Telephone: 00(39)-0498212210; FAX: 00(39)-049656145
Liver transplantation using suboptimal grafts: Impact of donor harvesting technique
Article first published online: 27 SEP 2007
Copyright © 2007 American Association for the Study of Liver Diseases
Volume 13, Issue 10, pages 1444–1450, October 2007
How to Cite
D'Amico, F., Vitale, A., Gringeri, E., Valmasoni, M., Carraro, A., Brolese, A., Zanus, G., Boccagni, P., D'Amico, D. F. and Cillo, U. (2007), Liver transplantation using suboptimal grafts: Impact of donor harvesting technique. Liver Transpl, 13: 1444–1450. doi: 10.1002/lt.21268
- Issue published online: 27 SEP 2007
- Article first published online: 27 SEP 2007
- Manuscript Accepted: 30 MAY 2007
- Manuscript Received: 19 DEC 2006
In recent years, an increasing number of suboptimal grafts has been used to reduce the gap between the supply and demand of organs for liver transplantation (LT). In this randomized prospective study, we tested the impact of donor harvesting technique on the posttransplantation outcome of suboptimal donor livers. A modified double perfusion (MDP) technique (aortic and portal cooling with tourniquet clamping of splenomesenteric vein inflow) was compared with the single aortic perfusion (SAP) technique. Between February and November 2005, 35 suboptimal grafts were randomly assigned to either technique (18 MDP livers and 17 SAP livers). Donor and recipient variables were comparable in the 2 study groups. The SAP group had significantly higher blood transaminases and bilirubin levels after LT. The prevalence of graft primary dysfunction (PDF) was also significantly higher (P = 0.01) in the SAP group (35%) than in the MDP group (5%). In 5 cases, all in the SAP group (P = 0.02), early re-LT (<30 days) was needed. The 6-month patient and graft survival rates were significantly higher in the MDP (100% in both cases) than in the SAP group (68% and 58%, respectively). The study was stopped in November 2005, when the interim analysis revealed such markedly significant differences between the two groups. In conclusion, the present study showed a very low prevalence of PDF, death, and re-LT after transplantation with suboptimal liver when a MDP technique was used to harvest the donor graft. Liver Transpl 13:1444–1450, 2007. © 2007 AASLD.
Liver transplantation (LT) is the treatment of choice for end-stage liver disease. One of the main limits of this therapeutic option, however, remains the disparity between organ availability and demand.1
In recent years, an increasing use has been made of suboptimal livers to help reduce this organ shortage, but the use of such marginal livers has had deleterious effects on initial graft function with a prevalence of primary dysfunction (PDF) ranging from 13 to 36%.2
Given such a high prevalence of liver dysfunction, it is of paramount importance to gain more insight into the factors potentially involved in the recovery of suboptimal grafts after LT.
In this study, we hypothesized a role for donor harvesting technique in determining the post-LT graft function and recovery profile when suboptimal livers are used.
The liver graft harvesting technique initially described by Starzl et al.3 consisted of a double perfusion of the liver through the aorta and portal vein. Single aortic perfusion (SAP) was introduced in recent years, proving a safe procedure with some potential advantages for multiple organ harvesting.4
To our knowledge, however, the issue of harvesting technique has never been thoroughly evaluated in the particular setting of suboptimal livers.
Since 1991, we have adopted the classic double perfusion procedure at our Institution and, as suggested by Starzl et al.,3 we consider it essential to reduce the splenomesenteric vein inflow in the cold retrieval phase, based on speculative assumption that this can reduce Kupffer cell activation by toxins of intestinal origin.5 With this aim, in 1993 we introduced the expedient of a tourniquet surrounding the portal vein to be tightened immediately after cross clamping on the portal cannula inserted via the inferior mesenteric vein.
At the same time, the SAP technique was adopted at our Institution in pediatric donors, for combined liver and pancreas harvesting, and when the split in situ technique was used.
The aim of the present study was to see whether a presumably better cold perfusion of the liver and the avoidance of vascular inflow—potentially afforded by our modified double perfusion (MDP) technique—are associated with a better graft survival and recovery profile than the SAP technique in LT using suboptimal grafts.
PATIENTS AND METHODS
From February 2005, all potential grafts considered suitable for LT at donor coordinator call were randomly assigned to either of the 2 procurement techniques. Cases of LT for acute liver disease, pediatric LT, multiple-organ transplantation, retransplantation, and LT with a partial liver (split, reduced, or living-related grafts) were excluded.
Suboptimal grafts were arbitrarily classified as follows:6–11 1 major suboptimal criterion (donor >60 yr old, biopsy-proven liver steatosis >20%, total ischemia time >10 hours) or at least 2 minor criteria (peak serum sodium >155 mEq/L, intensive care unit stay >7 days, dopamine doses >10 μg/kg/minute or any other vasoactive amines, abnormal liver tests, and major hemodynamic instability). The study protocol complied with the ethical guidelines of the 1975 Helsinki Declaration, as reflected in its a priori approval by the institutional review board of our institution. Follow-up visits were established on postoperative days 1, 2, 3, 5, and 7 after LT and then at 1, 3, 6, 9, and 12 months. The visits included a complete clinical examination and an evaluation of liver function parameters (total bilirubin, serum aspartate aminotransferase, alanine aminotransferase, prothrombin time, gamma-glutamyl transferase). An interim analysis was scheduled for November 2005.
With the exception of the above-mentioned cases, all donor livers were randomly assigned to the SAP or MDP techniques according to a computer-generated list. Randomization was done immediately after a phone call was received from our organ-sharing organization to advise us that a potential cadaveric liver donor was available. This meant that all livers “accepted” at the time of the phone call, but subsequently not transplanted because they were judged to be unsuitable for grafting at the time of harvesting, or at back-table biopsy, were included in the randomization list. From February 2005 to November 2005, 58 potential grafts were randomized (SAP group: 29; MDP group: 29). In 10 cases (5 randomized in SAP and 5 in MDP groups), the livers were subsequently considered unsuitable for transplantation due to high-grade steatosis in 5 cases (steatosis >50% after biopsy), liver cirrhosis in 2, and cancer in 3. The remaining 48 livers were transplanted into 48 recipients: 35 of them were considered suboptimal livers and form the object of the present study (SAP group: 17; MDP group: 18).
After randomization, donor livers were harvested using the randomly-assigned technique by 2 experienced surgeons.
Briefly, in the SAP technique, a single cannula was inserted in the aorta and the hepatic hilum was only marginally involved, whereas in the MDP technique, a cannula for perfusion was inserted in both the aorta and the inferior mesenteric vein, as in the original Starzl procedure. In addition, after confirming the presence of the cannula by palpation, the portal vein was surrounded with a tourniquet, which was clamped immediately after cross-clamping.
No heart and lung, pancreas, and small bowel procurements were associated with liver and kidney harvesting in our series and all donors were hemodynamically stable throughout the operation.
Donor livers were gravity-perfused in situ via the aorta and portal vein with Celsior solution at 4°C. About 60% of the volume of solution (30 mL/kg via the portal vein and 60 mL/kg via the aorta in the MDP procedure, 90 mL/kg via the aorta using the SAP technique) was infused rapidly (in 10–15 minutes) after aortic cross-clamping. Perfusion was then slowed down for the remaining 40% of solution until harvesting was completed (20–40 minutes). After hepatectomy, donor livers were further perfused at the back-table with Celsior (700 mL via the portal vein and 300 mL via the hepatic artery) and then stored in conventional bags containing the same solution at 4°C until transplantation.
LT was always done preserving the retrohepatic vena cava (piggy-back technique) and without using the biopump. Transplanted patients were given dual- or triple-drug immunosuppressive therapy, which included cyclosporine or tacrolimus combined with corticosteroids, with or without mycophenolate mofetil. A posttransplantation liver biopsy was only taken when clinically indicated.
We defined PDF, and 6-month post-LT graft and patient survival rates as the main end-points of the study.
PDF was defined according to Ploeg et al.6 as the sum of primary nonfunction (PNF) and initial poor function. PNF is a non-life-sustaining function of the graft leading to death or retransplantation within 7 days. Initial poor function was defined as aspartate aminotransferase >2,000 IU/L and prothrombin time >16 seconds (s) on postoperative days 2 to 7.
All demographic and baseline variables were described using statistical characteristics. Categorical data were described by frequency and percentage. Continuous data were described by mean and standard deviation. Missing data were excluded from the final analysis.
In the comparison of different subgroups, continuous variables were compared using logistic regression. Categorical variables were compared using chi-squared or Fisher's exact tests, as appropriate.
Follow-up and survival periods are expressed as median (range). Survival curves were calculated according to the Kaplan-Meier method and compared using log rank analysis.
For continuous variables, the cutoff was the median value. Statistical significance was set at P < 0.05. The calculations were done with the JMP package (1989–2003; SAS Institute, Cary, NC).
Baseline Characteristics of the Study Groups
The main liver donor variables were equally distributed in the 2 groups, as shown in Table 1. All our suboptimal livers had at least 1 major suboptimal criterion (and 2 in almost 50% of cases) in association with at least 1 minor suboptimal criterion. Table 2 shows that recipient data were also similar in the 2 study groups. There were more females in the SAP group, while the number of patients with hepatocellular carcinoma was higher in the MDP group, but these differences were not significant.
|Variables||MDP group: 18 patients (%)||SAP group: 17 patients (%)|
|Steatosis >20%||8 (44)||6 (35)|
|Body mass index||26.9 ± 2.5||25.7 ± 5.0|
|Age (yr)||63.1 ± 11.9||56.3 ± 18.7|
|CIT (minutes)||533 ± 77.0||583 ± 75.7|
|WIT (minutes)||32.6 ± 4.6||33.6 ± 7.4|
|Ischemia time >10 hours||9 (50)||5 (29)|
|Sodium (mEq/L)||148.5 ± 7.5||149.1 ± 6.4|
|ICU (days)||5.2 ± 3.1||5.3 ± 7.1|
|Dopamine doses (μg/kg/minute)||6.0 ± 4.4||6.5 ± 4.5|
|Relevant hemodynamic instability*||5 (28)||7 (41)|
|Abnormal liver tests§||9 (50)||8 (47)|
|Number of major criteria|
|1||9 (50)||11 (65)|
|2||7 (39)||6 (35)|
|Cumulative number of criteria||2.9 ± 0.8||2.3 ± 1.0|
|Vascular anomalies||4 (22)||3 (18)|
|Traumatic cause of death||7 (39)||6 (35)|
|Volume of solution (mL)||7666.4 ± 594.6||7058.8 ± 129.7|
|Variables||MDP group (18 patients)||SAP group (17 patients)|
|Age (yr)||54.4 ± 5.9||54.3 ± 6.3|
|MELD score||15.4 ± 5.0||16.2 ± 4.3|
|Blood transfusion (mL)||1675 ± 1307||1256 ± 1450|
The SAP group had higher transaminases and bilirubin levels on postoperative day 2, while the MDP group had higher gamma-glutamyl transferase values on postoperative day 7; all of these differences were statistically significant.
In the early post-LT period, we recorded 8 cases of PDF (23%) including 3 initial poor function and 5 PNF. The 5 PNF led all patients to re-LT, and only 2 survived 1 month after reoperation. Table 3 shows a significantly higher prevalence of PNF, and early re-LT in the SAP group, with consequently lower 6-month graft and patient survival figures (Table 3; Figs. 1 and 2). Table 4 shows the specific causes of graft loss and patient death. In 4 of 5 deaths, the initial event was primary liver dysfunction. Similarly, all 5 retransplantations were due to PNF.
|Variables||MDP group (18 patients)||SAP group (17 patients)|
|AST-POD 2 (U/L)*||763 ± 939||2125 ± 2747|
|ALT-POD 2 (U/L)*||614 ± 519||1580 ± 1687|
|GGT-POD 2 (U/L)||102.3 ± 65.1||71.7 ± 59.1|
|Bilirubin-POD 2 (μmol/L)§||46.6 ± 30.8||83.0 ± 65.1|
|PT-POD 2 (%)||43.7 ± 14.7||41.4 ± 19.8|
|AST-POD 7 (U/L)||41.7 ± 20.2||48.4 ± 20.4|
|ALT-POD 7 (U/L)||82.8 ± 56.6||116.1 ± 111.4|
|GGT-POD 7 (U/L)*||102.3 ± 65.1||71.7 ± 59.1|
|Bilirubin-POD 7 (μmol/L)||87.5 ± 5.1||87.9 ± 8.1|
|PT-POD 7 (%)||60.1 ± 12.4||57.3 ± 9.2|
|PDF (IPF+PNF) (%)*||1 (5)||7 (41)|
|PNF (%)*||0||5 (29)|
|Early re-LT (<30 days) (%)*||0||5 (29)|
|Patient||Group||Initial graft function||Cause of graft loss||Cause of death|
Outcome of Patients Transplanted With Optimal Livers
During the study period, 13 patients receiving optimal livers were also randomized to the 2 harvesting techniques: 7 in the SAP group, and 6 in the MDP group. There were 2 cases of PDF (15%), 1 in the SAP and 1 in the MDP group. There were no cases of PNF and re-LT among the patients transplanted with optimal livers. 6-month patient and graft survival rates were 100% (Fig. 1).
End of the Study
The interim analysis performed in November 2005 showed such marked differences between the study groups that the study was stopped and the MDP technique was adopted for all suboptimal livers thereafter.
Though less frequent than in the case of other organs, such as the kidney, post-LT PDF unrelated to any direct surgical, immunological, or other complications still occurs in a significant number of recipients.6–11 Grafts with a poor initial function have a higher than normal graft failure rate in the first 3 months after transplantation.6 The etiology of PDF has yet to be clarified and several factors are thought to be involved—primarily poor organ quality at the time of the transplantation, which may depend on the condition of the donor, the cold ischemia time and/or the organ preservation method.11
We previously considered this complex field and investigated the prognostic role of organ preservation method in LT,12 finding a very low prevalence of graft PDF after LT irrespective of the preservation solution used.
In the present prospective randomized study, we assumed that the technique used to harvest the liver has an impact on the outcome of LT when suboptimal livers are used. In particular, we compared our MDP procedure with the SAP technique, establishing as primary endpoints the prevalence of PDF in our 2 study groups.
In the last decade, progressive surgical, anesthesiological, and pharmacological advances have led to a significant reduction in the mortality and morbidity after LT. On the other hand, the rising demand for LT has lengthened the waiting lists, giving rise to higher pre-LT mortality figures.13 The shortage of organ donors has prompted many transplant centers to revise their donor selection criteria and consider using the so-called marginal donors.13, 14 Though there is a large body of evidence to show that older donors and fatty livers significantly increase the incidence of liver preservation injury,15 a first important issue concerns the criteria used to define suboptimal livers. A widely-accepted definition of marginal liver donors with homogeneous and well-defined cutoffs has yet to be established in the literature,16 but some insight has definitely been gained. First, the main factors that may have a negative fallout on post-LT liver function recovery, the donor's age, macroscopic steatosis, and ischemia time seem to be the most often reported.7 Second, it is generally accepted that a cumulative effect of several negative donor factors determining preservation injury may, in most cases, represent the pathophysiological mechanism behind post-LT dysfunction, which is seen as a multifactorial event.2 On these grounds, we defined suboptimal livers as those with 1 major negative factor and/or at least 2 minor factors. This definition, based on several studies published in recent years,2, 6–11, 16–20 revealed a statistically significant impact on post-LT graft and patient survival rates in a retrospective analysis of our last 400 cases (our unpublished data).
As suggested by Briceno et al.16, 17 and Pokorny et al.,2, 11 we considered ischemia time as an important donor variable for defining marginal livers, though it obviously cannot be known at the time of organ procurement. Verran et al.18 and Briceno et al.16, 17 have also suggested using liver steatosis (which can only be determined precisely after histological examination) as a major criterion in defining suboptimal grafts. However, the pathologist's response is generally only available when the graft is brought back to the Center; i.e., after several hours of cold ischemia. Consequently, both of these 2 major criteria cannot be known at the time of organ procurement. That is why, in our study design, randomization included all grafts considered potentially suitable for LT, while the 2 study groups were only identified after considering all data with a potential impact on graft quality, even if it was unavailable at the time of harvesting. This choice of study design aimed to avoid the potential bias deriving from excluding factors capable of significantly influencing graft performance from the analysis, and to recruit a study group of suboptimal grafts as objectively as possible, irrespective of any final considerations. In clinical practice, moreover, a lengthy ischemia time and severe steatosis can be suspected relatively early in the procurement process, the former based on logistic considerations, the latter on the surgeon's rough impressions during harvesting. This is more important in clinical practice than in the design of a randomized prospective study. Taking a pragmatic future perspective, in fact, the surgeon has the capacity to decide whether to use the MDP technique to improve graft outcome when these two factors are strongly suspected.
Taken individually, some of the criteria adopted in our study may also be seen in so-called “standard donors.” In Tector et al.,21 for example, only donor age >60 yr, as a single variable, had a significant prognostic impact on post-LT graft survival. As explained by Briceno et al.,16, 17 Pokorny et al.,2, 11 and Silberhumer et al.,22 however, it is the synergic sum of marginal factors, rather than their single contributions, that best predict severe liver preservation injury. In this light, Feng et al.,23 Ioannou,24 and Burroughs et al.25 have introduced a combination of donor variables in scores for predicting post-LT outcome in their recent works.
In our experience, too (Table 1), all our suboptimal livers had at least 1 major criterion (and 2 in almost 50% of our liver donors) in association with at least 1 minor criterion. It is the cumulative effect of several criteria that particularly characterizes the suboptimal nature of our grafts.
The incidence of PDF is reportedly higher than 30% when livers that are “at-risk ” are used.2, 15–19 This incidence is similar to the one observed for our SAP group. It is worth emphasizing that the prevalence of PNF, and consequently of graft/patient losses in the SAP group, are probably overestimated due to the small sample size. In Moore et al.,13 however, patients receiving a graft more than 60 yr old (as the only major criterion) had a 20% prevalence of retransplantation and a 1-yr patient survival around 73%. Verran et al.18 has also shown that the sum of 2 suboptimal donor factors, such as donor age >50 yr and liver steatosis >30%, carried a 50% risk of graft loss within 3 months of LT. Approximately the same probability of graft loss was estimated by Cameron et al.,26 when 3 suboptimal donor factors (age >55 yr, intensive care unit stay >5 days, and ischemia time >10 hours) were simultaneously involved.
In the light of this recent evidence from the literature confirming our 58% in the SAP group, the 100% graft and patient survival without PNF observed in the MDP clearly becomes more informative because its originality and uniqueness in the literature strongly suggest a protective role of the MDP technique in the event of donor risk factors, particularly when in combination. The main result of this study, however, is not the absolute number of PDF events in the 2 groups, but the marked difference in the rate of graft losses between the 2 groups, which was statistically significant even in the context of relatively low number of cases, inducing us to stop the study.
There is no clear evidence to explain such a disparity in PDF rates in our 2 study groups. The randomized nature of the study means that such a difference cannot be attributed to issues such as candidate selection or recipient surgical complications because these would be evenly distributed in the 2 groups. We can only assume some physiopathological advantages deriving from the use of the MDP technique, supported by findings in experimental models. Kupffer cells have revealed a relevant role in cold ischemia-reperfusion injury5, 27 and gut-derived substances can activate Kupffer cells and induce hepatic necrosis.28, 29 The graft may be exposed to intestinal toxins during LT for intestinal wall congestion (caused by portal vein occlusion), or during the cold phase of the liver harvesting procedure. With this in mind, the splenomesenteric inflow clamping involved in our MDP technique may reduce this exposure to intestinal toxins.
A second issue concerns the frequent occurrence of severe atherosclerosis of large splanchnic vessels in older donor organs. On a purely speculative basis, it may be that this pathological condition in the donor organ coincides with a better perfusion of the liver using the MDP technique than with the SAP procedure.
Exposure to intestinal toxins and severe atherosclerosis are, therefore, 2 potential marginal factors hindered by the MDP technique. From a physiopathological point of view, in optimal livers such factors alone are probably not enough to interfere with the recovery of liver function, whereas they would have a relevant harmful effect in association with other suboptimal variables.
As in other experiences,20 in fact, the patient and graft survival differences between our groups were not seen in our randomized cases using optimal livers (Fig. 1), and this supports the multifactorial nature of PDF. On the other hand, the MDP method is technically more complex because it involves dissecting the hepatic hilum in order to surround the portal vein. Cannulation of the portal system is also a potential hazard in terms of preserving the pancreas and intestines.30 In this context, the SAP technique was introduced as a safe and simple technique enabling combined liver, pancreas, and intestine retrieval. Since multiple organ harvesting is generally done using optimal donors, the good results of LT with optimal livers harvested using the SAP procedure in our study (Fig. 1) supports a selective use of the 2 harvesting methods, depending on the quality of the liver and the needs of multiple organ procurement.
In conclusion, our prospective randomized study showed that harvesting technique may have an important prognostic impact on post-LT outcome when suboptimal livers are used. In this particular setting, our MDP procedure was associated with a significantly better post-LT graft performance than the commonly-used SAP technique.
- 11Organ survival after primary dysfunction of liver grafts in clinical orthotopic liver transplantation. Transplant Int 2000; 13(Suppl): S154–S157., , , , , .
- 17A proposal for scoring marginal liver grafts. Transplant Int 2000; 13(Suppl 1): 249–252., , .