SUMMARYA prospective cohort study with a 1-year telephone follow-up was performed to estimate the incidence of surgical-site infections (SSIs) in hip and knee prostheses. Twenty-five public hospitals equipped with orthopaedic units in two Italian regions were involved. An 8-month surveillance period was set to obtain significant data at a regional level and data were collected by the infection control nurses of each centre. One-year follow-up was completed in 75% of cases. SSIs were recorded in 45 cases (incidence rate 1·9/100 person-years, 95% CI 1·4-2·5). Thirty-six percent of SSIs were diagnosed during hospitalization and 95·3% within 90 days post-operation. This incidence is similar to that reported by European and USA surveillance systems. The proposed investigative method proved to be reliable and feasible. A prolonged surveillance for at least 3 months post-operation yields a good estimate of SSI in joint replacements.