Objectives: To investigate an association between short-term mortality and donor age-associated worst outcomes in liver transplant.
Materials and Methods: A total of 178 consecutive patients underwent a liver transplant between 1999 and 2007. Among these patients, there were 172 liver transplants (donor age, 32.04 ± 16.66; range, 2-65 years) and 167 recipients. Mean recipient age was 39.16 ± 21.61 years (range, 6 months to 71 years), and 90 were males (53.8%).
Results: Among 172 transplants, 32.9% recipients died during follow-up (mean, 34.37 ± 20.50 months). A lower mean recipient and graft survival occurred in donors older than 50 years (P = .01) and 30 years (P = .02) at 7-year patient survival. At 6-month and 1-year recipient survival, cutoffs were 50 and 55 years (P < .05). Log-rank test showed no statistical difference among recipients, and graft survival from donors older/younger 50 and 30 years 1.5 years after liver transplant (P < .565 and P < .259).
Conclusions: Donor age is a key factor in liver transplant that carries prognostic impact in the recipients. Our data suggest that its harmful effects are exclusively elicited during the short-term, postoperative phase. We recommend careful and distinct management of recipients receiving grafts from elderly donors up to 1.5 years after liver transplant. Changes in the current early postoperative management of this selected group are encouraged.
Key words : Patient survival, Elderly donor
Orthotopic liver transplant has experienced successful results over the last years, however, it is followed by an increased number of patients on waiting lists. To expand the pool of liver donors, strategies have been used, particularly use of extended criteria donors, including elderly donors (1).
Previous studies have reported adverse effects in recipient survival associated with increased donor age (2). Although several studies have failed to deeply characterize the actual influence of this variable, a decline in the survival rates of recipients who received organs from donors older than 45, 50, or 60 years of age has been reported (3-5). Donor age greater than 60 years significantly increased 3-month mortality in the large analysis of European Liver Transplant Registry (6). Several factors may be involved, including a higher degree of steatosis, higher incidence of primary dysfunction, and reduced functional capacity (7-9). Additionally, elderly donors have been associated with more acute hepatitis C virus graft recurrence (10). As a consequence, advanced donor age is considered a contraindication for liver procurement in several transplant centers (11).
Despite the fact that elderly donors have been associated with worse recipient survivals, the exact time at which this variable may be associated with increased risk remains controversial. Based on previous studies of our group (12), we hypothesize that donor age-related poor outcomes are exclusively elicited in the short-term postoperative phase after liver transplant. We therefore aim to analyze a possible, exclusive, harmful effect of advanced donor age in short-term outcomes after liver transplant. To our knowledge, there have been no previous reports that assessed a correlation of these variables based on the statistical tools implemented in this study.
Materials and Methods
A total of 178 consecutive liver transplants performed in 173 patients from August 1999 to June 2007 were analyzed retrospectively. All patients were followed up until June 2007 for a minimum of 6 months (mean, 34.37 ± 20.50 months; range, 24.4 to 377.9 weeks). For survival analysis, the population was stratified by donor age cutoffs with intervals of 5 years. The study conforms with the ethical guidelines of the 1975 Helsinki Declaration. Written, informed consent was obtained from all subjects. The study was approved by The Oswaldo Cruz Hospital Ethics Research Committee.
Brain death was diagnosed in accordance with guidelines of Brazilian law. Whole liver grafts were recovered by the classic technique with resection of the vena cava, and flushed with cold Belzer solution. Both common iliac veins and common iliac arteries were recovered from each donor to be used, if appropriate. Deceased liver grafts were implanted using a standard conventional technique without venovenous bypass or in the piggyback fashion by anastomosing the hepatic vena cava to the junction of the 3 hepatic veins.
Pedicle elements were anastomosed using standard techniques. Biliary reconstruction was performed with choledochocholedochostomy or Roux-en-Y choledochojejunostomy as appropriate. Intraoperative biopsies were not acquired. Liver steatosis was predictable subjectively, considering liver color and characteristics of the liver border. The immunosuppressive protocol consisted of tacrolimus, prednisone, and mycophenolate mofetil for adults, and tacrolimus and prednisone for children.
Continuous variables are expressed as means ± standard deviation. Outcomes are expressed as means for continuous data, and as frequencies (percentages) for categoric data. The primary endpoint of the study was overall survival, which was defined as the time from liver transplant until death from any cause or last follow-up.
Cumulative patient and graft survival probabilities were estimated with the use of the Kaplan-Meier method. The 1-sample log-rank test (13) was used to compare survival among different donor age cutoffs. The significance level for overall mortality was significant if P < .05 after adjusting for 4 interim analyses with Fleming-Harrington group-sequential boundary (14). This analysis is performed by using P and q values where P = 0.0 and q > 0 representing analyses at long-term survival. This statistic tool is not capable of elucidating the point where the curves were compared. As a consequence, we empirically adopted the value of 542 days (1.5 years) to compare both curves by using log-rank test. At this point, there is less influence of early postoperative confounding factors.
The value for donor age was related to mortality from all causes by univariable Cox regression analyses (15). We performed a stratified analysis of prespecified subgroups defined according to sex, age of donor and recipient, and operative data. Statistical analyses were performed using Stata 9.2 software (StataCorp, College Station, TX, USA).
Donor characteristics are summarized in Table 1. Among 178 consecutive liver donors, 172 liver grafts were considered for the analysis (mean age, 32.04 ± 16.66 years; range, 2-65 years) with 88 males (51.2%). Death diagnosis was brain hemorrhage in 61 cases (35.5%), cranium-encephalic trauma in 54 cases (31.4%), nervous system neoplasm in 13 cases (7.6%), and other causes in 19 cases (11%). Ten living donors (5.8%) were included in this series. Intensive care unit time was more than 4 days in 35 of cases (20.3%). Twenty-six patients (15.1%) had a cardiac arrest before organ retrieval and 6 donors (3.4%) were classified as nonheartbeating donors.
Vasoactive drugs were used in 142 donors (82.5%). Twenty donors (11.6%) showed international normalized ratio over 2.0, and in 55 cases (31.9%), serum sodium was greater than 155 mEq/L. Concerning macroscopic steatosis, 62 donors (36.0%) were considered to have greater than 30% liver steatosis. Seven liver donors (4.0%) had body mass index over 30. Risk factors were calculated by univariate analysis as presented in Tables 1 and 2. The only factor that achieved significance was donor age (P = .01) in both recipient and graft analysis. Of 6 nonheartbeating donors, 4 were alive during follow-up (66.7%; mean survival, 839 ± 669 days; range, 49-1680 days).
Among 172 liver grafts, 3 grafts (1.7%) were from domino donors, 10 (5.8%) were from living donors, and 26 (15.1%) were from donors younger than 14 years of age.
Donor distribution by age groups was ≤ 25 years: 72 (41.9%); 26-35 years: 26 (15.1%); 36-45 years: 29 (16.9%); 46-55 years: 31 (18%); and > 55 years: 14 (8.1%). Minimum follow-up was 6 months, with a mean value of 34.37 ± 20.50 months. Of the 167 patients, 55 (32.9%) died during follow-up.
Recipient characteristics are summarized in Table 2. Among 173 recipients, 167 (male, 53.9% and female, 46.1%) were included in the analysis (mean age, 39.16 ± 21.61 years; range, 6 months to 71 years). The main diagnostic categories among patients listed for liver transplant were hepatitis virus C-related cirrhosis (24.6%); alcoholic cirrhosis (22.7%); hepatocellular carcinoma (7.1%); hepatocellular carcinoma associated with hepatitis virus C-related cirrhosis (6%); hepatitis virus B-related cirrhosis (2.4%); concomitant hepatitis virus C-related cirrhosis, and alcoholic cirrhosis (5.4%). According to the model for end-stage liver disease classification, the mean score was 18.31 ± 7.31 (range, 6-47), and 47 (28.1%) recipients presented model for end-stage liver disease score below 15.
Cold and warm ischemia times were 7.16 ± 3.01 hours (P < .091) and 0.98 ± 0.60 hours (.678). Liver grafts were implanted by standard conventional technique without venovenous bypass or by piggyback fashion in 71.3% and 29.6% of cases. Four patients (2.4%) required retransplant during follow-up, and 1 needed it twice. Three of these recipients died. In 18 cases, there were primary dysfunction of livers, and 6 primary nonfunctions. Most transplants were based on chronologic criteria (90%).
The 7-year patient survival was 67.1%. A lower mean survival rate was observed among recipients of grafts from donors older than 50 years (P = .01) than among those who were transplanted from younger donors. The difference remains continuously significant at ages of 45 (P = .05), 40 (P = .03), 35 (P = .04), 30 (P = .02), and 25 years (P = .01) (Figure 1-A).
Overall 6-month and 1-year patient survivals were 76.6% and 71.2%. By stratifying recipients according to different donor age cutoffs, no statistically significant difference was identified among recipients from donor age above and below 60 and 55 years at a 6-month interval, and 60 years at a 1-year patient survival (data not shown). The cutoffs for 6-month and 1-year survivals were 50 and 55 years (Figure 2-A/B). Table 3 summarizes the survival values according to cutoffs.
Four patients (2.9%) required retransplant during follow-up, and 1 needed it twice. Three of these recipients died. The 7-year graft survival was 64.5%. A significant difference was observed among graft survival from donors older than 30 years (P = .025) (Figure 1-B).
Overall 6-month and 1-year graft survivals were 75.0% and 69.1%. By stratifying donor age according to liver grafts, no statistically significant difference was identified among grafts from donor age above or below 60, 55, 50, 45, and 40 years at 6-month intervals, and above or below 60 and 55 years of age at 1-year patient survival (cutoffs, 40 years; P = .028; 45 years; P = .05; 50 years: P = .023) (data not shown). Cutoffs for 6-months and 1-year graft survivals were 35 and 50 years (Figure 2-C/D). Table 4 summarizes the graft survival values according to cutoffs.
Analyses of survival
Based on the comparison between the patient survival curves according to cutoff of 50 years of age, at values of p=.0; q=0.01, and p=.0; q= 0.1, the Fleming-Harrington test showed a statistically significant difference between the curves (P = .040 and P < .042). However, at values p=.0; q=1.0, the test revealed no statistical difference between the curves in the long-term (P < .111). Comparing graft survival and adopting 30 years of age as predetermined cutoff, at values p=.0; q=0.01, the Fleming-Harrington test showed a statistically significant difference between both curves (P < .04). At values p=.0; q=0.5, there was no statistical difference between graft survival in long-term analyses (P < .133).
Using the log-rank test, patient and graft survival curves were compared by empirically adopting initial time follow-up of 542 days (1.5 years). From this point, we found no statistical significant difference between survival curves in patient and graft analyses (P = .406 and P < .259) (Figures 3/4).
There is increasing awareness among the transplant community regarding the potential effect of aggressive-use practices on graft and patient outcomes as the frontiers of use continue to expand previously defined boundaries. There has been increased procurement of older donors in liver transplant accounting for about 13% in 1988 to 54% of adult liver transplants in 2003 (16). Although a substantial body of literature has long recognized that donor age is a significant factor related to graft failure and patient mortality, this is still a controversial issue. While several reports (12) have demonstrated worse associated outcomes, other single-center comparative studies have generally failed to demonstrate significant differences in graft and patient survival after elderly donor transplant (17, 18). Particularly, elderly livers present high crispness, which is partly related to the endothelial cell injury from cold ischemia and decreased adenosine triphosphate synthesis after reperfusion (19). Livers from elderly donors also present cholestatic patterns after transplant (20), an increased incidence of steatosis, and a higher progression of fibrosis and cirrhosis. Ultimately, development of cirrhosis in elderly grafts was associated with hepatocyte telomere shortening correlated with senescence (21).
Recently, great efforts have been made in the stratification of risk for orthotopic liver transplant outcomes. Surgical strategies have been proposed to increase the number and the suitability of donors older than 60 years of age, including selection of donors according to the presence of steatosis (22). Feng and associates (16), based on a study of 9882 deceased donors from the American Scientific Registry of Transplant Recipients, developed a donor risk index, which included 5 donors and 2 graft characteristics associated with higher risk of graft loss. Among those, donor age greater than 60 years was identified as the strongest risk factor for graft failure.
Increased results in liver transplant partly related to surgical and anesthetic improvements report a 5-year survival of 59% to 75% (23, 24). Using elderly donors, liver transplant appears to present survival rates approximately to 56% to 75% at 1 year, and 25% to 73% at 5 years (2, 19, 20, 22). Evidence has shown that these donors, without additional risk factors for liver transplant, presented similar survivals compared with younger donors (2). We present our 7-year overall survival of 67.1%, results that are comparable to those reported in the literature. Donor age cutoff associated with long-term lower recipient and graft survivals were 50 and 30 years. Considering 6-month and 1-year survivals, donor age cutoff observed was 50 and 55 years for patients and 35 and 50 years for graft survival. Because 6-month and 1-year patient survivals were 76% and 75%, and 6-month and 12-month graft survivals were 75% and 69%, patients died mainly from nonhepatic causes. In fact, there were only 4 deaths caused by primary graft failure in the present series.
The majority of previous published studies has examined the effect of donor age on survival in liver transplant and its inclusion in different risk models. In contrast, our study focuses on donor age at the point of time where this variable has the most influence for graft failure and determined the time at which donor age no longer has an effect on survival after liver transplant. In our analyses, several donor characteristics, previously identified by others as risk factors, achieved no significance. These included model for end-stage liver disease score, cold and warm ischemia times, donor sex, liver enzymes (ie, aspartate aminotransferase and alanine aminotransferase), sodium, living donors, and nonheartbeating donors. However, we acknowledge that there was trend toward a relation of donor sex and transplant outcomes in this series (P = .55, Table 1).
Furthermore, we observed that elderly female donors (> 50 years of age) presented with slightly lower survival rates compared with total elderly patients, but this did not reach statistical significance (858.5 ± 117.1 vs 1512.9 ± 138.8 days; P = .354). In this series, nonheartbeating donors were used in only 6 cases with acceptable outcomes (survival rate = 66.7%). Donor risk index includes age, donation after cardiac death, split/partial status, height, cause of brain death, macrosteatosis, and cold ischemia time as factors associated with transplant outcomes (16). In our case, owing to a well-organized effort of team synchronization, mean cold ischemia time was less than 8 hours. Some reports have demonstrated that by maintaining cold ischemia time at 8 hours or less, long-term graft function has similar outcomes among donors older or younger than 50 years of age (16). Each additional hour of cold ischemia time was associated with an additional 1% increased risk of graft loss. Furthermore, compared to grafts transplanted outside the local area but within the same region, they had an 11% increased risk of graft loss (16). Therefore, we believe that in elderly transplant, the share of grafts outside of local donor area service should be avoided as a means of reducing prolonged cold ischemia time.
Based on our previous studies (12), we identified donor age as significantly and independently associated with increased mortality of liver transplants on short- and long-term outcomes. However, we hypothesized that the advanced age presented a negative cumulative effect on the survival curves that could lead ultimately with no significance when analyzed separately, that is, short- and long-term. Based on our single-center data, we found no influence of donor age in both recipient and graft survivals after 1.5 years (542 days) after liver transplant. The harmful effects of advanced donor age were present exclusively in the short term after liver transplant.
We acknowledge that our study has several important limitations. In addition to relatively small numbers of patients, we faced issues regarding data collection. As a means, we decided to exclude 6 patients owing to a lack of relevant data. Additionally, the recipient selection, enforced by Brazilian law, was performed according to chronologic criteria until July 2006. Therefore, several patients with model for end-stage liver disease scores below 15 underwent orthotopic liver transplant at our institution. It has been stressed that the outcome of orthotopic liver transplant is related to the status of the recipient. Recipients with better health status are selected by chronologic criteria that have been slightly influenced by the use of nonoptimal donors with extended criteria.
Model for end-stage liver disease score addresses urgency, but it is not used for an individual patient. Use of a transplant still is determined by physicians. This judgment has no standardization and is influenced by urgency. Therefore, it remains unknown whether introduction of the model for end-stage liver disease-based prioritization system will change the efficiency of orthotopic liver transplant—a key parameter to be considered in elderly transplant (19). Additionally, most orthotopic liver transplants were performed by standard conventional technique without venovenous bypass. Ultimately, we decided to not perform a liver biopsy routinely in our center to estimate overall liver histologic characteristics. Although macrosteatosis has been associated with transplant outcomes (16, 25), the degree of macrosteatosis, which may or may not be known at the time of organ offer, was not significantly associated with graft failure (16). Furthermore, even using frozen section liver biopsy, estimation of steatosis is both difficult and subjective (26).
Given the current graft shortage, elderly donors should not be considered a contraindication. Those donors can be used with acceptable results, especially if they are carefully selected. We demonstrated that donor age is mostly, if not exclusively, associated with negative outcomes in the early postoperative phase, particularly up to 1.5 years after liver transplant. Considering that early mortality is representative of surgical outcomes, we recommend more careful attention to the intraoperative management of recipients of grafts from donors > 50 years of age. We believe that this finding may guide implementation of a specific and distinct early postoperative management of recipients of elderly grafts. Similarly, allocation of elderly grafts should be especially managed in a separate setting in the future as a means of improving a match between donor and recipient characteristics.
Volume : 8
Issue : 3
Pages : 202 - 209
From the 1Department of Surgery and Liver Transplantation, Oswaldo Cruz
University Hospital, University of Pernambuco School of Medicine and 2Division
of Biostatistics, Instituto de Medicina Integral de Pernambuco, Recife, Brazil
Address reprint requests to: Francisco Igor B Macedo, 310, Arnóbio Marques Street, Santo Amaro Recife-PE Brazil 50130 060
Phone: +55 81 9734 1314/+55 81 3223 3902
Fax: +55 81 3223 3902
E-mail: firstname.lastname@example.org / email@example.com
Table 1. Donor characteristics.
Table 2. Recipient characteristics.
Figure 1. A. Survival curve of patients transplanted with grafts from donors older than 50 years (cutoff; P < .05). B.At 7-year graft survival, cutoff of grafts transplanted with liver fromdonors older than 30 years (P < .05).
Figure 2. A/B.At 6-month recipient survival,Kaplan-Meier curves of patients transplanted with grafts from donors older or younger than 50 years and at 1-year patient survival, cutoff of 55 years of age (P < .05). C.At 6-month graft survival, curves from donors older and younger than 35 years of age. D. At 1-year graft survival, cutoff of 50 years of age (P < .05).
Table 3. Recipient survival according to donor age cutoff.
Table 4. Graft survival according to donor age cutoff.
Figure 3. Long-term patient survival curves according to donor age cutoff (50 years of age). Log-rank test found no statistical difference between the curves after 1.5 years (P < .406). t=0, 542 days (1.5 years).
Figure 4. Long-term graft survival curves according to donor age cutoff (30 years of age). Log-rank test found no statistical difference between the curves after 1.5 years (P < .259). t=0, 542 days (1.5 years).