Background. Kidney transplantation (Tx) is the treatment of choice for children on renal replacement therapy (RRT). However, graft loss remains an important problem and may affect the choice for living versus deceased donation and timing of transplantation. Therefore, the aim of this study is to identify subgroups of patients with markedly different graft survival times, and assess their association with the outcome, among a large sample of European paediatric transplant recipients.

Methods. 7839 paediatric Tx recipients included in the ESPN/ERA-EDTA Registry were analyzed. Factors considered were gender, age at start RRT, age at transplantation (Age-Tx), time on dialysis (Dial-Time), pre-emptive transplantation (Preemptive-Tx), and cause of renal failure classified according to the risk of disease recurrence (Recurrence-Risk). We used binary survival tree analysis to identify subgroups, defined by patterns of pre-transplant factors, with similar 1- and 5-years graft survival. We then fitted Cox models including dummy variables denoting membership to these groups to assess the association between pre-transplant factors, subgroups, and graft outcome.

Results. For 1-year graft survival, 4 subgroups of patients were identified defined by Age-Tx, Recurrence-Risk and Dial-Time (Fig.1). The Cox analysis revealed recipients presenting (i) Preemptive-Tx (HR=0.58, 95%CI 0.4-0.7), (ii) Age-TX<3.2years (2.45, 1.8-3.3) and (iii) older than 3 with High Recurrence-Risk (1.70, 1.4-2.1) to be significantly associated with 1 year graft failure.

For the 5 years graft survival, 6 subgroups were identified using Recurrence-Risk, Dial-Time and Age-Tx (Fig.2). The survival analysis showed subgroups with (i)High Recurrence-Risk (2.55, 2.1-3.1); (ii) Dial-Time<3.4yrs and young age (<4.1yrs) (2.22, 1.7-2.9); (iii)short-term dialysis (0-3.4yrs) and older age (>4.1yrs) (1.36, 1.1-1.6); (iv)long-term dialysis (>3.4yrs) and adolescent age (Age-Tx>12.8yrs) (2.91, 2.1-3.9) were all independently associated with 5 years graft prognosis.

Conclusions. Non pre-emptive transplantation and diseases with a high risk of disease-recurrence as well as long-term dialysis in combination with adolescent age were independently associated with worse renal graft outcome. The identification of these subgroups reveals interactions between factors, and carry added prognostic value for predicting renal graft prognosis. Additional analyses will include more graft-related factors like donor type and age.