Breast cancer survival in Soweto, Johannesburg, South Africa: A receptor-defined cohort of women diagnosed from 2009 to 11.

Background: South Africa's public healthcare system is better equipped to manage breast cancer than most other SSA countries, but survival rates are unknown.

Methods: A historical cohort of 602 women newly diagnosed with invasive breast carcinoma during 2009-2011 at Chris Hani Baragwanath Academic Hospital, Soweto, Johannesburg, was followed using health systems data to December 2014. 'Overall survival' time was defined from diagnosis to death or terminal illness. Cox regression was used to estimate hazard ratios (HR) associated with woman and tumour characteristics.

Results: During a median 2.1 years follow-up (IQR 0.5-3.8), 149 women died or were classified terminally ill; 287 were lost-to-follow-up. 3-year survival was 84% for early stage (I/II) and 56% for late stage (III/IV) tumours (late v early: HR 2.8 (95% confidence interval (CI): 1.9-4.1), however the 42% cumulative losses to follow-up over this period were greater for late stage, half of which occurred within 6 months of diagnosis. After mutual adjustment for stage, grade, age, receptor subtype and HIV status, lower survival was also associated with triple negative (HR 3.1 (95% CI: 1.9-5.0)) and HER2-enriched (2.5 (95% CI: 1.4-4.5)) compared to ER/PR+ HER2- tumours, but not with age or HIV-infection (1.4 (95% CI: 0.8, 2.3)).

Conclusion: In this South African cohort, breast cancer survival is suboptimal, but was better for early stage and hormone receptor-positive tumours. Efforts to reduce clinic losses in the immediate post-diagnosis period, in addition to early presentation and accelerated diagnosis and treatment, are needed to prevent breast cancer deaths, and survival improvements need to be monitored using prospective studies with active follow-up.