B vitamins and the aging brain


  • Jacob Selhub,

    Corresponding author
    1. Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, USA
    Search for more papers by this author
  • Aron Troen,

    1. Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, USA
    Search for more papers by this author
  • Irwin H Rosenberg

    1. Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, USA
    Search for more papers by this author

J Selhub, 711 Washington Street, Boston, MA 02111, USA. E-mail: Jacob.selhub@tufts.edu, Phone: +1-617-556-3191, Fax 617-556-3166.


Deficiencies of the vitamins folate, B12, and B6 are associated with neurological and psychological dysfunction and with congenital defects. In the elderly, cognitive impairment and incident dementia may be related to the high prevalence of inadequate B vitamin status and to elevations of plasma homocysteine. Plausible mechanisms include homocysteine neurotoxicity, vasotoxicity, and impaired S-adenosylmethionine-dependent methylation reactions vital to central nervous system function. In light of this, it is imperative to find safe ways of improving vitamin B status in the elderly without exposing some individuals to undue risk.


The vitamins folic acid, B6, and B12 serve as coenzymes in one-carbon metabolism. In this metabolism, a carbon unit from serine or glycine reacts with tetrahydrofolate (THF) to form methylene-THF. This is either used for the synthesis of thymidylate, a DNA nucleotide, oxidized to formylated-THF for use in purine synthesis (DNA and RNA), or is irreversibly reduced to methyl-THF, which is used for the methylation of homocysteine to form methionine. The latter reaction is catalyzed by a methyltransferase that has vitamin B12 (methyl-B12) as the prosthetic group. Homocysteine is also methylated by betaine in a reaction not involving vitamin B12; however, this reaction is confined mostly to the liver.

A considerable proportion of methionine is converted to S-adenosylmethionine (SAM), a universal methyl donor for numerous vital methylation reactions. Methionine synthesis is a most crucial part of the pathway for the health of brain tissue.1 In the brain, SAM-dependent methylations are extensive, and the products of these reactions include neurotransmitters (catecholamines and indoleamines), phospholipids, and myelin.2–5 Methylation is also important in the regulation of presenilin-1 expression, γ-secretase activity, and amyloid-β (Aβ) levels6,7 as well as in the regulation of leucine carboxyl methyltransferase-1 and protein phosphatase 2A (PP2A) Bα subunit expression, which correlates with the extent of tau phosphorylation.8–10

The importance of these vitamins for the well-being and normal function of the brain derives from data showing neurological and psychological dysfunction in vitamin deficiency states and in cases of congenital defects of one-carbon metabolism. These deficiencies, however, are rare, and the question that arises is the extent to which the neuropsychological disorders affecting the elderly, such as cognitive impairment, dementia, and Alzheimer's disease (AD), can be ascribed to impaired one-carbon metabolism.

There is ample evidence of a high prevalence of inadequate vitamin status in the elderly and, hence, impaired one-carbon metabolism is high in this population. Indicators of impaired folate and B12 status, including elevated homocysteine (Hcy) and methylmalonic acid (MMA), are prevalent in the elderly.11–13 Atrophic gastritis and increased antacid medications interfere with the intestinal absorption of food-derived vitamin B12.14 Transport of folate and B12 across the blood-brain barrier, which includes receptor-mediated endocytic pathways, may be damaged with age.14

Neuropsychiatric disorders in the elderly are multifactorial, and the contribution of impaired one-carbon metabolism to these disorders has, for the most part, been assessed through epidemiological studies.15 Preliminary associations linking low vitamin status or intake with lower neurocognitive function in the elderly was first reported in 1983 by Goodwin et al.16 These authors showed healthy elderly subjects who had low blood concentrations or intakes of folate, vitamin B12, vitamin C, and riboflavin scored poorly on tests of memory and nonverbal abstract thinking. A 1992 report elaborated on the effect of nutritional factors on cognitive function in the elderly and drew associations with Hcy, whose circulating levels were regulated by the B vitamins folate, B6, and B12.17 With the emergence of the importance of elevated plasma Hcy as a risk factor for cardiovascular disease, considerable effort was directed toward the study of associations between mild elevations of Hcy in plasma and aspects of neuropsychological diseases that afflict the elderly.15


The relationship between Hcy and neuropsychological disease was studied in conjunction with vitamin status. For example, the study by Lindenbaum et al.,18 which described patients with neuropsychiatric disorders caused by cobalamin deficiency, included measurements of plasma Hcy (and MMA) concentrations as part of patients' characteristics and showed a decrease in the concentrations of these indicators when patients were administered cobalamin. In the ensuing studies, however, relations with neuropsychiatric disorders used Hcy as a variable, either after adjusting for B vitamin status to indicate independence of Hcy effect or in combination with B vitamin status to assess potential interdependence between the various variables. A detailed summary of these studies was recently published by AD Smith,15 who estimated there have been, as of 2008, as many as 77 cross-sectional studies on more than 34,000 subjects and 33 prospective studies on more than 12,000 subjects showing associations between cognitive deficit or dementia and Hcy and/or B vitamin status.

Close associations between intake and plasma levels of folate, B6, and B12 were demonstrated in the Framingham cohort.19 In a following study in this elderly population, high plasma Hcy concentrations were shown to be associated with an increased risk of extracranial carotid-artery stenosis.20 A 10-year follow-up has shown incidents of stroke to be twofold higher among those with the highest quartile compared to the lower three quartiles of plasma Hcy levels.21

The next study, conducted along with researchers from the Framingham Study and Boston University, included over 1,000 cohort subjects who had no dementia at the time their baseline Hcy levels were determined.22 Over a median follow-up period of 8 years, dementia developed in 111 subjects, including 83 diagnosed with AD. The multivariable-adjusted relative risk of dementia was 1.4 (95% confidence interval [CI] 1.1–1.9) for each increase of one standard deviation in the log-transformed Hcy value. The relative risk of AD was 1.8 (95% CI 1.3–2.5) per increase of one standard deviation at baseline. With a plasma Hcy level >14 µmol/L, the risk of AD nearly doubled.

Cross-sectional data from magnetic resonance imaging (MRI) of healthy participants of the Framingham Offspring Cohort have suggested participants with plasma total Hcy (tHcy) levels in the highest age-specific (−0.37%, P = 0.01) or sex-specific (−0.48%, P = 0.001) quartile had smaller total cerebral brain volumes than participants with lower tHcy levels.23 High tHcy was also associated with smaller frontal (−0.14%, P = 0.001) and temporal lobar (−0.10%, P = 0.04) volumes. When these MRI data were analyzed against tHcy levels that were determined about 8 years earlier, significant associations were found only with respect to silent brain infarct. No significant relation was found between tHcy and the prevalence of extensive white matter hyperintensity.23

Taken together, these data suggest plasma tHcy levels have a role in the changes of brain aging and dementia, affecting not only the incidence of clinically overt stroke and dementia but also the prevalence of subclinical brain MRI changes in an apparently healthy population.


In the above studies, the relationship between brain function and plasma Hcy was unaltered even after adjusting for respective risk factors as well as for folate, B6, and B12 levels. These adjustments, particularly with the B vitamins, while pointing to the independence of Hcy as a risk factor for brain dysfunction, nevertheless do not preclude independent associations with the B vitamins. For example, in the early studies by both Nilsson et al.24 and Joosten et al.,25 high Hcy in patients with neuropsychiatric dementia or AD was also associated with lower folate status. Similarly, in the Oxford Project to Investigate Memory and Ageing (OPIMA) population, higher serum tHcy levels were seen in patients with dementia of the Alzheimer's type and in patients with histologically confirmed AD than in controls; they were also accompanied by both low serum folate and low B12 levels.26 The study by McCaddon et al.27 also showed higher Hcy levels to be accompanied by low B12 (but not low folate) levels in AD patients.

In the Normative Aging Study at the Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, poorer spatial copying skills were significantly related to higher concentrations of Hcy (P = 0.0009).28 Poorer spatial copying skills were also associated with lower concentrations of vitamin B12 (P = 0.04) and folate (P = 0.003), but these associations were definitely weaker than those seen with Hcy, and there was no evidence that this association with Hcy was strengthened by combining with vitamin status. In the Framingham Offspring Cohort, associations of high Hcy with a battery of tests for cognitive performance were not influenced by inclusion of the status of folate, B12, and B6.29


One approach that is gaining interest is assessing the status of the B vitamins as risk factors for brain malfunction and the use of Hcy plasma levels in addition to MMA and holotranscobalamin, the active form of B12 in plasma, as functional indicators of vitamin status.

Cognitive function in the McArthur Study of Successful Aging cohort was found to be (cross-sectionally) related to both high Hcy and low folate.30 However, the risk of cognitive decline at the 7-year follow-up remained significant with folate but not with Hcy after adjustment for each other.

Holotranscobalamin was found to be significantly lower in the plasma of AD patients than in controls.31 In the Sacramento Area Latino Study on Aging cohort, the association between high Hcy concentrations and increased risk of dementia or cognitive impairment without dementia was significantly modified by vitamin B12 levels, with the highest B12 tertile being associated with the attenuation of the Hcy relationship.32

In a prospective study of 107 community-dwelling elderly subjects, decrease in brain volume, as assessed by MRI, was greater among those with lower vitamin B12 and holotranscobalamin levels and higher plasma Hcy and MMA levels at baseline.33 Linear regression analysis showed associations with vitamin B12 and holotranscobalamin remained significant after adjustment for age, sex, creatinine, education, initial brain volume, cognitive test scores, systolic blood pressure, ApoE4 status, Hcy, and folate. Using the upper tertile (for the vitamins) or the lower tertile (for the metabolites) as a reference in logistic regression analysis and adjusting for the above covariates, vitamin B12 in the bottom tertile (≤308 pmol/L) was associated with increased rate of brain volume loss (odds ratio [OR] 6.17; 95% CI 1.25–30.47). The association was similar for low levels of holotranscobalamin or for low transcobalamin saturation.

In the Rotterdam Scan Study, poorer vitamin B12 status was significantly associated with greater severity of white-matter lesions, in particular periventricular white-matter lesions.34 Adjustment for common vascular risk factors did not alter the associations, while adjustment for tHcy and folate modestly weakened the associations. No association was observed between vitamin B12 status with presence of brain infarcts and baseline cognition or cognitive decline during follow-up. These results suggest vitamin B12 status in the normal range is associated with severity of white-matter lesions, especially periventricular lesions. Given the lack of correlation with cerebral infarcts, it is hypothesized that this association is explained by effects on myelin integrity in the brain rather than through vascular mechanisms.34

A combination of low plasma B12 levels (≤148 pmol/L) and/or high plasma MMA levels (≥210 nmol/L) – the maximum of the reference range for serum vitamin B12-replete participants with normal creatinine – was used to describe the association between vitamin B12 status and cognitive function in senior participants in the 1999–2002 US National Health and Nutrition Examination Survey.35 After controlling for a number of factors, low versus normal vitamin B12 status was associated with anemia (OR 2.7; 95% CI 1.7–4.2), macrocytosis (OR 1.8; 95% CI 1.01–3.3), and cognitive impairment (OR 2.5; 95% CI 1.6–3.8).

The intriguing observation in this study, however, was the presence of an adverse interaction between B12 status and folate status. In the group with a low vitamin B12 status, serum folate ≥59 nmol/L (80th percentile) was associated with an almost twofold higher prevalence of anemia and cognitive impairment than in those with low B12 status but normal plasma folate levels (≤59 nmol/L). In the normal vitamin B12 group, however, higher folate, i.e., >59 nmol/L, was associated with a lower prevalence of cognitive impairment than in the group with normal B12 status and normal folate.35

This interaction between folate and B12 status was further explored by determining the relationship between plasma folate and plasma B12 levels with respect to the concentrations of Hcy and MMA, the two functional indicators of vitamin B12 status.36 Using the same US National Health and Nutrition Examination Survey population, it was found that both plasma Hcy and MMA increase along with increases in plasma folate in people with low plasma B12 levels (≤148 pmol/L). At concentrations ≥59 nmol/L, Hcy concentrations are 30% higher and MMA concentrations are almost 100% higher than at normal plasma folate levels.

When plasma B12 concentrations are normal (>148 pmol/L), an increase in plasma folate is associated with decreased levels of Hcy and no significant change in MMA levels.36

Taken together, these data strongly imply that, with vitamin B12 deficiency, high plasma folate is associated with the exacerbation of both the biochemical and clinical status of B12.


The FACIT trial was a randomized, double-blind, placebo-controlled study conducted in the Netherlands and involving over 800 elderly persons who received either 800 µg/day folic acid or a placebo control for 3 years.37 After 3 years, plasma folate was substantially higher and tHcy was lower in participants who received folic acid compared to placebo. The 3-year change in memory (difference in Z scores 0.132, 95% CI 0.032–0.233), information processing speed (0.087, 95% CI 0.016–0.158), and sensorimotor speed (0.064, 95% CI −0.001–0.129) were significantly better in the folic acid group than in the placebo group.

The findings of a number of large-scale Hcy-lowering placebo-controlled trials for the reduction of vascular events were, for the most part, negative.38–42 However, subanalysis of two of these trials revealed vitamin supplementation may be effective for lowering the incidence of stroke.

  • 1) The primary analysis of the Vitamin Intervention for Stroke Prevention trial revealed no efficacy of combined vitamin therapy for recurrent vascular events in patients with nondisabling stroke.41 A secondary efficacy analysis excluded patients with low and very high B12 levels at baseline (≤250 and ≥637 pmol/L) to exclude those likely to have B12 malabsorption or to be taking B12 supplements outside the study as well as patients with significant renal impairment.43 For the combined endpoint of ischemic stroke, coronary disease, or death, there was a 21% reduction in the risk of events in the high-dose group compared with the low-dose group (unadjusted P = 0.049; adjusted for age, sex, blood pressure, smoking, and B12 level, P = 0.056). In Kaplan-Meier survival analyses, patients with a baseline B12 level at a median or higher level who were randomized to receive high-dose vitamin therapy had the best overall outcome, and those with a B12 level below the median level and who were assigned to low-dose vitamin therapy had the worst outcome (P = 0.02 for combined stroke, death, and coronary events; P = 0.03 for stroke and coronary events).
  • 2) In the Heart Outcomes Prevention Evaluation 2 trial, which involved more than 5,500 patients with a history of vascular disease or diabetes, supplementation with B vitamins compared to placebo control was without effect on pooled primary outcomes.38 A recent secondary analysis44 showed that the incidence rate of stroke was 0.88 per 100 person-years in the vitamin therapy group and 1.15 per 100 person-years in the placebo group (hazard ratio 0.75; 95% CI 0.59–0.97). Vitamin therapy also reduced the risk of nonfatal stroke (hazard ratio 0.72; 95% CI 0.54–0.95) but it did not impact neurological deficit at 24 hours (P = 0.45) or functional dependence at discharge or at 7 days (OR 0.95; 95% CI 0.57–1.56).

The response of stroke incidence to vitamin supplementation is also consistent with recent analyses on the outcome of fortification of staple products with folic acid.45 This fortification, which was implemented in the United States and Canada around 1998–1999, is associated with the doubling of plasma folate levels, an approximately 15% decrease in mean plasma Hcy, and a significant reduction in the incidence of neural tube defects (NTD). Analysis of stroke-related mortality revealed the ongoing decline in stroke mortality observed in the United States between 1990 and 1997 accelerated in 1998–2002 in nearly all population strata, with an overall change from −0.3% (95% CI −0.7–0.08) to −2.9% (95% CI −3.5–−2.3) per year (P = 0.0005). The decline in stroke mortality in Canada averaged −1.0% (95% CI −1.4–−0.6) per year from 1990 to 1997 and accelerated to −5.4% (95% CI −6.0–−4.7) per year in 1998–2002 (P ≤ 0.0001). In contrast, the decline in stroke mortality in England and Wales, where no fortification program was in place, did not change significantly between 1990 and 2002.

This improvement in stroke mortality observed after folic acid fortification in the United States and Canada but not in England and Wales is consistent with the hypothesis that folic acid fortification helps reduce deaths from stroke.


In a recent review, AD Smith15 listed several hypotheses that have been offered to explain the association between elevated Hcy and brain dysfunction, such as compromise of cerebrovascular circulation by elevated Hcy, leading to ischemia and reperfusion; neurotoxicity of Hcy, particularly after conversion to homocysteic acid; inhibition of the methylation reaction by Hcy through its conversion to S-adenosyl-Hcy; oxidative damage caused by Hcy; and others.

These hypotheses, although reasonable, rely to a large extent on animal and cell culture models, which do not truly represent the “physiology” of the small elevation of plasma Hcy levels epidemiologically associated with cardiovascular diseases and various brain dysfunctions. There is no supportive evidence to suggest the concentrations of Hcy in brain or cerebrospinal fluid reach levels considered neurotoxic, based on in vitro data.46

The alternative approach to study relations with B vitamins is likely to produce better understanding of the relationship between one-carbon metabolism and brain aging. Below are a few examples of the utility of this approach:

  • 1) PP2A, a major brain tau phosphate, consists of a dimeric core enzyme comprising the structural A and catalytic C subunits and a regulatory B(α) subunit. Assembly of the AC subunits with the B(α) subunit is controlled by methylation on Leu-309 by SAM-dependent leucine carboxyl methyltransferase-1 (LCMT-1)6. Disassembly of the PP2A holoenzyme is promoted by demethylation, a reaction catalyzed by protein phosphatase methyltransferase 1. Studies by Sontag et al.6 have shown that folate deprivation in N2a cells induced demethylation of the C subunit, a decrease in the B(α) subunit decrease in LCMT-1, and a substantial increase in phosphorylated tau. Overexpression of LCMT-1 or B(α) protects N2a cells against folate deficiency-induced tau phosphorylation and cell death. Mice maintained for 2 months on a folate-deficient diet have brain-region-specific alterations in metabolites of the methylation pathway. These are associated with downregulation of LCMT-1, methylated PP2A, and B(α) expression and enhanced tau phosphorylation in susceptible brain regions.6
  • 2) Presenilin 1 is a γ-secretase that mediates the formation of Aβ from amyloid precursor protein.
  • 3) Fuso et al.9,10 recently investigated the effect of combined B vitamin deficiency in transgenic mice (TgCRND8) that express Aβ and in a normal mice strain (127Sv). These studies showed vitamin B deficiency is associated with presenilin 1 and BACE (β-secretase) upregulation and Aβ deposition in the transgenic mice.
  • 4) The relation between B vitamins with and without excess methionine on brain function and metabolism has been studied. Young ApoE-deficient mice were fed one of four diets with differing methionine and B-vitamin content.47 After a period of 8 weeks the animals were subjected to psychomotor tests, the Morris water maze test of spatial memory and learning, and measurement of home-cage activity. B-vitamin (folate, B12, and B6) deficiency induced homocysteinemia and selectively impaired Morris water maze performance without affecting other behavioral measures. The cognitive deficits occurred in the absence of overt histologic neurodegeneration but in association with moderate impairments of brain methylation potential. Diets that yielded cognitive deficits were different from those that exacerbated aortic pathology. These findings are inconsistent with a single mechanism linking homocysteinemia to neurological dysfunctions mediated by Hcy vasotoxicity. Instead, they indicate that different “types” of homocysteinemia, or different impairments of nutritional metabolism affecting Hcy levels, may lead to different end-organ dysfunctions and/or diseases.
  • 5) In a second study, C57BL6/J mice fed a B-vitamin-deficient (folate, B12, and B6) diet for 10 weeks induced hyperhomocysteinemia, significantly impaired spatial learning and memory, and caused a significant rarefaction of hippocampal microvasculature without concomitant gliosis and neurodegeneration.48 Total hippocampal capillary length was inversely correlated with Morris water maze escape latencies (r = −0.757, P < 0.001) and with plasma total Hcy (r = −0.631, P = 0.007). Feeding mice a methionine-rich diet produced similar but less pronounced effects. The findings suggest cerebral microvascular rarefaction can cause cognitive dysfunction in the absence of or preceding neurodegeneration. Similar microvascular changes may mediate the association of hyperhomocysteinemia with human age-related cognitive decline.
  • 6) A third study focused on the role of folate deficiency (rather than deficiency of all B vitamins) in cognitive dysfunction.49 Young Sprague Dawley rats were fed folate-deficient diets (0 mg folic acid/kg) with or without supplemental L-methionine for 10 weeks, followed by cognitive testing and tissue collection for hematological and biochemical analysis. Folate-deficient rats with normal methionine showed impaired spatial memory and learning; however, this impairment was prevented when the folate-deficient diet was supplemented with methionine. Under conditions of folate deficiency, the brain membrane content of the methylated phospholipid phosphatidylcholine was significantly depleted, which was reversed with supplemental methionine. In contrast, neither elevated plasma Hcy nor brain S-adenosylmethionine and S-adenosylhomocysteine concentrations predicted cognitive impairment and its prevention by methionine. The correspondence of cognitive outcomes to changes in brain membrane phosphatidylcholine content suggests altered phosphatidylcholine and possibly choline metabolism might contribute to the manifestation of folate deficiency-related cognitive dysfunction.


Studies in the last 2–3 decades produced convincing evidence of the involvement of B vitamins in brain aging. Whether the data on Hcy relationship with brain dysfunction are consistent with the notion of Hcy-related causality has not been proven satisfactorily. Aging is an important independent determinant of plasma Hcy levels, which may be independent of vitamin status.11 Age adjustment may not be sufficient to discriminate between individuals who are of the same age and sex (and vitamin status) but who differ in their Hcy levels. It is quite possible the same factors that cause the brain dysfunction are also responsible for the rise in Hcy.

B vitamin metabolism is a defined entity that can be measured at all levels, particularly those associated with biochemical or molecular biological changes. It is this area that needs to be further explored to study changes in one-carbon metabolism and brain aging.


Any opinions, findings, conclusion, or recommendations expressed in this publication are those of authors and do not necessarily reflect the view of the USDA.

Funding.  This work was supported by the US Department of Agriculture (USDA) cooperative agreement nos. 58-1950-7-707 and 51520-008-04S.

Declaration of interest.  The author has no relevant interests to declare.