Pre-diagnostic smoking behaviour and poorer prognosis in a German breast cancer patient cohort - Differential effects by tumour subtype, NAT2 status, BMI and alcohol intake

BACKGROUND: Inconsistent associations of smoking and breast cancer-specific mortality might be explained by subgroups of patients with different susceptibility to harmful effects of smoking. METHODS: We used a prospective cohort of 3340 postmenopausal breast cancer patients aged 50-74 and diagnosed with invasive tumours 2001-2005 in Germany, with a median follow-up time of 6 years. The effect of pre-diagnostic smoking behaviour on mortality outcomes and risk of recurrence was investigated using delayed entry Cox regression analysis. Differential effects according to N-acetyltransferase (NAT2) status, BMI, alcohol consumption, and tumour subtypes were assessed. RESULTS: Overall, smoking at time of breast cancer diagnosis versus never/former smoking was non-significantly associated with increased breast cancer-specific mortality and risk of recurrence (HR 1.23, 95% CI 0.93-1.64, and HR 1.29, 95% CI 0.95-1.75, respectively). Associations were consistently stronger in NAT2 slow than in fast acetylators for all mortality outcomes. Breast cancer-specific mortality was significantly increased in smokers with NAT2 slow acetylating status (HR 1.77, 95% CI 1.13-2.79) but not in those with fast acetylating status (HR 1.09, 95% CI 0.60-1.98; Pheterogeneity=0.19). Smoking was associated with significantly poorer outcomes for triple negative and luminal A-like tumours (e.g. all-cause mortality: HR 1.93, 95% CI 1.02-3.65, and HR 2.08, 95% CI 1.40-3.10, respectively). Risk of recurrence was significantly increased for women with HER2 positive tumours (HR 3.64, 95% CI 1.22-10.8). There was significant heterogeneity by BMI for non-breast cancer-specific mortality (<25 kg/m(2): HR 2.52, 95% CI 1.52-4.15 vs. >/=25 kg/m(2): HR 0.94, 95% CI 0.38-2.36; Pheterogeneity=0.04). CONCLUSION: The harmful effects of smoking may be particularly relevant for certain subgroups of breast cancer patients. This may include patients with NAT2 slow acetylation status or with tumour subtypes other than luminal B, such as luminal A tumours who usually have a rather good prognosis. Emphasis on smoking cessation programmes for all cancer patients should be strengthened