Download Homocysteine Metabolism and Risk of Myocardial Infarction

American Journal of
EPIDEMIOLOGY
Volume 143
Copyright © 1996 by The Johns Hopkins University
Number 9
School of Hygiene and Public Health
May 1, 1996
Sponsored by the Society for Epldemlologlc Research
ORIGINAL CONTRIBUTIONS
Homocysteine Metabolism and Risk of Myocardial Infarction: Relation with
Vitamins Bfi, B i9 , and Folate
Petra Vertioef,1 Meir J. Stampfer, 23 Julie E. Buring,4'5 J. Michael Gaziano,4 Robert H. Allen,6
Sally P. Stabler,6 Robert D. Reynolds,7 Frans J. Kok,1 Charles H. Hennekens,4'5'8 and Walter C. Willett2-3
Elevated plasma homocyst(e)ine levels are an independent risk factor for vascular disease. In a case-control
study, the authors studied the associations of fasting plasma homocyst(e)ine and vitamins, which are
important cofactors in homocysteine metabolism, with the risk of myocardial infarction. The cases were 130
Boston area patients hospitalized with a first myocardial infarction and 118 population controls, less than 76
years of age, enrolled in 1982 and 1983. Dietary intakes of vitamins B6, B12, and folate were estimated from
a food frequency questionnaire. After adjusting for sex and age, the authors found that the geometric mean
plasma homocyst(e)ine level was 11 % higher in cases compared with controls (p = 0.006). There was no clear
excess of cases with extremely elevated levels. The age- and sex-adjusted odds ratio for each 3-/xmol/liter
(approximately 1 standard deviation) increase in plasma homocyst(e)ine was 1.35 (95% confidence interval
1.05-1.75; p trend = 0.007). After further control for several other risk factors, the odds ratio was not affected,
but the confidence interval was wider and the p value for trend was less significant. Dietary and plasma levels
of vitamin B 6 and folate were lower in cases than in controls, and these vitamins were inversely associated with
the risk of myocardial infarction, independently of other potential risk factors. Vitamin B12 showed no clear
association with myocardial infarction, although methylmalonic acid levels were significantly higher in cases.
Comparing the mean levels of several homocysteine metabolites among cases and controls, the authors found
that impairment of remethylation of homocyst(e)ine (dependent on folate and vitamin B 12 rather than on vitamin
Be-dependent transsulfuration) was the predominant cause of high homocyst(e)ine levels in cases. Accordingly, plasma folate and, to a lesser extent, plasma vitamin B 12 , but not plasma vitamin B6, correlated inversely
with plasma homocyst(e)ine, even for concentrations at the high end of normal values. These data provide
further evidence that plasma homomocyst(e)ine is an independent risk factor for myocardial infarction. In this
population, folate was the most important determinant of plasma homocyst(e)ine, even in subjects with
apparently adequate nutritional status of this vitamin. Am J Epidemiol 1996; 143:845-59.
folic acid; homocysteine; myocardial infarction; pyridoxine; vitamin B 12
Received for publication May 17, 1995, and in final form February 12, 1996.
Abbreviations: Cl, confidence interval; SE, standard error.
1
Department of Epidemiology and Public Health, Agricultural
University, Wageningen, Netherlands.
2
Channing Laboratory, Department of Medicine, Brigham and
Women's Hospital and Harvard Medical School, Boston, MA.
3
Departments of Epidemiology and Nutrition, Harvard School of
Public Health, Boston, MA.
4
Division of Preventive Medicine, Department of Medicine,
Brigham and Women's Hospital and Harvard Medical School,
Boston, MA.
5
Division of Ambulatory Care and Prevention, Brigham and
Women's Hospital and Harvard Medical School, Boston, MA
8
Department of Medicine, Division of Hematology, University of
Colorado Health Sciences Center, Denver, CO.
7
Department of Human Nutrition and Dietetics, University of
Illinois, Chicago, IL
8
Department of Epidemiology, Harvard School of Public Health,
Boston, M A
Reprint requests to Dr. M. J. Stampfer, Channing Laboratory,
180 Longwood Ave., Boston, MA 02115.
845
846
Vertioef et al.
Extremely elevated plasma levels of total homocysteine (generally referred to as homocyst(e)ine, the sum
of all homocysteine species in plasma including free
and protein-bound forms) are present in patients homozygous for genetic defects in enzymes of homocysteine metabolism, which may lead to premature
vascular disease. Plasma homocyst(e)ine can be moderately elevated in individuals heterozygous for these
defects or with inadequate intake of vitamins B 6 , B 12 ,
and folate, which serve as cofactors in the enzymatic
pathways of homocysteine metabolism (1, 2).
Homocysteine is a thiol-containing amino acid
derived from methionine metabolism that can be
degraded through two enzymatic pathways: transsulfuration and remethylation (figure 1). In the transsulfuration pathway, homocysteine is condensed with
serine to form cystathionine, an irreversible reaction
dependent on pyridoxal 5'-phosphate, the active form
of vitamin B 6 . Subsequently, cystathionine is converted to cysteine in another vitamin B6-dependent
reaction. In remethylation, homocysteine receives a
methyl group from N-S-methyltetrahydrofolate (a vitamin B12-dependent reaction) or from betaine to form
methionine. In the reaction involving betaine, which is
independent of folate and vitamin B 12 , Af,iV-dimethylglycine is formed. The methionine formed is activated
by adenosine triphosphate to form 5-adenosylmethionine, which serves as a methyl donor to a variety of
acceptors, one being glycine, leading to the formation
of yV-methylglycine. 5-Adenosylhomocysteine is
formed in a transmethylation reaction, which is then
hydrolyzed in homocysteine and adenosine (3, 4).
Even moderately elevated plasma homocyst(e)ine
levels appear to be an independent risk factor for
coronary artery disease (5-7). One prospective study
reported a 3.4-fold excess risk of myocardial infarction for men with levels above 15.8 /Ltmol/liter, the
95th percentile for controls in that population, as compared with those with levels below the 90th percentile
(8). A recent prospective study confirmed the association and also suggested a graded association rather
than a threshold effect (9). Suggested causative mech-
Polyamines
Acceptor
S-AdenosylMethioninc
(SAM)
CH3-Acceptor
ATP
-Proteins
Dimethylglycine
S-AdenosylHomocysteine
Choline
'
^ - * " Methionine
3
Serine
THF
•Glycine
MethyleneTHF
Adenosine
Cystathionine
PLP
a-Ketobutyrate •*-^
—•• Cysteine
I
Taurine
SO,
FIGURE 1. Homocyst(e)ine metabolism in humans and animals. Enzymes: 1, /V-5-methyttetrahydrofoIate:homocysteine methyttransferase;
2, methylenetetrahydrofolate reductase; 3, betalne:homocysteine methyttransferase; 4, choline dehydrogenase; 5, cystathionine 0-synthase;
6, 6-cystathionase. THF, tetrahydrofolate; PLP, pyridoxal 5'-phosphate; ATP, adenosine 5'-triphosphate; B12, vitamin B 12 . (Adapted from J.
Selhub and J. W. Miller. Am J Clin Nutr 1992;55:131-8).
Am J Epidemiol
Vol. 143, No. 9, 1996
Homocyst(e)ine, Vitamins, and Myocardial Infarction
anisms include increased uptake of low density lipoprotein cholesterol in the vascular wall, promotion of
vascular smooth muscle cell growth, and effects on
vascular coagulant mechanisms (2).
Several studies have demonstrated inverse associations of plasma homocyst(e)ine levels with blood levels of vitamins B 6 , B 12 , and folate and with intake
levels of vitamin B 6 and folate (10, 11). Also, in
human experimental studies, supplements of these vitamins (especially folate) reduce elevated homocyst(e)ine levels (12, 13). Some, but not all, epidemiologic studies have found inverse associations of the
vitamins with vascular disease (5, 14).
In the present case-control study, we investigated
the associations of plasma homocyst(e)ine, dietary and
plasma levels of vitamins B 6 , B 12 , and folate, and
dietary methionine with the risk of first myocardial
infarction. Also, we studied associations between
plasma homocyst(e)ine and the vitamins. Furthermore,
we compared plasma levels of several of the homocysteine derivatives shown in figure 1 between cases
of first myocardial infarction and control subjects to
study enzymatic disturbances in the pathways of homocysteine metabolism, either of nutritional or genetic
origin (15).
MATERIALS AND METHODS
Study population
Cases were white men and women less than 76
years of age with no history of previous myocardial
infarction or angina and living in the Boston area. All
admissions to the coronary or intensive care units of
six suburban hospitals between January 1, 1982, and
December 31, 1983, were reviewed to identify eligible
cases. The diagnosis of myocardial infarction was
confirmed from the hospital record, based on clinical
history and creatinine kinase rise (16). Informed consent was obtained from the patients after obtaining
permission from the admitting physician. A total of
450 cases were eligible for the study. Of these, nine
were not invited to participate in the study for lack of
physician consent, five could not be contacted after
discharge, and 70 (16 percent of those invited to
participate in the study) refused cooperation, leaving
366 potential cases. For each case, a control of the
same age (within 5 years) and sex was selected at
random from the residents' list of the town in which
the case resided. Controls were ineligible if they had
previous myocardial infarction or angina. Of 741 potential controls, 423 (57 percent) were willing to participate. For 340 cases, controls could be matched with
respect to age and sex, leaving 340 case-control pairs
(267 men in each group). Of these, sufficient plasma
Am J Epidemiol
Vol. 143, No. 9, 1996
847
was available for complete information on 130 cases
and 118 controls (see below).
Study measurements
Eligible and willing subjects were visited in their
homes by one of two nurses approximately 8 weeks
after hospital discharge (cases) or at about the same
time as the matching case (controls). These nurses
obtained fasting venous blood samples, and information on dietary intake and coronary risk factors related
specifically to the time period before the infarction for
the cases and before the interview for the controls was
obtained. The information included history of diabetes, high blood pressure, and high cholesterol; parental
history of ischemic heart disease at or before age 60;
weight and height (both self reported); physical activity; cigarette smoking; alcohol consumption; and diet.
A physical activity index, expressed in kilocalories per
week, was obtained by summing stairs climbed, blocks
walked, and recreation and leisure time activity (17).
Body mass index was calculated as the weight (kg)/
height (m)2.
Blood was drawn into 0.1 percent ethylenediaminetetraacetic acid-containing Vacutainer tubes
(Becton, Dickinson & Company, East Rutherford,
New Jersey), and fresh plasma was used to determine
total cholesterol and high density and low density
lipoprotein cholesterol, as previously described (16).
The rest of the plasma was stored at — 70°C, with
similar storage time for cases and controls. In 1986,
plasma pyridoxal 5'-phosphate was determined radiometrically by stimulation of tyrosine apodecarboxylase in 275 cases and 281 controls of the original study
population (including 125 cases and 117 controls of
the subsample described herein) (18). In 1993, we
assayed plasma samples for homocyst(e)ine, several of
its derivatives (cystathionine, cysteine, methionine,
iV, N-dimethylglycine, Af-methylglycine, serine, and
glycine), vitamin B, 2 , and folate, as previously described (15). A subsample of 130 cases (97 men, 33
women) and 118 control subjects (81 men, 33 women)
had a sufficient plasma volume left for these assays. In
addition to the homocysteine-related metabolites, we
measured methylmalonic acid and 2-methylcitric acids
I and n, which are usually elevated in subjects with
vitamin B 1 2 deficiency (15).
We had information on dietary intakes of methionine, the main metabolic precursor of homocysteine,
and vitamins B 6 , B 12 , and folate for both the subsample of 130 cases and 118 controls with plasma
assays and the original study population of 340 cases
and 339 controls (one female control subject was
excluded because of unreliable dietary data). Information on diet was collected using a semiquantitative
848
Verhoef et al.
food frequency questionnaire, which was an extended
and refined version of a previously validated questionnaire (19-22). The questionnaire included 116 food
items plus vitamin supplements. For each food, a
commonly used unit or portion size was specified, and
participants were asked how often on average over the
previous year they had consumed that amount. Nine
responses were possible, ranging from "less than one
time per month" to " six or more times per day." The
intakes of methionine, vitamin B 6 , vitamin B 12 , folate,
and other nutrients were computed by multiplying the
frequency of consumption of each unit of food by the
nutrient content of the specified portions. Composition
values of foods were primarily obtained from US
Department of Agriculture sources (23).
Statistical analysis
Mean values and proportions of various risk factors
for coronary disease were compared for 130 cases of
first myocardial infarction and 118 population controls
and tested for significance using Student's / test for
continuous variables and Pearson's x1 test for proportions. We also studied associations of homocyst(e)ine
levels with plasma and dietary variables and with risk
factors for coronary disease, controlling for age and
sex by calculating correlation coefficients between
residuals obtained from linear regression analyses of
age and sex on homocyst(e)ine and of age on sex on
the respective risk factor or vitamin level.
Mean plasma levels of homocyst(e)ine and metabolites, as well as plasma and dietary levels of vitamin
B 6 , vitamin B 12 , and folate, were compared for cases
and controls. The mean dietary intake of methionine
was compared as well. All distributions were skewed
to the right, so we used log-transformation to normalize them. After log-transformation, dietary intakes
were adjusted for total energy intake, using regression
analysis (24). We also show geometric means of cases
and controls, and p values are shown for the casecontrol difference in geometric means, adjusting for
age and sex differences between the groups by means
of linear regression analysis.
We used logistic regression analysis to study the
association of plasma homocyst(e)ine with the risk of
myocardial infarction. We calculated odds ratios plus
95 percent confidence intervals for both plasma homocyst(e)ine as a continuous variable (per 1 standard
deviation increase) and by quintiles (defined according
to the distribution among controls, with the lowest
quintile as a reference category). Multiple logistic
regression analysis was used to control for sex, age,
and several other coronary risk factors, some of which
were significantly associated with homocyst(e)ine.
Tests for trend were performed by adding log-
transformed plasma homocyst(e)ine to the logistic regression models continuously. In the same way as for
plasma homocyst(e)ine, we assessed the relations of
dietary intakes of methionine and dietary and plasma
levels of vitamins B 6 , B 12 , and folate with myocardial
infarction. Furthermore, we repeated the risk analyses
for plasma homocyst(e)ine as a continuous variable
(per 1 standard deviation) in strata of several risk
factors of coronary heart disease: age, smoking, history of high blood pressure, and history of hypercholesterolemia. All reported p values are two tailed.
RESULTS
Risk factors for coronary heart disease and
associations with homocyst(e)ine
Table 1 shows differences in major risk factors of
coronary disease between 130 cases and 118 controls.
Differences were mostly as expected, with cases having a significantly higher mean level of plasma total/
high density lipoprotein cholesterol, a higher mean
intake of saturated fat, a lower mean alcohol intake,
and a higher prevalence of hypercholesterolemia and
diabetes. Also, among cases there were significantly
more current smokers. Overall, the mean levels and
proportions were similar to those observed in the original study population of 340 cases and 339 controls
(data not shown), except for history of high blood
pressure and family history of myocardial infarction.
These were significantly lower in the larger group of
controls as compared with cases, whereas in the subset
controls had relatively high prevalences of a history of
high blood pressure and family history of myocardial
infarction.
Associations of plasma homocyst(e)ine with risk
factors of coronary disease and other possible confounding factors were evaluated, controlling for sex
and age (table 1). Correlations among cases and controls differed substantially. Homocyst(e)ine levels
were lower in women and directly associated with age,
especially in cases. In control subjects, we found direct
associations of homocyst(e)ine levels with alcohol
consumption and with a history of hypercholesterolemia. In cases, but not in controls, a history of high
blood pressure correlated directly with homocyst(e)ine.
Mean levels of homocyst(e)ine, metabolites,
vitamins, and methionine intake
Mean and geometric mean plasma levels of homocyst(e)ine and several metabolites are shown in table
2. The geometric mean of plasma homocyst(e)ine was
11 percent higher in cases than in control subjects
(p = 0.006). Homocyst(e)ine values of cases were
Am J Epidemiol
Vol. 143, No. 9, 1996
CO
p
o
*
t
$
§
II
0.01
45.4
25.6
0.001
0.13
0.10
r
0.15
0.27
J[
SmoMngt (current smokers)
0.26
-0.19
2,404 (843)
r
2,292 (712)
Mean
%
%
-0.16
-0.13
0.07
0.15
15.4
10.2
0.95
0.04
0.22
-0.09
-0.07
r
0.07
-0.15
-0.02
r
0.31
0.44
P
value
history oi diabetes
38.0(17.9)
33.2 (13.6)
Mean
42.6
43.2
%
Mean
3,005(3,111)
3,369 (2,947)
0.93
0.19
0.06
r
0.71
0.02
0.04
r
0.83
0.63
0.03
0.51
P
value
0.61
0.28
16.9
9.4
%
0.08
-0.04
0.22
r
Mean
0.62
0.02
16.9
19.5
%
Farrt
6.41 (1.82)
5.18(1.83)
Mean
p
value
0.82
0.0002
P
value
0.25
0.06
P
value
0.60
-0.19
-O.06
r
0.03
0.53
P
value
y history of rrlyocarcSaJ
0.0001
0.17
0.11
r
Total/HDL * cholesterol ratio
0.08
0.02
0.33
r
Alcohol consumption (g/day)
14.2(24.2)
20.0 (28.4)
history of high cholesterol
0.35
-0.05
-0.10
r
P
value
Physical adMty (kcal/week)
26.0 (4.3)
26.2 (4.6)
Mean
P
value
Body mass Index (kg/(m)i)
history of hypertension
0.09
0.81
P
value
Saturated tat Intake (g/day)
74.6
68.6
Energy Intake (kcaVday)
P
value
0.03
0.38
r
P
value
0.30
0.19
0.08
male
Sex
Matching factor
57.7 (9.3)§
57.7 (9.1)
r$
P
value
HDL, high density lipoprotein.
For calculation of the Pearson correlation coefficient, smoking was coded as follows: 1 =never, 2 = exsmokers, 3 = <1 pack/day, 4 - 1-2 packs/day, and 5 a >2 packs/day.
Pearson's correlation with homocyst(e)ine, adjusted for age and sex.
Numbers in parentheses, standard deviation.
p value of difference in risk factor between cases and controls.
p value for case-control
difference
Cases
Controls
p value for case-control
difference
Cases
Controls
p value for case-control
dfferencell
Cases
Controls
Mean
Age (years)
TABLE 1. Cardiovascular risk factors In 130 cases of myocardlal infarction and 118 controls and correlations of risk factors with plasma homocyst(e)ine, Boston Area
Health Study, 1982-1983
1
<;
CD
.2.
i
o
Lj
d.
ns, and
ial Infarction
850
Verhoef et al.
•a
88
g
§
oi c\i
1O
3
1
d
d
a
CO
CO
i - 00
•5
T3
73
<d
CD
CO
CO S
° ST
eg
8
q
d
in in
I
OJ
II
q
d
2-2-
I
8
•o
c
a
o
d
o
*
e
o
§
0
g
CO
a
c
d
«
cq in
S5
1
|
g
CD
E
OJ « -
•D
g
O
I8
a OS
li
oi co
2-52.
<q to
o o>
li
II
u
3 X
I*
m
as
ID c
2 8
if0) 0)
.= 5
shifted toward the right across the entire frequency
distribution (figure 2). There was no clear excess of
cases limited to those with extremely elevated levels.
For cysteine and iV, N-dimethylglycine, we observed
significantly higher geometric mean plasma levels
in cases than controls, whereas methionine and Nmethylglycine levels were lower in cases (all p values
below 0.05). Although not statistically significant,
geometric mean cystathionine was higher among cases
as compared with controls. Geometric mean levels of
serine and glycine did not show substantial differences
between the groups. The geometric mean level of
methylmalonic acid was significantly higher among
cases than among control subjects, possibly indicating
a higher prevalence of vitamin B I 2 inadequacy among
cases. However, the concentration of 2-methylcitric
acids I and II did not differ significantly between cases
and controls.
Table 3 shows the mean dietary intake of methionine, the metabolic precursor of homocysteine, and the
mean plasma and intake levels of vitamins B 6 , B 12 ,
and folate, which are all important cofactors in the
enzymatic pathways of homocysteine metabolism.
Adjusting for differences in sex and age, we observed
that both geometric mean dietary (including intake
from vitamin supplements) and plasma levels of vitamin B 6 and folate were significantly lower in cases
than in controls. There was no statistically significant
difference between cases and controls for dietary methionine and plasma or dietary vitamin B 12 , although
methionine intake was slightly higher among cases,
and plasma vitamin B 1 2 was slightly lower among
cases.
Associations of homocyst(e)ine with methionine
intake and vitamins
We calculated correlations between methionine intake and plasma homocyst(e)ine levels. Adjusting for
age and sex only, we found that these correlations
were -0.28 (p = 0.001) and -0.27 (p = 0.003) for
cases and controls, respectively. After additional adjustment for total intakes of vitamins B 6 , B 12 , and
folate, these correlations were —0.09 (p = 0.29) and
-0.22 (p = 0.01, data not shown).
Table 4 shows the correlation coefficients between
plasma levels of homocyst(e)ine and plasma and intake levels of vitamins, adjusted for age and sex.
Homocyst(e)ine correlated inversely with plasma levels of all three vitamins in cases as well as in controls.
Correlations were particularly strong for the association of plasma homocyst(e)ine with plasma folate,
with r = -0.38 (p = 0.0001) for cases and r = -0.49
(p = 0.0001) for controls. The association existed
across the entire range of folate levels, even in subjects
Am J Epidemiol
Vol. 143, No. 9, 1996
Homocyst(e)ine, Vitamins, and Myocardial Infarction
851
N
u
m
b
e
r
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Plasma Homocyst(e)ine (mnol/liter)
FIGURE 2. Smoothed frequency distribution of homocyst(e)ine levels (/imol/liter), rounded to the nearest Integer, in plasma from 130 cases
of first myocardial Infarction and 118 control subjects, Boston Area Health Study, 1982-1983.
with normal to high levels. In cases, but not in controls, an inverse association of plasma homocyst(e)ine
with plasma vitamin B 1 2 was observed (r = —0.35,
p = 0.0001). In cases, intakes of all three vitamins
showed significant inverse relations with plasma homocyst(e)ine, whereas in control subjects these associations were less strong.
Because of the use of vitamin supplements and, to a
lesser extent, because of similar dietary sources, vitamins were significantly intercorrelated in a similar
way for cases and controls. In control subjects, correlation coefficients were 0.44 (p = 0.0001) for plasma
pyridoxal 5'-phosphate and plasma folate, 0.30 (p =
0.0008) for plasma levels of pyridoxal 5'-phosphate
and vitamin B 12 , and 0.33 (p = 0.0003) for plasma
folate and plasma vitamin B 12 . Correlation coefficients were 0.57 (p = 0.0001) for intakes of vitamin
B 6 and folate, 0.64 (p = 0.0001) for vitamins B 6 and
B 12 , and 0.51 (p = 0.0001) for folate and vitamin B 12 .
Thus, we also calculated correlations between levels
of homocyst(e)ine with plasma and dietary levels of
each of the vitamins, controlling for levels of the other
two vitamins simultaneously (table 4). In control subjects, only for folate (plasma and dietary levels) did
the association with homocyst(e)ine remain, whereas
for vitamins B 6 and B 12 , associations disappeared. For
Am J Epidemiol
Vol. 143, No. 9, 1996
cases, both plasma folate and vitamin B 1 2 (plasma and
dietary levels) remained inversely related to plasma
homocyst(e)ine, when controlling for levels of each of
the other two vitamins simultaneously.
Associations of homocyst(e)ine with myocardial
infarction
We calculated risks by quintiles of plasma homocyst(e)ine, based on the distribution among controls.
For individuals with plasma homocyst(e)ine levels
higher than the 80th percentile of controls (11.2 ^i,mol/
liter), as compared with those with levels below the
20th percentile (7.2 /xmol/liter), we observed an ageand sex-adjusted odds ratio of 4.66 (95 percent confidence interval (CI) 1.90-11.40). The corresponding
multivariate adjusted odds ratio was 5.09 (95 percent
CI 1.84-14.10). For the three quintiles in the middle,
the risks were elevated as well. There was a strong
direct linear trend of plasma homocyst(e)ine with the
risk of myocardial infarction (p = 0.01, table 5).
In a logistic regression analysis with untransformed
plasma homocyst(e)ine and age and sex as independent variables, we estimated a coefficient of 0.101 ±
0.043 (/3 ± standard error (SE)), corresponding to an
odds ratio of 1.35 (95 percent CI 1.05-1.75) for each
CO
p
0.005
42.4
51.8
2.63
2.69
2.71 (0.67)
2.77 (0.69)
0.57
Geometric
mean
Mean
Food
0.04
3.29(1.37)
5.79 (22.99)
3.05
3.49
Mean
0.60
0.12
10.02
9.09
Geometric
moan
0.84
13.07(8.41)
14.20 (23.69)
11.26
11.13
Food plus
supplements
Geometric
Mean
moan
0.03
Food
0.01
340.9 (107.0)
372.0 (105.3)
Mean
8.23
9.11
Plasma folate
(nmol/ilter)
Geometric
Mean
mean
8.76 (3.30)
9.93 (4.93)
Dietary vitamin B,,
428
443
Food
11.52(7.56)
9.85 (4.02)
466 (199)
475(178)
Food phis
supplements
Geometric
Mean
mean
Dietary vitamin B, (mg/day)
49.3 (28.7)*
62.2 (40.5)
Plasma vitamin B,.
(pmol/Hter)t
Q so metric
Mean
mean
• PLP, pyrtdoxal 5'-phosphate; five cases and one control had missing values for plasma PLP.
t We excluded one case with a plasma B,, value of 2,712 pmol/liter, which was far outside the range of all other subjects.
t Numbers in parentheses, standard deviation.
§ Student's t test, based on the difference in geometric means between cases and controls, adjusting for age and sex differences.
p value for case-control
difference
Cases
Controls
p value for case-control
dlfference§
Cases
Controls
Plasma PLP*
(nmol/ilter)
Geometric
Mean
mean
0.25
323.7
357.7
Geometric
moan
2.50
2.41
Geometric
mean
0.002
423.0 (212.2)
493.6 (208.0)
377.4
453.7
Food plus
supplements
Geometric
Mean
moan
Dietary folate (jig/day)
2.58 (0.64)
2.49 (0.63)
Mean
Dietary methionlne
(g/day)
TABLE 3. Dietary Intake of methionlne and plasma levels and dietary intake of vitamin Bv vitamin B12> and folate In 130 cases of myocardlai infarction and 118 controls,
Boston Area Health Study, 1982-1983
a
Homocyst(e)ine, Vitamins, and Myocardial Infarction
3-^xmol/liter (about 1 standard deviation) increase in
plasma homocyst(e)ine. After additional adjustment
for standard risk factors, including a history of high
blood pressure, diabetes, plasma total/high density lipoprotein cholesterol, family history of ischemic heart
disease, cigarette smoking, alcohol consumption, body
mass index, and energy intake, we estimated a coefficient of 0.100 ± 0.051, corresponding to an odds ratio
of 1.35 (95 percent CI 1.00-1.82). This can be interpreted as an average 35 percent increase for each 1
standard deviation increase in plasma homocyst(e)ine.
To evaluate the possible effect modification of four
major risk factors of myocardial infarction (age, history of high blood pressure, history of hypercholesterolemia, smoking), we assessed the risk associated with
an increase of 3 /xmol/liter in plasma homocyst(e)ine
among subjects stratified for these factors. We did not
observe substantial differences among the strata of
these risk factors (data not shown).
CM
o
o co
p p
o o
o
n
o
O CO
in p
d d
m
JO
o
a
o
o
00
•a
c
a
o in
p r--
o
d d
o d
3
r-~ oo
co p
•a
a
2S
O <N
s
O
d d
O
853
_
CO
Associations of methionine intake and vitamins
with myocardial infarction
o
en CN
i
CM
f-
•a
c
a
£
m
o
o
p
cl d
§
o
i-
O
p p
d d
3
1
CO
o
c
O CM
O CO
d d
I
o
o
.c
a
eg
p
°°
«
a.
° o> en
a> CD co
||
E TJT3
co ca cd
p p
d d
*f
co cn
d d
(D r
O CO
IT
ii
LLJ
r-
aj
a
Am J Epidemiol
8I
58
..055
8S
to c
3(3
Vol. 143, No. 9, 1996
111 10
For plasma pyridoxal 5'-phosphate, the highest two
quintiles showed a trend toward an inverse relation
with the risk of myocardial infarction, even after adjustment for strong confounders like plasma total/high
density lipoprotein cholesterol and smoking habits
(table 5) (p trend = 0.12). For plasma folate, the
association was inverse as well, with a p value for
trend of 0.09 after control for other factors. Plasma
vitamin B 1 2 was not associated with the risk of myocardial infarction. For plasma pyridoxal 5'-phosphate,
we did the same analyses in the original study group,
of which 275 cases and 281 controls had pyridoxal
5'-phosphate levels measured. Both the age- and sexadjusted effect and the multivariate adjusted effect
were stronger in the smaller study group as compared
with the larger group. Also, there seemed to be more
confounding in the larger group, especially from
smoking and plasma total/high density lipoprotein
cholesterol (data not shown).
Table 6 shows odds ratios of myocardial infarction
by quintiles of energy-adjusted daily intakes of methionine and vitamins B 6 , B, 2 , and folate (with and
without intake from supplements). Intake of methionine showed no substantial association with the risk of
myocardial infarction. For folate intake, and to a lesser
extent for vitamin B 6 , we observed significant inverse
associations with myocardial infarction risk, after adjustment for age and sex, as well as in a multivariate
analysis, adjusting for possible confounders. For total
vitamin B 6 intake, the estimates were rather inconsistent. Intake of vitamin B 1 2 was not materially associated with the risk of myocardial infarction; odds ratios
CO
CO
CO
o
1.29
1.19(0.53-2.69)
£8.31
£6.56
1
1
0.71 (0.29-1.77)
1
£399
0.60
£321
1
0.97
0.97 (0.40-2.33)
1
1
£47.0
£29.7
0.41
0.47(0.18-1.22)
£9.95
Plasma folate (nmol/llter)
1.13
1.07 (0.46-2.48)
£503
Plasma vitamin B,, (pmot/Ilter)
0.94
1.68 (0.70-4.04)
£63.1
Plasma PLPt (nmoVUter)
3.23
4.61 (1.61-13.20)
£9.7
0.44
0.32 (0.12-0.84)
£12.0
0.55
0.57(0.22-1.49)
£610
0.37
0.37(0.14-1.01)
£88.9
2.10
2.50 (0.85-7.32)
£11.1
4
>610
0.53
0.57 (0.22-1.42)
>12.0
0.97 (0.41-2.28)
0.94
0.32
0.51 (0.19-1.36)
>88.9
4.66
5.09(1.84-14.10)
>11.1
5
-0.804 ± 0.365
-0.707 ±0.417
-0.175 ±0.328
-0.151 ± 0.364
-0.644 ± 0.233
-0.407 ± 0.259
1.258 ±0.467
1.352 ±0.548
Pt±SEt
P
0.03
0.09
0.59
0.68
0.006
0.12
0.007
0.01
value
* Quintiles are based on the distribution of control subjects; real scale boundaries of quintiles are given,
t Based on logistic regression analysis with log-transformed variables,
i SE, standard error; PLP, pyrldoxal 5'-phosphate.
§ Adjusted for age, sex, energy intake, history of diabetes, history of high blood pressure, family history of Ischemic heart disease, total/high density lipoprotein cholesterol ratio, body
mass index, cigarette smoking, and alcohol consumption.
II Numbers in parentheses, 95% confidence interval.
Age and sex adjusted
MultJvariate§ adjusted
Age and sex adjusted
Multjvariate§ adjusted
Age and sex adjusted
Multivariate§ adjusted
3
Qulntlle* of plasma level
Plasma homocyst(e)ine (nmol/liter)
1.81
1.80(0.62-5.21)11
£8.2
£7.2
1
1
2
1
Odda ratios of myocardlal Infarction by plasma levels of homocyst(e)ine, pyrldoxal 5 -phosphate, vitamin B12, and folate, Boston Area Health Study, 1982-1983
Age and sex adjusted
Multivariate§ adjusted
TABLE 5.
CO
p
1.51
1.60(0.65-3.91)
£319
1
1
£282
1
1
0.72
0.63(055-1.62)
0.88
0.66(057-1.63)
0.67
0.48 (0.19-155)
£3.79
>4.38
1.00
1.05(0.43-2.57)
44.38
0.17
0.17 (0.05-0.59)
1.09
1.17(0.46-3.01)
£9.81
>12.28
1.67
1.77(0.73-4.34)
£12.28
1.45
1.30(0.51-3.29)
0.67
0.70 (0.30-1.62)
£391
Folate Mate, food only (no'day)
1.36
1.13(0.46-2.78)
£12.61
>467
0.30
0.30(0.11-0.81)
£467
1.42
1.46(0.61-3.51)
0.72
0.73(058-1.92)
0.66
0.57(054-1.35)
>16.09
£16.09
0.45
0.44 (0.17-1.09)
£490
>682
0.43
0.38(0.15-0.95)
£682
0.31
054 (0.09-0.62)
Fotate Mate, tood and supplements (w>May)
0.61
0.63 (0.27-1.49)
£391
0.37
0.28 (0.11-0.74)
£9.70
£7.06
£310
>3.26
£3.28
Vitamin B1t Intake, tood and supplements (ntfday)
0.96
0.84(0.31-254)
1
1
1
1.00
1.00(0.43-2.36)
£257
Vtamki B1t Make, tood only (wjttay)
1.44
2.05 (0.86-4.90)
£3.07
£8.32
1
1.43
1.28(0.52-3.11)
1.37
156(0.52-3.09)
5
>256
4
£256
Vtamln B, Nate, tood only (mgAJay)
1.18
1.04(0.41-2.62)
£2.62
MetNonfrM Inlate (gAtay)
Vitamin B, Make, tood and supplements (mgMay)
0.55
0.53(0.21-1.33)
££50
0.77
0.58(0.22-1.53)11
4657
1
1
£2.31
1
1
£2.12
1
1
££24
£154
3
QuWlls' of nutrient intake
-0.924 ± 0.302
-1.012 ±0.347
-1.115 ±0.454
-1.086 ±0.514
0.048 ± 0 5 3 3
0.052 ± 0 5 6 0
0.442 ± 0.283
0.522 ± 0.321
-O.651 ± 0.324
-0.729 ± 0.377
-0.301 ± 0.523
-0.438 ± 0.597
0.584 ±0.510
0.552 ± 0.578
Pt±S£*
0.002
0.004
0.01
0.03
0.84
0.84
0.12
0.10
0.04
0.05
0.57
0.46
0.25
0.34
P
value
* Quintiles are based on the dstribution of control subjects; rsal scale boundaries of qulntles are given,
t Baaed on logistic regression analysis with log-transtormed variables.
t SE, standard error.
§ Adjusted tor age, sax, energy intake, history of dtebetes, history of high Wood pmsure, family history of ischemic heart disease, total/high density Upoprotein cholesterol ratio body
mass index, cigarette smoking, and alcohol consumption.
'
II Numbers In parentheses, 95% confidence interval.
Age and sex adjusted
Multivariate§ adjusted
Age and sex adjusted
Mutth/ariate§ adjusted
Age and sex adjusted
Multivariate§ adjusted
Age and sex adjusted
Multivariate§ adjusted
Age and sex adjusted
Muluvariate§ adjusted
Age and sex adjusted
Multfvaiiate§ adjusted
Age and sex adjusted
Multivariate§ adjusted
2
1
TABLE 6. Odda ratio* of myocardW infarction by level of energy-adjusted daily intake of methionine, vitamin B 8 , vitamin B12< and folate, Boston Area Health Study,
1982-1983
856
Verhoef et at.
for all quintiles were close to or above unity. Generally, for all risk analyses per quintiles, the small number of subjects for each quintile made the estimates
quite unstable and the confidence intervals wide. We
repeated the analyses shown for the smaller group in
table 6 in the original study population of 340 cases
and 339 controls; the associations were in the same
direction but weaker than those in the smaller population for whom we had plasma. Intake of folate and,
to a lesser extent, of vitamin B 6 was higher in the
smaller control group as compared with the original
control group (data not shown).
Because vitamin intakes were intercorrelated, all
variables were entered into the multivariate adjusted
model continuously as log-transformed variables. Association of the total dietary intake of folate with
myocardial infarction risk became somewhat stronger
(/3 ± SE: -1.378 ± 0.483, p = 0.004) when the
intakes of vitamin B 6 and vitamin B 12 were added to
the logistic model as continuous variables. Adjustment
for folate and vitamin B 1 2 weakened the association of
total intake of vitamin B 6 with myocardial infarction
(/3 ± SE: -0.3872 ± 0.4266, p = 0.36). This illustrates that increased intakes of both vitamins are associated with the decreased risk of myocardial infarction, but that part of the observed effect of vitamin B 6
is due to intercorrelation with folate.
Finally, when plasma folate, plasma pyridoxal 5'phosphate, and plasma vitamin B 1 2 were entered into
the logistic model together with plasma homocyst(e)ine (all as log-transformed continuous variables),
the coefficient for homocyst(e)ine became somewhat
smaller (/3 ± SE: 0.998 ± 0.620, p = 0.11) and for
folate, much smaller (/3 ± SE: -0.242 ± 0.511, p =
0.22). The coefficient of pyridoxal 5'-phosphate was
almost not affected (/3 ± SE: -0.311 ± 0.282, p =
0.27). This seems to indicate that folate and homocyst(e)ine are part of the same causal pathway (low folate
levels leading to elevated plasma homocyst(e)ine
which subsequently may lead to increased risk of
myocardial infarction), whereas in this population vitamin B 6 possibly exhibits an association with myocardial infarction through mechanisms other than influencing homocyst(e)ine levels.
DISCUSSION
In this study, we found plasma homocyst(e)ine to be
an independent risk factor for first myocardial infarction, and the effect seemed to be graded. Elevation of
homocyst(e)ine in cases was associated with increased
levels of cystathionine and cysteine, indicating increased catabolism of homocysteine through the transsulfuration pathway, which is dependent on vitamin
B 6 . Remethylation of homocysteine to methionine
seemed to be impaired in cases, as reflected by their
lower plasma methionine levels. This was most Likely
due to lower dietary intake of folate or subclinical
vitamin B 12 deficiency among cases as compared with
controls. Higher levels of AfJV-dimethylglycine suggest that there was increased demethylation of betaine
to A^yV-dimethylglycine, a reaction step which is independent of folate and vitamin B ] 2 , at which homocysteine is remethylated to methionine. Accordingly, folate exhibited a strong inverse association with plasma
homocyst(e)ine in cases as well as controls, whereas
vitamin B 6 showed no association with plasma homocyst(e)ine, controlling for the other two vitamins.
Therefore, although dietary and intake levels of vitamin B 6 were lower in cases than controls, this did not
appear to be the primary reason for elevated fasting
homocyst(e)ine levels. This is in line with the concept
that the fasting homocyst(e)ine level is usually determined by homocysteine remethylation and not by abnormalities in the transsulfuration pathway (25).
Total dietary intakes and plasma levels of vitamin
B 6 and folate showed inverse relations with the risk of
myocardial infarction, whereas for vitamin B 12 , no
clear associations with the risk of myocardial infarction were observed. Vitamin levels were intercorrelated, however, and the effect of dietary vitamin B 6
was attenuated after correction for dietary folate and
vitamin B 12 . The association between plasma pyridoxal 5'-phosphate and myocardial infarction risk was
independent of plasma folate and homocyst(e)ine, suggesting that protective mechanisms of vitamin B 6 ,
other than its role in homocysteine metabolism, may
be more important in this study population. Several
possible mechanisms have been proposed to explain
an apparent protective effect of vitamin B 6 against
coronary disease (26). Also, it may be possible that
low pyridoxal 5'-phosphate levels in cases are associated with reduced transsulfuration, maybe leading to
increased homocyst(e)ine levels after protein-rich
meals. However, without having measured plasma homocyst(e)ine in response to a methionine load, we
cannot draw any conclusion about this.
Among controls, we observed an inverse association
between methionine intake and plasma homocyst(e)ine, taking into account confounding by age, sex,
and intakes of vitamins B 6 , B 12 , and folate. One might
expect a positive association, since methionine is
the main precursor of homocysteine. However, a
methionine-rich diet may induce a more efficient catabolism, as has been shown in rats (27). This might
occur in humans as well, as suggested by our data.
This may explain why we did not find a strong direct
association of methionine intake with the risk of myocardial infarction in this population.
Am J Epidemiol
Vol. 143, No. 9, 1996
Homocyst(e)ine, Vitamins, and Myocardial Infarction
The effects of folate and, to some extent, of vitamin
B 6 (especially plasma pyridoxal 5'-phosphate) were
stronger in the subsample shown herein than in the
original study population, because of higher intake of
these vitamins among controls of the subsample as
compared with the original control group. Subjects in
the subsample were selected on the criterion of having
a sufficient volume of stored plasma. Since the same
amount of blood was drawn from each subject, independently of case-control status or dietary habits,
chance is the most plausible explanation for this observation. Another possibility may be residual confounding by age and sex, because cases and controls
were not matched by these factors in the subsample, as
was the case for the original sample. The selection bias
may in part explain the observed associations for folate and vitamin B 6 and maybe plasma homocyst(e)ine
with the risk of myocardial infarction. To get some
indication of the effect of this bias, we estimated the
expected homocyst(e)ine levels in cases and controls
of the original study group that had plasma levels of
pyridoxal 5'-phosphate available (281 cases, 275 controls), based on a linear regression equation obtained
from data in the subsample. Included independent
variables were the intake of the vitamins, plasma pyridoxal 5'-phosphate, age, sex, alcohol consumption,
plasma total/high density lipoprotein cholesterol, methionine intake, and history of high blood pressure. In
the original study group, estimated homocyst(e)ine
levels were significantly higher in cases than in controls. Also, the increase in risk per 1 standard deviation
increase in plasma homocyst(e)ine was similar to the
one observed in the smaller study population. Hence,
the selection bias might have strengthened the association of homocyst(e)ine with the risk of myocardial
infarction, but it cannot have been totally responsible
for the observed effect.
The case-control design of the study leaves the
possibility that the higher homocyst(e)ine level or
lower vitamin levels were caused by the disease or its
treatment. Cases were interviewed approximately 8
weeks after the event, and blood was drawn at the
same time. This time span of almost 2 months should
probably be enough to exclude any acute effect of the
myocardial infarction on plasma homocyst(e)ine or
vitamins. We tested this in 106 cases, for whom we
had plasma levels of homocyst(e)ine and vitamins
measured in blood that was drawn when they were still
in the hospital. In a paired analysis, we found the
hospital values of homocyst(e)ine to be significantly
lower (-1.4 ± 0.25 (SE) ^mol/liter, p = 0.0001) than
the values measured in blood obtained 8 weeks after
the event. Consequently, any temporary decrease in
homocyst(e)ine values in cases could only have caused
Am J Epidemiol
Vol. 143, No. 9, 1996
857
the association with myocardial infarction to be
weaker but certainly not stronger. For plasma pyridoxal 5'-phosphate, the hospital values were significantly lower than the levels measured in blood that
was drawn at the time of the interview. Therefore, any
temporary decrease in plasma pyridoxal 5'-phosphate
may have strengthened the inverse association of
plasma pyridoxal 5'-phosphate with myocardial infarction. Plasma levels of vitamin B 1 2 and folate did
not differ between blood samples drawn in the hospital
and at the interview.
The medications used by subjects could have affected their plasma levels of homocyst(e)ine or vitamins. However, most drugs that are known to have an
effect on plasma homocyst(e)ine or vitamins, such as
methotrexate, anticonvulsants, or penicillamine (1),
are not prescribed more frequently in myocardial infarction patients.
Recall bias or changes in dietary habits of the cases
since the event could have influenced homocyst(e)inemyocardial infarction or vitamin-myocardial infarction relations. However, we observed very similar
correlations between dietary and plasma levels of vitamin B 6 , vitamin B 12 , and folate for cases and controls, which reduces the possibility that either of those
two biases had occurred. Since dietary assessment by
means of a questionnaire is rather crude, this may
explain why the associations of plasma homocyst(e)ine with dietary intakes of the vitamins were weak,
as opposed to associations with plasma levels of the
vitamins.
Blood was stored for approximately 10 years at
—70°C before samples were assayed for homocyst(e)ine and vitamins, which might have had an effect
on homocyst(e)ine levels. However, because of similar
storage time, sample treatment, and laboratory measurement for samples of cases and controls, chances
are small that this could have affected the association.
Also, the strength of the observed effect of homocyst(e)ine and the similarity with findings from other
case-control (including prospective ones) and crosssectional studies on homocyst(e)ine and the risk of
coronary heart disease (8, 9, 28-34) make it unlikely
that chance or bias explain our findings. In a general
population in Norway, Arnesen et al. (9) showed a
graded association of homocyst(e)ine with coronary
heart disease. They observed a significant 32 percent
increase for each 1 standard deviation (4 /unol/liter)
increase in plasma homocyst(e)ine, after adjustment
for established coronary risk factors, similar to the 35
percent increase for each 3 jtimol/liter in our study.
Stampfer et al. (8), in the Physicians' Health Study,
probably a better nourished study population than that
of Arnesen et al., found a threshold level above which
858
Vertioef et al.
risk was increased. They observed that the higher
mean homocyst(e)ine level in cases was due to an
excess of cases with homocyst(e)ine values above the
95th percentile of control subjects (15.8 /xxnol/liter).
Whether the level of nourishment of the studied population influences the shape of the relation of homocyst(e)ine with risk of vascular disease needs to be
studied further.
Our study population represented a well-nourished
segment of a general population as well; both cases
and controls lived in upper middle class neighborhoods in the Boston suburbs and were more highly
educated than the general population. Only 10 subjects
with first myocardial infarction (eight men, two
women) and one male control subject had folate intakes less than 200 /xg/day, the current recommended
daily allowance for men. The nutritional status of
folate was less adequate, however. Thirty-nine cases
(30 percent, 27 men and 12 women) and 27 controls
(23 percent, 16 men and 11 women) had inadequate
folate levels (^6.8 nmol/liter, US National Health and
Nutrition Examination Survey II (NHANES II)).
When plotting folate intakes against plasma homocyst(e)ine levels, we observed that homocyst(e)ine
reached its nadir at folate intakes higher than 350-400
jig/day, similar to the findings of Selhub et al. (10).
Plasma folate was inversely associated with plasma
homocyst(e)ine across the entire distribution of plasma
folate levels, also above the critical level of 6.8 nmol/
liter. These findings indicate that, for optimizing homocyst(e)ine levels, recommended folate intake
should be at least 300-400 /xg/day.
Several explanations for the positive association of
homocyst(e)ine with vascular disease have been proposed. Homocysteine thiolactone, a reactive form of
homocysteine, can modify low density lipoprotein,
leading to aggregation and increased uptake of low
density lipoprotein by macrophages. Homocysteine
released from the low density lipoprotein within the
vascular wall promotes intimal injury and influences
vascular coagulant mechanisms (35). Furthermore, homocysteine may enhance the binding of lipoprotein(a)
to fibrin, diminishing fibrinolysis (2). A recent publication points at the possibility that homocysteine promotes vascular smooth muscle cell growth and inhibits
endothelial cell growth, both predisposing to atherosclerosis (36).
In conclusion, our study adds further epidemiologic
evidence to the hypothesis that plasma homocyst(e)ine
is an important independent risk factor for coronary
disease. The graded effect suggests that lowering of
plasma homocyst(e)ine can have important public
health effects for a large segment of the adult population. Furthermore, our study supports the view that
adequate folate intake from diet or supplements might
be an important step toward normalizing homocyst(e)ine levels. Supplementation with folate has been
shown to successfully reduce homocyst(e)ine levels
within a few weeks, even in subjects with genetically
caused hyperhomocysteinemia (12). Whether this will
also reduce the risk of coronary heart disease can only
be shown by observational studies and randomized
trials relating intake and status to the incidence of
coronary heart disease (37). In the meantime, it might
be worthwhile to reconsider the recommended daily
allowance of 200 pig/day for folate and perhaps restore
the previous value of 400 /xg/day.
ACKNOWLEDGMENTS
The research was supported by grants HL24423 and
HL21006 and institutional training grant HL07575 from the
National Heart, Lung, and Blood Institute, Bethesda, MD.
The authors thank the six Boston area hospitals that
participated in this study: Emerson Hospital, Concord (Dr.
Marvin H. Kendrick), Framingham Union Hospital, Framingham (Dr. Marvin Adner), Leonard Morse Hospital,
Natick (Dr. L. Frederick Kaplan), Mount Auburn Hospital,
Cambridge (Dr. Leonard Zir), Newton-Wellesley Hospital,
Newton (Dr. James Sidd), and Waltham Hospital, Waltham
(Dr. Solomon Gabbay). They also thank Stephanie Parker,
Tom Gaziano, Kathie Schneider, and Marty Vandenburgh,
who assisted in the research.
REFERENCES
1. Ueland PM, Refsum H. Plasma homocysteine, a risk factor for
vascular disease: plasma levels in health, disease, and drug
therapy. J Lab Clin Med 1989;114:473-501.
2. Rees MM, Rodgers GM. Homocysteinemia: association of a
metabolic disorder with vascular disease and thrombosis.
ThrombRes 1993;71:337-59.
3. Selhub J, Miller JW. The pathogenesis of homocysteinemia:
interruption of the coordinate regulation by S-adenosylmethionine of the remethylation and transsulfuration of homocysteine. Am J Clin Nutr 1992;55:131-8.
4. Mason JB, Miller JW. The effects of vitamins B, 2 , B 6 , and
folate on blood homocysteine levels. Ann N Y Acad Sci
1992;30:197-204.
5. Ueland PM, Refsum H, Brattstrbm L. Plasma homocysteine
and cardiovascular disease. In: Francis RBJ, ed. Atherosclerotic cardiovascular disease, hemostasis, and endothelial function. New York: Marcel Dekker, Inc, 1992:183-236.
6. Malinow MR. Hyperhomocyst(e)inemia. A common and easily reversible risk factor for occlusive atherosclerosis. Circulation 1990;81:2004-6.
7. Malinow MR. Homocyst(e)ine and arterial occlusive disease:
a mini review. Clin Chem 1994;40:173-6.
8. Stampfer MJ, Malinow MR, Willett WC, et al. A prospective
study of plasma homocyst(e)ine and risk of myocardial infarction in US physicians. JAMA 1992;268:877-81.
9. Amesen E, Refsum H, Bonaa KH, et al. Serum total homocysteine and coronary heart disease. Int J Epidemiol 1995;24:
704-9.
Am J Epidemiol Vol. 143, No. 9, 1996
Homocyst(e)ine, Vitamins, and Myocardial Infarction
10. Selhub J, Jacques PF, Wilson PWF, ct al. Vitamin status and
intake as primary determinants of homocysteinemia in an
elderly population. JAMA 1993,270:2693-8.
11. Ubbink JB, Vermaak WJH, van der Merwe A, et al. Vitamin
B| 2 , vitamin B 6 , and folate nutritional status in men with
hyperhomocysteinemia. Am J Clin Nutr 1993;57:47-53.
12. BrattstrOm LE, Israelsson B, Jeppsson JO, et al. Folic acid: an
innocuous means to reduce plasma homocysteine. Scand J
Clin Lab Invest 1988;48:215-21.
13. Ubbink JB, van der Merwe A, Vermaak WJH, et al. Hyperhomocysteinemia and the response to vitamin supplementation. Clin Investig 1993;71:993-8.
14. Pancharuniti N, Lewis CA, Sauberlich HE, et al. Plasma
homocyst(e)ine, folate, and vitamin B 1 2 concentrations and
risk for early-onset coronary artery disease. Am J Clin Nutr
1994;59:940-8.
15. Allen RH, Stabler SP, Savage DG, et al. Metabolic abnormalities in cobalamin (vitamin B12) and folate deficiency. FASEB
J 1993;7:1344-53.
16. Buring JE, O'Connor GT, Goldhaber SZ, et al. Decreased
HDLj and HDL, cholesterol, Apo A-I and Apo A-U, and
increased risk of myocardial infarction. Circulation 1992;85:
22-9.
17. Paffenbarger RS Jr, Wing AL, Hyde RT. Physical activity as
an index of heart attack risk in college alumni. Am J Epidemiol 1978; 108:161-75.
18. Reynolds RD. Vitamin B 6 analysis. In: Pesce AJ, Kaplan LA,
eds. Methods in clinical chemistry. St. Louis, MO: CV Mosby,
1987:558-68.
19. Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility
and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol 1985;122:51-65.
20. Willett WC, Sampson L, Browne ML, et al. The use of a
self-administered questionnaire to assess diet four years in the
past. Am J Epidemiol 1988;127:188-99.
21. Colditz GA, Willett WC, Stampfer MJ, et al. The influence of
age, relative weight, smoking, and alcohol intake on the rcproducibility of a dietary questionnaire. Int J Epidemiol 1987;
16:392-8.
22. Willett WC. Nutritional epidemiology. New York: Oxford
University Press, 1989.
23. US Department of Agriculture. Composition of foods: raw,
processed, prepared. Washington, DC: US GPO, 1963-1988.
(USDA handbook no. 8 series).
Am J Epidemiol
Vol. 143, No. 9, 1996
859
24. Willett W, Stampfer MJ. Total energy intake: implications for
epidemiologic analyses. Am J Epidemiol 1986;124:17-27.
25. BrattstrOm L, Israelsson B, Norrving B, et al. Impaired homocysteine metabolism in early-onset cerebral and peripheral
occlusive arterial disease. Atherosclerosis 1990;81:51-60.
26. Willett WC. Does low vitamin B 6 intake increase the risk of
coronary heart disease? In: Reynolds RD, Leklem JE, eds.
Vitamin B 6 : its role in health and disease. New York: Alan R
Liss, Inc, 1985:337-46.
27. Finkelstein JD, Martin JJ. Methionine metabolism in mammals. Adaptation to methionine excess. J Biol Chem 1986;
261:1582-7.
28. Israelsson B, BrattstrOm LE, Hultberg BL. Homocysteine and
myocardial infarction. Atherosclerosis 1988;71:227-33.
29. Malinow MR, Sexton G, Averbuch M, et al. Homocyst(e)ine
in daily practice: levels in coronary artery disease. Coron
Artery Dis 1990; 1:215-20.
30. Ubbink JB, Vermaak WHJ, Bennett JM, et al. The prevalence
of homocysteinemia and hypercholesterolemia in angiographically defined coronary heart disease. Klin Wochenschr
1991;69:527-34.
31. Genest JJ Jr, McNamara JR, Salem DN, et al. Plasma homocyst(e)ine levels in men with premature coronary artery disease. J Am Coll Cardiol 1990;16:l 114-19.
32. Kang SS, Wong PW, Cook HY, et al. Protein-bound homocyst(e)ine. A possible risk factor for coronary artery disease. J
Clin Invest 1986;77:1482-6.
33. Von Eckardstein A, Malinow MR, Upson B, et al. Effects of
age, lipoproteins, and hemostatic parameters on the role of
homocyst(e)inemia as a cardiovascular risk factor in men.
Arterioscler Thromb 1994;14:460-4.
34. Dalery K, Lussier-Cacan S, Selhub J, et al. Homocysteine and
coronary artery disease in French-Canadian subjects: relation
with vitamins B 12 , B 6 , pyridoxal phosphate, and folate. Am J
Cardiol 1995,75:1107-11.
35. McCully KS. Chemical pathology of homocysteine. I. Atherogenesis. Ann Clin Lab Sci 1993;23:477-93.
36. Tsai JC, Perrella MA, Yoshizumi M, et al. Promotion of
vascular smooth muscle cell growth by homocysteine: a link
to atherosclerosis. Proc Natl Acad Sci U S A 1994;91:
6369-73.
37. Stampfer MJ, Willett WC. Homocysteine and marginal vitamin deficiency. The importance of adequate vitamin intake.
(Editorial). JAMA 1993;270:2726-7.