br All consenting female patients aged years and
All consenting female patients aged 18 years and older, newly diagnosed with invasive breast cancer, were enrolled on the study. Patients who were HIV-unknown were counselled about HIV and were asked to give informed consent for testing. Patients newly diagnosed with HIV were offered post-test counselling. All HIV positive were asked to provide blood samples for CD 4 counts and HIV viral load testing. We recorded the time span since HIV diag-nosis, as per the first positive HIV serological test recorded on the National Health Laboratory system, and initiation of ART prior to histologically-confirmed breast cancer diagnosis. HIV positive
patients not yet on ART were sent to the HIV/ART clinic for initiation of ART prior to cancer treatment.
Our clinical staging of breast cancer followed the American Joint Committee on Cancer (AJCC) system . We also categorised pa-tients into early (Stage I/II) and late (Stage III/IV) stage disease. Pre-treatment pathology reporting included the histological diagnosis, tumour subtype, grade, oestrogen receptor (ER), progesterone re-ceptor (PR), human epidermal growth factor receptor 2 (HER2) and Ki 67. The Allred score was calculated from the intensity and pro-portion scores for oestrogen and progesterone receptors; speci-mens scored 3e8 were regarded as hormone receptor positive . HER2 was regarded as positive when the test showed 3 þ and negative when it was 1þ. A HER2 score of 2 þ was regarded as equivocal, and HER2 positive result was confirmed using in situ hybridization (FISH or SISH) . Ki 67 is a nuclear antigen used as a marker of cell proliferation; we categorised specimens in which <14% of MK-2206 expressed Ki67 as having low expression, as per the St Gallen 2011 guidelines [17,18]. The breast tumours were cat-egorised based on IHC 4 subtype into: Luminal A [ER and/or PR positive, HER2 negative, Ki 67 < 14%]; Luminal B [ER and/or PR positive and HER2 negative with Ki 67% 14% OR HER2 positive with any Ki 67]; HER2 positive subtype [ER and PR negative, HER2 positive]; and triple negative breast cancer (TNBC) [ ER, PR, and HER2 negative] .The Modified Bloom & Richardson grading system was used which is based on tubule formation, mitosis, and nuclear pleomorphism; a score of 5 denoted a Grade 1 (well differentiated) tumour, 6e7 a Grade 2 (moderately differentiated) tumour, and 8e9 a Grade 3 (poorly differentiated) tumour .
Patients were categorised as HIV positive, HIV negative or HIV unknown. Only 3.2% of the patients were HIV unknown and were excluded from further analysis. Demographic and clinico-pathological data were categorised and comparisons between HIV positive and HIV negative patients performed using c  tests. We obtained CD 4 counts and HIV viral loads for the HIV-infected pa-tients from the South African National Health Laboratory Service (NHLS) (www.nhls.ac.za) using results of tests performed closest to the date of breast cancer diagnosis. Viral load was categorised as either detectable (>50 copies/ml) or below detectable limits ( 50 copies/ml). Age, CD 4 count, and viral load (when detectable) were non-normally distributed and were thus represented as medians and inter-quartile ranges (IQRs) and compared using a Kruskal-Wallis test. Generalised linear models with binary outcomes and a log link function were used to determine prevalence ratios for non-missing variables to assess the relationship between stage at diagnosis, tumour grade, IHC4 subtype and presence or absence of metastatic disease with HIV status controlling for age at diagnosis (continuous) and ethnicity (black vs non-black). Collected data were analysed using STATA v12.1.
Of the 1050 patients newly diagnosed with invasive breast cancer enrolled in the study, 34 had unknown HIV status and were excluded from the analysis. The demographic and clinico-pathological characteristics of 1016 patients with and without HIV are shown in Table 1. Of these, 226 (22.2%) were HIV positive, 855 (84.2%) patients were self-reported as black. The median (interquartile range, IQR) age at diagnosis of axons analysed was 54 (IQR 44e64) years, and 560 (55.1%) were diagnosed at late stage disease (stage III/IV).