Background. Although risk factors of graft failure after renal transplantation have been widely considered, it is difficult to distinguish their etiological role: a direct association with graft failure or a relationship through the renal function leading to graft failure. This precise description of risk factors is of primary importance after the first acute year post-transplantation in order to help physician in the appraisal of the long-term risk of graft failure and the corresponding personalization of care.

Methods. We used a joint survival longitudinal model to simultaneously consider the serum creatinine evolution and the risk of graft failure (return to dialysis or patient death). We studied 2773 adult recipients from the DIVAT cohort, between 2000 and 2013 alive with a functioning graft at one year post-transplantation and maintained under Tacrolimus and MMF.

Results. As expected, we demonstrated that male recipients (p<0.0001), past history of diabetes (p<0.0001), old donor (p<0.0001), long cold ischemia time (p=0.0267) or high level of the 3-month creatinine (p<0.0001) were mainly associated with an increase of serum creatinine before the observance of graft failure. Serum creatinine level appears as a major risk factor of graft failure (HR=2.14, for a growth of 25%, p<0.0001). More interestingly, we demonstrated that elderly patients (HR=1.50, p<0.0001), who received a second transplantation (HR=1.41, p=0.0118), with an acute rejection episode during the first year (HR=1.43, p=0.0017), or with a history of cardiovascular disease (HR=1.37, p=0.0016) were associated with higher risk of graft failure without observing a first significant worsening renal function.

Conclusion. The use of joint model in kidney transplantation allows enhancing the knowledge of etiological pathways explaining the evolution of the patient and her/his graft. We especially demonstrated that physicians should pay a particular attention to elderly recipients, with second transplant recipients, with acute rejection episode, or with history of cardiovascular disease. Indeed, for such a profile of patients, the monitoring of the renal function did not appear useful to evaluate the risk of graft failure.

« Back to 2015 American Transplant Congress