Impact of Acute Rejection Episodes on Long-Term Graft Survival Following Simultaneous Kidney-Pancreas Transplantation
Abstract
Although it is well established that acute rejection is one of the major risk factors for chronic graft loss following kidney transplantation, its effect on long-term graft survival following simultaneous kidney-pancreas transplants (SKPTs) is less well known. We analyzed a large cohort of SKPTs and cadaver kidney transplants reported to the United Network for Organ Sharing database during 1988–97, to determine the impact of acute rejection episodes on long-term kidney and pancreas graft survival. Only patients whose kidney and pancreas grafts had survived for at least 1 year were included. Other potential risk factors influencing long-term graft survival were included in the analysis. Of the 4251 SKPTs, 45% had no acute rejection, 36% had kidney only rejection, 3% had pancreas only rejection, and 16% had both kidney and pancreas rejection within the 1st year post transplant. The 5-year kidney and pancreas graft survival rates adjusted for other risk factors were 91% and 85%, respectively; for those with no acute rejection episodes, 88% and 84%, respectively; for those with kidney only rejection, 94% and 83%, respectively; for those with pancreas only rejection; and 86% and 78%, respectively, for those with both kidney and pancreas rejection. The relative risk (RR) of kidney graft failure was 1.32 when acute rejection involved the kidney graft only, while the RR was 1.53 when the rejection involved both organs. We conclude that acute rejection episodes have a negative impact on the long-term kidney graft survival in the SKPT population similar to that in the cadaver kidney transplant population. Patients who had acute rejection episodes of both kidney and pancreas have the worst long-term graft survival.
Introduction
Simultaneous kidney-pancreas transplantation (SKPT) is a well-accepted treatment modality for selected patients with type 1 diabetes mellitus and renal failure. Successful SKPT improves the quality of life, stabilizes or reverses the secondary complications of diabetes and improves the long-term patient survival (1–8). The graft survival following SKPT has improved significantly over the past decade, primarily due to advances in immunosuppression. In SKPTs, 1-year pancreas graft survival improved from 72% during 1987-90 to 82% during 1998–2000, while 1-year kidney graft survival improved from 84% to 92% during the same time period (9). The 1-year immunological graft loss rate decreased from > 8% during 1987–90 to < 2% during 1998–2000 (9). Despite these improvements and potent immunosuppression, 15–30% of patients develop at least one episode of acute rejection following SKPT (10,11). Acute rejection has been demonstrated to be one of the important risk factors for the development of chronic rejection, impairing the long-term graft survival following kidney transplantation (12–15). The half-lives of cadaver and living donor kidney transplants with early rejection episodes are 7.9 and 11.3 years, respectively, compared to 9.8 and 16.2 years, respectively, for those without early rejection episodes (16). However, the impact of acute rejection on the long-term graft survival following SKPT has not been well studied.
We analyzed a large cohort of SKPT and cadaver kidney transplants reported to the United Network for Organ Sharing (UNOS) database to determine: (i) the impact of acute rejection episodes on long-term kidney and pancreas graft survival; (ii) the impact of acute rejection episodes of the kidney graft on long-term pancreas graft survival; (iii) the impact of acute rejection episodes of the pancreas on long-term kidney graft survival; and (iv) the impact of acute rejection episodes on long-term kidney graft survival in SKPT vs. cadaveric kidney alone transplant recipients.
Materials and Methods
The study population comprised all cadaveric kidney alone and SKPTs reported to the UNOS database during 1988–97 who had a diagnosis of type 1 diabetes mellitus and whose grafts (both kidney and pancreas in SKPT recipients) have survived at least 1 year post transplant. All living donor transplants, multiorgan transplants other than SKPT, and transplants in recipients less than 18 year of age were excluded from the analysis. There were 4251 SKPTs and 8609 cadaveric kidney alone transplants. Data about the occurrence of acute rejection episodes, defined according to the conventional criteria at each reporting institution, were obtained from the UNOS kidney transplant recipient follow-up form and/or the pancreas transplant recipient follow-up form. Follow-up was censored for patients who died with functioning grafts. Demographic data for the study groups are shown in Table 1.
| Kidney only transplants | SKP transplants | ||
|---|---|---|---|
| n=8609 | n=4251 | p-value | |
| Recipient gender – male | 5258 (61%) | 2503 (59%) | 0.02 |
| Recipient race/ethnicity: | |||
| White | 6146 (71%) | 3737 (88%) | <0.01 |
| Black | 1442 (17%) | 306 (7%) | <0.01 |
| Hispanic | 738 (8%) | 171 (4%) | <0.01 |
| Asian | 131 (1%) | 28 (1%) | <0.01 |
| Other | 147 (2%) | 8 (<1%) | <0.01 |
| Not reported | 5 (<1%) | 1 (<1%) | n.s. |
| Dialysis at time of transplant | 7786 (90%) | 3187 (75%) | <0.01 |
| Previous pancreas transplant | 110 (1%) | 21 (0.5%) | <0.01 |
| Previous kidney transplant | 447 (5%) | 99 (2%) | <0.01 |
| Previous kidney or pancreas transplant | 456 (5%) | 99 (2%) | <0.01 |
| Delayed graft function | 2022 (23%) | 1200 (28%) | <0.01 |
| Donor gender – male | 5359 (62%) | 2826 (66%) | <0.01 |
| Donor race/ethnicity: | |||
| White | 7260 (84%) | 3482 (82%) | <0.01 |
| Black | 638 (7%) | 429 (10%) | <0.01 |
| Hispanic | 592 (7%) | 275 (6%) | n.s. |
| Asian | 67 (1%) | 40 (1%) | n.s. |
| Other | 50 (1%) | 24 (1%) | n.s. |
| Not reported | 2 (<1%) | 1 (<1%) | n.s. |
| Donor history of hypertension | 316 (4%) | 102 (2%) | <0.01 |
| Mean donor terminal creatinine value (SEM)* | 1.19 (0.04) | 1.18 (0.04) | n.s. |
| Mean donor age (year) (SEM) | 31 (0.17) | 27 (0.17) | <0.01 |
| Mean recipient age (year) (SEM) | 43 (0.11) | 37 (0.11) | <0.01 |
| Mean cold ischemia time-kidney (h)** | 23 (0.11) | 14 (0.11) | <0.01 |
| Cold ischemia time-pancreas (h)*** | NA | 13 (0.09) | – |
- * Data missing from the kidney only group = 6347 (73%) and SKPT group = 1919 (45%).
- ** Data missing from the kidney only group = 188 (2%) and SKPT group = 208 (5%).
- *** Data missing 205 (5%). n.s.=not significant. NA=not applicable.
The following risk factors which could potentially influence the long-term graft survival were included in the analysis: donor age, donor gender, donor race/ethnicity, donor cause of death, donor history of hypertension, donor creatinine, recipient age, recipient gender, recipient race/ethnicity, recipient requirement of dialysis prior to transplant, previous pancreas, kidney or kidney-pancreas transplant, cold ischemia time of kidney and pancreas, HLA mismatch, delayed graft function of kidney, pretransplant transfusions, and administration of antilymphocyte induction therapy.
A Cox proportional hazard model was used to determine the effect of graft rejection on graft survival. The model was adjusted for the effect of other variables. Projected graft half-lives (t½) were estimated using PROC LIFEREG (SAS, v.8.02) with the assumption that survival times beyond 5 years were distributed exponentially. All tests of significance were two-sided and alpha of 0.05 was considered to indicate statistical significance. All statistical analyses were performed with SAS version 8.02 for Windows NT (Cary, NC, 1995). The data element was set to the group with the highest frequency if information was missing for categorical data while it was set the mean value by organ transplant type if the information missing was for continuous data.
Results
Incidence of rejection
Of the 8609 cadaveric kidney transplant alone recipients, 3000 (35%) had at least one rejection episode during the first-year post-transplant. Of the 4251 SKPT recipients, 1523 (36%) had at least one rejection episode of the kidney within the first year post transplant, 127 (3%) had at least one rejection episode of the pancreas, while 669 (16%) had rejection episodes involving both kidney and pancreas grafts. Forty-five percent of the SKPT group and 65% of the cadaveric kidney transplant alone group had no rejection episodes within the first year post transplant. In the SKPT group, the incidence of kidney rejection decreased from 47% in the 1988–92 era to 31% in the 1993–97 era. The incidence of combined kidney and pancreas rejection decreased from 18% to 15% during the same period, while the incidence of pancreas alone rejection remained similar at 3% in the two eras.
Risk factors for graft failure
Significant risk factors for kidney graft failure for the whole cohort of transplants (both cadaveric kidney and SKPTs) are shown in Table 2. As indicated in the table, there was a 61% increase in the probability of graft failure in cadaveric kidney transplant alone group, compared to 42% increase in the probability of graft failure in SKPT group in which the graft has survived at least 1 year and had at least one rejection episode within the first year post transplant. This difference in the probability of graft failure was not significant by contrast analysis (p = 0.068).
| Relative risk | 95% CI* | p-value | |
|---|---|---|---|
| SKPT with kidney rejection episode within 1year** | 1.423 | 1.36–1.90 | <0.01 |
| Cadaver kidney transplant alone with rejection episode within 1year** | 1.610 | 1.20–1.68 | <0.01 |
| Donor age (linear) | 1.014 | 1.01–1.02 | <0.01 |
| Donor African American | 1.341 | 1.16–1.54 | <0.01 |
| Recipient age (linear) | 0.980 | 0.97-.098 | <0.01 |
| Recipient Hispanic | 1.238 | 1.05–1.45 | <0.01 |
| Recipient African American | 1.882 | 1.69–2.10 | <0.01 |
| Cold ischemia time- pancreas | 0.978 | 0.96–0.99 | <0.01 |
| Delayed graft function | 1.193 | 1.07–1.32 | <0.01 |
| HLA mismatch 6 | 1.191 | 1.03–1.38 | 0.018 |
| Pretransplant transfusion | 1.094 | 1.00–1.19 | 0.037 |
- * CI: confidence interval.
- ** **Reference group: SKPTs with no kidney rejection during the first year post transplant.
Table 3 shows the relative risk of graft failure in the SKPT group according to occurrence of acute rejection of the kidney, pancreas or both allografts. As shown in the table, rejection episodes involving both the organs were associated with the highest risk for both kidney and pancreas graft failure. The findings were similar when the relative risk was calculated with death treated as graft failure (data not shown).
| Relative risk | 95% CI* | p-value | |
|---|---|---|---|
| Kidney graft failure | |||
| Kidney only rejection | 1.32 | 1.09–1.58 | <0.01 |
| Pancreas only rejection | n.s. | n.s. | n.s. |
| Both kidney and pancreas rejection | 1.53 | 1.22–1.91 | <0.01 |
| No rejection of either organ (reference group) | 1.00 | – | – |
| Pancreas graft failure | |||
| Kidney only rejection | n.s. | n.s. | n.s. |
| Pancreas only rejection | n.s. | n.s. | n.s. |
| Both kidney and pancreas rejection | 1.54 | 1.25–1.88 | <0.01 |
| No rejection of either organ (reference group) | 1.00 | – | – |
- * CI: confidence interval. n.s.=not significant.
Graft survival and graft half-life
Table 4 shows the adjusted death censored 5-year kidney graft survival and graft half-life for the various groups in the SKPT population. Half-life of the kidney graft was significantly reduced when the patients had acute rejection of the kidney or both the kidney and pancreas. Occurrence of acute rejection in the pancreas graft did not affect the kidney graft survival or half-life. The findings were similar when the graft survival and half-life were calculated treating death as graft failure (Table 5). Table 6 shows the kidney graft half-life according to the era of transplant. Kidney graft half-lives have improved in the no rejection, kidney rejection, and combined kidney and pancreas rejection groups in the 1993–97 era compared to those in the 1988–92 era. The kidney graft survival in the pancreas alone rejection group is similar between the two eras, although the sample size is quite small. As shown in Table 6, acute rejection has greater impact in the later era; with kidney alone rejection, graft half-life decreased from 24.5 years to 18.7 years in the 1988–92 era, while it decreased from 44.4 years to 25.8 years in the 1993–97 era.
| 5year kidney | Kidney graft years | |||
|---|---|---|---|---|
| n | graft survival (CI*) | half-life in (CI*) | p-value | |
| No rejection | 1906 | 91% (90–93) | 33.6 (29–39) | Ref. |
| Kidney rejection | 1536 | 88% (88–89) | 21.0 (18–24) | <0.001 |
| Pancreas rejection | 128 | 93% (89–97) | 38.5 (17–61) | 0.650 |
| Kidney and pancreas rejection | 681 | 86% (84–89) | 19.6 (16–23) | <0.001 |
- *CI: confidence interval. Ref.= reference group.
| 5year kidney | Kidney graft half-life | |||
|---|---|---|---|---|
| n | graft survival (CI*) | in years(CI*) | p-value | |
| No rejection | 1906 | 68% (63–73) | 16.0 (13–19) | Ref. |
| Kidney rejection | 1536 | 60% (55–66) | 11.9 (10–14) | <0.001 |
| Pancreas rejection | 128 | 68% (59–79) | 15.4 (10–21) | 0.83 |
| Kidney and pancreas rejection | 681 | 54% (47–62) | 10.3 (8–12) | <0.001 |
- * CI: confidence interval. Ref.= reference group.
| n | 1988–92 | p-value | n | 1993–97 | p-value | |
|---|---|---|---|---|---|---|
| No rejection | 435 | 24.5 | Ref. | 1471 | 44.4 | Ref. |
| Kidney rejection | 621 | 18.7 | 0.03 | 915 | 25.8 | <0.001 |
| Pancreas rejection | 33 | 42.9 | 0.27 | 95 | 37.9 | 0.67 |
| Kidney and pancreas rejection | 231 | 18.3 | 0.06 | 450 | 21.8 | <0.001 |
- Ref.= reference group.
Table 7 shows the adjusted 5-year pancreas graft survival and graft half-life for the various groups in the SKPT population. Half-life of the pancreas graft was significantly reduced when the patients had acute rejection of both the kidney and pancreas, while kidney alone or pancreas alone rejection did not affect the pancreas graft half-life. Like the kidney, pancreas graft half-life has improved over time but the change in the impact of rejection on graft half-life over time was not as apparent as in the kidney (Table 8).
| 5-year pancreas | Pancreas graft | |||
|---|---|---|---|---|
| n | graft survival (CI*) | half-life in years (CI*) | p-value | |
| No rejection | 1906 | 85% (83–87) | 20.0 (17–23) | Ref. |
| Kidney rejection | 1536 | 84% (82–86) | 17.7 (15–20) | 0.185 |
| Pancreas rejection | 128 | 82% (76–90) | 16.3 (9–23) | 0.379 |
| Kidney and pancreas rejection | 681 | 78% (75–82) | 12.8 (11–15) | <0.001 |
- * CI: confidence interval. Ref.= reference group.
| n | 1988–92 | p-value | n | 1993–97 | p-value | |
|---|---|---|---|---|---|---|
| No rejection | 435 | 16.0 | Ref. | 1471 | 28.2 | Ref. |
| Kidney rejection | 621 | 15.7 | 0.88 | 915 | 23.9 | 0.21 |
| Pancreas rejection | 33 | 18.5 | 0.71 | 95 | 18.3 | 0.12 |
| Kidney and pancreas rejection | 231 | 11.3 | 0.01 | 450 | 17.2 | <0.01 |
- Ref.= reference group.
Discussion
The present study using a large cohort of patients from the UNOS database demonstrates that acute rejection episodes involving the kidney graft have a significant negative impact on the long-term kidney graft survival (both 5-year graft survival and graft half-life) in the SKPT population similar to that in cadaver kidney transplant population. Patients who had acute rejection episodes of both kidney and pancreas had the worst long-term survival. Long-term pancreas graft survival was influenced by acute rejection episodes involving both the kidney and pancreas, but not by acute rejection episodes involving the pancreas graft alone. One explanation for this discrepancy could be related to the relatively small number of patients with isolated pancreas rejection (n = 128), leading to a possible type 2 error. It is also possible that rejection episodes involving the pancreas graft are of more severe degree when both organs are involved, compared to isolated pancreas graft rejection episodes. It has been shown in previous studies with kidney transplantation that severe rejection episodes had a greater impact on the long-term graft survival (17).
Several studies have demonstrated that acute rejection episodes have a deleterious effect on the long-term graft survival following kidney transplantation (12–15,17–19). Matas et al. reported that a single rejection episode during the first year reduced the graft half-life by 20 years (45 ± 11 years in those with no rejection vs. 25 ± 8 years in those with 1 rejection episode) (12). Patients with multiple rejection episodes (where t½ = 5 ± 11 years) and first rejection after the first year (where t½ = 3 ± 1 years) had the worst long-term survival (12). Leggat et al. analyzed graft survival in 31 600 cadaver kidney transplants reported to the US Renal Data System (18). The 4-year death censored graft survival rates adjusted for other covariables were 78% for those with no rejection episode during the first year, 72% for patients those who had an acute rejection episode before discharge, 69% for patients with acute rejection between discharge and month 6, and 54% for those with first acute rejection episode in months 7–12. In patients who had acute rejection in more than one time period, later episodes were associated with worse long-term survival, an observation that was independent of previous acute rejection episodes. McLaren et al. performed multivariate logistic regression analysis on 862 renal allografts to identify risk factors for the development of biopsy-proven chronic allograft failure (17). Acute rejection episodes, proteinuria and serum triglycerides were found to be significant risk factors in their analysis. Acute rejection after 6 months was more detrimental, where the risk for allograft failure was 6 times greater with rejection occurring at > 6 months, compared to 2 times greater for rejection episodes occurring within 3 months. The North American Pediatric Renal Transplant Cooperative Study (NAPRTCS) group, in their analysis of 6395 pediatric renal transplants using the proportional hazards model, reported acute rejection, African–American race, repeat transplant and cyclosporine dose < 5 mg/kg/day as risk factors for chronic rejection (19). In this study, the risk of graft failure from chronic rejection was 3 times higher in patients with one acute rejection episode, 12 times higher in patients with 2 or more acute rejection episodes, and 6 times higher for patients with acute rejection episode beyond the first year post transplant.
Two groups have reported the impact of acute rejection episodes in the kidney-pancreas transplant population (20,21). Tesi et al. reported 53% incidence of acute rejection in 160 consecutive technically successful bladder-drained SKPTs (20). Four-year censored kidney and pancreas graft survival rates were: 88% and 97%, respectively, for those without a rejection episode; 97% and 90%, respectively, for those with 1 rejection episode; and 56% and 67% for those with >1 rejection episode. Pirsch analyzed risk factors for long-term pancreas graft loss in 545 SKPT recipients (21). In this multivariate analysis involving both pretransplant as well as the time-dependent factors, risk factors for pancreas graft loss included occurrence of pancreas rejection, kidney graft loss, duration of the disease, and donor age. In the group with pancreas rejection, the risk of graft loss was highest if rejection occurred after 1 month post transplant.
Three per cent of patients experienced isolated pancreas rejection episodes in the present study, similar to that reported in the previous studies. Some investigators have, however, questioned the occurrence of isolated pancreas rejection. Shapiro et al. performed renal biopsies in seven SKPT recipients who had elevated lipase and normal creatinine, and found mild acute cellular rejection in all of them (22). Gruessner et al. have biopsied both the grafts following SKPT in animal experiments using pigs and have found isolated rejection episodes as well as different degrees of severity of rejection in the two organs (23). In the present study, acute rejection episodes of the kidney did not affect long-term pancreas graft survival, and likewise acute rejection episodes of the pancreas did not affect long-term kidney graft survival, suggesting that isolated kidney and pancreas rejection episodes do exist.
One of the limitations of this study, apart from its retrospective nature, is the way the diagnosis of rejection is established. Diagnosis of acute rejection is based on the individual center's criteria, and is not necessarily biopsy proven. Histopathology is the gold standard for investigating episodes of graft dysfunction, while the diagnosis of rejection based solely on clinical criteria can be inaccurate (24). In one study, renal allograft biopsy findings altered patient management recommendations in 40% of patients in whom a presumptive diagnosis had been made on the basis of clinical and laboratory findings (25). However, diagnosis of rejection is well standardized, and criteria used are similar at many centers. In addition, an important question such as this can be answered only through large data analysis.
In the present study, 55% of the SKPT population experienced acute rejection episodes (36% kidney, 3% pancreas and 16% involving both organs) compared to 35% of the cadaver kidney transplant group. It is possible that the incidence of rejection involving both organs is higher than reported, as the pancreas biopsy is not performed in most patients with kidney rejection. As mentioned earlier, rejection involving both organs may be a marker for more severe pancreatic rejection. In addition, the actual incidence of rejection is likely to be higher, as patients who lost their graft(s) during the first year post transplant were excluded from the study. This is consistent with several previously published reports where the incidence of rejection following SKPT was higher than that following cadaver kidney transplantation (20,26). However, with the introduction of potent immunosuppressive agents, such as tacrolimus and mycophenolate mofetil, several centers have reported a decreased incidence of acute rejection, in the 15–30% range, following SKPT (10,11). Several investigators have reported a decreased incidence of chronic allograft nephropathy and an increased half-life with a decrease in the incidence of acute rejection following kidney transplantation in the most recent era (27–29). One could expect an increase in long-term graft survival following SKPT as well with the improvement in the acute rejection rate in the most recent era.
In summary, the present study, utilizing a large cohort of patients reported to the UNOS database, has demonstrated that acute rejection episodes have a negative impact on long-term kidney graft survival in the SKPT population, similar to that in the cadaver kidney transplant population. Patients who had acute rejection episodes of both the kidney and pancreas have the worst long-term graft survival. Acute rejection episodes of the kidney did not affect long-term pancreas graft survival, suggesting that isolated rejection episodes do exist.
Acknowledgments
Presented at the American Transplant Congress, Chicago, May 2001.
