The purpose of this study was to review our institutional experience with colorflow duplex scanning in detecting significant renal artery stenosis and to validate the criteria used: renal artery peak systolic velocity (PSV) >/=200 cm/sec and renal-to-aortic peak systolic ratio (RAR) >/=3.5. The results of renal artery duplex and arteriography in 58 patients (107 kidneys) who underwent both exams were reviewed. Arteriography revealed 32 main renal arteries with >/=60% stenosis. The PSV criterion detected 29, for a sensitivity of 91%, specificity of 75%, positive predictive value (PPV) of 60%, negative predictive value (NPV) of 95%, and accuracy of 79%. Using RAR >/=3.5 provided a sensitivity of 72%, specificity of 92%, PPV of 79%, NPV of 88%, and accuracy of 86%. In a subset of 36 kidneys that had hilar scans, the criteria of acceleration time (AT) >/=100 cm/sec and index (AI) </=3.78 kHz/sec were evaluated. The AT and AI yielded sensitivity of 50% and 36%, specificity of 86% and 100%, PPV of 70% and 100%, NPV of 73% and 71%, and accuracy of 72% and 75%, respectively. Colorflow duplex scanning is clinically useful in screening for hemodynamically significant renal artery stenosis. The renal artery PSV criterion is highly sensitive, with a high NPV that obviates the need for arteriography in most cases of a negative duplex. The criteria of RAR, AT, and AI are of less value.