Joint trajectories of adiposity and systemic inflammation and incident diabetes among older adults: a bi-cohort prospective study with mediation analysis.

BACKGROUND: Obesity and chronic low-grade inflammation are established risk factors for type 2 diabetes, yet their joint longitudinal evolution and combined contribution to diabetes incidence remain poorly characterized in aging populations. OBJECTIVE: To identify joint trajectories of adiposity and systemic inflammation, evaluate their associations with incident diabetes, and quantify the mediating role of glycated hemoglobin (HbA1c) across two independent cohorts. METHODS: This prospective study included 3,733 diabetes-free adults aged ≥50 years from the English Longitudinal Study of Ageing (ELSA, n = 1,778) and the Health and Retirement Study (HRS, n = 1,955), with approximately 8 years of follow-up. Group-based multi-trajectory modeling identified joint trajectory groups under two strategies: general adiposity (body mass index paired with log-transformed C-reactive protein) and central adiposity (waist-to-height ratio paired with log-transformed C-reactive protein). Associations with incident diabetes were estimated using Cox proportional hazards models pooled via random-effects meta-analysis. Mediation analysis quantified the proportion statistically mediated through mid-wave HbA1c. RESULTS: Four trajectory groups were identified under each strategy in both cohorts. A graded association was observed: the highest-risk group exhibited substantially elevated diabetes risk under both general adiposity (pooled HR 6.00, 95% CI 3.71-9.70) and central adiposity strategies (pooled HR 5.69, 95% CI 3.37-9.59) after full adjustment. Under the conservative outcome definition, HbA1c mediated 13.7-27.5% of the total effect (25.7-44.8% under the original definition representing an upper bound). Trajectory classification yielded modestly higher C-statistics over baseline cross-classification (ΔC 0.061-0.067, statistically significant in HRS), with significant net reclassification improvement primarily driven by event cases. Results were consistent across strategies and cohorts. CONCLUSION: Co-evolving trajectories of adiposity and systemic inflammation exhibit a graded association with incident diabetes in older adults, partially mediated through glycemic deterioration. Trajectory modeling offered modest discriminative gains over single-time-point assessment. These findings support concurrent adiposity-inflammation monitoring for diabetes risk stratification in aging populations.