Non-ionic dimers induce less diuresis than non-ionic monomers, resulting in increased opacification of the urinary tract in intravenous (i.v.) urography. This double blind, comparative, randomized, parallel trial compared the efficacy of iodixanol (non-ionic dimer) and iopromide (non-ionic monomer) in 100 patients with mild to moderate renal insufficiency (serum creatinine of 135 to 265 mumol l-1) who underwent i.v. urography. A total dose of 600 mgI kg-1 bw of iodixanol (320 mgI ml-1) or iopromide (300 mgI ml-1) was injected. Radiographs were blindly evaluated by three radiologists who analysed different parameters (renal border visualization, nephrogram density, calyceal filling and density, papillary blush detection, delineation of collecting ducts, renal pelvis opacification, visualization of ureters, bladder density, bladder distention). Densitometric evaluation on the renal pelvis and bladder was also performed. Iodixanol showed better filling and density of the calyces (p = 0.004), more frequent detection of papillary blush (p = 0.003) and better opacification of the renal pelvis (p = 0.006). No significant differences between the two contrast media were found in regard to other parameters. In conclusion, the results confirmed theoretical expectations. The non-ionic dimer iodixanol is to be preferred to a non-ionic monomer such as iopromide in i.v. urography on patients with impaired renal function.