To evaluate the influence of delay between first symptom and first treatment upon survival the medical records of 596 patients with breast cancer were reviewed. The following intervals were considered: less than 3 months; 3-6 months and greater than 6 months. Patients in the less than 3 months delay group had a better distribution by clinical stages and a 10-year survival rate higher than those in the longer delay groups (p = 0.034). However, within each stage no statistically significant difference in survival according to delay was observed. A Cox multivariate analysis revealed that performance status and stage of disease were independent predictors of survival, but not delay. Assuming the best prognosis for patients with clinical stages I and II and less than 3 months delay, the group with longer delay times had 15 deaths over what would have been predicted. This adverse effect was observed almost exclusively among patients over age 50 (14/15).