To evaluate the relationship between the clinical and pathological parameters of the primary tumor and disease-free survival (DFS) in a sample of hospital cases of invasive breast cancer.We performed a retrospective cohort study in 635 patients recruited at San Cecilio University Hospital in Granada (Spain) between 1994 and 2006. Information on the primary tumor and the outcomes of patients was collected by reviewing the medical records. Predictors of recurrence and/or metastasis and DFS (follow up of 3, 5 and 10 years) were analyzed by using Cox regression analysis.Multivariate models adjusted for age, tumor size, lymph nodal status, histological grade and estrogen and progesterone receptor expression showed a higher risk of recurrence and/or metastasis and lower DFS (adjusted relative risk, 95% confidence intervals) with tumor size (3 yrs: 3.00, 1.79-5.03; 5 yrs: 2.56, 1.65-3.98; 10 yrs: 2.16, 1.44-3.24), lymph nodal status (3 yrs: 4.58, 2.42-8.65; 5 yrs: 3.84, 2.35-6.30; 10 yrs: 3.08, 2.05-4.61), lymphovascular invasion (5 yrs: 1.88, 1.16-3.04; 10 yrs: 2.19, 1.43-3.35), multifocal and/or multicenter tumors (3 yrs: 2.69, 1.46-4.96; 5 yrs: 1.90, 1.08-3.35) and p53 protein expression (3 yrs: 2.03, 1.00-4.09). DFS was positively associated with an increased expression of progesterone receptor (3 yr: 0.48, 0.26-0.89; 5 yrs: 0.58, 0.35-0.97; 10 yrs: 0.59, 0.38-0.90).The biological characteristics of the primary tumor can be used to identify patients with distinctive prognoses and DFS, and could be helpful in making individual follow up strategies.