Determining early surrogate markers of long-term graft outcome is important for optimal medical management. In order to identify such markers, we used clinical information from a cross-validated French database (Données Informatisées et VAlidées en Transplantation) of 2169 kidney transplant recipients to construct a composite score 1 year after transplantation. This Kidney Transplant Failure Score took into account a series of eight accepted pre- and post-transplant risk factors of graft loss, and was subsequently evaluated for its ability to predict graft failure at 8 years. This algorithm outperformed the traditional surrogates of serum creatinine and the estimated graft filtration rate, with an area under the receiver–operator characteristic curve of 0.78. Validation on an independent database of 317 graft recipients had the same predictive capacity. Our algorithm was also able to stratify patients into two groups according to their risk: a high-risk group of 81 patients with 25% graft failure and a low-risk group of 236 patients with an 8% failure rate. Thus, although this clinical composite score predicts long-term graft survival, it needs validation in different patient groups throughout the world.