BACKGROUND: Hydrocolonic ultrasonography--abdominal ultrasonography in conjunction with the retrograde instillation of water into the colon--has been advocated as an alternative to colonoscopy for detecting colorectal polyps and cancer. We conducted a prospective, blinded trial to evaluate the procedure further. METHODS: Fifty-two consecutive patients (50 men and 2 women; average age, 62 years) who were referred for colonoscopy underwent hydrocolonic ultrasonography followed by colonoscopy. The physicians performing colonoscopy were blinded to the ultrasound results. Patients who had a history of colonic polyps or tumors or who had previously undergone flexible sigmoidoscopy or colonoscopy were excluded. RESULTS: Twenty-two patients had normal results on colonoscopy, 26 had polyps, 3 had cancer and polyps, and 1 had cancer alone. Twenty patients had polyps less than 7 mm in diameter, eight had polyps 7 mm or more in diameter, and one had a polyp of unknown size. Hydrocolonic ultrasonography did not detect any cancers and detected only one polyp > or = 7 mm and one polyp < 7 mm in diameter. The overall sensitivity of ultrasonography for identifying any polyp was 6.9 percent, and for identifying a polyp > or = 7 mm, it was 12.5 percent. Ultrasonography suggested the presence of five masses and five polyps that were not confirmed by colonoscopy. Six patients had incomplete ultrasound studies because of discomfort or the inability to retain water. There were two complications: one patient had two vasovagal episodes, and another had diaphoresis. CONCLUSIONS: Hydrocolonic ultrasonography was less useful than colonoscopy for detecting colorectal polyps and cancers. The usefulness of the technique in screening for colonic polyps and tumors appears to be limited.