Background: Studies have shown an increased risk of breast cancer (BC) among women with type 2 diabetes (T2D). This association could be causal, related to hyperglycaemia or hyperinsulinaemia, or could be triggered by other factors such as obesity and physical inactivity which are known risk factors for T2D and for BC in post-menopausal women. Methods: A meta-analysis was performed to assess the risk of BC in T2D patients compared to non-diabetic women, with special attention to changes in the risk of BC when the body mass index (BMI) was included in multivariate analyses. Studies were selected if they had a prospective design. Studies that compared BC incidence in T2D women to the incidence in the general population were excluded. Summary relative risks (SRR) and 95% confidence intervals (CI) were computed using random-effects models. Results: Eighteen studies were included in the meta-analysis, based on 28,230,143 person-years of follow-up and 320,111 BC cases. Compared to non-diabetic women, the SRR of BC among T2D women was 1.13 (95% CI: 1.04, 1.24). There was a large amount of unexplained heterogeneity of results across studies (I² = 95%), but no indication of publication bias. Three studies reported the risk of BC by BMI category, with a consistently higher risk of BC associated with increasing BMI. In the 9 studies that adjusted for adiposity, the SRR decreased to 1.05 (95% CI: 0.97, 1.14) while the heterogeneity of results across studies reduced to I² = 21%. Only two studies reported data by menopausal status. In contrast, in the 9 studies that did not adjust for adiposity, the SRR increased to SRR = 1.19 (95% CI: 1.01, 1.39), while the heterogeneity remained high (I² = 98%). Five studies reported data for post-menopausal women only, with a SRR of 1.13 (95% CI: 0.89; 1.44) and high heterogeneity (I² = 94%) with one study representing 82% of the weight in the meta-analysis. Conclusions: This analysis provides evidence for a moderately increased risk of BC in T2D women. The effect of the adjustment for BMI on the SRR and on the heterogeneity suggests that the higher risk of BC among women with T2D may not be due to the diabetes itself but to adiposity. New studies should examine the relationship between BC, T2D and adiposity in premenopausal women.