PURPOSE: Previous studies suggest that larger birth size is associated with a higher breast cancer incidence, but studies on birth measures and mortality in breast cancer cases are scarce. This study investigates survival of women after breast cancer diagnosis (n = 437) in the Uppsala Birth Cohort born in 1915-1929. METHODS: Cox regression was used to analyze mortality from any cause after a breast cancer diagnosis. Birth measures including gestational age (GA), birth weight (BW), BW for GA, birth length, and ponderal index (PI) were converted to standard deviation (SD) scores, and all analyses were adjusted for age and calendar time at diagnosis. Analyses were performed with and without adjustment for other birth measures, reproductive history, and adult socioeconomic position. RESULTS: In fully adjusted analyses, one SD increase in GA was associated with 17% [95% confidence interval (CI) 6-26%] lower mortality and one SD increase in BW was associated with 29% (7-56%) higher mortality. PI showed a weaker trend in the same direction: hazard ratio = 1.16 (95% CI 1.03-1.30). CONCLUSIONS: Our results bring in new evidence that both high GA and low BW predict a better survival in breast cancer cases. Further studies need to investigate mediation of these associations.