Objective To examine the age-related trajectories of the energy-adjusted dietary inflammatory index (E-DII) and their longitudinal association with the incidence of type 2 diabetes mellitus (T2DM) among middle-aged and elderly Chinese residents, providing scientific evidence for dietary guidance in T2DM prevention.

Methods Data were obtained from the China Health and Nutrition Survey (CHNS) cohort (1991–2018). A total of 4 342 participants aged ≥ 40 years without baseline diabetes, who had completed ≥ 3 rounds of dietary surveys and ≥ 2 rounds of assessments on sociodemographics, lifestyle, medical history, anthropometry, and blood glucose, were included. Group-based trajectory modeling was employed to identify E-DII trajectory patterns, and a two-level Cox proportional hazards regression model was used to analyze the association between E-DII trajectories and T2DM incidence.

Results The study included 4 342 middle-aged and elderly participants aged ≥ 40 years, contributing 29 979 person-years of follow-up, mean follow-up (6.90 ± 2.54) person-years. During the follow-up, 401 T2DM cases were identified, yielding an incidence density of 13.38 per 1 000 person-years and a cumulative incidence of 9.24%. Three distinct E-DII trajectory groups were identified for both sexes. Among males, the three groups were labeled as "moderate pro-inflammatory–increasing" (78.82%), "moderate pro-inflammatory–decreasing" (6.10%), and "high pro-inflammatory–stable" (15.08%) groups, with cumulative T2DM incidence rates of 10.05%, 14.17%, and 11.45%, respectively. Among females, the three groups were "low pro-inflammatory–increasing" (16.60%), "moderate pro-inflammatory–decreasing" (6.74%), and "moderate pro-inflammatory–stable" (76.66%) groups, with cumulative T2DM incidence rates of 9.90%, 3.75%, and 8.19%, respectively. After adjusting for baseline age, educational background, marital status, residence, annual per capita family income, smoking, alcohol consumption, physical activity, sleep duration, chronic disease history, mean daily energy intake, body mass index, systolic/diastolic blood pressure, and follow-up duration, a two-level Cox proportional hazards regression model revealed that females in the "moderate pro-inflammatory–stable" and "low pro-inflammatory–increasing" groups had 3.22-fold (95%CI: 1.37–7.55) and 4.81-fold (95%CI: 1.96–11.83) higher risks of T2DM, respectively, than those in the "moderate pro-inflammatory–decreasing" group. No significant association was observed in males.

Conclusions Pro-inflammatory dietary patterns are prevalent among middle-aged and elderly Chinese residents. In females, both an upward trend in E-DII with age and sustained high pro-inflammatory levels are associated with an elevated risk of T2DM.