Telephone: 414-955-6766; FAX: 414-955-6215
Recipient survival and graft survival are not diminished by simultaneous liver-kidney transplantation: An analysis of the united network for organ sharing database†
Article first published online: 24 JUL 2012
Copyright © 2012 American Association for the Study of Liver Diseases
Volume 18, Issue 8, pages 914–929, August 2012
How to Cite
Martin, E. F., Huang, J., Xiang, Q., Klein, J. P., Bajaj, J. and Saeian, K. (2012), Recipient survival and graft survival are not diminished by simultaneous liver-kidney transplantation: An analysis of the united network for organ sharing database. Liver Transpl, 18: 914–929. doi: 10.1002/lt.23440
The statistical analysis was supported by grant 1UL1RR031973 from the Clinical and Translational Science Award program of the National Center for Research Resources in association with the National Institutes of Health.
- Issue published online: 24 JUL 2012
- Article first published online: 24 JUL 2012
- Accepted manuscript online: 29 MAR 2012 06:13AM EST
- Manuscript Accepted: 20 MAR 2012
- Manuscript Received: 20 SEP 2011
Recipients of solitary liver and kidney transplants are living longer, and this increases their risk of long-term complications such as recurrent hepatitis C virus (HCV) and drug-induced nephrotoxicity. These complications may require retransplantation. Since the adoption of the Model for End-Stage Liver Disease, the number of simultaneous liver-kidney transplantation (SLK) procedures has increased. However, there are no standardized criteria for organ allocation to SLK candidates. The aims of this study were to retrospectively compare recipient and graft survival with liver transplantation alone (LTA), SLK, kidney after liver transplantation (KALT), and liver after kidney transplantation (LAKT) and to identify independent risk factors affecting recipient and graft survival. The United Network for Organ Sharing/Organ Procurement and Transplantation Network database (1988-2007) was queried for adult LTA (66,026), SLK (2327), KALT (1738), and LAKT procedures (242). After adjustments for potential confounding demographic and clinical variables, there was no difference in recipient mortality rates with LTA and SLK (P = 0.02). However, there was a 15% decreased risk of graft loss with SLK versus LTA (hazard ratio = 0.85, P < 0.001). The recipient and graft survival rates with SLK were higher than the rates with both KALT (P <0.001 and P <0.001) and LAKT (P = 0.003 and P < 0.001). The following were all identified as independent negative predictors of recipient mortality and graft loss: recipient age ≥ 65 years, male sex, black race, HCV/diabetes mellitus status, donor age ≥ 60 years, serum creatinine level ≥2.0 mg/dL, cold ischemia time > 12 hours, and warm ischemia time > 60 minutes. Although the recent increase in the number of SLK procedures performed each year has effectively decreased the number of potential donor kidneys available to patients with end-stage renal disease (ESRD) awaiting kidney transplantation, SLK in patients with end-stage liver disease and ESRD is justified because of the lower risk of graft loss with SLK versus LTA as well as the superior recipient and graft survival with SLK versus serial liver-kidney transplantation. Liver Transpl, 2012. © 2012 AASLD.
See Editorial on Page 875
Improvements in posttransplant management and particularly immunosuppressive therapy have led to a dramatic increase in patient survival after solid organ transplantation. Specifically, immunosuppressive therapy with the calcineurin inhibitors (CNIs) cyclosporine and tacrolimus has enhanced survival after orthotopic liver transplantation (OLT).1, 2 Despite their association with improved survival, CNIs are inherently nephrotoxic, and they are often cited as the main cause of chronic renal failure (CRF) after OLT.3-7 In the first 6 months after OLT, CNIs have been associated with a nearly 30% decline in the glomerular filtration rate (GFR).3 Moreover, CNIs contribute to the development of CRF in approximately 18% of OLT patients after 13 years; this number includes patients (9.5%) who progress to end-stage renal disease (ESRD).3 Because nephrotoxicity may lead to the discontinuation of CNIs, less effective immunosuppressive agents are used, and this may result in more frequent liver and renal graft dysfunction. For these reasons, CNI toxicity and the acceleration of underlying liver and renal disease may necessitate subsequent liver and/or kidney retransplantation.
Since the implementation of the Model for End-Stage Liver Disease (MELD) as an objective allocation system by the United Network for Organ Sharing (UNOS) in 2002, priority has shifted to end-stage liver disease (ESLD) patients with renal insufficiency. This shift in priority has led to a rapid (albeit unintentional) increase in the number of simultaneous liver-kidney transplantation (SLK) procedures. Specifically, since 2001, not only has the number of SLK procedures increased by more than 300%, but the proportion of SLK procedures with respect to the overall number of OLT procedures more than doubled from 2.38% in 2001 to 5.5% in 2006.8 Although significant renal impairment was previously considered a contraindication for OLT, SLK has become a well-established therapeutic option for ESRD and ESLD since SLK was first performed by Margreiter et al.9 in 1984. Few studies have analyzed liver graft survival after SLK versus liver transplantation alone (LTA); however, an increasing number of studies have reviewed renal graft survival, and they have produced mixed results.10-16 There is compelling evidence supporting the theory of an immunoprotective role assumed by the transplanted liver in preventing renal allograft rejection with SLK when the same donor is used, but there is no definitive evidence demonstrating improved recipient or graft survival. Despite the significant increase in the number of SLK procedures performed each year, there are currently no standardized criteria for organ allocation to SLK candidates. The current national organ shortage, the increasing demand for liver and kidney transplantation, and the increasing number of patients undergoing transplantation magnify the need for robust data on outcomes of SLK so that scarce resources can be appropriately allocated and posttransplant care can be optimized.
The principal aim of this study was to compare recipient and liver graft survival with LTA, SLK, kidney after liver transplantation (KALT), and liver after kidney transplantation (LAKT) according to the UNOS/Organ Procurement and Transplantation Network (OPTN) database. A secondary aim of this study was to identify independent risk factors affecting recipient and graft survival.
PATIENTS AND METHODS
The UNOS/OPTN database was queried for all first-time adult LTA, SLK, KALT, and LAKT procedures performed between January 1, 1988 and December 31, 2007 with at least 2 years of follow-up. Pediatric recipients and recipients of other organs were excluded from the analysis. SLK was defined as any combination of liver and kidney transplants in which the second transplant occurred within 2 days of the initial transplant; serial transplantation was defined as any combination of liver and kidney transplants in which the second transplant occurred more than 2 days after the initial transplant. Various demographic and clinical characteristics of the recipients and donors were identified: recipient age, sex, and ethnicity; hepatitis C virus (HCV) and diabetes mellitus (DM) status; pretransplant serum creatinine level; indication for liver transplantation; time on the waiting list; time between serial transplants; cold ischemia time (CIT); warm ischemia time (WIT); donor age and ethnicity; and graft type. Patients were identified from the liver files and then were matched to the kidney files to ensure accuracy.
Cox regression models were used to estimate recipient and graft survival rates so that Kaplan-Meier curves could be constructed for both recipient survival and graft survival. Survival was calculated from the time of liver transplantation. A time-dependent and covariate Cox proportional hazards regression model was used for the KALT arm to reflect the fact that the patients were in the LTA arm until they underwent kidney transplantation, at which time they moved from the LTA arm to the KALT arm. Multivariate Cox regression models were used to assess the associations of recipient mortality and graft loss with different demographic and clinical variables. All patient deaths were considered graft failures regardless of the graft status at the time of death. Therefore, graft failure was defined as a patient's death with the original graft or retransplantation (whichever came first). Cumulative incidence probabilities were calculated for both death (which was defined as the probability of death with the first liver transplant still functioning before time t) and retransplantation (which was defined as the probability of receiving a new liver before time t). In the setting of a large national database, P < 0.01 was considered statistically significant. Statistical analysis was performed with SAS 9.2. This protocol was reviewed and approved by the institutional review board of the Medical College of Wisconsin.
The demographic and clinical characteristics of the 4 transplantation groups are shown in Table 1. In all, 70,333 liver transplants were analyzed, and they included LTA (n = 66,026), SLK (n = 2327), KALT (n = 1738), and LAKT (n = 242). Nearly 50% of all liver transplants were performed between the ages of 50 and 64 years; however, the largest age group for KALT was 35 to 49 years (45.4%). Male recipients were more common in each of the transplant groups and received 62.8% of all liver transplants. With respect to ethnicity, Caucasian recipients were more common in each of the transplant groups and received 76.7% of all liver transplants. Although a plurality of all transplant recipients were HCV-negative (38.7%), 30.5% were HCV-positive, and the HCV status was not available for 30.8%. Similarly, a majority of all transplant recipients were DM-negative (63.7%), 15.2% were DM-positive, and the DM status was not available for 21.2%. HCV with cirrhosis was the most common indication for liver transplantation in each transplant group and was responsible for 24.5% of all liver transplants. Nearly all organs were obtained from deceased donors (96.7%). A large majority of the transplant recipients were on the liver transplant waiting list for less than 1 year; the proportions ranged from 81.0% (LAKT) to 83.4% (KALT). Similarly, many transplant recipients were on the kidney transplant waiting list for less than 1 year; the proportions ranged from 45.9% (LAKT) to 73.8% (SLK). The majority of the serial transplants were completed more than 5 years after the initial organ transplants: 56.3% for KALT and 53.3% for LAKT. A serum creatinine level <2.0 mg/dL at the time of transplantation was found in as many as 86.9% of recipients (LTA) and in as few as 13.1% of recipients (SLK). On the other hand, 74.9% of KALT recipients had a pretransplant serum creatinine level <2.0 mg/dL, whereas 69.0% of LAKT recipients had a pretransplant serum creatinine level ≥2.0 mg/dL.
|Characteristic||LTA (n = 66,026)||SLK (n = 2327)||KALT (n = 1738)||LAKT (n = 242)||Total (n = 70,333)|
|Age [n (%)]|
|18-34 years||5206 (7.9)||143 (6.2)||194 (11.2)||39 (16.1)||5582 (7.9)|
|35-49 years||22,698 (34.4)||629 (27.0)||789 (45.4)||95 (39.3)||24,211 (34.4)|
|50-64 years||32,447 (49.1)||1346 (57.8)||703 (40.4)||98 (40.5)||34,594 (49.2)|
|≥65 years||5675 (8.6)||209 (9.0)||52 (3.0)||10 (4.1)||5946 (8.5)|
|Sex [n (%)]|
|Male||41,349 (62.6)||1525 (65.5)||1146 (65.9)||156 (64.5)||44,176 (62.8)|
|Female||24,677 (37.4)||802 (34.5)||592 (34.1)||86 (35.5)||26,157 (37.2)|
|Ethnicity [n (%)]|
|Caucasian||50,826 (77.0)||1555 (66.8)||1367 (78.7)||176 (72.7)||53,924 (76.7)|
|Hispanic||7164 (10.9)||356 (15.3)||172 (9.9)||14 (5.8)||7706 (11.0)|
|Black||4816 (7.3)||288 (12.4)||136 (7.8)||36 (14.9)||5276 (7.5)|
|Other||3220 (4.9)||128 (5.5)||63 (3.6)||16 (6.6)||3427 (4.9)|
|HCV status [n (%)]|
|Positive||20,216 (30.6)||756 (32.5)||417 (24.0)||80 (33.1)||21,469 (30.5)|
|Negative||25,466 (38.6)||1123 (48.3)||510 (29.3)||112 (46.3)||27,211 (38.7)|
|Unknown||20,344 (30.8)||448 (19.3)||811 (46.7)||50 (20.7)||21,653 (30.8)|
|DM status [n (%)]|
|Positive||9550 (14.5)||753 (32.4)||300 (17.3)||55 (22.7)||10,658 (15.2)|
|Negative||42,594 (64.5)||1295 (55.7)||722 (41.5)||162 (66.9)||44,773 (63.7)|
|Unknown||13,882 (21.0)||279 (12.0)||716 (41.2)||25 (10.3)||14,902 (21.2)|
|HCV/DM status [n (%)]|
|HCV-positive/DM-positive||3334 (5.0)||298 (12.8)||110 (6.3)||21 (8.7)||3763 (5.4)|
|HCV-positive/DM-negative||16,104 (24.4)||440 (18.9)||278 (16.0)||55 (22.7)||16,877 (24.0)|
|HCV-negative/DM-positive||4683 (7.1)||382 (16.4)||138 (7.9)||22 (9.1)||5225 (7.4)|
|HCV-negative/DM-negative||19,679 (29.8)||703 (30.2)||331 (19.0)||85 (35.1)||20,798 (29.6)|
|Unknown||22,226 (33.7)||504 (21.7)||881 (50.7)||59 (24.4)||23,670 (33.7)|
|Indication for liver transplantation [n (%)]|
|HCV with cirrhosis||16,088 (24.4)||634 (27.3)||411 (23.7)||77 (31.8)||17,210 (24.5)|
|Alcoholic cirrhosis||9361 (14.2)||389 (16.7)||274 (15.8)||4 (1.7)||10,028 (14.3)|
|Primary sclerosing cholangitis||4476 (6.8)||47 (2.0)||125 (7.2)||1 (0.4)||4649 (6.6)*|
|Alcoholic cirrhosis with HCV||3993 (6.1)||113 (4.9)||90 (5.2)||1 (0.4)||4197 (6.0)|
|Hepatocellular carcinoma and cirrhosis||3208 (4.9)||70 (3.0)||28 (1.6)||6 (2.5)||3312 (4.7)|
|Primary biliary cirrhosis||4100 (6.2)||61 (2.6)||103 (5.9)||3 (1.2)||4267 (6.1)|
|Cryptogenic cirrhosis||3302 (5.0)||167 (7.2)||122 (7.0)||7 (2.9)||3598 (5.1)|
|Other||21,498 (32.6)||846 (36.4)||585 (33.7)||143 (59.1)||23,072 (32.8)|
|Time on the waiting list [n (%)]†|
|Liver transplant||53,550 (81.1)||1918 (82.4)||1449 (83.4)||196 (81.0)||57,112 (81.2)|
|Kidney transplant||0||1717 (73.8)||1014 (58.3)||111 (45.9)||2842 (4.0)|
|Liver transplant||10,082 (15.3)||320 (13.8)||245 (14.1)||38 (15.7)||10,682 (15.2)|
|Kidney transplant||0||142 (6.1)||481 (27.7)||78 (32.2)||701 (1.0)|
|Liver transplant||2394 (3.6)||91 (3.9)||46 (2.6)||11 (4.5)||2539 (3.6)|
|Kidney transplant||0||21 (0.9)||14 (0.8)||5 (2.1)||40 (0.1)|
|Missing||0||447 (19.2)||229 (13.2)||48 (19.8)||724 (1.0)|
|Time between serial transplants [n (%)]|
|<1 year||NA||NA||177 (10.2)||24 (9.9)||201 (0.3)|
|1-5 years||NA||NA||582 (33.5)||89 (36.8)||671 (1.0)|
|>5 years||NA||NA||979 (56.3)||129 (53.3)||1108 (1.6)|
|Donor type [n (%)]|
|Deceased donor||63,729 (96.5)||2319 (99.7)||1699 (97.8)||238 (98.3)||67,985 (96.7)|
|Living donor||2297 (3.5)||8 (0.3)||39 (2.2)||4 (1.7)||2348 (3.3)|
|Serum creatinine [n (%)]|
|<2.0 mg/dL||57,377 (86.9)||306 (13.1)||1301 (74.9)||71 (29.3)||59,055 (84.0)|
|≥2.0 mg/dL||7997 (12.1)||2003 (86.1)||418 (24.1)||167 (69.0)||10,585 (15.0)|
|Missing||652 (1.0)||18 (0.8)||19 (1.1)||4 (1.7)||693 (1.0)|
|Donor age [n (%)]|
|<60 years||57,959 (87.8)||2194 (94.3)||1575 (90.6)||223 (92.1)||61,951 (88.1)|
|≥60 years||8040 (12.2)||133 (5.7)||162 (9.3)||19 (7.9)||8354 (11.9)|
|Missing||27 (0.04)||0||1 (0.06)||0||28 (0.04)|
|Donor race [n (%)]|
|Caucasian||49,452 (74.9)||1611 (69.2)||1347 (77.5)||183 (75.6)||52,593 (74.8)|
|Black||7854 (11.9)||302 (13.0)||195 (11.2)||31 (12.8)||8382 (11.9)|
|Hispanic||6771 (10.3)||325 (14.0)||147 (8.5)||21 (8.7)||7264 (10.3)|
|Other||1949 (3.0)||89 (3.8)||49 (2.8)||7 (2.9)||2094 (3.0)|
|CIT [n (%)]|
|<6 hours||13,946 (21.1)||606 (26.0)||248 (14.3)||43 (17.8)||14,843 (21.1)|
|6-12 hours||36,225 (54.9)||1267 (54.4)||990 (57.0)||144 (59.5)||38,626 (54.9)|
|>12 hours||8657 (13.1)||190 (8.2)||346 (19.9)||21 (8.7)||9214 (13.1)|
|Missing||7198 (10.9)||264 (11.3)||154 (8.9)||34 (14.0)||7650 (10.9)|
|WIT [n (%)]|
|<30 minutes||5837 (8.8)||288 (12.4)||123 (7.1)||28 (11.6)||6276 (8.9)|
|30-60 minutes||33,186 (50.3)||1060 (45.6)||916 (52.7)||114 (47.1)||35,276 (50.2)|
|>60 minutes||10,734 (16.3)||221 (9.5)||447 (25.7)||31 (12.8)||11,433 (16.3)|
|Missing||16,269 (24.6)||758 (32.6)||252 (14.5)||69 (28.5)||17,348 (24.7)|
Donors less than 60 years old were commonly seen in all liver transplant groups, and they accounted for 88.1% of all liver transplants. The commonest donor ethnicity for all liver transplant groups was Caucasian, and these donors accounted for 74.8% of all transplants. A majority of the liver transplants in each group had a CIT of 6 to 12 hours (54.9%) and a WIT of 30 to 60 minutes (50.2%).
As shown in Fig. 1, there has been a steady increase in the number of SLK procedures since 1988, with a distinct deviation developing between SLK and KALT after the adoption of the MELD score in 2002. Accordingly, there was a 233% increase in the number of SLK procedures, and this coincided with a 74% decrease in the number of KALT procedures from 2001 to 2007.
After we controlled for potential confounding demographic and clinical variables in the multivariate model, independent predictors of overall recipient mortality were identified, and they are listed in Table 2. Compared to Caucasians, black had a 29% increased risk of mortality, whereas Hispanics had a 9% decreased risk of mortality. A longer time on the waiting list was more protective, with a 17% decrease in mortality for those on the waiting list longer than 3 years versus those on the list for less than 1 year. Female sex was also protective and was associated with a 5% decreased risk of mortality. The risk of mortality increased with age: recipients who were 65 years old or older had an 86% increased risk of mortality versus recipients who were 18 to 34 years old. Transplant recipients with a serum creatinine level ≥ 2.0 mg/dL at the time of transplantation had a 41% increased risk of mortality in comparison with recipients with a serum creatinine level < 2.0 mg/dL. The HCV/DM status was also found to be an independent predictor of recipient mortality. Specifically, in comparison with HCV-negative/DM-negative recipients, HCV-positive/DM-negative, HCV-negative/DM-positive, and HCV-positive/DM-positive recipients had 44%, 32%, and 72% increased risks of mortality, respectively. Donor variables such as an age greater than 60 years, a Hispanic donor, a CIT > 6 hours, and a WIT > 30 minutes were also found to be independent predictors of overall recipient mortality.
|Variable||HR||95% CI||P Value|
|Sex: female versus male||0.95||0.92-0.97||<0.001|
|Age (versus 18-34 years)|
|Race (versus Caucasian)|
|HCV/DM status (versus HCV-negative/DM-negative)|
|Time on the waiting list (versus <1 year)|
|Donor type: deceased versus living||1.03||0.95-1.12||0.43|
|Transplant group (versus LTA)|
|KALT (<3 months)||2.32||1.74-3.10||<0.001|
|KALT (3-6 months)||3.04||1.52-6.08||0.002|
|KALT (>6-12 months)||0.85||0.47-1.54||0.59|
|KALT (>12 months)||1.40||1.28-1.52||<0.001|
|Donor age: ≥60 versus <60 years||1.40||1.35-1.45||<0.001|
|Donor race (versus Caucasian)|
|Serum creatinine at transplant: ≥2.0 versus <2.0 mg/dL||1.41||1.36-1.46||<0.001|
|CIT (versus <6 hours)|
|WIT (versus <30 minutes)|
Although the difference approached statistical significance, there was no increased risk of overall recipient mortality with SLK versus LTA [hazard ratio (HR) = 0.92, P = 0.02; Table 2]. On the other hand, there was a 43% increased risk of recipient mortality with LAKT versus LTA (HR = 1.43, P = 0.001). Although there was no increased risk to recipient survival in the KALT (>6-12 months) subgroup versus the LTA group, there were considerable increases in recipient mortality in all other KALT subgroups in comparison with the LTA group.
Unadjusted Kaplan-Meier curves for overall recipient survival are shown in Fig. 2. The survival curve for the LTA group is significantly higher than the curves for all other transplant groups (P < 0.001). The survival curve for the SLK group is significantly higher than the curves for the both the KALT group (P < 0.001) and the LAKT group (P = 0.003). There was no difference between KALT and LAKT (P = 0.64). Similar trends were present after adjustments for potential confounding demographic and clinical variables, except that there was no statistically significant difference in overall recipient mortality between LTA and SLK (P = 0.02).
To further subdivide the relatively heterogeneous KALT group, we separated its members according to the timing of kidney transplantation with respect to the preceding liver transplant: <3 months (n = 72), 3 to 6 months (n = 12), >6 to 12 months (n = 30), and >12 months (n = 1624). Unadjusted Kaplan-Meier curves for overall recipient survival in these KALT subgroups and the SLK group are shown in Fig. 3. Just as the survival rate was significantly higher for SLK recipients versus KALT recipients overall (Fig. 2), the survival rate was significantly higher for SLK recipients versus the recipients in each KALT subgroup except for the KALT (>6-12 months) subgroup (P = 0.84). There were no statistical differences between the survival curves of the different KALT subgroups except for the KALT (<3 months) subgroup versus the KALT (>12 months) subgroup (P = 0.003). Similar trends were present after adjustments for potential confounding variables.
The overall recipient survival rates for each transplant group 1, 3, 5, and 10 years after transplantation are shown in Table 3. With the exception of LAKT recipients at 10 years (P = 0.01), LTA recipients had significantly higher overall survival rates throughout the 10-year posttransplant period in comparison with the SLK, LAKT, and KALT (overall) recipients (P ≤ 0.005). Specifically, the 1-, 3-, 5-, and 10-year survival rates for the LTA group were 84.4%, 75.4%, 68.4%, and 51.4%, respectively. However, the survival rates for the LTA group and the KALT subgroups showed that the rates for the LTA group were significantly higher only in comparison with the KALT (<3 months) subgroup 1 year after transplantation (P < 0.001) and the KALT (<3 months) and KALT (>12 months) subgroups 3, 5, and 10 years after transplantation (P < 0.001). Although the SLK survival rates were significantly lower, they were similar to the LTA rates: 81.8%, 71.7%, 64.1%, and 46.5% at 1, 3, 5, and 10 years, respectively. Similarly to the LTA survival rates, the SLK survival rates were significantly higher than the survival rates for only the KALT (<3 months) subgroup at 1, 3, 5, and 10 years (P < 0.001) and the KALT (>12 months) subgroup at 3, 5, and 10 years (P ≤ 0.003). On average, the LAKT survival rates were approximately 10% worse than the SLK rates (P < 0.001). The KALT (>12 months) subgroup, which accounted for 93.4% of all KALT cases, had 1-, 3-, 5-, and 10-year survival rates of 75.5%, 53.9%, 46.1%, and 30.5%, respectively, which were higher than the corresponding rates for the KALT (<3 months) subgroup: 36.5% (P < 0.001), 31.1% (P = 0.008), 28.1% (P = 0.02), and 24.6% (P = 0.38).
|Transplant Group||Overall Recipient Survival|
|1 Year||3 Years||5 Years||10 Years|
|Survival Estimate (%)||95% CI (%)||Survival Estimate (%)||95% CI (%)||Survival Estimate (%)||95% CI (%)||Survival Estimate (%)||95% CI (%)|
The HRs for graft loss (Table 4) were similar to the HRs for recipient mortality (Table 2) with respect to recipient variables (eg, sex, ethnicity, pretransplant serum creatinine level, time on the waiting list, and HCV/DM status) as well as donor variables (eg, age, CIT, and WIT). In contrast to the results for recipient mortality, recipient age was not predictive of increased graft loss until the age of 65 years. Although the donor type was not associated with an increased risk of recipient mortality, deceased donors were associated with a 10% reduction in the risk of graft loss in comparison with living donors. In comparison with Caucasian donors, black donors brought a 9% increased risk of graft loss, which was absent in the multivariate analysis of recipient mortality. Similar patterns of increased risk of graft loss were also seen in the LAKT group and the KALT subgroups in comparison with the LTA group. In agreement with the results for recipient mortality, the KALT (>12 months) subgroup had a 39% increased risk of graft loss in comparison with the LTA group. Furthermore, LAKT was associated with a 34% increased risk of graft loss in comparison with LTA. However, in contrast to the findings for recipient mortality, SLK was associated with a 15% decreased risk of graft loss in comparison with LTA (HR = 0.85, P < 0.001). Just as with recipient mortality, a donor age greater than or equal to 60 years, a Hispanic donor, a serum creatinine level ≥2.0 mg/dL, a CIT > 12 hours, and a WIT > 60 minutes were also found to be independent risk factors for graft loss.
|Variable||HR||95% CI||P Value|
|Sex: female versus male||0.96||0.93-0.98||<0.001|
|Age (versus 18-34 years)|
|Race (versus Caucasian)|
|HCV/DM status (versus HCV-negative/DM-negative)|
|Time on the waiting list (versus <1 year)|
|Donor type: deceased versus living||0.90||0.83-0.97||0.004|
|Transplant group (versus LTA)|
|KALT (< 3 months)||1.60||0.91-2.81||0.11|
|KALT (3-6 months)||2.97||0.75-11.77||0.12|
|KALT (>6-12 months)||0.81||0.37-1.81||0.61|
|KALT (>12 months)||1.39||1.25-1.54||<0.001|
|Donor age: ≥60 versus <60 years||1.51||1.46-1.56||<0.001|
|Donor race (versus Caucasian)|
|Serum creatinine at transplant: ≥2.0 versus <2.0 mg/dL||1.35||1.31-1.39||<0.001|
|CIT (versus <6 hours)|
|WIT (versus <30 minutes)|
Unadjusted Kaplan-Meier curves for overall graft survival are shown in Fig. 4. Similarly to overall recipient survival (Fig. 2), the overall graft survival curves for LTA and SLK are significantly higher than the curves for both KALT and LAKT (P < 0.001). Unlike the overall recipient survival curves, the overall graft survival curve for LAKT is significantly higher than the curve for KALT (P < 0.001). Although there is no difference between the survival curves for LTA and SLK in the unadjusted model (P = 0.27), after adjustments for potential confounding demographic and clinical variables, overall graft survival was found to be significantly higher in the SLK group versus the LTA group (P < 0.001).
Unadjusted Kaplan-Meier curves for the overall graft survival of the KALT subgroups and the SLK group are shown in Fig. 5. The SLK graft survival curve is significantly higher than the curves for the KALT (<3 months) subgroup (P < 0.001) and the KALT (>12 months) subgroup (P < 0.001). Similar trends were present after adjustments for potential confounding variables, except that the SLK curve was significantly higher than only the curve for the KALT (>12 months) subgroup (P < 0.001).
The graft survival rates for each transplant group 1, 3, 5, and 10 years after transplantation are shown in Table 5. Although the graft survival rates followed a trend similar to that for the recipient survival rates, there was no difference between LTA and SLK (P = 0.51 at 1 year, P = 0.33 at 3 years, P = 0.71 at 5 years, and P = 0.76 at 10 years). Similarly to recipient survival, graft survival in the SLK group was significantly higher than graft survival in both the KALT group (overall) and the LAKT group for each of the same time periods. The graft survival rate for the KALT (>12 months) subgroup (the largest KALT subgroup) was significantly higher than the graft survival rates for the LTA, SLK, and LAKT groups only 3 years after transplantation (P < 0.001). Otherwise, the graft survival rates for the KALT (>12 months) subgroup were significantly higher than only the graft survival rates for the LAKT group and the KALT (< 3 months) and KALT (3-6 months) subgroups at 1 year (P ≤ 0.006), for the LAKT group and the KALT (<3 months) subgroup at 5 years (P ≤ 0.009), and for the KALT (<3 months) and KALT (>6-12 months) subgroups at 10 years (P = 0.004). Although the KALT (3-6 months) subgroup had the lowest graft survival rates, the differences were rarely significant. The KALT (<3 months) subgroup, on the other hand, had the lowest significant graft survival rates: 38.8%, 24.8%, 17.8%, and 10.8% at 1, 3, 5, and 10 years, respectively.
|Transplant Group||Overall Graft Survival|
|1 Year||3 Years||5 Years||10 Years|
|Survival Estimate (%)||95% CI (%)||Survival Estimate (%)||95% CI (%)||Survival Estimate (%)||95% CI (%)||Survival Estimate (%)||95% CI (%)|
The cumulative probabilities of death with the original graft (before liver retransplantation) and liver retransplantation are listed in Table 6 and are displayed in Figs. 6 and 7, respectively. There was a slightly higher probability of death before retransplantation for SLK recipients versus LTA recipients (P < 0.001). On the other hand, LTA recipients had a significantly higher probability of death before retransplantation in comparison with KALT recipients overall (P < 0.001). LAKT recipients had a higher probability of death before retransplantation in comparison with both LTA (P < 0.001) and SLK recipients (P = 0.002). Although the incidence of death for SLK recipients was only slightly higher than the incidence for LTA recipients, the incidence of liver retransplantation was nearly 50% higher for LTA recipients (P < 0.001). Liver retransplantation for KALT recipients accounted for approximately 50% of graft losses, and this rate was statistically higher than only the rate for SLK recipients (P = 0.007).
|Death (%)||Retransplantation (%)||Graft Survival (%)||Death (%)||Retransplantation (%)||Graft Survival (%)|
We report here the long-term follow-up of 70,333 liver transplant recipients; to our knowledge, this is the largest retrospective review of recipient and graft survival with LTA, SLK, and serial liver-kidney transplantation.
The 1- and 5-year recipient and graft survival rates demonstrated in our study are similar to those presented in previous studies (see Table 7), including the largest previous study of SLK recipients by Simpson et al.12 (n = 1136). Although the recipient survival rates with SLK 1, 3, 5, and 10 years after transplantation in our study were on average 4% worse than the rates with LTA, there was no increased risk of overall recipient mortality with SLK versus LTA in our multivariate analysis model. Although there was no difference in the graft survival rates between LTA and SLK recipients 1, 3, 5, and 10 years after transplantation, the risk of graft loss was significantly decreased in SLK recipients versus LTA recipients.
|Study||Patients (n)||Recipient Survival (%)||Liver Graft Survival (%)|
|1 Year||5 Years||1 Year||5 Years|
|Jeyarajah et al.17 (1997)||29||78.6 (2 years)||48.1||62.7 (2 years)||41.2|
|Demirci et al.15 (2003)||38||73.7||64.1||68*||56*|
|Creput et al.10 (2003)||45||85||82 (2 years)||NA||NA|
|Fong et al.11 (2003)||800||74*||64||NA||NA|
|Simpson et al.12 (2006)||1136||80*||65*||NA||NA|
|Ruiz et al.18 (2006)||99||76||70||70||65|
|Locke et al.14 (2008)||1032||82||NA||79.6||NA|
|Schmitt et al.19 (2009)||959||79.4,† 81.0‡||69,*† 65*‡ (3 years)||NA||NA|
|Van Wagner et al.20 (2009)||38§||73.7||68.1||73.7||53.0|
Many previous studies have also shown a reduced incidence of renal allograft rejection after SLK versus KALT.10-12 In the largest such study by Simpson et al.,12 it was hypothesized that the liver allograft provides via immunogenetic identity a form of renal graft immunoprotection after SLK that is absent with KALT. In the same study, there was no difference in the recipient or renal graft survival rates between SLK and KALT. Although the recipient survival rates for SLK in their study closely matched those calculated in our study, we found that survival was significantly better after SLK versus KALT.
Although previous studies have evaluated kidney graft survival after SLK, few have reported liver graft survival. In those studies reviewing kidney graft survival, the main causes of mortality and subsequent graft failure have been severe infectious complications.10 Decreased survival after SLK versus LTA may be due to the inherent severity of illness in recipients with ESLD and ESRD (which is evident from the higher MELD scores at the time of transplantation) as well as the associated subsequent risks and complications immediately after transplantation.
Before SLK was first performed in 1984 by Margreiter et al.,9 significant renal impairment was considered a contraindication for OLT. Because of the allocation of organs according to the MELD score, which is heavily affected by renal disease, liver transplant candidates are now given priority for renal dysfunction. The rationale for SLK is easily justified in patients with ESLD in the setting of ESRD requiring chronic renal replacement therapy (RRT). However, the matter is more complicated in patients with ESLD who develop acute renal failure when the duration or potential reversibility of the renal failure is uncertain, as is often the case with hepatorenal syndrome (HRS) type II. HRS is generally not considered an indication for SLK because of the well-documented reversibility of renal failure after OLT.21, 22 It has also been shown that kidneys transplanted from donors with HRS result in good renal function.23 However, there is currently no consensus about the duration of HRS or RRT after which renal dysfunction is not reversible. This lack of consensus sets the stage for the debate between proceeding with SLK up front and proceeding with LTA with the potential need for sequential renal transplantation.
The advantages of SLK appear to outweigh the disadvantages. As previously mentioned, one main advantage of SLK is the enhanced outcome in comparison with outcomes after KALT and LAKT, which is likely due to the apparent immunoprotection afforded to the kidney by the transplanted liver. Moreover, a well-functioning kidney allows the optimal dosing of necessary immunosuppressants. Lastly, several studies have confirmed the presence of immune complex–mediated glomerulonephropathy in HCV-positive patients undergoing OLT even in the absence of clinical signs or symptoms.24-27 Particularly in recipients with HCV, the underlying renal disease, which would improve after SLK, is surely exacerbated after OLT.
The need for better standardization of organ allocation for SLK candidates is perhaps best illustrated by the large disparity in the number of SLK procedures performed in US transplant centers: SLK accounted for just more than 6% of all OLT procedures in 2007 according to UNOS/OPTN data, but the rate was as high as 45% at some transplant centers.28 Furthermore, in a separate review of UNOS/OPTN data, a renal diagnosis was missing or was not specified for nearly 40% of SLK recipients.29 Also, in a study of 1032 SLK recipients between 2002 and 2006, only 318 (30.8%) were on chronic hemodialysis (HD),14 and this suggests that nearly 70% of the kidneys given to SLK recipients were transplanted too early. These examples alone are clear evidence that the allocation of renal allografts to SLK candidates requires closer monitoring. A lack of standardized criteria for the allocation of renal grafts to SLK candidates has the potential to adversely affect patients awaiting kidney transplants in the MELD era. It follows that efforts should be made to optimize both organ utilization and outcomes in liver candidates with impaired renal function.
Two national, multidisciplinary consensus conferences were recently held to review current data and recommend to UNOS standard listing criteria for SLK.28, 30 These meetings were instrumental to a recent UNOS policy proposal that sets forth minimum criteria for candidates listed for SLK.31 In brief, the first part of the proposal (188.8.131.52) recommends SLK listing for liver candidates according to the following criteria: chronic kidney disease stage 4 or 5, acute renal failure with a GFR < 25 mL/minute for 6 consecutive weeks (regardless of the need for RRT), and metabolic diseases (ie, hyperoxaluria). The second part of the proposal (184.108.40.206) was established to clarify listing requirements for sequential KALT, which is recommended for LTA recipients who remain HD-dependent for 3 months after OLT.
Sequential Liver-Kidney Transplantation
The management of OLT recipients who develop post-OLT ESRD often includes KALT. Unfortunately, only a few relatively small studies have reviewed survival outcomes of serial liver-kidney transplantation, with numbers ranging from 9 to 352 recipients (Table 8); the largest of these studies reported post-KALT 1- and 5-year recipient survival rates of 91% and 64%, respectively.12 This is in contrast to the current study, which reports post-KALT 1- and 5-year recipient survival rates of 33.4% and 22.0%, respectively; these rates are considerably lower than previously reported data. However, we have included much larger numbers of patients (1738 KALT patients and 242 LAKT patients), so our data may represent a more accurate picture of serial liver-kidney transplantation outcomes.
|Study||Patients (n)||Patient Survival (%)||Liver Graft Survival (%)|
|KALT||LAKT||1 Year||5 Year||1 Year||5 Years|
|Molmenti et al.32 (2001)||17||NA||88||71 (3 years)||88||61 (3 years)|
|Gonwa et al.3 (2001)||16||NA||95*||62||NA||NA|
|Demirci et al.15 (2003)||5||4||80,† 50‡||80,† 25‡ (4 years)||NA||NA|
|Simpson et al.12 (2006)||352||NA||91§||64§||NA||NA|
|This study||1738||242||33.4,† 71.0‡||22.0,† 54.2‡||36.3,† 68.4‡||27.6,† 53.3‡|
Varying degrees of renal insufficiency are common both before and after liver transplantation. According to a well-cited retrospective review of the UNOS database by Ojo et al.33 in 2003, 27% of liver transplant recipients had abnormal renal function (GFR < 60 mL/minute) at the time of transplantation, with 2.2% requiring HD.33 A more recent review by Eason et al.28 in 2008 reported that 8.2% of all OLT patients required HD at the time of transplantation.28 Furthermore, studies have shown that 22% of OLT recipients develop CRF within 5 years of transplantation,34 and 2% to 10% of these cases progress to ESRD within 10 years after transplantation.3 For OLT recipients who develop ESRD, the 6-year survival rate after the onset of ESRD has been reported to be as high as 27% for those who are maintained on HD and 71.4% for those who subsequently undergo KALT.3 In a similar study by Paramesh et al.,35 among 186 OLT recipients who developed acute renal failure requiring HD, 43 (23%) developed ESRD after a mean of 5.6 years. Eight of these 43 recipients (18.6%) subsequently underwent KALT with a 10-year recipient survival rate of 71%, whereas the rate for those who developed ESRD and were maintained on HD was 21%. This is in clear contrast to the 10-year recipient survival rate of 14.7% for KALT recipients observed in our study.
We identified multiple demographic and clinical characteristics that negatively affect both recipient and graft survival: recipient age and ethnicity, HCV/DM status, pretransplant serum creatinine level, donor age and ethnicity, WIT, and CIT. An increased risk of recipient mortality was seen in recipients who were 65 years old or older. Although recipient age is not considered an absolute contraindication for OLT, there are conflicting outcome studies in the current literature about the impact of recipient age on liver transplantation. One such study by Herrero et al.36 concluded that recipients older than 60 years had a significantly lower survival rate than matched controls younger than 60 years. This is in contrast to much of the recent literature, which has failed to demonstrate an increased risk of recipient mortality or graft failure in elderly recipients undergoing OLT,37-39 except in high-risk elderly patients who had poor hepatic synthetic function, had elevated bilirubin levels, or were admitted to the hospital before transplantation; these patients had significantly lower survival rates than sicker younger patients or less ill older patients.39
Donor age, much like recipient age, has shown mixed results with respect to its impact on recipient and graft survival.40-43 Unfortunately, most studies have been retrospective chart reviews with relatively small sample sizes. Interestingly, a prospective study that did not analyze recipient or graft survival observed an increased number of hospitalizations and longer initial stays for recipients who were 60 years or older versus recipients who were less than 60 years old.44
Our study also shows that both black recipients and donors are associated with increased risks of recipient and graft mortality. This is in contrast to a recent study by Asrani et al.,45 who demonstrated that donor race (ie, black and Asian/Pacific Islanders) was not associated with graft survival after OLT. Although our study did not assess donor-recipient race mismatch, a significant interaction between donor race and recipient race was noted in a recent study by Layden et al.46 Specifically, black recipients with Caucasian donors had a 66% higher risk of recipient mortality than Caucasian recipients with Caucasian donors. However, neither the survival of Caucasian recipients with black donors nor the survival of black recipients with black donors was significantly different from the survival of Caucasian recipients with Caucasian donors.
Similarly to previous studies, our results demonstrate a strong negative impact of both HCV and DM on recipient and graft survival.19, 47-51 Our results also demonstrate a profound negative impact of serum creatinine on both recipient survival and graft survival that is consistent with previously reported data.29, 30, 52 Lastly, our findings concerning the impact of WIT and CIT on recipient and graft survival are comparable to previously published data. Specifically, a WIT longer than 45 minutes and a CIT longer than 12 hours have been associated with a significant increase in graft dysfunction.53 Furthermore, a recent large meta-analysis demonstrated significantly worse patient and graft survival with CITs longer than 10 and 12.5 hours, respectively.54
Our study suffers from the limitations inherent to a retrospective analysis of the UNOS database: the inability to validate the provided data and the inability to control the immunosuppressive therapies and all the undocumented other potential interventions used. In an effort to overcome these limitations, P < 0.01 was considered statistically significant to allow a stricter assessment of statistical significance with a large national database. On the basis of the available UNOS data, we were able to identify only the presence or absence of HCV and DM and not the length of either disease, their severity, or their respective overall diabetic controls (ie, glycosylated hemoglobin or insulin versus dietary controls). Furthermore, the HCV status and the DM status were not recorded by UNOS until April 1, 1994 (ie, some 6 years into our study); this in part explains the relatively large number of missing values in Table 1. In addition, we stratified the results according to the level of renal impairment with only serum creatinine levels; we did not use the need for HD or its duration. Despite our intentions to include HD in our analysis, too many missing data prevented an accurate analysis of the impact of HD on recipient and graft survival. Similarly, our study does not make a distinction between outcomes in the pre-MELD and MELD eras.
There is also discrepancy between the number of SLK procedures according to the UNOS database and the number of SLK procedures in our study. Although a large majority of the SLK patients listed by UNOS received a liver and a kidney on the same day, 255 patients received the organs 1 day apart, and 5 patients received the organs 2 days apart. No liver or kidney transplants were performed 2 days after the initial date of transplantation if the patients were listed by UNOS as undergoing SLK. Similarly, 204 of those patients not listed by UNOS as undergoing SLK received a kidney 1 day after liver transplantation, and 5 received a kidney 2 days after liver transplantation. These 209 patients were added to the SLK group instead of the KALT group. In addition, 52 of those patients not listed as undergoing SLK received a liver 1 day after kidney transplantation, and 1 patient received a liver 2 days after kidney transplantation. These 53 patients were also added to the SLK group rather than the LAKT group. For these reasons, we defined SLK as any combination of liver and kidney transplants in which the second transplant occurred within 2 days of the first transplant.
In reality, the true prevalence of renal insufficiency in patients with ESLD is unknown. Because most liver transplant candidates have decreased creatinine production, a poor nutritional status, low muscle mass, weight loss, and edema, the use of serum creatinine alone results in an overestimation of GFR.55 Although the 6-variable Modification of Diet in Renal Disease equation was found to be the most accurate equation for estimating GFR (versus the Cockcroft-Gault and Nankivell formulas), all the equations consistently overestimated kidney function in patients with a GFR < 40 mL/minute (as measured by I125-iothalamate).56 In addition, although GFR measurements are the current gold standard, they are both expensive and time-consuming, and this is why serum creatinine continues to be used in most studies examining kidney function. In fact, the diagnostic criteria for HRS according to the International Ascites Club include a serum creatinine level > 1.5 mg/dL.57 Therefore, the use of serum creatinine is an inherent limitation in most studies. If cost continues to limit the use of GFR measurements, more studies will be needed to evaluate the potential role of serum creatinine alternatives such as cystatin C. Cystatin C, a protein that is produced at a constant rate and that is freely filtered and is neither secreted nor reabsorbed, has been shown to be more accurate than serum creatinine in estimating renal function in patients with advanced liver disease because the serum levels are independent of sex, muscle mass, and age.58
In conclusion, this is the largest retrospective review of recipient and graft survival after LTA, SLK, and serial liver-kidney transplantation. We have provided a succinct review and compelling evidence supporting the use of SLK as an effective and appropriate therapy in patients with ESLD with proven irreversible or chronic ESRD. Although the recent increase in the number of SLK procedures performed each year has effectively decreased the number of potential donor kidneys available to patients with ESRD awaiting kidney transplantation, SLK in patients with ESLD and ESRD is justified because of the lower risk of graft loss with SLK versus LTA as well as the superior recipient and graft survival with SLK versus serial liver-kidney transplantation. With the apparent and perhaps inevitable trend of increases in SLK, prospective outcome studies are needed to further optimize organ allocation to liver transplant candidates with renal insufficiency. Efforts need to be routinely made to identify liver transplant candidates with irreversible renal insufficiency who would most benefit from SLK.
The authors formally thank Katarina Linden and Denise Tripp for their assistance in acquiring the necessary data from UNOS/OPTN for the completion of this study.
- 2A comparison of tacrolimus (FK 506) and cyclosporine for immunosuppression in liver transplantation. The U.S. Multicenter FK506 Liver Study Group. N Engl J Med 1994; 331: 1110-1115.
- 11Analysis of the United Network for Organ Sharing database comparing renal allografts and patient survival in combined liver-kidney transplantation with the contralateral allografts in kidney alone or kidney-pancreas transplantation. Transplantation 2003; 76: 348-353., , , , .
- 31United Network for Organ Sharing. http://optn.transplant. hrsa.gov/PublicComment/pubcommentPropSub_237.pdf. Accessed March 2012.
- 38Outcomes after liver transplant in patients aged 70 years or older compared with those younger than 60 years. Mayo Clin Proc 2009; 84: 973-978., , , , , , et al.