OBJECTIVES Type 2-diabetes and excess weight incur large costs to healthcare systems, but the association between weight progression in diabetes and healthcare costs is unknown. We investigated those relations using real world data in a sample of newly diagnosed diabetes patients in Sweden by using repeated body mass index (BMI) measurement and healthcare resource utilization data. METHODS Patients with a BMI (kg/m) measure at diagnosis and subsequent BMI measures at 12, 24, 36, 48, and 60 months were identified from a previously conducted register study. Individuals were classified into three groups based on their BMI change over 5 years: increase (>1 BMI unit increase), decrease (>1 BMI unit decrease), and stable (≤1 BMI unit change). Each group was stratified by BMI at diagnosis (BMI 18-25; 25-30; ≥30). Healthcare costs for each group were estimated by applying Swedish unit costs to the healthcare resource data extracted from electronic patient journals and a national patient register. RESULTS The study included 903 T2D patients (women, 43%; mean age, 62; mean HbA1c, 6.78%; mean BMI, 30.9). The BMI increased, decreased, or remained stable in 178 (20%), 387 (43%), and 338 (37%) patients, respectively. Among patients with baseline BMI 18-25 (n=104), the five year cumulative healthcare costs were €13,695, €9,059, €8,936, in the increase, decrease, and stable group, respectively. Corresponding figures were €11,470, €7,950, €8,683 for patients with baseline BMI 25-30 (n=321), €14,387, €9,465, and €9,302 for patients with BMI ≥30 (n=478). CONCLUSIONS In newly diagnosed diabetes, an increase in BMI lead to increased healthcare costs, irrespective of baseline BMI. Linking registry data on repeated BMI measurements and healthcare utilization is a valuable approach to investigate the association between weight changes and costs. Costs of interventions that maintain weight in patients with diabetes should be considered in the context of costs associated with weight gain.