Prolonged dialysis duration is associated with graft failure and mortality after kidney transplantation: results from the French transplant database

Prolonged dialysis duration is associated with graft failure and mortality after kidney... Abstract Background Kidney transplantation (KT) is the treatment of choice for end-stage renal disease. Preemptive KT (PKT) should be considered when glomerular filtration rate is <15 mL/min/1.73 m2 but European reports on the results of PKT and the effect of pretransplant dialysis are scarce. Methods We analysed all first kidney-only transplants performed in adults in France between 2002 and 2012. A Cox multivariable model was used to investigate the association of PKT and of pretransplant dialysis time with the hazard of graft failure defined as death, return to dialysis or retransplant, whichever occurred first. Results We included 22 345 patients, with a mean ± SD age at KT of 50.5 ± 13.4 years; 61.9% were men and 3112 (14.0%) received a PKT. Median time of follow-up was 4.7 years. Graft failure occurred in 4952 patients up to 31 December 2013. After adjustment for recipients’ age and sex, primary kidney disease, donor type (living or deceased donor, expanded criteria donor), HLA mismatches, cold ischaemia time, centre and year of transplantation, PKT was associated with a decreased hazard of graft failure when compared with pretransplant dialysis [hazard ratio (HR) 0.57; 95% confidence interval (CI) 0.51–0.63], whatever the duration of dialysis, even in the first 6 months. The effect of PKT on the hazard of graft failure was stronger in living kidney donors (HR 0.32; 95% CI 0.19–0.55). Conclusions In France, PKT was associated with a lower risk of graft failure than KT performed after the initiation of dialysis, whatever the duration of dialysis. graft survival, kidney transplantation, patient survival, pre-emptive transplantation, pretransplant dialysis time INTRODUCTION Kidney transplantation (KT) is associated with decreased mortality compared with dialysis [1, 2] and is thus considered the treatment of choice for end-stage renal disease (ESRD). Preemptive KT (PKT) should be considered when the glomerular filtration rate is <15 mL/min/1.73 m2 [3] or a few months before the need to initiate renal replacement therapy (RRT) [4]. In France in 2009–14, only 6.6% of all patients with ESRD were already registered on the waiting list for transplantation at the time of dialysis initiation [5]. In the USA, large studies found that KT performed after a period of dialysis was associated with a higher risk of graft failure and patient death compared with PKT [6, 7]. Results were not consistent regarding the effect of a short period of dialysis of <6 months prior to KT [8–10]. In Japan, pretransplant dialysis time was also associated with an increased risk of graft failure but the impact of a short period of dialysis was not investigated [11]. In Europe, the majority of reports on the results of PKT did not find any graft survival difference between PKT and non-PKT [12–14]. One study found an advantage of PKT in deceased donors only [15] and one study found that the risk of death was increased in patients with dialysis treatment of >1 year [16]. More recently, Haller et al. found that longer waiting time in dialysis was not associated with a higher rate of graft loss, but was associated with a higher rate of death [17]. These differences in results could be explained by some limitations observed in European reports, such as incomplete coverage rates of the registries [12] or small cohorts [13]. Another explanation might be the higher morbidity and mortality of patients treated by dialysis in the USA than in Europe. For instance, the 2-year survival probability of incident dialysis patients in the USA was 62.6% in 2002 [16] and 64.4% in 2008 [18], whereas it was 73% in France between 2002 and 2013 [19]. These differences may also be partly explained by disparities in access to PKT. Indeed, in the USA, PKT occurs most often in Caucasian patients with private insurance or with Medicare as primary payer [6, 20], whereas in France, the access to PKT is not related to medical coverage, which is universal. Finally, despite adjustment for multiple relevant factors, many previous US and European studies may have been limited by residual confounding bias due to unadjusted unbalances between the two groups [4]. Indeed, patients who receive a PKT arise from a group of patients registered on the transplant waiting list before dialysis and are likely to differ from non-PKT patients, in particular, those registered after a period of dialysis. Restricting the analysis to patients registered on the transplant waiting list before initiation of RRT, while still adjusting for observed confounders, should allow a better balance between the PKT and non-PKT groups. However, to our knowledge, no study has performed a comparison of graft survival among patients pre-emptively registered on the waiting list. The primary objective of this study was thus to evaluate the impact of PKT and of pretransplant dialysis duration on patient and graft survival among French adult first kidney transplant recipients. The secondary objective was to evaluate the impact of PKT on graft survival among patients registered on the transplant waiting list before the start of RRT. MATERIALS AND METHODS Data source and ethics statement Data were collected through the French organ transplant registry, which includes details on all KT performed in France since 1994. All 33 French transplantation centres gave their consent to participate in the study. The registry has received approval from the French data protection authority as well as a waiver of obtaining written informed consent. Study population and administrative censoring We included all first kidney-only transplantation performed in adults in France from 1 January 2002 to 31 December 2012. Patients whose age at the registration on the waiting list was under 18 years were excluded. Administrative censoring was performed on 31 December 2013. Data on recipients, donors and transplantation characteristics Data on recipients included age at KT, date of registration on the waiting list, date of KT and date of dialysis initiation if applicable, as well as initial dialysis modality, date of death if applicable, sex, primary kidney disease and blood group. For patients registered on the waiting list after 2007, data on the presence of diabetes, hypertension and cardiovascular comorbidities (defined as previous history of myocardial infarction, cardiac arrhythmia, peripheral arterial disease, stroke or heart failure) at the time of registration were available. Data on human leucocyte antigen (HLA) sensitization, weight and height at the time of registration on the waiting list were available for patients registered after 2011. Data on donors included age, sex and type (deceased or living donor and expanded or standard criteria donor). Data on KT included transplant centre, year of KT, HLA matching and cold ischaemia time. Outcomes The primary event of interest was graft failure defined as death, (return to) dialysis or retransplant, whichever occurred first. Secondary events of interest were (i) (return to) dialysis or retransplant, whichever occurred first and (ii) death with a functioning graft. These two specific secondary events were in competition and were thus analysed using competing risk methods. Statistical analyses We compared patients’ characteristics between those who did and those who did not receive a PKT using chi-squared tests and t-tests when appropriate. In all survival analyses, the time axis was the time elapsed since KT. We estimated the probabilities to survive with a functioning graft at different times after KT and in different groups of patients (PKT, non-PKT and strata of pretransplant dialysis duration), using the Kaplan–Meier estimator. The log-rank test was used to compare the hazard of graft failure in the different groups without any adjustment. The risk of the secondary events of interest was estimated using the Aalen–Johansen estimator to account for competing risks [21], and the different groups were compared using the Gray test [22]. To estimate adjusted hazard ratios (HRs) of graft failure between PKT and non-PKT patients as a whole or according to their pretransplant dialysis duration, we used standard Cox proportional hazard models adjusted for recipient age at KT, recipient sex, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of KT. The adjustment factors were chosen before any analysis based on a priori considerations, using clinical expertise, literature review and directed acyclic graphs [23]. Adjustment for comorbidities (diabetes history and cardiovascular comorbidities history) was performed in a sensitivity analysis within patients receiving a first KT between 2007 and 2012. We also estimated the HR of graft failure between PKT and non-PKT in patients who received a KT from living and deceased donors separately, because of a significant interaction between PKT and donor type. Furthermore, to potentially reduce residual confounding due to unmeasured confounders, we restricted the group of non-PKT to those registered on the waiting list for KT before the initiation of pretransplant dialysis. Finally, to estimate the adjusted HR of the secondary events of interest, we used cause-specific proportional hazards models censoring at competing events [21, 24], adjusted for the same potential confounders as the analysis for the primary event of interest. Log-linearity of the effects of all quantitative variables was checked using spline functions [25, 26]. Proportional hazard assumptions were checked using Schoenfeld residuals. Violation of the linearity assumption for age at transplant and pretransplant dialysis duration was handled by categorizing the variables using quartiles for age at transplant and finer categories for dialysis duration. Violation of proportional hazard assumption for recipient sex was handled by using a stratified Cox model on sex. Time-dependent effect of PKT was handled by using appropriate interaction terms with time since transplantation, resulting in time-dependent HR. The analyses were performed with SAS software 9.4 (SAS Institute Inc., Cary, NC, USA) and R version 3.2.1 (The R Foundation for Statistical Computing, Vienna, Austria). RESULTS Patients’ characteristics During the study period, 22 345 adult patients received a first KT in France, including 22 288 (99.7%) with available data for the analysis. Mean ± SD age of recipients at KT was 50.5 ± 13.4 years, 61.9% were men and 3112 (14.0%) received a PKT. Median time of follow-up was 4.7 years [interquartile range (IQR) 2.2–7.3]. Median pretransplant dialysis duration was 2.3 years (IQR 1.3–4.1) in non-PKT patients (Table 1). Recipients of PKT were younger than non-PKT patients (mean age of 48.8 versus 50.8 years), more frequently female (41.0% versus 37.6%), and less likely to be diabetics, to have cardiovascular comorbidities or to have blood group O (Table 1). PKT recipients had fewer HLA mismatches. The proportion of living donors was greater in PKT (22.2%) than in non-PKT patients (7.0%) (Table 1) and increased over time in PKT, from 16.2% in 2002 to 30.2% in 2012 (data not shown). Donors for PKT were slightly younger and less likely to be expanded criteria donors. Among the subgroup of patients registered on the renal transplant waiting list before RRT, there was no more difference on diabetes history and cardiovascular comorbidities history at the time of registration on the waiting list between patients who did or did not receive a PKT (Supplementary data, Table S1). Table 1 Patients characteristics at time of transplantation in each group (PKT and non-PKT), France, 2002–13 Variables  Available data  PKT  Non-PKT  P-valuea  n = 3112 (14.0%)  n = 19 176 (86.0%)  Recipients characteristics           Male, n (%)  22 288  1835 (59.0)  11 961 (62.4)  0.0003   Age at transplantation (years), mean (SD)  22 288  48.8 (13.8)  50.8 (13.3)  <0.0001   Primary kidney disease, n (%)  21 628      <0.0001    Glomerulonephritis    928 (29.8)  6143 (32.0)      Cystic kidney disease    790 (25.4)  3174 (16.6)      Vascular    113 (3.6)  1202 (6.3)      Diabetics    155 (5.0)  1692 (8.8)      Tubulointerstitial    435 (14.0)  2032 (10.6)      Other or unknown    691 (22.2)  4933 (25.7)     Waiting time on transplantation list (months), median (IQR)  22 288  8.0 (3.1–17.9)  13.8 (5.4–29.4)  <0.0001   Diabetes history, n (%)  9918  171 (10.2)  1493 (18.1)  <0.0001   Cardiovascular comorbidities history, n (%)  8768  230 (15.2)  1864 (25.7)  <0.0001   Hypertension, n (%)  8854  1029 (68.6)  4988 (67.8)  0.56   Panel reactive antibodies, mean (SD)  3750  28.0 (34.3)  31.8 (35.6)  0.02   Blood group, n (%)  22 288      <0.0001    O    1079 (34.7)  8071 (42.1)      A    1619 (52.0)  8251 (43.0)      B    270 (8.7)  2029 (10.6)      AB    144 (4.6)  825 (4.3)     Dialysis modality before transplantb, n (%)  17 929          Peritoneal dialysis      2219 (12.4)      Haemodialysis      15 710 (87.6)     Pretransplant dialysis durationb  18 916          Median in years (IQR)      2.3 (1.3–4.1)      0–6 months, n (%)      918 (4.2)      6–12 months, n (%)      2223 (11.7)      12–24 months; n (%)      4932 (26.1)      24–36 months, n (%)      3979 (21.0)      36–60 months, n (%)      3558 (18.8)      > 60 months, n (%)      3306 (17.5)     BMI (kg/m2), mean (SD)  2725  24.7 (4.1)  25.2 (4.6)  0.04  Donors characteristics           Age (years), mean (SD)  21 845  49.5 (15.7)  50.8 (15.8)  <0.0001   Living donor, n (%)  22 288  690 (22.2)  1341 (7.0)  <0.0001   Expanded criteria donor, n (%)  21 611  1264 (43.6)  9287 (49.5)  <0.0001  KT characteristics         Number of HLA mismatches, n (%)  22 262      <0.0001    0    121 (3.9)  492 (2.5)      1–2    645 (20.8)  3619 (18.9)      3–4    1825 (58.8)  11 893 (62.1)      5–6    514 (16.5)  3153 (16.5)     Cold ischaemia time (h), mean (SD)  21 549  14.9 (8.9)  17.3 (8.1)  <0.0001  Variables  Available data  PKT  Non-PKT  P-valuea  n = 3112 (14.0%)  n = 19 176 (86.0%)  Recipients characteristics           Male, n (%)  22 288  1835 (59.0)  11 961 (62.4)  0.0003   Age at transplantation (years), mean (SD)  22 288  48.8 (13.8)  50.8 (13.3)  <0.0001   Primary kidney disease, n (%)  21 628      <0.0001    Glomerulonephritis    928 (29.8)  6143 (32.0)      Cystic kidney disease    790 (25.4)  3174 (16.6)      Vascular    113 (3.6)  1202 (6.3)      Diabetics    155 (5.0)  1692 (8.8)      Tubulointerstitial    435 (14.0)  2032 (10.6)      Other or unknown    691 (22.2)  4933 (25.7)     Waiting time on transplantation list (months), median (IQR)  22 288  8.0 (3.1–17.9)  13.8 (5.4–29.4)  <0.0001   Diabetes history, n (%)  9918  171 (10.2)  1493 (18.1)  <0.0001   Cardiovascular comorbidities history, n (%)  8768  230 (15.2)  1864 (25.7)  <0.0001   Hypertension, n (%)  8854  1029 (68.6)  4988 (67.8)  0.56   Panel reactive antibodies, mean (SD)  3750  28.0 (34.3)  31.8 (35.6)  0.02   Blood group, n (%)  22 288      <0.0001    O    1079 (34.7)  8071 (42.1)      A    1619 (52.0)  8251 (43.0)      B    270 (8.7)  2029 (10.6)      AB    144 (4.6)  825 (4.3)     Dialysis modality before transplantb, n (%)  17 929          Peritoneal dialysis      2219 (12.4)      Haemodialysis      15 710 (87.6)     Pretransplant dialysis durationb  18 916          Median in years (IQR)      2.3 (1.3–4.1)      0–6 months, n (%)      918 (4.2)      6–12 months, n (%)      2223 (11.7)      12–24 months; n (%)      4932 (26.1)      24–36 months, n (%)      3979 (21.0)      36–60 months, n (%)      3558 (18.8)      > 60 months, n (%)      3306 (17.5)     BMI (kg/m2), mean (SD)  2725  24.7 (4.1)  25.2 (4.6)  0.04  Donors characteristics           Age (years), mean (SD)  21 845  49.5 (15.7)  50.8 (15.8)  <0.0001   Living donor, n (%)  22 288  690 (22.2)  1341 (7.0)  <0.0001   Expanded criteria donor, n (%)  21 611  1264 (43.6)  9287 (49.5)  <0.0001  KT characteristics         Number of HLA mismatches, n (%)  22 262      <0.0001    0    121 (3.9)  492 (2.5)      1–2    645 (20.8)  3619 (18.9)      3–4    1825 (58.8)  11 893 (62.1)      5–6    514 (16.5)  3153 (16.5)     Cold ischaemia time (h), mean (SD)  21 549  14.9 (8.9)  17.3 (8.1)  <0.0001  a t-test or chi-square test. b Among patients who received a non-PKT. BMI, body mass index. Table 1 Patients characteristics at time of transplantation in each group (PKT and non-PKT), France, 2002–13 Variables  Available data  PKT  Non-PKT  P-valuea  n = 3112 (14.0%)  n = 19 176 (86.0%)  Recipients characteristics           Male, n (%)  22 288  1835 (59.0)  11 961 (62.4)  0.0003   Age at transplantation (years), mean (SD)  22 288  48.8 (13.8)  50.8 (13.3)  <0.0001   Primary kidney disease, n (%)  21 628      <0.0001    Glomerulonephritis    928 (29.8)  6143 (32.0)      Cystic kidney disease    790 (25.4)  3174 (16.6)      Vascular    113 (3.6)  1202 (6.3)      Diabetics    155 (5.0)  1692 (8.8)      Tubulointerstitial    435 (14.0)  2032 (10.6)      Other or unknown    691 (22.2)  4933 (25.7)     Waiting time on transplantation list (months), median (IQR)  22 288  8.0 (3.1–17.9)  13.8 (5.4–29.4)  <0.0001   Diabetes history, n (%)  9918  171 (10.2)  1493 (18.1)  <0.0001   Cardiovascular comorbidities history, n (%)  8768  230 (15.2)  1864 (25.7)  <0.0001   Hypertension, n (%)  8854  1029 (68.6)  4988 (67.8)  0.56   Panel reactive antibodies, mean (SD)  3750  28.0 (34.3)  31.8 (35.6)  0.02   Blood group, n (%)  22 288      <0.0001    O    1079 (34.7)  8071 (42.1)      A    1619 (52.0)  8251 (43.0)      B    270 (8.7)  2029 (10.6)      AB    144 (4.6)  825 (4.3)     Dialysis modality before transplantb, n (%)  17 929          Peritoneal dialysis      2219 (12.4)      Haemodialysis      15 710 (87.6)     Pretransplant dialysis durationb  18 916          Median in years (IQR)      2.3 (1.3–4.1)      0–6 months, n (%)      918 (4.2)      6–12 months, n (%)      2223 (11.7)      12–24 months; n (%)      4932 (26.1)      24–36 months, n (%)      3979 (21.0)      36–60 months, n (%)      3558 (18.8)      > 60 months, n (%)      3306 (17.5)     BMI (kg/m2), mean (SD)  2725  24.7 (4.1)  25.2 (4.6)  0.04  Donors characteristics           Age (years), mean (SD)  21 845  49.5 (15.7)  50.8 (15.8)  <0.0001   Living donor, n (%)  22 288  690 (22.2)  1341 (7.0)  <0.0001   Expanded criteria donor, n (%)  21 611  1264 (43.6)  9287 (49.5)  <0.0001  KT characteristics         Number of HLA mismatches, n (%)  22 262      <0.0001    0    121 (3.9)  492 (2.5)      1–2    645 (20.8)  3619 (18.9)      3–4    1825 (58.8)  11 893 (62.1)      5–6    514 (16.5)  3153 (16.5)     Cold ischaemia time (h), mean (SD)  21 549  14.9 (8.9)  17.3 (8.1)  <0.0001  Variables  Available data  PKT  Non-PKT  P-valuea  n = 3112 (14.0%)  n = 19 176 (86.0%)  Recipients characteristics           Male, n (%)  22 288  1835 (59.0)  11 961 (62.4)  0.0003   Age at transplantation (years), mean (SD)  22 288  48.8 (13.8)  50.8 (13.3)  <0.0001   Primary kidney disease, n (%)  21 628      <0.0001    Glomerulonephritis    928 (29.8)  6143 (32.0)      Cystic kidney disease    790 (25.4)  3174 (16.6)      Vascular    113 (3.6)  1202 (6.3)      Diabetics    155 (5.0)  1692 (8.8)      Tubulointerstitial    435 (14.0)  2032 (10.6)      Other or unknown    691 (22.2)  4933 (25.7)     Waiting time on transplantation list (months), median (IQR)  22 288  8.0 (3.1–17.9)  13.8 (5.4–29.4)  <0.0001   Diabetes history, n (%)  9918  171 (10.2)  1493 (18.1)  <0.0001   Cardiovascular comorbidities history, n (%)  8768  230 (15.2)  1864 (25.7)  <0.0001   Hypertension, n (%)  8854  1029 (68.6)  4988 (67.8)  0.56   Panel reactive antibodies, mean (SD)  3750  28.0 (34.3)  31.8 (35.6)  0.02   Blood group, n (%)  22 288      <0.0001    O    1079 (34.7)  8071 (42.1)      A    1619 (52.0)  8251 (43.0)      B    270 (8.7)  2029 (10.6)      AB    144 (4.6)  825 (4.3)     Dialysis modality before transplantb, n (%)  17 929          Peritoneal dialysis      2219 (12.4)      Haemodialysis      15 710 (87.6)     Pretransplant dialysis durationb  18 916          Median in years (IQR)      2.3 (1.3–4.1)      0–6 months, n (%)      918 (4.2)      6–12 months, n (%)      2223 (11.7)      12–24 months; n (%)      4932 (26.1)      24–36 months, n (%)      3979 (21.0)      36–60 months, n (%)      3558 (18.8)      > 60 months, n (%)      3306 (17.5)     BMI (kg/m2), mean (SD)  2725  24.7 (4.1)  25.2 (4.6)  0.04  Donors characteristics           Age (years), mean (SD)  21 845  49.5 (15.7)  50.8 (15.8)  <0.0001   Living donor, n (%)  22 288  690 (22.2)  1341 (7.0)  <0.0001   Expanded criteria donor, n (%)  21 611  1264 (43.6)  9287 (49.5)  <0.0001  KT characteristics         Number of HLA mismatches, n (%)  22 262      <0.0001    0    121 (3.9)  492 (2.5)      1–2    645 (20.8)  3619 (18.9)      3–4    1825 (58.8)  11 893 (62.1)      5–6    514 (16.5)  3153 (16.5)     Cold ischaemia time (h), mean (SD)  21 549  14.9 (8.9)  17.3 (8.1)  <0.0001  a t-test or chi-square test. b Among patients who received a non-PKT. BMI, body mass index. Risk of graft failure in the whole cohort Graft failure occurred in 4952 patients up to 31 December 2013. The probability of surviving with a functioning graft for ≥10 years after KT was 79.9% in PKT patients and 73.8, 70.4, 62.4, 62.8, 59.5 and 56.0% in patients with a pretransplant duration of dialysis of ≤6, >6–12, >12–24, >24–36, >36–60 and >60 months, respectively (P < 0.0001) (Figure 1). The probability of surviving with a functioning graft for ≥10 years after KT was 61.4% in non-PKT patients (Figure 2A). After adjustment for recipient sex and age at KT, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), number of HLA mismatches, cold ischaemia time, centre and year of KT, pretransplant dialysis was associated with an increased hazard of graft failure when compared with PKT, whatever the duration of dialysis (Table 2). Compared with PKT patients, recipients with a pretransplant dialysis duration < 6 months had even a 27% increased hazard of graft failure at any time after KT [HR 1.27; 95% confidence interval (CI) 1.03–1.55]. The HR gradually increased with dialysis duration. Table 2 Effects of PKT and pretransplant dialysis duration on the hazard of graft failure (death, return to dialysis or retransplant, whichever came first), results of Cox proportional hazard model, France, 2002–13 Variables  na  Number of events  HRb  95% CI  P-value  Pretransplant dialysis duration  20 717  4580      <0.0001   0 (PKT)  2770  338  1       0–6 months  818  127  1.27  1.03–1.55     6–12 months  2061  394  1.38  1.19–1.59     12–24 months  4659  1101  1.67  1.48–1.90     24–36 months  3420  794  1.64  1.44–1.86     36–60 months  3823  955  1.82  1.61–2.07     > 60 months  3166  871  2.04  1.79–2.32    Type of transplantation             At any time after transplantation  20 953  4734          Non-PKT  18 183  4396  1        PKT  2770  338  0.57  0.51–0.63  <0.0001   In the first year post-transplant  20 953  1484          Non-PKT  18 183  1358  1        PKT  2770  126  0.69  0.57–0.83  < 0.0001   After 1-year post-transplant  20 953  3250          Non-PKT  18 183  3038  1        PKT  2770  212  0.51  0.45–0.59  0.013   In living donors  1203  116          Non-PKT  791  100  1        PKT  412  16  0.32  0.19–0.55  <0.0001   In deceased donors  19 750  4618          Non-PKT  17 392  4296  1        PKT  2358  322  0.59  0.53–0.64  <0.0001   In patients registered on the waiting list before RRT  4162  512          Non-PKT  1392  174  1        PKT  2770  338  0.70  0.57–0.86  0.0007  Variables  na  Number of events  HRb  95% CI  P-value  Pretransplant dialysis duration  20 717  4580      <0.0001   0 (PKT)  2770  338  1       0–6 months  818  127  1.27  1.03–1.55     6–12 months  2061  394  1.38  1.19–1.59     12–24 months  4659  1101  1.67  1.48–1.90     24–36 months  3420  794  1.64  1.44–1.86     36–60 months  3823  955  1.82  1.61–2.07     > 60 months  3166  871  2.04  1.79–2.32    Type of transplantation             At any time after transplantation  20 953  4734          Non-PKT  18 183  4396  1        PKT  2770  338  0.57  0.51–0.63  <0.0001   In the first year post-transplant  20 953  1484          Non-PKT  18 183  1358  1        PKT  2770  126  0.69  0.57–0.83  < 0.0001   After 1-year post-transplant  20 953  3250          Non-PKT  18 183  3038  1        PKT  2770  212  0.51  0.45–0.59  0.013   In living donors  1203  116          Non-PKT  791  100  1        PKT  412  16  0.32  0.19–0.55  <0.0001   In deceased donors  19 750  4618          Non-PKT  17 392  4296  1        PKT  2358  322  0.59  0.53–0.64  <0.0001   In patients registered on the waiting list before RRT  4162  512          Non-PKT  1392  174  1        PKT  2770  338  0.70  0.57–0.86  0.0007  a Subjects with complete data. b Adjusted for age of recipients at time of transplantation, sex of recipients, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of transplantation. Table 2 Effects of PKT and pretransplant dialysis duration on the hazard of graft failure (death, return to dialysis or retransplant, whichever came first), results of Cox proportional hazard model, France, 2002–13 Variables  na  Number of events  HRb  95% CI  P-value  Pretransplant dialysis duration  20 717  4580      <0.0001   0 (PKT)  2770  338  1       0–6 months  818  127  1.27  1.03–1.55     6–12 months  2061  394  1.38  1.19–1.59     12–24 months  4659  1101  1.67  1.48–1.90     24–36 months  3420  794  1.64  1.44–1.86     36–60 months  3823  955  1.82  1.61–2.07     > 60 months  3166  871  2.04  1.79–2.32    Type of transplantation             At any time after transplantation  20 953  4734          Non-PKT  18 183  4396  1        PKT  2770  338  0.57  0.51–0.63  <0.0001   In the first year post-transplant  20 953  1484          Non-PKT  18 183  1358  1        PKT  2770  126  0.69  0.57–0.83  < 0.0001   After 1-year post-transplant  20 953  3250          Non-PKT  18 183  3038  1        PKT  2770  212  0.51  0.45–0.59  0.013   In living donors  1203  116          Non-PKT  791  100  1        PKT  412  16  0.32  0.19–0.55  <0.0001   In deceased donors  19 750  4618          Non-PKT  17 392  4296  1        PKT  2358  322  0.59  0.53–0.64  <0.0001   In patients registered on the waiting list before RRT  4162  512          Non-PKT  1392  174  1        PKT  2770  338  0.70  0.57–0.86  0.0007  Variables  na  Number of events  HRb  95% CI  P-value  Pretransplant dialysis duration  20 717  4580      <0.0001   0 (PKT)  2770  338  1       0–6 months  818  127  1.27  1.03–1.55     6–12 months  2061  394  1.38  1.19–1.59     12–24 months  4659  1101  1.67  1.48–1.90     24–36 months  3420  794  1.64  1.44–1.86     36–60 months  3823  955  1.82  1.61–2.07     > 60 months  3166  871  2.04  1.79–2.32    Type of transplantation             At any time after transplantation  20 953  4734          Non-PKT  18 183  4396  1        PKT  2770  338  0.57  0.51–0.63  <0.0001   In the first year post-transplant  20 953  1484          Non-PKT  18 183  1358  1        PKT  2770  126  0.69  0.57–0.83  < 0.0001   After 1-year post-transplant  20 953  3250          Non-PKT  18 183  3038  1        PKT  2770  212  0.51  0.45–0.59  0.013   In living donors  1203  116          Non-PKT  791  100  1        PKT  412  16  0.32  0.19–0.55  <0.0001   In deceased donors  19 750  4618          Non-PKT  17 392  4296  1        PKT  2358  322  0.59  0.53–0.64  <0.0001   In patients registered on the waiting list before RRT  4162  512          Non-PKT  1392  174  1        PKT  2770  338  0.70  0.57–0.86  0.0007  a Subjects with complete data. b Adjusted for age of recipients at time of transplantation, sex of recipients, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of transplantation. FIGURE 1 View largeDownload slide Probability of surviving without retransplant or dialysis after transplantation (with 95% CI) according to the pretransplant dialysis duration (Kaplan–Meier estimator). FIGURE 1 View largeDownload slide Probability of surviving without retransplant or dialysis after transplantation (with 95% CI) according to the pretransplant dialysis duration (Kaplan–Meier estimator). FIGURE 2 View largeDownload slide Probability of surviving without retransplant or dialysis after transplantation (with 95% CI) in each group of transplantation (Kaplan–Meier estimator). (A) Analysis in the whole cohort. (B) Analysis in the subgroup of patients registered pre-emptively on the waiting list. FIGURE 2 View largeDownload slide Probability of surviving without retransplant or dialysis after transplantation (with 95% CI) in each group of transplantation (Kaplan–Meier estimator). (A) Analysis in the whole cohort. (B) Analysis in the subgroup of patients registered pre-emptively on the waiting list. Overall, PKT was associated with a 43% reduction in the hazard of graft failure at any time after KT when compared with patients treated by dialysis prior to KT (HR 0.57; 95% CI 0.51–0.63) (Table 2). This reduction in the hazard of graft failure was greater after the first year post-KT (Table 2). Specifically, the hazard of graft failure was reduced by 31% during the first year post-KT in patients with PKT compared with those with non-PKT (HR 0.69; 95% CI 0.57–0.83) and by 49% beyond 1 year (HR 0.51; 95% CI 0.45–0.59). Among patients who received a first KT between 2007 and 2012, further adjustment for cardiovascular comorbidities and diabetes did not change the association of PKT with the hazard of graft failure (data not shown). The reduction of hazard of graft failure in PKT patients was stronger in living than in deceased donor recipients. In living donor recipients, PKT was associated with a 68% reduction in the hazard of graft failure when compared with non-PKT (HR 0.32; 95% CI 0.19–0.55), whereas it was associated with a 41% reduction (HR 0.59; 95% CI 0.53–0.64) in deceased donor recipients (Table 2). Risk of graft failure among patients registered on the waiting list before RRT In the subgroup of patients pre-emptively registered, the association of PKT with graft failure was weaker than in the whole study population but remained significant (Figure 2B). Specifically, the probability of surviving with a functioning graft for ≥10 years after KT was 79.9% in PKT patients and 75.0% in non-PKT patients (P < 0.001) (Figure 2B). After adjustment for the same covariables as for the whole cohort, PKT was associated with a 30% reduction in the hazard of graft failure at any time after KT, when compared with non-PKT (HR 0.70; 95% CI 0.57–0.86) (Table 2). This result was observed while the median pretransplant dialysis duration was lower in patients with non-PKT registered on the waiting list before RRT than in the whole population of non-PKT patients (1.1 years in Supplementary data, Table S1 versus 2.3 years in Table 1). Risk of dialysis/retransplantation and death with a functioning graft The probabilities to return to dialysis or to be retransplanted, and to die with a functioning graft within 10 years after KT were, respectively, 9.5% and 10.6% in patients with PKT and, 21.2% and 17.4% in patients with non-PKT (P < 0.0001) (Supplementary data, Figure S1). Compared with non-PKT, PKT was associated with a 45% reduction in the hazard of dialysis/retransplantation (HR 0.55; 95% CI 0.47–0.64) and with a 40% reduction in the hazard of death with a functioning graft (HR 0.60; 95% CI 0.50–0.71) (Table 3), after adjustment for the same factors as previously. These results were confirmed in the subgroup of patients registered on the waiting list before dialysis for the event dialysis/retransplantation (HR 0.59; 95% CI 0.45–0.78) but not for death with a functioning graft (HR 0.87; 95% CI 0.62–1.23) (Table 3). Table 3 Effect of pre-emptive transplantation on the hazard of retransplant or return to dialysis, and death with a functioning graft, results of cause-specific proportional hazard model, France, 2002–13 Variables  Retransplant or return to dialysis, whichever came first (censored at death with a functioning graft)   Death with a functioning graft (censored at retransplant or return to dialysis, whichever came first)   HRa  95% CI  P-value  HRa  95% CI  P-value  Type of transplantation   At any time after transplantation (n = 20 953)    Non-PKT  1      1        PKT  0.55  0.47–0.64  <0.0001  0.60  0.50–0.71  <0.0001   In patients registered on the waiting list before RRT (n = 4162)    Non-PKT  1      1        PKT  0.59  0.45–0.78  0.0002  0.87  0.62–1.23  0.44  Variables  Retransplant or return to dialysis, whichever came first (censored at death with a functioning graft)   Death with a functioning graft (censored at retransplant or return to dialysis, whichever came first)   HRa  95% CI  P-value  HRa  95% CI  P-value  Type of transplantation   At any time after transplantation (n = 20 953)    Non-PKT  1      1        PKT  0.55  0.47–0.64  <0.0001  0.60  0.50–0.71  <0.0001   In patients registered on the waiting list before RRT (n = 4162)    Non-PKT  1      1        PKT  0.59  0.45–0.78  0.0002  0.87  0.62–1.23  0.44  a Adjusted for age of recipients at time of transplantation, sex of recipients, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of transplantation. Table 3 Effect of pre-emptive transplantation on the hazard of retransplant or return to dialysis, and death with a functioning graft, results of cause-specific proportional hazard model, France, 2002–13 Variables  Retransplant or return to dialysis, whichever came first (censored at death with a functioning graft)   Death with a functioning graft (censored at retransplant or return to dialysis, whichever came first)   HRa  95% CI  P-value  HRa  95% CI  P-value  Type of transplantation   At any time after transplantation (n = 20 953)    Non-PKT  1      1        PKT  0.55  0.47–0.64  <0.0001  0.60  0.50–0.71  <0.0001   In patients registered on the waiting list before RRT (n = 4162)    Non-PKT  1      1        PKT  0.59  0.45–0.78  0.0002  0.87  0.62–1.23  0.44  Variables  Retransplant or return to dialysis, whichever came first (censored at death with a functioning graft)   Death with a functioning graft (censored at retransplant or return to dialysis, whichever came first)   HRa  95% CI  P-value  HRa  95% CI  P-value  Type of transplantation   At any time after transplantation (n = 20 953)    Non-PKT  1      1        PKT  0.55  0.47–0.64  <0.0001  0.60  0.50–0.71  <0.0001   In patients registered on the waiting list before RRT (n = 4162)    Non-PKT  1      1        PKT  0.59  0.45–0.78  0.0002  0.87  0.62–1.23  0.44  a Adjusted for age of recipients at time of transplantation, sex of recipients, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of transplantation. DISCUSSION Using data from the French organ transplant registry, we found that pretransplant dialysis duration was associated with an increased hazard of graft failure at any time after transplantation when compared with patients receiving a PKT, whatever dialysis duration. PKT was associated with a decreased risk of graft failure whether or not death with a functioning graft was included in the definition of graft failure. We also found a substantial reduction of the hazard of graft failure in PKT patients even after restricting the non-PKT group to more comparable patients, that is, those registered on the waiting list before dialysis initiation. A further reduction of the risk of graft failure was observed when kidney transplant was from a living donor. These results are consistent with American studies [6, 7], despite differences in access to transplantation and a lower mortality in dialysis in France. One of the major differences between previous studies and our work is that our results suggest that even a short period on dialysis (≤6 months) is associated with poorer outcome after KT. Indeed, American, Japanese and Finnish cohort studies found an increased risk of graft failure after 6 months of dialysis or did not explore the effect of a short period of dialysis [9, 11, 16]. Avoiding dialysis as far as possible should thus always be considered, especially for patients waiting for a living donor graft. This message is important, because it is counter-intuitive for many physicians. Our results differ more from previous European reports, which mostly did not find significant graft failure risk differences between PKT and non-PKT [12, 13]. This may partly be explained by the fact that our study is based on a larger sample size, a longer inclusion period and an exhaustive national transplant registry. As all previous studies that compared the risk of graft failure in PKT and non-PKT patients, our study may suffer from important residual confounding due to unmeasured confounders. First, PKT recipients may have a higher level of residual function of their native kidney at transplantation. This could also explain why we observed a stronger reduction in the hazard of graft failure after the first year post-transplantation than within the first year after KT. As most previous studies, we could not adjust for residual renal function because these data were not available in patients who started dialysis before transplantation. However, two studies found that a higher level of residual renal function was not associated with a better graft survival in patients who received a PKT [27, 28], suggesting that residual renal function might not be a major factor. A second reason that could explain the observed lower risk of graft failure in PKT recipients is that these patients might be less likely to have comorbidities, as in our study, than those transplanted after initiation of dialysis. However, after adjustment for cardiovascular comorbidities and diabetes, our results were similar. The reduction of hazard in PKT also persisted, although to a lesser extent, after restriction of the non-PKT patients to those registered on the waiting list before KT. Other potential factors may explain the observed lower risk in PKT recipients, such as socio-economic status, nutritional status, sensitization before transplant and adherence to immunosuppressive treatment. For example, social deprivation was associated with a decreased access to waiting list even in countries with universal health coverage, including France, and with increased risk of death after KT [29–32]. The lack of information about sensitization of the waitlisted patients is an important weakness, because preformed for donor-specific HLA antibodies (HLA-DSAs) are a considerable risk factor for graft survival. Furthermore, patients without preformed HLA-DSAs are transplanted faster and it is possible that PKT patients in our study were less likely to be sensitized than non-PKT patients. However, adjustment for all these factors could not be performed because they were not, or only recently, collected in the registry. Furthermore, it could have been interesting to adjust our analyses for time from the first nephrology consultation to transplant but these data are not recorded in the French RRT registry. To reduce residual confounding due to unmeasured confounders, we performed an analysis where we restricted the non-PKT group to patients registered on the transplant waiting list prior to the initiation of dialysis. Indeed, patients who have been registered on the waiting list before dialysis initiation are more likely to be comparable with PKT patients, as we observed in our data with respect to history of diabetes and cardiovascular diseases at baseline. We may assume that if this analysis reduced unbalances between PKT and non-PKT patients in terms of these comorbidities, then it might have also reduced unbalances on other unmeasured confounders. As expected, we showed a lower reduction of the hazard of graft failure in PKT patients compared with non-PKT patients. However, a reduction of 30% of the hazard at each time after transplantation is still substantial, even if we acknowledge that this analysis could not solve all the residual confounding issue. We found that the risk of graft failure, and not the risk of death with a functioning graft, was statistically significantly reduced with PKT in pre-emptively registered patients, suggesting that PKT has a stronger impact on graft survival than on patient survival. This also suggests that patients registered pre-emptively on the waiting list are selected among patients with a better health status. It should also be noted that we included only patients who were transplanted and did not intend to assess the effect of PKT in the population of all patients registered on the waiting list before RRT. Indeed, such a study would have to account for the fact that some patients of the target population may die prior to KT, and thus will never be at risk of graft failure. In conclusion, although our study suffers from potential residual confounding due to unmeasured confounders as previous studies, we believe that our findings tend to confirm that KT performed after the initiation of dialysis may be associated with an increased risk of graft failure at any time after transplantation, compared with PKT, even for a dialysis duration of less than 6 months. Overall, our results suggest that nephrologists should more frequently consider the registration of their patients on the waiting list before dialysis initiation. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS All authors thank all the 33 French kidney transplant centres that participated in the study. FUNDING M.P.-R. reports grants from la Fondation pour la Recherche Médicale (FRM) during the conduct of the study. The funding source did not have any impact on the study design or the decision to submit the paper. AUTHORS’ CONTRIBUTIONS M.P.-R., J.H. and K.L. contributed to literature search, figures, study design, data analysis, data interpretation and writing of the paper. C.C., P.M. and L.C. contributed to study design, data interpretation and writing of the paper. C.J. contributed to study design, data collection and writing of the paper. REFERENCES 1 Wolfe RA, Ashby VB, Milford EL et al.   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  2 Schnuelle P, Lorenz D, Trede M et al.   Impact of renal cadaveric transplantation on survival in end-stage renal failure: evidence for reduced mortality risk compared with hemodialysis during long-term follow-up. J Am Soc Nephrol  1998; 9: 2135– 2141 Google Scholar PubMed  3 Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 Clinical Practice Guideline for the Evaluation and the Management of Chronic Kidney Disease. Kidney Int Suppl  2013; 3: 1– 150 CrossRef Search ADS   4 Abramowicz D, Hazzan M, Maggiore U et al.   Does pre-emptive transplantation versus post start of dialysis transplantation with a kidney from a living donor improve outcomes after transplantation? A systematic literature review and position statement by the Descartes Working Group and ERBP. Nephrol Dial Transplant  2016; 31: 691– 697 Google Scholar CrossRef Search ADS PubMed  5 Registre Epidémiologie et Information en Néphrologie. Rapport REIN 2014. https://www.agence-biomedecine.fr/IMG/pdf/rapportrein2014.pdf (25 August 2017, date last accessed) 6 Kasiske BL, Snyder JJ, Matas AJ et al.   Preemptive kidney transplantation: the advantage and the advantaged. J Am Soc Nephrol  2002; 13: 1358– 1364 Google Scholar CrossRef Search ADS PubMed  7 Mange KC, Joffe MM, Feldman HI. Effect of the use or non-use of long-term dialysis on the subsequent survival of renal transplants from living donors. N Engl J Med  2001; 344: 726– 731 Google Scholar CrossRef Search ADS PubMed  8 Meier-Kriesche HU, Port FK, Ojo AO et al.   Effect of waiting time on renal transplant outcome. Kidney Int  2000; 58: 1311– 1317 Google Scholar CrossRef Search ADS PubMed  9 Goldfarb-Rumyantzev A, Hurdle JF, Scandling J et al.   Duration of end-stage renal disease and kidney transplant outcome. Nephrol Dial Transplant  2005; 20: 167– 175 Google Scholar CrossRef Search ADS PubMed  10 Jay C, Dean P, Helmick R et al.   Reassessing preemptive kidney transplantation in the United States: are we making progress? Transplantation  2016; 100: 1120– 1127 Google Scholar CrossRef Search ADS PubMed  11 Goto N, Okada M, Yamamoto T et al.   Association of dialysis duration with outcomes after transplantation in a Japanese cohort. Clin J Am Soc Nephrol  2016; 11: 497– 504 Google Scholar CrossRef Search ADS PubMed  12 Berthoux FC, Jones EH, Mehls O et al.   Transplantation report. 2: preemptive renal transplantation in adults aged over 15 years. The EDTA-ERA Registry. European Dialysis and Transplant Association-European Renal Association. Nephrol Dial Transplant  1996; 11 (Suppl 1): 41– 43 Google Scholar CrossRef Search ADS PubMed  13 Kessler M, Ladriere M, Giral M et al.   Does pre-emptive kidney transplantation with a deceased donor improve outcomes? Results from a French transplant network: outcome of pre-emptive kidney transplantation. Transpl Int  2011; 24: 266– 275 Google Scholar CrossRef Search ADS PubMed  14 Pérez-Flores I, Sánchez-Fructuoso A, Calvo N et al.   Preemptive kidney transplant from deceased donors: an advantage in relation to reduced waiting list. Transplant Proc  2007; 39: 2123– 2124 Google Scholar CrossRef Search ADS PubMed  15 Witczak BJ, Leivestad T, Line PD et al.   Experience from an active preemptive kidney transplantation program–809 cases revisited. Transplantation  2009; 88: 672– 677 Google Scholar CrossRef Search ADS PubMed  16 Helanterä I, Salmela K, Kyllönen L et al.   Pretransplant dialysis duration and risk of death after kidney transplantation in the current era. Transplantation  2014; 98: 458– 464 Google Scholar CrossRef Search ADS PubMed  17 Haller MC, Kainz A, Baer H et al.   Dialysis vintage and outcomes after kidney transplantation: a retrospective cohort study. Clin J Am Soc Nephrol  2017; 12: 122– 130 Google Scholar CrossRef Search ADS PubMed  18 United States Renal Data System. 2015 USRDS annual data report: epidemiology of kidney disease in the United States. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 2015 19 Registre Epidémiologie et Information en Néphrologie. Rapport REIN 2013. https://www.agence-biomedecine.fr/IMG/pdf/rapport_rein2013.pdf (25 August 2017, date last accessed) 20 Grams ME, Chen BP-H, Coresh J et al.   Preemptive deceased donor kidney transplantation: considerations of equity and utility. Clin J Am Soc Nephrol  2013; 8: 575– 582 Google Scholar CrossRef Search ADS PubMed  21 Andersen PK, Geskus RB, de Witte T et al.   Competing risks in epidemiology: possibilities and pitfalls. Int J Epidemiol  2012; 41: 861– 870 Google Scholar CrossRef Search ADS PubMed  22 Gray RJA. Class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat  1988; 16: 1141– 1154 Google Scholar CrossRef Search ADS   23 Suttorp MM, Siegerink B, Jager KJ et al.   Graphical presentation of confounding in directed acyclic graphs. Nephrol Dial Transplant  2015; 30: 1418– 1423 Google Scholar CrossRef Search ADS PubMed  24 Noordzij M, Leffondré K, van Stralen KJ et al.   When do we need competing risks methods for survival analysis in nephrology? Nephrol Dial Transplant  2013; 28: 2670– 2677 Google Scholar CrossRef Search ADS PubMed  25 Eilers PHC, Marx BD. Flexible smoothing with B-splines and penalties. Statist Sci  1996; 11: 89– 121 Google Scholar CrossRef Search ADS   26 Leffondré K, Jager KJ, Boucquemont J et al.   Representation of exposures in regression analysis and interpretation of regression coefficients: basic concepts and pitfalls. Nephrol Dial Transplant  2014; 29: 1806– 1814 Google Scholar CrossRef Search ADS PubMed  27 Akkina SK, Connaire JJ, Snyder JJ et al.   Earlier is not necessarily better in preemptive kidney transplantation. Am J Transplant  2008; 8: 2071– 2076 Google Scholar CrossRef Search ADS PubMed  28 Grams ME, Massie AB, Coresh J et al.   Trends in the timing of pre-emptive kidney transplantation. J Am Soc Nephrol  2011; 22: 1615– 1620 Google Scholar CrossRef Search ADS PubMed  29 Riffaut N, Lobbedez T, Hazzan M et al.   Access to preemptive registration on the waiting list for renal transplantation: a hierarchical modeling approach. Transpl Int  2015; 28: 1066– 1073 Google Scholar CrossRef Search ADS PubMed  30 Kasiske BL, London W, Ellison MD. Race and socioeconomic factors influencing early placement on the kidney transplant waiting list. J Am Soc Nephrol  1998; 9: 2142– 2147 Google Scholar PubMed  31 Udayaraj U, Ben-Shlomo Y, Roderick P et al.   Social deprivation, ethnicity, and access to the deceased donor kidney transplant waiting list in England and Wales. Transplantation  2010; 90: 279– 285 Google Scholar CrossRef Search ADS PubMed  32 Begaj I, Khosla S, Ray D et al.   Socioeconomic deprivation is independently associated with mortality post kidney transplantation. Kidney Int  2013; 84: 803– 809 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

Prolonged dialysis duration is associated with graft failure and mortality after kidney transplantation: results from the French transplant database

Loading next page...
 
/lp/ou_press/prolonged-dialysis-duration-is-associated-with-graft-failure-and-p9yn2K0ecH
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/gfy039
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Kidney transplantation (KT) is the treatment of choice for end-stage renal disease. Preemptive KT (PKT) should be considered when glomerular filtration rate is <15 mL/min/1.73 m2 but European reports on the results of PKT and the effect of pretransplant dialysis are scarce. Methods We analysed all first kidney-only transplants performed in adults in France between 2002 and 2012. A Cox multivariable model was used to investigate the association of PKT and of pretransplant dialysis time with the hazard of graft failure defined as death, return to dialysis or retransplant, whichever occurred first. Results We included 22 345 patients, with a mean ± SD age at KT of 50.5 ± 13.4 years; 61.9% were men and 3112 (14.0%) received a PKT. Median time of follow-up was 4.7 years. Graft failure occurred in 4952 patients up to 31 December 2013. After adjustment for recipients’ age and sex, primary kidney disease, donor type (living or deceased donor, expanded criteria donor), HLA mismatches, cold ischaemia time, centre and year of transplantation, PKT was associated with a decreased hazard of graft failure when compared with pretransplant dialysis [hazard ratio (HR) 0.57; 95% confidence interval (CI) 0.51–0.63], whatever the duration of dialysis, even in the first 6 months. The effect of PKT on the hazard of graft failure was stronger in living kidney donors (HR 0.32; 95% CI 0.19–0.55). Conclusions In France, PKT was associated with a lower risk of graft failure than KT performed after the initiation of dialysis, whatever the duration of dialysis. graft survival, kidney transplantation, patient survival, pre-emptive transplantation, pretransplant dialysis time INTRODUCTION Kidney transplantation (KT) is associated with decreased mortality compared with dialysis [1, 2] and is thus considered the treatment of choice for end-stage renal disease (ESRD). Preemptive KT (PKT) should be considered when the glomerular filtration rate is <15 mL/min/1.73 m2 [3] or a few months before the need to initiate renal replacement therapy (RRT) [4]. In France in 2009–14, only 6.6% of all patients with ESRD were already registered on the waiting list for transplantation at the time of dialysis initiation [5]. In the USA, large studies found that KT performed after a period of dialysis was associated with a higher risk of graft failure and patient death compared with PKT [6, 7]. Results were not consistent regarding the effect of a short period of dialysis of <6 months prior to KT [8–10]. In Japan, pretransplant dialysis time was also associated with an increased risk of graft failure but the impact of a short period of dialysis was not investigated [11]. In Europe, the majority of reports on the results of PKT did not find any graft survival difference between PKT and non-PKT [12–14]. One study found an advantage of PKT in deceased donors only [15] and one study found that the risk of death was increased in patients with dialysis treatment of >1 year [16]. More recently, Haller et al. found that longer waiting time in dialysis was not associated with a higher rate of graft loss, but was associated with a higher rate of death [17]. These differences in results could be explained by some limitations observed in European reports, such as incomplete coverage rates of the registries [12] or small cohorts [13]. Another explanation might be the higher morbidity and mortality of patients treated by dialysis in the USA than in Europe. For instance, the 2-year survival probability of incident dialysis patients in the USA was 62.6% in 2002 [16] and 64.4% in 2008 [18], whereas it was 73% in France between 2002 and 2013 [19]. These differences may also be partly explained by disparities in access to PKT. Indeed, in the USA, PKT occurs most often in Caucasian patients with private insurance or with Medicare as primary payer [6, 20], whereas in France, the access to PKT is not related to medical coverage, which is universal. Finally, despite adjustment for multiple relevant factors, many previous US and European studies may have been limited by residual confounding bias due to unadjusted unbalances between the two groups [4]. Indeed, patients who receive a PKT arise from a group of patients registered on the transplant waiting list before dialysis and are likely to differ from non-PKT patients, in particular, those registered after a period of dialysis. Restricting the analysis to patients registered on the transplant waiting list before initiation of RRT, while still adjusting for observed confounders, should allow a better balance between the PKT and non-PKT groups. However, to our knowledge, no study has performed a comparison of graft survival among patients pre-emptively registered on the waiting list. The primary objective of this study was thus to evaluate the impact of PKT and of pretransplant dialysis duration on patient and graft survival among French adult first kidney transplant recipients. The secondary objective was to evaluate the impact of PKT on graft survival among patients registered on the transplant waiting list before the start of RRT. MATERIALS AND METHODS Data source and ethics statement Data were collected through the French organ transplant registry, which includes details on all KT performed in France since 1994. All 33 French transplantation centres gave their consent to participate in the study. The registry has received approval from the French data protection authority as well as a waiver of obtaining written informed consent. Study population and administrative censoring We included all first kidney-only transplantation performed in adults in France from 1 January 2002 to 31 December 2012. Patients whose age at the registration on the waiting list was under 18 years were excluded. Administrative censoring was performed on 31 December 2013. Data on recipients, donors and transplantation characteristics Data on recipients included age at KT, date of registration on the waiting list, date of KT and date of dialysis initiation if applicable, as well as initial dialysis modality, date of death if applicable, sex, primary kidney disease and blood group. For patients registered on the waiting list after 2007, data on the presence of diabetes, hypertension and cardiovascular comorbidities (defined as previous history of myocardial infarction, cardiac arrhythmia, peripheral arterial disease, stroke or heart failure) at the time of registration were available. Data on human leucocyte antigen (HLA) sensitization, weight and height at the time of registration on the waiting list were available for patients registered after 2011. Data on donors included age, sex and type (deceased or living donor and expanded or standard criteria donor). Data on KT included transplant centre, year of KT, HLA matching and cold ischaemia time. Outcomes The primary event of interest was graft failure defined as death, (return to) dialysis or retransplant, whichever occurred first. Secondary events of interest were (i) (return to) dialysis or retransplant, whichever occurred first and (ii) death with a functioning graft. These two specific secondary events were in competition and were thus analysed using competing risk methods. Statistical analyses We compared patients’ characteristics between those who did and those who did not receive a PKT using chi-squared tests and t-tests when appropriate. In all survival analyses, the time axis was the time elapsed since KT. We estimated the probabilities to survive with a functioning graft at different times after KT and in different groups of patients (PKT, non-PKT and strata of pretransplant dialysis duration), using the Kaplan–Meier estimator. The log-rank test was used to compare the hazard of graft failure in the different groups without any adjustment. The risk of the secondary events of interest was estimated using the Aalen–Johansen estimator to account for competing risks [21], and the different groups were compared using the Gray test [22]. To estimate adjusted hazard ratios (HRs) of graft failure between PKT and non-PKT patients as a whole or according to their pretransplant dialysis duration, we used standard Cox proportional hazard models adjusted for recipient age at KT, recipient sex, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of KT. The adjustment factors were chosen before any analysis based on a priori considerations, using clinical expertise, literature review and directed acyclic graphs [23]. Adjustment for comorbidities (diabetes history and cardiovascular comorbidities history) was performed in a sensitivity analysis within patients receiving a first KT between 2007 and 2012. We also estimated the HR of graft failure between PKT and non-PKT in patients who received a KT from living and deceased donors separately, because of a significant interaction between PKT and donor type. Furthermore, to potentially reduce residual confounding due to unmeasured confounders, we restricted the group of non-PKT to those registered on the waiting list for KT before the initiation of pretransplant dialysis. Finally, to estimate the adjusted HR of the secondary events of interest, we used cause-specific proportional hazards models censoring at competing events [21, 24], adjusted for the same potential confounders as the analysis for the primary event of interest. Log-linearity of the effects of all quantitative variables was checked using spline functions [25, 26]. Proportional hazard assumptions were checked using Schoenfeld residuals. Violation of the linearity assumption for age at transplant and pretransplant dialysis duration was handled by categorizing the variables using quartiles for age at transplant and finer categories for dialysis duration. Violation of proportional hazard assumption for recipient sex was handled by using a stratified Cox model on sex. Time-dependent effect of PKT was handled by using appropriate interaction terms with time since transplantation, resulting in time-dependent HR. The analyses were performed with SAS software 9.4 (SAS Institute Inc., Cary, NC, USA) and R version 3.2.1 (The R Foundation for Statistical Computing, Vienna, Austria). RESULTS Patients’ characteristics During the study period, 22 345 adult patients received a first KT in France, including 22 288 (99.7%) with available data for the analysis. Mean ± SD age of recipients at KT was 50.5 ± 13.4 years, 61.9% were men and 3112 (14.0%) received a PKT. Median time of follow-up was 4.7 years [interquartile range (IQR) 2.2–7.3]. Median pretransplant dialysis duration was 2.3 years (IQR 1.3–4.1) in non-PKT patients (Table 1). Recipients of PKT were younger than non-PKT patients (mean age of 48.8 versus 50.8 years), more frequently female (41.0% versus 37.6%), and less likely to be diabetics, to have cardiovascular comorbidities or to have blood group O (Table 1). PKT recipients had fewer HLA mismatches. The proportion of living donors was greater in PKT (22.2%) than in non-PKT patients (7.0%) (Table 1) and increased over time in PKT, from 16.2% in 2002 to 30.2% in 2012 (data not shown). Donors for PKT were slightly younger and less likely to be expanded criteria donors. Among the subgroup of patients registered on the renal transplant waiting list before RRT, there was no more difference on diabetes history and cardiovascular comorbidities history at the time of registration on the waiting list between patients who did or did not receive a PKT (Supplementary data, Table S1). Table 1 Patients characteristics at time of transplantation in each group (PKT and non-PKT), France, 2002–13 Variables  Available data  PKT  Non-PKT  P-valuea  n = 3112 (14.0%)  n = 19 176 (86.0%)  Recipients characteristics           Male, n (%)  22 288  1835 (59.0)  11 961 (62.4)  0.0003   Age at transplantation (years), mean (SD)  22 288  48.8 (13.8)  50.8 (13.3)  <0.0001   Primary kidney disease, n (%)  21 628      <0.0001    Glomerulonephritis    928 (29.8)  6143 (32.0)      Cystic kidney disease    790 (25.4)  3174 (16.6)      Vascular    113 (3.6)  1202 (6.3)      Diabetics    155 (5.0)  1692 (8.8)      Tubulointerstitial    435 (14.0)  2032 (10.6)      Other or unknown    691 (22.2)  4933 (25.7)     Waiting time on transplantation list (months), median (IQR)  22 288  8.0 (3.1–17.9)  13.8 (5.4–29.4)  <0.0001   Diabetes history, n (%)  9918  171 (10.2)  1493 (18.1)  <0.0001   Cardiovascular comorbidities history, n (%)  8768  230 (15.2)  1864 (25.7)  <0.0001   Hypertension, n (%)  8854  1029 (68.6)  4988 (67.8)  0.56   Panel reactive antibodies, mean (SD)  3750  28.0 (34.3)  31.8 (35.6)  0.02   Blood group, n (%)  22 288      <0.0001    O    1079 (34.7)  8071 (42.1)      A    1619 (52.0)  8251 (43.0)      B    270 (8.7)  2029 (10.6)      AB    144 (4.6)  825 (4.3)     Dialysis modality before transplantb, n (%)  17 929          Peritoneal dialysis      2219 (12.4)      Haemodialysis      15 710 (87.6)     Pretransplant dialysis durationb  18 916          Median in years (IQR)      2.3 (1.3–4.1)      0–6 months, n (%)      918 (4.2)      6–12 months, n (%)      2223 (11.7)      12–24 months; n (%)      4932 (26.1)      24–36 months, n (%)      3979 (21.0)      36–60 months, n (%)      3558 (18.8)      > 60 months, n (%)      3306 (17.5)     BMI (kg/m2), mean (SD)  2725  24.7 (4.1)  25.2 (4.6)  0.04  Donors characteristics           Age (years), mean (SD)  21 845  49.5 (15.7)  50.8 (15.8)  <0.0001   Living donor, n (%)  22 288  690 (22.2)  1341 (7.0)  <0.0001   Expanded criteria donor, n (%)  21 611  1264 (43.6)  9287 (49.5)  <0.0001  KT characteristics         Number of HLA mismatches, n (%)  22 262      <0.0001    0    121 (3.9)  492 (2.5)      1–2    645 (20.8)  3619 (18.9)      3–4    1825 (58.8)  11 893 (62.1)      5–6    514 (16.5)  3153 (16.5)     Cold ischaemia time (h), mean (SD)  21 549  14.9 (8.9)  17.3 (8.1)  <0.0001  Variables  Available data  PKT  Non-PKT  P-valuea  n = 3112 (14.0%)  n = 19 176 (86.0%)  Recipients characteristics           Male, n (%)  22 288  1835 (59.0)  11 961 (62.4)  0.0003   Age at transplantation (years), mean (SD)  22 288  48.8 (13.8)  50.8 (13.3)  <0.0001   Primary kidney disease, n (%)  21 628      <0.0001    Glomerulonephritis    928 (29.8)  6143 (32.0)      Cystic kidney disease    790 (25.4)  3174 (16.6)      Vascular    113 (3.6)  1202 (6.3)      Diabetics    155 (5.0)  1692 (8.8)      Tubulointerstitial    435 (14.0)  2032 (10.6)      Other or unknown    691 (22.2)  4933 (25.7)     Waiting time on transplantation list (months), median (IQR)  22 288  8.0 (3.1–17.9)  13.8 (5.4–29.4)  <0.0001   Diabetes history, n (%)  9918  171 (10.2)  1493 (18.1)  <0.0001   Cardiovascular comorbidities history, n (%)  8768  230 (15.2)  1864 (25.7)  <0.0001   Hypertension, n (%)  8854  1029 (68.6)  4988 (67.8)  0.56   Panel reactive antibodies, mean (SD)  3750  28.0 (34.3)  31.8 (35.6)  0.02   Blood group, n (%)  22 288      <0.0001    O    1079 (34.7)  8071 (42.1)      A    1619 (52.0)  8251 (43.0)      B    270 (8.7)  2029 (10.6)      AB    144 (4.6)  825 (4.3)     Dialysis modality before transplantb, n (%)  17 929          Peritoneal dialysis      2219 (12.4)      Haemodialysis      15 710 (87.6)     Pretransplant dialysis durationb  18 916          Median in years (IQR)      2.3 (1.3–4.1)      0–6 months, n (%)      918 (4.2)      6–12 months, n (%)      2223 (11.7)      12–24 months; n (%)      4932 (26.1)      24–36 months, n (%)      3979 (21.0)      36–60 months, n (%)      3558 (18.8)      > 60 months, n (%)      3306 (17.5)     BMI (kg/m2), mean (SD)  2725  24.7 (4.1)  25.2 (4.6)  0.04  Donors characteristics           Age (years), mean (SD)  21 845  49.5 (15.7)  50.8 (15.8)  <0.0001   Living donor, n (%)  22 288  690 (22.2)  1341 (7.0)  <0.0001   Expanded criteria donor, n (%)  21 611  1264 (43.6)  9287 (49.5)  <0.0001  KT characteristics         Number of HLA mismatches, n (%)  22 262      <0.0001    0    121 (3.9)  492 (2.5)      1–2    645 (20.8)  3619 (18.9)      3–4    1825 (58.8)  11 893 (62.1)      5–6    514 (16.5)  3153 (16.5)     Cold ischaemia time (h), mean (SD)  21 549  14.9 (8.9)  17.3 (8.1)  <0.0001  a t-test or chi-square test. b Among patients who received a non-PKT. BMI, body mass index. Table 1 Patients characteristics at time of transplantation in each group (PKT and non-PKT), France, 2002–13 Variables  Available data  PKT  Non-PKT  P-valuea  n = 3112 (14.0%)  n = 19 176 (86.0%)  Recipients characteristics           Male, n (%)  22 288  1835 (59.0)  11 961 (62.4)  0.0003   Age at transplantation (years), mean (SD)  22 288  48.8 (13.8)  50.8 (13.3)  <0.0001   Primary kidney disease, n (%)  21 628      <0.0001    Glomerulonephritis    928 (29.8)  6143 (32.0)      Cystic kidney disease    790 (25.4)  3174 (16.6)      Vascular    113 (3.6)  1202 (6.3)      Diabetics    155 (5.0)  1692 (8.8)      Tubulointerstitial    435 (14.0)  2032 (10.6)      Other or unknown    691 (22.2)  4933 (25.7)     Waiting time on transplantation list (months), median (IQR)  22 288  8.0 (3.1–17.9)  13.8 (5.4–29.4)  <0.0001   Diabetes history, n (%)  9918  171 (10.2)  1493 (18.1)  <0.0001   Cardiovascular comorbidities history, n (%)  8768  230 (15.2)  1864 (25.7)  <0.0001   Hypertension, n (%)  8854  1029 (68.6)  4988 (67.8)  0.56   Panel reactive antibodies, mean (SD)  3750  28.0 (34.3)  31.8 (35.6)  0.02   Blood group, n (%)  22 288      <0.0001    O    1079 (34.7)  8071 (42.1)      A    1619 (52.0)  8251 (43.0)      B    270 (8.7)  2029 (10.6)      AB    144 (4.6)  825 (4.3)     Dialysis modality before transplantb, n (%)  17 929          Peritoneal dialysis      2219 (12.4)      Haemodialysis      15 710 (87.6)     Pretransplant dialysis durationb  18 916          Median in years (IQR)      2.3 (1.3–4.1)      0–6 months, n (%)      918 (4.2)      6–12 months, n (%)      2223 (11.7)      12–24 months; n (%)      4932 (26.1)      24–36 months, n (%)      3979 (21.0)      36–60 months, n (%)      3558 (18.8)      > 60 months, n (%)      3306 (17.5)     BMI (kg/m2), mean (SD)  2725  24.7 (4.1)  25.2 (4.6)  0.04  Donors characteristics           Age (years), mean (SD)  21 845  49.5 (15.7)  50.8 (15.8)  <0.0001   Living donor, n (%)  22 288  690 (22.2)  1341 (7.0)  <0.0001   Expanded criteria donor, n (%)  21 611  1264 (43.6)  9287 (49.5)  <0.0001  KT characteristics         Number of HLA mismatches, n (%)  22 262      <0.0001    0    121 (3.9)  492 (2.5)      1–2    645 (20.8)  3619 (18.9)      3–4    1825 (58.8)  11 893 (62.1)      5–6    514 (16.5)  3153 (16.5)     Cold ischaemia time (h), mean (SD)  21 549  14.9 (8.9)  17.3 (8.1)  <0.0001  Variables  Available data  PKT  Non-PKT  P-valuea  n = 3112 (14.0%)  n = 19 176 (86.0%)  Recipients characteristics           Male, n (%)  22 288  1835 (59.0)  11 961 (62.4)  0.0003   Age at transplantation (years), mean (SD)  22 288  48.8 (13.8)  50.8 (13.3)  <0.0001   Primary kidney disease, n (%)  21 628      <0.0001    Glomerulonephritis    928 (29.8)  6143 (32.0)      Cystic kidney disease    790 (25.4)  3174 (16.6)      Vascular    113 (3.6)  1202 (6.3)      Diabetics    155 (5.0)  1692 (8.8)      Tubulointerstitial    435 (14.0)  2032 (10.6)      Other or unknown    691 (22.2)  4933 (25.7)     Waiting time on transplantation list (months), median (IQR)  22 288  8.0 (3.1–17.9)  13.8 (5.4–29.4)  <0.0001   Diabetes history, n (%)  9918  171 (10.2)  1493 (18.1)  <0.0001   Cardiovascular comorbidities history, n (%)  8768  230 (15.2)  1864 (25.7)  <0.0001   Hypertension, n (%)  8854  1029 (68.6)  4988 (67.8)  0.56   Panel reactive antibodies, mean (SD)  3750  28.0 (34.3)  31.8 (35.6)  0.02   Blood group, n (%)  22 288      <0.0001    O    1079 (34.7)  8071 (42.1)      A    1619 (52.0)  8251 (43.0)      B    270 (8.7)  2029 (10.6)      AB    144 (4.6)  825 (4.3)     Dialysis modality before transplantb, n (%)  17 929          Peritoneal dialysis      2219 (12.4)      Haemodialysis      15 710 (87.6)     Pretransplant dialysis durationb  18 916          Median in years (IQR)      2.3 (1.3–4.1)      0–6 months, n (%)      918 (4.2)      6–12 months, n (%)      2223 (11.7)      12–24 months; n (%)      4932 (26.1)      24–36 months, n (%)      3979 (21.0)      36–60 months, n (%)      3558 (18.8)      > 60 months, n (%)      3306 (17.5)     BMI (kg/m2), mean (SD)  2725  24.7 (4.1)  25.2 (4.6)  0.04  Donors characteristics           Age (years), mean (SD)  21 845  49.5 (15.7)  50.8 (15.8)  <0.0001   Living donor, n (%)  22 288  690 (22.2)  1341 (7.0)  <0.0001   Expanded criteria donor, n (%)  21 611  1264 (43.6)  9287 (49.5)  <0.0001  KT characteristics         Number of HLA mismatches, n (%)  22 262      <0.0001    0    121 (3.9)  492 (2.5)      1–2    645 (20.8)  3619 (18.9)      3–4    1825 (58.8)  11 893 (62.1)      5–6    514 (16.5)  3153 (16.5)     Cold ischaemia time (h), mean (SD)  21 549  14.9 (8.9)  17.3 (8.1)  <0.0001  a t-test or chi-square test. b Among patients who received a non-PKT. BMI, body mass index. Risk of graft failure in the whole cohort Graft failure occurred in 4952 patients up to 31 December 2013. The probability of surviving with a functioning graft for ≥10 years after KT was 79.9% in PKT patients and 73.8, 70.4, 62.4, 62.8, 59.5 and 56.0% in patients with a pretransplant duration of dialysis of ≤6, >6–12, >12–24, >24–36, >36–60 and >60 months, respectively (P < 0.0001) (Figure 1). The probability of surviving with a functioning graft for ≥10 years after KT was 61.4% in non-PKT patients (Figure 2A). After adjustment for recipient sex and age at KT, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), number of HLA mismatches, cold ischaemia time, centre and year of KT, pretransplant dialysis was associated with an increased hazard of graft failure when compared with PKT, whatever the duration of dialysis (Table 2). Compared with PKT patients, recipients with a pretransplant dialysis duration < 6 months had even a 27% increased hazard of graft failure at any time after KT [HR 1.27; 95% confidence interval (CI) 1.03–1.55]. The HR gradually increased with dialysis duration. Table 2 Effects of PKT and pretransplant dialysis duration on the hazard of graft failure (death, return to dialysis or retransplant, whichever came first), results of Cox proportional hazard model, France, 2002–13 Variables  na  Number of events  HRb  95% CI  P-value  Pretransplant dialysis duration  20 717  4580      <0.0001   0 (PKT)  2770  338  1       0–6 months  818  127  1.27  1.03–1.55     6–12 months  2061  394  1.38  1.19–1.59     12–24 months  4659  1101  1.67  1.48–1.90     24–36 months  3420  794  1.64  1.44–1.86     36–60 months  3823  955  1.82  1.61–2.07     > 60 months  3166  871  2.04  1.79–2.32    Type of transplantation             At any time after transplantation  20 953  4734          Non-PKT  18 183  4396  1        PKT  2770  338  0.57  0.51–0.63  <0.0001   In the first year post-transplant  20 953  1484          Non-PKT  18 183  1358  1        PKT  2770  126  0.69  0.57–0.83  < 0.0001   After 1-year post-transplant  20 953  3250          Non-PKT  18 183  3038  1        PKT  2770  212  0.51  0.45–0.59  0.013   In living donors  1203  116          Non-PKT  791  100  1        PKT  412  16  0.32  0.19–0.55  <0.0001   In deceased donors  19 750  4618          Non-PKT  17 392  4296  1        PKT  2358  322  0.59  0.53–0.64  <0.0001   In patients registered on the waiting list before RRT  4162  512          Non-PKT  1392  174  1        PKT  2770  338  0.70  0.57–0.86  0.0007  Variables  na  Number of events  HRb  95% CI  P-value  Pretransplant dialysis duration  20 717  4580      <0.0001   0 (PKT)  2770  338  1       0–6 months  818  127  1.27  1.03–1.55     6–12 months  2061  394  1.38  1.19–1.59     12–24 months  4659  1101  1.67  1.48–1.90     24–36 months  3420  794  1.64  1.44–1.86     36–60 months  3823  955  1.82  1.61–2.07     > 60 months  3166  871  2.04  1.79–2.32    Type of transplantation             At any time after transplantation  20 953  4734          Non-PKT  18 183  4396  1        PKT  2770  338  0.57  0.51–0.63  <0.0001   In the first year post-transplant  20 953  1484          Non-PKT  18 183  1358  1        PKT  2770  126  0.69  0.57–0.83  < 0.0001   After 1-year post-transplant  20 953  3250          Non-PKT  18 183  3038  1        PKT  2770  212  0.51  0.45–0.59  0.013   In living donors  1203  116          Non-PKT  791  100  1        PKT  412  16  0.32  0.19–0.55  <0.0001   In deceased donors  19 750  4618          Non-PKT  17 392  4296  1        PKT  2358  322  0.59  0.53–0.64  <0.0001   In patients registered on the waiting list before RRT  4162  512          Non-PKT  1392  174  1        PKT  2770  338  0.70  0.57–0.86  0.0007  a Subjects with complete data. b Adjusted for age of recipients at time of transplantation, sex of recipients, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of transplantation. Table 2 Effects of PKT and pretransplant dialysis duration on the hazard of graft failure (death, return to dialysis or retransplant, whichever came first), results of Cox proportional hazard model, France, 2002–13 Variables  na  Number of events  HRb  95% CI  P-value  Pretransplant dialysis duration  20 717  4580      <0.0001   0 (PKT)  2770  338  1       0–6 months  818  127  1.27  1.03–1.55     6–12 months  2061  394  1.38  1.19–1.59     12–24 months  4659  1101  1.67  1.48–1.90     24–36 months  3420  794  1.64  1.44–1.86     36–60 months  3823  955  1.82  1.61–2.07     > 60 months  3166  871  2.04  1.79–2.32    Type of transplantation             At any time after transplantation  20 953  4734          Non-PKT  18 183  4396  1        PKT  2770  338  0.57  0.51–0.63  <0.0001   In the first year post-transplant  20 953  1484          Non-PKT  18 183  1358  1        PKT  2770  126  0.69  0.57–0.83  < 0.0001   After 1-year post-transplant  20 953  3250          Non-PKT  18 183  3038  1        PKT  2770  212  0.51  0.45–0.59  0.013   In living donors  1203  116          Non-PKT  791  100  1        PKT  412  16  0.32  0.19–0.55  <0.0001   In deceased donors  19 750  4618          Non-PKT  17 392  4296  1        PKT  2358  322  0.59  0.53–0.64  <0.0001   In patients registered on the waiting list before RRT  4162  512          Non-PKT  1392  174  1        PKT  2770  338  0.70  0.57–0.86  0.0007  Variables  na  Number of events  HRb  95% CI  P-value  Pretransplant dialysis duration  20 717  4580      <0.0001   0 (PKT)  2770  338  1       0–6 months  818  127  1.27  1.03–1.55     6–12 months  2061  394  1.38  1.19–1.59     12–24 months  4659  1101  1.67  1.48–1.90     24–36 months  3420  794  1.64  1.44–1.86     36–60 months  3823  955  1.82  1.61–2.07     > 60 months  3166  871  2.04  1.79–2.32    Type of transplantation             At any time after transplantation  20 953  4734          Non-PKT  18 183  4396  1        PKT  2770  338  0.57  0.51–0.63  <0.0001   In the first year post-transplant  20 953  1484          Non-PKT  18 183  1358  1        PKT  2770  126  0.69  0.57–0.83  < 0.0001   After 1-year post-transplant  20 953  3250          Non-PKT  18 183  3038  1        PKT  2770  212  0.51  0.45–0.59  0.013   In living donors  1203  116          Non-PKT  791  100  1        PKT  412  16  0.32  0.19–0.55  <0.0001   In deceased donors  19 750  4618          Non-PKT  17 392  4296  1        PKT  2358  322  0.59  0.53–0.64  <0.0001   In patients registered on the waiting list before RRT  4162  512          Non-PKT  1392  174  1        PKT  2770  338  0.70  0.57–0.86  0.0007  a Subjects with complete data. b Adjusted for age of recipients at time of transplantation, sex of recipients, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of transplantation. FIGURE 1 View largeDownload slide Probability of surviving without retransplant or dialysis after transplantation (with 95% CI) according to the pretransplant dialysis duration (Kaplan–Meier estimator). FIGURE 1 View largeDownload slide Probability of surviving without retransplant or dialysis after transplantation (with 95% CI) according to the pretransplant dialysis duration (Kaplan–Meier estimator). FIGURE 2 View largeDownload slide Probability of surviving without retransplant or dialysis after transplantation (with 95% CI) in each group of transplantation (Kaplan–Meier estimator). (A) Analysis in the whole cohort. (B) Analysis in the subgroup of patients registered pre-emptively on the waiting list. FIGURE 2 View largeDownload slide Probability of surviving without retransplant or dialysis after transplantation (with 95% CI) in each group of transplantation (Kaplan–Meier estimator). (A) Analysis in the whole cohort. (B) Analysis in the subgroup of patients registered pre-emptively on the waiting list. Overall, PKT was associated with a 43% reduction in the hazard of graft failure at any time after KT when compared with patients treated by dialysis prior to KT (HR 0.57; 95% CI 0.51–0.63) (Table 2). This reduction in the hazard of graft failure was greater after the first year post-KT (Table 2). Specifically, the hazard of graft failure was reduced by 31% during the first year post-KT in patients with PKT compared with those with non-PKT (HR 0.69; 95% CI 0.57–0.83) and by 49% beyond 1 year (HR 0.51; 95% CI 0.45–0.59). Among patients who received a first KT between 2007 and 2012, further adjustment for cardiovascular comorbidities and diabetes did not change the association of PKT with the hazard of graft failure (data not shown). The reduction of hazard of graft failure in PKT patients was stronger in living than in deceased donor recipients. In living donor recipients, PKT was associated with a 68% reduction in the hazard of graft failure when compared with non-PKT (HR 0.32; 95% CI 0.19–0.55), whereas it was associated with a 41% reduction (HR 0.59; 95% CI 0.53–0.64) in deceased donor recipients (Table 2). Risk of graft failure among patients registered on the waiting list before RRT In the subgroup of patients pre-emptively registered, the association of PKT with graft failure was weaker than in the whole study population but remained significant (Figure 2B). Specifically, the probability of surviving with a functioning graft for ≥10 years after KT was 79.9% in PKT patients and 75.0% in non-PKT patients (P < 0.001) (Figure 2B). After adjustment for the same covariables as for the whole cohort, PKT was associated with a 30% reduction in the hazard of graft failure at any time after KT, when compared with non-PKT (HR 0.70; 95% CI 0.57–0.86) (Table 2). This result was observed while the median pretransplant dialysis duration was lower in patients with non-PKT registered on the waiting list before RRT than in the whole population of non-PKT patients (1.1 years in Supplementary data, Table S1 versus 2.3 years in Table 1). Risk of dialysis/retransplantation and death with a functioning graft The probabilities to return to dialysis or to be retransplanted, and to die with a functioning graft within 10 years after KT were, respectively, 9.5% and 10.6% in patients with PKT and, 21.2% and 17.4% in patients with non-PKT (P < 0.0001) (Supplementary data, Figure S1). Compared with non-PKT, PKT was associated with a 45% reduction in the hazard of dialysis/retransplantation (HR 0.55; 95% CI 0.47–0.64) and with a 40% reduction in the hazard of death with a functioning graft (HR 0.60; 95% CI 0.50–0.71) (Table 3), after adjustment for the same factors as previously. These results were confirmed in the subgroup of patients registered on the waiting list before dialysis for the event dialysis/retransplantation (HR 0.59; 95% CI 0.45–0.78) but not for death with a functioning graft (HR 0.87; 95% CI 0.62–1.23) (Table 3). Table 3 Effect of pre-emptive transplantation on the hazard of retransplant or return to dialysis, and death with a functioning graft, results of cause-specific proportional hazard model, France, 2002–13 Variables  Retransplant or return to dialysis, whichever came first (censored at death with a functioning graft)   Death with a functioning graft (censored at retransplant or return to dialysis, whichever came first)   HRa  95% CI  P-value  HRa  95% CI  P-value  Type of transplantation   At any time after transplantation (n = 20 953)    Non-PKT  1      1        PKT  0.55  0.47–0.64  <0.0001  0.60  0.50–0.71  <0.0001   In patients registered on the waiting list before RRT (n = 4162)    Non-PKT  1      1        PKT  0.59  0.45–0.78  0.0002  0.87  0.62–1.23  0.44  Variables  Retransplant or return to dialysis, whichever came first (censored at death with a functioning graft)   Death with a functioning graft (censored at retransplant or return to dialysis, whichever came first)   HRa  95% CI  P-value  HRa  95% CI  P-value  Type of transplantation   At any time after transplantation (n = 20 953)    Non-PKT  1      1        PKT  0.55  0.47–0.64  <0.0001  0.60  0.50–0.71  <0.0001   In patients registered on the waiting list before RRT (n = 4162)    Non-PKT  1      1        PKT  0.59  0.45–0.78  0.0002  0.87  0.62–1.23  0.44  a Adjusted for age of recipients at time of transplantation, sex of recipients, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of transplantation. Table 3 Effect of pre-emptive transplantation on the hazard of retransplant or return to dialysis, and death with a functioning graft, results of cause-specific proportional hazard model, France, 2002–13 Variables  Retransplant or return to dialysis, whichever came first (censored at death with a functioning graft)   Death with a functioning graft (censored at retransplant or return to dialysis, whichever came first)   HRa  95% CI  P-value  HRa  95% CI  P-value  Type of transplantation   At any time after transplantation (n = 20 953)    Non-PKT  1      1        PKT  0.55  0.47–0.64  <0.0001  0.60  0.50–0.71  <0.0001   In patients registered on the waiting list before RRT (n = 4162)    Non-PKT  1      1        PKT  0.59  0.45–0.78  0.0002  0.87  0.62–1.23  0.44  Variables  Retransplant or return to dialysis, whichever came first (censored at death with a functioning graft)   Death with a functioning graft (censored at retransplant or return to dialysis, whichever came first)   HRa  95% CI  P-value  HRa  95% CI  P-value  Type of transplantation   At any time after transplantation (n = 20 953)    Non-PKT  1      1        PKT  0.55  0.47–0.64  <0.0001  0.60  0.50–0.71  <0.0001   In patients registered on the waiting list before RRT (n = 4162)    Non-PKT  1      1        PKT  0.59  0.45–0.78  0.0002  0.87  0.62–1.23  0.44  a Adjusted for age of recipients at time of transplantation, sex of recipients, primary kidney disease, donor type (living or deceased donor, expanded criteria donor or standard), HLA mismatches, cold ischaemia time, centre and year of transplantation. DISCUSSION Using data from the French organ transplant registry, we found that pretransplant dialysis duration was associated with an increased hazard of graft failure at any time after transplantation when compared with patients receiving a PKT, whatever dialysis duration. PKT was associated with a decreased risk of graft failure whether or not death with a functioning graft was included in the definition of graft failure. We also found a substantial reduction of the hazard of graft failure in PKT patients even after restricting the non-PKT group to more comparable patients, that is, those registered on the waiting list before dialysis initiation. A further reduction of the risk of graft failure was observed when kidney transplant was from a living donor. These results are consistent with American studies [6, 7], despite differences in access to transplantation and a lower mortality in dialysis in France. One of the major differences between previous studies and our work is that our results suggest that even a short period on dialysis (≤6 months) is associated with poorer outcome after KT. Indeed, American, Japanese and Finnish cohort studies found an increased risk of graft failure after 6 months of dialysis or did not explore the effect of a short period of dialysis [9, 11, 16]. Avoiding dialysis as far as possible should thus always be considered, especially for patients waiting for a living donor graft. This message is important, because it is counter-intuitive for many physicians. Our results differ more from previous European reports, which mostly did not find significant graft failure risk differences between PKT and non-PKT [12, 13]. This may partly be explained by the fact that our study is based on a larger sample size, a longer inclusion period and an exhaustive national transplant registry. As all previous studies that compared the risk of graft failure in PKT and non-PKT patients, our study may suffer from important residual confounding due to unmeasured confounders. First, PKT recipients may have a higher level of residual function of their native kidney at transplantation. This could also explain why we observed a stronger reduction in the hazard of graft failure after the first year post-transplantation than within the first year after KT. As most previous studies, we could not adjust for residual renal function because these data were not available in patients who started dialysis before transplantation. However, two studies found that a higher level of residual renal function was not associated with a better graft survival in patients who received a PKT [27, 28], suggesting that residual renal function might not be a major factor. A second reason that could explain the observed lower risk of graft failure in PKT recipients is that these patients might be less likely to have comorbidities, as in our study, than those transplanted after initiation of dialysis. However, after adjustment for cardiovascular comorbidities and diabetes, our results were similar. The reduction of hazard in PKT also persisted, although to a lesser extent, after restriction of the non-PKT patients to those registered on the waiting list before KT. Other potential factors may explain the observed lower risk in PKT recipients, such as socio-economic status, nutritional status, sensitization before transplant and adherence to immunosuppressive treatment. For example, social deprivation was associated with a decreased access to waiting list even in countries with universal health coverage, including France, and with increased risk of death after KT [29–32]. The lack of information about sensitization of the waitlisted patients is an important weakness, because preformed for donor-specific HLA antibodies (HLA-DSAs) are a considerable risk factor for graft survival. Furthermore, patients without preformed HLA-DSAs are transplanted faster and it is possible that PKT patients in our study were less likely to be sensitized than non-PKT patients. However, adjustment for all these factors could not be performed because they were not, or only recently, collected in the registry. Furthermore, it could have been interesting to adjust our analyses for time from the first nephrology consultation to transplant but these data are not recorded in the French RRT registry. To reduce residual confounding due to unmeasured confounders, we performed an analysis where we restricted the non-PKT group to patients registered on the transplant waiting list prior to the initiation of dialysis. Indeed, patients who have been registered on the waiting list before dialysis initiation are more likely to be comparable with PKT patients, as we observed in our data with respect to history of diabetes and cardiovascular diseases at baseline. We may assume that if this analysis reduced unbalances between PKT and non-PKT patients in terms of these comorbidities, then it might have also reduced unbalances on other unmeasured confounders. As expected, we showed a lower reduction of the hazard of graft failure in PKT patients compared with non-PKT patients. However, a reduction of 30% of the hazard at each time after transplantation is still substantial, even if we acknowledge that this analysis could not solve all the residual confounding issue. We found that the risk of graft failure, and not the risk of death with a functioning graft, was statistically significantly reduced with PKT in pre-emptively registered patients, suggesting that PKT has a stronger impact on graft survival than on patient survival. This also suggests that patients registered pre-emptively on the waiting list are selected among patients with a better health status. It should also be noted that we included only patients who were transplanted and did not intend to assess the effect of PKT in the population of all patients registered on the waiting list before RRT. Indeed, such a study would have to account for the fact that some patients of the target population may die prior to KT, and thus will never be at risk of graft failure. In conclusion, although our study suffers from potential residual confounding due to unmeasured confounders as previous studies, we believe that our findings tend to confirm that KT performed after the initiation of dialysis may be associated with an increased risk of graft failure at any time after transplantation, compared with PKT, even for a dialysis duration of less than 6 months. Overall, our results suggest that nephrologists should more frequently consider the registration of their patients on the waiting list before dialysis initiation. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS All authors thank all the 33 French kidney transplant centres that participated in the study. FUNDING M.P.-R. reports grants from la Fondation pour la Recherche Médicale (FRM) during the conduct of the study. The funding source did not have any impact on the study design or the decision to submit the paper. AUTHORS’ CONTRIBUTIONS M.P.-R., J.H. and K.L. contributed to literature search, figures, study design, data analysis, data interpretation and writing of the paper. C.C., P.M. and L.C. contributed to study design, data interpretation and writing of the paper. C.J. contributed to study design, data collection and writing of the paper. REFERENCES 1 Wolfe RA, Ashby VB, Milford EL et al.   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  2 Schnuelle P, Lorenz D, Trede M et al.   Impact of renal cadaveric transplantation on survival in end-stage renal failure: evidence for reduced mortality risk compared with hemodialysis during long-term follow-up. J Am Soc Nephrol  1998; 9: 2135– 2141 Google Scholar PubMed  3 Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 Clinical Practice Guideline for the Evaluation and the Management of Chronic Kidney Disease. Kidney Int Suppl  2013; 3: 1– 150 CrossRef Search ADS   4 Abramowicz D, Hazzan M, Maggiore U et al.   Does pre-emptive transplantation versus post start of dialysis transplantation with a kidney from a living donor improve outcomes after transplantation? A systematic literature review and position statement by the Descartes Working Group and ERBP. Nephrol Dial Transplant  2016; 31: 691– 697 Google Scholar CrossRef Search ADS PubMed  5 Registre Epidémiologie et Information en Néphrologie. Rapport REIN 2014. https://www.agence-biomedecine.fr/IMG/pdf/rapportrein2014.pdf (25 August 2017, date last accessed) 6 Kasiske BL, Snyder JJ, Matas AJ et al.   Preemptive kidney transplantation: the advantage and the advantaged. J Am Soc Nephrol  2002; 13: 1358– 1364 Google Scholar CrossRef Search ADS PubMed  7 Mange KC, Joffe MM, Feldman HI. Effect of the use or non-use of long-term dialysis on the subsequent survival of renal transplants from living donors. N Engl J Med  2001; 344: 726– 731 Google Scholar CrossRef Search ADS PubMed  8 Meier-Kriesche HU, Port FK, Ojo AO et al.   Effect of waiting time on renal transplant outcome. Kidney Int  2000; 58: 1311– 1317 Google Scholar CrossRef Search ADS PubMed  9 Goldfarb-Rumyantzev A, Hurdle JF, Scandling J et al.   Duration of end-stage renal disease and kidney transplant outcome. Nephrol Dial Transplant  2005; 20: 167– 175 Google Scholar CrossRef Search ADS PubMed  10 Jay C, Dean P, Helmick R et al.   Reassessing preemptive kidney transplantation in the United States: are we making progress? Transplantation  2016; 100: 1120– 1127 Google Scholar CrossRef Search ADS PubMed  11 Goto N, Okada M, Yamamoto T et al.   Association of dialysis duration with outcomes after transplantation in a Japanese cohort. Clin J Am Soc Nephrol  2016; 11: 497– 504 Google Scholar CrossRef Search ADS PubMed  12 Berthoux FC, Jones EH, Mehls O et al.   Transplantation report. 2: preemptive renal transplantation in adults aged over 15 years. The EDTA-ERA Registry. European Dialysis and Transplant Association-European Renal Association. Nephrol Dial Transplant  1996; 11 (Suppl 1): 41– 43 Google Scholar CrossRef Search ADS PubMed  13 Kessler M, Ladriere M, Giral M et al.   Does pre-emptive kidney transplantation with a deceased donor improve outcomes? Results from a French transplant network: outcome of pre-emptive kidney transplantation. Transpl Int  2011; 24: 266– 275 Google Scholar CrossRef Search ADS PubMed  14 Pérez-Flores I, Sánchez-Fructuoso A, Calvo N et al.   Preemptive kidney transplant from deceased donors: an advantage in relation to reduced waiting list. Transplant Proc  2007; 39: 2123– 2124 Google Scholar CrossRef Search ADS PubMed  15 Witczak BJ, Leivestad T, Line PD et al.   Experience from an active preemptive kidney transplantation program–809 cases revisited. Transplantation  2009; 88: 672– 677 Google Scholar CrossRef Search ADS PubMed  16 Helanterä I, Salmela K, Kyllönen L et al.   Pretransplant dialysis duration and risk of death after kidney transplantation in the current era. Transplantation  2014; 98: 458– 464 Google Scholar CrossRef Search ADS PubMed  17 Haller MC, Kainz A, Baer H et al.   Dialysis vintage and outcomes after kidney transplantation: a retrospective cohort study. Clin J Am Soc Nephrol  2017; 12: 122– 130 Google Scholar CrossRef Search ADS PubMed  18 United States Renal Data System. 2015 USRDS annual data report: epidemiology of kidney disease in the United States. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 2015 19 Registre Epidémiologie et Information en Néphrologie. Rapport REIN 2013. https://www.agence-biomedecine.fr/IMG/pdf/rapport_rein2013.pdf (25 August 2017, date last accessed) 20 Grams ME, Chen BP-H, Coresh J et al.   Preemptive deceased donor kidney transplantation: considerations of equity and utility. Clin J Am Soc Nephrol  2013; 8: 575– 582 Google Scholar CrossRef Search ADS PubMed  21 Andersen PK, Geskus RB, de Witte T et al.   Competing risks in epidemiology: possibilities and pitfalls. Int J Epidemiol  2012; 41: 861– 870 Google Scholar CrossRef Search ADS PubMed  22 Gray RJA. Class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat  1988; 16: 1141– 1154 Google Scholar CrossRef Search ADS   23 Suttorp MM, Siegerink B, Jager KJ et al.   Graphical presentation of confounding in directed acyclic graphs. Nephrol Dial Transplant  2015; 30: 1418– 1423 Google Scholar CrossRef Search ADS PubMed  24 Noordzij M, Leffondré K, van Stralen KJ et al.   When do we need competing risks methods for survival analysis in nephrology? Nephrol Dial Transplant  2013; 28: 2670– 2677 Google Scholar CrossRef Search ADS PubMed  25 Eilers PHC, Marx BD. Flexible smoothing with B-splines and penalties. Statist Sci  1996; 11: 89– 121 Google Scholar CrossRef Search ADS   26 Leffondré K, Jager KJ, Boucquemont J et al.   Representation of exposures in regression analysis and interpretation of regression coefficients: basic concepts and pitfalls. Nephrol Dial Transplant  2014; 29: 1806– 1814 Google Scholar CrossRef Search ADS PubMed  27 Akkina SK, Connaire JJ, Snyder JJ et al.   Earlier is not necessarily better in preemptive kidney transplantation. Am J Transplant  2008; 8: 2071– 2076 Google Scholar CrossRef Search ADS PubMed  28 Grams ME, Massie AB, Coresh J et al.   Trends in the timing of pre-emptive kidney transplantation. J Am Soc Nephrol  2011; 22: 1615– 1620 Google Scholar CrossRef Search ADS PubMed  29 Riffaut N, Lobbedez T, Hazzan M et al.   Access to preemptive registration on the waiting list for renal transplantation: a hierarchical modeling approach. Transpl Int  2015; 28: 1066– 1073 Google Scholar CrossRef Search ADS PubMed  30 Kasiske BL, London W, Ellison MD. Race and socioeconomic factors influencing early placement on the kidney transplant waiting list. J Am Soc Nephrol  1998; 9: 2142– 2147 Google Scholar PubMed  31 Udayaraj U, Ben-Shlomo Y, Roderick P et al.   Social deprivation, ethnicity, and access to the deceased donor kidney transplant waiting list in England and Wales. Transplantation  2010; 90: 279– 285 Google Scholar CrossRef Search ADS PubMed  32 Begaj I, Khosla S, Ray D et al.   Socioeconomic deprivation is independently associated with mortality post kidney transplantation. Kidney Int  2013; 84: 803– 809 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: Mar 22, 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