ABSTRACT: To assess 7-year associations between magnesium intake and incident prediabetes and/or insulin resistance (IR), and progression from these states to type 2 diabetes.In 2,582 community-dwelling participants 26-81 years old at baseline, magnesium intake and risk of incident "metabolic impairment," defined as impaired fasting glucose (FG) (?5.6 to <7.0 mmol/L), impaired glucose tolerance (2-h postload glucose ?7.8 to <11.1 mmol/L), IR, or hyperinsulinemia (?90th percentile of homeostasis model assessment of IR or fasting insulin, respectively), was estimated among those with normal baseline status, and risk of incident diabetes was estimated among those with baseline metabolic impairment. In participants without incident diabetes, we examined magnesium intake in relation to 7-year changes in fasting and postload glucose and insulin, IR, and insulin sensitivity.After adjusting for age, sex, and energy intake, compared with those with the lowest magnesium intake, those with the highest intake had 37% lower risk of incident metabolic impairment (P trend = 0.02), and in those with baseline metabolic impairment, higher intake was associated with 32% lower risk of incident diabetes (P trend = 0.05). In the combined population, the risk in those with the highest intake was 53% (P trend = 0.0004) of those with the lowest intake. Adjusting for risk factors and dietary fiber attenuated associations in the baseline normal population but did not substantially affect associations in the metabolically impaired. Higher magnesium intake tended to associate with lower follow-up FG and IR, but not fasting insulin, postload values, or insulin sensitivity.Magnesium intake may be particularly beneficial in offsetting risk of developing diabetes among those at high risk. Magnesium's long-term associations with non-steady-state (dynamic) measures deserve further research.