Colonoscopy, done to elucidate the cause of obscure rectal bleeding not diagnosable on the basis of previous proctoscopic and barium-enema examinations, resulted in a 17 per cent diagnostic yield in this series. Eighteen of 105 patients were found to have lesions: three had carcinomas, six had polyps, eight had colitis, and one had proctitis.
When bleeding had been demonstrated by the physician, there was a higher diagnostic yield (22.5 per cent) than when bleeding was reported only by the patient (5.9 per cent). Descriptions of the bleeding by the patients were found to be somewhat unreliable.
Sixteen of the 18 lesions revealed by the colonoscope were in the left colon.
It is suggested that the basic work-up for undiagnosed rectal bleeding should include history, physical examination, sigmoidoscopy, and barium-enema study. When no diagnosis is found, subsequent work-up must be individualized. When bleeding has been demonstrated to originate from the colon, colonoscopic examination should be performed.