Featured ArticleSerum homocysteine and dementia: Meta-analysis of eight cohort studies including 8669 participants

BackgroundProspective cohort studies have not been consistent in showing an association between serum homocysteine and dementia.ObjectiveTo conduct a meta-analysis of cohort studies that examined the relationship between serum homocysteine and dementia, and to estimate the change in risk of dementia for a unit change in serum homocysteine.MethodsThe data from eight cohort studies (involving 8669 participants; range of mean ages, 47-81 years; median duration of study, 5 years) of serum homocysteine on the incidence of dementia were combined and the odds ratio of dementia per 5 μmol/L increase in serum homocysteine was determined.ResultsThere was a statistically significant association between serum homocysteine and the incidence of dementia: the odds ratio for a 5 μmol/L increase in serum homocysteine was 1.35 (95% confidence interval, 1.02-1.79) or 1.50 (1.13-2.00) adjusted for regression dilution bias. The odds ratio for a 3 μmol/L decrease in serum homocysteine (the average reduction expected using folic acid and B12) was 0.78 (0.66-0.93).ConclusionThe meta-analysis of epidemiological cohort studies shows a positive association between serum homocysteine and dementia. Although the results do not provide evidence of cause and effect, they do provide an estimate of the expected effect if the relationship were causal; an approximate 20% reduction in risk of dementia from treatment with folic acid and B12.