPA-LRVL group (0.34 cm vs. 0.84 cm; P = 0.01).There was also a trend toward a larger number of iterativeLTs and combined liver-kidney grafts in the RPA group.Moreover patients with RPA were transplanted with largerand more steatotic grafts which are associated with higherpost-LT cytolysis34 and higher postoperative complications,such as renal insufficiency,35,36 respectively. Despite these dif-ferences in favor of PPA-LRVL, the overall survival and post-operative course as well as long-term morbidity andmortality were similar. Even if survival is the result of con-founding medical and environmental factors, this confirmsthat RPA is an effective and reliable technique, especially inseverely ill cirrhotic patients. Of note, one of the patientswho underwent an RPA on POD 9 after a PPA-LRVL (earlyPVT) was included in the PPA-LRVL group on an intent-to-treat basis. However, his long-term prognosis is linked tothe RPA procedure. When analyzing the data on a per-protocol basis, the difference of survival between RPA andPPA-LRVLbecomes significantly (P = 0.047) in favor of RPA.

The difference found on the liver weights between bothgroups reflects the frequently observed mismatch betweenthe expected portal flow in the presence of an enlarged liverand the one obtained after LRVL, prompting us to chooseanother method of revascularisation to ensure better flow(i.e., RPA).

The rate of postoperative complications was similar in the2 groups despite a trend toward more rejection, renal failureand cytolysis (ALAT levels) in the RPA group. The lowerprevalence of rejections during the second part of the study(PPA-LRVL) can be explained, at least in part, by modifica-tions to the immunosuppressive regimen, leading to globallyimproved management. The same bias of period may alsohave explained the markedly worsened renal function in theRPA group (earlier period, adverse effects of cyclosporine).However Marubashi et al37 reported 100% temporary renalimpairment after RPA in 3 patients. The second explanationcould have been the high rate of graft steatosis seen in theRPA group, as steatosis was correlated with early post-LTrenal dysfunction in two recent series.35,36 The vascular

FIGURE 5. Decisional algorithm for portal revascularisation in case of p


anatomy of the left kidney allows ligation of the LRVwithoutamajor risk of venous ischemia due to possible derivations ofthe exorenal circle38,39 (diaphragmatic, adrenal, gonadic andazygos veins), although a temporary increase in the size of theleft kidney has been reported after LRVL with ipsilateral de-creased function.33 This questions the outflow capacity afterLRVL, at least during the initial phase. Most series of PPA-LRVL have indicated some degree of temporary and rapidlyregressing renal insufficiency. The recent report regardingPPA-LRVL on a single left kidney confirmed initial venous is-chemia before development of the exorenal circle.40

Two patients in each group presented post-LT ascitis. Inboth group, 1 patient had small renal vein caliber, thus indi-cating poor pre-LT SRS efficiency and predicting poorpost-LT portal inflow (i.e., persistent PHT). The second pa-tient in RPA group had no morphometric analysis. In thePPA-LRVL group, the second patient presenting ascitis hada huge spleen (3150 cm3), requiring post-LT radiologic em-bolization. These 3 situations illustrate the necessity of an ad-equate portal inflow to limit post-LT PHT. In our initialexperience (not reported in this series), we performed 5RPA in patients with complete PVTwithout SRS. All patientspresented persistent ascitis and severe complications. The ab-sence of SRS has since been considered to be an absolute con-traindication to RPA in our team.

One may question the place of cavoportal hemitransposition(CPHT) in themanagement of pre-LT PVTwith SRS. This tech-nique allows the revascularisation of the graft with systemicblood but mesenterical flow is insufficiently diverted into thegraft, even in the presence of portosystemic shunt. Since 1996,we performed 4CPHTwith calibration of the retrohepatic venacava. It is a feasible procedure but 3 patients had severe postop-erative complications related to persistent PHT (ascites, esopha-geal, and rectal variceal bleeding), and one patient had an earlyCPHT thrombosis requiring re-LT. After a median follow-up of15 (6-120) months, 1 patient died of chronic rejection, and2 patients died in a context of severe and persistent PHT. Otherteams reported lower torso edema and pulmonary embolism.Because of the high morbi-mortality rate, we do not promoteanymore this demanding technique. Patients with diffuse PVTwithout surgical or spontaneous portosystemic derivation re-main a contraindication to LT in our center. In the presence ofSRS, the RPA or the LRVL should be evaluated because of

ortal vein thrombosis and splenorenal shunt.



encouraging results. The algorithm presented summarizes ourmanagement in case of PVTwith SRS (Figure 5). External vali-dation is needed to confirm on larger series its efficiency.

This is a retrospective study comparing 2 nonrandomizedtechniques and thus presents some limits. First, RPAwas pref-erentially used when the PV could not be used (completethrombosis or small diameter). Second, the small sample ineach group should lead to cautiously analyse the results. Nomultivariate analysis could be performed because of the co-hort was too small. Lastly, the nonhomogeneous inclusionperiod implies a longer follow-up in the RPA group.

A further limitation of this work was the absence of intra-operative flowmetry findings. Intraoperative portal flowmetrymeasurements could have theoretically helped us to determinethe need for an SRS occlusion and to evaluate the effects ofLRV clamping. To be clinically relevant, flow should only havebeenmeasured after graft implantation (leading to dismantlingof the PPA in the event of poor flowmetry findings) or duringthe anhepatic phase with a temporary portocaval shunt. How-ever, this procedure is not performed routinely in our centre inthis particular population because portal clamping is generallyvery well tolerated (spontaneous SRS). This explains why weconsidered that intraoperative flow measurements would notfundamentally modify the global management of SRS, and itwas therefore not usually used during transplantation. Themanagement of portosystemic but non-splenorenal shuntsmay benefit from an intraoperative portal flowmetry, indicat-ing an increase in the flow after temporary shunt closure andportal steal removing. This situation was not evaluated inour study.

This descriptive series should be confirmed by furtherevaluations, including randomization when both techniquesmay be proposed (patent PV, initially or after thrombectomy).However due to the scarcity of patients in this situation, thisproject remains very theoretical. The following algorithmmay be proposed (Figure 5): the RPA is the only technique fea-sible in case of complete PVTwith SRS. When thrombectomy

FIGURE 6. Example of pre-LT small portal vein diameter with largeleft renal vein. Ideal situation for an efficient renoportal anastomosis.


allows restoration of the portal flow but SRS is greater than1 cm, 2 alternatives should be balanced: (1) PPA-LRVL in caseof small graft to recipient weight ratio (cutoff to be deter-mined), no residual PVT nor hypoplastic PV and (2) RPA incase of small PVand large graft (Figure 6). The impact of thepre-LT renal function has to be evaluated.

CONCLUSIONS

Pre-LT PV patency and the presence of SRS need to becarefully evaluated to anticipate any intraoperative portal re-vascularization difficulties. In the presence of a large SRS,RPA appears to be as efficient as LRVL in revascularizingthe graft. In the event of a small diameter and partiallythrombosed PV, or complete PVT with the failure ofthrombectomy, LRVL cannot be used, whereas RPA offersan opportunity for satisfactory long-term results.

REFERENCES1. Spitzer AL, Dick AAS, Bakthavatsalam R, et al. Intraoperative portal vein

blood flow predicts allograft and patient survival following livertransplantation. HPB (Oxford). 2009;12(3):166–173.

2. Ishizaki Y, Kawasaki S, Sugo H, et al. Left lobe adult-to-adult living donorliver transplantation: Should portal inflow modulation be added? LiverTranspl. 2012;18(3):305–314.

3. Awad N, Horrow MM, Parsikia A, et al. Perioperative management ofspontaneous splenorenal shunts in orthotopic liver transplant patients.Exp Clin Transplant. 2012;10(5):475–481.

4. Cherqui D, Panis Y, Gheung P, et al. Spontaneous portosystemic shunts incirrhotics: implications for orthotopic liver transplantation. Transplant Proc.1993;25(1 Pt 2):1120–1121.

5. Chikamori F, Nishida S, Selvaggi G, et al. Effect of liver transplantation onspleen size, collateral veins, and platelet counts. World J Surg. 2010;34(2):320–326.

6. Ploeg RJ, D'Alessandro AM, Stegall MD, et al. Effect of surgical andspontaneous portasystemic shunts on liver transplantation. TransplantProc. 1993;25(2):1946–1948.

7. Tallón Aguilar L, Jiménez Riera G, Suárez Artacho G, et al.Posttransplantation portal thrombosis secondary to splenorenal shuntpersistence. Transplant Proc. 2010;42(8):3169–3170.

8. Braun MM, Bar-Nathan N, Shaharabani E, et al. Postshunt hepaticencephalopathy in liver transplant recipients. Transplantation. 2009;87(5):734–739.

9. De Carlis L, Del Favero E, Rondinara G, et al. The role of spontaneousportosystemic shunts in the course of orthotopic liver transplantation.Transpl Int. 1992;5(1):9–14.

10. Castillo-Suescun F, OniscuGC, Hidalgo E. Hemodynamic consequences ofspontaneous splenorenal shunts in deceased donor liver transplantation.Liver Transpl. 2011;17(8):891–895.

11. Paskonis M, Jurgaitis J, Mehrabi A, et al. Surgical strategies for livertransplantation in the case of portal vein thrombosis—current role ofcavoportal hemitransposition and renoportal anastomosis. ClinTransplant. 2006;20(5):551–562.

12. Kato T, Levi DM, DeFaria W, et al. Liver transplantation with renoportalanastomosis after distal splenorenal shunt. Arch Surg. 2000;135(12):1401–1404.

13. Bhangui P, Lim C, Salloum C, et al. Caval inflow to the graft for livertransplantation in patients with diffuse portal vein thrombosis: a 12-yearexperience. Ann Surg. 2011;254(6):1008–1016.

14. Azoulay D, Adam R, Castaing D, et al. Liver transplantation withcavoportal or renoportal anastomosis: a solution in cases of diffuseportal thrombosis. Gastroenterol Clin Biol. 2002;26(4):325–330.

15. Hajjaj Al A, Bonatti H, Krishna M, et al. Percutaneous transfemoralembolization of a spontaneous splenorenal shunt presenting withischemic graft dysfunction 18 months post-transplant. Transpl Int.2008;21(8):816–819.

16. Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survivalin patients with end-stage liver disease.Hepatology. 2001;33(2):464–470.

17. Halldorson JB, Bakthavatsalam R, Fix O, et al. D-MELD, a simplepredictor of post liver transplant mortality for optimization of donor/recipient matching. Am J Transplant. 2009;9(2):318–326.



18. Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindoclassification of surgical complications: five-year experience. Ann Surg.2009;250(2):187–196.

19. Yerdel MA, Gunson B, Mirza D, et al. Portal vein thrombosis in adultsundergoing liver transplantation: risk factors, screening, management,and outcome. Transplantation. 2000;69(9):1873–1881.

20. BismuthH, Chiche L, Adam R, et al. Liver resection versus transplantationfor hepatocellular carcinoma in cirrhotic patients. Ann Surg. 1993;218(2):145–151.

21. Azoulay D, Bhangui P, Andreani P, et al. Short- and long-term donormorbidity in right lobe living donor liver transplantation: 91 consecutivecases in a European Center. Am J Transplant. 2011;11(1):101–110.

22. Azoulay D, Castaing D, Adam R, et al. Split-liver transplantation for twoadult recipients: feasibility and long-term outcomes. Ann Surg. 2001;233(4):565–574.

23. Azoulay D, Samuel D, Castaing D, et al. Domino liver transplants formetabolic disorders: experience with familial amyloidotic polyneuropathy.J Am Coll Surg. 1999;189(6):584–593.

24. Rodríguez-Castro KI, Porte RJ, Nadal E, et al. Management ofnonneoplastic portal vein thrombosis in the setting of liver transplantation.Transplantation. 2012;94(11):1145–1153.

25. Francoz C, Valla D, Durand F. Portal vein thrombosis, cirrhosis, and livertransplantation. J Hepatol. 2012;57(1):203–212.

26. Englesbe MJ, Schaubel DE, Cai S, et al. Portal vein thrombosis and livertransplant survival benefit. Liver Transpl. 2010;16(8):999–1005.

27. Englesbe MJ, Kubus J, Muhammad W, et al. Portal vein thrombosis andsurvival in patients with cirrhosis. Liver Transpl. 2010;16(1):83–90.

28. Doenecke A, Tsui T-Y, Zuelke C, et al. Pre-existent portal vein thrombosisin liver transplantation: influence of pre-operative disease severity. ClinTransplant. 2010;24(1):48–55.

29. http://www.agence-biomedecine.fr/annexes/bilan2011/donnees/organes/05-foie/synthese.htm.


30. Litvin S, Atar E, Knizhnik M, et al. Stent graft closure of a high flowsplenorenal shunt after liver transplantation. Diagn Interv Radiol. 2010;16(4):312–314.

31. Esquivel CO, Klintmalm G, Iwatsuki S, et al. Liver transplantation inpatients with patent splenorenal shunts. Surgery. 1987;101(4):430–432.

32. Samimi F, Irish WD, Eghtesad B, et al. Role of splenectomy in human livertransplantation under modern-day immunosuppression. Dig Dis Sci.1998;43(9):1931–1937.

33. Lee S-G, Moon D-B, Ahn C-S, et al. Ligation of left renal vein for largespontaneous splenorenal shunt to prevent portal flow steal in adult livingdonor liver transplantation. Transpl Int. 2007;20(1):45–50.

34. Levesque E, Duclos J, Ciacio O, et al. Influence of larger graft weight torecipient weight on the post-liver transplantation course. Clin Transplant.2013;27(2):239–247.

35. de Graaf EL, Kench J, Dilworth P, et al. Grade of deceased donor livermacrovesicular steatosis impacts graft and recipient outcomes morethan the Donor Risk Index. J Gastroenterol Hepatol. 2012;27(3):540–546.

36. Mccormack L, Petrowsky H, Jochum W, et al. Use of severely steatoticgrafts in liver transplantation: a matched case-control study. Ann Surg.2007;246(6):940–946.

37. Marubashi S, Dono K, Nagano H, et al. Living-donor liver transplantationwith renoportal anastomosis for patients with large spontaneoussplenorenal shunts. Transplantation. 2005;80(12):1671–1675.

38. Baptista-Silva JC, VeríssimoMJ, Castro MJ, et al. Anatomical study of therenal veins observed during 342 living-donor nephrectomies. Sao PauloMed J. 1997;115(3):1456–1459.

39. Satyapal KS, Kalideen JM, Haffejee AA, et al. Left renal vein variations.Surg Radiol Anat. 1999;21(1):77–81.

40. Genzini T, Trevizol AP, Yamashita ET, et al. Left renal vein ligation duringliver transplantation in a recipient with a single kidney. Liver Transpl.2013;19(5):563–564.


http://www.agence-biomedecine.fr/annexes/bilan2011/donnees/organes/05-foie/synthese.htm

http://www.agence-biomedecine.fr/annexes/bilan2011/donnees/organes/05-foie/synthese.htm

Spontaneous Splenorenal Shunt in Liver Transplantation

Documents

Transcript of Spontaneous Splenorenal Shunt in Liver Transplantation