: We have examined whether metabolic syndrome is associated with intermediate risk of impaired cognition between people with and without diabetes. Men aged 45 to 59 years were identified from Caerphilly in South Wales, United Kingdom. Participation rate was 89% (41% of the original cohort) and 2,512 men were examined in phase one from July 1979 until September 1983. Follow-up examinations occurred at four intervals until 2004 when 1,225 men participated. Participants were categorized on the basis of their exposure to metabolic syndrome not diabetes (MSND) and diabetes (with or without metabolic syndrome) at each of the first three phases. Neuropsychological outcomes and clinical diagnosis of cognitive impairment not dementia (CIND) and dementia were assessed at phase five. The prevalence of MSND increased from 1% to 5% and for diabetes from 3% to 9% between phase one and phase three. 15% of participants had CIND and 8% dementia. People with diabetes, but not those with MSND, at phases one, two, or three had poorer cognition at phase five (adjusted ? coefficient AH4 -4.3 95% CI -7.9, -0.7; phase two: -2.5 95% CI -4.7, -0.3; phase three: -2.3 95% CI -4.2, -0.5). The adjusted odds ratio (phase one) for diabetes and CIND was 4.0 (95% CI 1.4, 11.5) and dementia 0.61 (95% CI 0.07, 5.37). After adjustment, higher systolic blood pressure was the only component of the metabolic syndrome associated with worse cognitive outcomes. Diabetes in mid-life, but not MSND, is associated with impaired cognition and increased odds of CIND in later life.