Objective: To investigate the relations between lifestyle factors (diet and exercise), glycated haemoglobin (HbA 1c ) and body mass index (BMI) in older adults with diabetes. Design and setting: A community hospital-based cross-sectional study of 150 noninstitutionalized, ambulatory adults (Z65 y) with diabetes, residing within New Zealand's Kapiti region. Subjects: Patients were recruited from all general practices; two diabetes clinics; local diabetes society and through advertisements in community newspapers. A total of 211 eligible people were identified, but 60 refused to participate and one withdrew. In all, 150 people completed the study (71% participation rate). Methods: Nutrient intakes were calculated by a food-frequency questionnaire. Physical activity was assessed by interview using a validated questionnaire. Medical history and demographic data were obtained by interview or self-completed questionnaires; height, weight and HbA 1c were measured. Multivariate models using bootstrapping and stepwise linear regression were used to select factors associated with HbA 1c and BMI. Results: Each five-unit increase in energy from dietary saturated fat and five-unit increase in BMI were associated with 6% (95% confidence interval ¼ 2-10%; P ¼ 0.004) and 4% (0.3-7%; P ¼ 0.031) increases in HbA 1c , respectively. For females with moderate, compared with low overall activity, there was a 14% (7-20%; P ¼ 0.000) reduction in BMI while for males the reduction was only 5% (À1-11%; P ¼ 0.116). BMI decreased 5% (2-9%; P ¼ 0.004) with each 10-y increase in age, while a fiveunit increment in energy from dietary sucrose was associated with a 6% (1-11%; P ¼ 0.025) increase in BMI. Conclusions: Reducing dietary saturated fat and excess body weight may be useful means of improving glycaemic control in older adults with diabetes. Increasing physical activity and reducing energy from dietary sucrose may assist weight control, the former particularly in women.