The Journal of Nutrition, Health & Aging©

Volume 16, Number 8, 2012

CONVERSION FROM MILD COGNITIVE IMPAIRMENT TO DEMENTIA:

INFLUENCE OF FOLIC ACID AND VITAMIN B12 USE IN THE VITA COHORT
I. Blasko1, M. HInterBerger2, g. keMMler1, s. JungwIrtH2, w. kraMpla3, t. leItHa4,
k. HeInz tragl2, p. FIscHer2,5
1. Department of psychiatry and psychotherapy, Division of general and social psychiatry, Innsbruck Medical university, austria; 2 . ludwig Boltzmann society, l. Boltzmann Institute
of aging research, Vienna, austria; 3. Department of radiology, Danube Hospital, Vienna, austria; 4. Department of nuclear Medicine, Danube Hospital, Vienna, austria;
5. Department of psychiatry, Medical research society, Danube Hospital, Vienna, austria. corresponding author: Imrich Blasko, MD, Department of psychiatry and psychotherapy,
Division of general and social psychiatry, Innsbruck Medical university, anichstrasse 35, 6020 Innsbruck, austria, tel.: +43 512 504 81608, Fax: +43 512 504 23628,
e-mail: imrich.blasko@i-med.ac.at

Abstract: Objective: Increased serum homocysteine and low folate levels are associated with a higher rate of
conversion to dementia. this study examined the influence of vitamin B12/folic acid intake on the conversion
from mild cognitive impairment (McI) to dementia. Participants: a community dwelling cohort of older adults
(n=81) from the Vienna transdanube aging study with McI. Design: prospective study with a retrospective
evaluation of vitamin intake. Measurements: laboratory measurements, brain magnetic resonance imaging, and
cognitive functioning were assessed at baseline and at five-year follow-up. Results: the self-reported combined
use of folic acid and vitamin B12 for more than one year was associated with a lower conversion rate to
dementia. serum levels of homocysteine and vitamin B12 as measured at baseline or at five years were not
associated with conversion. Higher folate levels at baseline in females predicted a lower conversion rate to
dementia. the assessment of brain morphological parameters by magnetic resonance imaging revealed higher
serum folate at baseline, predicting lower medial temporal lobe atrophy and higher levels of homocysteine at
baseline, predicting moderate/severe global brain atrophy at five years. users of vitamin B12 or folate,
independent of time and pattern of use, had lower grades of periventricular hyperintensities and lower grades of
deep white matter lesions as compared to non-users. Conclusions: these results from a middle european study
support observations on the protective ability of folate in McI patients with respect to conversion to dementia;
they also point to a participation of homocysteine metabolism on processes associated with brain atrophy.

Key words: Folic acid, vitamin B12, homocysteine, mild cognitive impairment, dementia.

Introduction subjects, have shown associations between cognitive deficit or

current therapy of age-associated dementia includes the use
of acetylcholine esterase inhibitors and n-methyl-d-aspartate
(nMDa) receptor antagonists memantine. In the past decade no
further substances were added for clinical use, although a
number of treatment strategies were tested in clinical studies.
new treatment options and improved selection criteria of
patients are needed. For these reasons, well known substances
which influence important steps in neurodegeneration, such as
oxidative stress and inflammation, have been tested for their
efficacy.
the aging brain shows a slow progressive atrophy which can
be accelerated by the conversion from mild cognitive
impairment (McI) to dementia. several efforts were made to
identify possible contributing factors to this decline. a higher
serum level of total homocysteine was shown to be one of the
contributing factors. Homocysteine, a sulfur-containing amino
acid derived from methionine, is a risk factor for developing
vascular disease, brain atrophy, cognitive impairment and
alzheimer’s Disease (aD) (1). In the elderly, cognitive
impairment and incident dementia may be related to the high
prevalence of an inadequate B vitamin status and an elevation
in plasma homocysteine levels (2, 3).
as many as 77 cross-sectional studies on more than 34.000
subjects, as well as 33 prospective studies on more than 12.000
dementia and homocysteine and/or B vitamin status (4-6).
these studies suggest that elevated concentrations of total
homocysteine and/or low-normal concentrations of B vitamins
(folate, vitamin B6, and vitamin B12) are associated with an
increased risk of brain atrophy and the development of
cognitive impairment and dementia in the elderly (7).
surprisingly, a recent cochrane Database review has
concluded that there is no consistent evidence that folate, with
or without vitamin B12, has a beneficial effect on cognitive
function of unselected healthy or cognitively impaired older
people (8).
In a randomized, double-blind controlled trial of
homocysteine-lowering B vitamins, McI patients using B
vitamins showed slowing of the accelerated rate of brain
atrophy (9). the treatment response was related to baseline
homocysteine levels, with more severe atrophy occurring in
persons with higher homocysteine levels. recently, the same
trial (VItacog) has reported positive cognitive outcomes in
which, B vitamins appear to slow cognitive and clinical decline
in people with McI, in particular in those with elevated
homocysteine (10).
we analyzed a sub-group of the middle european cohort of
the Vienna transdanube aging study (VIta). these
individuals were already showing mild cognitive impairment at
baseline. the aim of our study was to determine whether the

687
Received July 6, 2011
Accepted for publication October 11, 2011
The Journal of Nutrition, Health & Aging©
Volume 16, Number 8, 2012

INFLUENCE OF FOLIC ACID AND VITAMIN B12 USE IN THE VITA COHORT

use of folic acid and vitamin B12 influenced the metabolism of
homocysteine and the rate of conversion from McI to dementia
five years later. we also tested whether the supplementation of
folic acid and vitamin B12 might influence the rate of brain
atrophy and white matter degeneration, as measured by
magnetic resonance imaging (MrI). we selected 81 persons
with McI and followed their conversion to dementia five years
later with respect to their vitamin B status, plasma
homocysteine levels, as well as their subjective information on
vitamins intake. It had previously been shown in the same
cohort that the conversion to dementia after two and a half
years was associated with increased levels of homocysteine
(11).
Methods
Study population, recruitment and rate of conversion
these analyses originate from the Vienna transdanube
aging study (VIta), a population-based study of 75 year-old
residents of Vienna, austria. the recruiting procedures are
described in detail elsewhere (12). Briefly, VIta began in
2000-2002 with a baseline examination of 606 persons (40 % of
1505 contacted) born between May 1925 and april 1926.
medication to the appointment. In those persons who
supplemented vitamins, the mean daily dose of folic acid was
1.8±2.1 mg/day (±sD), range 0.1-5 mg and in case of vitamin
B12 it was 146.3±258.8 μg/day, range 1-1000 μg.
corroborating information on drug exposure was obtained
through medical records, as well as through information of
care-givers or relatives. when a subject was determined to have
not previously consumed folic acid and vitamin B12, yet their
serum levels were markedly increased (folate >20 ng/ml);
vitamin B12 > 1200 pg/ml), they were categorised as having
consumed folic acid and vitamin B12 for a period of one month
only.
Figure 1
Flow chart of studied population, which represents selected
persons with mild cognitive impairment at baseline. For
classifications see Methods; aD - alzheimer´s disease, FtlD -
Frontotemporal lobe dementia, DlB - Dementia with lewy
bodies
baseline mild cognitive impairment
n=141 excluded n=58
refused to participate n=29
follow-up 5 years not reachable n= 1

participants were re-examined every 2.5 years from baseline.

partial investigation n= 2
deceased n=26
Data from the baseline and five-year follow-up is included in
the following analyses. the study was approved by the local
total N of investigated
not classifiable
persons
ethics committee of the Vienna Medical university. subjects
n=2

were informed about the specific procedures and gave their

written informed consent.
all 141 persons with mild cognitive impairment (McI) at converters to dementia
baseline from the VIta cohort were included in this study. at cognitive unimpaired mild cognitive n = 49

five-year follow-up, 83 persons were re-investigated. as

n = 12 impairment
n = 20 AD n=40
presented in the flow chart (Fig. 1), 29 probands declined to vascular dementia
FTLD
n= 6
n= 2
participate. the reasons for this included lack of interest, DLB n= 1
somatic morbidity, obvious affective difficulties, and lack of
51 of 81 individuals with McI at baseline did not report the
time. two persons underwent a partial investigation (telephone
use of either folic acid or vitamin B12 during the observation
interview) without blood sampling (13). 20 individuals (25%)
period of five years. 10 individuals who had used a
were diagnosed as having stable McI, 12 (15%) had improved
combination of folic acid and vitamin B12 for more than one
to cognitive health, 49 (60%) had converted to dementia and 2
year were assigned to the group of “combination users”. 14
persons could not be classified. the current analysis is
individuals who used either folic acid or vitamin B12 for more
comprised of 81 persons.
than one year, 4 individuals who used either folic acid or
vitamin B12 for less than one year, and 2 individuals who used
Clinical evaluation and vitamin use
a combination of both for a period of less than one year were
the VIta study included neurological and psychiatric
assigned to the group of “inconsistent users”.
examination and neuropsychological testing which was
of 49 individuals with dementia, 2 (4.1%) belonged to the
administered in face-to-face interviews by experienced
group of “combination users”, 15 (30.6%) were “inconsistent
specialist raters, lasting approximately 9 hours per subject.
users” and 32 (65.3%) had not consumed folic acid and/or
Medical history was obtained from the tested subjects, their
vitamin B12 in the past five years. In the McI/cognitively
medical documentation and, if necessary, from relatives. the
unimpaired group, 8 persons (25%) belonged to the group of
serum samples collected for homocysteine, folate, and vitamin
“combination users”, 5 persons (15.6%) were “inconsistent
B12 were drawn following overnight fasting.
users” and 19 persons (59.4%) had not used either substance.
neuropsychological testing and an MrI were performed on the
Variables used in this study include serum levels of
same day.
homocysteine, folate, vitamin B12, creatinine, years of
the information on folic acid and vitamin B12
education, as well as the presence of at least one apoe ε4
supplementation was obtained from study participants during a
allele.
clinical interview. the participants were asked to bring all
688
The Journal of Nutrition, Health & Aging©
Volume 16, Number 8, 2012
JNHA: CLINICAL NEUROSCIENCES
Classification criteria and psychometric tests used to
determine the cognitive status
In accordance with the current guidelines (14), McI was
assigned to non-demented subjects who scored below the 1.5
standard deviation of the age-normative values of the VIta-
population in any of the 6 psychometric tests. these
procedures, as well as the corresponding cut-off values, are
described elsewhere (13,15). Briefly, for neuropsychological
testing the consortium to establish a registry for alzheimer's
Disease battery (ceraD, (16)), in its standardized german
translation (17), and a test of executive functioning (trail
Making test B, (18)) were used. the ceraD contains five
cognitive tests: (1) verbal memory (word list recall), (2) non-
verbal memory (constructional recall), (3) verbal fluency, (4)
naming (Boston naming test), and (5) constructional praxis.
probable or possible aD was diagnosed according to
national Institute of neurological and communicative
Disorders and stroke-aD and related Disorders association
(nIncDs-aDrDa) criteria (19). Frontotemporal lobe
dementia (FtlD) was diagnosed according to the Manchester-
lund criteria (20). Dementia with lewy bodies (DlB) was
diagnosed according to the revised criteria of consortium for
DlB (21).
at five-year follow up, 49 probands had developed
dementia; 22 had possible dementia, and 18 had probable aD,
6 had probable or possible vascular dementia, 2 had
frontotemporal dementia and one person dementia with levy
bodies. 8 of 40 probands with probable or possible aD also
fulfilled the nInDs-aIren criteria (22) for possible vascular
dementia, suggesting the presence of a mixed vascular plus
neurodegenerative type of disorder.
Magnetic resonance imaging (MRI) scans
at five-year follow-up, 55 of 81 subjects (68%) had MrI
scans. the MrI investigation was performed on a siemens
Impact expert 1.0-t device (siemens Impact expert; siemens
Medical systems, Inc., south Iselin, nJ) with a circular
polarized head coil. the sequences obtained and protocol
performed is described in detail elsewhere (23). the images
were assessed by a radiologist (w.k.) who had more than ten
years of experience in the field of diagnostic neuroradiology.
at the time of the investigation the radiologist had no
information about the probands. to determine whether the
radiologist’s evaluation was investigator-independent, the first
100 probands from the entire VIta-cohort at baseline were
assessed by two independent radiologists; a high degree of
congruency was registered (p<0.001).
among other factors, the presence of hyperintensities in the
white matter on the t2-weighted sequence was evaluated
according to a semi-quantitative method (24). a distinction was
made between periventricular changes in the white matter
(periventricular hyperintensities; pVH) and lesions
predominantly in the sub-cortical white matter (sub-cortical
white matter hyperintensities; wMH). this scale provides 2
different scores rated on a 4-point scale according to the
689
following criteria: pVH = 0 (absence), 1 (caps or pencil-thin
lining), 2 (smooth halo), or 3 (irregular pVH extending into the
deep white matter), and wMH = 0 (absence), 1 (punctate foci),
2 (beginning confluence of foci), or 3 (large confluent areas).
signs of general and circumscribed brain atrophy were also
assessed. on the coronary t1-weighted gradient echo sequence
(Mprage image), atrophy of the mesial temporal lobe
(hippocampus and hippocampal gyrus) was assessed according
to visual criteria. atrophy of the medical temporal lobe (Mta)
was graded into one of five stages (none, mild, moderate,
severe, and very severe atrophy) according to the criteria of
scheltens et al. (25).
global brain atrophy was assessed by four-point visual scale
and was graded by 0 (as none), by 1 (as mild, when brain sulci
were not in contact with each other but no wider than the
adjacent width of the gray matter) by 2 (as moderate, when
brain sulci were maximally as broad as the adjacent brain gyri),
by 3 (as severe, when brain sulci were broader than the
adjacent brain gyri). this global brain atrophy scale was
created in alignment to previously published data (26-28). Its
values at five years correlated negatively with corresponding
cella media index (cMI, rho=-0.371, p<0.01) and positively
with corresponding Mta (rho=0.752, p<0.001). the cMI was
calculated as maximal head diameter divided by maximal
diameter of the cella media in transversal slides (29). the
smaller the cMI, the greater the atrophy and the smaller the
remaining brain volume.
Biochemical and genetic analysis
serum folate (ng/ml) and vitamin B12 (pg/ml) were
measured by radioimmunoassay or chemiluminescence assay
(Dpc, los angeles, ca, usa). Homocysteine (μmol/l) was
detected in IMX-assay (abbot, wiesbaden, germany). clinical
chemistry procedures are described in detail elsewhere (30).
the presence of particular alleles of apolipoprotein e (apoe)
genotypes was determined with the roche light cycler apoe
assay kit (Basel, switzerland). participants were divided into
two groups, one comprised of persons with an apoe-ε4 allele
(ε2/ε4, ε3/ε4, or ε4/ε4 genotype) and another comprised of
those without an apoe-ε4 allele. Data on the apolipoprotein e
genotype was available for 79 of 81 subjects (97.5%).
Data analysis
statistical analysis was performed using the spss statistical
package (version 18.0). the comparison of folate, vitamin B12
and homocysteine serum levels between users and non-users
was performed by Mann-whitney u-tests (due to the skewed
distribution non-parametric tests were indicated). associations
between folate, vitamin B12 and homocysteine (baseline, five-
year follow-up) and categories of medication use were
analyzed by non-parametric correlation analysis using
spearman's rank correlation coefficients. For further evaluation
of changes in the time course, Δ-changes were calculated for
folate, vitamin B12 and homocysteine distribution. Δ-change is
defined as a measurement at five year follow-up minus baseline
measurement. Δ-change is positive when follow-up values
The Journal of Nutrition, Health & Aging©
Volume 16, Number 8, 2012

INFLUENCE OF FOLIC ACID AND VITAMIN B12 USE IN THE VITA COHORT

increase and negative when follow-up values decrease from showed significantly lower values.
baseline. associations between categorical variables (persons
with stable McI/cognitively unimpaired vs converters to Table 1
dementia and persons with or without brain atrophy or with or Demographic and clinical measures of 81 persons with mild
without white matter lesions) were analyzed by means of the χ- cognitive impairment at baseline
square or Fisher´s exact test (two-sided). the distribution of
variable N or x ± SD Range or % Total N
vitamin serum levels between non-users, inconsistent, and
combination users was compared by kruskal-wallis test. male 25 30.9 % -
to study the combined effect of several co-variates on female 56 69.1 % -
binary dependent variables characterizing cognition (e.g., age (years) 75.8 ± 0.4 75.0 - 76.7 81
persons remaining McI/cognitively unimpaired vs converters education (years) 9.5 ± 2.2 0 - 16 81
to dementia) or brain morphology, we complemented the above > 1 apoe ε4 allele 17 21.0 % 79
body mass index (kg/m2) 27.1 ± 3.9 18.6 - 37.2 81
analysis by binary logistic regression. Variables of brain diabetes 2 2.5 81
morphology were coded either as probands with no versus any myocardial infarction 11 13.6 81
Mta/diffuse brain atrophy/deep white matter lesions and stroke or cerebral infarction in MrI 18 22.2 81
periventricular hyperintensities. In some calculations, probands MMse 27.0 ± 1.8 23 - 30 80
were grouped as having no and a low degree of Mta/diffuse notes: the data represents mean ± standard deviation (x±sD). apoe-ε4: presence of at
brain atrophy/deep white matter lesions versus moderate and least one apoe-ε4 allele. MMse - Mini Mental examination scores
severe degree of investigated parameters. significant predictors
were selected using the step-wise backward elimination method The combined use of folate and vitamin B12 reduced
(the initial set of potential predictors considered in the conversion to dementia
individual analyses depended on the context). the level of p ≤ Firstly, we tested the predictability of conversion to aD by
0.05 was considered as statistically significant. odds ratios and vitamin use. conversion rate to dementia at 5 years was 20% in
95% confidence intervals were calculated to quantify the effect combination users, 75% in inconsistent users and 62.7% in non-
of significant predictors. users (p=0.013, χ-square test). the logistic regression on
conversion to dementia versus the stable McI/cognitively
Results unimpaired group at five year follow-up revealed significant
differences for all three categories of vitamin use (p=0.03). a
Descriptive statistics at baseline
pair-wise comparison of the individual categories revealed that
Baseline characteristic of 81 elderly persons with McI are
there was no significant difference between non-users and
presented in table 1. considering that this study was a
inconsistent users, whereas combination users were
retrospective evaluation of individual drug intake during a five
significantly more likely to remain in the McI/cognitively
year time frame, measurement of vitamin B12 and folate serum
unimpaired group than non-users (p=0.023, table 3). In a
level was necessary to validate the subject´s information
multiple logistic regression model, gender, years of education,
regarding intake. as described in methods, the users were
and the presence of at least one apoe ε4 allele showed no
subdivided into non-users, “inconsistent users” and
significant effect on conversion rate to dementia (data not
“combination users”. table 2 demonstrates the distribution of
shown).
serum vitamins and homocysteine levels in non-users,
inconsistent users, and combination users at baseline, 5 years, Serum levels of homocysteine, folate, vitamin B12 and
as well as any changes between baseline and 5 years. Baseline prediction of conversion to dementia
vitamin B12, folate, and homocysteine levels at 5 years, as well In a second model we investigated whether the serum levels
as the Δ-changes of folate significantly differed between of homocysteine, folate or vitamin B12 at baseline, at five years
groups (kruskal-wallis-test). compared to non-users, the folate or their Δ-changes over five years might be associated with a
levels at 5 years were significantly higher in inconsistent and lower conversion rate to dementia. In univariate logistic
combination users. the Δ-changes of folate in inconsistent regression, higher levels of folate at baseline and at five-years
users were also significantly higher than those of non-users. showed a trend association with a lower conversion rate to
the mean of vitamin B12 level at 5 years in combination users dementia (p<0.10; table 4). neither homocysteine nor vitamin
was also higher compared to non-users, but this increase did not B12, as measured at baseline or at five years or as Δ-changes,
reach a level of significance. Interestingly, already at baseline were significantly associated with conversion to dementia (data
the combination users showed higher levels of vitamin B12 not shown).
than non-users. the joint effect of baseline and 5-year folate levels on
Finally, as an outcome parameter of vitamin conversion to dementia was tested in a logistic regression
supplementations, the homocysteine level was followed to time model, with adjustment for the patient’s characteristics, gender,
and manner of intake. compared to non-users, the mean of years of education, creatinine (baseline and 5 years), and the
homocysteine level in inconsistent and combination users presence of at least one apoe ε4 allele. gender correction was

690
The Journal of Nutrition, Health & Aging©
Volume 16, Number 8, 2012

JNHA: CLINICAL NEUROSCIENCES

Table 2
serum folate (ng/ml), vitamin B12 (pg/ml) and homocysteine (µmol/l) levels in nonusers, inconsistent and combination users.

users statistics
non inconsistent combination overall non vs non vs
inconsistent combination

folate
baseline 8.4 ± 3.6 9.4 ± 5.1 12.9 ± 7.2 - - -
5 years 8.5 ± 4.3 14.7 ± 8.5 18.4 ± 6.0 0.001 0.024 <0.000
Δ -change 0.1 ± 4.7 5.5 ± 7.3 3.7 ± 11.7 0.016 0.005 -
vitamin B12
baseline 479 ± 242 752 ± 593 616 ± 174 0.034 - 0.029
5 years 457 ± 205 712 ± 600 564 ± 143 - - -
Δ-change -28 ± 240 -5.8 ± 302 -61 ± 194 - - -
homocysteine
baseline 15.0 ± 4.7 14.0 ± 4.4 12.3 ± 2.5 - - -
5 years 15.2 ± 6.3 12.7 ± 4.7 11.1 ± 1.6 0.012 0.044 0.007
Δ-change 1.1 ± 5.8 -1.4 ± 4.9 -0.6 ± 2.9 - - -

notes: Data are presented as mean ± standard deviation; overall – kruskal-wallis test; group comparisons between non-users versus inconsistent and combination users were calculated
in Mann-whitney u-test. – indicates non significant differences, p > 0.05.

necessary since most probands were women. Higher levels of plasma levels or a subject´s information on vitamin use better
folate at baseline predicted lower conversion rates to dementia predict the conversion to dementia. In a logistic regression,
at five years (table 5.). More years of education also showed a including categories of vitamin intake (non-users, inconsistent
trend association with lower conversion rates to dementia. users and combination users) and plasma levels of
Folate at 5 years failed to reach a level of significance. In order homocysteine, folate, vitamin B12 as independent variables, the
to investigate whether the effect of folate on conversion to aD results were the same as in the second model, where higher
might be moderated by gender, we studied the interaction levels of folate were associated with a lower conversion rate.
between the two factors. the results showed a significant the variable, vitamin intake was not found to be an additional
interaction between gender and folate at baseline (p=0.042, significant predictor once the variable folate level had been
or=0.61, 95% cI: 0.38-0.98). to further explore this entered into the model (data not shown).
interaction, a sub-group analysis by gender was performed. In
this analysis an almost significant association of higher folate Table 4
with lower conversion rate to dementia was found in women univariate logistic regressions model for association with lower
(p=0.055, or=0.27, 95% cI: 0.07-1.03) but not in men conversion to dementia at 5 years. single tested variable were
(p=0.40). folate, vitamin B12 or homocysteine levels at baseline, at 5
years or their Δ-changes. Variables with only p<0.10 are
Table 3 presented
logistic regression on prediction of conversion to dementia at 5
year follow up in dependence on vitamin use variable ß coefficient p-value Odds Ratio
95% Confidence
Interval
MCI/cognitive healthy ß coefficient p-value Odds Ratio
versus converters to dementia 95% Confidence ln folate at baseline -0.97 0.060 0.38 (0.14-1.0)
Interval ln folate at 5 years -0.83 0.067 0.44 (0.18-1.1)

all categories? - 0.030 - notes: Variables with p>0.10 : Δ-changes of ln folate, ln vitamin B12 at baseline, 5-years
non users vs inconsistent users 0.58 0.330 1.8 (0.6-5.6) or its Δ-changes, ln homocysteine at baseline, 5-years or its Δ-changes
non users vs combination users - 1.93 0.023 0.15 (0.03-0.77)
The use of folic acid and vitamin B12 and brain
morphologic parameters
notes: the exposure to folic acid or vitamin B12 in months during the 5 years was coded
as variable with categories of non users, inconsistent users and combination users for
longer than 12 months. the group of inconsistent users consists from persons using either Having observed that higher levels of folate at baseline were
associated with lower conversion rates to dementia, we tested
folic acid or vitamin B12 in single use irrespective on time and persons using combination
of both substances for less than one year.
whether this effect might be accompanied by less global brain
In addition, a significant interaction of folate levels at 5 or medial temporal lobe atrophy and decreased distribution of
years and gender was found (p=0.011, or=0.52, 95% cI: 0.32- periventricular hyperintensities or deep white matter lesions.
0.86). By testing the 5 year folate levels in a sub-group analysis as in the above analysis on cognition, we applied logistic
by gender, a significant association with lower conversion rates regression using the presence of atrophy in the global brain or
to dementia was found in women (p=0.023, or=0.19, 95% cI: in the medial temporal lobe, as well as the extent of
0.04-0.80), but not in men (p=0.76). periventricular hyperintensities or white matter lesions as the
Finally, we investigated whether objectively measured dependent variable. Independent variables considered were
691
The Journal of Nutrition, Health & Aging©
Volume 16, Number 8, 2012

INFLUENCE OF FOLIC ACID AND VITAMIN B12 USE IN THE VITA COHORT

serum levels of homocysteine, folate and vitamin B12 at Table 6

baseline, at five-years, and Δ-changes of each parameter over univariate logistic regressions models for prediction or
five years. association with brain morphologic parameter at 5 years. single
tested variable were folate, vitamin B12 or homocysteine levels
Table 5 at baseline, at 5 years or their Δ-changes. Variables with only
Multivariate logistic regression model for prediction of lower p<0.10 are presented
conversion to dementia. Independent variable tested were
gender, education years, the presence of at least one apoe ε4 ß coefficient p-value Odds Ratio
95% Confidence
allele, folate and creatinine at baseline as well as folate and Interval
creatinine at 5 years medial temporal lobe
ln folate at baseline -1.40 0.024 0.25 (0.09-0.83)
variable ß coefficient p-value Odds Ratio global brain atrophy
ln homocysteine at baseline 2.40 0.023 18.2 (1.5-220.03)
95% Confidence ln folate at 5 years -1.00 0.050 0.37 (0.13-0.99)
Interval ln vitamin B12 at 5 years -1.45 0.040 0.23 (0.06-0.93)
deep white matter lesions
ln folate at baseline -1.77 0.013 0.17 (0.04-0.69) Δ-change of ln vitamin B12 -1.96 0.045 0.14 (0.02-0.96)
gender* -1.93 0.010 0.15 (0.03-0.62)
notes: the dependent variable medial temporal lobe consisted of persons with
education years -0.26 0.062 0.77 (0.58-1.01)
normal+mild versus persons with moderate+severe+very severe atrophy. the dependent
variable global brain athrophy consisted of persons with normal+mild versus persons with
notes: * males=0, female=1 moderate+severe atrophy. the dependent variable deep white matter lesions consisted of
persons without any lesions versus persons with “punctate foci” and “beginning confluence
In a univariate logistic regression, higher folate levels at of foci” and “large confluent areas”.
baseline were found to predict lower atrophy of medial
temporal lobes (table 6). Higher serum levels of homocysteine Table 7
at baseline predicted higher rates of global brain atrophy at Distribution of periventricular hyperintensities and deep white
five-years (no atrophy/low vs moderate/severe atrophy group). matter lesions at 5 year follow up in vitamin B12 and/or folic
Higher folate and vitamin B12 levels at 5 years were associated acid users versus non-users
with lower rates of moderate/severe global brain atrophy. no
associations were found for periventricular hyperintensities (not folic acid /
vitamin B12 users
non-users statistics*
shown). For deep white matter lesions, higher values of Δ-
p = 0.043
change of vitamin B12 levels were associated with lower rates
periventricular hyperintensities
absent 11 (55.0%) 9 (25.7%)
of deep white matter lesions. any
total
9 (45.0%)
20 (100.0%)
26 (74.3%)
35 (100.0%)
the joint effect of variables associated (p<0.10) with deep white matter lesions p = 0.024
absence + punctate foci 15 (75.0%) 14 (40.0%)
particular brain morphologic parameters was tested in logistic beginning confluence of foci + 5 (25.0%) 21 (60.0%)
regression models, with adjustment for gender and presence of large confluent areas
total 20 (100.0%) 35 (100.0%)
at least one apoe ε4 allele. Higher folate at baseline remained
the only predictor of no/ low medial temporal lobe atrophy at 5 notes: * Fisher’s exact test; the dependent variable any periventricular hyperintensities

years (p=0.013). For global brain atrophy, a higher

homocysteine level at baseline was the only significant
predictor for moderate/severe global brain atrophy at five-years
(p=0.046). the association of Δ-change of vitamin B12 with
the absence of deep white matter lesions became less
pronounced and reached only trend-level significance
(p=0.064).
Finally, we tested the effects of vitamin use on brain
morphological parameters. the distribution of vitamin use was
converted to a dichotomous variable (any use of folic acid
and/or vitamin B12 versus no use) because only 55 probands
had MrI-scans. the use of vitamin B12 or folic acid in the past
five years, independent of its duration or intake manner, was
associated with a lower extent of periventricular
hyperintensities and with a lower extent of deep white matter
lesions compared to non use of vitamins (table 7. Fisher´s
exact test p=0.043 and p=0.024, resp). the distribution of
global brain and medial temporal lobe atrophy did not differ
significantly between users and non-users of vitamin B12 or
folic acid (data not shown).
692
consisted of persons with “caps and periventricular lines” plus persons with
“periventricular bands” plus persons with “irregular hyperintensities with extension in deep
white matter”.
Discussion
at the 5-year evaluation time the serum levels of folate and
vitamin B12 were 116% and 19% higher, respectively, in those
75-year old persons who reported taking supplementary B
vitamins. although the change of vitamin B12 was not
significant, homocysteine levels were significantly lower by
27% as compared to baseline. this effect can be regarded as a
response to the vitamin intake. the group of combination users
for more than one year showed lowest homocysteine levels,
which suggests a time and dose dependent effect of the
probands´ voluntary vitamin use.
60% of initial McI-persons had developed dementia during
the previous five years. this conversion rate, calculated over
five years, by an annual conversion rate of 12%, corresponds
well to other populations (31). 15% of initially mildly impaired
persons improved to cognitively unimpaired after five years of
observation. such an improvement of cognition in old age
The Journal of Nutrition, Health & Aging©
Volume 16, Number 8, 2012
JNHA: CLINICAL NEUROSCIENCES
frequently occurs and reflects the natural course of the disorder
(32, 33).
the data were analyzed in two ways. First, as subjective
information on vitamin intake and secondly as objective
information based on serum measurements. testing the
predictability of subjective information, we divided users into
three sub-groups: non-users, inconsistent, and combination
users. this division separation already revealed significant
differences for all categories in respect to dementia conversion.
Information on vitamin use separated combination users from
inconsistent and non-users in respect to conversion to dementia.
compared to non-users, the conversion rate in combinations
users declined to one third (i.e 20% vs 63%). the reason for
high conversion rate in inconsistent users (75%) is less clear
and might be due to factors not covered by this study, but it was
not significantly different from the rate in non-users (table 3).
By direct comparison of predictors, serum levels were a
stronger predictor than subjective information on vitamin
intake.
when analyzing serum levels, higher folate at baseline and
five years was associated with lower rates of dementia (table
4). In multiple regression analysis, higher baseline levels were
the only significant predictor, concluding that baseline values
are a better predictor than five-year values in respect to
dementia conversion. nevertheless, both baseline and five-year
folate levels showed significant interaction with gender.
However, in a sub-group analysis by gender, only 5-year levels
showed a clear, significant association with a lower rate of
dementia in females. Baseline levels failed to reach a
significance (p=0.055). aside from this, the female probands,
in contrast to male probands, clearly showed a lower
conversion rate. an association of higher folate levels with
lower conversion rate was described in previous studies in
populations where both gender were more equally distributed
or higher total numbers of subjects were included (34, 35). the
limitation on females in our study is less clear and probably
results from selection bias, since nearly 70% of the subjects
were women. a lower conversion rate to dementia was
associated with higher levels of folate, but was independent of
the presence of apoe ε4 allele in multiple testing (table 5.).
this may correspond to the observation where the B vitamin
status did not influence the risk of overall cognitive dysfunction
in apoe ε4 allele affected older adults (36).
serum levels of homocysteine and vitamin B12 measured at
different times or as change-over-time were not found to be
significant predictors for dementia conversion in this study.
this is in contrast with several other studies (35, 37), and also
with a publication based on the same cohort (11), where the
increase of homocysteine over 2.5 years was significantly
associated with conversion from cognitive health to
alzheimer´s disease. In addition, persons with stable McI over
2.5 years in this previous study showed lower Δ-changes of
homocysteine when compared to converters to aD from
baseline McI. the reason for the difference in these two studies
based on the same cohort might be due to the different lengths
693
of the studies, i.e. 2.5 versus 5 years and due to a lower number
of selected persons in the present study (487 versus 81 of re-
investigated persons).
Investigations of the association of objectively measured
vitamin levels with brain morphology showed that higher folate
levels at baseline were a predictor of lower medial temporal
lobe atrophy at five years and higher serum level of
homocysteine at baseline predicted a higher rate of global brain
atrophy. this observation corresponds to a prospective study
showing that low B12 levels are associated with a faster rate of
brain atrophy (38). lower brain volumes were previously
associated with higher homocysteine levels in volumetric
studies (9, 39); these findings could be corroborated in our
semi-quantitative measurements by a visually-based scale.
testing the association of vitamin use with brain
morphology, the use of vitamin B12 and/or folic acid in the
past five years was associated with a lower rate of
hyperintensities and with lower rate of deep white matter
lesions compared to non use of vitamins. In a previous study,
high homocysteine levels were significantly associated with
white matter hyperintensities at least in males (40). regarding
the decrease of homocysteine as an effect of vitamin intake in
our McI-patients, those persons taking folic acid or vitamin
B12 explained the lower extent of periventricular
hyperintensities in MrI. this is also partly supported by
univariate regression results where higher values of vitamin
B12 change over 5 years were associated with lower rates of
deep white matter lesions. the results on vitamin B12 are
consistent with findings from the rotterdam scan study (41).
our study supports previous results from other cohorts. the
strengths of this particular study are the selection of McI
persons of the same age from a population-based cohort, the
long follow-up period of five years, and the evaluation of
homocysteine, folate and vitamin B12 in relation to dementia
conversion. the 81 subjects are from a well-characterized
longitudinal study (VIta) designed to investigate risk factors
for dementia and aD. the decrease of the homocysteine serum
level in persons with a subjective indication for vitamin
supplementation suggests some validity of such information;
however, several limitations remain: the identification of prior
folic acid and vitamin B12 supplementation was based on
patient´s self-report without the use of an administrative
database. this is problematic due to a retrospective reporting
bias (42). the main restriction for our study includes the small
sample size which made the pooling of McI and cognitive
unimpaired individuals necessary. Moreover, the low number
of converters in each dementia sub-group precluded the
analysis with respect to efficacy of each vitamin. some of the
calculations of medication intake were dichotomously
classified into users or non-users; as a result, dose-response
relations could have been missed. In addition, less than 60% of
the approached probands underwent follow-up investigations at
5 years, which decreases the generalizability of these results.
this study indicates the involvement of folate in the
conversion of McI into dementia and of homocysteine by
The Journal of Nutrition, Health & Aging©
Volume 16, Number 8, 2012
INFLUENCE OF FOLIC ACID AND VITAMIN B12 USE IN THE VITA COHORT
increase of global brain atrophy. a few randomized, controlled
trials have so far shown vitamin B supplementation to be of
value in preventing conversion to dementia (6, 9, 10, 43).
However, another study did not support associations of serum
concentrations of homocysteine, vitamin B12, or folic acid with
the rate of cognitive decline in unselected elderly persons (44).
this discrepancy emphasizes the need for further studies
identifying individuals who are at increased risk of developing
dementia. regarding McI-patients as a possible target
population for treatment intervention, supplementation with
folate or vitamin B12 may be most effective during a critical
time window. Future randomized controlled trials need to
determine the critical serum levels of homocysteine, folate or
vitamin B12 on which to base the recommendation to
supplement. Moreover, such treatment must be based on the
cognitive and brain morphological parameters of each
individual.
Acknowledgements: we acknowledge the funding and organization of the Vienna
transdanube aging (VIta) study by the ludwig Boltzmann society, ludwig Boltzmann
Institute of aging research (head: prof. karl-Heinz tragl), p. Bauer for statistical advice
in planning of the VIta study, w. kirchmeyr and s. torma for medical exploration of
subjects and H. Hinterhuber for continuous support of our work.
Disclosure Statement: all authors of this manuscript declare no conflict of interest.
References
1. sachdev ps. Homocysteine and brain atrophy. prog neuropsychopharmacol Biol
psychiatry. 2005 sep;29:1152-61.
2. selhub J, Bagley lc, Miller J, rosenberg IH. B vitamins, homocysteine, and
neurocognitive function in the elderly. am J clin nutr. 2000 Feb;71:614s-20s.
3. weir Dg, scott JM. Brain function in the elderly: role of vitamin B12 and folate. Br Med
Bull. 1999;55:669-82.
4. goodwin Js, goodwin JM, garry pJ. association between nutritional status and cognitive
functioning in a healthy elderly population. JaMa. 1983 Jun 3;249:2917-21.
5. clarke r, smith aD, Jobst ka, refsum H, sutton l, ueland pM. Folate, vitamin B12,
and serum total homocysteine levels in confirmed alzheimer disease. arch neurol. 1998
nov;55:1449-55.
6. smith aD. the worldwide challenge of the dementias: a role for B vitamins and
homocysteine? Food nutr Bull. 2008 Jun;29:s143-s172.
7. Dangour aD, whitehouse pJ, rafferty k, Mitchell sa, smith l, Hawkesworth s, Vellas
B. B-vitamins and fatty acids in the prevention and treatment of alzheimer's disease and
dementia: a systematic review. J alzheimers Dis. 2010;22:205-24.
8. Malouf r, grimley eJ. Folic acid with or without vitamin B12 for the prevention and
treatment of healthy elderly and demented people. cochrane Database syst rev.
2008;cD004514.
9. smith aD, smith sM, de Jager ca, whitbread p, Johnston c, agacinski g, oulhaj a,
Bradley kM, Jacoby r, refsum H. Homocysteine-lowering by B vitamins slows the rate
of accelerated brain atrophy in mild cognitive impairment: a randomized controlled trial.
plos one. 2010;5:e12244.
10. de Jager ca, oulhaj a, Jacoby r, refsum H, smith aD. cognitive and clinical outcomes
of homocysteine-lowering B-vitamin treatment in mild cognitive impairment: a
randomized controlled trial. Int J geriatr psychiatry. 2011 Jul 21.
11. Blasko I, Jellinger k, kemmler g, krampla w, Jungwirth s, wichart I, tragl kH, Fischer
p. conversion from cognitive health to mild cognitive impairment and alzheimer's
disease: prediction by plasma amyloid beta 42, medial temporal lobe atrophy and
homocysteine. neurobiol aging. 2008 Jan;29:1-11.
12. Fischer p, Jungwirth s, krampla w, weissgram s, kirchmeyr w, schreiber w, Huber k,
rainer M, Bauer p, tragl kH. Vienna transdanube aging "VIta": study design,
recruitment strategies and level of participation. J neural transm suppl. 2002;105-16.
13. Fischer p, Jungwirth s, zehetmayer s, weissgram s, Hoenigschnabl s, gelpi e, krampla
w, tragl kH. conversion from subtypes of mild cognitive impairment to alzheimer
dementia. neurology. 2007 Jan 23;68:288-91.
14. winblad B, palmer k, kivipelto M, Jelic V, Fratiglioni l, wahlund lo, nordberg a,
Backman l, albert M, et al. Mild cognitive impairment-beyond controversies, towards a
consensus: report of the International working group on Mild cognitive Impairment. J
Intern Med. 2004 sep;256:240-6.
15. Jungwirth s, weissgram s, zehetmayer s, tragl kH, Fischer p. VIta: subtypes of mild
cognitive impairment in a community-based cohort at the age of 75 years. Int J geriatr
psychiatry. 2005 May;20:452-8.
16. Morris Jc, Heyman a, Mohs rc, Hughes Jp, van Belle g, Fillenbaum g, Mellits eD,
clark c. the consortium to establish a registry for alzheimer's Disease (ceraD). part I.
clinical and neuropsychological assessment of alzheimer's disease. neurology.
694
1989;39:1159-65.
17. Berres M, Monsch au, Bernasconi F, thalmann B, stahelin HB. normal ranges of
neuropsychological tests for the diagnosis of alzheimer's disease. stud Health technol
Inform. 2000;77:195-9.
18. reitan rM. Validity of the trail Making test as an indicator of organic brain damage.
perceptual and Motor skills. 1958;8:271-6.
19. Mckhann g, Drachman D, Folstein M, katzman r, price D, stadlan eM. clinical
diagnosis of alzheimer's disease: report of the nIncDs-aDrDa work group under the
auspices of Department of Health and Human services task Force on alzheimer's
Disease. neurology. 1984 Jul;34:939-44.
20. clinical and neuropathological criteria for frontotemporal dementia. the lund and
Manchester groups. J neurol neurosurg psychiatry. 1994 apr;57:416-8.
21. Mckeith Ig. consensus guidelines for the clinical and pathologic diagnosis of dementia
with lewy bodies (DlB): report of the consortium on DlB International workshop. J
alzheimers Dis. 2006;9:417-23.
22. roman gc, tatemichi tk, erkinjuntti t, cummings Jl, Masdeu Jc, garcia JH,
amaducci l, orgogozo JM, Brun a, et al. Vascular dementia: diagnostic criteria for
research studies. report of the nInDs-aIren International workshop. neurology. 1993
Feb;43:250-60.
23. Fischer p, krampla w, Mostafaie n, zehetmayer s, rainer M, Jungwirth s, Huber k,
Bauer k, Hruby w, et al. VIta study: white matter hyperintensities of vascular and
degenerative origin in the elderly. J neural transm suppl. 2007;181-8.
24. Fazekas F, chawluk JB, alavi a, Hurtig HI, zimmerman ra. Mr signal abnormalities at
1.5 t in alzheimer's dementia and normal aging. aJr am J roentgenol. 1987
aug;149:351-6.
25. scheltens p, Barkhof F, leys D, pruvo Jp, nauta JJ, Vermersch p, steinling M, Valk J. a
semiquantative rating scale for the assessment of signal hyperintensities on magnetic
resonance imaging. J neurol sci. 1993 Jan;114:7-12.
26. Bracco l, piccini c, Manfredi g, Fonda c, Falcini M, amaducci l. Magnetic resonance
measures in alzheimer disease: their utility in early diagnosis and evaluating disease
progression. alzheimer Dis assoc Disord. 1999 Jul;13:157-64.
27. Fox nc, Freeborough pa, rossor Mn. Visualisation and quantification of rates of atrophy
in alzheimer's disease. lancet. 1996 Jul 13;348:94-7.
28. luxenberg Js, Haxby JV, creasey H, sundaram M, rapoport sI. rate of ventricular
enlargement in dementia of the alzheimer type correlates with rate of neuropsychological
deterioration. neurology. 1987 Jul;37:1135-40.
29. Meese w, kluge w, grumme t, Hopfenmuller w. ct evaluation of the csF spaces of
healthy persons. neuroradiology. 1980 apr;19:131-6.
30. Huber kr, Mostafaie n, stangl g, worofka B, kittl e, Hofmann J, Hejtman M, Michael
r, weissgram s, et al. clinical chemistry reference values for 75-year-old apparently
healthy persons. clin chem lab Med. 2006;44:1355-60.
31. Mitchell aJ, shiri-Feshki M. rate of progression of mild cognitive impairment to
dementia--meta-analysis of 41 robust inception cohort studies. acta psychiatr scand. 2009
apr;119:252-65.
32. ganguli M, snitz Be, saxton Ja, chang cc, lee cw, Vander BJ, Hughes tF,
loewenstein Da, unverzagt Fw, petersen rc. outcomes of mild cognitive impairment
by definition: a population study. arch neurol. 2011 Jun;68:761-7.
33. Manly JJ, tang MX, schupf n, stern Y, Vonsattel Jp, Mayeux r. Frequency and course
of mild cognitive impairment in a multiethnic community. ann neurol. 2008 apr;63:494-
506.
34. annerbo s, wahlund lo, lokk J. the relation between homocysteine levels and
development of alzheimer's disease in mild cognitive impairment patients. Dement geriatr
cogn Disord. 2005;20:209-14.
35. luchsinger Ja, tang MX, Miller J, green r, Mayeux r. relation of higher folate intake to
lower risk of alzheimer disease in the elderly. arch neurol. 2007 Jan;64:86-92.
36. Brown B, Huang MH, karlamangla a, seeman t, kado D. Do the effects of apoe-
epsilon4 on cognitive function and decline depend upon vitamin status? Macarthur
studies of successful aging. J nutr Health aging. 2011 Mar;15:196-201.
37. Quadri p, Fragiacomo c, pezzati r, zanda e, tettamanti M, lucca u. Homocysteine and
B vitamins in mild cognitive impairment and dementia. clin chem lab Med.
2005;43:1096-100.
38. Vogiatzoglou a, refsum H, Johnston c, smith sM, Bradley kM, de Jc, Budge MM,
smith aD. Vitamin B12 status and rate of brain volume loss in community-dwelling
elderly. neurology. 2008 sep 9;71:826-32.
39. seshadri s, wolf pa, Beiser as, selhub J, au r, Jacques pF, Yoshita M, rosenberg IH,
D'agostino rB, Decarli c. association of plasma total homocysteine levels with
subclinical brain injury: cerebral volumes, white matter hyperintensity, and silent brain
infarcts at volumetric magnetic resonance imaging in the Framingham offspring study.
arch neurol. 2008 May;65:642-9.
40. sachdev ps, parslow r, wen w, anstey kJ, easteal s. sex differences in the causes and
consequences of white matter hyperintensities. neurobiol aging. 2009 Jun;30:946-56.
41. de lau lM, smith aD, refsum H, Johnston c, Breteler MM. plasma vitamin B12 status
and cerebral white-matter lesions. J neurol neurosurg psychiatry. 2009 Feb;80:149-57.
42. Bradburn nM, rips lJ, shevell sk. answering autobiographical questions: the impact of
memory and inference on surveys. science. 1987 apr 10;236:157-61.
43. chan a, remington r, kotyla e, lepore a, zemianek J, shea tB. a vitamin/nutriceutical
formulation improves memory and cognitive performance in community-dwelling adults
without dementia. J nutr Health aging. 2010 Mar;14:224-30.
44. Mooijaart sp, gussekloo J, Frolich M, Jolles J, stott DJ, westendorp rg, de craen aJ.
Homocysteine, vitamin B-12, and folic acid and the risk of cognitive decline in old age: the
leiden 85-plus study. am J clin nutr. 2005 oct;82:866-71.