Although transurethral resection of the prostate provides an effective treatment for obstructive voiding symptoms associated with prostate cancer, there is growing concern about the possible role of transurethral resection in the dissemination of this malignancy. To determine the effect of transurethral resection on the rate of development of distant metastasis, we analyzed a large series of patients (379) treated at our institution with definitive radiotherapy for localized prostate cancer that was diagnosed by either needle biopsy or transurethral prostatic resection. In our series the presence of lymph node metastasis was documented by pelvic lymph node dissection in all patients. An initial univariate analysis suggested that patients diagnosed by transurethral resection had distant metastases significantly more rapidly than patients diagnosed by needle biopsy. However, transurethral resection usually was performed because of the presence of obstructive voiding symptoms and such patients were much more likely to have positive lymph node dissections than patients without obstructive voiding symptoms. A proportional hazards regression analysis showed that nodal status and the degree of obstructive voiding symptoms at diagnosis were independent and powerful predictors of the interval to distant metastases, along with stage and grade. The type of initial biopsy (transurethral prostatic resection versus needle biopsy) had no independent prognostic significance in this analysis. Among patients who had substantial obstructive voiding symptoms there was no significant difference in interval to distant metastases between the transurethral prostatic resection and needle biopsy groups. We conclude that the apparent adverse effect of transurethral prostatic resection results from the poor prognosis of tumors causing obstructive voiding symptoms rather than as a direct result of the resection itself.