METHODS Retrospective review. OBJECTIVE An annual screening cystoscopy protocol was begun at our institution in an attempt to minimize the morbidity and mortality of bladder cancer in the chronically catheterized spinal cord injured (SCI) population. The objectives of this study are: (1) to present the results of 6 years of screening for primary bladder cancer in this population, and (2) examine the suitability of this protocol based upon accepted principles of cancer screening. METHODS Veterans hospital, Seattle, WA, USA. METHODS SCI patients selected for screening cystoscopy were those who had been continuously catheterized for 10 or more years, or were smokers who had been catheterized for 5 or more years. Biopsies and/or urine cytologies were taken at the surgeon's discretion. RESULTS Fifty-nine patients underwent 156 cystoscopy procedures from January 1992 through December 1997. The vast majority of patients were at risk for autonomic dysreflexia, so cystoscopy was performed with anesthesia. No bladder cancers were diagnosed by screening cystoscopy. All bladder biopsies and cytology specimens were benign. During the same period of time four SCI patients presented with symptomatic bladder cancers. Two patients did not fit the criteria for surveillance, one patient was not being followed by the SCI unit and presented to an outside physician, and one patient had a screening cystoscopy 4 months prior to presenting with bladder cancer. CONCLUSIONS Cystoscopy does not fulfil the accepted criteria for screening for primary bladder cancer in SCI patients. The disease does not appear to be amenable to screening, the population to be screened is not easily definable, and the costs are excessive compared to the low cancer detection rate.