This was a retrospective, longitudinal, 5-year epidemiological study to evaluate clinical and demographic variables impacting survival in pancreas transplant recipients at eight pancreas transplantation centres in Spain. Although all 12 Spanish centres that perform pancreas transplantation were invited to participate, four centres declined to participate for administrative or other internal reasons. All patients transplanted between 1 January 2008 and 31 December 2012 were included; data were collected until 31 December 2015. Due to protracted development of the clinical study report and manuscript, the data were not available for publication until 2020.

The study was conducted in accordance with the Declaration of Helsinki, the International Council of Harmonisation Good Clinical Practice guidelines for observational studies, and local ethical committee regulations in all participating centres. Each patient provided informed consent for inclusion in this study. Retrospective data were collected and anonymized by investigators, preserving confidentiality according to local regulation. Data collection was performed from medical records if the patient met one of the following inclusion criteria: SPK transplantation performed in patients with diabetes mellitus and end-stage renal disease; PAK performed in patients with diabetes mellitus and a functioning kidney graft from a living-related or deceased donor; or PTA performed in patients with labile diabetes mellitus and preserved renal function. In patients undergoing PAK who had received more than one kidney transplant, the last kidney transplant was the reference for data collection.

The co-primary endpoints were patient survival and pancreas graft survival. Patient survival was calculated from the time of pancreas transplantation until either the date of death or study end. Pancreas graft survival was calculated based on the time from transplantation until pancreas graft loss, patient death with functioning pancreas graft, or study end. Pancreas graft failure was defined as C-peptide<1ng/mL and exogenous insulin requirements>0.5IU/kg/day.

Secondary endpoints included: kidney graft survival (defined as time from kidney transplantation until return to dialysis or kidney graft loss), acute pancreatic rejection and biopsy-proven acute rejection (BPAR), acute kidney rejection and BPAR, renal function, glucose metabolism, lipid levels and blood pressure. With the exception of kidney transplant outcomes (i.e. kidney graft survival, and acute kidney rejection and BPAR), secondary endpoints were assessed in all patients irrespective of transplant type.

Patient and donor parameters were assessed to determine possible associations with patient survival and pancreas graft survival. These included the recipient variables: age, sex, weight, body mass index (BMI), type of transplant (SPK, PAK, PTA) and cardiovascular events; and the donor variables: age, sex and pancreas ischaemic time. Safety events occurring during follow-up were also recorded.

Continuous variables were summarized as mean±standard deviation (SD) or median [interquartile range (IQR)], as appropriate. Categorical variables were summarized as incidence counts or rates. The Kaplan–Meier method was used for all time-to-event analyses, including patient survival, graft survival, acute rejection, and BPAR. For graft survival analysis, in cases of death with functioning graft, patients were censored without any event on the date of death. For acute rejection and BPAR, patients were censored without any event on the date of death or graft loss. Univariable and multivariable analyses (Cox proportional hazards model) were conducted to assess the association between baseline clinical and demographic characteristics and patient/graft survival. Factors found to be associated with patient/graft survival in univariable analyses were included in the multivariable analysis and backward elimination applied to identify the variables that significantly affected patient survival. There were no imputations for missing data. All analyses were performed using SAS® Version 9.3, and p≤0.05 was considered statistically significant.

Of 304 patients recruited into the study, three were excluded from the analysis because no post-transplant data were available. Of the 301 patients included in the analysis, there were 241 (80.1%) SPKs, 56 (18.6%) PAKs and four (1.3%) PTAs. Nineteen patients died during the study period. The majority of patients (65.4%) were followed for 3 years after pancreatic transplantation, and 87 (28.9%) were followed for 5 years (Supplementary Fig. 1). The median follow-up period from pancreas transplantation was 3.2 years (maximum 5.6 years).

Patient demographics and baseline characteristics for all transplant recipients are presented in Table 1. The mean±SD age of patients at pancreas transplantation was 40.3±7.5 years; 67.1% were male and 94.4% were Caucasian. The majority of transplant recipients had diabetes mellitus type I (96.7%) and were receiving dialysis (82.3%). The mean±SD time from diagnosis of diabetes (including type I, type II and maturity onset diabetes of the young) until transplantation was 26.1±7.5 years. For the 56 transplant recipients in the PAK group, the median (range) time from kidney transplantation to pancreas transplantation was 2.7 (0.6–14.1) years.

Pancreas donors had a mean±SD age of 32.6±10.6 years (range, 5–54 years), and the majority (57.6%) were male, with a mean±SD BMI of 23.6±2.6kg/m2. Most pancreas donors died as a result of traumatic brain injury (43.3%) or from cerebrocardiovascular causes (42.0%) (Table 2).

Of the 301 patients, all except two received induction therapy with either anti-interleukin-2 (IL-2) receptor (n=161; 53.5%) or thymoglobulin (n=138; 45.8%). All 301 patients received prednisone at baseline; by the end of the study, of 87 patients with available data at 5 years, 65 (74.7%) were receiving prednisone (Supplementary Table 1). All patients were treated with tacrolimus; at baseline, of the 251 patients with available formulation data, 216 (86.1%) received immediate-release tacrolimus. By the end of the study, of the 85 patients with available formulation data at 5 years, 57 (67.1%) were receiving prolonged-release tacrolimus (Supplementary Table 1). The dose of tacrolimus decreased over the 5 years of the study. All 301 patients received concomitant immunosuppression with mycophenolate at baseline [193 (64.1%) received mycophenolate mofetil (MMF) and 108 (35.9%) received mycophenolic acid (MPA)]. Of 85 patients with available data at the end of the study at 5 years, 80 (94.1%) were taking these therapies (Supplementary Table 1).

A total of 282 recipients (93.7%) survived from the date of transplantation until the study end, with estimated 12-month, 3-year and 5-year survival rates of 98.7%, 96.1% and 94.4%, respectively (Fig. 1). A total of 19 (6.3%) recipients died during the study period, due to infection (n=10, 52.6%), vascular disease (n=4, 21.1%), sudden death (n=2, 10.5%), gastrointestinal haemorrhage (n=2, 10.5%) and road traffic accident (n=1, 5.3%). The overall rate of vascular-related death was 1.3% at 5 years post-transplant.

Fig. 1.

Kaplan–Meier curve of overall patient survival.

(0.08MB).

A total of 246 (81.7%) pancreas grafts survived from transplantation date to study end, with estimated 12-month, 3-year and 5-year pancreas graft survival rates of 87.8%, 84.9% and 83.0%, respectively (Fig. 2). Fifty-five (18.3%) patients experienced pancreas graft loss during the study. In these patients, primary non-function occurred in 22 transplants (40.0%), due to vascular thrombosis (n=19, 34.5%), poor graft perfusion (n=2, 3.6%) and bile leak (n=1, 1.8%). The remaining causes of graft loss were acute (n=6, 10.9%) and chronic rejection (n=10, 18.2%), infection (n=2, 3.6%), 10 (18.2%) due to other reasons, and death with a functioning graft (n=5, 9.1%).

Fig. 2.

Kaplan–Meier curve of pancreas graft survival.

(0.08MB).

Of the 297 recipients who received either SPK or PAK transplants, 271 kidney grafts (91.2%) survived from transplantation date until the end of the observation period (Fig. 3). For the 26 recipients (8.8%) who developed graft loss during the observation period, the reasons for graft loss included: death with a functioning graft (n=7, 26.9%), chronic humoral rejection (n=4, 15.4%), thrombosis (n=4, 15.4%), cellular rejection (n=3, 11.5%), acute humoral rejection (n=1, 3.8%), glomerular disease (recurrent or de novo) (each n=1, 3.8%), and other reasons (n=6, 23.1%).

Fig. 3.

Kaplan–Meier curve of kidney graft survival.

(0.08MB).

Of 77 episodes of acute pancreatic rejection in 55 recipients (18.3%), 29 were BPAR (37.7%). The estimated proportion of patients who had not experienced acute rejection was 83.2% at 12 months post-transplant, and 81.7% at 5 years (Fig. 4). At the end of the study, 91.7% of patients had not experienced a BPAR event.

Fig. 4.

Kaplan–Meier curve of pancreas acute rejection-free survival.

(0.13MB).

Of 67 episodes of acute kidney rejection in 49 recipients (16.5%; based on clinical criteria and/or laboratory analyses), 49 were episodes of BPAR in 39 patients (13.1%). At the end of the study, 83.5% and 86.9% of patients had not experienced acute kidney rejection or BPAR, respectively.

Renal function improved from the day of pancreas transplantation up to Month 6 and remained generally stable thereafter. At 5 years, mean±SD serum creatinine concentration was 1.3±0.5mg/dL and mean±SD estimated glomerular filtration rate (eGFR; Modification of Diet in Renal Disease-4 [MDRD-4]) was 70.2±23.0mL/min/1.73m2 (Table 3). Proteinuria remained stable throughout the study period, with a median (IQR) of 170.0 (100, 310)mg/24h at 5 years. Median urine protein-to-creatinine ratio decreased gradually throughout the study period. Glucose metabolism parameters generally remained stable over the 5 years of follow-up (Table 3), with the exception of C-peptide which decreased throughout the study period from 5.4±9.5ng/mL at Month 1 to 1.8±1.1ng/mL at 5 years.

BMI increased from Month 1 post-transplant to the end of the study (22.9±3.5 to 25.0±4.4kg/m2). Lipid profiles remained stable from the day of transplantation to the end of the study: total cholesterol (167.9±42.2–165.2±31.9mg/dL), high density lipoprotein cholesterol (47.5±16.2–54.3±12.3mg/dL), low density lipoprotein cholesterol (90.0±27.2–92.6±23.1mg/dL); triglycerides decreased from 135.4±71.3 on the day of transplantation to 103.1±108.2mg/dL at 5 years. A similar proportion of recipients were being treated with statins at transplant (51.2%) and at study end (51.7%).

Systolic and diastolic blood pressure (SBP and DBP) remained stable from Month 1 post-transplant to the end of the study (SBP: 127.5±18.4–128.2±16.6mmHg; DBP: 74.2±12.6–74.9±11.0mmHg). The proportion of patients with hypertension was 22.2% on the day of transplantation, which decreased and remained stable post-transplant (range, 5.3–8.4%). Post-transplant, the most commonly used hypertension treatments were beta blockers (range, 20.4–24.9%) and calcium antagonists (range, 18.4–25.4%) (Supplementary Table 2). Compared with 1 month post-transplant, the proportion of patients receiving an ACEi or ARB numerically increased by 5 years, while the proportion of patients receiving calcium antagonists decreased.

Given the high patient survival rate, data for the association of risk factors with patient survival are not reported.

Univariable analysis showed that patient age and weight, donor age, previous kidney transplantation, previous cardiovascular events and need for insulin more than 48h post-transplantation were significantly associated with pancreas graft survival. Using a Cox proportional hazards model that included these variables, with backward elimination, previous kidney transplant was the only significant risk factor for graft failure (HR=3.3, 95% CI: 1.8, 5.9; p<0.0001). Kaplan–Meier analysis confirmed that pancreas graft survival was inferior in patients who had received a kidney transplant prior to pancreas transplantation compared with those who had not received a kidney transplant prior to pancreas transplantation (log-rank test, p=0.0002) (Fig. 5).

Fig. 5.

Kaplan–Meier curve of pancreas graft survival according to previous kidney transplant (No, SPK or PTA; Yes, PAK).

(0.1MB).

Table 4 shows the reports of post-transplant surgical intervention, cardiovascular events, infections and tumours occurring during the study. A total of 143 surgical procedures were required in 104 patients (34.6%); 25% of these patients required a surgical procedure during the first 6 months post-transplant. There were 25 cardiovascular events in 21 patients (7.0%), including cerebrovascular accident, coronary disease and amputations. There were 246 infections in 150 patients (49.8%), the majority of which were recorded as ‘bacterial’ in origin (n=199), occurring in 124 patients (41.2%). The median time from pancreas transplantation until the first infection was 2.3 years. Only eleven patients (3.7%) developed tumours during the study.

While this study included a large number of pancreas transplant recipients, there was a considerable difference in the number and types of transplantations performed at the different centres, and four national transplant centres did not participate. Despite this, the setting reflects national clinical practice in Spain for this patient population. The broad inclusion criteria were necessary to ensure all three pancreas transplantation types were considered during the study. Given the high patient survival rate, data for the association of risk factors with patient survival were not reported. Additionally, kidney graft survival was not analyzed according to pancreas transplant type. Additionally, the proportion of patients with missing data increased over time, either due to a lack of follow-up assessments or to omission of results from the medical records. The data relating to renal function and glucose metabolism should therefore be interpreted with caution.