Risk factors for death in kidney transplant patients: analysis from a large protocol biopsy registry

Risk factors for death in kidney transplant patients: analysis from a large protocol biopsy registry Abstract Background Identification and quantification of the relevant factors for death can improve patients’ individual risk assessment and decision-making. We used a well-documented patient cohort (n = 892) in a renal transplant programme with protocol biopsies to establish multivariable Cox models for risk assessment at 3 and 12 months post-transplantation. Methods Patients transplanted between 2000 and 2007 were observed up to 11 years (total observation 5227 patient-years; median 5.9 years). Loss to follow-up was negligible (n = 15). A total of 2251 protocol biopsies and 1214 biopsies for cause were performed. All rejections and clinical borderline rejections in protocol biopsies were treated. Results Overall 10-year patient survival was 78%, with inferior survival of patients with graft loss and superior survival of patients with living-donor transplantation. Eight factors were common in the models at 3 and 12 months, including age, pre-transplant heart failure and a score of cardiovascular disease and type 2 diabetes, post-transplant urinary tract infection, treatment of rejection, new-onset heart failure, coronary events and malignancies. Additional variables of the model at 3 months included deceased donor transplantation, transplant lymphocele, BK virus nephropathy and severe infections. Graft function and graft loss were significant factors of the model at 12 months. Internal validation and validation with a separate cohort of patients (n = 349) demonstrated good discrimination of the models. Conclusions The identified factors indicate the important areas that need special attention in the pre- and post-transplant care of renal transplant patients. On the basis of these models, we provide nomograms as a tool to weigh individual risks that may contribute to decreased survival. kidney transplantation, protocol biopsies, risk factors, survival INTRODUCTION Patients with end-stage renal disease have a significant excess mortality compared with the general population. A high prevalence of cardiovascular disease is one of the major comorbidity factors for this increased mortality [1]. Kidney transplantation can lower the mortality in the long term [2]. However, in addition to the peri-operative risks, new hazards can develop including infections, malignancies, metabolic disorders and hypertension [3]. Consideration of individual pre- and post-transplant risks can help to tailor patient’s specific diagnostic and therapeutic requirements. In addition, quantification of risks by multivariable decision models may be a supportive tool for balancing the risks and benefits of potentially harmful treatments for individual patients. The latter is of special importance with regard to the increasing implementation of protocol biopsies and treatment of subclinical rejections. The aim of our study is to establish robust decision models that help in estimating the risks stemming from pre- and post-transplant factors. Therefore, we used a well-documented patient cohort of ∼900 patients in a specialized post-transplant care programme with protocol biopsies and a long-term follow-up of up to 11 years. Unlike previous studies, assessment of the risks for death included patients with graft loss and return to dialysis. MATERIALS AND METHODS Patients In this retrospective cohort study, 892 adult patients were included who received a kidney transplant alone or in combination with another solid organ at Hannover Medical School between 2000 and 2007 and who participated in our protocol biopsy programme. Protocol biopsies were performed 6 weeks, 3 and 6 months after transplantation. Data were collected prior to and at the time of transplantation, at the time points of protocol biopsies and any additional biopsies, and in yearly intervals after transplantation (complete list of variables in Supplementary data, Table S1). For patients who were followed-up elsewhere, data were retrieved by contacting their local caregivers. The total observation time was 5227 patient-years (25/50/75 percentile: 4.0, 5.9 and 8.0 years). Fifteen patients were lost to follow-up at a median of 3.0 years (minimum observation time 1.9 years) after transplantation. These were included in the modelling and censored at their last observation time. A separate cohort of 349 patients transplanted in 2008–13 and followed up until June 2017 was used to validate the obtained results. Data collection and analysis were performed with informed consent of the patients and with approval of the ethic board (no. 2765) of the Hannover Medical School. Pre-transplant and peri-operative data are depicted in Table 1. Combined organ transplantations included kidney with pancreas (n = 65), liver (n = 4), heart (n = 7) and lung (n = 5). The estimated glomerular filtration rate (eGFR; mL/min/1.7 m2) was calculated with the Cockcroft–Gault formula. Urinary tract infection was defined by leucocyturia and a urine culture with >104 bacterial or >102 fungal colonies. Urological interventions excluded removal of ureteral stents that had been implanted during transplantation. Death is defined as all-cause mortality with functioning graft or after graft loss. Table 1 Pre-transplant and peri-operative data of the whole patient group and of surviving and deceased patients   All patients, n = 892  Surviving patients, n = 760  Deceased patients, n = 132  P-value  Pre-transplant data  Recipient          Age (years)  50 ± 13  48 ± 13  59 ± 11  <0.001  Gender (%; male/female)  59/41  59/41  61/39  0.774  Main reason for endstage renal disease (%)        0.056   Glomerulonephritis/vasculitis  26  27  17     Tubulointerstitial diseases  9  9  8     Hypertensive/diabetic nephropathy  14  13  18     Congenital disease  15  15  14     Other known causes  4  4  2     Unknown  33  31  40    Comorbidity before and at Tx (%)           Arterial hypertension  96  96  97  0.806   Heart failure  2  1  8  <0.001   Coronary heart disease  16  14  31  <0.001    History of myocardial infarction  5  4  11  0.002    History of coronary bypass surgery  6  5  13  0.003    History of coronary angioplasty  5  4  12  0.001   Peripheral arterial disease  13  10  32  <0.001   Ischaemic cerebral stroke  4  3  6  0.116   Replicative hepatitis B  1  1  1  1.000   Replicative hepatitis C  6  7  4  0.246   Type I diabetes  8  9  8  1.000   Type II diabetes  7  5  13  0.003   Hypercholesterolaemia  60  60  62  0.908   Hyperparathyroidism  67  68  62  0.429   Parathyroidectomy before Tx  22  23  17  0.091   CMV IgG positivity  58  58  61  0.502   EBV IgG positivity  91  91  89  0.501   Malignancies (excluding non-melanotic skin cancer)  6  6  7  0.689   Monoclonal gammopathy  1  1  3  0.087   Previous cytotoxic therapies  2  2  2  1.000   History of cigarette smoking  41  41  38  0.396   Time on dialysis before Tx (months)  70 ± 41  69 ± 42  78 ± 33  0.009    Deceased donor recipients  77 ± 38  77 ± 39  80 ± 31  0.474    Living donor recipients  29 ± 33  28 ± 33  37 ± 31  0.366   Re-transplanted patients  13  13  11  0.579   Body weight at Tx (kg)  72 ± 13  72 ± 13  73 ± 12  0.356   BMI at Tx  24 ± 4  24 ± 4  25 ± 3  0.063  Peri-operative data  Transplant-related factors           ESP (%)  10  8  24  <0.001   Combined kidney/pancreas Tx (%)  7  8  6  0.717   Pre-formed antibodies >0% (%)  7  7  8  0.854   Donor age (years)  49 ± 16  48 ± 15  53 ± 17  <0.001   Donor gender (%; male/female)  52/48  53/47  48/52  0.343   Donor type (%; deceased/living)  85/15  83/17  95/5  <0.001   Donor CMV IgG positivity (%)  58  58  59  0.924   Donor S-creatinine (µmol/L)  83 ± 46  84 ± 48  81 ± 34  0.890   Mean number of HLA mismatches  2.4 ± 1.7  2.3 ± 1.7  2.8 ± 1.8  0.005    HLA-A mismatches  0.71 ± 0.71  0.69 ± 0.69  0.81 ± 0.78  0.127    HLA-B mismatches  0.91 ± 0.73  0.89 ± 0.72  1.03 ± 0.75  0.048    HLA-DR mismatches  0.78 ± 0.70  0.75 ± 0.69  0.96 ± 0.70  0.001   CIT (h)  14.5 ± 7.4  14.0 ± 7.4  17.3 ± 6.8  <0.001    Deceased donor recipients  16.7 ± 5.9  16.4 ± 5.8  18.0 ± 6.1  0.004    Living donor recipients  2.5 ± 0.9  2.5 ± 0.9  2.7 ± 0.5  0.208  Initial immunosupressive therapy (%)           Induction therapy        0.660    Anti-lymphocyte globulin  9  9  11      IL-2 antibodies  79  79  80      None  10  10  7      Other/unknown  1/1  1/1  2/2     Cyclosporine A  75  75  74  0.664   Tacrolimus  19  20  13  0.055   MMF  66  67  61  0.197   Azathioprine  1  1  0  0.602   Everolimus or rapamycin  5  4  6  0.360   Steroids  95  96  95  0.653   Study drugs  9  9  9  1.000  Early post-Tx course           DGF (%)  29  27  43  <0.001   Best eGFR in the first 6 weeks post-Tx (mL/min/1.7m2)  59 ± 24  61 ± 24  49 ± 22  <0.001    All patients, n = 892  Surviving patients, n = 760  Deceased patients, n = 132  P-value  Pre-transplant data  Recipient          Age (years)  50 ± 13  48 ± 13  59 ± 11  <0.001  Gender (%; male/female)  59/41  59/41  61/39  0.774  Main reason for endstage renal disease (%)        0.056   Glomerulonephritis/vasculitis  26  27  17     Tubulointerstitial diseases  9  9  8     Hypertensive/diabetic nephropathy  14  13  18     Congenital disease  15  15  14     Other known causes  4  4  2     Unknown  33  31  40    Comorbidity before and at Tx (%)           Arterial hypertension  96  96  97  0.806   Heart failure  2  1  8  <0.001   Coronary heart disease  16  14  31  <0.001    History of myocardial infarction  5  4  11  0.002    History of coronary bypass surgery  6  5  13  0.003    History of coronary angioplasty  5  4  12  0.001   Peripheral arterial disease  13  10  32  <0.001   Ischaemic cerebral stroke  4  3  6  0.116   Replicative hepatitis B  1  1  1  1.000   Replicative hepatitis C  6  7  4  0.246   Type I diabetes  8  9  8  1.000   Type II diabetes  7  5  13  0.003   Hypercholesterolaemia  60  60  62  0.908   Hyperparathyroidism  67  68  62  0.429   Parathyroidectomy before Tx  22  23  17  0.091   CMV IgG positivity  58  58  61  0.502   EBV IgG positivity  91  91  89  0.501   Malignancies (excluding non-melanotic skin cancer)  6  6  7  0.689   Monoclonal gammopathy  1  1  3  0.087   Previous cytotoxic therapies  2  2  2  1.000   History of cigarette smoking  41  41  38  0.396   Time on dialysis before Tx (months)  70 ± 41  69 ± 42  78 ± 33  0.009    Deceased donor recipients  77 ± 38  77 ± 39  80 ± 31  0.474    Living donor recipients  29 ± 33  28 ± 33  37 ± 31  0.366   Re-transplanted patients  13  13  11  0.579   Body weight at Tx (kg)  72 ± 13  72 ± 13  73 ± 12  0.356   BMI at Tx  24 ± 4  24 ± 4  25 ± 3  0.063  Peri-operative data  Transplant-related factors           ESP (%)  10  8  24  <0.001   Combined kidney/pancreas Tx (%)  7  8  6  0.717   Pre-formed antibodies >0% (%)  7  7  8  0.854   Donor age (years)  49 ± 16  48 ± 15  53 ± 17  <0.001   Donor gender (%; male/female)  52/48  53/47  48/52  0.343   Donor type (%; deceased/living)  85/15  83/17  95/5  <0.001   Donor CMV IgG positivity (%)  58  58  59  0.924   Donor S-creatinine (µmol/L)  83 ± 46  84 ± 48  81 ± 34  0.890   Mean number of HLA mismatches  2.4 ± 1.7  2.3 ± 1.7  2.8 ± 1.8  0.005    HLA-A mismatches  0.71 ± 0.71  0.69 ± 0.69  0.81 ± 0.78  0.127    HLA-B mismatches  0.91 ± 0.73  0.89 ± 0.72  1.03 ± 0.75  0.048    HLA-DR mismatches  0.78 ± 0.70  0.75 ± 0.69  0.96 ± 0.70  0.001   CIT (h)  14.5 ± 7.4  14.0 ± 7.4  17.3 ± 6.8  <0.001    Deceased donor recipients  16.7 ± 5.9  16.4 ± 5.8  18.0 ± 6.1  0.004    Living donor recipients  2.5 ± 0.9  2.5 ± 0.9  2.7 ± 0.5  0.208  Initial immunosupressive therapy (%)           Induction therapy        0.660    Anti-lymphocyte globulin  9  9  11      IL-2 antibodies  79  79  80      None  10  10  7      Other/unknown  1/1  1/1  2/2     Cyclosporine A  75  75  74  0.664   Tacrolimus  19  20  13  0.055   MMF  66  67  61  0.197   Azathioprine  1  1  0  0.602   Everolimus or rapamycin  5  4  6  0.360   Steroids  95  96  95  0.653   Study drugs  9  9  9  1.000  Early post-Tx course           DGF (%)  29  27  43  <0.001   Best eGFR in the first 6 weeks post-Tx (mL/min/1.7m2)  59 ± 24  61 ± 24  49 ± 22  <0.001  Fifteen of the 892 patients were lost to follow-up during the post-transplant course. Tx, transplantation; CMV, cytomegalovirus; EBV, Epstein–Barr virus. Heart failure: any degree of insufficiency. Hyperparathyroidism is defined as an elevation of the most recent parathormone value within the year before transplantation by at least 2-fold of the upper normal. Pre-formed antibodies were determined by the lymphocytotoxic panel reactive antibody test. Serum creatinine of the donor represents the last known value before organ explantation and was available in 371 of the cases. DGF was defined as <500 mL urine within the first 24 h after transplantation and/or need of dialysis within the first week. Means are given with SD. Table 1 Pre-transplant and peri-operative data of the whole patient group and of surviving and deceased patients   All patients, n = 892  Surviving patients, n = 760  Deceased patients, n = 132  P-value  Pre-transplant data  Recipient          Age (years)  50 ± 13  48 ± 13  59 ± 11  <0.001  Gender (%; male/female)  59/41  59/41  61/39  0.774  Main reason for endstage renal disease (%)        0.056   Glomerulonephritis/vasculitis  26  27  17     Tubulointerstitial diseases  9  9  8     Hypertensive/diabetic nephropathy  14  13  18     Congenital disease  15  15  14     Other known causes  4  4  2     Unknown  33  31  40    Comorbidity before and at Tx (%)           Arterial hypertension  96  96  97  0.806   Heart failure  2  1  8  <0.001   Coronary heart disease  16  14  31  <0.001    History of myocardial infarction  5  4  11  0.002    History of coronary bypass surgery  6  5  13  0.003    History of coronary angioplasty  5  4  12  0.001   Peripheral arterial disease  13  10  32  <0.001   Ischaemic cerebral stroke  4  3  6  0.116   Replicative hepatitis B  1  1  1  1.000   Replicative hepatitis C  6  7  4  0.246   Type I diabetes  8  9  8  1.000   Type II diabetes  7  5  13  0.003   Hypercholesterolaemia  60  60  62  0.908   Hyperparathyroidism  67  68  62  0.429   Parathyroidectomy before Tx  22  23  17  0.091   CMV IgG positivity  58  58  61  0.502   EBV IgG positivity  91  91  89  0.501   Malignancies (excluding non-melanotic skin cancer)  6  6  7  0.689   Monoclonal gammopathy  1  1  3  0.087   Previous cytotoxic therapies  2  2  2  1.000   History of cigarette smoking  41  41  38  0.396   Time on dialysis before Tx (months)  70 ± 41  69 ± 42  78 ± 33  0.009    Deceased donor recipients  77 ± 38  77 ± 39  80 ± 31  0.474    Living donor recipients  29 ± 33  28 ± 33  37 ± 31  0.366   Re-transplanted patients  13  13  11  0.579   Body weight at Tx (kg)  72 ± 13  72 ± 13  73 ± 12  0.356   BMI at Tx  24 ± 4  24 ± 4  25 ± 3  0.063  Peri-operative data  Transplant-related factors           ESP (%)  10  8  24  <0.001   Combined kidney/pancreas Tx (%)  7  8  6  0.717   Pre-formed antibodies >0% (%)  7  7  8  0.854   Donor age (years)  49 ± 16  48 ± 15  53 ± 17  <0.001   Donor gender (%; male/female)  52/48  53/47  48/52  0.343   Donor type (%; deceased/living)  85/15  83/17  95/5  <0.001   Donor CMV IgG positivity (%)  58  58  59  0.924   Donor S-creatinine (µmol/L)  83 ± 46  84 ± 48  81 ± 34  0.890   Mean number of HLA mismatches  2.4 ± 1.7  2.3 ± 1.7  2.8 ± 1.8  0.005    HLA-A mismatches  0.71 ± 0.71  0.69 ± 0.69  0.81 ± 0.78  0.127    HLA-B mismatches  0.91 ± 0.73  0.89 ± 0.72  1.03 ± 0.75  0.048    HLA-DR mismatches  0.78 ± 0.70  0.75 ± 0.69  0.96 ± 0.70  0.001   CIT (h)  14.5 ± 7.4  14.0 ± 7.4  17.3 ± 6.8  <0.001    Deceased donor recipients  16.7 ± 5.9  16.4 ± 5.8  18.0 ± 6.1  0.004    Living donor recipients  2.5 ± 0.9  2.5 ± 0.9  2.7 ± 0.5  0.208  Initial immunosupressive therapy (%)           Induction therapy        0.660    Anti-lymphocyte globulin  9  9  11      IL-2 antibodies  79  79  80      None  10  10  7      Other/unknown  1/1  1/1  2/2     Cyclosporine A  75  75  74  0.664   Tacrolimus  19  20  13  0.055   MMF  66  67  61  0.197   Azathioprine  1  1  0  0.602   Everolimus or rapamycin  5  4  6  0.360   Steroids  95  96  95  0.653   Study drugs  9  9  9  1.000  Early post-Tx course           DGF (%)  29  27  43  <0.001   Best eGFR in the first 6 weeks post-Tx (mL/min/1.7m2)  59 ± 24  61 ± 24  49 ± 22  <0.001    All patients, n = 892  Surviving patients, n = 760  Deceased patients, n = 132  P-value  Pre-transplant data  Recipient          Age (years)  50 ± 13  48 ± 13  59 ± 11  <0.001  Gender (%; male/female)  59/41  59/41  61/39  0.774  Main reason for endstage renal disease (%)        0.056   Glomerulonephritis/vasculitis  26  27  17     Tubulointerstitial diseases  9  9  8     Hypertensive/diabetic nephropathy  14  13  18     Congenital disease  15  15  14     Other known causes  4  4  2     Unknown  33  31  40    Comorbidity before and at Tx (%)           Arterial hypertension  96  96  97  0.806   Heart failure  2  1  8  <0.001   Coronary heart disease  16  14  31  <0.001    History of myocardial infarction  5  4  11  0.002    History of coronary bypass surgery  6  5  13  0.003    History of coronary angioplasty  5  4  12  0.001   Peripheral arterial disease  13  10  32  <0.001   Ischaemic cerebral stroke  4  3  6  0.116   Replicative hepatitis B  1  1  1  1.000   Replicative hepatitis C  6  7  4  0.246   Type I diabetes  8  9  8  1.000   Type II diabetes  7  5  13  0.003   Hypercholesterolaemia  60  60  62  0.908   Hyperparathyroidism  67  68  62  0.429   Parathyroidectomy before Tx  22  23  17  0.091   CMV IgG positivity  58  58  61  0.502   EBV IgG positivity  91  91  89  0.501   Malignancies (excluding non-melanotic skin cancer)  6  6  7  0.689   Monoclonal gammopathy  1  1  3  0.087   Previous cytotoxic therapies  2  2  2  1.000   History of cigarette smoking  41  41  38  0.396   Time on dialysis before Tx (months)  70 ± 41  69 ± 42  78 ± 33  0.009    Deceased donor recipients  77 ± 38  77 ± 39  80 ± 31  0.474    Living donor recipients  29 ± 33  28 ± 33  37 ± 31  0.366   Re-transplanted patients  13  13  11  0.579   Body weight at Tx (kg)  72 ± 13  72 ± 13  73 ± 12  0.356   BMI at Tx  24 ± 4  24 ± 4  25 ± 3  0.063  Peri-operative data  Transplant-related factors           ESP (%)  10  8  24  <0.001   Combined kidney/pancreas Tx (%)  7  8  6  0.717   Pre-formed antibodies >0% (%)  7  7  8  0.854   Donor age (years)  49 ± 16  48 ± 15  53 ± 17  <0.001   Donor gender (%; male/female)  52/48  53/47  48/52  0.343   Donor type (%; deceased/living)  85/15  83/17  95/5  <0.001   Donor CMV IgG positivity (%)  58  58  59  0.924   Donor S-creatinine (µmol/L)  83 ± 46  84 ± 48  81 ± 34  0.890   Mean number of HLA mismatches  2.4 ± 1.7  2.3 ± 1.7  2.8 ± 1.8  0.005    HLA-A mismatches  0.71 ± 0.71  0.69 ± 0.69  0.81 ± 0.78  0.127    HLA-B mismatches  0.91 ± 0.73  0.89 ± 0.72  1.03 ± 0.75  0.048    HLA-DR mismatches  0.78 ± 0.70  0.75 ± 0.69  0.96 ± 0.70  0.001   CIT (h)  14.5 ± 7.4  14.0 ± 7.4  17.3 ± 6.8  <0.001    Deceased donor recipients  16.7 ± 5.9  16.4 ± 5.8  18.0 ± 6.1  0.004    Living donor recipients  2.5 ± 0.9  2.5 ± 0.9  2.7 ± 0.5  0.208  Initial immunosupressive therapy (%)           Induction therapy        0.660    Anti-lymphocyte globulin  9  9  11      IL-2 antibodies  79  79  80      None  10  10  7      Other/unknown  1/1  1/1  2/2     Cyclosporine A  75  75  74  0.664   Tacrolimus  19  20  13  0.055   MMF  66  67  61  0.197   Azathioprine  1  1  0  0.602   Everolimus or rapamycin  5  4  6  0.360   Steroids  95  96  95  0.653   Study drugs  9  9  9  1.000  Early post-Tx course           DGF (%)  29  27  43  <0.001   Best eGFR in the first 6 weeks post-Tx (mL/min/1.7m2)  59 ± 24  61 ± 24  49 ± 22  <0.001  Fifteen of the 892 patients were lost to follow-up during the post-transplant course. Tx, transplantation; CMV, cytomegalovirus; EBV, Epstein–Barr virus. Heart failure: any degree of insufficiency. Hyperparathyroidism is defined as an elevation of the most recent parathormone value within the year before transplantation by at least 2-fold of the upper normal. Pre-formed antibodies were determined by the lymphocytotoxic panel reactive antibody test. Serum creatinine of the donor represents the last known value before organ explantation and was available in 371 of the cases. DGF was defined as <500 mL urine within the first 24 h after transplantation and/or need of dialysis within the first week. Means are given with SD. Biopsies A total number of 2251 protocol biopsies and 1214 biopsies for cause were performed. Acute T-cell-mediated rejections including borderline cases were treated with steroid boli, except subclinical borderline cases in protocol biopsies, defined by an increase in serum creatinine by <20% at biopsy. In addition, in patients on a dual therapy mycophenolate mofetil was added. Patients with acute rejections occurring at 6 months or later or with vascular rejection at any time point were switched from cyclosporine to tacrolimus. No standardized treatment was defined for humoral rejection. Within the first transplant year, 316 rejections detected in protocol biopsies and 249 rejections detected in biopsies for cause were treated. Twenty-one cases of recurrent renal disease occurred among the surviving patients and eight cases among deceased patients, with only five recurrences within the first year (all in the group of surviving patients, with one case each with membranoproliferative and membranous glomerulonephritis, focal segmental glomerulosclerosis and two cases with thrombotic microangiopathy). Statistical analysis The IBM SPSS statistical software package version 24 and the rms package from R software version 3.2.1 [4] were used. The number of missing data was low except for the pre-transplant hyperparathyroidism (n = 344), hypercholesterolaemia (n = 301), blood transfusions (n = 103) and smoking (n = 79), which were therefore not included in the multivariable modelling. For a further 18 variables, a mean of 16 values was missing. These were replaced by the median of the whole group (continuous variables) and the most common attribute value (categorical variables) [5]. Continuous variables with normal distribution are given as means ± SD, data without normal distribution as medians with 25/75% quartiles. Categorical data were analysed with Fisher’s exact test and chi-square test for two or more samples. Continuous data were compared with the Kruskal–Wallis and Mann–Whitney test. Kaplan–Meier analysis and the log-rank test were used to describe patient and graft survival. For the Cox regression analyses, linearity of continuous variables was confirmed by categorizing the variables and comparing the β-coefficients from univariable Cox regression. Proportional hazards assumptions were confirmed graphically and by testing the scaled Schoenfeld residuals with the global PH-Test. Variables differing with a P-value of  ≤0.05 in univariable Cox analyses were considered for multivariable modelling, which was performed by stepwise backward selection (P-value threshold ≤0.05) to calculate the hazard ratios (HRs) for death. Validation of the models was performed by the bootstrapping procedures in the R software. Harrell’s concordance index was taken as measure of discrimination. Survival probabilities in the validation cohort were depicted by Kaplan–Meier curves for four discrete risks groups obtained from the linear predictor, using cut-points on the prognostic index determined by Cox’s method [6]. Nomograms were built using the estimators from the multivariable Cox modelling. Statistical significance was assumed for P < 0.05 (two tailed). RESULTS Patient survival Patient survival was 89% after 5 years and 78% after 10 years. Ninety-nine of 776 patients with functioning graft and 33 of 116 patients with graft failure and return to dialysis died (Figure 1A and B). Death events in patients with living-donor transplants were >50% lower, and twice as high in patients transplanted in the Eurotransplant Senior Programme (ESP; [7]) (Figure 1C). However, patient survival was not significantly different between patients in the ESP and patients aged >65 years who received an organ from a deceased donor aged <65 years (P = 0.20; Figure 1D). Two deaths occurred in the 16 cases with combined transplantation of kidney with heart, liver or lung, both after renal graft failure. Infection was the leading cause of death (23%), followed by malignancies (14%) and cardiovascular disease (14%). Malignancies included carcinoma of unknown origin (n = 7), three bronchial carcinomas, four carcinomas and one sarcoma of the gut, two liver carcinomas, one prostate carcinoma and one non-Hodgkin lymphoma. In 42%, there was no reliable information on the cause of death. Other specified death causes included one suicide, two traffic accidents, one fatal bleeding after transplant kidney rupture and four gastrointestinal causes (multiorgan failures after pancreatitis, bowel perforation after endoscopy, oesophageal varicose bleeding). The median time to death was similar among the different causes of death (3.2–3.7 years) except for other causes (2.2 years). FIGURE 1 View largeDownload slide Overall survival and survival in different subgroups. Dashed lines indicate 95% upper and lower confidence limits. (A) Overall survival of the 892 patients. (B) Survival in patients with functioning graft and with graft loss. Patients with graft loss had significant worse survival (P< 0.001). (C) Survival in patients with combined kidney with liver, heart or lung transplantations (two deaths in 16 patients), living-donor transplantations (n = 6 deaths in 134 patients), combined pancreas/kidney transplantations (n = 8 deaths in 65 patients) and patients transplanted in the ESP (n = 32 deaths in 93 patients). Groups were significantly different in the log-rank test (P< 0.001). (D) Survival in patients transplanted in the ESP compared with patients aged >65 years who received transplants from donors aged <65 years (P = 0.20). FIGURE 1 View largeDownload slide Overall survival and survival in different subgroups. Dashed lines indicate 95% upper and lower confidence limits. (A) Overall survival of the 892 patients. (B) Survival in patients with functioning graft and with graft loss. Patients with graft loss had significant worse survival (P< 0.001). (C) Survival in patients with combined kidney with liver, heart or lung transplantations (two deaths in 16 patients), living-donor transplantations (n = 6 deaths in 134 patients), combined pancreas/kidney transplantations (n = 8 deaths in 65 patients) and patients transplanted in the ESP (n = 32 deaths in 93 patients). Groups were significantly different in the log-rank test (P< 0.001). (D) Survival in patients transplanted in the ESP compared with patients aged >65 years who received transplants from donors aged <65 years (P = 0.20). Pre- and post-transplant factors with association to death Table 1 shows the comparison of pre-transplant and peri-operative variables between surviving patients and patients who died. Deceased patients were older and had spent more time on dialysis. Pre-existing cardiovascular disease, heart failure and type 2 diabetes were more prevalent. Transplants from deceased or older donors were more common and HLA-mismatches were higher. Cold ischaemic times (CITs) were longer in deceased patients receiving a transplant from a deceased donor. Delayed graft function (DGF) was more prevalent and the eGFR within the first 6 weeks after transplantation was lower. Supplementary data, Table S2 shows these variables descriptively in the subgroups with and without graft loss. In an explorative analysis, data of the further post-transplant course (complete list in Supplementary data, Table S1) were compared between survivors and deceased patients (Supplementary data, Table S3). Graft function was lower in deceased patients and graft loss more prevalent. Deceased patients had a higher graft arterial resistance index, more urinary tract infections, a higher frequency of BK virus nephropathy, treatment for acute T-cell-mediated and antibody-mediated rejection within the first transplant year and urological interventions. New-onset heart failure and myocardial infarction, coronary artery angioplasty or surgical revascularization, severe other diseases including severe infections and newly discovered malignancies were more prevalent in the post-transplant course of deceased patients. HRs and multivariable modelling of factors for death at 3 and 12 months after transplantation including validation of the models Univariable HRs were calculated for all pre- and post-transplant variables at 3 and 12 months after transplantation (Tables 2 and 3). Due to deaths before 3 months (n = 2) and 12 months (n = 17), these calculations were based on 890 and 875 patients, respectively. Variables with a P-value of ≤0.05 were used to create multivariable models at the two time points. Graft function, expressed as eGFR, was linearly correlated with survival. Graft loss was included in this linear predictor (annotated as an eGFR of 5 mL/min/1.7 m2) instead of using it as a separate variable in the modelling because eGFR and graft loss were highly collinear. Pre-transplant coronary heart disease, peripheral arterial disease and type 2 diabetes were combined in a score because this yielded the most stable results in the modelling and prevented multicollinearity. Linearity of this score was confirmed by comparing the β-coefficients from univariable Cox regression. Transplantation in the ESP was not included in the modelling as this variable was represented by recipient and donor age and HLA matching. Table 2 HRs of pre- and post-transplant variables at 3 months after transplantation from uni- and multivariable Cox regression analysis   Univariable analysis   Multivariable analysis   HR  95% CI  P-value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.067  1.050 1.085  <0.001  1.051  1.033 1.069  <0.001   Heart failure  5.320  2.855 9.914  <0.001  2.749  1.393 5.424  0.004   DM/PAD/CHD score  1.917  1.595 2.304  <0.001  1.412  1.143 1.745  0.001   Coronary heart disease  2.338  1.606 3.402  <0.001         Peripheral arterial disease  3.419  2.362 4.951  <0.001         Type 2 diabetes  2.560  1.514 4.328  <0.001         Monoclonal gammopathy  2.711  1.000 7.347  0.050         Time on dialysis (years)  1.005  1.001 1.009  0.017        Variables at Tx               BMI  1.049  1.000 1.101  0.050         Deceased donor Tx  3.707  1.634 8.410  0.002  2.640  1.100 6.336  0.030   Tx within the ESP  3.203  2.147 4.779  <0.001         Donor age (years)  1.026  1.014 1.038  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.076 1.320  <0.001          HLA-A mismatches  1.298  1.024 1.645  0.031          HLA-B mismatches  1.321  1.040 1.677  0.022          HLA-DR mismatches  1.549  1.217 1.972  <0.001          CIT (h)  1.045  1.022 1.068  <0.001        Post-Tx until 3 months              Graft factors               DGF  1.887  1.334 2.671  <0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.987  <0.001         eGFR at 3 months post-Tx (mL/min/1.7 m2)  0.971  0.962 0.981  <0.001         Graft loss  1.848  0.457 7.473  0.389         Renal graft arterial resistance index >0.75  2.836  1.963 4.097  <0.001         Urinary tract infection (%)  1.013  1.009 1.018  0.000  1.007  1.002 1.012  0.002   Transplant lymphocele  2.056  1.044 4.050  0.037  2.122  1.055 4.270  0.035   Urological interventions  1.582  1.069 2.340  0.022         Number of treated rejection episodes  1.342  1.122 1.604  0.001  2.055  1.428 2.957  <0.001   BK virus nephropathy  5.172  1.906 14.029  0.01  3.481  1.191 10.175  0.023  Comorbidities               New-onset heart failure  6.026  2.224 16.323  <0.001  5.675  2.047 15.736  0.001   Myocardial infarction, percutaneous or surgical coronary intervention  12.230  5.313 28.153  <0.001  5.815  2.468 13.703  <0.001   Severe acute diseases  3.807  2.285 6.341  <0.001         Severe infections  2.436  1.421 4.176  0.001  1.901  1.072 3.369  0.028   Malignancies  14.908  6.509 34.143  <0.001  19.404  7.892 47.709  <0.001   BMI change since Tx  0.892  0.818 0.973  0.010          Univariable analysis   Multivariable analysis   HR  95% CI  P-value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.067  1.050 1.085  <0.001  1.051  1.033 1.069  <0.001   Heart failure  5.320  2.855 9.914  <0.001  2.749  1.393 5.424  0.004   DM/PAD/CHD score  1.917  1.595 2.304  <0.001  1.412  1.143 1.745  0.001   Coronary heart disease  2.338  1.606 3.402  <0.001         Peripheral arterial disease  3.419  2.362 4.951  <0.001         Type 2 diabetes  2.560  1.514 4.328  <0.001         Monoclonal gammopathy  2.711  1.000 7.347  0.050         Time on dialysis (years)  1.005  1.001 1.009  0.017        Variables at Tx               BMI  1.049  1.000 1.101  0.050         Deceased donor Tx  3.707  1.634 8.410  0.002  2.640  1.100 6.336  0.030   Tx within the ESP  3.203  2.147 4.779  <0.001         Donor age (years)  1.026  1.014 1.038  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.076 1.320  <0.001          HLA-A mismatches  1.298  1.024 1.645  0.031          HLA-B mismatches  1.321  1.040 1.677  0.022          HLA-DR mismatches  1.549  1.217 1.972  <0.001          CIT (h)  1.045  1.022 1.068  <0.001        Post-Tx until 3 months              Graft factors               DGF  1.887  1.334 2.671  <0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.987  <0.001         eGFR at 3 months post-Tx (mL/min/1.7 m2)  0.971  0.962 0.981  <0.001         Graft loss  1.848  0.457 7.473  0.389         Renal graft arterial resistance index >0.75  2.836  1.963 4.097  <0.001         Urinary tract infection (%)  1.013  1.009 1.018  0.000  1.007  1.002 1.012  0.002   Transplant lymphocele  2.056  1.044 4.050  0.037  2.122  1.055 4.270  0.035   Urological interventions  1.582  1.069 2.340  0.022         Number of treated rejection episodes  1.342  1.122 1.604  0.001  2.055  1.428 2.957  <0.001   BK virus nephropathy  5.172  1.906 14.029  0.01  3.481  1.191 10.175  0.023  Comorbidities               New-onset heart failure  6.026  2.224 16.323  <0.001  5.675  2.047 15.736  0.001   Myocardial infarction, percutaneous or surgical coronary intervention  12.230  5.313 28.153  <0.001  5.815  2.468 13.703  <0.001   Severe acute diseases  3.807  2.285 6.341  <0.001         Severe infections  2.436  1.421 4.176  0.001  1.901  1.072 3.369  0.028   Malignancies  14.908  6.509 34.143  <0.001  19.404  7.892 47.709  <0.001   BMI change since Tx  0.892  0.818 0.973  0.010        Heart failure: any grade of insufficiency. Presence of type 2 diabetes, peripheral arterial disease (any grade) and coronary heart disease is additionally expressed as a score (DM/PAD/CHD score), which was later used in the multivariable analyses (1 point for each disease entity, adding up to a discrete score of 0–3). HLA mismatch on locus A, B and DR is entered as a numerical variable as 0, 1 or 2 according to the number of mismatches. eGFR: a value of 5 mL/min/1.7 m2 was assigned to patients with graft loss. Renal graft arterial resistance index was calculated as the mean of all available measurements. Urinary tract infection is expressed as the percentage of positive urine tests of the total number of examined urine samples. Transplant lymphocele: any singular or repeated finding. Urological interventions: any interventions (except ureteral stent withdrawal) compared with none. Number of treated rejection episodes includes intravenous steroid boli and increases in the maintenance immunosuppression. Blood pressure is expressed as the average of all available ambulatory measurements. Severe acute diseases include any major surgery and other severe, noninfectious illness, or life-threatening event compared with none. BMI change is defined as current BMI minus BMI at transplantation. All variables, with a P-value of ≤0.05 in the univariable analysis are included in the table. The 95% confidence intervals (95% CIs) are given with the upper and lower boundaries. The Harrel’s concordance index as a measure of model performance was 0.81 and 0.79 after 200-fold bootstrapping. Table 2 HRs of pre- and post-transplant variables at 3 months after transplantation from uni- and multivariable Cox regression analysis   Univariable analysis   Multivariable analysis   HR  95% CI  P-value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.067  1.050 1.085  <0.001  1.051  1.033 1.069  <0.001   Heart failure  5.320  2.855 9.914  <0.001  2.749  1.393 5.424  0.004   DM/PAD/CHD score  1.917  1.595 2.304  <0.001  1.412  1.143 1.745  0.001   Coronary heart disease  2.338  1.606 3.402  <0.001         Peripheral arterial disease  3.419  2.362 4.951  <0.001         Type 2 diabetes  2.560  1.514 4.328  <0.001         Monoclonal gammopathy  2.711  1.000 7.347  0.050         Time on dialysis (years)  1.005  1.001 1.009  0.017        Variables at Tx               BMI  1.049  1.000 1.101  0.050         Deceased donor Tx  3.707  1.634 8.410  0.002  2.640  1.100 6.336  0.030   Tx within the ESP  3.203  2.147 4.779  <0.001         Donor age (years)  1.026  1.014 1.038  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.076 1.320  <0.001          HLA-A mismatches  1.298  1.024 1.645  0.031          HLA-B mismatches  1.321  1.040 1.677  0.022          HLA-DR mismatches  1.549  1.217 1.972  <0.001          CIT (h)  1.045  1.022 1.068  <0.001        Post-Tx until 3 months              Graft factors               DGF  1.887  1.334 2.671  <0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.987  <0.001         eGFR at 3 months post-Tx (mL/min/1.7 m2)  0.971  0.962 0.981  <0.001         Graft loss  1.848  0.457 7.473  0.389         Renal graft arterial resistance index >0.75  2.836  1.963 4.097  <0.001         Urinary tract infection (%)  1.013  1.009 1.018  0.000  1.007  1.002 1.012  0.002   Transplant lymphocele  2.056  1.044 4.050  0.037  2.122  1.055 4.270  0.035   Urological interventions  1.582  1.069 2.340  0.022         Number of treated rejection episodes  1.342  1.122 1.604  0.001  2.055  1.428 2.957  <0.001   BK virus nephropathy  5.172  1.906 14.029  0.01  3.481  1.191 10.175  0.023  Comorbidities               New-onset heart failure  6.026  2.224 16.323  <0.001  5.675  2.047 15.736  0.001   Myocardial infarction, percutaneous or surgical coronary intervention  12.230  5.313 28.153  <0.001  5.815  2.468 13.703  <0.001   Severe acute diseases  3.807  2.285 6.341  <0.001         Severe infections  2.436  1.421 4.176  0.001  1.901  1.072 3.369  0.028   Malignancies  14.908  6.509 34.143  <0.001  19.404  7.892 47.709  <0.001   BMI change since Tx  0.892  0.818 0.973  0.010          Univariable analysis   Multivariable analysis   HR  95% CI  P-value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.067  1.050 1.085  <0.001  1.051  1.033 1.069  <0.001   Heart failure  5.320  2.855 9.914  <0.001  2.749  1.393 5.424  0.004   DM/PAD/CHD score  1.917  1.595 2.304  <0.001  1.412  1.143 1.745  0.001   Coronary heart disease  2.338  1.606 3.402  <0.001         Peripheral arterial disease  3.419  2.362 4.951  <0.001         Type 2 diabetes  2.560  1.514 4.328  <0.001         Monoclonal gammopathy  2.711  1.000 7.347  0.050         Time on dialysis (years)  1.005  1.001 1.009  0.017        Variables at Tx               BMI  1.049  1.000 1.101  0.050         Deceased donor Tx  3.707  1.634 8.410  0.002  2.640  1.100 6.336  0.030   Tx within the ESP  3.203  2.147 4.779  <0.001         Donor age (years)  1.026  1.014 1.038  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.076 1.320  <0.001          HLA-A mismatches  1.298  1.024 1.645  0.031          HLA-B mismatches  1.321  1.040 1.677  0.022          HLA-DR mismatches  1.549  1.217 1.972  <0.001          CIT (h)  1.045  1.022 1.068  <0.001        Post-Tx until 3 months              Graft factors               DGF  1.887  1.334 2.671  <0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.987  <0.001         eGFR at 3 months post-Tx (mL/min/1.7 m2)  0.971  0.962 0.981  <0.001         Graft loss  1.848  0.457 7.473  0.389         Renal graft arterial resistance index >0.75  2.836  1.963 4.097  <0.001         Urinary tract infection (%)  1.013  1.009 1.018  0.000  1.007  1.002 1.012  0.002   Transplant lymphocele  2.056  1.044 4.050  0.037  2.122  1.055 4.270  0.035   Urological interventions  1.582  1.069 2.340  0.022         Number of treated rejection episodes  1.342  1.122 1.604  0.001  2.055  1.428 2.957  <0.001   BK virus nephropathy  5.172  1.906 14.029  0.01  3.481  1.191 10.175  0.023  Comorbidities               New-onset heart failure  6.026  2.224 16.323  <0.001  5.675  2.047 15.736  0.001   Myocardial infarction, percutaneous or surgical coronary intervention  12.230  5.313 28.153  <0.001  5.815  2.468 13.703  <0.001   Severe acute diseases  3.807  2.285 6.341  <0.001         Severe infections  2.436  1.421 4.176  0.001  1.901  1.072 3.369  0.028   Malignancies  14.908  6.509 34.143  <0.001  19.404  7.892 47.709  <0.001   BMI change since Tx  0.892  0.818 0.973  0.010        Heart failure: any grade of insufficiency. Presence of type 2 diabetes, peripheral arterial disease (any grade) and coronary heart disease is additionally expressed as a score (DM/PAD/CHD score), which was later used in the multivariable analyses (1 point for each disease entity, adding up to a discrete score of 0–3). HLA mismatch on locus A, B and DR is entered as a numerical variable as 0, 1 or 2 according to the number of mismatches. eGFR: a value of 5 mL/min/1.7 m2 was assigned to patients with graft loss. Renal graft arterial resistance index was calculated as the mean of all available measurements. Urinary tract infection is expressed as the percentage of positive urine tests of the total number of examined urine samples. Transplant lymphocele: any singular or repeated finding. Urological interventions: any interventions (except ureteral stent withdrawal) compared with none. Number of treated rejection episodes includes intravenous steroid boli and increases in the maintenance immunosuppression. Blood pressure is expressed as the average of all available ambulatory measurements. Severe acute diseases include any major surgery and other severe, noninfectious illness, or life-threatening event compared with none. BMI change is defined as current BMI minus BMI at transplantation. All variables, with a P-value of ≤0.05 in the univariable analysis are included in the table. The 95% confidence intervals (95% CIs) are given with the upper and lower boundaries. The Harrel’s concordance index as a measure of model performance was 0.81 and 0.79 after 200-fold bootstrapping. Table 3 HRs of pre- and post-transplant variables at 12 months after transplantation from uni- and multivariable Cox regression analysis   Univariable analysis   Multivariable analysis   HR  95% CI  P- value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.071  1.053 1.089  <0.001  1.048  1.029 1.068  <0.001   Heart failure  6.022  3.220 11.264  <0.001  3.020  1.570 5.809  0.001   DM/PAD/CHD score  1.999  1.656 2.413  <0.001  1.523  1.218 1.904  <0.001   Coronary heart disease  2.469  1.674 3.640  <0.001         Peripheral arterial disease  3.796  2.595 5.554  <0.001         Type 2 diabetes  2.690  1.562 4.633  <0.001         Monoclonal gammopathy  3.022  1.113 8.203  0.030         Time on dialysis (years)  1.004  1.000 1.008  0.039        Variables at transplantation               BMI  1.054  1.003 1.109  0.038         Deceased donor Tx  3.379  1.486 7.680  0.004         Tx within the ESP  3.357  2.217 5.084  <0.001         Donor age (years)  1.028  1.015 1.041  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.071 1.327  0.001          HLA-A mismatches  1.305  1.019 1.671  0.035          HLA-B mismatches  1.335  1.040 1.714  0.023          HLA-DR mismatches  1.524  1.184 1.961  0.001         CIT (h)  1.039  1.015 1.063  0.001        Post-Tx variables until 12 months              Graft factors               Delayed graft function  1.871  1.301 2.691  0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.988  <0.001         eGFR at 12 months post-Tx (mL/min/1.7 m2)  0.969  0.958 0.979  <0.001  0.987  0.977 0.997  0.011   Graft loss  5.956  3.337 10.630  0.000         Renal graft arterial resistance index >0.75  2.722  1.882 3.937  <0.001         Urinary tract infection (%)  1.027  1.022 1.033  <0.001  1.019  1.012 1.026  <0.001   Urological interventions  1.597  1.079 2.362  0.019         Number of treated rejection episodes  1.422  1.195 1.693  <0.001  1.351  1.102 1.656  0.004   BK virus nephropathy  2.731  1.111 6.710  0.029        Comorbidities               Mean of diastolic blood pressure (mmHg)  0.948  0.922 0.975  <0.001         New-onset heart failure  4.113  1.808 9.364  0.001  5.435  2.294 12.875  <0.001   Myocardial infarction, percutaneous or surgical   coronary intervention  4.934  2.406 10.121  <0.001  3.619  1.706 7.678  <0.001   Severe acute diseases  1.905  1.178 3.080  0.009         Severe infections  2.433  1.607 3.683  <0.001         Malignancies  2.907  1.185 7.131  0.045  3.284  1.317 8.191  0.011   BMI change since Tx  0.827  0.764 0.896  0.000          Univariable analysis   Multivariable analysis   HR  95% CI  P- value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.071  1.053 1.089  <0.001  1.048  1.029 1.068  <0.001   Heart failure  6.022  3.220 11.264  <0.001  3.020  1.570 5.809  0.001   DM/PAD/CHD score  1.999  1.656 2.413  <0.001  1.523  1.218 1.904  <0.001   Coronary heart disease  2.469  1.674 3.640  <0.001         Peripheral arterial disease  3.796  2.595 5.554  <0.001         Type 2 diabetes  2.690  1.562 4.633  <0.001         Monoclonal gammopathy  3.022  1.113 8.203  0.030         Time on dialysis (years)  1.004  1.000 1.008  0.039        Variables at transplantation               BMI  1.054  1.003 1.109  0.038         Deceased donor Tx  3.379  1.486 7.680  0.004         Tx within the ESP  3.357  2.217 5.084  <0.001         Donor age (years)  1.028  1.015 1.041  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.071 1.327  0.001          HLA-A mismatches  1.305  1.019 1.671  0.035          HLA-B mismatches  1.335  1.040 1.714  0.023          HLA-DR mismatches  1.524  1.184 1.961  0.001         CIT (h)  1.039  1.015 1.063  0.001        Post-Tx variables until 12 months              Graft factors               Delayed graft function  1.871  1.301 2.691  0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.988  <0.001         eGFR at 12 months post-Tx (mL/min/1.7 m2)  0.969  0.958 0.979  <0.001  0.987  0.977 0.997  0.011   Graft loss  5.956  3.337 10.630  0.000         Renal graft arterial resistance index >0.75  2.722  1.882 3.937  <0.001         Urinary tract infection (%)  1.027  1.022 1.033  <0.001  1.019  1.012 1.026  <0.001   Urological interventions  1.597  1.079 2.362  0.019         Number of treated rejection episodes  1.422  1.195 1.693  <0.001  1.351  1.102 1.656  0.004   BK virus nephropathy  2.731  1.111 6.710  0.029        Comorbidities               Mean of diastolic blood pressure (mmHg)  0.948  0.922 0.975  <0.001         New-onset heart failure  4.113  1.808 9.364  0.001  5.435  2.294 12.875  <0.001   Myocardial infarction, percutaneous or surgical   coronary intervention  4.934  2.406 10.121  <0.001  3.619  1.706 7.678  <0.001   Severe acute diseases  1.905  1.178 3.080  0.009         Severe infections  2.433  1.607 3.683  <0.001         Malignancies  2.907  1.185 7.131  0.045  3.284  1.317 8.191  0.011   BMI change since Tx  0.827  0.764 0.896  0.000        All variables, with a P-value of  ≤0.05 in the univariable analysis are included in the table. The 95% confidence intervals (95% CIs) are given with the upper and lower boundaries. The Harrel’s concordance index as a measure of model performance was 0.81 and 0.80 after 200-fold bootstrapping. For definition of the clinical terms, see legend to Table 2. Table 3 HRs of pre- and post-transplant variables at 12 months after transplantation from uni- and multivariable Cox regression analysis   Univariable analysis   Multivariable analysis   HR  95% CI  P- value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.071  1.053 1.089  <0.001  1.048  1.029 1.068  <0.001   Heart failure  6.022  3.220 11.264  <0.001  3.020  1.570 5.809  0.001   DM/PAD/CHD score  1.999  1.656 2.413  <0.001  1.523  1.218 1.904  <0.001   Coronary heart disease  2.469  1.674 3.640  <0.001         Peripheral arterial disease  3.796  2.595 5.554  <0.001         Type 2 diabetes  2.690  1.562 4.633  <0.001         Monoclonal gammopathy  3.022  1.113 8.203  0.030         Time on dialysis (years)  1.004  1.000 1.008  0.039        Variables at transplantation               BMI  1.054  1.003 1.109  0.038         Deceased donor Tx  3.379  1.486 7.680  0.004         Tx within the ESP  3.357  2.217 5.084  <0.001         Donor age (years)  1.028  1.015 1.041  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.071 1.327  0.001          HLA-A mismatches  1.305  1.019 1.671  0.035          HLA-B mismatches  1.335  1.040 1.714  0.023          HLA-DR mismatches  1.524  1.184 1.961  0.001         CIT (h)  1.039  1.015 1.063  0.001        Post-Tx variables until 12 months              Graft factors               Delayed graft function  1.871  1.301 2.691  0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.988  <0.001         eGFR at 12 months post-Tx (mL/min/1.7 m2)  0.969  0.958 0.979  <0.001  0.987  0.977 0.997  0.011   Graft loss  5.956  3.337 10.630  0.000         Renal graft arterial resistance index >0.75  2.722  1.882 3.937  <0.001         Urinary tract infection (%)  1.027  1.022 1.033  <0.001  1.019  1.012 1.026  <0.001   Urological interventions  1.597  1.079 2.362  0.019         Number of treated rejection episodes  1.422  1.195 1.693  <0.001  1.351  1.102 1.656  0.004   BK virus nephropathy  2.731  1.111 6.710  0.029        Comorbidities               Mean of diastolic blood pressure (mmHg)  0.948  0.922 0.975  <0.001         New-onset heart failure  4.113  1.808 9.364  0.001  5.435  2.294 12.875  <0.001   Myocardial infarction, percutaneous or surgical   coronary intervention  4.934  2.406 10.121  <0.001  3.619  1.706 7.678  <0.001   Severe acute diseases  1.905  1.178 3.080  0.009         Severe infections  2.433  1.607 3.683  <0.001         Malignancies  2.907  1.185 7.131  0.045  3.284  1.317 8.191  0.011   BMI change since Tx  0.827  0.764 0.896  0.000          Univariable analysis   Multivariable analysis   HR  95% CI  P- value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.071  1.053 1.089  <0.001  1.048  1.029 1.068  <0.001   Heart failure  6.022  3.220 11.264  <0.001  3.020  1.570 5.809  0.001   DM/PAD/CHD score  1.999  1.656 2.413  <0.001  1.523  1.218 1.904  <0.001   Coronary heart disease  2.469  1.674 3.640  <0.001         Peripheral arterial disease  3.796  2.595 5.554  <0.001         Type 2 diabetes  2.690  1.562 4.633  <0.001         Monoclonal gammopathy  3.022  1.113 8.203  0.030         Time on dialysis (years)  1.004  1.000 1.008  0.039        Variables at transplantation               BMI  1.054  1.003 1.109  0.038         Deceased donor Tx  3.379  1.486 7.680  0.004         Tx within the ESP  3.357  2.217 5.084  <0.001         Donor age (years)  1.028  1.015 1.041  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.071 1.327  0.001          HLA-A mismatches  1.305  1.019 1.671  0.035          HLA-B mismatches  1.335  1.040 1.714  0.023          HLA-DR mismatches  1.524  1.184 1.961  0.001         CIT (h)  1.039  1.015 1.063  0.001        Post-Tx variables until 12 months              Graft factors               Delayed graft function  1.871  1.301 2.691  0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.988  <0.001         eGFR at 12 months post-Tx (mL/min/1.7 m2)  0.969  0.958 0.979  <0.001  0.987  0.977 0.997  0.011   Graft loss  5.956  3.337 10.630  0.000         Renal graft arterial resistance index >0.75  2.722  1.882 3.937  <0.001         Urinary tract infection (%)  1.027  1.022 1.033  <0.001  1.019  1.012 1.026  <0.001   Urological interventions  1.597  1.079 2.362  0.019         Number of treated rejection episodes  1.422  1.195 1.693  <0.001  1.351  1.102 1.656  0.004   BK virus nephropathy  2.731  1.111 6.710  0.029        Comorbidities               Mean of diastolic blood pressure (mmHg)  0.948  0.922 0.975  <0.001         New-onset heart failure  4.113  1.808 9.364  0.001  5.435  2.294 12.875  <0.001   Myocardial infarction, percutaneous or surgical   coronary intervention  4.934  2.406 10.121  <0.001  3.619  1.706 7.678  <0.001   Severe acute diseases  1.905  1.178 3.080  0.009         Severe infections  2.433  1.607 3.683  <0.001         Malignancies  2.907  1.185 7.131  0.045  3.284  1.317 8.191  0.011   BMI change since Tx  0.827  0.764 0.896  0.000        All variables, with a P-value of  ≤0.05 in the univariable analysis are included in the table. The 95% confidence intervals (95% CIs) are given with the upper and lower boundaries. The Harrel’s concordance index as a measure of model performance was 0.81 and 0.80 after 200-fold bootstrapping. For definition of the clinical terms, see legend to Table 2. Eight factors were present in both models, including age, pre-transplant heart failure, the score of cardiovascular disease and type 2 diabetes, post-transplant urinary tract infection, treatment of rejection (either in protocol biopsies or biopsies for cause), new-onset heart failure, coronary events and malignancies. Additional variables at 3 months included deceased donor transplantation, transplant lymphocele and BK virus nephropathy. Graft function was a significant factor of the model at 12 months. The Harrell’s concordance index indicated satisfactory performance, with 0.81 for both models, and 0.79 (3 months) and 0.80 (12 months) after 200-fold bootstraps. Validation of the two models was performed on a separate group of 349 patients transplanted between 2008 and 2013, with a follow-up until June 2017. These patients differed in several aspects (Supplementary data, Table S4) from the patient cohort that was used to build the models. Peripheral arterial disease was less and type 2 diabetes more prevalent. More patients received living-donor transplantations. CIT was shorter and DGF and urological interventions were less common. All patients had mycophenolate mofetil and a higher proportion received tacrolimus instead of cyclosporine. Less rejection treatments were performed (0.46 ±0.81 versus 0.63 ±0.89/patient). The validation analysis showed satisfactory discrimination of patient risks for death (Figure 2), with a concordance index of 0.73 for the model at 3 months and 0.76 for the model at 12 months. Estimators of both models are graphically depicted by a nomogram, which is intended as a tool to weigh individual risks that may contribute to decreased survival (Figure 3). FIGURE 2 View largeDownload slide Survival probability in the separate validation group of 349 patients, based on the prediction of the multivariable models at 3 months (A) and at 12 months (B) post-transplantation. Kaplan Meier curves are shown for four discrete risks groups (‘very low risk’ up to ‘high risk’ for death) obtained from the linear predictor, using cut-points on the prognostic index determined by Cox’s method (cut points: 16th, 50th and 84th percentiles of the prognostic index). The boundaries of the prognostic index were: very low risk; −2.83 to −1.30, low risk; −1.31 to −0.23, moderate risk; −0.24 to 0.81, high risk; 0.82–2.69 (3 months); and very low risk; −2.46 to −1.14, low risk; −1.15 to −0.14, moderate risk; −0.15 to 0.90, high risk; 0.91–2.93 (12 months). Number of patients observed at each time is shown above the abscissa. FIGURE 2 View largeDownload slide Survival probability in the separate validation group of 349 patients, based on the prediction of the multivariable models at 3 months (A) and at 12 months (B) post-transplantation. Kaplan Meier curves are shown for four discrete risks groups (‘very low risk’ up to ‘high risk’ for death) obtained from the linear predictor, using cut-points on the prognostic index determined by Cox’s method (cut points: 16th, 50th and 84th percentiles of the prognostic index). The boundaries of the prognostic index were: very low risk; −2.83 to −1.30, low risk; −1.31 to −0.23, moderate risk; −0.24 to 0.81, high risk; 0.82–2.69 (3 months); and very low risk; −2.46 to −1.14, low risk; −1.15 to −0.14, moderate risk; −0.15 to 0.90, high risk; 0.91–2.93 (12 months). Number of patients observed at each time is shown above the abscissa. FIGURE 3 View largeDownload slide Nomograms to estimate patient’s individual risks of reduced survival at 3 months (A) and 12 months post-transplantation (B). For each factor, the individual value is located on the corresponding scale. By drawing a vertical line from this position to the most upper scale (‘points’), the number of points is read from this scale. The number of points for each variable is added up to give the total number of points. On the scale for total points, the patient’s total points are located, and a vertical line is drawn to the survival scale which indicates the effect of the individual risk factors on the 10-year survival. DM/PAD/CHD score: a value of 1 for each condition at Tx (type 2 diabetes, any grade of peripheral arterial disease, coronary heart disease). Heart failure: any grade of heart failure. Percentage of urinary tract infections is calculated by relating the number of positive urine tests to the total number of examined urine samples. New coronary event: myocardial infarction, percutaneous or surgical coronary intervention. Rejection treatments: number of treated episodes. Tx, transplantation. FIGURE 3 View largeDownload slide Nomograms to estimate patient’s individual risks of reduced survival at 3 months (A) and 12 months post-transplantation (B). For each factor, the individual value is located on the corresponding scale. By drawing a vertical line from this position to the most upper scale (‘points’), the number of points is read from this scale. The number of points for each variable is added up to give the total number of points. On the scale for total points, the patient’s total points are located, and a vertical line is drawn to the survival scale which indicates the effect of the individual risk factors on the 10-year survival. DM/PAD/CHD score: a value of 1 for each condition at Tx (type 2 diabetes, any grade of peripheral arterial disease, coronary heart disease). Heart failure: any grade of heart failure. Percentage of urinary tract infections is calculated by relating the number of positive urine tests to the total number of examined urine samples. New coronary event: myocardial infarction, percutaneous or surgical coronary intervention. Rejection treatments: number of treated episodes. Tx, transplantation. DISCUSSION In this retrospective study of patients from our large kidney protocol biopsy programme, we confirm established risk factors for death like age, heart failure, type 2 diabetes, cardiovascular disease and graft function. Novel factors include post-transplant urinary tract infections and BK virus nephropathy. Unlike most previous studies, our analysis included patients with graft loss. We believe that this integral view is best suited to a rational and patient-centred risk assessment, since graft loss is not uninformative regarding death. Further, study of patients with protocol biopsies allowed evaluating treatment of subclinical rejections in the analyses. Possible limitations include the large proportion of unknown causes of death that precluded in-depth subanalyses for different causes. Regarding the generalizability of our results, based on the pre-transplant data, we believe that our cohort of approximately 900 patients is well comparable to patients from many other transplant programmes that treat mainly Caucasian patients. Strengths of our single-centre study include consistent data collection with a high degree of completeness and a low number of lost to follow-up cases within up to 11 years. In addition, we aimed to employ robust definitions for each variable and to retain the highest possible informational value of continuous predictors by avoiding categorization of these variables [8]. The obtained multivariable models were robust as confirmed by bootstrapping analysis and by the discriminatory performance in an independent validation cohort. Patient survival was 89% after 5 years and 78% after 10 years, similar to figures from the ERA-EDTA registry [9]. Yet, it has to be noted that patient survival reported in this study included patients with graft loss which have an excess risk for death [10]. Similar to the International Pancreas Transplant Registry [11], we observed favourable survival results in the highly selected patient group with combined pancreas/kidney transplants. One- to four-year results on the ESP showed similar survival compared with recipients of the same age who received organs from younger donors [12, 13]. Nevertheless, the benefit for older patients receiving organs from extended criteria donors is controversial [14, 15]. Our data do not indicate a significant survival difference (P = 0.2) between transplantations within the ESP compared with transplantations with younger deceased donors. In recent registry studies from the UK [16] and from Australia/New Zealand [17], cardiovascular disease accounted for 23% and 16% of the deaths, respectively, compared with our rate of 14%. However, the cause of death was unavailable in 42% in our data so that the true number might be higher. Rates of infectious deaths were very similar, with 23% in our centre and 22–23% in the aforementioned studies. We did not observe relevant time effects among different causes of death like Mazuecos et al. [18], who reported higher rates of infectious death in the first transplant year and increasing malignancy-related deaths in later years. Similar to recent studies [19–23], age, pre-transplant type 2 diabetes, coronary and peripheral arterial disease and heart failure were important predictors of death. In our study, pre-transplant heart failure was prevalent with 2%, with similar proportions of New York Heart Association Grades I and II. The degree of pre-transplant heart failure was not predictive of death (data not shown), reflecting the difficulty of assessing the true severity of heart failure in the pre-transplant setting with reduced hydration status and of predicting the effect of hydration after transplantation in these patients. New-onset heart failure after transplantation was a strong risk factor, with three deaths among 10 patients with Grade II, two deaths in two patients with Grade III, four deaths in the nine patients with Grade IV and no deaths in four patients with Grade I. The importance of post-transplant coronary events has been highlighted by a previous study [23]. Newly discovered malignancy was an important risk factor, including one case each with carcinoma of the prostate, kidney, vulva and oesophagus, one invasive squamous skin cancer, and one Kaposi sarcoma in the group of deceased patients. Excluding the patient with Kaposi sarcoma, it remains open whether these malignancies had been present pre-transplant and missed by the regular waitlist examinations. A previous study in nearly 5000 kidney transplant recipients reported an even higher rate of 12.7% malignancies within the first transplant year [24], compared with 2% in our study. In a large registry analysis from the UK, malignancy-related death accounted for 7.4% of deaths within the first transplant year, and two-thirds of these patients had a pre-transplant history of malignancy [25]. Reduced graft function is an established risk factor for death [23, 26, 27]. In our study, the eGFR at 12 months was linearly correlated with the risk of death, including patients with graft loss (which were assigned to an eGFR of 5 mL/min/1.7 m2). Graft function was not a risk factor in the model at 3 months probably because graft function was more variable at this time due to ischaemia/reperfusion injury, toxicity of the medication and other factors. In fact, the model at 3 months included factors with known effects on graft function, namely deceased donor transplantations, BK virus nephropathy, rejection and urinary tract infection. The detrimental effect of BK virus nephropathy on renal graft survival is well established [28]. We are not aware of studies linking BK virus replication or nephropathy to decreased patient survival in kidney transplant recipients. A large study in allogeneic hematopoietic stem cell recipients identified BK virus replication as an independent risk factor for death. It remains open whether BK virus replication directly affected survival or was a surrogate of other risks, e.g. stronger suppression of the host’s immune system. Yet, we did not find a higher proportion of deaths due to infection in patients with BK virus nephropathy. The negative impact of early acute rejection—even if subclinical—on graft function and graft survival has been demonstrated by recent studies, without studying effects on patient survival [29, 30] or with reporting lacking effect on patient survival [31, 32]. In one study [20], acute rejection was associated with death. Treated rejection episodes—either in protocol biopsies or biopsies for cause—were a significant factor in our models. Yet, rejection treatments were not specifically related to infectious or malignancy-related deaths in univariable analyses, with an average of 0.75 and 0.94 treatments/patient with infectious and malignancy-associated death compared with 1.1 treatments/patient with cardiovascular death. Most likely, acute rejection represented therefore a predictor of the future course of graft function and loss. Severe infections were a significant factor in the 3 months’ model. In recipients of kidney, liver or heart transplants, post-transplant infections were significantly associated with death, and urinary tract infections accounted for one-third of all infections [33]. Only one study specifically addressed urinary tract infection in kidney transplant patients and found an association with death [34]. In our study, we established a linear relationship between the frequency of urinary tract infections and death. Patients with infectious or malignancy-related deaths did not have a higher rate of urinary tract infections than patients dying from cardiovascular disease (0.16 and 0.26 versus 0.20, respectively). Previous studies have attempted to identify and integrate risk factors for death into predictive models, including the pre-transplant variables age, gender, race, body mass index (BMI), time on dialysis, cause of end-stage renal disease, panel reactive antibodies, HLA mismatches, comorbidities such as diabetes, cardiovascular disease and heart failure, and donor age. In some models, the post-transplant factors DGF, acute rejection and graft function were included [35–42]. A model performance with an area under the curve of 0.63 after validation with a separate patient cohort was reported by using age, pre-transplant coronary heart disease, left ventricular hypertrophy, vascular calcification, diabetes, time on dialysis, and acute tubular necrosis and GFR at discharge after transplantation as factors [36]. Established indices with numerous comorbidities were tested without validation on independent patient groups, reporting a c-statistics value of 0.71 [39] or no performance measures [38]. Another model to predict death until 3 years post-transplantation included age, weight, time on dialysis, diabetes, hepatitis C at transplantation, DGF, and diabetes, proteinuria, renal function and immunosuppressants within the first year, with a c-statistics value of 0.74 on validation in an independent patient cohort [37]. In patients aged >65 years, a model with age, gender, race, post-transplant smoking, eGFR, albumin excretion, diabetes, heart failure and stroke predicted 5-year mortality risk with a c-statistics value of 0.69 on validation with an independent patient cohort [42]. In none of these studies was the fate of patients with graft loss and return to dialysis considered in the prediction. Our results emphasize the importance of pre- and post-transplant monitoring for cardiovascular disease, heart failure and malignancies. Further, preservation of renal graft function, particularly avoidance of acute rejection and BK nephropathy appear to be important treatment goals. The proposed models with a comparatively high predictive performance may help to estimate and balance existing risks for death in individual patients better. Although some risks may not be modifiable or may have undetermined causality, such risk assessment can help to identify individually important areas that need particular attention in the pre-transplant preparation and post-transplant care. Further validation of the models is required, preferentially in cohorts from other transplant centres. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS The help of A. Henkel and B. Boes in the collection of the data is greatly appreciated. This study was supported by intramural funding. AUTHORS’ CONTRIBUTIONS T.A. and W.G. were responsible for the research design, writing and data analysis. I.S. and A. Karch were responsible for the data analysis, research design and writing. V.B. was responsible for the research design and writing. A. Koch was responsible for the data analysis and research design. H.H. and A.S. were responsible for the research design and writing. CONFLICT OF INTEREST STATEMENT None declared. The results presented in this paper have not been published previously in whole or part, except in abstract format. REFERENCES 1 Foley RN, Parfrey PS, Sarnak MJ. Clinical epidemiology of cardiovascular disease in chronic renal disease. Am J Kidney Dis  1998; 32: S112– S119 Google Scholar CrossRef Search ADS PubMed  2 Wolfe RA, Ashby VB, Milford EL. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med  1999; 341: 1725– 1730 Google Scholar CrossRef Search ADS PubMed  3 Zsom L, Wagner L, Fulop T. Minimization vs tailoring: where do we stand with personalized immunosuppression during renal transplantation in 2015? World J Transplant  2015; 5: 73– 80 Google Scholar CrossRef Search ADS PubMed  4 R Development Core Team. R R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2015. www.r-project.org 5 Kaiser J. Dealing with missing values in data. J Syst Integr  2014; 5: 42– 51 Google Scholar CrossRef Search ADS   6 Cox DR. Note on grouping. J Am Stat Assoc  1957; 52: 543– 547 Google Scholar CrossRef Search ADS   7 Smits JM, Persijn GG, van Houwelingen HC et al.   Evaluation of the Eurotransplant Senior Program: the results of the first year. Am J Transplant  2002; 2: 664– 670 Google Scholar CrossRef Search ADS PubMed  8 Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors in multiple regression: a bad idea. Statist Med  2006; 25: 127– 141 Google Scholar CrossRef Search ADS   9 Pippias M, Stel VS, Abad Diez JM et al.   Renal replacement therapy in Europe: a summary of the 2012 ERA-EDTA Registry Annual Report. Clin Kidney J  2015; 8: 248– 261 Google Scholar CrossRef Search ADS PubMed  10 Brar A, Markell M, Stefanov DG et al.   Mortality after renal allograft failure and return to dialysis. Am J Nephrol  2017; 45: 180– 186 Google Scholar CrossRef Search ADS PubMed  11 Gruessner AC. 2011 update on pancreas transplantation: comprehensive trend analysis of 25, 000 cases followed up over the course of twenty-four years at the International Pancreas Transplant Registry (IPTR). Rev Diabet Stud  2011; 8: 6– 16 Google Scholar CrossRef Search ADS PubMed  12 Bentas W, Jones J, Karaoguz A et al.   Renal transplantation in the elderly: surgical complications and outcome with special emphasis on the Eurotransplant Senior Programme. Nephrol Dial Transplant  2008; 23: 2043– 2051 Google Scholar CrossRef Search ADS PubMed  13 Fabrizii V, Kovarik J, Bodingbauer M et al.   Long-term patient and graft survival in the eurotransplant senior program: a single-center experience. Transplantation  2005; 80: 582– 589 Google Scholar CrossRef Search ADS PubMed  14 Veroux M, Grosso G, Corona D et al.   Age is an important predictor of kidney transplantation outcome. Nephrol Dial Transplant  2012; 27: 1663– 1671 Google Scholar CrossRef Search ADS PubMed  15 Merion RM, Ashby VB, Wolfe RA et al.   Deceased-donor characteristics and the survival benefit of kidney transplantation. JAMA  2005; 294: 2726– 2733 Google Scholar CrossRef Search ADS PubMed  16 Steenkamp R, Castledine C, Feest T. Chapter 6 Survival and causes of death of UK adult patients on renal replacement therapy in 2010: national and centre-specific analyses. Nephron Clin Pract  2012; 120 (Suppl. 1): c105– c135 Google Scholar CrossRef Search ADS PubMed  17 Pilmore H, Dent H, Chang S et al.   Reduction in cardiovascular death after kidney transplantation. Transplantation  2010; 89: 851– 857 Google Scholar CrossRef Search ADS PubMed  18 Mazuecos A, Munoz Terol JM, Garcia Alvarez T et al.   Increase in malignancies as cause of death in renal transplant patients. Transplant Proc  2009; 41: 2159– 2162 Google Scholar CrossRef Search ADS PubMed  19 Karim A, Farrugia D, Cheshire J et al.   Recipient age and risk for mortality after kidney transplantation in England. Transplantation  2014; 97: 832– 838 Google Scholar CrossRef Search ADS PubMed  20 Faravardeh A, Eickhoff M, Jackson S et al.   Predictors of graft failure and death in elderly kidney transplant recipients. Transplantation  2013; 96: 1089– 1096 Google Scholar CrossRef Search ADS PubMed  21 Hayer MK, Farrugia D, Begaj I et al.   Infection-related mortality is higher for kidney allograft recipients with pretransplant diabetes mellitus. Diabetologia  2014; 57: 554– 561 Google Scholar CrossRef Search ADS PubMed  22 Brar A, Jindal RM, Elster EA et al.   Effect of peripheral vascular disease on kidney allograft outcomes: a study of U.S. Renal data system. Transplantation  2013; 95: 810– 815 Google Scholar CrossRef Search ADS PubMed  23 Kahwaji J, Bunnapradist S, Hsu JW et al.   Cause of death with graft function among renal transplant recipients in an integrated healthcare system. Transplantation  2011; 91: 225– 230 Google Scholar CrossRef Search ADS PubMed  24 Cheung CY, Lam MF, Chu KH et al.   Malignancies after kidney transplantation: Hong Kong renal registry. Am J Transplant  2012; 12: 3039– 3046 Google Scholar CrossRef Search ADS PubMed  25 Farrugia D, Cheshire J, Begaj I et al.   Death within the first year after kidney transplantation: an observational cohort study. Transpl Int  2014; 27: 262– 270 Google Scholar CrossRef Search ADS PubMed  26 Moranne O, Maillard N, Fafin C et al.   Rate of renal graft function decline after one year is a strong predictor of all-cause mortality. Am J Transplant  2013; 13: 695– 706 Google Scholar CrossRef Search ADS PubMed  27 Soveri I, Holdaas H, Jardine A et al.   Renal transplant dysfunction–importance quantified in comparison with traditional risk factors for cardiovascular disease and mortality. Nephrol Dial Transplant  2006; 21: 2282– 2289 Google Scholar CrossRef Search ADS PubMed  28 Dharnidharka VR, Cherikh WS, Abbott KC. An OPTN analysis of national registry data on treatment of BK virus allograft nephropathy in the United States. Transplantation  2009; 87: 1019– 1026 Google Scholar CrossRef Search ADS PubMed  29 Loupy A, Vernerey D, Tinel C et al.   Subclinical rejection phenotypes at 1 year post-transplant and outcome of kidney allografts. J Am Soc Nephrol  2015; 26: 1721– 1731 Google Scholar CrossRef Search ADS PubMed  30 Koo EH, Jang HR, Lee JE et al.   The impact of early and late acute rejection on graft survival in renal transplantation. Kidney Res Clin Pract  2015; 34: 160– 164 Google Scholar CrossRef Search ADS PubMed  31 Wu K, Budde K, Lu H et al.   The severity of acute cellular rejection defined by Banff classification is associated with kidney allograft outcomes. Transplantation  2014; 97: 1146– 1154 Google Scholar CrossRef Search ADS PubMed  32 Krisl JC, Alloway RR, Shield AR et al.   Acute rejection clinically defined phenotypes correlate with long-term renal allograft survival. Transplantation  2015; 99: 2167– 2173 Google Scholar CrossRef Search ADS PubMed  33 Cervera C, Fernandez-Ruiz M, Valledor A et al.   Epidemiology and risk factors for late infection in solid organ transplant recipients. Transpl Infect Dis  2011; 13: 598– 607 Google Scholar CrossRef Search ADS PubMed  34 Abbott KC, Swanson SJ, Richter ER et al.   Late urinary tract infection after renal transplantation in the United States. Am J Kidney Dis  2004; 44: 353– 362 Google Scholar CrossRef Search ADS PubMed  35 Swanson SJ, Hypolite IO, Agodoa LY et al.   Effect of donor factors on early graft survival in adult cadaveric renal transplantation. Am J Transplant  2002; 2: 68– 75 Google Scholar CrossRef Search ADS PubMed  36 Hernandez D, Rufino M, Bartolomei S et al.   A novel prognostic index for mortality in renal transplant recipients after hospitalization. Transplantation  2005; 79: 337– 343 Google Scholar CrossRef Search ADS PubMed  37 Hernandez D, Sanchez-Fructuoso A, Gonzalez-Posada JM et al.   A novel risk score for mortality in renal transplant recipients beyond the first posttransplant year. Transplantation  2009; 88: 803– 809 Google Scholar CrossRef Search ADS PubMed  38 Jassal SV, Schaubel DE, Fenton SS. Predicting mortality after kidney transplantation: a clinical tool. Transplant Int  2005; 18: 1248– 1257 Google Scholar CrossRef Search ADS   39 Machnicki G, Pinsky B, Takemoto S et al.   Predictive ability of pretransplant comorbidities to predict long-term graft loss and death. Am J Transplant  2009; 9: 494– 505 Google Scholar CrossRef Search ADS PubMed  40 Ojo AO, Hanson JA, Wolfe RA et al.   Long-term survival in renal transplant recipients with graft function. Kidney Int  2000; 57: 307– 313 Google Scholar CrossRef Search ADS PubMed  41 Schaeffner ES, Fodinger M, Kramar R et al.   Prognostic associations between lipid markers and outcomes in kidney transplant recipients. Am J Kidney Dis  2006; 47: 509– 517 Google Scholar CrossRef Search ADS PubMed  42 Bansal N, Katz R, De Boer IH et al.   Development and validation of a model to predict 5-year risk of death without ESRD among older adults with CKD. Clin J Am Soc Nephrol  2015; 10: 363– 371 Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nephrology Dialysis Transplantation Oxford University Press

Risk factors for death in kidney transplant patients: analysis from a large protocol biopsy registry

Loading next page...
 
/lp/ou_press/risk-factors-for-death-in-kidney-transplant-patients-analysis-from-a-DRiC71nr28
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
ISSN
0931-0509
eISSN
1460-2385
D.O.I.
10.1093/ndt/gfy131
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Identification and quantification of the relevant factors for death can improve patients’ individual risk assessment and decision-making. We used a well-documented patient cohort (n = 892) in a renal transplant programme with protocol biopsies to establish multivariable Cox models for risk assessment at 3 and 12 months post-transplantation. Methods Patients transplanted between 2000 and 2007 were observed up to 11 years (total observation 5227 patient-years; median 5.9 years). Loss to follow-up was negligible (n = 15). A total of 2251 protocol biopsies and 1214 biopsies for cause were performed. All rejections and clinical borderline rejections in protocol biopsies were treated. Results Overall 10-year patient survival was 78%, with inferior survival of patients with graft loss and superior survival of patients with living-donor transplantation. Eight factors were common in the models at 3 and 12 months, including age, pre-transplant heart failure and a score of cardiovascular disease and type 2 diabetes, post-transplant urinary tract infection, treatment of rejection, new-onset heart failure, coronary events and malignancies. Additional variables of the model at 3 months included deceased donor transplantation, transplant lymphocele, BK virus nephropathy and severe infections. Graft function and graft loss were significant factors of the model at 12 months. Internal validation and validation with a separate cohort of patients (n = 349) demonstrated good discrimination of the models. Conclusions The identified factors indicate the important areas that need special attention in the pre- and post-transplant care of renal transplant patients. On the basis of these models, we provide nomograms as a tool to weigh individual risks that may contribute to decreased survival. kidney transplantation, protocol biopsies, risk factors, survival INTRODUCTION Patients with end-stage renal disease have a significant excess mortality compared with the general population. A high prevalence of cardiovascular disease is one of the major comorbidity factors for this increased mortality [1]. Kidney transplantation can lower the mortality in the long term [2]. However, in addition to the peri-operative risks, new hazards can develop including infections, malignancies, metabolic disorders and hypertension [3]. Consideration of individual pre- and post-transplant risks can help to tailor patient’s specific diagnostic and therapeutic requirements. In addition, quantification of risks by multivariable decision models may be a supportive tool for balancing the risks and benefits of potentially harmful treatments for individual patients. The latter is of special importance with regard to the increasing implementation of protocol biopsies and treatment of subclinical rejections. The aim of our study is to establish robust decision models that help in estimating the risks stemming from pre- and post-transplant factors. Therefore, we used a well-documented patient cohort of ∼900 patients in a specialized post-transplant care programme with protocol biopsies and a long-term follow-up of up to 11 years. Unlike previous studies, assessment of the risks for death included patients with graft loss and return to dialysis. MATERIALS AND METHODS Patients In this retrospective cohort study, 892 adult patients were included who received a kidney transplant alone or in combination with another solid organ at Hannover Medical School between 2000 and 2007 and who participated in our protocol biopsy programme. Protocol biopsies were performed 6 weeks, 3 and 6 months after transplantation. Data were collected prior to and at the time of transplantation, at the time points of protocol biopsies and any additional biopsies, and in yearly intervals after transplantation (complete list of variables in Supplementary data, Table S1). For patients who were followed-up elsewhere, data were retrieved by contacting their local caregivers. The total observation time was 5227 patient-years (25/50/75 percentile: 4.0, 5.9 and 8.0 years). Fifteen patients were lost to follow-up at a median of 3.0 years (minimum observation time 1.9 years) after transplantation. These were included in the modelling and censored at their last observation time. A separate cohort of 349 patients transplanted in 2008–13 and followed up until June 2017 was used to validate the obtained results. Data collection and analysis were performed with informed consent of the patients and with approval of the ethic board (no. 2765) of the Hannover Medical School. Pre-transplant and peri-operative data are depicted in Table 1. Combined organ transplantations included kidney with pancreas (n = 65), liver (n = 4), heart (n = 7) and lung (n = 5). The estimated glomerular filtration rate (eGFR; mL/min/1.7 m2) was calculated with the Cockcroft–Gault formula. Urinary tract infection was defined by leucocyturia and a urine culture with >104 bacterial or >102 fungal colonies. Urological interventions excluded removal of ureteral stents that had been implanted during transplantation. Death is defined as all-cause mortality with functioning graft or after graft loss. Table 1 Pre-transplant and peri-operative data of the whole patient group and of surviving and deceased patients   All patients, n = 892  Surviving patients, n = 760  Deceased patients, n = 132  P-value  Pre-transplant data  Recipient          Age (years)  50 ± 13  48 ± 13  59 ± 11  <0.001  Gender (%; male/female)  59/41  59/41  61/39  0.774  Main reason for endstage renal disease (%)        0.056   Glomerulonephritis/vasculitis  26  27  17     Tubulointerstitial diseases  9  9  8     Hypertensive/diabetic nephropathy  14  13  18     Congenital disease  15  15  14     Other known causes  4  4  2     Unknown  33  31  40    Comorbidity before and at Tx (%)           Arterial hypertension  96  96  97  0.806   Heart failure  2  1  8  <0.001   Coronary heart disease  16  14  31  <0.001    History of myocardial infarction  5  4  11  0.002    History of coronary bypass surgery  6  5  13  0.003    History of coronary angioplasty  5  4  12  0.001   Peripheral arterial disease  13  10  32  <0.001   Ischaemic cerebral stroke  4  3  6  0.116   Replicative hepatitis B  1  1  1  1.000   Replicative hepatitis C  6  7  4  0.246   Type I diabetes  8  9  8  1.000   Type II diabetes  7  5  13  0.003   Hypercholesterolaemia  60  60  62  0.908   Hyperparathyroidism  67  68  62  0.429   Parathyroidectomy before Tx  22  23  17  0.091   CMV IgG positivity  58  58  61  0.502   EBV IgG positivity  91  91  89  0.501   Malignancies (excluding non-melanotic skin cancer)  6  6  7  0.689   Monoclonal gammopathy  1  1  3  0.087   Previous cytotoxic therapies  2  2  2  1.000   History of cigarette smoking  41  41  38  0.396   Time on dialysis before Tx (months)  70 ± 41  69 ± 42  78 ± 33  0.009    Deceased donor recipients  77 ± 38  77 ± 39  80 ± 31  0.474    Living donor recipients  29 ± 33  28 ± 33  37 ± 31  0.366   Re-transplanted patients  13  13  11  0.579   Body weight at Tx (kg)  72 ± 13  72 ± 13  73 ± 12  0.356   BMI at Tx  24 ± 4  24 ± 4  25 ± 3  0.063  Peri-operative data  Transplant-related factors           ESP (%)  10  8  24  <0.001   Combined kidney/pancreas Tx (%)  7  8  6  0.717   Pre-formed antibodies >0% (%)  7  7  8  0.854   Donor age (years)  49 ± 16  48 ± 15  53 ± 17  <0.001   Donor gender (%; male/female)  52/48  53/47  48/52  0.343   Donor type (%; deceased/living)  85/15  83/17  95/5  <0.001   Donor CMV IgG positivity (%)  58  58  59  0.924   Donor S-creatinine (µmol/L)  83 ± 46  84 ± 48  81 ± 34  0.890   Mean number of HLA mismatches  2.4 ± 1.7  2.3 ± 1.7  2.8 ± 1.8  0.005    HLA-A mismatches  0.71 ± 0.71  0.69 ± 0.69  0.81 ± 0.78  0.127    HLA-B mismatches  0.91 ± 0.73  0.89 ± 0.72  1.03 ± 0.75  0.048    HLA-DR mismatches  0.78 ± 0.70  0.75 ± 0.69  0.96 ± 0.70  0.001   CIT (h)  14.5 ± 7.4  14.0 ± 7.4  17.3 ± 6.8  <0.001    Deceased donor recipients  16.7 ± 5.9  16.4 ± 5.8  18.0 ± 6.1  0.004    Living donor recipients  2.5 ± 0.9  2.5 ± 0.9  2.7 ± 0.5  0.208  Initial immunosupressive therapy (%)           Induction therapy        0.660    Anti-lymphocyte globulin  9  9  11      IL-2 antibodies  79  79  80      None  10  10  7      Other/unknown  1/1  1/1  2/2     Cyclosporine A  75  75  74  0.664   Tacrolimus  19  20  13  0.055   MMF  66  67  61  0.197   Azathioprine  1  1  0  0.602   Everolimus or rapamycin  5  4  6  0.360   Steroids  95  96  95  0.653   Study drugs  9  9  9  1.000  Early post-Tx course           DGF (%)  29  27  43  <0.001   Best eGFR in the first 6 weeks post-Tx (mL/min/1.7m2)  59 ± 24  61 ± 24  49 ± 22  <0.001    All patients, n = 892  Surviving patients, n = 760  Deceased patients, n = 132  P-value  Pre-transplant data  Recipient          Age (years)  50 ± 13  48 ± 13  59 ± 11  <0.001  Gender (%; male/female)  59/41  59/41  61/39  0.774  Main reason for endstage renal disease (%)        0.056   Glomerulonephritis/vasculitis  26  27  17     Tubulointerstitial diseases  9  9  8     Hypertensive/diabetic nephropathy  14  13  18     Congenital disease  15  15  14     Other known causes  4  4  2     Unknown  33  31  40    Comorbidity before and at Tx (%)           Arterial hypertension  96  96  97  0.806   Heart failure  2  1  8  <0.001   Coronary heart disease  16  14  31  <0.001    History of myocardial infarction  5  4  11  0.002    History of coronary bypass surgery  6  5  13  0.003    History of coronary angioplasty  5  4  12  0.001   Peripheral arterial disease  13  10  32  <0.001   Ischaemic cerebral stroke  4  3  6  0.116   Replicative hepatitis B  1  1  1  1.000   Replicative hepatitis C  6  7  4  0.246   Type I diabetes  8  9  8  1.000   Type II diabetes  7  5  13  0.003   Hypercholesterolaemia  60  60  62  0.908   Hyperparathyroidism  67  68  62  0.429   Parathyroidectomy before Tx  22  23  17  0.091   CMV IgG positivity  58  58  61  0.502   EBV IgG positivity  91  91  89  0.501   Malignancies (excluding non-melanotic skin cancer)  6  6  7  0.689   Monoclonal gammopathy  1  1  3  0.087   Previous cytotoxic therapies  2  2  2  1.000   History of cigarette smoking  41  41  38  0.396   Time on dialysis before Tx (months)  70 ± 41  69 ± 42  78 ± 33  0.009    Deceased donor recipients  77 ± 38  77 ± 39  80 ± 31  0.474    Living donor recipients  29 ± 33  28 ± 33  37 ± 31  0.366   Re-transplanted patients  13  13  11  0.579   Body weight at Tx (kg)  72 ± 13  72 ± 13  73 ± 12  0.356   BMI at Tx  24 ± 4  24 ± 4  25 ± 3  0.063  Peri-operative data  Transplant-related factors           ESP (%)  10  8  24  <0.001   Combined kidney/pancreas Tx (%)  7  8  6  0.717   Pre-formed antibodies >0% (%)  7  7  8  0.854   Donor age (years)  49 ± 16  48 ± 15  53 ± 17  <0.001   Donor gender (%; male/female)  52/48  53/47  48/52  0.343   Donor type (%; deceased/living)  85/15  83/17  95/5  <0.001   Donor CMV IgG positivity (%)  58  58  59  0.924   Donor S-creatinine (µmol/L)  83 ± 46  84 ± 48  81 ± 34  0.890   Mean number of HLA mismatches  2.4 ± 1.7  2.3 ± 1.7  2.8 ± 1.8  0.005    HLA-A mismatches  0.71 ± 0.71  0.69 ± 0.69  0.81 ± 0.78  0.127    HLA-B mismatches  0.91 ± 0.73  0.89 ± 0.72  1.03 ± 0.75  0.048    HLA-DR mismatches  0.78 ± 0.70  0.75 ± 0.69  0.96 ± 0.70  0.001   CIT (h)  14.5 ± 7.4  14.0 ± 7.4  17.3 ± 6.8  <0.001    Deceased donor recipients  16.7 ± 5.9  16.4 ± 5.8  18.0 ± 6.1  0.004    Living donor recipients  2.5 ± 0.9  2.5 ± 0.9  2.7 ± 0.5  0.208  Initial immunosupressive therapy (%)           Induction therapy        0.660    Anti-lymphocyte globulin  9  9  11      IL-2 antibodies  79  79  80      None  10  10  7      Other/unknown  1/1  1/1  2/2     Cyclosporine A  75  75  74  0.664   Tacrolimus  19  20  13  0.055   MMF  66  67  61  0.197   Azathioprine  1  1  0  0.602   Everolimus or rapamycin  5  4  6  0.360   Steroids  95  96  95  0.653   Study drugs  9  9  9  1.000  Early post-Tx course           DGF (%)  29  27  43  <0.001   Best eGFR in the first 6 weeks post-Tx (mL/min/1.7m2)  59 ± 24  61 ± 24  49 ± 22  <0.001  Fifteen of the 892 patients were lost to follow-up during the post-transplant course. Tx, transplantation; CMV, cytomegalovirus; EBV, Epstein–Barr virus. Heart failure: any degree of insufficiency. Hyperparathyroidism is defined as an elevation of the most recent parathormone value within the year before transplantation by at least 2-fold of the upper normal. Pre-formed antibodies were determined by the lymphocytotoxic panel reactive antibody test. Serum creatinine of the donor represents the last known value before organ explantation and was available in 371 of the cases. DGF was defined as <500 mL urine within the first 24 h after transplantation and/or need of dialysis within the first week. Means are given with SD. Table 1 Pre-transplant and peri-operative data of the whole patient group and of surviving and deceased patients   All patients, n = 892  Surviving patients, n = 760  Deceased patients, n = 132  P-value  Pre-transplant data  Recipient          Age (years)  50 ± 13  48 ± 13  59 ± 11  <0.001  Gender (%; male/female)  59/41  59/41  61/39  0.774  Main reason for endstage renal disease (%)        0.056   Glomerulonephritis/vasculitis  26  27  17     Tubulointerstitial diseases  9  9  8     Hypertensive/diabetic nephropathy  14  13  18     Congenital disease  15  15  14     Other known causes  4  4  2     Unknown  33  31  40    Comorbidity before and at Tx (%)           Arterial hypertension  96  96  97  0.806   Heart failure  2  1  8  <0.001   Coronary heart disease  16  14  31  <0.001    History of myocardial infarction  5  4  11  0.002    History of coronary bypass surgery  6  5  13  0.003    History of coronary angioplasty  5  4  12  0.001   Peripheral arterial disease  13  10  32  <0.001   Ischaemic cerebral stroke  4  3  6  0.116   Replicative hepatitis B  1  1  1  1.000   Replicative hepatitis C  6  7  4  0.246   Type I diabetes  8  9  8  1.000   Type II diabetes  7  5  13  0.003   Hypercholesterolaemia  60  60  62  0.908   Hyperparathyroidism  67  68  62  0.429   Parathyroidectomy before Tx  22  23  17  0.091   CMV IgG positivity  58  58  61  0.502   EBV IgG positivity  91  91  89  0.501   Malignancies (excluding non-melanotic skin cancer)  6  6  7  0.689   Monoclonal gammopathy  1  1  3  0.087   Previous cytotoxic therapies  2  2  2  1.000   History of cigarette smoking  41  41  38  0.396   Time on dialysis before Tx (months)  70 ± 41  69 ± 42  78 ± 33  0.009    Deceased donor recipients  77 ± 38  77 ± 39  80 ± 31  0.474    Living donor recipients  29 ± 33  28 ± 33  37 ± 31  0.366   Re-transplanted patients  13  13  11  0.579   Body weight at Tx (kg)  72 ± 13  72 ± 13  73 ± 12  0.356   BMI at Tx  24 ± 4  24 ± 4  25 ± 3  0.063  Peri-operative data  Transplant-related factors           ESP (%)  10  8  24  <0.001   Combined kidney/pancreas Tx (%)  7  8  6  0.717   Pre-formed antibodies >0% (%)  7  7  8  0.854   Donor age (years)  49 ± 16  48 ± 15  53 ± 17  <0.001   Donor gender (%; male/female)  52/48  53/47  48/52  0.343   Donor type (%; deceased/living)  85/15  83/17  95/5  <0.001   Donor CMV IgG positivity (%)  58  58  59  0.924   Donor S-creatinine (µmol/L)  83 ± 46  84 ± 48  81 ± 34  0.890   Mean number of HLA mismatches  2.4 ± 1.7  2.3 ± 1.7  2.8 ± 1.8  0.005    HLA-A mismatches  0.71 ± 0.71  0.69 ± 0.69  0.81 ± 0.78  0.127    HLA-B mismatches  0.91 ± 0.73  0.89 ± 0.72  1.03 ± 0.75  0.048    HLA-DR mismatches  0.78 ± 0.70  0.75 ± 0.69  0.96 ± 0.70  0.001   CIT (h)  14.5 ± 7.4  14.0 ± 7.4  17.3 ± 6.8  <0.001    Deceased donor recipients  16.7 ± 5.9  16.4 ± 5.8  18.0 ± 6.1  0.004    Living donor recipients  2.5 ± 0.9  2.5 ± 0.9  2.7 ± 0.5  0.208  Initial immunosupressive therapy (%)           Induction therapy        0.660    Anti-lymphocyte globulin  9  9  11      IL-2 antibodies  79  79  80      None  10  10  7      Other/unknown  1/1  1/1  2/2     Cyclosporine A  75  75  74  0.664   Tacrolimus  19  20  13  0.055   MMF  66  67  61  0.197   Azathioprine  1  1  0  0.602   Everolimus or rapamycin  5  4  6  0.360   Steroids  95  96  95  0.653   Study drugs  9  9  9  1.000  Early post-Tx course           DGF (%)  29  27  43  <0.001   Best eGFR in the first 6 weeks post-Tx (mL/min/1.7m2)  59 ± 24  61 ± 24  49 ± 22  <0.001    All patients, n = 892  Surviving patients, n = 760  Deceased patients, n = 132  P-value  Pre-transplant data  Recipient          Age (years)  50 ± 13  48 ± 13  59 ± 11  <0.001  Gender (%; male/female)  59/41  59/41  61/39  0.774  Main reason for endstage renal disease (%)        0.056   Glomerulonephritis/vasculitis  26  27  17     Tubulointerstitial diseases  9  9  8     Hypertensive/diabetic nephropathy  14  13  18     Congenital disease  15  15  14     Other known causes  4  4  2     Unknown  33  31  40    Comorbidity before and at Tx (%)           Arterial hypertension  96  96  97  0.806   Heart failure  2  1  8  <0.001   Coronary heart disease  16  14  31  <0.001    History of myocardial infarction  5  4  11  0.002    History of coronary bypass surgery  6  5  13  0.003    History of coronary angioplasty  5  4  12  0.001   Peripheral arterial disease  13  10  32  <0.001   Ischaemic cerebral stroke  4  3  6  0.116   Replicative hepatitis B  1  1  1  1.000   Replicative hepatitis C  6  7  4  0.246   Type I diabetes  8  9  8  1.000   Type II diabetes  7  5  13  0.003   Hypercholesterolaemia  60  60  62  0.908   Hyperparathyroidism  67  68  62  0.429   Parathyroidectomy before Tx  22  23  17  0.091   CMV IgG positivity  58  58  61  0.502   EBV IgG positivity  91  91  89  0.501   Malignancies (excluding non-melanotic skin cancer)  6  6  7  0.689   Monoclonal gammopathy  1  1  3  0.087   Previous cytotoxic therapies  2  2  2  1.000   History of cigarette smoking  41  41  38  0.396   Time on dialysis before Tx (months)  70 ± 41  69 ± 42  78 ± 33  0.009    Deceased donor recipients  77 ± 38  77 ± 39  80 ± 31  0.474    Living donor recipients  29 ± 33  28 ± 33  37 ± 31  0.366   Re-transplanted patients  13  13  11  0.579   Body weight at Tx (kg)  72 ± 13  72 ± 13  73 ± 12  0.356   BMI at Tx  24 ± 4  24 ± 4  25 ± 3  0.063  Peri-operative data  Transplant-related factors           ESP (%)  10  8  24  <0.001   Combined kidney/pancreas Tx (%)  7  8  6  0.717   Pre-formed antibodies >0% (%)  7  7  8  0.854   Donor age (years)  49 ± 16  48 ± 15  53 ± 17  <0.001   Donor gender (%; male/female)  52/48  53/47  48/52  0.343   Donor type (%; deceased/living)  85/15  83/17  95/5  <0.001   Donor CMV IgG positivity (%)  58  58  59  0.924   Donor S-creatinine (µmol/L)  83 ± 46  84 ± 48  81 ± 34  0.890   Mean number of HLA mismatches  2.4 ± 1.7  2.3 ± 1.7  2.8 ± 1.8  0.005    HLA-A mismatches  0.71 ± 0.71  0.69 ± 0.69  0.81 ± 0.78  0.127    HLA-B mismatches  0.91 ± 0.73  0.89 ± 0.72  1.03 ± 0.75  0.048    HLA-DR mismatches  0.78 ± 0.70  0.75 ± 0.69  0.96 ± 0.70  0.001   CIT (h)  14.5 ± 7.4  14.0 ± 7.4  17.3 ± 6.8  <0.001    Deceased donor recipients  16.7 ± 5.9  16.4 ± 5.8  18.0 ± 6.1  0.004    Living donor recipients  2.5 ± 0.9  2.5 ± 0.9  2.7 ± 0.5  0.208  Initial immunosupressive therapy (%)           Induction therapy        0.660    Anti-lymphocyte globulin  9  9  11      IL-2 antibodies  79  79  80      None  10  10  7      Other/unknown  1/1  1/1  2/2     Cyclosporine A  75  75  74  0.664   Tacrolimus  19  20  13  0.055   MMF  66  67  61  0.197   Azathioprine  1  1  0  0.602   Everolimus or rapamycin  5  4  6  0.360   Steroids  95  96  95  0.653   Study drugs  9  9  9  1.000  Early post-Tx course           DGF (%)  29  27  43  <0.001   Best eGFR in the first 6 weeks post-Tx (mL/min/1.7m2)  59 ± 24  61 ± 24  49 ± 22  <0.001  Fifteen of the 892 patients were lost to follow-up during the post-transplant course. Tx, transplantation; CMV, cytomegalovirus; EBV, Epstein–Barr virus. Heart failure: any degree of insufficiency. Hyperparathyroidism is defined as an elevation of the most recent parathormone value within the year before transplantation by at least 2-fold of the upper normal. Pre-formed antibodies were determined by the lymphocytotoxic panel reactive antibody test. Serum creatinine of the donor represents the last known value before organ explantation and was available in 371 of the cases. DGF was defined as <500 mL urine within the first 24 h after transplantation and/or need of dialysis within the first week. Means are given with SD. Biopsies A total number of 2251 protocol biopsies and 1214 biopsies for cause were performed. Acute T-cell-mediated rejections including borderline cases were treated with steroid boli, except subclinical borderline cases in protocol biopsies, defined by an increase in serum creatinine by <20% at biopsy. In addition, in patients on a dual therapy mycophenolate mofetil was added. Patients with acute rejections occurring at 6 months or later or with vascular rejection at any time point were switched from cyclosporine to tacrolimus. No standardized treatment was defined for humoral rejection. Within the first transplant year, 316 rejections detected in protocol biopsies and 249 rejections detected in biopsies for cause were treated. Twenty-one cases of recurrent renal disease occurred among the surviving patients and eight cases among deceased patients, with only five recurrences within the first year (all in the group of surviving patients, with one case each with membranoproliferative and membranous glomerulonephritis, focal segmental glomerulosclerosis and two cases with thrombotic microangiopathy). Statistical analysis The IBM SPSS statistical software package version 24 and the rms package from R software version 3.2.1 [4] were used. The number of missing data was low except for the pre-transplant hyperparathyroidism (n = 344), hypercholesterolaemia (n = 301), blood transfusions (n = 103) and smoking (n = 79), which were therefore not included in the multivariable modelling. For a further 18 variables, a mean of 16 values was missing. These were replaced by the median of the whole group (continuous variables) and the most common attribute value (categorical variables) [5]. Continuous variables with normal distribution are given as means ± SD, data without normal distribution as medians with 25/75% quartiles. Categorical data were analysed with Fisher’s exact test and chi-square test for two or more samples. Continuous data were compared with the Kruskal–Wallis and Mann–Whitney test. Kaplan–Meier analysis and the log-rank test were used to describe patient and graft survival. For the Cox regression analyses, linearity of continuous variables was confirmed by categorizing the variables and comparing the β-coefficients from univariable Cox regression. Proportional hazards assumptions were confirmed graphically and by testing the scaled Schoenfeld residuals with the global PH-Test. Variables differing with a P-value of  ≤0.05 in univariable Cox analyses were considered for multivariable modelling, which was performed by stepwise backward selection (P-value threshold ≤0.05) to calculate the hazard ratios (HRs) for death. Validation of the models was performed by the bootstrapping procedures in the R software. Harrell’s concordance index was taken as measure of discrimination. Survival probabilities in the validation cohort were depicted by Kaplan–Meier curves for four discrete risks groups obtained from the linear predictor, using cut-points on the prognostic index determined by Cox’s method [6]. Nomograms were built using the estimators from the multivariable Cox modelling. Statistical significance was assumed for P < 0.05 (two tailed). RESULTS Patient survival Patient survival was 89% after 5 years and 78% after 10 years. Ninety-nine of 776 patients with functioning graft and 33 of 116 patients with graft failure and return to dialysis died (Figure 1A and B). Death events in patients with living-donor transplants were >50% lower, and twice as high in patients transplanted in the Eurotransplant Senior Programme (ESP; [7]) (Figure 1C). However, patient survival was not significantly different between patients in the ESP and patients aged >65 years who received an organ from a deceased donor aged <65 years (P = 0.20; Figure 1D). Two deaths occurred in the 16 cases with combined transplantation of kidney with heart, liver or lung, both after renal graft failure. Infection was the leading cause of death (23%), followed by malignancies (14%) and cardiovascular disease (14%). Malignancies included carcinoma of unknown origin (n = 7), three bronchial carcinomas, four carcinomas and one sarcoma of the gut, two liver carcinomas, one prostate carcinoma and one non-Hodgkin lymphoma. In 42%, there was no reliable information on the cause of death. Other specified death causes included one suicide, two traffic accidents, one fatal bleeding after transplant kidney rupture and four gastrointestinal causes (multiorgan failures after pancreatitis, bowel perforation after endoscopy, oesophageal varicose bleeding). The median time to death was similar among the different causes of death (3.2–3.7 years) except for other causes (2.2 years). FIGURE 1 View largeDownload slide Overall survival and survival in different subgroups. Dashed lines indicate 95% upper and lower confidence limits. (A) Overall survival of the 892 patients. (B) Survival in patients with functioning graft and with graft loss. Patients with graft loss had significant worse survival (P< 0.001). (C) Survival in patients with combined kidney with liver, heart or lung transplantations (two deaths in 16 patients), living-donor transplantations (n = 6 deaths in 134 patients), combined pancreas/kidney transplantations (n = 8 deaths in 65 patients) and patients transplanted in the ESP (n = 32 deaths in 93 patients). Groups were significantly different in the log-rank test (P< 0.001). (D) Survival in patients transplanted in the ESP compared with patients aged >65 years who received transplants from donors aged <65 years (P = 0.20). FIGURE 1 View largeDownload slide Overall survival and survival in different subgroups. Dashed lines indicate 95% upper and lower confidence limits. (A) Overall survival of the 892 patients. (B) Survival in patients with functioning graft and with graft loss. Patients with graft loss had significant worse survival (P< 0.001). (C) Survival in patients with combined kidney with liver, heart or lung transplantations (two deaths in 16 patients), living-donor transplantations (n = 6 deaths in 134 patients), combined pancreas/kidney transplantations (n = 8 deaths in 65 patients) and patients transplanted in the ESP (n = 32 deaths in 93 patients). Groups were significantly different in the log-rank test (P< 0.001). (D) Survival in patients transplanted in the ESP compared with patients aged >65 years who received transplants from donors aged <65 years (P = 0.20). Pre- and post-transplant factors with association to death Table 1 shows the comparison of pre-transplant and peri-operative variables between surviving patients and patients who died. Deceased patients were older and had spent more time on dialysis. Pre-existing cardiovascular disease, heart failure and type 2 diabetes were more prevalent. Transplants from deceased or older donors were more common and HLA-mismatches were higher. Cold ischaemic times (CITs) were longer in deceased patients receiving a transplant from a deceased donor. Delayed graft function (DGF) was more prevalent and the eGFR within the first 6 weeks after transplantation was lower. Supplementary data, Table S2 shows these variables descriptively in the subgroups with and without graft loss. In an explorative analysis, data of the further post-transplant course (complete list in Supplementary data, Table S1) were compared between survivors and deceased patients (Supplementary data, Table S3). Graft function was lower in deceased patients and graft loss more prevalent. Deceased patients had a higher graft arterial resistance index, more urinary tract infections, a higher frequency of BK virus nephropathy, treatment for acute T-cell-mediated and antibody-mediated rejection within the first transplant year and urological interventions. New-onset heart failure and myocardial infarction, coronary artery angioplasty or surgical revascularization, severe other diseases including severe infections and newly discovered malignancies were more prevalent in the post-transplant course of deceased patients. HRs and multivariable modelling of factors for death at 3 and 12 months after transplantation including validation of the models Univariable HRs were calculated for all pre- and post-transplant variables at 3 and 12 months after transplantation (Tables 2 and 3). Due to deaths before 3 months (n = 2) and 12 months (n = 17), these calculations were based on 890 and 875 patients, respectively. Variables with a P-value of ≤0.05 were used to create multivariable models at the two time points. Graft function, expressed as eGFR, was linearly correlated with survival. Graft loss was included in this linear predictor (annotated as an eGFR of 5 mL/min/1.7 m2) instead of using it as a separate variable in the modelling because eGFR and graft loss were highly collinear. Pre-transplant coronary heart disease, peripheral arterial disease and type 2 diabetes were combined in a score because this yielded the most stable results in the modelling and prevented multicollinearity. Linearity of this score was confirmed by comparing the β-coefficients from univariable Cox regression. Transplantation in the ESP was not included in the modelling as this variable was represented by recipient and donor age and HLA matching. Table 2 HRs of pre- and post-transplant variables at 3 months after transplantation from uni- and multivariable Cox regression analysis   Univariable analysis   Multivariable analysis   HR  95% CI  P-value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.067  1.050 1.085  <0.001  1.051  1.033 1.069  <0.001   Heart failure  5.320  2.855 9.914  <0.001  2.749  1.393 5.424  0.004   DM/PAD/CHD score  1.917  1.595 2.304  <0.001  1.412  1.143 1.745  0.001   Coronary heart disease  2.338  1.606 3.402  <0.001         Peripheral arterial disease  3.419  2.362 4.951  <0.001         Type 2 diabetes  2.560  1.514 4.328  <0.001         Monoclonal gammopathy  2.711  1.000 7.347  0.050         Time on dialysis (years)  1.005  1.001 1.009  0.017        Variables at Tx               BMI  1.049  1.000 1.101  0.050         Deceased donor Tx  3.707  1.634 8.410  0.002  2.640  1.100 6.336  0.030   Tx within the ESP  3.203  2.147 4.779  <0.001         Donor age (years)  1.026  1.014 1.038  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.076 1.320  <0.001          HLA-A mismatches  1.298  1.024 1.645  0.031          HLA-B mismatches  1.321  1.040 1.677  0.022          HLA-DR mismatches  1.549  1.217 1.972  <0.001          CIT (h)  1.045  1.022 1.068  <0.001        Post-Tx until 3 months              Graft factors               DGF  1.887  1.334 2.671  <0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.987  <0.001         eGFR at 3 months post-Tx (mL/min/1.7 m2)  0.971  0.962 0.981  <0.001         Graft loss  1.848  0.457 7.473  0.389         Renal graft arterial resistance index >0.75  2.836  1.963 4.097  <0.001         Urinary tract infection (%)  1.013  1.009 1.018  0.000  1.007  1.002 1.012  0.002   Transplant lymphocele  2.056  1.044 4.050  0.037  2.122  1.055 4.270  0.035   Urological interventions  1.582  1.069 2.340  0.022         Number of treated rejection episodes  1.342  1.122 1.604  0.001  2.055  1.428 2.957  <0.001   BK virus nephropathy  5.172  1.906 14.029  0.01  3.481  1.191 10.175  0.023  Comorbidities               New-onset heart failure  6.026  2.224 16.323  <0.001  5.675  2.047 15.736  0.001   Myocardial infarction, percutaneous or surgical coronary intervention  12.230  5.313 28.153  <0.001  5.815  2.468 13.703  <0.001   Severe acute diseases  3.807  2.285 6.341  <0.001         Severe infections  2.436  1.421 4.176  0.001  1.901  1.072 3.369  0.028   Malignancies  14.908  6.509 34.143  <0.001  19.404  7.892 47.709  <0.001   BMI change since Tx  0.892  0.818 0.973  0.010          Univariable analysis   Multivariable analysis   HR  95% CI  P-value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.067  1.050 1.085  <0.001  1.051  1.033 1.069  <0.001   Heart failure  5.320  2.855 9.914  <0.001  2.749  1.393 5.424  0.004   DM/PAD/CHD score  1.917  1.595 2.304  <0.001  1.412  1.143 1.745  0.001   Coronary heart disease  2.338  1.606 3.402  <0.001         Peripheral arterial disease  3.419  2.362 4.951  <0.001         Type 2 diabetes  2.560  1.514 4.328  <0.001         Monoclonal gammopathy  2.711  1.000 7.347  0.050         Time on dialysis (years)  1.005  1.001 1.009  0.017        Variables at Tx               BMI  1.049  1.000 1.101  0.050         Deceased donor Tx  3.707  1.634 8.410  0.002  2.640  1.100 6.336  0.030   Tx within the ESP  3.203  2.147 4.779  <0.001         Donor age (years)  1.026  1.014 1.038  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.076 1.320  <0.001          HLA-A mismatches  1.298  1.024 1.645  0.031          HLA-B mismatches  1.321  1.040 1.677  0.022          HLA-DR mismatches  1.549  1.217 1.972  <0.001          CIT (h)  1.045  1.022 1.068  <0.001        Post-Tx until 3 months              Graft factors               DGF  1.887  1.334 2.671  <0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.987  <0.001         eGFR at 3 months post-Tx (mL/min/1.7 m2)  0.971  0.962 0.981  <0.001         Graft loss  1.848  0.457 7.473  0.389         Renal graft arterial resistance index >0.75  2.836  1.963 4.097  <0.001         Urinary tract infection (%)  1.013  1.009 1.018  0.000  1.007  1.002 1.012  0.002   Transplant lymphocele  2.056  1.044 4.050  0.037  2.122  1.055 4.270  0.035   Urological interventions  1.582  1.069 2.340  0.022         Number of treated rejection episodes  1.342  1.122 1.604  0.001  2.055  1.428 2.957  <0.001   BK virus nephropathy  5.172  1.906 14.029  0.01  3.481  1.191 10.175  0.023  Comorbidities               New-onset heart failure  6.026  2.224 16.323  <0.001  5.675  2.047 15.736  0.001   Myocardial infarction, percutaneous or surgical coronary intervention  12.230  5.313 28.153  <0.001  5.815  2.468 13.703  <0.001   Severe acute diseases  3.807  2.285 6.341  <0.001         Severe infections  2.436  1.421 4.176  0.001  1.901  1.072 3.369  0.028   Malignancies  14.908  6.509 34.143  <0.001  19.404  7.892 47.709  <0.001   BMI change since Tx  0.892  0.818 0.973  0.010        Heart failure: any grade of insufficiency. Presence of type 2 diabetes, peripheral arterial disease (any grade) and coronary heart disease is additionally expressed as a score (DM/PAD/CHD score), which was later used in the multivariable analyses (1 point for each disease entity, adding up to a discrete score of 0–3). HLA mismatch on locus A, B and DR is entered as a numerical variable as 0, 1 or 2 according to the number of mismatches. eGFR: a value of 5 mL/min/1.7 m2 was assigned to patients with graft loss. Renal graft arterial resistance index was calculated as the mean of all available measurements. Urinary tract infection is expressed as the percentage of positive urine tests of the total number of examined urine samples. Transplant lymphocele: any singular or repeated finding. Urological interventions: any interventions (except ureteral stent withdrawal) compared with none. Number of treated rejection episodes includes intravenous steroid boli and increases in the maintenance immunosuppression. Blood pressure is expressed as the average of all available ambulatory measurements. Severe acute diseases include any major surgery and other severe, noninfectious illness, or life-threatening event compared with none. BMI change is defined as current BMI minus BMI at transplantation. All variables, with a P-value of ≤0.05 in the univariable analysis are included in the table. The 95% confidence intervals (95% CIs) are given with the upper and lower boundaries. The Harrel’s concordance index as a measure of model performance was 0.81 and 0.79 after 200-fold bootstrapping. Table 2 HRs of pre- and post-transplant variables at 3 months after transplantation from uni- and multivariable Cox regression analysis   Univariable analysis   Multivariable analysis   HR  95% CI  P-value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.067  1.050 1.085  <0.001  1.051  1.033 1.069  <0.001   Heart failure  5.320  2.855 9.914  <0.001  2.749  1.393 5.424  0.004   DM/PAD/CHD score  1.917  1.595 2.304  <0.001  1.412  1.143 1.745  0.001   Coronary heart disease  2.338  1.606 3.402  <0.001         Peripheral arterial disease  3.419  2.362 4.951  <0.001         Type 2 diabetes  2.560  1.514 4.328  <0.001         Monoclonal gammopathy  2.711  1.000 7.347  0.050         Time on dialysis (years)  1.005  1.001 1.009  0.017        Variables at Tx               BMI  1.049  1.000 1.101  0.050         Deceased donor Tx  3.707  1.634 8.410  0.002  2.640  1.100 6.336  0.030   Tx within the ESP  3.203  2.147 4.779  <0.001         Donor age (years)  1.026  1.014 1.038  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.076 1.320  <0.001          HLA-A mismatches  1.298  1.024 1.645  0.031          HLA-B mismatches  1.321  1.040 1.677  0.022          HLA-DR mismatches  1.549  1.217 1.972  <0.001          CIT (h)  1.045  1.022 1.068  <0.001        Post-Tx until 3 months              Graft factors               DGF  1.887  1.334 2.671  <0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.987  <0.001         eGFR at 3 months post-Tx (mL/min/1.7 m2)  0.971  0.962 0.981  <0.001         Graft loss  1.848  0.457 7.473  0.389         Renal graft arterial resistance index >0.75  2.836  1.963 4.097  <0.001         Urinary tract infection (%)  1.013  1.009 1.018  0.000  1.007  1.002 1.012  0.002   Transplant lymphocele  2.056  1.044 4.050  0.037  2.122  1.055 4.270  0.035   Urological interventions  1.582  1.069 2.340  0.022         Number of treated rejection episodes  1.342  1.122 1.604  0.001  2.055  1.428 2.957  <0.001   BK virus nephropathy  5.172  1.906 14.029  0.01  3.481  1.191 10.175  0.023  Comorbidities               New-onset heart failure  6.026  2.224 16.323  <0.001  5.675  2.047 15.736  0.001   Myocardial infarction, percutaneous or surgical coronary intervention  12.230  5.313 28.153  <0.001  5.815  2.468 13.703  <0.001   Severe acute diseases  3.807  2.285 6.341  <0.001         Severe infections  2.436  1.421 4.176  0.001  1.901  1.072 3.369  0.028   Malignancies  14.908  6.509 34.143  <0.001  19.404  7.892 47.709  <0.001   BMI change since Tx  0.892  0.818 0.973  0.010          Univariable analysis   Multivariable analysis   HR  95% CI  P-value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.067  1.050 1.085  <0.001  1.051  1.033 1.069  <0.001   Heart failure  5.320  2.855 9.914  <0.001  2.749  1.393 5.424  0.004   DM/PAD/CHD score  1.917  1.595 2.304  <0.001  1.412  1.143 1.745  0.001   Coronary heart disease  2.338  1.606 3.402  <0.001         Peripheral arterial disease  3.419  2.362 4.951  <0.001         Type 2 diabetes  2.560  1.514 4.328  <0.001         Monoclonal gammopathy  2.711  1.000 7.347  0.050         Time on dialysis (years)  1.005  1.001 1.009  0.017        Variables at Tx               BMI  1.049  1.000 1.101  0.050         Deceased donor Tx  3.707  1.634 8.410  0.002  2.640  1.100 6.336  0.030   Tx within the ESP  3.203  2.147 4.779  <0.001         Donor age (years)  1.026  1.014 1.038  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.076 1.320  <0.001          HLA-A mismatches  1.298  1.024 1.645  0.031          HLA-B mismatches  1.321  1.040 1.677  0.022          HLA-DR mismatches  1.549  1.217 1.972  <0.001          CIT (h)  1.045  1.022 1.068  <0.001        Post-Tx until 3 months              Graft factors               DGF  1.887  1.334 2.671  <0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.987  <0.001         eGFR at 3 months post-Tx (mL/min/1.7 m2)  0.971  0.962 0.981  <0.001         Graft loss  1.848  0.457 7.473  0.389         Renal graft arterial resistance index >0.75  2.836  1.963 4.097  <0.001         Urinary tract infection (%)  1.013  1.009 1.018  0.000  1.007  1.002 1.012  0.002   Transplant lymphocele  2.056  1.044 4.050  0.037  2.122  1.055 4.270  0.035   Urological interventions  1.582  1.069 2.340  0.022         Number of treated rejection episodes  1.342  1.122 1.604  0.001  2.055  1.428 2.957  <0.001   BK virus nephropathy  5.172  1.906 14.029  0.01  3.481  1.191 10.175  0.023  Comorbidities               New-onset heart failure  6.026  2.224 16.323  <0.001  5.675  2.047 15.736  0.001   Myocardial infarction, percutaneous or surgical coronary intervention  12.230  5.313 28.153  <0.001  5.815  2.468 13.703  <0.001   Severe acute diseases  3.807  2.285 6.341  <0.001         Severe infections  2.436  1.421 4.176  0.001  1.901  1.072 3.369  0.028   Malignancies  14.908  6.509 34.143  <0.001  19.404  7.892 47.709  <0.001   BMI change since Tx  0.892  0.818 0.973  0.010        Heart failure: any grade of insufficiency. Presence of type 2 diabetes, peripheral arterial disease (any grade) and coronary heart disease is additionally expressed as a score (DM/PAD/CHD score), which was later used in the multivariable analyses (1 point for each disease entity, adding up to a discrete score of 0–3). HLA mismatch on locus A, B and DR is entered as a numerical variable as 0, 1 or 2 according to the number of mismatches. eGFR: a value of 5 mL/min/1.7 m2 was assigned to patients with graft loss. Renal graft arterial resistance index was calculated as the mean of all available measurements. Urinary tract infection is expressed as the percentage of positive urine tests of the total number of examined urine samples. Transplant lymphocele: any singular or repeated finding. Urological interventions: any interventions (except ureteral stent withdrawal) compared with none. Number of treated rejection episodes includes intravenous steroid boli and increases in the maintenance immunosuppression. Blood pressure is expressed as the average of all available ambulatory measurements. Severe acute diseases include any major surgery and other severe, noninfectious illness, or life-threatening event compared with none. BMI change is defined as current BMI minus BMI at transplantation. All variables, with a P-value of ≤0.05 in the univariable analysis are included in the table. The 95% confidence intervals (95% CIs) are given with the upper and lower boundaries. The Harrel’s concordance index as a measure of model performance was 0.81 and 0.79 after 200-fold bootstrapping. Table 3 HRs of pre- and post-transplant variables at 12 months after transplantation from uni- and multivariable Cox regression analysis   Univariable analysis   Multivariable analysis   HR  95% CI  P- value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.071  1.053 1.089  <0.001  1.048  1.029 1.068  <0.001   Heart failure  6.022  3.220 11.264  <0.001  3.020  1.570 5.809  0.001   DM/PAD/CHD score  1.999  1.656 2.413  <0.001  1.523  1.218 1.904  <0.001   Coronary heart disease  2.469  1.674 3.640  <0.001         Peripheral arterial disease  3.796  2.595 5.554  <0.001         Type 2 diabetes  2.690  1.562 4.633  <0.001         Monoclonal gammopathy  3.022  1.113 8.203  0.030         Time on dialysis (years)  1.004  1.000 1.008  0.039        Variables at transplantation               BMI  1.054  1.003 1.109  0.038         Deceased donor Tx  3.379  1.486 7.680  0.004         Tx within the ESP  3.357  2.217 5.084  <0.001         Donor age (years)  1.028  1.015 1.041  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.071 1.327  0.001          HLA-A mismatches  1.305  1.019 1.671  0.035          HLA-B mismatches  1.335  1.040 1.714  0.023          HLA-DR mismatches  1.524  1.184 1.961  0.001         CIT (h)  1.039  1.015 1.063  0.001        Post-Tx variables until 12 months              Graft factors               Delayed graft function  1.871  1.301 2.691  0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.988  <0.001         eGFR at 12 months post-Tx (mL/min/1.7 m2)  0.969  0.958 0.979  <0.001  0.987  0.977 0.997  0.011   Graft loss  5.956  3.337 10.630  0.000         Renal graft arterial resistance index >0.75  2.722  1.882 3.937  <0.001         Urinary tract infection (%)  1.027  1.022 1.033  <0.001  1.019  1.012 1.026  <0.001   Urological interventions  1.597  1.079 2.362  0.019         Number of treated rejection episodes  1.422  1.195 1.693  <0.001  1.351  1.102 1.656  0.004   BK virus nephropathy  2.731  1.111 6.710  0.029        Comorbidities               Mean of diastolic blood pressure (mmHg)  0.948  0.922 0.975  <0.001         New-onset heart failure  4.113  1.808 9.364  0.001  5.435  2.294 12.875  <0.001   Myocardial infarction, percutaneous or surgical   coronary intervention  4.934  2.406 10.121  <0.001  3.619  1.706 7.678  <0.001   Severe acute diseases  1.905  1.178 3.080  0.009         Severe infections  2.433  1.607 3.683  <0.001         Malignancies  2.907  1.185 7.131  0.045  3.284  1.317 8.191  0.011   BMI change since Tx  0.827  0.764 0.896  0.000          Univariable analysis   Multivariable analysis   HR  95% CI  P- value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.071  1.053 1.089  <0.001  1.048  1.029 1.068  <0.001   Heart failure  6.022  3.220 11.264  <0.001  3.020  1.570 5.809  0.001   DM/PAD/CHD score  1.999  1.656 2.413  <0.001  1.523  1.218 1.904  <0.001   Coronary heart disease  2.469  1.674 3.640  <0.001         Peripheral arterial disease  3.796  2.595 5.554  <0.001         Type 2 diabetes  2.690  1.562 4.633  <0.001         Monoclonal gammopathy  3.022  1.113 8.203  0.030         Time on dialysis (years)  1.004  1.000 1.008  0.039        Variables at transplantation               BMI  1.054  1.003 1.109  0.038         Deceased donor Tx  3.379  1.486 7.680  0.004         Tx within the ESP  3.357  2.217 5.084  <0.001         Donor age (years)  1.028  1.015 1.041  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.071 1.327  0.001          HLA-A mismatches  1.305  1.019 1.671  0.035          HLA-B mismatches  1.335  1.040 1.714  0.023          HLA-DR mismatches  1.524  1.184 1.961  0.001         CIT (h)  1.039  1.015 1.063  0.001        Post-Tx variables until 12 months              Graft factors               Delayed graft function  1.871  1.301 2.691  0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.988  <0.001         eGFR at 12 months post-Tx (mL/min/1.7 m2)  0.969  0.958 0.979  <0.001  0.987  0.977 0.997  0.011   Graft loss  5.956  3.337 10.630  0.000         Renal graft arterial resistance index >0.75  2.722  1.882 3.937  <0.001         Urinary tract infection (%)  1.027  1.022 1.033  <0.001  1.019  1.012 1.026  <0.001   Urological interventions  1.597  1.079 2.362  0.019         Number of treated rejection episodes  1.422  1.195 1.693  <0.001  1.351  1.102 1.656  0.004   BK virus nephropathy  2.731  1.111 6.710  0.029        Comorbidities               Mean of diastolic blood pressure (mmHg)  0.948  0.922 0.975  <0.001         New-onset heart failure  4.113  1.808 9.364  0.001  5.435  2.294 12.875  <0.001   Myocardial infarction, percutaneous or surgical   coronary intervention  4.934  2.406 10.121  <0.001  3.619  1.706 7.678  <0.001   Severe acute diseases  1.905  1.178 3.080  0.009         Severe infections  2.433  1.607 3.683  <0.001         Malignancies  2.907  1.185 7.131  0.045  3.284  1.317 8.191  0.011   BMI change since Tx  0.827  0.764 0.896  0.000        All variables, with a P-value of  ≤0.05 in the univariable analysis are included in the table. The 95% confidence intervals (95% CIs) are given with the upper and lower boundaries. The Harrel’s concordance index as a measure of model performance was 0.81 and 0.80 after 200-fold bootstrapping. For definition of the clinical terms, see legend to Table 2. Table 3 HRs of pre- and post-transplant variables at 12 months after transplantation from uni- and multivariable Cox regression analysis   Univariable analysis   Multivariable analysis   HR  95% CI  P- value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.071  1.053 1.089  <0.001  1.048  1.029 1.068  <0.001   Heart failure  6.022  3.220 11.264  <0.001  3.020  1.570 5.809  0.001   DM/PAD/CHD score  1.999  1.656 2.413  <0.001  1.523  1.218 1.904  <0.001   Coronary heart disease  2.469  1.674 3.640  <0.001         Peripheral arterial disease  3.796  2.595 5.554  <0.001         Type 2 diabetes  2.690  1.562 4.633  <0.001         Monoclonal gammopathy  3.022  1.113 8.203  0.030         Time on dialysis (years)  1.004  1.000 1.008  0.039        Variables at transplantation               BMI  1.054  1.003 1.109  0.038         Deceased donor Tx  3.379  1.486 7.680  0.004         Tx within the ESP  3.357  2.217 5.084  <0.001         Donor age (years)  1.028  1.015 1.041  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.071 1.327  0.001          HLA-A mismatches  1.305  1.019 1.671  0.035          HLA-B mismatches  1.335  1.040 1.714  0.023          HLA-DR mismatches  1.524  1.184 1.961  0.001         CIT (h)  1.039  1.015 1.063  0.001        Post-Tx variables until 12 months              Graft factors               Delayed graft function  1.871  1.301 2.691  0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.988  <0.001         eGFR at 12 months post-Tx (mL/min/1.7 m2)  0.969  0.958 0.979  <0.001  0.987  0.977 0.997  0.011   Graft loss  5.956  3.337 10.630  0.000         Renal graft arterial resistance index >0.75  2.722  1.882 3.937  <0.001         Urinary tract infection (%)  1.027  1.022 1.033  <0.001  1.019  1.012 1.026  <0.001   Urological interventions  1.597  1.079 2.362  0.019         Number of treated rejection episodes  1.422  1.195 1.693  <0.001  1.351  1.102 1.656  0.004   BK virus nephropathy  2.731  1.111 6.710  0.029        Comorbidities               Mean of diastolic blood pressure (mmHg)  0.948  0.922 0.975  <0.001         New-onset heart failure  4.113  1.808 9.364  0.001  5.435  2.294 12.875  <0.001   Myocardial infarction, percutaneous or surgical   coronary intervention  4.934  2.406 10.121  <0.001  3.619  1.706 7.678  <0.001   Severe acute diseases  1.905  1.178 3.080  0.009         Severe infections  2.433  1.607 3.683  <0.001         Malignancies  2.907  1.185 7.131  0.045  3.284  1.317 8.191  0.011   BMI change since Tx  0.827  0.764 0.896  0.000          Univariable analysis   Multivariable analysis   HR  95% CI  P- value  HR  95% CI  P-value  Pre-Tx variables               Recipient age (years)  1.071  1.053 1.089  <0.001  1.048  1.029 1.068  <0.001   Heart failure  6.022  3.220 11.264  <0.001  3.020  1.570 5.809  0.001   DM/PAD/CHD score  1.999  1.656 2.413  <0.001  1.523  1.218 1.904  <0.001   Coronary heart disease  2.469  1.674 3.640  <0.001         Peripheral arterial disease  3.796  2.595 5.554  <0.001         Type 2 diabetes  2.690  1.562 4.633  <0.001         Monoclonal gammopathy  3.022  1.113 8.203  0.030         Time on dialysis (years)  1.004  1.000 1.008  0.039        Variables at transplantation               BMI  1.054  1.003 1.109  0.038         Deceased donor Tx  3.379  1.486 7.680  0.004         Tx within the ESP  3.357  2.217 5.084  <0.001         Donor age (years)  1.028  1.015 1.041  <0.001         Mean number of HLA mismatches on locus A, B, DR  1.192  1.071 1.327  0.001          HLA-A mismatches  1.305  1.019 1.671  0.035          HLA-B mismatches  1.335  1.040 1.714  0.023          HLA-DR mismatches  1.524  1.184 1.961  0.001         CIT (h)  1.039  1.015 1.063  0.001        Post-Tx variables until 12 months              Graft factors               Delayed graft function  1.871  1.301 2.691  0.001         Best eGFR in the first 6 weeks post-Tx (mL/min/1.7 m2)  0.980  0.972 0.988  <0.001         eGFR at 12 months post-Tx (mL/min/1.7 m2)  0.969  0.958 0.979  <0.001  0.987  0.977 0.997  0.011   Graft loss  5.956  3.337 10.630  0.000         Renal graft arterial resistance index >0.75  2.722  1.882 3.937  <0.001         Urinary tract infection (%)  1.027  1.022 1.033  <0.001  1.019  1.012 1.026  <0.001   Urological interventions  1.597  1.079 2.362  0.019         Number of treated rejection episodes  1.422  1.195 1.693  <0.001  1.351  1.102 1.656  0.004   BK virus nephropathy  2.731  1.111 6.710  0.029        Comorbidities               Mean of diastolic blood pressure (mmHg)  0.948  0.922 0.975  <0.001         New-onset heart failure  4.113  1.808 9.364  0.001  5.435  2.294 12.875  <0.001   Myocardial infarction, percutaneous or surgical   coronary intervention  4.934  2.406 10.121  <0.001  3.619  1.706 7.678  <0.001   Severe acute diseases  1.905  1.178 3.080  0.009         Severe infections  2.433  1.607 3.683  <0.001         Malignancies  2.907  1.185 7.131  0.045  3.284  1.317 8.191  0.011   BMI change since Tx  0.827  0.764 0.896  0.000        All variables, with a P-value of  ≤0.05 in the univariable analysis are included in the table. The 95% confidence intervals (95% CIs) are given with the upper and lower boundaries. The Harrel’s concordance index as a measure of model performance was 0.81 and 0.80 after 200-fold bootstrapping. For definition of the clinical terms, see legend to Table 2. Eight factors were present in both models, including age, pre-transplant heart failure, the score of cardiovascular disease and type 2 diabetes, post-transplant urinary tract infection, treatment of rejection (either in protocol biopsies or biopsies for cause), new-onset heart failure, coronary events and malignancies. Additional variables at 3 months included deceased donor transplantation, transplant lymphocele and BK virus nephropathy. Graft function was a significant factor of the model at 12 months. The Harrell’s concordance index indicated satisfactory performance, with 0.81 for both models, and 0.79 (3 months) and 0.80 (12 months) after 200-fold bootstraps. Validation of the two models was performed on a separate group of 349 patients transplanted between 2008 and 2013, with a follow-up until June 2017. These patients differed in several aspects (Supplementary data, Table S4) from the patient cohort that was used to build the models. Peripheral arterial disease was less and type 2 diabetes more prevalent. More patients received living-donor transplantations. CIT was shorter and DGF and urological interventions were less common. All patients had mycophenolate mofetil and a higher proportion received tacrolimus instead of cyclosporine. Less rejection treatments were performed (0.46 ±0.81 versus 0.63 ±0.89/patient). The validation analysis showed satisfactory discrimination of patient risks for death (Figure 2), with a concordance index of 0.73 for the model at 3 months and 0.76 for the model at 12 months. Estimators of both models are graphically depicted by a nomogram, which is intended as a tool to weigh individual risks that may contribute to decreased survival (Figure 3). FIGURE 2 View largeDownload slide Survival probability in the separate validation group of 349 patients, based on the prediction of the multivariable models at 3 months (A) and at 12 months (B) post-transplantation. Kaplan Meier curves are shown for four discrete risks groups (‘very low risk’ up to ‘high risk’ for death) obtained from the linear predictor, using cut-points on the prognostic index determined by Cox’s method (cut points: 16th, 50th and 84th percentiles of the prognostic index). The boundaries of the prognostic index were: very low risk; −2.83 to −1.30, low risk; −1.31 to −0.23, moderate risk; −0.24 to 0.81, high risk; 0.82–2.69 (3 months); and very low risk; −2.46 to −1.14, low risk; −1.15 to −0.14, moderate risk; −0.15 to 0.90, high risk; 0.91–2.93 (12 months). Number of patients observed at each time is shown above the abscissa. FIGURE 2 View largeDownload slide Survival probability in the separate validation group of 349 patients, based on the prediction of the multivariable models at 3 months (A) and at 12 months (B) post-transplantation. Kaplan Meier curves are shown for four discrete risks groups (‘very low risk’ up to ‘high risk’ for death) obtained from the linear predictor, using cut-points on the prognostic index determined by Cox’s method (cut points: 16th, 50th and 84th percentiles of the prognostic index). The boundaries of the prognostic index were: very low risk; −2.83 to −1.30, low risk; −1.31 to −0.23, moderate risk; −0.24 to 0.81, high risk; 0.82–2.69 (3 months); and very low risk; −2.46 to −1.14, low risk; −1.15 to −0.14, moderate risk; −0.15 to 0.90, high risk; 0.91–2.93 (12 months). Number of patients observed at each time is shown above the abscissa. FIGURE 3 View largeDownload slide Nomograms to estimate patient’s individual risks of reduced survival at 3 months (A) and 12 months post-transplantation (B). For each factor, the individual value is located on the corresponding scale. By drawing a vertical line from this position to the most upper scale (‘points’), the number of points is read from this scale. The number of points for each variable is added up to give the total number of points. On the scale for total points, the patient’s total points are located, and a vertical line is drawn to the survival scale which indicates the effect of the individual risk factors on the 10-year survival. DM/PAD/CHD score: a value of 1 for each condition at Tx (type 2 diabetes, any grade of peripheral arterial disease, coronary heart disease). Heart failure: any grade of heart failure. Percentage of urinary tract infections is calculated by relating the number of positive urine tests to the total number of examined urine samples. New coronary event: myocardial infarction, percutaneous or surgical coronary intervention. Rejection treatments: number of treated episodes. Tx, transplantation. FIGURE 3 View largeDownload slide Nomograms to estimate patient’s individual risks of reduced survival at 3 months (A) and 12 months post-transplantation (B). For each factor, the individual value is located on the corresponding scale. By drawing a vertical line from this position to the most upper scale (‘points’), the number of points is read from this scale. The number of points for each variable is added up to give the total number of points. On the scale for total points, the patient’s total points are located, and a vertical line is drawn to the survival scale which indicates the effect of the individual risk factors on the 10-year survival. DM/PAD/CHD score: a value of 1 for each condition at Tx (type 2 diabetes, any grade of peripheral arterial disease, coronary heart disease). Heart failure: any grade of heart failure. Percentage of urinary tract infections is calculated by relating the number of positive urine tests to the total number of examined urine samples. New coronary event: myocardial infarction, percutaneous or surgical coronary intervention. Rejection treatments: number of treated episodes. Tx, transplantation. DISCUSSION In this retrospective study of patients from our large kidney protocol biopsy programme, we confirm established risk factors for death like age, heart failure, type 2 diabetes, cardiovascular disease and graft function. Novel factors include post-transplant urinary tract infections and BK virus nephropathy. Unlike most previous studies, our analysis included patients with graft loss. We believe that this integral view is best suited to a rational and patient-centred risk assessment, since graft loss is not uninformative regarding death. Further, study of patients with protocol biopsies allowed evaluating treatment of subclinical rejections in the analyses. Possible limitations include the large proportion of unknown causes of death that precluded in-depth subanalyses for different causes. Regarding the generalizability of our results, based on the pre-transplant data, we believe that our cohort of approximately 900 patients is well comparable to patients from many other transplant programmes that treat mainly Caucasian patients. Strengths of our single-centre study include consistent data collection with a high degree of completeness and a low number of lost to follow-up cases within up to 11 years. In addition, we aimed to employ robust definitions for each variable and to retain the highest possible informational value of continuous predictors by avoiding categorization of these variables [8]. The obtained multivariable models were robust as confirmed by bootstrapping analysis and by the discriminatory performance in an independent validation cohort. Patient survival was 89% after 5 years and 78% after 10 years, similar to figures from the ERA-EDTA registry [9]. Yet, it has to be noted that patient survival reported in this study included patients with graft loss which have an excess risk for death [10]. Similar to the International Pancreas Transplant Registry [11], we observed favourable survival results in the highly selected patient group with combined pancreas/kidney transplants. One- to four-year results on the ESP showed similar survival compared with recipients of the same age who received organs from younger donors [12, 13]. Nevertheless, the benefit for older patients receiving organs from extended criteria donors is controversial [14, 15]. Our data do not indicate a significant survival difference (P = 0.2) between transplantations within the ESP compared with transplantations with younger deceased donors. In recent registry studies from the UK [16] and from Australia/New Zealand [17], cardiovascular disease accounted for 23% and 16% of the deaths, respectively, compared with our rate of 14%. However, the cause of death was unavailable in 42% in our data so that the true number might be higher. Rates of infectious deaths were very similar, with 23% in our centre and 22–23% in the aforementioned studies. We did not observe relevant time effects among different causes of death like Mazuecos et al. [18], who reported higher rates of infectious death in the first transplant year and increasing malignancy-related deaths in later years. Similar to recent studies [19–23], age, pre-transplant type 2 diabetes, coronary and peripheral arterial disease and heart failure were important predictors of death. In our study, pre-transplant heart failure was prevalent with 2%, with similar proportions of New York Heart Association Grades I and II. The degree of pre-transplant heart failure was not predictive of death (data not shown), reflecting the difficulty of assessing the true severity of heart failure in the pre-transplant setting with reduced hydration status and of predicting the effect of hydration after transplantation in these patients. New-onset heart failure after transplantation was a strong risk factor, with three deaths among 10 patients with Grade II, two deaths in two patients with Grade III, four deaths in the nine patients with Grade IV and no deaths in four patients with Grade I. The importance of post-transplant coronary events has been highlighted by a previous study [23]. Newly discovered malignancy was an important risk factor, including one case each with carcinoma of the prostate, kidney, vulva and oesophagus, one invasive squamous skin cancer, and one Kaposi sarcoma in the group of deceased patients. Excluding the patient with Kaposi sarcoma, it remains open whether these malignancies had been present pre-transplant and missed by the regular waitlist examinations. A previous study in nearly 5000 kidney transplant recipients reported an even higher rate of 12.7% malignancies within the first transplant year [24], compared with 2% in our study. In a large registry analysis from the UK, malignancy-related death accounted for 7.4% of deaths within the first transplant year, and two-thirds of these patients had a pre-transplant history of malignancy [25]. Reduced graft function is an established risk factor for death [23, 26, 27]. In our study, the eGFR at 12 months was linearly correlated with the risk of death, including patients with graft loss (which were assigned to an eGFR of 5 mL/min/1.7 m2). Graft function was not a risk factor in the model at 3 months probably because graft function was more variable at this time due to ischaemia/reperfusion injury, toxicity of the medication and other factors. In fact, the model at 3 months included factors with known effects on graft function, namely deceased donor transplantations, BK virus nephropathy, rejection and urinary tract infection. The detrimental effect of BK virus nephropathy on renal graft survival is well established [28]. We are not aware of studies linking BK virus replication or nephropathy to decreased patient survival in kidney transplant recipients. A large study in allogeneic hematopoietic stem cell recipients identified BK virus replication as an independent risk factor for death. It remains open whether BK virus replication directly affected survival or was a surrogate of other risks, e.g. stronger suppression of the host’s immune system. Yet, we did not find a higher proportion of deaths due to infection in patients with BK virus nephropathy. The negative impact of early acute rejection—even if subclinical—on graft function and graft survival has been demonstrated by recent studies, without studying effects on patient survival [29, 30] or with reporting lacking effect on patient survival [31, 32]. In one study [20], acute rejection was associated with death. Treated rejection episodes—either in protocol biopsies or biopsies for cause—were a significant factor in our models. Yet, rejection treatments were not specifically related to infectious or malignancy-related deaths in univariable analyses, with an average of 0.75 and 0.94 treatments/patient with infectious and malignancy-associated death compared with 1.1 treatments/patient with cardiovascular death. Most likely, acute rejection represented therefore a predictor of the future course of graft function and loss. Severe infections were a significant factor in the 3 months’ model. In recipients of kidney, liver or heart transplants, post-transplant infections were significantly associated with death, and urinary tract infections accounted for one-third of all infections [33]. Only one study specifically addressed urinary tract infection in kidney transplant patients and found an association with death [34]. In our study, we established a linear relationship between the frequency of urinary tract infections and death. Patients with infectious or malignancy-related deaths did not have a higher rate of urinary tract infections than patients dying from cardiovascular disease (0.16 and 0.26 versus 0.20, respectively). Previous studies have attempted to identify and integrate risk factors for death into predictive models, including the pre-transplant variables age, gender, race, body mass index (BMI), time on dialysis, cause of end-stage renal disease, panel reactive antibodies, HLA mismatches, comorbidities such as diabetes, cardiovascular disease and heart failure, and donor age. In some models, the post-transplant factors DGF, acute rejection and graft function were included [35–42]. A model performance with an area under the curve of 0.63 after validation with a separate patient cohort was reported by using age, pre-transplant coronary heart disease, left ventricular hypertrophy, vascular calcification, diabetes, time on dialysis, and acute tubular necrosis and GFR at discharge after transplantation as factors [36]. Established indices with numerous comorbidities were tested without validation on independent patient groups, reporting a c-statistics value of 0.71 [39] or no performance measures [38]. Another model to predict death until 3 years post-transplantation included age, weight, time on dialysis, diabetes, hepatitis C at transplantation, DGF, and diabetes, proteinuria, renal function and immunosuppressants within the first year, with a c-statistics value of 0.74 on validation in an independent patient cohort [37]. In patients aged >65 years, a model with age, gender, race, post-transplant smoking, eGFR, albumin excretion, diabetes, heart failure and stroke predicted 5-year mortality risk with a c-statistics value of 0.69 on validation with an independent patient cohort [42]. In none of these studies was the fate of patients with graft loss and return to dialysis considered in the prediction. Our results emphasize the importance of pre- and post-transplant monitoring for cardiovascular disease, heart failure and malignancies. Further, preservation of renal graft function, particularly avoidance of acute rejection and BK nephropathy appear to be important treatment goals. The proposed models with a comparatively high predictive performance may help to estimate and balance existing risks for death in individual patients better. Although some risks may not be modifiable or may have undetermined causality, such risk assessment can help to identify individually important areas that need particular attention in the pre-transplant preparation and post-transplant care. Further validation of the models is required, preferentially in cohorts from other transplant centres. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS The help of A. Henkel and B. Boes in the collection of the data is greatly appreciated. This study was supported by intramural funding. AUTHORS’ CONTRIBUTIONS T.A. and W.G. were responsible for the research design, writing and data analysis. I.S. and A. Karch were responsible for the data analysis, research design and writing. V.B. was responsible for the research design and writing. A. Koch was responsible for the data analysis and research design. H.H. and A.S. were responsible for the research design and writing. CONFLICT OF INTEREST STATEMENT None declared. The results presented in this paper have not been published previously in whole or part, except in abstract format. REFERENCES 1 Foley RN, Parfrey PS, Sarnak MJ. Clinical epidemiology of cardiovascular disease in chronic renal disease. Am J Kidney Dis  1998; 32: S112– S119 Google Scholar CrossRef Search ADS PubMed  2 Wolfe RA, Ashby VB, Milford EL. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med  1999; 341: 1725– 1730 Google Scholar CrossRef Search ADS PubMed  3 Zsom L, Wagner L, Fulop T. Minimization vs tailoring: where do we stand with personalized immunosuppression during renal transplantation in 2015? World J Transplant  2015; 5: 73– 80 Google Scholar CrossRef Search ADS PubMed  4 R Development Core Team. R R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2015. www.r-project.org 5 Kaiser J. Dealing with missing values in data. J Syst Integr  2014; 5: 42– 51 Google Scholar CrossRef Search ADS   6 Cox DR. Note on grouping. J Am Stat Assoc  1957; 52: 543– 547 Google Scholar CrossRef Search ADS   7 Smits JM, Persijn GG, van Houwelingen HC et al.   Evaluation of the Eurotransplant Senior Program: the results of the first year. Am J Transplant  2002; 2: 664– 670 Google Scholar CrossRef Search ADS PubMed  8 Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors in multiple regression: a bad idea. Statist Med  2006; 25: 127– 141 Google Scholar CrossRef Search ADS   9 Pippias M, Stel VS, Abad Diez JM et al.   Renal replacement therapy in Europe: a summary of the 2012 ERA-EDTA Registry Annual Report. Clin Kidney J  2015; 8: 248– 261 Google Scholar CrossRef Search ADS PubMed  10 Brar A, Markell M, Stefanov DG et al.   Mortality after renal allograft failure and return to dialysis. Am J Nephrol  2017; 45: 180– 186 Google Scholar CrossRef Search ADS PubMed  11 Gruessner AC. 2011 update on pancreas transplantation: comprehensive trend analysis of 25, 000 cases followed up over the course of twenty-four years at the International Pancreas Transplant Registry (IPTR). Rev Diabet Stud  2011; 8: 6– 16 Google Scholar CrossRef Search ADS PubMed  12 Bentas W, Jones J, Karaoguz A et al.   Renal transplantation in the elderly: surgical complications and outcome with special emphasis on the Eurotransplant Senior Programme. Nephrol Dial Transplant  2008; 23: 2043– 2051 Google Scholar CrossRef Search ADS PubMed  13 Fabrizii V, Kovarik J, Bodingbauer M et al.   Long-term patient and graft survival in the eurotransplant senior program: a single-center experience. Transplantation  2005; 80: 582– 589 Google Scholar CrossRef Search ADS PubMed  14 Veroux M, Grosso G, Corona D et al.   Age is an important predictor of kidney transplantation outcome. Nephrol Dial Transplant  2012; 27: 1663– 1671 Google Scholar CrossRef Search ADS PubMed  15 Merion RM, Ashby VB, Wolfe RA et al.   Deceased-donor characteristics and the survival benefit of kidney transplantation. JAMA  2005; 294: 2726– 2733 Google Scholar CrossRef Search ADS PubMed  16 Steenkamp R, Castledine C, Feest T. Chapter 6 Survival and causes of death of UK adult patients on renal replacement therapy in 2010: national and centre-specific analyses. Nephron Clin Pract  2012; 120 (Suppl. 1): c105– c135 Google Scholar CrossRef Search ADS PubMed  17 Pilmore H, Dent H, Chang S et al.   Reduction in cardiovascular death after kidney transplantation. Transplantation  2010; 89: 851– 857 Google Scholar CrossRef Search ADS PubMed  18 Mazuecos A, Munoz Terol JM, Garcia Alvarez T et al.   Increase in malignancies as cause of death in renal transplant patients. Transplant Proc  2009; 41: 2159– 2162 Google Scholar CrossRef Search ADS PubMed  19 Karim A, Farrugia D, Cheshire J et al.   Recipient age and risk for mortality after kidney transplantation in England. Transplantation  2014; 97: 832– 838 Google Scholar CrossRef Search ADS PubMed  20 Faravardeh A, Eickhoff M, Jackson S et al.   Predictors of graft failure and death in elderly kidney transplant recipients. Transplantation  2013; 96: 1089– 1096 Google Scholar CrossRef Search ADS PubMed  21 Hayer MK, Farrugia D, Begaj I et al.   Infection-related mortality is higher for kidney allograft recipients with pretransplant diabetes mellitus. Diabetologia  2014; 57: 554– 561 Google Scholar CrossRef Search ADS PubMed  22 Brar A, Jindal RM, Elster EA et al.   Effect of peripheral vascular disease on kidney allograft outcomes: a study of U.S. Renal data system. Transplantation  2013; 95: 810– 815 Google Scholar CrossRef Search ADS PubMed  23 Kahwaji J, Bunnapradist S, Hsu JW et al.   Cause of death with graft function among renal transplant recipients in an integrated healthcare system. Transplantation  2011; 91: 225– 230 Google Scholar CrossRef Search ADS PubMed  24 Cheung CY, Lam MF, Chu KH et al.   Malignancies after kidney transplantation: Hong Kong renal registry. Am J Transplant  2012; 12: 3039– 3046 Google Scholar CrossRef Search ADS PubMed  25 Farrugia D, Cheshire J, Begaj I et al.   Death within the first year after kidney transplantation: an observational cohort study. Transpl Int  2014; 27: 262– 270 Google Scholar CrossRef Search ADS PubMed  26 Moranne O, Maillard N, Fafin C et al.   Rate of renal graft function decline after one year is a strong predictor of all-cause mortality. Am J Transplant  2013; 13: 695– 706 Google Scholar CrossRef Search ADS PubMed  27 Soveri I, Holdaas H, Jardine A et al.   Renal transplant dysfunction–importance quantified in comparison with traditional risk factors for cardiovascular disease and mortality. Nephrol Dial Transplant  2006; 21: 2282– 2289 Google Scholar CrossRef Search ADS PubMed  28 Dharnidharka VR, Cherikh WS, Abbott KC. An OPTN analysis of national registry data on treatment of BK virus allograft nephropathy in the United States. Transplantation  2009; 87: 1019– 1026 Google Scholar CrossRef Search ADS PubMed  29 Loupy A, Vernerey D, Tinel C et al.   Subclinical rejection phenotypes at 1 year post-transplant and outcome of kidney allografts. J Am Soc Nephrol  2015; 26: 1721– 1731 Google Scholar CrossRef Search ADS PubMed  30 Koo EH, Jang HR, Lee JE et al.   The impact of early and late acute rejection on graft survival in renal transplantation. Kidney Res Clin Pract  2015; 34: 160– 164 Google Scholar CrossRef Search ADS PubMed  31 Wu K, Budde K, Lu H et al.   The severity of acute cellular rejection defined by Banff classification is associated with kidney allograft outcomes. Transplantation  2014; 97: 1146– 1154 Google Scholar CrossRef Search ADS PubMed  32 Krisl JC, Alloway RR, Shield AR et al.   Acute rejection clinically defined phenotypes correlate with long-term renal allograft survival. Transplantation  2015; 99: 2167– 2173 Google Scholar CrossRef Search ADS PubMed  33 Cervera C, Fernandez-Ruiz M, Valledor A et al.   Epidemiology and risk factors for late infection in solid organ transplant recipients. Transpl Infect Dis  2011; 13: 598– 607 Google Scholar CrossRef Search ADS PubMed  34 Abbott KC, Swanson SJ, Richter ER et al.   Late urinary tract infection after renal transplantation in the United States. Am J Kidney Dis  2004; 44: 353– 362 Google Scholar CrossRef Search ADS PubMed  35 Swanson SJ, Hypolite IO, Agodoa LY et al.   Effect of donor factors on early graft survival in adult cadaveric renal transplantation. Am J Transplant  2002; 2: 68– 75 Google Scholar CrossRef Search ADS PubMed  36 Hernandez D, Rufino M, Bartolomei S et al.   A novel prognostic index for mortality in renal transplant recipients after hospitalization. Transplantation  2005; 79: 337– 343 Google Scholar CrossRef Search ADS PubMed  37 Hernandez D, Sanchez-Fructuoso A, Gonzalez-Posada JM et al.   A novel risk score for mortality in renal transplant recipients beyond the first posttransplant year. Transplantation  2009; 88: 803– 809 Google Scholar CrossRef Search ADS PubMed  38 Jassal SV, Schaubel DE, Fenton SS. Predicting mortality after kidney transplantation: a clinical tool. Transplant Int  2005; 18: 1248– 1257 Google Scholar CrossRef Search ADS   39 Machnicki G, Pinsky B, Takemoto S et al.   Predictive ability of pretransplant comorbidities to predict long-term graft loss and death. Am J Transplant  2009; 9: 494– 505 Google Scholar CrossRef Search ADS PubMed  40 Ojo AO, Hanson JA, Wolfe RA et al.   Long-term survival in renal transplant recipients with graft function. Kidney Int  2000; 57: 307– 313 Google Scholar CrossRef Search ADS PubMed  41 Schaeffner ES, Fodinger M, Kramar R et al.   Prognostic associations between lipid markers and outcomes in kidney transplant recipients. Am J Kidney Dis  2006; 47: 509– 517 Google Scholar CrossRef Search ADS PubMed  42 Bansal N, Katz R, De Boer IH et al.   Development and validation of a model to predict 5-year risk of death without ESRD among older adults with CKD. Clin J Am Soc Nephrol  2015; 10: 363– 371 Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Nephrology Dialysis TransplantationOxford University Press

Published: May 31, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off