Split and whole liver transplantation outcomes: A comparative cohort study
Abstract
A specific split liver transplantation (SLT) program has been pursued in the North Italian Transplant program (NITp) since November 1997. After 5 yr, 1,449 liver transplants were performed in 7 transplant centers, using 1,304 cadaveric donors. Whole liver transplantation (WLT) and SLT were performed in 1,126 and 323 cases, respectively. SLTs were performed in situ as 147 left lateral segments (LLS), 154 right trisegment liver (RTL) grafts, and 22 modified split livers (MSL), used for couples of adult recipients. After a median posttransplant follow-up of 22 months, SLTs achieved a 3-yr patient and graft survival not significantly different from the entire series of transplants (79.4 and 72.2% vs. 80.6 and 74.9%, respectively). Recipients receiving a WLT or a LLS showed significantly better outcomes than patients receiving RTL and MSL (P < 0.03 for patients and P < 0.04 for graft survival). At the multivariate analysis, donor age of >60 yr, RTL transplant, <50 annual transplants volume, urgent transplantation (United Network for Organ Sharing (UNOS) status I and IIA), ischemia time of >7 hours, and retransplantation were factors independently related to graft failure and to significantly worst patient survival. Right grafts procured from RTL and either split procured as MSL had a similar outcome of marginal whole livers. In conclusion, in 5 yr, the increased number of pediatric transplants due to split liver donation reduced to 3% the in-list children mortality, and a decrease in the adult patient dropout rate from 27.2 to 16.2% was observed. Such results justify a more widespread adoption of SLT protocols, organizational difficulties not being a limit for the application of such technique. Liver Transpl 12:402–410, 2006. © 2006 AASLD.
Over the last decade, constant improvement in the results of liver transplantation (LT) in patients with end-stage liver disease has led to a remarkable increase in indications and in number of transplants performed every year.1 However, the major drawback of LT success has been the worldwide lengthening of the pre-LT waiting time, with particular reference to adult patients, and despite any effort to reduce the waiting list mortality through use of marginal donors and implementation of living donor programs. In the particular case of the North Italian Transplant program (NITp), the number of cadaveric donors has more than doubled in the last decade (from 11.9 to 23.3/million population/yr), with the number of LT per year grown from 163 to 370 (127% increment), a result mainly due to a 5-fold increase in the percentage of donors over 60 yr of age (from 6.9 to 32.7%). However, despite extensive education campaigns, at the end of 2004 only 41.8% of the LT list was fulfilled, and about 400 patients in an area of nearly 18 million people are still waiting for a liver, a result that is described in other European procurement agencies.2, 3
Pediatric LT represented the main factor influencing the development of an integrated split liver transplantation (SLT) program throughout the world, with possible advantages on both the pediatric and adult waiting lists.4 Following the early experiences of the 1980s,5, 6 a SLT program has been implemented also in the NITp area7 since 1997 using the conventional splitting procedure, which assigns the left lateral segment (LLS) of the donated liver to pediatric recipients and the right trisegment liver (RTL) to adults. More recently, a modified split liver (MSL) technique based on the division of the donated liver in 2 similarly sized parts has also been implemented8, 9 to allow LT in 2 small adults.
In the present report, the results of such a split liver-oriented attitude developed throughout the transplant centers in northern Italy have been analyzed in detail and compared with the results obtained using conventional whole liver transplantation (WLT) in the same territory centers and time span.
Abbreviations
SLT, split liver transplantation; NITp, North Italian Transplant program; WLT, whole liver transplantation; RTL, right trisegment liver (segments 4 to 8); MSL, modified split liver (segments 1 to 4 and 5 to 8); UNOS, United Network for Organ Sharing; LT, liver transplantation; LLS, left lateral segment (segments 1 and 2).
PATIENTS AND METHODS
Split Liver Protocol Definition
Outcomes of different groups of recipients transplanted with cadaveric graft, whether procured as a whole organ or through splitting techniques, were compared within a period of 5 yr. The study protocol gathered a number of 7 LT centers belonging to the NITp, a reference organization serving about 18 million inhabitants with about 60 donor hospitals.7 Donor eligibility criteria for the split procedure included age below 60 yr, intensive care unit stay of less than 5 days, low inotropic support (dopamine ≤5 μg/kg/minute, dobutamine ≤10 μg/kg/minute, and no epinephrine or norepinephrine), and near-normal liver function tests.7 The anatomic classification of the liver described by Couinaud10 and refined by Bismuth11 served as guideline for the description of the splitting procedure and graft sectioning.
As previously described,12-14 the surgical in situ conventional split procedure foresees the separation of the left lateral liver graft (segments 2 and 3) from the RTL graft (segments 1 and 4 to 8). The left hepatic artery and the celiac axis were generally maintained with the left liver, and the arterial revascularization of the right graft required in some cases an allogenic arterial graft interposition.15 When the modified splitting procedure was applied to comply with 2 adult recipients, the in situ liver separation was obtained along the Cantlie line in order to have 2 grafts similar in size, that is, segments 5 to 8 (RTL) to the right side and segments 1 to 4 (left trisegment) to the left side, in continuity with the celiac axis, hepatic artery, left portal trunk, left hepatic duct, and median and left hepatic veins with the infrahepatic vena cava. The graft-to-recipient body weight ratio was calculated on the basis of donor height and weight. A graft-to-recipient body weight ratio of at least 0.8 was pursued in all cases in order to minimize the risk of small-for-size transplants.16 Besides such volume restrictions, livers derived from split procedures were allocated according to the same clinical and immunological criteria used for the whole livers.
Patient Populations
In the 5-yr period under study (November 1997-October 2002), 1,449 liver transplants were performed in the NITp area using 1,304 cadaveric donors. In fact, 178 donors (13.6%) underwent different kinds of split procedures for liver procurement. Overall, WLT was performed in 1,126 cases, while split livers were used in 323 patients, namely, LLS in 147 cases and conventional right (RTL) in 154 patients, with a further 22 MSL for couples of adult recipients. In at least 13 cases, 1 split organ was exported to other Italian or European areas, while in 20 of the attempted splits (5.5%), 1 of the 2 half livers was discarded for a variety of anatomic and logistic reasons. No donor for both potential liver recipients was lost because of irreversible graft damage during the split procedure.
The 7 LT centers belonging to the NITp allocation area actively participated in the split liver program, although with different experience in performing such procurement (the center volume for split liver harvest ranged from 8 to 142). A single center gathered 77.6% of pediatric recipients, and consequently 85.7% of the LLS grafts, biliary atresia being the main indication for liver replacement in children (66.6%). Cirrhosis (either viral or alcoholic) was the main evidence indicating transplantation among adults, with the percentage of conventional RTL grafts used for different indications ranging from 5.1 to 26.3%. Overall, the mean age of the recipients was 3.6 ± 4.3 and 48.8 ± 10.7 in the pediatric and adult populations, respectively, with a male:female ratio of 119:113 and 851:366 in children and adults, respectively.
Information on 32 different pre- and posttransplant variables in both donors and recipients were collected, including the United Network for Organ Sharing (UNOS) recipient status at the time of transplantation, operation techniques (donor/recipient), graft outcome, and recipient morbidity/mortality. Data were centralized and protected in a database that did not disclose individual or center-specific information.
Cohorts Identification and Study Design
Four different cohorts of patients were recruited during the study period and compared on the basis of graft procurement: WLT and split livers (LLS and RTL for child/adult couples and MSL for adult/adult transplantation). For further analysis, WLT cases were divided into optimal and marginal categories: whether the donor was younger or older than 60 yr and had more or less than 12 hours of cold ischemia time. Outcome of each cohort (graft and patient survival) was assessed in relationship with recipient conditions (as expressed by UNOS status grouping, that is, status I and IIA for urgent patients and status IIB and III for semielective/elective cases), recipient primary diagnosis, annual whole liver volume of each LT center, donor/recipient gender and blood-type match, donor and recipient age and weight, recipient panel reactive antibodies, waiting time, liver cold ischemic time, and combined transplants. Each cohort of patients was collected consecutively within the same time frame and allocation area, and no restrictions were applied to any of the donor and recipient variables at the time of organ allocation and splitting.
Statistical Analysis
This was a retrospective analysis of 4 different cohorts of patients consecutively collected and compared over a time span of 5 yr. Arithmetic means and standard deviations of the demographic variables were calculated for the whole sample and stratified according to different types of transplant technique. Comparisons of continuous variables among groups were analyzed using 1-way analysis of variance (t-test for parametric data and Wilcoxon rank sum test for nonparametric data), whereas differences among categorical variables have been calculated with the chi-squared test or, when necessary, Monte Carlo simulation.17 Patient and graft survival were evaluated using the Kaplan-Meier method.18 Survival estimations were compared among cohorts of patients by the log-rank test for nonparametric data. The independent contribution of several risk factors on patient and graft survival was evaluated by Cox multivariate regression analysis.19 Values of P ≤ 0.05 were considered significant. Analysis was supported by the use of the SAS statistical package, version 8 (SAS Institute, Cary, NC).
RESULTS
Utilization of SLT: Influence of Donor, Recipient, and Center Volume Variables
At first analysis, SLT represented 22.3% of the total number of LTs performed in the area under study (323/1,449 cases).
Significant differences were observed in recipient and donor distribution for several variables among the 4 possible classes of liver grafts (Table 1). In particular, patients receiving LLS were obviously younger (2.5 ± 3.9 yr), with a shorter waiting time and mostly females, compared with patients who underwent RTL or WLT. However, younger donor age was significantly associated with any kind of split liver procurement when compared to WLT (30.8 ± 15.1 vs. 43.3 ± 19.6, respectively; P < 0.0001). Transplant procedures were performed among identical donor-recipient blood types in 91.7% of cases (1,329 procedures), compatible match in 8.2% (118 patients), and incompatible blood type in only 2 cases (0.1%), both represented by patients retransplanted after a primary nonfunction graft (and receiving, as adult and child, a WLT and LLS, respectively). Noteworthy, retransplantation accounted for 9.2 to 13.6% of the series in all graft categories, with nonsignificant differences among them, with the single exception of MSL, never allocated for retransplantation and assigned to urgent patients in only 1 of 22 cases (4.5%), a significantly lower percentage with respect to other groups. MSL donors were also younger and exposed to a shorter cold ischemia time than WLT. Since adequate donor-recipient size matching was strongly pursued in case of right and left lobes procured from MSL, recipient weight in such a group was significantly lower and donor weight higher with respect to other procurement procedures for adult patients. The mean graft-to-recipient body weight ratio showed significant variations among split and whole liver groups, but it was always within the predetermined limits, also in case of MSL (1.3 ± 0.2).
| Overall | Whole liver transplantation | Left lateral segment | Right tri-segmental liver | Modified split liver | |
|---|---|---|---|---|---|
| Number of cases | 1,449 | 1,126 | 147 | 154 | 22 |
| Primary diagnosis | |||||
| Acute hepatic failure | 48 (3.3%) | 39 (3.4%) | 6 (4.1%) | 2 (1.3%) | 1 (4.6%) |
| Cholestatic diseases | 63 (4.3%) | 48 (4.3%) | 5 (3.4%) | 7 (4.6%) | 3 (13.5%) |
| Congenital biliary diseases | 114 (7.9%) | 36 (3.2%) | 75 (51.0%) | 3 (1.9%) | 0 (0%) |
| Cirrhosis | 820 (56.6%) | 699 (62.1%) | 5 (3.4%) | 100 (64.9%) | 16 (72.7%) |
| Primary liver tumors | 104 (7.2%) | 89 (7.9%) | 4 (2.7%) | 11 (7.1%) | 0 (0%) |
| Metabolic diseases | 34 (2.3%) | 20 (1.8%) | 8 (5.5%) | 5 (3.3%) | 1 (4.6%) |
| Hemangioma, polycystic, others | 126 (8.7%) | 91 (8.1%) | 24 (16.3%) | 10 (6.5%) | 1 (4.6%) |
| Re-transplantation | 140 (9.7%) | 104 (9.2%) | 20 (13.6%) | 16 (10.4%) | 0 (0%) |
| Donor age (yr) | 40.5 ± 19.4 | 43.6 ± 19.6 | 30.4 ± 15.6 | 31.2 ± 15.1 | 30.2 ± 11.5 |
| Donor weight (kg) | 68.0 ± 16.0 | 67.7 ± 16.4 | 67.3 ± 15.4 | 68.6 ± 14.1 | 79.6 ± 9.9 |
| Recipient age (yr) | 41.6 ± 19.3 | 45.8 ± 15.2 | 2.5 ± 3.9 | 47.8 ± 13.1 | 47.5 ± 10.8 |
| Recipient weight (kg) | 61.3 ± 24.2 | 67.4 ± 18.6 | 12.2 ± 10.9 | 64.9 ± 14.3 | 58.7 ± 9.3 |
| Graft-recipient body weight ratio (%) | 2.4 ± 2.0 | 2.3 ± 2.0 | 4.0 ± 1.8 | 1.7 ± 0.5 | 1.3 ± 0.2 |
| Mean waiting time (days) | 243.3 ± 311.1 | 266.9 ± 325.6 | 67.5 ± 133.5 | 232.2 ± 253.2 | 287.1 ± 389.5 |
| Cold ischemia time (hours) | 7.9 ± 2.6 | 8.1 ± 2.6 | 6.1 ± 1.9 | 8.2 ± 2.5 | 5.6 ± 2.7 |
| Combined transplant | 21 (1.6%) | 13 (1.2%) | 4 (2.7%) | 4(2.6%) | — |
| Donor/recipient gender match | |||||
| M/M | 625 (43.1%) | 495 (44.0%) | 52 (35.3%) | 70 (45.4%) | 8 (36.4%) |
| M/F | 260 (18.0%) | 157 (13.9%) | 49 (33.3%) | 42 (27.3%) | 12 (54.6%) |
| F/M | 345 (23.8%) | 297 (26.4%) | 23 (15.7%) | 24 (15.6%) | 1 (4.5%) |
| F/F | 219 (15.1%) | 177 (15.7%) | 23 (15.7%) | 18 (11.7%) | 1 (4.5%) |
| Donor/recipient AB0 match | |||||
| Identical | 1,329 (91.7%) | 1,046 (92.9%) | 122 (83.0%) | 142 (92.2%) | 19 (86.4%) |
| Compatible | 118 (8.2%) | 79 (7.0%) | 24 (16.3%) | 12 (7.8%) | 3 (13.6%) |
| Incompatible | 2 (0.1%) | 1 (0.17%) | 1 (0.7%) | 0 (0.0%) | 0 (0.0%) |
| Panel reactive antibodies | |||||
| <30 % | 1,400 (97.8%) | 1,092 (98.4%) | 142 (96.6%) | 147 (96.7%) | 19 (86.4%) |
| ≥30 % | 31 (2.2%) | 18 (1.6%) | 5 (3.4%) | 5 (3.3%) | 3 (13.6%) |
| Recipient status | |||||
| UNOS I-IIA | 217 (15%) | 167 (14.8%) | 28 (19.0%) | 21 (13.6%) | 1 (4.5%) |
| UNOS IIB-III | 1,232 (85%) | 959 (85.2%) | 119 (81.0%) | 133 (86.4%) | 21 (95.5%) |
| Center activity | |||||
| Annual ≤50 transplants | 873 (60.2%) | 717 (63.7%) | 12 (8.2%) | 125 (81.1%) | 19 (86.4%) |
| Annual >50 transplants | 576 (39.8%) | 409 (36.3%) | 135 (91.8%) | 29 (18.9%) | 3 (13.6%) |
- NOTE: Congenital biliary disease in pediatric patients was the main indication for the LLS compared to other groups (P < 0.0001). GRBWR was higher and mean waiting time lower in the LLS compared with RTL and MSL groups (P < 0.0001). Lower donor age was associated with any kind of split-liver if compared to WLT (30.8 ± 15.1 vs. 43.6 ± 19.6, respectively, P < 0.0001). In the MSL group, recipient weight and donor weight were respectively lower and higher with respect to WLT and RTL (P < 0.0001). Centers performing >50 transplants/y utilized LLS for children, more than lower-volume centers (P < 0.0001). Categorical variables are expressed as number and percentage, continuous variables are expressed as mean ± standard deviation.
- Abbreviations: UNOS, United Network for Organ Sharing; GRBWR, graft-recipient body weight ratio.
Besides the particular category of MSL, the remaining 301 split livers were allocated to urgent recipients (status I and IIA) in 16.3% of cases, with nonsignificant differences among split groups and WLT. In fact, 85% of the transplant procedures were performed in status IIB and III recipients, regardless of the technique of graft procurement. Correlation between different techniques in liver procurement, UNOS status of the recipients, and annual center volume of WLT did not show any difference, with the remarkable exception of pediatric transplantation using LLS, which was more efficient in centers performing more than 50 transplants per year with respect to lower-volume centers.
Outcomes
The median posttransplant follow-up of the presented cohorts was 22 months.
As summarized in Table 2, the overall mortality rate was 18.4%, with some significant differences among the groups. In general, LLS achieved the best results on both patient outcome (observed mortality = 12.9%,with only 3 deaths related to graft failure) and organ function (the incidence of LLS graft malfunction was 20.4%, lower than any other kind of graft; P = 0.05). As a whole, 144 grafts were lost within 15 days due to early failure (9.9% of the entire series of transplants and 41.6% of the total number of graft failures), and there was no significant correlation with different methods of graft procurement. In case of graft failure, 64.2% of recipients died (that is, 222 deaths/346 graft failures), but only 69 of 222 cases (31.1%) were lost for reasons directly related to graft failure. Among the studied cohorts, RTL and MSL were associated with a significantly worst outcome (31.8% of graft failures; P = 0.05) and higher mortality (in the range of 25-27%) when compared with WLT and LLS (for which graft failure was in the range of 20-23% and the mortality rate at the level of 13-18%).
| Prognostic factors | Overall | Whole liver transplantation | Left lateral segment | Right tri-segmental liver | Modified split liver | P |
|---|---|---|---|---|---|---|
| Graft failure | ||||||
| Early (within 15 days) | 144 (41.6%) | 106 (40.8%) | 16 (53.3%) | 20 (40.8%) | 2 (28.6%) | 0.5173 |
| Late | 202 (58.4%) | 154 (59.2%) | 14 (46.7%) | 29 (59.2%) | 5 (71.4%) | |
| Total | 346 (23.9%) | 260 (23.1%) | 30 (20.4%) | 49 (31.8) | 7 (31.8) | 0.0585 |
| ReLT in case of graft failure | ||||||
| Yes | 124 (35.8%) | 85 (32.7%) | 14 (46.7%) | 24 (49.0%) | 1 (14.3%) | 0.0516 |
| No | 222 (64.2%) | 175 (67.3%) | 16 (53.3%) | 25 (51.0%) | 6 (85.7%) | |
| Recipient death | ||||||
| Graft related | 95 (35.6%) | 75 (36.8%) | 3 (15.8%) | 14 (36.8) | 3 (50.0%) | 0.2678 |
| Nongraft related | 172 (64.4%) | 129 (63.2%) | 16 (84.2%) | 24 (63.2%) | 3 (50.0%) | |
| Total | 267 (18.4%) | 204 (18.1%) | 19 (12.9%) | 38 (24.7%) | 6 (27.3%) | 0.0425 |
- NOTE: RTL and MSL procedures were associated with a significantly worst outcome (31.8% of graft failures, P = 0.05), if compared with WLT and LLS procedures. MSL suffered a high rate of late graft failures and a consequent lower percentage of retransplantation when compared to other procurement procedures (P = 0.05). Patient mortality was higher in the RTL and MSL if compared with the other 2 groups (P = 0.04).
- Abbreviation: LT, liver transplantation.
The overall retransplantation rate in case of graft failure was 35.8%, with a nonsignificant difference observed among groups. One-third of the retransplant procedures were performed using a second split organ liver (20.8% of RTL and 57.1% of LLS, the latter being referred to 66.7% of pediatric recipients), namely, with figures that witness the general trust in split liver donation among the 7 centers of the NITp area.
A particular case was represented by the MSL techniques applied to 2 adults. These procedures suffered several limitations, mainly due to the limited experience of the LT centers (only 22 transplants performed). In fact, MSL reported a high rate of late graft failures (71.4% at >15 days) and a consequent significantly lower percentage of retransplantation when compared to other failing grafts (P = 0.05). Sepsis and multiple organ failures due to late definition of unsuccessful transplants in case of MSL accounted for the ineligibility of most recipients (5/7) for retransplantation.
Overall 3-yr patient and graft survival of the entire series of transplants were 80.6 and 74.9%, respectively, while 79.4% of the SLT patients were alive and 72.2% of the split grafts were functioning at the same interval. When survival was referred to different kinds of graft procurement, significant differences were found between 2 main subgroups of patients. In fact, patients receiving a WLT or LLS showed a 3-yr survival of 81.0 or 86.3%, respectively (Fig. 1, left), while at the same interval RTL and MSL did not set over 74.1 and 71.1% (P = 0.03). The same trend was observed when the overall 3-yr graft survival of 74.9% was parceled into the 4 groups of procured grafts (Fig. 1, right). Again, organs from WLT and LLS (that reached a 3-yr survival of 75.7 and 78.8%) confirmed a significantly better outcome (P = 0.04) than RTL and MSL (66.8 and 66.4%), with no significant difference between left/right side in case of MSL (i.e., patient receiving MSL had a 70.1 and 72.2% survival at 3 yr, while at the same interval right and left hemilivers for 2 adults reached 70.1 and 63.0% graft survival).
Patient (left) and graft (right) survival on 1,449 cases of LT: intention-to-treat analysis by type of liver procurement. WLT, whole liver transplant; LLS, left lateral segment; RTL, right trisegmental liver; MSL, modified split liver (right/left lobe) for 2 adults.
A separate survival analysis comparing WLT and RTL groups, with the exclusion of pediatric patients, was conducted: the 3-yr patient survival was 80.8 and 74.1% (P = 0.04) respectively, while graft survival was 75.9 and 66.8% (P = 0.01) at the same interval. Such a result confirmed that intrinsic differences among adult and pediatric recipients did not influence the worst outcome of RLT when compared to WLT in general (although RTLs were comparable to marginal whole grafts; see later).
There was a wide range of success among different centers in case of RTL, with a 3-yr recipient survival rate ranging from 57.8 to 87.5%. Such a difference was not significantly correlated to the volume of cases, even if a trend toward better outcome was observed in centers performing more than 50 liver transplants per year.
After adjustment of recipient age, panel reactive antibodies, and ABO blood-type compatibility, 6 factors emerged at the Cox multivariate analysis as independently related to both graft failure and patient survival (Table 3) : 1) donor age greater than 60 yr (that doubled the risk of failures with respect to younger donors); 2) right segmental grafts from split procedures (with risk almost doubling in case of RTL and MSL, compared to WLT); 3) center annual transplant volume (above or below 50 cases); 4) recipient conditions requiring urgent transplantation (UNOS status I and IIA); 5) cold ischemia time above 7 hours; and 6) case of retransplantation.
| Factors analyzed | RR | 95% CI | P |
|---|---|---|---|
| Donor age (yr) | |||
| ≤15 vs. 16–60 | 0.831 | 0.469–1.473 | 0.5270 |
| >60 vs. 16–60 | 2.004 | 1.541–2.607 | <0.0001 |
| Graft type | |||
| Left lateral segment vs. whole liver transplant | 0.913 | 0.457–1.826 | 0.7975 |
| Right tri-segment liver vs. whole liver transplant | 1.755 | 1.247–2.468 | 0.0012 |
| Modified split liver vs. whole liver transplant | 2.117 | 0.925–4.849 | 0.0760 |
| Center transplant volume | |||
| Annual ≤50 vs. >50 transplants | 1.341 | 1.034–1.741 | 0.0272 |
| Recipient UNOS status | |||
| I-IIA vs. IIB-III | 2.248 | 1.642–3.076 | <0.0001 |
| Cold ischemia time (hours) | |||
| >7 vs. ≤7 | 1.404 | 1.115–1.768 | 0.0039 |
| Number of transplants | |||
| First transplant vs. retransplantation | 1.836 | 1.289–2.615 | 0.0008 |
- NOTE: Adjusted for recipient age, panel reactive antibodies, and ABO blood group match.
- Abbreviations: UNOS, United Network for Organ Sharing; RR, relative risk; CI, confidence interval.
Comparison of Split Livers to Marginal and Nonmarginal Whole Livers
Since split liver procedures were mostly performed in otherwise optimal donors, and having previous data confirmed that partial right liver grafts (RTL and MSL) suffered a worst outcome than whole and left livers, a further comparison was made between right liver grafts and marginal whole liver donors. After stratification of recipient conditions and adjustment of the other variables, RTL and MSL were compared to the whole livers procured from donors of >60 yr or implanted after a long (>12 hours) ischemia time. The results of the multivariate analysis after adjustment of the unmatched variables among groups are reported in Table 4.
| Right segmental grafts vs. marginal whole livers | Number of Patients | Graft failure | Patient death | ||||
|---|---|---|---|---|---|---|---|
| RR | 95% CI | P | RR | 95% CI | P | ||
| Overall | 505 | 0.980 | 0.696–1.380 | 0.9064 | 0.886 | 0.606–1.294 | 0.5307 |
| UNOS I-IIA | 60 | 1.368 | 0.655–2.855 | 0.4041 | 1.017 | 0.459–2.256 | 0.9666 |
| UNOS IIB-III | 445 | 0.828 | 0.551–1.243 | 0.3616 | 0.798 | 0.506–1.260 | 0.3335 |
- NOTE: Right grafts procured from RTL and MSL had a similar outcome of “marginal” whole livers, with an overall graft failure and mortality risk not statistically different among groups (RR 0.980 and 0.886, respectively). Graft failure and mortality risk of the 2 groups were also comparable in the subset of urgent (60 cases of UNOS status I-IIA vs. 445 patients IIB-III) and elective recipients. Adjusted for indication to transplant, donor/recipient gender, ABO blood group compatibility, retransplantation, recipient age, and center transplant volume.
- Abbreviations: UNOS, United Network for Organ Sharing; RR, relative risk; CI, confidence interval.
Right grafts procured from RTL and MSL had a similar outcome of marginal whole livers, with an overall graft failure rate and mortality not statistically different among groups. Noteworthy, patient and graft survival of right livers and marginal whole liver donors were comparable also in the case of urgent recipients (60 cases of UNOS status I and IIA vs. 445 patients of status IIB and III). In both urgent and elective settings, the relative risk of receiving a whole liver from a marginal donor was almost identical (ranging from 0.8-1.0) to the one calculated for a RTL graft or a hemiliver for 2 adults.
Besides such a limitation in outcome suffered by RTL and MSL, no other differences were detected among groups, confirming the previous observation that the splitting procedures applied on optimal donor grafts yield right grafts that behave similarly to whole livers procured from marginal donors.20 However, a significant influence of SLT was registered in the area under study, as expressed by the reduction in the dropout rate (for disease progression or death) in patients listed for LT during 5 consecutive yr (from 27.2 to 16.2% at the end of the study period; P = 0.001). During the same period, a 21.2% increase in cadaveric donor procurement rate was observed in the NITp area, and that has to be acknowledged as a significant contribution to the general decreases in the waiting list mortality.
DISCUSSION
Despite intensive education programs on cadaveric donation and increasing cooperative efforts among transplant centers and procurement agencies, the practice of split liver procurement is still under-performing with respect to its potential, both in Europe and the United States. Split liver grafts represent 4% of the total number of transplants in Europe21 (as reported in the European Liver Transplant Registry: 124 centers from 1968 to 2001), and according to a national survey conducted in the United States, only 45% of the responding teams reported any experience with cadaver SLT, with two-thirds of the centers being experienced in SLT with less than 5 procedures done.20
The present report on 323 patients transplanted with a cadaveric split liver graft represents the second series ever reported worldwide and the largest in Europe.21 In fact, in the NITp, 21.8% of candidates collected from 1997 to 2002 were transplanted either by the use of older donors (248 cases > 60 yr of age, that is, 19% of the whole transplant population) or by split organs in 22.3% of the cases. Such a trust in split liver procedures has been higher in the NITp area than in other large programs, even if the protocol applied to cadaveric split liver donation in the present experience was quite similar to those adopted by other groups.22, 23 An effective cooperation among LT units and surgical teams procuring organs other than the liver has to be acknowledged as the main reason for the success of the presented experience.
In our setting, the in situ splitting technique was extensively used with no detrimental effect on other organs procured from the same donor. In fact, the donors eligible for splitting were optimal by any means, and the in situ splitting allowed (as in living donation) a dreadful reduction in cold ischemia time, with an ultimate increase in the possibility of graft sharing over longer distances.
At the multivariate analysis, older donor age significantly influenced graft survival (Table 3). Although the impact of donor age on posttransplant outcome in case of adult recipients has already been reported,21, 24 it has to be emphasized that any possible drawback associated with older donors is balanced by the magnitude of the shortage of organs. Experienced centers confirmed that, in case of proper selection of the donor with exclusion of other detrimental factors during the procedure, age should not be taken per se as a contraindication for split liver procedures.25
Cold ischemia time under 7 hours was a significant prognostic factor for successful liver function (Table 3), and the prolongation of organ retrieval time due to additional in situ liver divisions did not influence the posttransplant function of thoracic organs harvested from the same donor, with respect to whole liver procurements. In fact, survival of hearts or lungs collected in the NITp area over the same period of time was not significantly different, whether or not a split liver procedure was added during procurement. Heart or lung 5-yr survivals from split liver donors were 81.0 and 58.9%, respectively, not significantly different from 79.5 and 59.1% survival registered when hearts or lungs were procured from non-split liver donors.
Both patient and graft survival after SLT were adequate (79.4 and 72.2% for patient and graft 3-yr survival, respectively) and generally better than those reported previously.20, 21, 23 In fact, of the approximately 1,770 split liver cases reported in the European Liver Transplant Registry, 5-yr patient survival was 57%.21 However, with respect to other indicators of outcome, the NITp split liver series was associated with a 11.1% retransplantation rate (that is, higher than the 7% reported in the European Liver Transplant Registry) and with a 2.1 and 2.7 risk of graft failure in case of split liver used for retransplantation and/or emergency settings (UNOS status I and IIA). A similar detrimental effect of urgent recipient status was observed in other large series of SLTs, confirming a 4-fold risk of death in case of UNOS status I receiving LLS, compared with nonurgent recipients. Moreover, a 25% mortality was observed in UNOS status I and IIA patients who were transplanted with RTL.20
The present large experience confirms how rewarding the implementation of a split policy among liver transplant centers can be. Due to such an attitude, a reduction of patient mortality was observed in the NITp area. In particular, 3-yr actuarial patient survival significantly improved whether or not the patients received a liver graft (80.6 vs. 50.3%, respectively; P = 0.001), and a progressive reduction in the dropout rate (for disease progression or death) in patients listed for LT was achieved during the 5 yr of the study (from 27.2 to 16.2%). Such effects, although not exclusively due to SLT, were registered only in the period under study, namely, when the procedures of in situ liver division were implemented.
However, distinct outcomes were noticed with respect to different kinds of grafts procured. In general, LLS behaved better than RTL, exceeding the results pioneered in the early 1990s with such a liver division technique.12-14 It has to be emphasized that the division of the arterial supply to the grafts applied in the North Italian experience gives some advantages in terms of revascularization to the LLS compared with the RTL. Division of the hilar structures to give preference to the pediatric recipients, while incurring added risk to the right-lobe adult recipients (Table 3 and Fig. 1), could be questionable, since the present series of RTL had poorer outcome than other large series (e.g., UCLA, Hamburg), where the hilar structures remain with the right side.12, 20, 23 In fact, causes of graft failures (Table 5) showed that splitting techniques leaving hilar structures with the left side exposed RTL to a significantly higher risk of vascular complication, and ultimately to infection with respect to the more preserved vascular anatomy of LLS. However, in the particular framework of the North Italian procurement area, the split liver policy in liver donation was implemented, with all centers being aware that reduction of the pediatric list mortality was the primary endpoint of the program. In addition to technical reasons, other factors related to adult recipients' conditions and to different donated/expected liver mass must have affected the generally worst results of right liver transplants, compared to left splits (Fig. 1). Nevertheless, a 3-yr survival of 67% of the RTL (and of 74% for the recipients) seems reasonable if compared to the life gain achieved on both pediatric and adult recipients.
| Whole liver transplantation | Left lateral segment | Right trisegmental liver | WLT-LLS comparison (P value) | WLT-RTL comparison (P value) | |
|---|---|---|---|---|---|
| Number of cases | 1126 | 147 | 154 | ||
| Causes of failure | |||||
| Hepatic artery thrombosis | 37 (3.3%) | 5 (3.4%) | 11 (7.1%) | 0.9996 | 0.0100 |
| Hepatic vein-portal thrombosis | 4 (0.4%) | 1 (0.7%) | 3 (1.9%) | 0.5776 | 0.0070 |
| Primary nonfunction | 64 (5.7%) | 8 (5.4%) | 10 (6.5%) | 0.8411 | 0.4744 |
| Hemorrage (intra-peri operative) | 12 (1.1%) | 2 (1.4%) | 3 (1.9%) | 0.7848 | 0.2583 |
| Biliary complications | 7 (0.6%) | 0 (0.0%) | 1 (0.6%) | 0.3310 | 0.8785 |
| Infection | 40 (3.6%) | 4 (2.7%) | 12 (7.8%) | 0.5714 | 0.0068 |
| Rejection | 10 (0.9%) | 1 (0.7%) | 2 (1.3%) | 0.7744 | 0.5176 |
| Recurrent disease (viral-cancer) | 24 (2.1%) | 2 (1.4%) | 4 (2.6%) | 0.5112 | 0.5613 |
| Cardiovascular complications | 13 (1.2%) | 2 (1.4%) | 1 (0.6%) | 0.8652 | 0.6599 |
| Gastrointestinal complications | 5 (0.4%) | 0 (0.0%) | 0 (0.0%) | 0.4113 | 0.4364 |
| Cerebral complications | 14 (1.2%) | 1 (0.7%) | 0 (0.0%) | 0.5331 | 0.1930 |
| De novo tumor | 5 (0.4%) | 1 (0.7%) | 1 (0.6%) | 0.7197 | 0.6458 |
| Other | 25 (2.2%) | 3 (2.0%) | 1 (0.6%) | 0.8480 | 0.2554 |
- NOTE: Right trisegmental liver grafts (RTL) were exposed to a significantly higher rate of vascular complications (and ultimately to infections) with respect to whole liver transplants (WLT). That was possibly related to Center experience and to the hilar division adopted in this study, in which preference was given to the left lateral segment (LLS) to be used for pediatric recipients (see text).
Even though the decision of transplanting a right liver was burdened by a risk 1.7 higher than that of a whole organ (Table 3), RTL and MSL outcomes were quite similar to that of marginal whole liver graft procured from older donors (Table 4). Interestingly, the parallelism between right liver from split and marginal whole liver was independent from the clinical conditions of the recipient (whether urgent or elective cases), meaning that in situ liver division is a procedure that has to be performed under stringent conditions of donor eligibility20 and by experienced surgical teams.
Similarly to the European Liver Transplant Registry report,21 also in the present series, short- and long-term survival of split livers (particularly right livers) were higher in the most experienced centers. However, it is likely that such a difference would become less significant as the learning phase is completed in all centers, with at least 50 cases of SLT being performed by each group.
In the NITp area, a small, although specific, experience for modified splitting techniques (MSL) was collected and applied to 2 small-size, nonsensitized, first-transplant and nonurgent adult recipients. Such a procedure, performed by experienced teams under favorable logistic conditions, in young donors with adequate liver mass and a very short cold ischemia time, was applied in only 22 cases, and therefore in a largely experimental mode. Several biases in application of MSL have to be acknowledged, and that is confirmed by the observed 2-fold increment in the risk of graft failure in case of modified splitting, when compared to whole organ transplants. Nevertheless, technical refinements in split liver for 2 adults (for both right or left hemilivers) should be pursued, although under stringent conditions and by very experienced teams, since such a technique owns the potential for a future increase in organ availability for adult liver transplant candidates.26
As reported previously,27, 28 the NITp experience confirms that an extensively applied split liver policy for organ donation is crucial in expanding the pool of organs available for pediatric recipients. Thanks to split livers, the pediatric in-list mortality of the area under study has dropped to 3% in 2004, as well as the need of living-related pediatric liver transplant. Moreover, when 2 consecutive 5-yr periods were compared (presplit era, 1992-1997 vs. split era, 1998-2002), the length of the waiting list for children dropped from 6.8 to 3.9 months, and the mortality rate from 16.6 to 8.1% (P = 0.001), even though the number of pediatric donors did not change over time (173 vs. 175, respectively). On the other hand, the availability of right livers for adults has been confirmed as a sound resource for adult recipients, with a moderate increase in the risk of failure that is certainly balanced by the life gain given to patients in the adult waiting list.
In conclusion, the presented split liver-oriented program (7 centers serving an area of 18 million inhabitants) has been quite effective in increasing the number of available liver grafts, especially for children. The MSL technique designed to comply with 2 small adult recipients is still experimental, although worth the effort by experienced teams under stringent logistic conditions. A sound cooperation among LT centers is the key factor to overcome the organizational difficulties that limit a more extensive application of the split procedures. Even though split liver sharing is still an uncommon event,20 allocation policies promoting application of splitting procedures should be encouraged.
Acknowledgements
Crucial comments by Claudia Pizzi, Francesca Poli, and Francesca Marangoni were greatly appreciated, and the help of Daniela Guarneri and Elena Bertocchi is acknowledged in editing the manuscript. Authors' special thanks goes to the personnel of the donor and recipient hospitals belonging to the NITp, and to the families who gave their consent to organ donation.
