To explore uric acid (UA) trajectories in different body mass index (BMI) populations and to determine their associations with incident diabetes.

A total of 4566 adults without diabetes in 2011 were enrolled. All participants underwent a medical examination every year until 2016, and were classified into three subgroups based on BMI: non-obese (BMI<24kg/m²); overweight (BMI ≥24kg/m² but<28kg/m²); and obese (BMI ≥28kg/m²). Distinct UA trajectories were identified through group-based trajectory modelling (GBTM). Cox proportional-hazards models were applied to evaluate the associations between UA trajectories and risk of incident diabetes.

UA trajectories were identified in the three BMI subgroups: ‘low’ (42.4% in non-obese, 22.1% in overweight, 22.0% in obese); ‘moderate’ (32.5%, 41.1%, 34.8%); ‘moderate-high’ (18.6%, 29.5%, 30.8%); and ‘high’ (6.5%, 7.3%, 12.4%). After a 5-year follow-up, 170 (3.7%) participants had developed diabetes. The prevalence of new-onset diabetes increased progressively with the higher UA trajectories in the BMI groups (P values<0.05). Whereas compared with the low trajectory, a significant association between a high UA trajectory and incidence of diabetes was observed only in the overweight population [hazard ratio (HR): 6.95, 95% confidence interval (CI): 1.90–25.45], with no significant associations found in either the non-obese (HR: 0.67, 95% CI: 0.13–3.52) or obese (HR: 0.40, 95% CI: 0.06–2.64) populations, in the fully adjusted model.

Higher UA trajectories are significantly associated with an increased risk of incident diabetes, thereby suggesting that monitoring UA trajectories over time may assist in the identification of prediabetes and diabetes, particularly in the overweight population.