Bone 40 (2007) 730 736

www.elsevier.com/locate/bone

The association of homocysteine and its determinants MTHFR genotype,

folate, vitamin B12 and vitamin B6 with bone mineral density in
postmenopausal British women
M. Baines , M.-B. Kredan, J. Usher, A. Davison, G. Higgins, W. Taylor,
C. West, W.D. Fraser, L.R. Ranganath
Department of Clinical Biochemistry and Metabolic Medicine, Royal Liverpool University Hospital, Liverpool, L7 8XP, UK
Received 14 July 2006; revised 8 September 2006; accepted 1 October 2006
Available online 1 December 2006

Abstract
We studied the association between plasma total homocysteine (tHcy), its determinants folate, vitamin B12, vitamin B6 and MTHFR genotype,
and bone mineral density (BMD) in 328 postmenopausal British women. When the subjects were assigned to one of 3 groups (control, osteopenic
or osteoporotic) according to their BMD at the os calcis, those in the osteoporotic group had, compared with the controls, a significantly lower
serum folate concentration, a significantly higher % of current smokers and a significantly higher incidence of recent fracture.
In the population as a whole, we found significant associations of BMD with tHcy (r = 0.130, p = 0.033, log tHcy) and folate (r = 0.132,
p = 0.025, log folate). The association of folate with BMD was maintained after correction for age, weight and height (r = 0.124, p = 0.042, log
folate), but the association of tHcy with BMD weakened after correction for age, weight, height and creatinine (r = 0.117, p = 0.059, log tHcy).
Vitamins B12 and B6 were not associated with BMD, but were significantly associated with tHcy, vitamin B12 (r = 0.34, p < 0.0001), vitamin B6
(r = 0.16, p = 0.007), as was folate (r = 0.41, p < 0.0001). There was an increasing frequency of the MTHFR TT genotype across the 3 BMD
groups, but this did not attain significance. Individuals with the TT genotype had significantly higher plasma tHcy but there was no difference
between the genotypes (CC, CT, TT) for folate or BMD. Smoking was associated with a highly significant reduction in BMD and lower weight,
and a significant reduction in circulating folate and vitamin B6 concentrations, but no change in tHcy or vitamin B12 concentrations when
compared with non-smokers.
We conclude that low serum folate is a significant risk factor for osteoporosis, with plasma tHcy having a lesser effect. Both vitamins B12 and
B6, by acting through tHcy, may also have an effect on the skeleton, albeit a weaker one than folate. Cigarette smoking is a strong determinant of
BMD, and may act through effects on folate and vitamin B6.
2006 Elsevier Inc. All rights reserved.
Keywords: Bone mineral density; Homocysteine; Folate; Smoking; MTHFR genotype

Introduction
The link between homocysteine and skeletal abnormality was
first established in studies of the classical metabolic disorder,
Homocystinuria, caused by deficiency of the first enzyme of
homocysteine trans-sulfuration, cystathionine -synthase (E.C.
4.2.1.22) [1].Since this finding, further investigations have been
limited, but have been stimulated by a renewed interest in
homocysteine (tHcy) and more accessible methods for its
Corresponding author. Fax: +44 0151 706 4230.
E-mail address: malcolm.baines@rlbuht.nhs.uk (M. Baines).
8756-3282/$ - see front matter 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.bone.2006.10.008

measurement. Two recent population-based studies, from the

Netherlands and Framingham, USA, have shown an association
between increased concentrations of plasma total tHcy, and risk
of osteoporotic fracture in both male and female subjects [2,3].
Previously, studies have investigated determinants of tHcy, the
vitamins folate, B6, B12 and polymorphism of the methylenetetrahydrofolate reductase (MTHFR) enzyme and their relation
with markers of fracture risk, particularly bone mineral density
(BMD). An association of BMD with MTHFR polymorphism
was reported in a group of 307 postmenopausal Japanese women
[4], where those homozygous for the thermo labile TT variant
had significantly lower BMD of both lumbar spine and total

M. Baines et al. / Bone 40 (2007) 730736

body when compared with the CC and CT carriers. Similar

findings were reported in a group of 1748 postmenopausal
women in the Danish Osteoporosis Prevention Study, where
those with the TT genotype had a significantly lower BMD at
femoral neck, hip and spine, and also a 2-fold increased fracture
risk when compared with those with the CC or CT genotypes [5].
However, another study of postmenopausal Danish women,
where 74 women with forearm fracture and 41 with hip fracture
were compared with 207 controls, did not find an association of
MTHFR genotype with peripheral BMD, but paradoxically
found a relative protection against risk of fracture in those with
the TT genotype [6].
The vitamin determinants folate and vitamin B12, together
with tHcy itself were measured in a group of 161 postmenopausal Italian women, where folate, but not vitamin B12 or tHcy,
was both significantly lower in osteoporotic women as
compared with controls and also independently related to
BMD [7]. However, a study of folate, tHcy and MTHFR
polymorphism with BMD in 271 postmenopausal Iranian
women concluded that hyperhomocysteinemia, as a result of
folate deficiency but not MTHFR polymorphism, was independently associated with low BMD[8]. Thus findings have not
been entirely consistent, and may be dependent upon populations studied, and site or method of assessment of BMD. We
report here the first British study to investigate together MTHFR
polymorphism, tHcy and its determinants folate, vitamin B12
and vitamin B6, the lifestyle indicators alcohol intake and
smoking habit and peripheral BMD in a group of postmenopausal women.
Subjects and methods
328 British women at least 12 months since their last menstrual period were
selected at random to attend our Metabolic Bone Unit following take up of a
Peripheral Instantaneous X-ray Imager (PIXI) (GELunar Inc, Madison, WI) scan
of the os calcis as part of a GP osteoporosis screening program. All subjects gave
written consent to the study which was approved by the local research ethics
committee. The demographic details of the subjects are shown in Table 1. The

731

subjects were assigned to one of three groups according to their T-score, a

control group (N) with a T-score of 0.6 or higher, an osteopenic group (OPN)
with a T-score between 0.7 and 1.6, and an osteoporotic group (OP) with a Tscore of lower than 1.6 [8]. Recruitment continued prospectively until each
group had approximately equal numbers. Following an overnight fast of 12 h,
blood samples were taken for tHcy, folate and vitamin B12, vitamin B6, basic
bone, liver and renal biochemical profiles and MTHFR genotype. tHcy was
measured by a dedicated HPLC system (Drew DS 30, Drew Scientific, UK) [9]
on an EDTA plasma sample which had been separated from the cells within 1 h of
venepuncture and stored at 20C. The intra-assay coefficient of variation (CV)
was 2.7% at a concentration of 16.9 mol/L, and the inter-assay CV 7.4% at a
concentration of 17.8 mol/L. Serum folate and vitamin B12 were measured by
competitive binding receptor assay on a Beckman Access analyzer ( Beckman
Coulter UK Ltd., High Wycombe, UK). The inter-assay CV was 7.4% at a
concentration of 4.9 ng/mL for folate and 6.6% at a concentration of 276 pmol/L
for vitamin B12. Vitamin B6 was measured as pyridoxal-5-phosphate by an HPLC
method (Chromsystems, Munich, Germany) on a heparin plasma sample. The
intra-assay CV was 2.6% at a concentration of 78 nmol/L. The 677 point
mutation in the MTHFR gene was analyzed by PCR amplification and melting
curve analysis using hybridization probes monitored by the LightCycler (Roche
Molecular Biochemicals, Lewes, UK) after DNA isolation from an EDTA sample
with the QIAamp Kit [11]. Accuracy of phenotyping was assessed by inclusion of
an internal quality control sample of known genotype with each run. Plasma
creatinine was measured with a kinetic Jaffe reaction on the Roche Modular
system (Roche Diagnostic, Lewes, UK).
Statistical analysis was carried out using SPSS software, version 13 (SPSS,
Chicago, IL, USA). P values were calculated by one-way ANOVA for
continuous variables and the chi-square test for categorical variables. The
Pearson correlation coefficient was calculated as a measure of association, after
correction for skewness (where appropriate) as indicated by the Kolmogorov
Smirnov test. Partial correlation analysis was used to determine the association
after correction for confounding variables age, weight and height (for BMD) and
additionally plasma creatinine for tHcy. Differences between smokers and nonsmokers were assessed by the MannWhitney U-test. To negate any bias due to
high potency vitamin supplementation usage, values >3SD from the population
mean were deleted as outliers for folate (2 outliers), vitamin B6 (6 outliers) and
vitamin B12 (8 outliers). A p value of <0.05 was considered statistically
significant for all analyses.

Results
Table 1 summarizes the variables distributed between the
groups ranked according to BMD. The gene frequencies of group

Table 1
Demographic and measured analytes of the groups, ranked by BMD

Age, years
Weight, kg
Height, m
BMI, kg/m2
Alcohol (Units/week)
Current smoker (%)
Fracture in previous year (%)
Frequency of MTHFR CC genotype (%)
Frequency of MTHFR CT genotype (%)
Frequency of MTHFR TT genotype (%)
Homocysteine, mol/L
Folate, g/L
Vitamin B12, pmol/L
Vitamin B6, nmol/L
Creatinine mol/L

Normal (N)
BMD = >0.5 n = 110

Osteopenia (OPN)
BMD 1.6 to 0.6 n = 108

Osteoporosis (OP)
BMD = < 1.7 n = 110

P (ANOVA or
Chi-square)

67.6 (4584)
72.16 (12.31)
1.608 (0.067)
27.84 (4.96)
2.2 (3.81)
7.3
16.4
43.4
48.1
8.5
12.00 (4.87)
9.05 (4.69)
255.0 (94.2)
43.85 (22.54)
89.4 (19.22)

66.1 (4080)
64.02 (11.03)
1.590 (0.058)
25.14 (3.53)
2.4 (4.35)
18.3
23.1
39.8
47.2
11.1
11.16 (4.17)
10.16 (4.63)
269.7 (85.4)
47.19 (24.46)
88.3 (13.5)

68.9 (4186)
60.06 (10.12)
1.562 (0.072)
24.52 (4.09)
3.1 (5.88)
26.4
30.9
42.7
44.5
12.7
13.22 (5.32)
7.41 (3.67)
247.5 (99.9)
41.29 (24.53)
86.7 (17.38)

0.014c
<0.001a,b, 0.009c
<0.0001a, 0.041b, 0.003c
<0.0001a,b
NS
0.0003a, 0.027b
0.017a

Data given as mean (SD), except age given as mean (range).

Legend for significance values: aNvOP; bNvOPN; cOPNvOP; NS = no significant difference between any group.

0.003c
0.007a, <0.0001c
NS
NS
NS

732

M. Baines et al. / Bone 40 (2007) 730736

Table 2
Correlations between BMD, tHcy and other variables for the whole population

BMD
BMD
BMD
BMD
BMD
BMD
BMD
BMD
BMD
BMD
Log tHcy
Log tHcy
Log tHcy

Age
Weight
Height
BMI
Alcohol
Log tHcy
Log folate
Vitamin B6
Vitamin B12
Creatinine
Log folate
B6
B12

Pearson correlation
coefficent, r

0.064
0.453
0.237
0.370
0.083
0.130
0.132
0.006
0.014
0.011
0.440
0.154
0.278

0.249
>0.0001
>0.0001
>0.0001
0.137
0.033
0.025
0.919
0.820
0.843
<0.0001
0.010
<0.0001

N were 0.67 for the C allele and 0.33 for the T allele while both the
OPN group and the OP group had gene frequencies of 0.65 for the
C allele and 0.35 for the T allele. The distribution was compatible
with HardyWeinberg equilibrium (p = 0.59). Though there was a
trend of increasing frequency of the MTHFR TT genotype across
the groups, this was not significant (p = 0.601).
The OP group was significantly older, lighter and shorter
than the OPN group, and there was an increasing incidence of
both current smoking and recent fracture across the groups,
those in the OP group having significantly higher percentage of
both. There was no significant difference between the groups in
alcohol intake or plasma creatinine concentration.
Folate was significantly lower in the OP group then both the
other groups, and tHcy was significantly higher in the OP group
relative to OPN. The groups did not differ in vitamin B6 or
vitamin B12 status.
The bivariate associations of BMD and tHcy in the
population as a whole are shown in Table 2. Partial correlation
analysis, correcting for age, weight and height, confirmed the
significant association of folate with BMD (r = 0.124,

p = 0.042), but a similar analysis of tHcy with BMD,

additionally correcting for creatinine, reduced the significance
of this association (r = 0.117, p = 0.059, log tHcy) (Fig. 1).
Similarly, correction of BMD for age, weight and height, and
tHcy additionally for folate weakened the association (r =
0.075,p = 0.248). Correction of tHcy for folate, vitamin B6 and
vitamin B12 produced a non-significant association of tHcy with
BMD (r = 0.057,p = 0.382), confirming the dependence of
plasma tHcy on these determinants (Fig. 2).
Table 3 shows the distribution of tHcy, folate and BMD
according to MTHFR genotype. The TT genotype was
associated with an increased tHcy when compared with
the other genotypes (p = 0.05). Though there was a trend to
decreasing folate across the groups, this was not significant,
and there was no difference in BMD between the genotypes.
Lack of association of BMD with MTHFR genotype was
confirmed by a KruskalWallis test (chi-square 0.366,
p = 0.835).
Table 4 shows that compared with non-smokers, smokers
had significantly lower BMD, weight, folate and vitamin B6
concentrations, but no significant differences in tHcy or vitamin
B12 concentration.
Table 5 shows the effect of interaction of vitamin status with
MTHFR genotype on BMD. In those with the CT genotype,
those at the lowest quintile of folate, vitamin B6 and vitamin B12
concentrations have significantly higher tHcy concentrations,
which in the case of folate is associated with a significant
reduction in BMD.
Discussion
Homocysteine has been a candidate risk factor for
osteoporosis on the basis of observations linking homocystinuria and skeletal abnormality. In subjects with classical
homocystinuria, features including skeletal abnormality and
early-onset osteoporosis, as well as vascular complications,
have been observed [12]. With the economic and social burden

Fig. 1. Correlation of BMD with tHcy and folate.

M. Baines et al. / Bone 40 (2007) 730736

Fig. 2. Correlation of tHcy with folate, vitamin B6 and vitamin B12.

of OP rising with increasing life expectancy [16], the

identification of risk factors for OP, particularly if they are
modifiable, becomes more important.
In our population of British postmenopausal women, our
data showed a significant association of tHcy with BMD (r =
0.130, p = 0.033, log tHcy), though the association weakened
when corrections were made for age, weight, height and
creatinine ( r = 0.117, p = 0.059, log tHcy). The OP group had

733

a significantly higher tHcy than the OPN group (p = 0.003). The

association between tHcy and BMD has been controversial. van
Meurs et al. [2] did not find evidence of a association between
tHcy and BMD of femoral neck and lumbar spine, though there
was a strong relationship of tHcy with fracture incidence.
Cagnacci et al. [7] found that folate, but not tHcy, was
independently related to BMD of lumbar spine in Italian
women, whereas Golbahar et al. [10] did find a correlation
between tHcy and BMD of both femoral neck and lumbar spine
in 271 postmenopausal Iranian women. These differences may
therefore be due to genetic differences in the populations
studied, or the site of the BMD measurement. Our peripheral
BMD measurement at the os calcis differed from those studies
measuring at femoral neck or lumbar spine, but has been shown
to have comparable efficacy in predicting fracture in elderly
women [8]. Alternately, there may be variable confounding
factors in the populations studied. Our population, for example,
had a significantly higher % of current smokers in the OP group
than, for example, those studied by Cagnacci [7].
Reported associations of MTHFR genotype and BMD have
also been inconsistent. Positive associations were found in
postmenopausal Japanese women [4], and in postmenopausal
Danish women [5], but not in postmenopausal Iranian women
[8] or in postmenopausal or elderly Chinese women [13].
Villadsen et al. [14] found an association of MTHFR genotype
with osteoporotic vertebral fractures in 338 Danish osteoporotic
patients when compared with controls, but only a weak
association with lumbar spine BMD. McLean et al. [15],
using data from 1632 male and female subjects in the
Framingham Osteoporosis Study, did find an association
between MTHFR polymorphism and BMD, but suggested
that this was dependent upon folate status. Our data in Table 5
lend support to this view; among those with the CT genotype,
those in the lowest quintile for folate (less than 5.0 g/L), had a
significantly higher plasma tHcy and a significantly lower
BMD.
Folate status is known to be a determinant of plasma tHcy
concentration. Our data (Table 2) confirm a highly significant
correlation of folate with tHcy (r = 0.44, p < 0.0001). Both
dietary folate insufficiency, and variants of MTHFR can
contribute to a reduced 5-MTHF pool, leading to a hyperhomocysteinemia. Additionally, 5-MTHF deficiency can lead to
reduced methylation of DNA and consequent reduction of bone
synthesis and turnover. Folate deficiency may therefore have an
indirect effect on bone metabolism via hyperhomocysteinemia,
or a direct effect via other mechanisms. In our population, folate
had a significant positive correlation with BMD (r = 0.132,
Table 3
Distribution of homocysteine, folate and BMD between the MTHFR genotypes
MTHFR genotype
tHcy, mol/L
Folate, g/L
BMD

p (ANOVA)

CC, n = 122

CT, n = 135

TT, n = 32

11.97 (4.88)
9.06 (4.66)
1.00 (1.27)

12.00 (4.80)
8.82 (4.43)
0.87 (1.27)

14.49 (8.88)
7.95 (4.05)
1.02 (1.22)

Data given as mean (SD).

0.050
0.479
0.637

734

M. Baines et al. / Bone 40 (2007) 730736

Table 4
Differences in some measured analytes between smokers and non-smokers

Smokers, n = 57
Non-smokers, n = 267
p

BMD

Weight, Kg

tHcy, mol/L

Folate, g/L

Vitamin B6, nmol/L

Vitamin B12, pmol/L

1.50 (0.94)
0.80 (1.290
<0.001

62.59 (10.86)
66.10 (12.30)
0.042

11.46 (4.25)
12.47 (5.69)
0.327

7.89 (4.71)
9.02 (4.46)
0.030

39.6 (26.1)
44.8 (23.3)
0.038

255.1 (95.1)
257.6 (94.6)
0.948

Data presented as mean (SD).

p = 0.025, log folate) which was maintained after correction for

age, weight and height (r = 0.124, p = 0.042), suggesting that
folate may be an independent risk factor for osteoporosis. As
with tHcy, studies of the link of folate with osteoporosis risk
have been inconsistent. Several studies [7,10,15] have shown a
significant positive correlation of folate with BMD, whereas a
recent study of 1550 older American subjects concluded that
neither serum nor red blood cell folate was related to BMD or
osteoporosis [17], possibly because of a higher mean age and
higher mean plasma folate concentration. None of these studies,
including our own, investigated populations who were subject
to grain-enhancement with folic acid, the 2 USA-based studies
using data obtained before mandatory enrichment began in
1998, and most had population mean folate concentrations of at
least 7.6 g/L, the exception being the study of Iranian women
where the mean folate concentration was 5.1 g/L [8]. Whilst
the mean age of the population studied by Morris et al. [17] was
68 years, significantly higher than the Italian, Iranian or
Framingham populations, it was almost identical to the
population mean in our own study. As the mean folate
concentrations were comparable between our study and that
of Morris et al. [17], the observed difference in correlation
between folate and BMD may be due to population mix (ours
was all women whereas that of Morris was both women and
men), or site of BMD measurement.
Vitamin B12, a cofactor for methionine synthase in the
remethylation of tHcy to methionine, is another determinant of
tHcy status. We confirmed the strong association of tHcy with
vitamin B12 (r = 0.278, p < 0.0001, log tHcy), but were unable
to show any significant correlation with BMD (r = 0.014,
p = 0.820). These findings are in agreement with those of
Cagnacci et al. [7] and Macdonald et al. [18]. In contrast, Morris
et al. [17] found that serum vitamin B12 (up to 220 pmol/L) was
related to BMD in a dose-related manner in their 1550 older
Americans enrolled in the third US National Health and
Nutrition Examination Survey, but that increases in serum
vitamin B12 beyond 220 pmol/L were not so associated. Tucker
et al. [19] found that both men and women with vitamin B12
concentrations below 148 pmol/L had lower average BMD than
those with higher concentrations, but with significance varying
with site of BMD measurement (men at most hip sites, women
at the spine). Dhonukshe-Rutten et al. [20] found that elderly
Dutch women, but not men, who were classified as having a
marginal serum vitamin B12 status ( 211319 pmol/L) or
deficient vitamin B12 status (< or = 210 pmol/L) had an
increased prevalence odds ratio for osteoporosis, and Stone et
al. [21] found that in white women aged over 65 y, those that
had a serum vitamin B12 concentration below 207 pmol/L had a

higher annual rate of reduction in total hip BMD, but not

calcaneal BMD, than those with higher concentrations. Thus
there appears to be some evidence for a threshold effect of
vitamin B12 on BMD, with serum concentrations of less than
200220 pmol/L being associated with decreased BMD at some
sites. However, our data (Table 5) failed to show any significant
change in BMD between Vitamin B12 quintiles. Difference in
findings of the association of serum vitamin B12 with BMD may
therefore be explained by population demographic differences,
and, more particularly, the site of assessment of BMD.
Vitamin B6, as pyridoxal-5-phosphate, acts as a cofactor to
two enzymes of the trans-sulfuration pathway of tHcy,
cystathionine -synthase and cystathionase, and as such is
another determinant of tHcy status. As with vitamin B12, we
were able to confirm a significant association of tHcy with
vitamin B6 (r = 0.154, p = 0.010, log tHcy), but no significant
association of vitamin B6 with BMD (r = 0.006, p = 0.919). To
our knowledge, ours is the only study to measure serum vitamin
B6 as a possible risk factor for osteoporosis, two other recent
studies [18,27] using dietary records to conclude that no
significant association of BMD with vitamin B6 could be
established. Whilst vitamin B6 may have an indirect effect on
BMD through tHcy, it seems unlikely that it has a direct effect
on BMD despite speculation regarding tHcy-independent
pathways [28].
Our population had a significantly higher % of current
smokers in the OP group. Smoking has been suggested to reduce
BMD in elderly women in a dose-related manner, possibly as a
result of decreased calcium absorption and consequent secondary hyperparathyroidism [22]. Our results are strongly supportive of this suggestion, showing a highly significant reduction in
BMD in the smokers, compared with non-smokers, though the
lower weight in smokers may be a contributor here. The relation
between smoking and tHcy remains largely unresolved [2326].
The findings in our population of postmenopausal women is that
smoking is associated with significantly reduced plasma
concentrations of folate and vitamin B6, but not vitamin B12,
though these changes did not lead to an increase in tHcy. Perhaps
surprisingly, the mean plasma tHcy concentration was 1 mol/L
lower in the smoking group, though this was not significant.
Thus smoking appears to be a major determinant of BMD,
though this is not modulated through effects on tHcy.
If tHcy is toxic to bone or bone formation, by what
mechanism does this occur? In studies in patients with
homocystinuria, the early onset OP was considered to be due
to impaired cross-linking of collagen, and not a deficiency in its
synthesis [29]. Later in vitro studies have shown that
homocysteine-thiolactone, a natural metabolite of tHcy present

M. Baines et al. / Bone 40 (2007) 730736

Table 5
Distribution of tHcy and BMD in MTHFR genotypes, divided into tertiles (TT
genotype) or quintiles (CT and CC genotypes) of folate, vitamin B6 and vitamin
B12 concentration
tHcy mol/L
Folate
TT T1
T2
T3
CT Q1
Q2
Q3
Q4
Q5
CC Q1
Q2
Q3
Q4
Q5

18.80 (14.91)
13.05 (6.69)
13.08 (4.85)
16.21 (7.25)**
12.78 (1.71)**
11.33 (3.02)**
11.11 (3.37)**
8.22 (2.33)
14.31 (5.53)*
11.59 (3.47)
12.38 (4.27)
12.07 (4.74)
9.95 (4.51)

Vitamin B6
TT T1
13.16 (5.87)
T2
12.18 (5.95)
T3
15.30 (4.73)
CT Q1
14.14 (6.37)**
Q2
12.44 (3.65)
Q3
11.98 (4.93)
Q4
11.74 (3.84)
Q5
9.64 (2.72)
CC Q1
13.49 (5.38)
Q2
10.99 (5.48)
Q3
12.52 (3.88)
Q4
12.64 (6.10)
Q5
11.02 (4.17)
Vitamin B12
TT T1
21.74 (14.55)
T2
11.22 (4.88)
T3
12.83 (4.81)
CT Q1
14.31 (4.69)*
Q2
12.25 (3.71)
Q3
11.20 (2.72)
Q4
11.50 (7.48)
Q5
10.47 (3.52)
CC Q1
14.87 (5.03)
Q2
13.48 (4.43)
Q3
9.96 (2.56)
Q4
11.45 (4.77)
Q5
11.70 (5.30)

Group
ANOVA, p
0.387

<0.0001

0.059

0.471

0.008

0.374

0.055

0.064

0.007

BMD

Group
ANOVA, p

0.525 (1.559)
0.756 (1.217)
1.278 (1.215)
1.652 (1.076)*
0.677 (1.095)
0.684 (1.452)
0.688 (1.557)
0.600 (1.253)
0.892 (1.297)
1.470 (1.017)
0.739 (1.177)
1.038 (1.548)
0.561 (1.360)

0.495

0.900 (0.990)
1.044 (1.389)
0.800 (1.678)
0.785 (1.921)
0.892 (1.049)
0.770 (1.325)
0.807 (1.162)
0.914 (1.131)
1.028 (1.485)
0.713 (1.335)
0.900 (1.048)
1.272 (1.560)
1.131 (1.139)
0.363 (1.342)
0.600 (1.722)
1.178 (0.573)
0.956 (1.448)
0.596 (1.152)
0.636 (1.659)
1.285 (1.177)
0.712 (1.271)
1.039 (1.455)
1.291 (1.466)
0.459 (1.253)
1.026 (1.276)
1.067 (0.989)

0.023

0.170

0.928

0.993

0.635

735

al. to speculate that homocysteine interferes with the development of the microarchitecture of bone independent of the
amount of mineral in the bone [2] and would account for the
lack of firm association between plasma tHcy concentrations
and BMD. Recently, Herrmann et al. have shown that increased
tHcy concentrations stimulate osteoclast activity in vitro, and
may therefore stimulate increased bone resorption in vivo [33].
By whatever mechanism, determinants of tHcy, by influencing
circulating tHcy concentrations, may have a deleterious effect
on bone health.
Whilst the etiology of OP is acknowledged to be multifactorial, if easily modifiable risk factors can be identified this
would offer the possibility of population-based preventative
measures. This study has shown that in our population of
postmenopausal British women, the risk of OP was associated
with a reduced serum folate concentration and an increased
plasma tHcy concentration. Additionally, those with lower
circulating concentrations of vitamin B6 or vitamin B12, or
having the MTHFR TT genotype were likely to have higher
tHcy concentrations, and adverse effects on the bone. Smoking
also has a detrimental effect on BMD, which may in part be
modulated by reductions in circulating folate and vitamin B6
concentrations. What is not known is whether supplementation
with folate, vitamin B6 or vitamin B12 would reduce the rate of
loss of BMD, or fracture risk. The optimization of B-vitamin
status within a clinical trial may be one area where further effort
may be directed.
Acknowledgments

0.420

0.332

0.294

Vitamin concentrations:
Folate, g/L TT, T1 2.05.8:T2 5.98.7: T3 8.820.0.
CT, Q1 1.75.0: Q2 5.16.7: Q3 6.88.8: Q4 8.911.2: Q5 11.420.0
CC, Q1 2.45.5; Q2 5.66.8: Q3 6.98.9: Q4 9.012.7: Q5 12.820.3
Vitamin B6, TT,T1 838: T2 3957: T3 58109
Ct, Q1 1032: Q2 2332: Q3 3339: Q4 4055: Q5 56143
nmol/L
CC, Q1 725: Q2 2634: Q3 3544: Q4 4558: Q5 59118
Vitamin B12 TT,T1 80211: T2 230320: T3 322455
CT,Q1112183: Q2 184213: Q3 214259: Q4 260334: Q5 335529
pmol/L
CC, Q1 52175: Q2 176213: Q3 214264: Q4 265319: Q5 320492
Significance levels, against highest tertile/quintile, *p < 0.01, **p < 0.001. Data
given as mean (SD).

in nanomolar concentrations in human plasma [30], inhibits

lysyl oxidase, an enzyme involved in collagen cross-linking
[31,32]. These and other related observations lead van Meurs et

We acknowledge with thanks a grant from REMEDI, which

supported aspects of this work.
References
[1] Schimke RN, McKusick VA, Huang T, Pollock AD. Homocystinuria.
Studies of 20 families with 38 affected members. JAMA 1965;
193:8795.
[2] van Meurs JBJ, Dhonukshe-Rutten RAM, Pluijm SMF, van der Klift M, de
Jonge R, et al. Homocysteine levels and the risk of osteoporotic fracture. N
Engl J Med 2004;350:203341.
[3] McLean RR, Jacques PF, Selhub J, Tucker KL, Samelson EJ, et al.
Homocysteine as a predictive factor for hip fracture in older persons. N
Engl J Med 2004;350:20429.
[4] Miyao M, Morita H, Hosai T, Kurihara H, Inoue S, et al. Association of
methylenetetrehydrofolate reductase (MTHFR) polymorphism with bone
mineral density in postmenopausal Japanese women. Calcif Tissue Int
2000;66:1904.
[5] Abrahamsen B, Madsen JS, Tofteng CL, Stilgren L, Bladjberg EM. A
common methylenetetrahydrofolate reductase (C677T) polymorphism is
associated with low bone mineral density and increased fracture incidence
after menopause: longitudinal data from the Danish Osteoporosis
Prevention Study. J Bone Miner Res 2003;18:7239.
[6] Jorgenssn HL, Madsen JS, Madsen B, Saleh MM, Abrahamsen B, Fenger
M, et al. Association of a common allelic polymorphism (C677T) in the
methylenetetrahydrofolate reductase gene with a reduced risk of
osteoporotic fractures. A case control study in Danish postmenopausal
women. Calcif Tissue Int 2002;71:38692.
[7] Cagnacci A, Baldassari F, Rivolta G, Arangino S, Volpe A. Relation of
homocysteine, folate, and vitamin B12 to bone mineral density of
postmenopausal women. Bone 2003;33:9569.

736

M. Baines et al. / Bone 40 (2007) 730736

[8] Barr RJ, Adebajo A, Fraser WD, Halsey JP, Kelsey C, Stewart A, et
al. Can peripheral DXA measurements be used to predict fractures in
elderly women living in the community? Osteoporos Int 2005;16:
117783.
[9] Zhang M, Gunter EW, Pfeiffer CM. Evaluation of the Drew Scientific
DS30 homocysteine assay in comparison with the Centers for Disease
Control and Prevention Reference HPLC method. Clin Chem 2001;47:
9667.
[10] Golbahar J, Hamidi A, Aminzadeh MA, Omrani GR. Association of plasma
folate, plasma total homocysteine, but not methylenetetrahydrofolate
reductase C677T polymorphism, with bone mineral density in
postmenopausal Iranian women: a cross-sectional study. Bone 2004;35:
7605.
[11] von Ahsen N, Oellerisch M, Schutz E. A method for homogenous
color-compensated genotyping of Factor V (G1691A) and
methylenetetrahydrofolate reductase (C677T) mutations using realtime multiplex fluorescence PCR. Clin Biochem 2000;33:5359.
[12] Mudd SH, Skovby F, Levy HL, Pettigrew KD, Wilcken B, et al. The
natural history of homocystinuria due to cystathione beta-synthase
deficiency. Am J Hum Genet 1984;37:131.
[13] Lim M, Lau EMC, Woo J. Methylenetetrahydrofolate reductase
polymorphism (MTHFR C677T) and bone mineral density in Chinese
men and women. Bone 2004;35:136974.
[14] Villadsen MM, Bunger MH, Carstens M, Stenkjaer L, Langdahl BL.
Methyleneterrahydrofolate reductase (MTHFR) C677T polymorphism is
associated with osteoporotic vertebral fractures, but is a weak predictor of
BMD. Osteoporos Int 2005;16:4116.
[15] McLean RR, Karasik D, Selhub J, Tucker KL, Ordovas JM, et al.
Association of a common polymorphism in the Methylenetetrahydrofolate
reductase (MTHFR) gene with bone phenotypes depends on plasma folate
status. J Bone Miner Res 2004;19:4108.
[16] Fraser WD. The burden of osteoporosis and the case for disease
management. Dis Manag Health Outcomes 2004;12:40918.
[17] Morris MS, Jacques PF, Selhub J. Relation between homocysteine and Bvitamin status indicators and bone mineral density in older Americans.
Bone 2005;37:23442.
[18] Macdonald HM, McGuigan FE, Fraser WD, New SA, Ralston SH, Reid
DM. Methylenetetrahydrofolate reductase polymorphism interacts with
riboflavin intake to influence bone mineral density. Bone 2004;35:95764.
[19] Tucker KL, Hannan MT, Qiao N, Jacques PF, Selhub J, et al. Low plasma
vitamin B12 is associated with lower BMD: The Framingham Osteoporosis
Study. J Bone Miner Res 2005;20:1528.

[20] Dhonukshe-Rutten RAM, Lips M, de Jong N, Chin A, Paw MJM, Hiddink

GJ, et al. Vitamin B-12 status is associated with bone mineral content and
bone mineral density in frail elderly women but not in men. J Nutr
2003;133:8017.
[21] Stone KL, Bauer DC, Sellmeyer D, Cummings SR. Low serum vitamin B12 levels are associated with increased hip bone loss in older women: a
prospective study. J Clin Endocrinol Metab 2004;89:121721.
[22] Rapuri PB, Gallagher JC, Balhorn KE, Rayschon KL. Smoking and bone
metabolism in elderly women. Bone 2000;27:42936.
[23] Pagan K, Hou J, Goldenberg RL, Cliver SP, Tamura T. Effect of smoking
on serum concentrations of total homocysteine and B vitamins in midpregnancy. Clin Chim Acta 2001;306:1039.
[24] Sobszak A. The effects of tobacco smoke on the homocysteine levelA
risk factor of atherosclerosis. Addict Biol 2003;8:14758.
[25] Sobszak A, Wardas W, Zielinska-Danch W, Pawlicki K. The influence of
smoking on plasma homocysteine and cysteine levels in passive and active
smokers. Clin Chem Lab Med 2004;42:40814.
[26] Gabriel HE, Crott JW, Ghandour H, Dallal GE, Choi S-W, Keyes MK, et
al. Chronic cigarette smoking is associated with diminished folate status,
altered folate form distribution, and increased genetic damage in the buccal
mucosa of healthy adults. Am J Clin Nutr 2006;83:83541.
[27] Abrahamsen B, Madsen JS, Tofteng CL, Stilgren L, Bladbjerg EM, et al.
Are effects of MTHFR (C677T) genotype on BMD confined to women
with low folate and riboflavin intake ? Analysis of food records from the
Danish osteoporosis prevention study. Bone 2005;36:57783.
[28] Hermann M, Widmann T, Hermann W. HomocysteineA newly
recognised risk factor for osteoporosis. Clin Chem Lab Med 2005;43:
11117.
[29] Lubec B, Fang-Kircher S, Lubec T, Blom HJ, Boers GHJ. Evidence for
McKuisik's hypothesis of deficient collagen cross-linking in patients with
homocystinuria. Biochim Biophys Acta 1996;1315:15962.
[30] Chwatko G, Jakubowski H. The determination of homocysteine
thiolactone in human plasma. Anal Biochem 2005;337:2717.
[31] Liu G, Nellaiappan K, Kagan HM. Irreversible inhibition of lysyl oxidase
by homocysteine thiolactone and its selenium and oxygen analogues.
Implications for homocystinuria. J Biol Chem 1997;272:323707.
[32] Raposo B, Rodriguez C, Martinez-Gonzalez J, Badimon L. High levels of
homocysteine inhibit lysyl oxidase (LOX) and downregulate LOX
expression in vascular endothelial cells. Atherosclerosis 2004;177:18.
[33] Herrmann M, Widmann T, Colaianna G, Colucci S, Zallone A, Herrman
W. Increased osteoclast activity in the presence of increased homocysteine
concentrations. Clin Chem 2005;51:234853.