Assessing the association between homocysteine and cognition: Reflections on Bradford Hill, meta-analyses, and causality

Abstract

Hyperhomocysteinemia is a recognized risk factor for cognitive decline and incident dementia in older adults. Two recent reports addressed the cumulative epidemiological evidence for this association but expressed conflicting opinions. Here, the evidence is reviewed in relation to Sir Austin Bradford Hill's criteria for assessing "causality," and the latest meta-analysis of the effects of homocysteine-lowering on cognitive function is critically examined. The meta-analysis included 11 trials, collectively assessing 22 000 individuals, that examined the effects of B vitamin supplements (folic acid, vitamin B12, vitamin B6) on global or domain-specific cognitive decline. It concluded that homocysteine-lowering with B vitamin supplements has no significant effect on cognitive function. However, careful examination of the trials in the meta-analysis indicates that no conclusion can be made regarding the effects of homocysteine-lowering on cognitive decline, since the trials typically did not include individuals who were experiencing such decline. Further definitive trials in older adults experiencing cognitive decline are still urgently needed.