Purpose Previous studies have suggested an association between metformin use and improved outcome in patients with diabetes and breast cancer. In the current study, we aimed to explore this association in human epidermal growth factor receptor 2 (HER2 ) -positive primary breast cancer in the context of a large, phase III adjuvant trial. Patients and Methods The ALTTO trial randomly assigned patients with HER2-positive breast cancer to receive 1 year of either trastuzumab alone, lapatinib alone, their sequence, or their combination. In this substudy, we evaluated whether patients with diabetes at study entry-with or without metformin treatment-were associated with different disease-free survival (DFS), distant disease-free survival (DDFS), and overall survival (OS) compared with patients without diabetes. Results A total of 8,381 patients were included in the current analysis: 7,935 patients (94.7%) had no history of diabetes at diagnosis, 186 patients (2.2%) had diabetes with no metformin treatment, and 260 patients (3.1%) were diabetic and had been treated with metformin. Median follow-up was 4.5 years (0.16 to 6.31 years), at which 1,205 (14.38%), 929 (11.08%), and 528 (6.3%) patients experienced DFS, DDFS, and OS events, respectively. Patients with diabetes who had not been treated with metformin experienced worse DFS (multivariable hazard ratio [HR], 1.40; 95% CI, 1.01 to 1.94; P = .043), DDFS (multivariable HR, 1.56; 95% CI, 1.10 to 2.22; P = .013), and OS (multivariable HR, 1.87; 95% CI, 1.23 to 2.85; P = .004). This effect was limited to hormone receptor-positive patients. Whereas insulin treatment was associated with a detrimental effect, metformin had a salutary effect in patients with diabetes who had HER2-positive and hormone receptor-positive breast cancer. Conclusion Metformin may improve the worse prognosis that is associated with diabetes and insulin treatment, mainly in patients with primary HER2-positive and hormone receptor-positive breast cancer.