THE PREDICTIVE VALUE OF RENAL VASCULAR RESISTANCE FOR RENAL GRAFT LOSS

Abstract

P319 Aims: The renal artery resistance index (RI), assessed by Doppler sonography, was recently identified as the premier predictor of graft loss.1 Since RI in that study was not measured at a fixed time-point after transplantation, and since Doppler sonography is very operator-dependent, we investigated the predictive value of renal vascular resistance (RVR), a less operator-dependent method, at fixed time-points. Methods: In 793 consecutive first-time cadaveric transplant recipients, transplanted between 1970-2003, RVR was determined at 1 (and 5) years posttransplant by dividing the mean arterial pressure (average of 3 automated measurements) by renal blood flow (renal clearance of 131Hippuran divided by 1-hematocrit). The predictive value of RVR was compared to that of, among others, 24h urinary creatinine clearance (24h CrCl) and proteinuria. The primary endpoint of the study was 15-year death-censored graft loss. Results: Mean RVR at 1 year was 0.34 ± 0.17 mmHg/mL/min. Median follow-up time was 6.7 [3.6-10.5] yrs. A total of 101 patients (13%) reached the endpoint, whereas 157 patients (20%) were censored because of death. Recipients in the highest 3rd and 4th quartiles of RVR at 1 year posttransplant had significantly lower 15-y death-censored graft survival (figure 1). RVR, 24h CrCl, and proteinuria at 1 year posttransplant, being the best predictors, all predicted the endpoint similarly (figure 2). All risk factors were only moderate predictors of the endpoint, as assessed by the low AUC value of ROC curves (e.g. RVR: 0.64 [0.58-0.70]). Stratifying RVR ROC-curves for donor and recipient sex and age, and for 24h CrCl and proteinuria did not change our findings. Interestingly, RVR at 5 years predicted the endpoint better than at 1 year, albeit still modest (AUC 0.74 [0.66-0.83], p=0.05). Not censoring for death did not change our results essentially.FigureFigureConclusions: RVR is among the most prominent, albeit modest, predictors for 15-year graft loss. However, because RVR at 1 year posttransplant is not a better predictor than 24h CrCl or proteinuria, and because the determination of RVR is time-consuming and costly, RVR holds no clinical merit in our view for the follow-up of renal transplant outpatients. Besides the fact RVR may not equal RI, the difference between our results and previous findings may lie in the fact that we measured RVR at a fixed time-point close to transplantation. 1) N Engl J Med 2003; 349(2):115-124.