B vitamins and cognitive function: do we need more and larger trials?1,2

Abstract

The failure to directly translate promising observational associations between disease and intakes or blood concentrations of vitamins into effective chemoprevention or clinical treatment modalities is not new. The first major disappointment came a decade ago after studies of -carotene and lung cancer (1). Contrary to expectations, the large Alpha-Tocopherol Beta-Carotene Cancer Prevention Study and Carotene and Retinol Efficacy Trial showed that this provitamin may increase lung cancer rates and all-cause mortality. Also, a recent meta-analysis of cardiovascular disease trials showed that high doses of vitamin E could increase all-cause mortality (2). There is a clear association between total homocysteine (tHcy) concentrations and cardiovascular disease, and a series of large randomized clinical trials in patients with coronary heart disease or stroke are ongoing and a few studies have been completed. In these trials, the effects of treatment with folic acid in combination with vitamin B-6 and vitamin B-12 were compared with those of placebo. A large study of stroke patients in the United States showed no beneficial effect of B vitamins (3). In patients undergoing coronary angioplasty, one trial showed a clear reduction in restenosis (4), whereas another trial showed an increase in restenosis (5) with B vitamin therapy. On the positive side, solid evidence from randomized trials and intervention studies has shown that folic acid prevents neural tube defects (6). This knowledge provides the basis for recommendations that women take folic acid if they plan to become pregnant and early in pregnancy and is why food-fortification programs have been implemented in several countries in North and South America. Observational studies, including a recent investigation in 2871 subjects (7), have consistently reported that an elevated concentration of tHcy in serum or plasma is a risk factor for dementia and impaired cognitive function (8). This association has been explained by neurotoxic effects of homocysteine or the ability of elevated homocysteine to cause vascular lesions. Alternatively, the effects are mediated by the impaired function of the B vitamins involved in homocysteine metabolism, including vitamin B-12, folate, and vitamin B-6. Impaired vitamin B-12 status is known to be prevalent among the elderly, and vitamin B-12 deficiency may cause severe myelopathy and also prominent mental symptoms and memory loss (9). Poor folate status has been associated with depression and dementia in the elderly, and folate metabolism is linked to a variety of neurochemical processes (10). Vitamin B-6 status declines with age, and low blood concentrations of vitamin B-6 have been associated with impaired cognitive function and Alzheimer disease. Such associations could be explained by the involvement of vitamin B-6 in the synthesis of several neurotransmitters (11). However, recent Cochrane Library reviews of randomized trials with folic acid, vitamin B-12, and vitamin B-6 provide no evidence of any improvement of cognition or dementia (11–13). The Cochrane reviewers, however, concluded that the current evidence was based on small trials and that the research question was sufficiently important to motivate more randomized trials. In an elegant study from Sweden in this issue of the Journal, Lewerin et al (14) failed to show any effect of a combination treatment with folic acid, vitamin B-6, and vitamin B-12 on cognitive function or movement performance tests among elderly community-dwelling men and women. They conducted a double-blind, placebo-controlled randomized trial with 126 subjects allocated to the intervention and 69 subjects allocated to placebo. Their study also had an observational component, which showed strong relations of plasma tHcy and serum methylmalonic acid (MMA) with movement and cognitive performance tests. The Swedish authors used an impressive panel of tests that assessed movement and cognitive performance in elderly subjects. The authors noted different associations between tHcy and MMA and different aspects of movement and cognitive performance. However, multiple comparisons of independent variables in a limited number of subjects may produce chance findings. Therefore, one should be cautious when interpreting the data and associations in terms of specific biological mechanisms. The authors discuss their results relative to those of the now retired professor Ranjiit Kumar Chandra (15). Their failure to reproduce Chandra’s results is not surprising in light of the recent controversy surrounding his research, which has received considerable attention both in the media (16) and in prominent medical journals (17). When should we stop pursuing a treatment option with more randomized trials? Neither previous meta-analyses nor the single trial published in this issue of the Journal found evidence that B