Aims: Data on factors associated with prognostically unfavourable breast cancer using population-based breast cancer family studies is scarce. Our aims were to study survival by selected demographic and tumour characteristics, and to examine whether oestrogen-receptor status is an independent predictor of survival, especially in young women with breast cancer.

Methods: We included 1,196 participants of the Australian Breast Cancer Family Study, a population-based family study of early-onset breast cancer carried out in Melbourne and Sydney, aged <60 years at diagnosis of an incident, first primary invasive breast cancer during 1992-95. Overall survival (time from date of invasive breast cancer diagnosis to death from any cause) was assessed using Kaplan-Meier curves. A flexible parametric survival model was fitted to overall survival with oestrogen-receptor status as a time-dependent covariate. Multivariable-adjusted mortality hazard ratios (mHR) for oestrogen-receptor status were derived using Cox regression. All statistical tests were two-sided. P-values <0.05 were considered statistically significant.

Results: After a total follow-up of 14,717 person-years, there were 375 deaths. The estimated 5- and 10-year overall survival for all study participants were 85.7% and 75.8%, respectively. With increasing age at diagnosis and with decreasing tumour grade, the overall survival was higher. At both 5 and 10 years, women with oestrogen-receptor negative tumours showed lower rates of survival compared with oestrogen-receptor positive tumours (p-value <0.001). The flexible parametric survival model gave a mHR of 1.34 (1.08-1.65) and confirmed a greater risk of death for women with oestrogen-receptor negative tumours up to 5 years over oestrogen-receptor positive tumours, however, adjusting for tumour characteristics including tumour grade gave a mHR of 1.12 (95% CI: 0.90-1.40) for oestrogen-receptor negative tumours compared with oestrogen-receptor positive tumours. 

Conclusions: Results confirm lower survival rates in younger women diagnosed with high grade, oestrogen-receptor negative breast cancers. However, preliminary regression analyses show that the oestrogen-receptor status may not be an independent predictor of survival.