To investigate a cohort of patients undergoing active surveillance (AS) for localized prostate cancer (PCa) with regard to possible differences in discontinuation rates, subsequent therapies, reasons for intervention and pathological findings after deferred surgery after patient stratification into very-low-risk, low-risk and intermediate-/high-risk PCa groups. The HAROW study was a non-interventional, observational, outcomes research study on the management of localized PCa in the community setting. A total of 468 patients were prospectively enrolled in the HAROW study, with a mean follow-up of 28.5 months. Treating urologists reported clinical variables, information on therapy and clinical course of disease at 6-month intervals. Of 468 patients under AS, 244 were stratified into very-low-risk, 142 into low-risk and 82 into intermediate-/high-risk groups. Of these patients, 112 discontinued AS. Discontinuation rates were 25.4% in the very-low-risk, 21.1% in the low-risk and 24.4% in the intermediate-/high-risk groups (P = 0.633). The main reasons for intervention were biopsy upgrade and/or prostate-specific antigen elevation in the very-low- and low-risk groups, and patient preference in the intermediate-/high-risk group (P < 0.05). No significant differences were found regarding subsequent therapies and pathological findings after deferred surgery. Our results show no differences in the outcome of risk-stratified patients in the specified risk groups managed with AS, while switching to an invasive treatment on the patient's request was more frequent in the intermediate-/high-risk group.