Serum homocysteine increases with age and is also considered a marker for low serum vitamin B(12) and folate. Furthermore, raised serum total homocysteine has been associated with atrophic changes in the brain. An association between serum vitamin B(12)/folate and cognitive impairment would be of considerable public health importance in view of the increasing numbers of elderly people. To systematically review published studies on the relationship between serum vitamin B(12), folate and total homocysteine and cognitive function in the elderly. A systematic review was undertaken of published evidence in English, examining the association between low serum vitamin B(12)/folate and raised total homocysteine with cognitive impairment (as indicated by low scores on neuropsychological testing) in subjects aged over 60 years. Sixteen electronic databases and cited articles were searched. Of 383 potential articles, six fulfilled the eligibility criteria: three case control and three cohort studies were identified. 'The Cochrane Non-Randomized Studies Methods Group' guidelines were used for assessment and extraction of data from these studies. All three case control studies found that serum total homocysteine was significantly higher in cases when compared with controls, and there was wide variation for both serum vitamin B(12) and folate in both groups of participants. The relationship of serum folate and vitamin B(12) status with cognitive impairment was heterogeneous and one case control study reported decreasing cognitive scores with increasing serum vitamin B(12). In the cohort studies, although serum total homocysteine could predict the rate of decline in neuropsychological testing, the overall odds ratio/relative risk (RR) of developing cognitive impairment in relation to levels of serum B(12) and serum folate were not significant. Although one study reported a significant RR of developing Alzheimer's disease when both serum folate and B(12) levels were low. One cohort study reported an increased prevalence of Alzheimer's type dementia in subjects who had normal serum vitamin B(12) at baseline. Serum total homocysteine is negatively correlated with neuropsychological tests scores. But the evidence does not support a correlation between serum vitamin B(12) or folate and cognitive impairment in people aged over 60 years. Hence, there is little evidence to justify treating cognitive impairment with vitamin B(12) or folate supplementation. This is consistent with the findings from recent systematic reviews of randomized double-blind trials, which have not found any evidence of potential benefit of vitamin supplementation. Further research is required in order to establish whether raised serum total homocysteine is a cause or consequence of disease.