The best approximation of the natural course of early prostate cancer comes from studies on deferred treatment. In such studies, the 10-year disease-specific survival rate is 85-90% for patients with clinically localized tumours and 74% for those with tumours not confined to the prostate gland; 15-year data are sparse and diverging. Patients with poorly differentiated tumours have a worse prognosis than those with well or moderately differentiated tumours. Endocrine treatment of early prostate cancer is a revitalized concept, with the majority of the evidence for a beneficial effect coming from uncontrolled studies. Recently, however, data from two randomized studies have suggested that endocrine treatment may prolong survival in patients with non-metastatic prostate cancer. Also, in studies of radiotherapy for localized prostate cancer, a longer survival has been seen in patients randomized to neoadjuvant or adjuvant endocrine treatment compared with placebo. The current trend for earlier treatment means that the total time on endocrine therapy has increased and patients are younger when treatment is initiated. Therefore, randomized studies should not only substantiate a possible survival benefit of endocrine treatment, but also assess potential long-term side effects and their impact on quality of life.