Characterizing trajectories of diabetes-related health parameters before diabetes diagnosis in diabetes subtypes: analysis of a 20-year long prospective cohort study in Sweden

BACKGROUND: Evidence is limited on whether alterations in diabetes-related health parameters are detectable before clinical diagnosis in novel diabetes subtypes. We investigated trajectories of diabetes-related health parameters in individuals with recently diagnosed type 2 diabetes (T2D). METHODS: Using data from the Stockholm Diabetes Prevention Programme cohort (SDPP) participants (n = 215) with recent onset T2D were classified as having severe insulin-deficient diabetes (SIDD, 9%), severe insulin-resistant diabetes (SIRD, 15%), mild obesity-related diabetes (MOD, 14%) and mild age-related diabetes (MARD, 62%). Participants without a family history of diabetes who remained diabetes-free throughout the study served as the controls (n = 2531). Multilevel longitudinal mixed-effects models were used to analyse the trajectories of fasting plasma glucose (FPG) and insulin, body mass index (BMI), homeostasis model assessment estimates of beta-cell function (HOMA2-B) and insulin resistance (HOMA2-IR), waist-to hip-ratio (WHR), diastolic blood pressure (DBP) and systolic blood pressure (SBP) up to 20 years before and 10 years after T2D diagnosis. Pairwise comparisons of the estimated marginal means were used to assess differences between all groups. RESULTS: Individuals with SIDD consistently exhibited the highest FPG concentrations (P < 0.001) and the steepest decline in HOMA2-B levels among all subtypes. BMI was higher in MOD and SIRD than in SIDD and MARD throughout the study period (P < 0.01). Individuals with SIRD showed the highest fasting insulin concentrations and higher HOMA2-IR than those with MOD and MARD (P < 0.001). WHR and DBP were comparable between subgroups, while SIDD had higher SBP than MOD (P = 0.03). The control group exhibited the mildest trajectories across all parameters except for HOMA2-B. Notably, these changes were visible up to 20 years prior to diagnosis. CONCLUSIONS: In a Swedish population, trajectories of diabetes-related health parameters differed up to 20 years before diagnosis between the T2D-related subtypes and controls. This might support early prediction of subtype-specific risks for long-term complications, allowing early initiation of personalized treatment strategies.