Obesity is a risk factor for the development of new cases of breast cancer and also affects survival in women who have already been diagnosed with breast cancer. Early studies of obesity and breast cancer survival have been summarised in two meta-analyses, but the latest of these only included studies that recruited women diagnosed as recently as 1991. The primary aim of this study was to conduct a meta-analysis that included the more recent studies. A systematic search of MEDLINE, EMBASE and CINAHL was conducted to identify original data evaluating the effects of obesity on survival in newly diagnosed breast cancer patients. Adjusted hazard ratios (HR) from individual studies were pooled using a random effects model. A series of pre-specified sensitivity analyses were conducted on factors such as overall versus breast cancer survival and treatment versus observational cohort. The meta-analysis included 43 studies that enrolled women diagnosed with breast cancer between 1963 and 2005. Sample size ranged from 100 to 424168 (median 1192). The meta-analysis showed poorer survival among obese compared with non-obese women with breast cancer, which was similar for overall (HR = 1.33; 95% confidence interval (CI): 1.21, 1.47) and breast cancer specific survival (HR = 1.33; 95% CI: 1.19, 1.50). The survival differential varied only slightly, depending on whether body mass index (1.33; 1.21, 1.47) or waist–hip ratio (1.31; 1.08, 1.58) was used as the measure of obesity. There were larger differences by whether the woman was pre-menopausal (1.47) or post-menopausal (1.22); whether the cohort included women diagnosed before (1.31) or after 1995 (1.49); or whether the women were in a treatment (1.22) or observational cohort (1.36), but none of the differences were statistically significant. Women with breast cancer, who are obese, have poorer survival than women with breast cancer, who are not obese. However, no study has elucidated the causal mechanism and there is currently no evidence that weight loss after diagnosis improves survival. Consequently, there is currently no reason to place the additional burden of weight loss on women already burdened with a diagnosis of cancer. Further research should concentrate on assessing whether factors such as diabetes or type of chemotherapy modify the obesity effect and on understanding the causal mechanism, in particular the role of relative under-dosing.