Use of 12-Month Renal Function and Baseline Clinical Factors to Predict Long-Term Graft Survival: Application to BENEFIT and BENEFIT-EXT Trials

Schnitzler, Mark A.1; Lentine, Krista L.1,5; Axelrod, David2; Gheorghian, Adrian1; You, Min3; Kalsekar, Anupama3; L'Italien, Gilbert3,4

doi: 10.1097/TP.0b013e31823ec02a
Clinical and Translational Research

Background. Innovation in renal transplant management would benefit from identification of early markers that accurately predict long-term graft survival.

Methods. Data from the United States Renal Data System for kidney transplant recipients (1995–2004) were analyzed to develop prediction models for all-cause graft survival based on estimated glomerular filtration rate (eGFR), the presence or absence of acute rejection within 1 year, and recipient and donor demographic characteristics. The prediction models were applied to participants in the Belatacept Evaluation of Nephroprotection and Efficacy as First-line Immunosuppression Trial and Belatacept Evaluation of Nephroprotection and Efficacy as First-line Immunosuppression Trial—EXTended criteria donors trials comparing belatacept with cyclosporine in standard criteria donor (SCD) and expanded criteria donor (ECD) graft recipients, respectively, as an external validation of the model predictions in a diverse population.

Results. Compared with eGFR 60 mL/min/1.73 m2, the relative hazard for all-cause graft loss increased in an accelerating pattern with lower GFR to approximately eight and seven times, respectively, among SCD and ECD recipients with eGFR less than 15 mL/min/1.73 m2. When applied to the clinical trial samples, the predicted differences in all-cause graft survival of less intensive belatacept versus cyclosporine at the second transplant anniversary (SCD: 3.9%, 95% confidence interval [CI]: 3.6% to 4.2%; ECD: 4.1%, 95% CI: 3.5% to 4.7%) were similar to observed differences (SCD: 4.2%, 97.3% CI: −1.3% to 10.1%; ECD: 1.4%, 97.3% CI: −7.5% to 10.2%).

Conclusions. Accurate models of long-term graft survival can be developed using eGFR, donor, and recipient characteristics. Long-term survival prediction models may provide an efficient method for assessing the impact of novel pharmaceutical agents and clinical management protocols.

Author Information

1Saint Louis University Center for Outcomes Research, St. Louis, MO.

2Dartmouth-Hitchcock Medical Center, Hanover, NH.

3Bristol-Myers Squibb Health Services, Plainsboro, NJ.

4Internal Medicine, Yale University School of Medicine, New Haven, CT.

This work was supported, in part, by a grant from Bristol-Myers Squibb.

The sponsor's support of the research does not cover publication, nor is there any restriction of the authors' publication rights by the sponsor.

The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the U.S. government.

Address correspondence to: Krista L. Lentine, M.D., M.S., Saint Louis University Center for Outcomes Research, 3545 Lafayette Avenue, Salus Center, 4th Floor, SLUCOR, St. Louis, MO 63104. E-mail:

M.A.S. participated in study design, data acquisition, data analysis, interpretation, and writing of the manuscript; G.L. and A.K. participated in study design, data acquisition, interpretation, and writing of the manuscript; K.L.L., D.A., A.G., participated in study design, interpretation, and writing of the manuscript; M.Y. participated in study design, data analysis, interpretation and writing of the manuscript; all authors agreed to publish the manuscript; and K.L.L. and M.A.S. wrote the first draft of the manuscript.

Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal's Web site (

Received 18 May 2011. Revision requested 2 June 2011.

Accepted 12 October 2011.

© 2012 Lippincott Williams & Wilkins, Inc.