Pancreas After Living Donor Kidney Versus Simultaneous Pancreas-Kidney Transplant: An Analysis of the Organ Procurement Transplant Network/United Network of Organ Sharing Database

Abstract

Background. 

One of the alternative options to simultaneous pancreas-kidney transplantation (SPKT) for type I diabetics with renal failure is sequential transplant of a living donor kidney followed by a deceased donor pancreas transplant (pancreas after living donor kidney transplant [PALK]). We retrospectively compared the outcomes of SPKT versus PALK.

Methods. 

Adults (age 18–59 years) with type I diabetes who were waitlisted for kidney-pancreas and received a SPKT or PALK between 2000 and 2007 were studied. We compared patient, kidney graft, and pancreas graft survival. Multivariate analysis was performed, and the results were expressed as hazard ratios (HRs) of graft loss and death of PALK, with SPKT as a reference.

Results. 

Of 11,966 patients who received a kidney transplant, 807 received a PALK and 5580 received a SPKT. Median time to pancreas from kidney transplant was 336 (25%–75%: 185–602 days) days. Average hospital stay for SPKT recipients was 13.2±15 days, whereas for PALK recipients was 5.7±4 days and 9.5±8 days for kidney and pancreas transplants, respectively. After controlling for confounding factors, patients receiving PALK had better patient survival (HR 0.52; 95% confidence interval [CI] 0.39–0.70) and kidney survival (HR 0.48; 95% CI 0.39–0.60) but worse pancreas survival (HR 1.37; 95% CI 1.16–1.62) compared with SPKT.

Conclusion. 

Among those who were waitlisted for a kidney-pancreas transplant, 53% received a kidney-pancreas transplant. Of those who received a kidney-pancreas transplant, 87% patients underwent SPKT and 13% underwent PALK. PALK was associated with better kidney graft and patient survival compared with SPKT. We found an inferior pancreas graft survival and longer total transplant hospitalization in PALK.

More than 20,000 patients with type I diabetes mellitus reached end-stage renal disease (ESRD) between 2002 and 2006 according to the 2008 United States Renal Data System Annual Report (¹). Renal replacement options include dialysis, deceased donor kidney transplant (DDKT), living donor kidney transplant (LDKT), or simultaneous pancreas-kidney transplant (SPKT). Kidney transplantation has been shown to increase long-term survival compared with maintenance dialysis (²) and has become the treatment of choice. Given the increased mortality observed in patients on dialysis, LDKT is the preferred type of transplant as it offers preemptive transplantations, minimizes excess dialysis times, and has better renal allograft survival compared with DDKT (³).

SPKT is associated with better kidney allograft and patient survival compared to DDKT (⁴). It also has the advantage of eliminating the need for exogenous insulin and preventing the progression of secondary complications. In particular, a functional pancreas transplant has been shown to reverse lesions of diabetic nephropathy (⁵), improve diabetic polyneuropathy (⁶), stabilize diabetic retinopathy (⁷), and halt progression of macrovascular disease (⁸). The pancreas graft survival rate has previously been reported significantly higher in SPKT compared with pancreas transplant alone (⁹). Hence, SPKT has the potential to improve the quality of life and significantly reduce long-term morbidity and mortality.

The kidney graft and patient survivals between SPKT and LDKT have been reported previously as comparable (¹⁰). A recent Organ Procurement Transplant Network (OPTN)/United Network of Organ Sharing (UNOS) registry analysis by Young et al. (³) on patients with type I diabetics who received a primary kidney transplant from 2000 to 2007 revealed that LDKT was associated with better kidney graft (hazard ratio [HR] 0.71, 95% confidence interval [CI] 0.61–0.83) and patient survival (HR 0.78, 95% CI 0.65–0.94) compared with SPKT after adjusting for older age and comorbid illnesses more common in the LDKT cohort.

Based on the results of this study, an attractive transplant option is the sequential transplantation of a living donor kidney followed by a deceased donor pancreas transplant (pancreas after living donor kidney transplant [PALK]). It has the advantages of a living donor kidney and a functional pancreas and expanding the number of deceased donor kidneys that would have been used in a SPKT. The disadvantages include the risks associated with a second surgical procedure and additional antigenic challenge.

The purpose of this study was to compare the outcomes and utility of PALK to SPKT in type I diabetics with ESRD. We performed a retrospective analysis using the OPTN/UNOS of type I diabetic kidney transplant recipients initially waitlisted for SPKT and receiving SPKT or PALK to examine short- and long-term kidney graft, pancreas graft, and patient survival in the current era of transplantation.

METHODS

By using the OPTN/UNOS database, we selected patients with type I diabetes according to diagnosis codes, aged 18 to 59 years, who were waitlisted for kidney-pancreas and received a primary kidney transplant between January 2000 and December 2007 with follow-up data available through August 2008. Dual organ transplants other than kidney-pancreas transplants were excluded. Of the 11,966 patients in the study population, 5580 (47%), 3461 (29%), and 2925 (24%) patients received a SPKT, LDKT, and DDKT, respectively. Of the LDKT, 807 patients (23%) received a subsequent PALK and 2654 patients (77%) received a LDKT alone during the study period (Fig. 1). PALK was compared with SPKT, whereas LDKT alone and DDKT were excluded.

FIGURE 1.: Kidney graft survival. (A) Overall kidney graft survival. Log-rank test: PALK versus SPKT, P less than 0.01; SPKT versus LDKT, P=0.03. (B) Death censored kidney graft survival. Log-rank test: PALK versus SPKT, P less than 0.01; SPKT versus LDKT, P less than 0.01. LDKT includes living kidney donor alone and PALK. PALK, pancreas after living kidney transplant; SPKT, simultaneous pancreas kidney transplant; LDKT, living donor kidney transplant.

Donor, recipient, and transplant characteristics were presented as means±SD or frequencies. Significant times were presented as medians. The Wilcoxon rank-sum test was used to test for significant differences in continuous variables. The chi-square test was used to compare categorical variables. Univariate comparisons of graft and patient survival were performed using the Kaplan-Meier product limit method, with significance tested using the log-rank test. Overall kidney graft survival was determined from the time of kidney transplantation until death, kidney retransplantation, or return to dialysis. Overall pancreas graft survival was determined from the time of pancreas transplantation until death, pancreas retransplantation, or return to exogenous insulin use. Patients were censored at the end of the study. Patient survival was determined from time of kidney transplantation until death or end of follow-up.

Covariates examined on univariate analyses were transplant type along with kidney donor (age, gender, race, hypertension, and body mass index), pancreas donor (age, gender, race, cause of death, and hypertension), and recipient (age, gender, race, hypertension, cardiovascular disease, cerebrovascular disease, peripheral vascular disease, dialysis time, peak panel reactive antibody, human leukocyte antigen (HLA) mismatch to kidney and pancreas allograft, type of exocrine pancreas drainage, and year of kidney and pancreas transplant) characteristics. UNOS does not provide center identification. To address center differences, we created a variable to separate the US states according to the number of transplant performed in each one. The state was considered with a high number of transplants, if numbers of PALK were more than 50 and SPKT were more than 200 during 2000 to 2007. Using this variable, PALK and SPKT were identified to be performed in a state of a high volume of transplant and possibly reflecting the most experienced transplant centers. All covariates with a significance level less than or equal to 0.1 in the univariate analysis were initially entered into multivariate analyses. The model was then optimized by keeping only those covariates with a P less than or equal to 0.05. Variables related to the transplant groups were kept or forced in the model. Multivariate estimates of hazards of kidney graft loss, pancreas graft loss, and patient mortality were calculated using Cox proportional hazards regression model. All analyses were conducted using STATA Statistical Software.

RESULTS

Baseline Characteristics and Waiting Times

The characteristics of the recipients and donors are listed in Table 1. Donor kidneys for PALK were older living donors compared with younger deceased donors of SPKT. SPKT had a lower proportion of female donors and a higher proportion of African American, hypertensive, and obese donors. Donor pancreas factors were similar between groups except hypertension, which was more likely in PALK. PALK recipients were younger at the time of kidney transplantation, had a higher proportion with coronary artery disease, and more likely to be female. SPKT recipients had a higher proportion of African Americans. PALK transplants had fewer HLA mismatch, shorter cold ischemia time, and less induction antibody therapy compared with SPKT. Pancreas exocrine bladder drainage was more common in PALK, and HLA mismatch to the pancreas allograft was increased in SPKT. Thirty-seven and 61% of the PALK and SPKT transplants, respectively, were performed in the states considered of a high transplant volume. Median waiting time from listing for kidney-pancreas to kidney transplantation was 143 (25%–75%: 62–276 days) days and to pancreas transplantation was 535 (25%–75%: 314–826 days) days for PALK transplant compared with median waiting time for SPKT of 269 (25%–75%: 103–512 days) days. Median time from kidney to pancreas transplant for PALK was 336 (25%–75%: 185–602 days) days.

TABLE 1: Characteristics of donors, recipients, and grafts

TABLE 1: Continued

Kidney Transplant Outcomes

Early outcomes of kidney transplantation are listed in Table 2. PALK had shorter initial hospitalizations for the kidney transplant and higher percentage of immediate graft function with less rejection up to 1 year compared with SPKT despite comparable serum creatinine at discharge and 1 year. Overall kidney graft survival rates (Fig. 1A) were significantly higher for PALK compared with SPKT (P<0.01) up to 60 months. This difference persisted after censoring for death (P<0.01) (Fig. 1B). Multivariate regression analysis of potential donor or recipient risk factors suggest that PALK was independently associated with better kidney graft survival with 52% risk reduction compared with SPKT (HR 0.48; 95% CI 0.39–0.60; Table 3).

TABLE 2: Early outcomes of kidney and pancreas transplantation

TABLE 3: Multivariate Cox regression analysis for kidney and pancreas graft survival

Pancreas Transplant Outcomes

Early outcomes of pancreas transplantation are listed in Table 2. PALK had shorter initial hospitalization for pancreas transplant (9.49±8.77) compared with SPKT (13.17±15.27), but when added to length of hospitalization for kidney transplantation (5.70±4.01), total mean hospital stay was 15 days. PALK transplants also had significantly higher rejection rates at 1 year and had higher percentages of thrombosis, infections, and rejections as causes of pancreas graft failure. Overall pancreas graft survival rates (Fig. 2A) were significantly lower for PALK compared with SPKT (P<0.01) up to 60 months. This difference persisted after censoring for death (P<0.01; Fig. 2B). Multivariate regression analysis suggests that PALK was independently associated with inferior pancreas graft survival with 37% increased risk compared with SPKT (HR 1.37; 95% CI 1.16–1.62; Table 3). There was no association between type of pancreas drainage and HLA mismatch for the pancreas with the pancreas failure.

FIGURE 2.: Pancreas graft survival. (A) Overall pancreas graft survival. (B) Death censored pancreas graft survival. PALK, pancreas after living kidney transplant; SPKT, simultaneous pancreas kidney transplant.

Patient Survival

PALK transplants were associated with better patient survival compared with SPKT (Fig. 3A), which was confirmed with multivariate regression analysis (HR 0.52, 95% CI 0.39–0.70; Table 3). One-year survival for PALK and SPKT was 99.24% and 95.55%, respectively. Comparing patient death rate among those with a functioning renal graft at 1 year after kidney transplantation revealed no significant difference between PALK and SPKT (Fig. 3B). This was confirmed with multivariate analysis (HR 0.90, 95% CI 0.64–1.26) (data not shown). Similarly, there was no significant difference in death rate between PALK and SPKT from the time of pancreas transplantation (Fig. 3C). This was confirmed with multivariate analysis (HR 0.82, 95% CI 0.55–1.23) (data not shown).

FIGURE 3.: Patient survival. (A) Patient survival from time of kidney transplant. PALK versus SPKT, P less than 0.01; SPKT versus LDKT, P=0.61. (B) Patient survival among those with a functioning renal graft at 1 year. PALK versus SPKT, P=0.23; SPKT versus LDKT, P less than 0.01. (C) Patient survival from time of pancreas transplant, P=0.29. LDKT includes living donor kidney alone and PALK. PALK, pancreas after living kidney transplant; SPKT, simultaneous pancreas kidney transplant; LDKT, living donor kidney transplant.

DISCUSSION

Patients with type I diabetics with ESRD are placed on the waitlist for kidney-pancreas transplant to eliminate the need for both dialysis and exogenous insulin. During the years 2000 to 2007 in the United States, 11,966 patients with type I diabetics waitlisted for a kidney-pancreas received a primary kidney transplant. Among these, approximately half (53%) received a kidney-pancreas transplant, 46% underwent simultaneous pancreas-kidney transplant (SPKT), and 7% received a living donor kidney followed by a PALK. At 5 years after kidney transplant, patient survival for PALK and SPKT was 91% and 87%, respectively. Overall kidney graft survival for PALK and SPKT was 86% and 77% and overall pancreas graft survival was 55% and 72%, respectively. After adjusting for potential confounding risk factors, PALK transplants were associated with a 52% reduction in kidney graft loss, a 37% increase in pancreas graft loss, and a 48% reduction in death compared with SPKT in relative risk.

Kidney Transplant Outcomes

Previous studies show better kidney graft outcomes in the LDKT compared with SPKT in the early posttransplant follow-up (^{3, 10, 11}). A subgroup of LDKT subsequently received a deceased donor pancreas (PALK). Although a second antigenic challenge and enhanced immunosuppression associated with PALK may adversely affect renal graft survival, we found overall superior short- and long-term renal allograft outcomes in the PALK compared with the SPKT recipients up to 5 years follow-up. Incidence of delayed graft function and rejection at 1 year was lower in PALK transplant in part because of shorter cold ischemia time, better organ quality, and less HLA mismatch. It is likely that LDKT recipients with poor outcomes after kidney transplant, specifically those with rejection, were excluded. This bias may contribute to the better renal allograft outcomes in the PALK recipients.

Pancreas Transplant Outcomes

PALK was associated with inferior pancreas graft outcomes compared with SPKT. Our results are in agreement with the pancreas' graft survival rates reported previously (^{9, 12, 13}). At 1 year posttransplantation, pancreas graft survival was 86% and 80% for SPKT and PALK recipients, respectively, which is comparable with the 85% and 78% reported by Gruessner et al. for recipients transplanted from 2000 to 2004. PALK transplants had a higher rate of short-term graft loss, in particular from thrombosis, which may partly be explained by the reversal of uremia. Frequency of rejection at 1 year, hyperacute rejection, acute rejection, and chronic rejection as a cause of graft loss was also higher in PALK transplants. This may be explained by the prior sensitization from kidney transplant encountered in PALK. In addition, SPKT can use kidney graft function and biopsy pathology to monitor for rejection as both organs are from a single donor. This surveillance strategy is not an option in PALK recipients and may contribute to the inferior graft outcomes (¹⁴).

Patient Survival

We previously showed that LDKT have comparable overall patient survival to SPKT (³). We found that PALK recipients had better overall patient survival compared with SPKT recipients from the time of kidney transplant. The major difference was seen early posttransplant in SPKT, with a 2.3% difference at 3 months and 3.7% difference at 12 months. This can in part be explained by PALK recipients must survive until the time of pancreas transplant. The median time to pancreas transplant was 336 (25%–75%: 185–602 days) days from kidney transplant. This introduced bias toward better patient survival in PALK transplant. Furthermore, the complicated nature of the SPKT surgical procedure and its higher rate of early posttransplant complications including myocardial infarction (¹⁵), particularly with the older age of SPKT recipients, likely contributed to early mortality. Therefore, we examined the death rate for living donor kidney alone recipients between 6 and 12 months postkidney transplant, which was 0.5% per 3 months, comparable with the PALK death rate of 0.3% per 3 months. Among those with a functioning kidney graft at 12 months, we found no significant difference in death rate between PALK and SPKT in the subsequent 4 years. Patient survival was also similar in PALK and SPKT when the pancreas transplant was considered the index date (Fig. 3C). These two different analyses diminished the bias that favor a selection of recipients with better characteristics in PALK population.

Limitations

As with all retrospective, observational studies, our results should be interpreted with care. The objective of this study was to compare the outcomes between PALK and SPKT. We limited our cohort to PALK recipients, which only represented 23% of the LDKT recipients initially waitlisted for pancreas transplant. PALK recipients must survive long enough to receive a subsequent pancreas transplant, which introduced a favorable selection bias. LDKT recipients who develop early postkidney transplant complications likely did not receive a subsequent pancreas transplant. We could not examine the patients who successfully underwent LDKT but later did not wish to pursue a pancreas transplant or were removed for medical reasons. Therefore, the results of this study should not be generalized to all recipients of LDKT awaiting subsequent pancreas transplantation. In addition, we considered only patients transplanted after 2000 to reflect outcomes using current practices. Our study did not examine the effects of different immunosuppressive protocols on graft survival. We also did not account for the use of induction in the subsequent pancreas transplant. Because UNOS does not provide center identification, we tried to adjust results for center differences using the number of transplants performed in each state. It may or may not reflect differences of high volume transplant centers from low volume centers. Finally, although our multivariate analysis adjusted for known differences in donor and recipient characteristics, there may be other factors not considered in this study that are also important in assessing solid organ transplant outcome and patient survival.

CONCLUSIONS

Living donor kidney transplantation has the potential to expand the number of kidney transplants and minimize excess wait times. We previously reported that kidney graft and patient survival was superior in those who underwent LDKT when compared with simultaneous pancreas-kidney transplant. In this study, we found that PALK was associated with better renal allograft, shorter dialysis time, and more preemptive transplants compared with SPKT. Because of this, we recommend, in general, that patients with type I diabetics with a potential living donor should consider living donor kidney followed by pancreas transplant. However, it is also important to note that of patients with the type I diabetes who were waitlisted for kidney-pancreas and received a kidney transplant between 2000 and 2007, only 7% received a living donor kidney followed by a PALK. Moreover, PALK was associated with inferior pancreas graft outcomes compared with to SPKT. Despite worse pancreas survival, no significant difference in mortality rates was noted whether the pancreas transplant occurred with or after kidney transplant. Thus, it is reasonable to recommend an SPKT over a PALK in a selective group of brittle type I diabetics who require or wish to have a functional pancreas, especially if anticipated local wait time for SPKT is short.