Purpose The purpose was to establish the impact on survival of early detection of a local recurrence of breast cancer as compared to late detection. Design A meta-analysis was carried out using Cochrane review manager software (RevMan version 4.2). Studies were included if women were treated for primary breast cancer without evidence of distant metastasis at primary diagnosis and if these concerned routine follow-up strategies focusing on the early detection of curable recurrences. Data regarding the risk for death were derived from each study. Multi level models were used to study heterogeneity by using MLWin. Results Thirteen studies concerning 2,263 patients were included. Early detection of breast cancer recurrences during follow-up gave a significantly better survival as compared to late detected recurrences (HR: 1.68 (95% CI: 1.48-1.91)). Survival was better when the recurrence was found by mammography instead of physical examination or in patients without symptoms as compared to those with symptoms (HR: 2.44 (95% CI: 1.78-3.35); HR: 1.56 (95% CI: 1.36-1.79), respectively). If all breast cancer recurrences would be detected earlier, that 5-8 deaths (i.e. an absolute reduction in mortality of 17-28%) would be avoided by performing routine follow-up during a 10 year-period for 1,000 breast cancer patients. Conclusion These data support the hypothesis that detection of isolated loco-regional or contra-lateral breast cancer recurrences in patients without symptoms has beneficial impact on survival of breast cancer patients when compared to late symptomatic detection.