Background
Parity and age at first pregnancy are well-established risk factors for breast cancer, but the effects of other characteristics of pregnancies are uncertain and the literature is inconsistent.

Methods
In a cohort of 83,451 parous women from the general population of the UK, which collected detailed information on each pregnancy and a wide range of potential confounders, we investigated the associations of length of gestation and birthweight of offspring in a woman’s pregnancies with her breast cancer risk, adjusting for a full range of non-reproductive as well as reproductive risk factors unlike in previous large studies.

Results
Gestation of the first-born offspring was significantly inversely related to the risk of pre-menopausal breast cancer (p trend = 0.03; hazard ratio (HR) for 26–31 compared with 40–41 weeks, the baseline group, = 2.38, 95% confidence interval (CI) 1.26–4.49), and was borderline significantly related to risk of breast cancer overall (p trend = 0.05). Risk was significantly raised in mothers of high birthweight first-born (HR for breast cancer overall = 1.53, 95% CI 1.06–2.21 for ≥ 4500 g compared with 3000–3499 g, the baseline group). For gestation and birthweight of most recent birth, there were no clear effects. Analyses without adjustment for confounders (other than age) gave similar results.

Conclusions
Our data add to evidence that short gestation pregnancies may increase the risk of breast cancer, at least pre-menopausally, perhaps by hormonal stimulation and breast proliferation early in pregnancy without the opportunity for the differentiation that occurs in late pregnancy. High birthweight first pregnancies may increase breast cancer risk, possibly through the association of birthweight with oestrogen and insulin-like growth factor 1 levels.