MDRD-estimated GFR at one year post-renal transplant is a predictor of long-term graft function

Abstract

Renal transplantation is the optimal mode of renal replacement. Improvements in graft survival and acute rejection rates have made these outcomes less useful for prognostication and as end-points in clinical trials; accurate surrogate markers of long-term graft outcome are therefore increasingly important. This study examines the relationship between both serum creatinine (SCr(1 yr)) and MDRD estimated glomerular filtration rate measured at one year (eGFR(MDRD)(1 yr)) as predictors of graft survival. Data on 1,110 patients who received a renal transplant between 1989 and 2005 were extracted from the Irish Renal Transplant Registry. The study group was divided into quartiles of patient numbers according to SCr(1 yr) and eGFR(MDRD)(1 yr). Kaplan-Meier estimates of the primary end-point graft survival were constructed for each quartile. Additionally, a Cox Regression restricted cubic spline model was performed for both eGFR(MDRD)(1 yr) and SCr(1 yr). Both overall graft outcome and outcome censored for death with a functioning graft (CDWFG) were used as endpoints. Cox regression analysis was performed along with tests for the proportionality assumption to compare the predictive value of eGFR(MDRD)(1 yr)and SCr(1 yr). Both eGFR(MDRD)(1yr) and SCr(1 yr) were independently associated with long-term renal transplant survival. eGFR(MDRD)(1 yr) and SCr(1 yr) had similarly strong associations with long-term outcome when the quartile variables were compared using the Bayesian Information Criterion method. The Cox regression restricted cubic spline modeling demonstrated that an eGFR(MDRD)(1 yr) value < 27 mLs/min/1.73 m(2) and a SCr(1 yr) value > 229 micromol/L were associated with poor graft survival.