Original Article  /  Transplantation

Hepatobiliary & Pancreatic Diseases International

Different cava reconstruction techniques in liver transplantation: piggyback versus cava resection Volker Schmitz, Wenzel Schoening, Ines Jelkmann, Brigitta Globke, Andreas Pascher, Marcus Bahra, Peter Neuhaus and Gero Puhl Berlin, Germany

BACKGROUND:  Originally, cava reconstruction (CR) in liver transplantation meant complete resection and reinsertion of the donor cava. Alternatively, preservation of the recipients inferior vena cava (IVC) with side-to-side anastomosis (known as "piggyback") can be performed. Here, partial clamping maintains blood flow of the IVC, which may improve cardiovascular stability, reduce blood loss and stabilize kidney function. The aim of this study was to compare both techniques with particular focus on kidney function.

vs 1.35±0.96 (piggyback) and 1.45±1.03 mg/dL (CR); P=0.102]. Accordingly, the proportion of patients displaying RIFLE stages ≥2 was the highest in CR/CR-B (26%/19%) when compared to piggyback (18%).

CONCLUSIONS:  Piggyback revealed a shorter warm ischemic time, a reduced blood loss, and a decreased risk of acute kidney failure. Thus, piggyback is a useful technique, which should be applied in standard procedures. When piggyback is unfeasible, cava replacement, which displayed a lower incidence of vascular METHODS:  A series of 414 patients who had had adult liver and biliary complications in our study, remains as a safe transplantations (2006-2009) were included. Among them, alternative. 176 (42.5%) patients had piggyback and 238 had classical (Hepatobiliary Pancreat Dis Int 2014;13:242-249) CR operation, 112 (27.1%) of the patients underwent CR accompanied with veno-venous bypass (CR-B) and 126 (30.4%) KEY WORDS:  liver transplantation; without a bypass. The choice of either technique was based on anastomosis; the surgeons' individual preference. Kidney function [serum surgical procedure creatinine, calculated glomerular filtration rate (GFR), RIFLE stages] was assessed over 14 days. RESULTS:  Lab-MELD scores were significantly higher in CR-B Introduction (22.5±11.0) than in CR (17.3±9.0) and piggyback (18.8±10.0) he technique first described for orthotopic liver (P=0.008). Unexpectedly, the incidences of arterial stenoses transplantation consisted of a complete resection (P=0.045) and biliary leaks (P=0.042) were significantly of the recipients inferior vena cava (IVC) and increased in piggyback. Preoperative serum creatinine levels interposition of the donor intrahepatic part of the were the highest in CR-B [1.45±1.17 vs 1.25±0.85 (piggyback) and 1.13±0.60 mg/dL (CR); P=0.033]. Although a worsening of vena cava with two end-to-end anastomoses including postoperative kidney function was observed among all groups, the use of a veno-venous bypass for hemodynamic this was most pronounced in CR-B [creatinine day 14: 1.67±1.40 stabilization.[1] This approach is still performed as

T

Author Affiliations: Department of General, Visceral and Transplantation Surgery, Charité, Campus Virchow, Berlin, Germany (Schmitz V, Schoening W, Globke B, Pascher A, Bahra M, Neuhaus P and Puhl G); Department of General Surgery, Luebeck, Germany (Jelkmann I) Corresponding Author: Volker Schmitz, MD, Department of General, Visceral and Transplantation Surgery, Charité, Campus Virchow, Augustenburger Platz 1, Berlin 13353, Germany (Tel: 49-30-450-652194; Fax: 49-30-450552900; Email: [email protected]) © 2014, Hepatobiliary Pancreat Dis Int. All rights reserved. doi: 10.1016/S1499-3872(14)60250-2

a standard in many centers. A modification for cava reconstruction, called the piggyback technique, was later introduced by Tzakis,[2] which preserved the full length of the recipients cava with subsequent anastomosis of the suprahepatic donor hepatic veins to the ostia of the recipient left and middle hepatic veins. Several modifications eventually lead to the side-to-side cavocavostomy, which is referred to as piggyback today.[3] In this method, the distal and caudal orifices of the donor cava segment are closed and after partial clamping of the recipient inferior cava segment, which still

242 • Hepatobiliary Pancreat Dis Int，Vol 13，No 3 • June 15，2014 • www.hbpdint.com

Piggyback versus cava replacement in liver transplantation

preserves the blood flow from the lower body part to the test the hemodynamic stability. For the use of a venoheart, a side-to-side anatomosis can be made. As first venous bypass, preservation and operating time (all published by Tzakis,[2] all patients were simultaneously in minutes) during transplantation were registered. stabilized by using a veno-venous bypass. Although this Hemodynamic instability was defined when mean became basically unnecessary with the latter described arterial blood pressure decreased by approximately modification, some surgeons still perform a temporary more than 30% during a trial of clamping of the portal vein and IVC. During this 3 to 5 minutes trial, fluids portocaval shunt to minimize portal congestion.[4] (colloids, crystalloids) were administered to restore Advantages of the piggyback procedure are considered to be the shorter operation time (saving preclamping central venous pressure. The number of one anastomosis), a shorter anhepatic phase/warm perioperatively administered units of transfusions or ischemia, a reduction of blood loss and thus better fresh frozen plasmas (FFP) were included. Postoperative hemodynamic stabilization with a lower incidence measurements contained the number of units of packed of kidney dysfunction. All these aspects have been red blood cells (RBC) and FFP administered and the analyzed previously but only a few publications with highest postoperative serum creatinine levels on the small and in general uneven distribution of patients and day immediately after operation and day 7 and 14 posttransplant. The creatinine levels at these time points controversial results are available.[5-12] were compared with the preoperative creatinine levels. In our center, the standard for cava reconstruction In the piggyback operation, trial clamping was was classic replacement for many years. This concept was changed towards piggyback in the beginning of 2006 tolerated by the patients at all times, and thus in until the middle of 2007 when cava replacement was these patients, after partial clamping of the IVC, a reinstalled as our standard because of the impression of longitudinal cavotomy on the donor and recipient vena cava followed by a side-to-side running suture increasing complications in the piggyback era. The present study was to compare the outcome and anastomosis was achieved. Arterial, portal venous and were performed thereafter as complications of the two techniques (piggyback versus biliary reconstructions [3] described elsewhere. cava replacement) based on our centers' experience, we In the remaining 238 patients (57.5%), a conventional specifically focused on the influence of each technique cava reconstruction[13] was performed with full replacement on kidney function. of the recipients' cava and therefore two anastomoses. Depending on the surgeons' decision and the extent of hemodynamic instability after intra-operative clamping Methods of the vena cava as described before, a veno-venous Surgical technique of liver transplantation bypass (CR-B) was performed in 112 patients (27.1%), In our retrospective analysis, 414 adult patients that and a cava replacement was applied without venohad received a full-size liver transplantation at our venous bypass in 126 (30.4%). The patients usually received a T-tube for biliary institution between January 2006 and September 2009 stenting and decompression. Those who had not were included. Patients with combined liver-kidney transplantation, transplantation in children, living- received T-tubing were enrolled in a randomized trial to determine its necessity, which had been incidentally donor and split liver transplantation were excluded. Of these, 176 (42.5%) patients underwent cava conducted within the observation period. reconstruction using the modified piggyback technique (side to side cavo-cavostomy) as described by Belghiti[3] Outcome parameters without complete occlusion of the IVC and thus, no Patients' demographics included age, gender, primary veno-venous bypass. For that, the vena cava of the diagnosis for liver transplantation and (laboratory) model hepatic graft was sutured one centimeter above the for end-stage liver disease (MELD) score. confluence of the hepatic veins on the back table. In the Perioperative morbidity was assessed by analyzing recipient, a complete dissection of the retrohepatic vena the incidence of vascular (stenosis, thrombosis, bleeding), cava was performed, transecting and ligating all the biliary (leaks, stenosis) and infectious (cholangitis, sepsis) short hepatic veins draining the posterior part of the complications. Special emphasis was given to the degree right liver lobe. Eventually, the three hepatic veins were of pre- and post-operative kidney function. This was transected and oversewn. Initially, a vascular clamp was obtained by serum creatinine levels on days 0, 1, 7 and applied laterally on the anterior part of the IVC (with 14, and the requirement of hemodialysis. The incidence preserved cava blood flow) for a couple of minutes to of hepatorenal syndrome (HRS), which was defined Hepatobiliary Pancreat Dis Int，Vol 13，No 3 • June 15，2014 • www.hbpdint.com • 243

Hepatobiliary & Pancreatic Diseases International

by criteria described elsewhere, was also registered.[14] Based on serum creatinine values at different time points, glomerular filtration rate (GFR) was calculated using the 4-parameter-MDRD-formula.[15] Dynamic outcome of renal dysfunction was further categorized by calculating the changes of serum creatinine levels on days 1, 7 and 14 from baseline (day 0). According to the extent of change, these values were divided into 7 different groups: 1: decrease >1.0 mg/dL; 2: decrease 0.51-1.0 mg/dL; 3: decrease ≤0.5 mg/dL; 4: no change; 5: increase ≤0.5 mg/dL; 6: increase 0.51-1.0 mg/dL; and 7: increase >1.0 mg/dL. Furthermore, to better display the proportion of different degrees of acute kidney deterioration at each time point, the patients were grouped according to the RIFLE criteria for acute kidney injury [stage 0=no change, stage 1 (Risk)=increase in creatinine × 1.5 or GFR decrease >25%, stage 2 (Injury)=increase in creatinine × 2 or GFR decrease >50%, stage 3 (Failure)= increase in creatinine ×3 or creatinine >4 mg/dL or GFR decrease >75%, stages 4 (Loss)/5 (End-stage renal failure)], defined as persistent kidney failure >4 weeks/>3 months, which were summarized within stage 3, since observation period did not exceed 14 days.[15] Postoperative liver graft function was characterized by the levels of transaminases [alanine aminotransferase (ALT)/aspartate aminotransferase (AST)] and total bilirubin on corresponding time-points (days 0, 1, 7, 14). All grafts were procured from heart-beating, brain dead, and ABO compatible donors with standard procurement techniques. All grafts were flushed and preserved in HTK solution. Specific operative characteristics included anastomosis time (warm ischemic time), preceding cold ischemic time of the graft and number of intraoperative blood (packed cells) and plasma (FFP) units.

Table 1. Patient demographics and renal function based on specific type of cava reconstruction (n, %) Demographics

Piggyback CR no bypass CR bypass P value (n=176) (n=126) (n=112) 111 (63.1) 98 (77.8)* 54±10 56±9 51 (40.5) 24 (19.0) 6 (4.8) 2 (1.6) 13 (10.3) 8 (6.3) 3 (2.4) 19 (15.1) 52 (41.3) 17.3±9.0*

65 (58.0) 0.0212 56±10 0.1071 0.0012 35 (31.3) 0.2622 14 (12.5) 0.1792 1 (0.9) 0.1592 2 (1.8) 0.0082 16 (14.3) 0.246 19 (17.0) 0.0062 8 (7.1) 0.0342 22 (19.6) 0.9762 31 (27.7) 0.0152 22.5±11.0* 0.0081

Hepatorenal syndrome*

42 (23.9) 24 (19.0)

33 (29.5) 0.0062

Preop HD* Postop HD*

13 (7.4) 9 (7.1) 55 (31.3) 41 (32.5)

17 (15.2) 0.0352 50 (44.6) 0.5302

Gender (male) Age (yr) Diagnosis for OLTx Alcohol HCV HBV Acute liver failure Kryptogenic Graft failure/ITBL PSC/PBC Others Hepatocellular carcinoma Lab-MELD Renal function

58 (33.0) 39 (22.2) 9 (5.1) 14 (8.0) 13 (7.4) 12 (6.8) 3 (1.7) 28 (15.9) 47 (26.7) 18.8±10.0

Age and lab-MELD as mean±standard deviation. Remaining figures as absolute numbers and percentage. 1: One-way ANOVA. *: significant in pairwise multiple comparison using the Holm-Sidak post-hoc test. 2: Chisquare. CR: cava replacement; OLTx: orthotopic liver transplantation; MELD: model of end-stage liver disease; HD: hemodialysis; HCV: hepatitis C virus; HBV: hepatitis B virus; ITBL: ischemic type biliary lesions; PSC: primary sclerosing cholangitis; PBC: primary biliary cirrhosis.

There were a higher proportion of male patients in the CR group (P=0.021) and a significant overall difference for the primary diagnoses for liver transplantation (P=0.001), which was more often acute liver failure (8.0%) in piggyback and more often alcoholic cirrhosis in CR (40.5%). There was also a higher proportion of retransplantations (17.0%) in the CR-B group. Also, hepatocellular carcinoma (41.3%) was found (P=0.015) Statistical analysis more often in CR (vs 27.7% in CR-B and 26.7% in All numerical data were presented as mean±standard piggyback). Subsequently, with this uneven distribution deviation. An analysis of variance (ANOVA) was used of carcinoma, a higher percentage of "standard-exceptto compare quantitative differences and the Chi-square MELD" patients could be seen in CR as well, thus test to compare qualitative ones. If applicable, additional leading to significant differences in lab-MELD scores group comparison was performed using a Holm-Sidak [17.3 in CR vs 18.8 in piggyback and 22.5 in CR-B post-hoc analysis (Software IBM SPSS Statistics 20 (P=0.008)]. by IBM® Germany). A P value 0.5 mg/dL in both cava replacement groups (day 1: CR 25%, CR-B 26%; day 7: CR 29%, CR-B 28%; day 14: CR 26%, CR-B 31%) compared with the piggyback group (day 1: 10%, day 7: 20%, day 14: 18%). On the other hand, on day 7 post-transplant, there was also the highest proportion of creatinine decreases >0.5 mg/dL in the cava replacement plus the bypass group (CR-B: 31% vs 6% in CR and 11% in piggyback). Another classification according to the RIFLE criteria (Fig.) revealed that before transplantation there were not only most patients with high degrees (stage ≥2) of kidney failure in the CR-B; this was also the group with the highest proportion of normal kidney function (stage 0). However, within the observation period, although all groups experienced a worsening of kidney function, the most severe changes (stage ≥3) were seen in both the CR and CR-B groups at all time-points. This demonstrated that the majority of patients after piggyback reconstruction obviously experienced less pronounced changes of serum creatinine which stayed

Hepatobiliary Pancreat Dis Int，Vol 13，No 3 • June 15，2014 • www.hbpdint.com • 245

Hepatobiliary & Pancreatic Diseases International

within the range of ±0.5 mg/dL (day 1: 87%, day 7: 70%, day 14: 68%) compared with CR (day 1-changes±0.5 mg/dL: 79%, day 7: 65%, day 14: 65%) or CR-B (day 1- changes±0.5 mg/dL: 67%, day 7: 50%, day 14: 54%).

In almost the same manner, stages reflected that most piggyback patients only developed mild (stage 1=risk) or no (stage 0) acute renal failure (e.g. day 1: 48% compared with CR: 32%, CR-B: 38%).

Discussion

Fig. Classification of renal dysfunction according to the RIFLE criteria (stage 0=no change, 1=increase in creatinine ×1.5 or GFR decrease >25%, 2=increase in creatinine ×2 or GFR decrease >50%, 3=increase in creatinine ×3 or creatinine >4 mg/dL or GFR decrease >75%, 4/5 persistent kidney failure >4 weeks/3 months; since observation period did not exceed 14 days, stages 3-5 were summarized as ≥3).

The initially described surgical procedure of liver transplantation, which naturally included a complete resection of the recipient vena cava and its replacement with the graft, has been widely replaced in many centers by the so-called piggyback technique.[12, 16, 17] Presumably, the main advantages of this newer method are the preservation of the venous backflow and avoidance of a veno-venous bypass resulting in a shorter operation time due to the avoidance of one anastomosis. Also, resection of the vena cava as in cava replacement might be associated with an increased risk of retroperitoneal hemorrhage especially in retransplantation. Thus, in theory, the preservation of the recipient cava should better stabilize the patient hemodynamically and this could also contribute to the decrease of the incidence of acute kidney failure. However, results of different studies on the impact of cava reconstruction techniques on kidney function are controversial with some authors describing no difference following either cava

Table 4. Summary of current literatures evaluating cava reconstruction in liver transplantation Definition of renal failure

Renal failure

Bypass

Benefit in renal function for PB

500 Piggyback (Belghiti) only

Temporary (150 µmol/L at day 1

13%

Always temporary portocaval shunt

/

Authors

n

Mehrabi et al[9] Cherqui et al[6]

Technique

Nikeghbalian et al[11] 253 Piggyback: 94% CR: 6%

Creatinine >1.5 mg/dL, increase Piggyback: 13% by 50% or renal replacement CR: 47% therapy during first week

Piggyback: 0 CR: 0

Yes

Jovine et al[8]

39 Piggyback: 51% CR: 49%

Not defined

Piggyback: 0 CR: 19%

Yes (P1.5 mg/dL; increase Piggyback: 18% CR: 50% by 50%, renal replacement therapy

Piggyback: 0 CR: 19.2%

Yes (P