We determined the effect of cystoscopy (flexible and rigid), transrectal ultrasonography (with and without needle biopsy of the prostate) and transurethral resection of the prostate or bladder tumour on the serum prostate-specific antigen (PSA) concentration. Samples were taken from 60 men before and up to 14 days following these procedures. Flexible cystoscopy did not result in a significant increase in serum PSA concentration, with a median increase of 0.1 microgram/L (P > 0.05). Small but statistically significant increases in serum PSA levels 1 day post-procedure were observed following rigid cystoscopy and transrectal ultrasound without biopsy. The median increase in serum PSA concentration following rigid cystoscopy was 0.15 microgram/L (P = 0.04) and following transrectal ultrasound was 0.3 microgram/L (P = 0.01). In both cases the serum PSA level had normalized by 2 days post-procedure. Transurethral resection of bladder tumours resulted in a variable rise in serum PSA, with a median increase of 2.6 micrograms/L after 1 day, which returned to normal over 7-14 days. Ultrasound-guided needle biopsy of the prostate and transurethral resection of the prostate produced significant increases in serum PSA levels, which took up to fourteen days to return to normal. The median increase in serum PSA following needle biopsy was 6.0 micrograms/L and following transurethral resection of the prostate (TURP) was 13 micrograms/L. Samples for PSA measurement may safely be taken within 24-48 h of flexible cystoscopy and transrectal ultrasonography (TRUS) providing prostatic biopsy is not carried out. For other procedures it is necessary to wait for at least 14 days to ensure that false positive PSA results are not obtained.