Observational studies have found higher long-term mortality after transurethral prostatectomy (TURP) than after open prostatectomy (OP) and that this difference remained after statistical adjustment for comorbidity. This higher mortality has been attributed to the transurethral procedure itself. This association is reassessed here. Time-sequenced hospital and death records were analysed for 13,815 men undergoing prostatectomy (not for cancer) during the years 1963-85. TURP had a lower 30-day case-fatality rate than had OP, a similar 90-day rate, but TURP had a higher one-year rate. By inference, any excess mortality after TURP must begin shortly after the first postoperative month. However, plotting mortality expressed as SMRs, for the three years after operation showed no increase in long-term mortality after TURP, nor was there any concomitant increase in one-year death rates after prostatectomy as TURP replaced OP. Long-term mortality after TURP was close to that expected from background population rates: after SMR for TURP, for the second and third postoperative years, was 100 (95% CI 93-107). In contrast, long-term mortality after OP was lower than expected from population rates with a corresponding SMR of 79 (95% CI 71-88). The apparent excess in long-term mortality after TURP is unlikely to be caused by the operation itself. It is more likely to reflect relatively low long-term mortality in OP patients as a consequence of OP patients having been relatively fitter than those having TURP.