PURPOSE: To distinguish between contributions to dementia made by homocysteine, folate, B12 and antioxidant micronutrients. METHODS: This is a follow-up study of a sample reported in 2002. Homocysteine was measured at baseline in 201 individuals born in 1921 and without dementia at age 77 years and followed up to age 88 years. Baseline macro- and micronutrient status was estimated from BMI, the MONICA food frequency questionnaire, plasma folate, B12 and, in a subgroup (N = 173), plasma antioxidant micronutrients. Time to dementia onset during follow-up was compared between participants grouped by homocysteine concentration using Cox regression. Model 1 adjusted for age, sex, childhood IQ, education, socioeconomic deprivation, presence of heart disease, hypertension, plasma folate and B12. In model 2 plasma, antioxidants were added to these covariables. RESULTS: During a mean follow-up of about 5 years, there were 39 incident dementia cases among 201 participants. In model 1, being in the highest homocysteine group (>14 mumol/L) was associated with a 234 % increased risk (HR 3.34, 95 % CI 1.16-9.57) of any dementia. After inclusion of plasma antioxidants in model 2, there were 32 incident dementia cases from a subsample (N = 173). Homocysteine >14 mumol was associated with a 272 % increased dementia risk (HR = 3.72, 95 % CI 1.06-13.08). CONCLUSIONS: High homocysteine increases the risk of dementia. The association between tHcy and dementia is independent of plasma folate, B12 and antioxidant micronutrient status.