OBJECTIVE: To assess whether the role of selenium on pre-diabetes is differential by age, given comorbidities and decreased β-cell function in older adults. RESEARCH DESIGN AND METHODS: We evaluated the cross-sectional association of blood selenium with the homeostatic model assessment for insulin resistance (HOMA-IR) and β-cell function (HOMA-β) in middle-aged (Aragon Workers Health Study [AWHS], N = 1186), and older (Seniors ENRICA [Study on Nutrition and Cardiovascular Risk in Spain]-2 [SEN-2], N = 915) diabetes-free adults. A subsample of participants from AWHS (N = 571) and SEN-2 (N = 603) had glucose and insulin repeated measurements for longitudinal analysis. We validated the cross-sectional dose-response associations in the 2011-2018 National Health and Nutrition Examination Survey (NHANES, N = 1317 middle age and N = 960 older) participants. Selenium was measured in whole blood with ICP-MS in AWHS, SEN-2 and NHANES. RESULTS: The cross-sectional geometric mean ratios (95% confidence intervals) per two-fold selenium increase were 1.09 (1.01, 1.19) for HOMA-IR and 1.15 (1.06, 1.24) for HOMA-β in AWHS
and 1.13 (0.98, 1.31) and 1.03 (0.90, 1.18), in SEN-2. The cross-sectional dose-response associations were consistent in NHANES, with mostly increasingly positive trends for both HOMA endpoints in younger adults and a plateau at levels >
~150 μg/L in older adults. The longitudinal dose-response consistently showed positive associations at high selenium dose for both HOMA endpoints in the younger, but not the older, study population. CONCLUSIONS: Increased blood selenium was associated with increased insulin resistance and β-cell function in middle-aged, but not in older individuals, especially for β-cell function. The results suggest that selenium-associated insulin resistance might induce compensatory increased β-cell function at younger ages, being this compensatory capacity decreased with aging.