Background: Gastric bypass surgery induces early remission or significant improvement in type 2 diabetes (T2D).Aim: To assess effectiveness of stopping glucose-lowering treatment at the time of surgery. Design: Observational cohort analysis. Methods: We identified 101 patients (62 women) with T2D who had undergone gastric bypass surgery at a mean (SD, standard deviation) age of 51.4 (9.0) years. We recorded weight, body mass index (BMI), glycosylated haemoglobin (HbA1c), blood pressure (BP), total and high-density lipoprotein (HDL) cholesterol preoperatively and at a median 4, 12 and 24 months postoperatively, and changes to glucose-lowering therapy. Results: Mean (SD) baseline BMI was 50.3 (6.3) kg/m2, HbA1c 65.3 (18.5) mmol/mol, systolic BP 146.0 (18.0) mmHg, diastolic BP 87.0 (10.8) mmHg and total cholesterol-to-HDL cholesterol ratio 4.0 (1.2). Mean (95% confidence interval) reduction in BMI was 16.4 (14.1-18.7) kg/m2, HbA1c 23.6 (17.6-29.6) mmol/mol, systolic BP 12.9 (5.9-19.8) mmHg, diastolic BP 6.1 (1.8-10.5) mmHg and total cholesterol-to-HDL cholesterol ratio 1.1 (0.6-1.5) at 24 months (P <0.001 for all measures). Although 91% of patients were receiving glucose-lowering therapies preoperatively, complete (HbA1c <42 mmol/mol) and partial (HbA1c 42-48 mmol/mol) remissions of T2D were seen in 62.1% and 5.2% at 2 years postoperatively. Conclusions: Cessation of glucose-lowering therapies in people with T2D at the time of gastric bypass surgery was clinically effective. The majority of patients remained in complete or partial remission of diabetes up to 2 years postoperatively. © The Author 2014. Published by Oxford University Press on behalf of the Association of Physicians. All rights reserved.