BACKGROUND: Breast cancer remains the leading cause of cancer death in French women in spite of continuously improving management. The objectives of this study were to analyse trends in the metastasis-free interval over the past 30 years and to identify the prognostic factors of survival, while accounting for time dependency. METHODS: A total of 1613 patients diagnosed with invasive non-metastatic breast cancer at Saint Vincent de Paul Hospital, Lille, France between 1977 and 2013, were followed for outcome (metastasis-free interval). Cohort entry time delay, a continuous temporal covariate, was defined to assess improvement of outcome. Data were analysed using the Cox proportional hazards model and presented as hazard ratio (HR). RESULTS: Metastatic disease developed during follow-up in 446 (27.6%) patients. Cohort entry time delay exhibited strong independent prognostic value while accounting for multiple prognostic factors including: tumour size (HR = 1.62, 95 %CI 1.37-1.91); rapid tumour growth (HR = 1.59, 95%CI 1.17-2.16); lymph node ratio (HR = 2.29, 95%CI 1.97-2.66); histological grade (grade 2 was significant only during the first 10 years after diagnosis, grade 3 and progesterone receptor status only during the first 5 years after diagnosis); and oestrogen receptor status (significant only during the first 8 years (HR = 0.75, 95%CI 0.58-0.96)). CONCLUSIONS: The current study showed an improvement in the prognosis of breast cancer patients over the past 30 years and pointed to the importance of evaluating covariates with time-varying effects.