Annual recurrence rates (ARR) are used to assess changes in the risk of breast cancer recurrence following surgery. In this retrospective study, ARR were calculated from the clinical records of 2,209 breast cancer patients who had undergone surgery. The time-course changes of ARR associated with prognostic/predictive factors were calculated. Overall, ARR decreased for 5 years following surgery and then remained almost constant. In hormone receptor (HR)-negative patients, ARR peaked after 2 years and peaked again at 6-7 years. In HR-positive patients, ARR peaked at 2 years. ARR increased in relation to the number of lymph-node metastases for 5 years, and peaked after 2 years in the absence and presence of venous invasion. The log-rank test demonstrated significant differences in recurrence between HR-negative and HR-positive cancer up to 5 years post-surgery. The presence of venous invasion had a significant effect on recurrence in the first 5 years, and the presence of lymph-node metastasis had a significant effect on recurrence up to and after 5 years. In conclusion, prognostic/predictive factors affected breast cancer recurrence in the first 5 years but had a lesser effect on recurrence more than 5 years post-surgery.