We performed a prospective study to establish the optimal background correction algorithm for the determination of differential renal function (DRF), using 99Tc(m)-diethylenetriamine pentaacetate (99Tc(m)-DTPA) in the presence of unilateral hydronephrosis, with 24 h 99Tc(m)-dimercaptosuccinic acid (99Tc(m)-DMSA) uptake as the 'gold standard'. From September 1996 to June 1997, 12 males and 4 females (mean age 10 years, range 1 month to 72 years), presenting with unilateral hydronephrosis, were studied. All patients underwent both DTPA renography and quantitative DMSA scintigraphy within 24 h. In all patients, using a surface method, the DRF of the obstructed kidney was determined using infrarenal, suprarenal and perirenal background correction, time intervals of 60-180 s (t1), 120-180 s (t2) and 80-140 s (t3), and the application or non-application of a Rutland-Patlak correction (RPC). In the absence of RPC, for all three types of background correction, no difference in DTPA DRF for any of the three time intervals was noted; higher DTPA DRF values were found (mean +/- S.D.: overestimates of 7.8+/-24.4%, 6.5+/-9.5% and 3.3+/-14.9% for suprarenal, infrarenal and perirenal background correction, respectively). Application of RPC resulted in an overall decrease in both the mean and standard deviation values, which was most pronounced with infrarenal background correction: -0.38+/-6.5% for t1, 0.31+/-6.3% for t2 and -1.3 +/-6.9% for t3 (t1 vs t2, P = 0.06; t3 vs t1 or t2, P = 0.04). Our results suggest that infrarenal background correction using t1 or t2 and RPC is the best algorithm for DRF estimation using 99Tc(m)-DTPA renography.