Anda di halaman 1dari 7

Plasma lycopene, other carotenoids, and retinol and the risk of

cardiovascular disease in women13

Howard D Sesso, Julie E Buring, Edward P Norkus, and J Michael Gaziano
Background: Growing evidence suggests that lycopene has significant in vitro antioxidant potential. Lycopene has rarely been
tested in prospective studies for its role in cardiovascular disease
(CVD) prevention.
Objective: We examined the association between plasma lycopene and the risk of CVD in middle-aged and elderly women.
Design: A prospective, nested, case-control study was conducted
in 39 876 women initially free of CVD and cancer in the Womens
Health Study. Baseline blood samples were collected from 28 345
(71%) of the women. During a mean of 4.8 y of follow-up, we
identified 483 CVD cases and 483 control subjects matched by
age, smoking status, and follow-up time. Plasma lycopene, other
carotenoids, retinol, and total cholesterol were measured.
Results: In analyses matched for age and smoking, with adjustment for plasma cholesterol, the relative risks (RRs) and 95% CIs
of CVD in increasing quartiles of plasma lycopene were 1.00
(referent), 0.78 (95% CI: 0.55, 1.11), 0.56 (0.39, 0.82), and 0.62
(0.43, 0.90). In multivariate models, the RRs were 1.00 (referent),
0.94 (0.60, 1.49), 0.62 (0.39, 1.00), and 0.67 (0.41, 1.11); those in
the upper compared with the lower half of plasma lycopene had an
RR of 0.66 (0.47, 0.95). For CVD, exclusive of angina, women in
the upper 3 quartiles had a significant multivariate 50% risk
reduction compared with those in the lowest quartile. The stepwise
addition of individual plasma carotenoids did not affect the RRs.
Conclusions: Higher plasma lycopene concentrations are associated with a lower risk of CVD in women. These findings require
confirmation in other cohorts, and the determinants of plasma lycopene concentrations need to be better understood.
Am J Clin Nutr
Lycopene, cardiovascular disease, women,
prospective studies, nutrition


There is growing evidence from many studies that lycopene,

a carotenoid without provitamin A activity and found in high
concentrations in a small number of plant foods (tomato, watermelon, pink grapefruit, papaya, and apricot), has significant
antioxidant potential in vitro, which suggests a role in preventing cardiovascular disease (CVD) (1). More than 80% of the
lycopene intake in the American diet is from the consumption
of tomato products, including ketchup, tomato juice, and tomato sauces (2). Lycopene may have a cholesterol synthesis
inhibiting effect that may enhance LDL degradation (3, 4).

Some (4, 5), but not all (6), dietary intervention studies
involving either lycopene-containing foods or lycopene supplementation have shown potential short-term improvements in
LDL oxidation. However, whether these apparent short-term
benefits translate into long-term improvements in health, manifested by a reduction in the risk of CVD, remains unknown.
Limited data suggest an inverse association between lycopene
and CVD (713). Data concerning the relation between plasma
lycopene concentrations and CVD risk are limited, particularly
in women. Lingering questions remain about whether all carotenoids (9), specific carotenoids such as lutein/zeaxanthin (14),
or carotenoids in general (15) are associated with a reduced risk
of CVD.
Therefore, we evaluated baseline concentrations of plasma
lycopene, other carotenoids, and retinol in middle-aged and
elderly women who were initially free of CVD in the Womens
Health Study (WHS) to examine their relation with the subsequent risk of CVD in female health professionals in the United


Study population
The WHS is an ongoing randomized, double-blind, placebocontrolled 2 2 factorial trial of the relation between low-dose
aspirin and vitamin E and the primary prevention of CVD and
cancer (16). The -carotene component of the trial was terminated in 1996 after a median follow-up of 2.1 y, in part because
of the lack of effect of -carotene on cancer incidence in the
From the Division of Preventive Medicine, Department of Medicine,
Brigham and Womens Hospital and Harvard Medical School, Boston
(HDS, JEB, and JMG); the Department of Epidemiology, Harvard School
of Public Health, Boston (HDS and JEB); the Massachusetts Veterans
Epidemiology Research and Information Center, VA Boston Healthcare
System, Boston (HDS and JMG); the Department of Ambulatory Care and
Prevention, Harvard Medical School, Boston (JEB); and the Department of
Medical Research, Our Lady of Mercy Medical Center, and Community
and Preventive Medicine, New York Medical College, Bronx, NY (EPN).
Supported by grants CA-47988 and HL-43851 from the National
Institutes of Health, Bethesda, MD, and a grant from DSM Nutritional
Products, Inc.
Address reprint requests to HD Sesso, Brigham and Womens Hospital, 900 Commonwealth Avenue East, Boston, MA 02215-1204. E-mail:
Received March 31, 2003.
Accepted for publication July 7, 2003.

Am J Clin Nutr 2004;79:4753. Printed in USA. 2004 American Society for Clinical Nutrition




Physicians Health Study (17, 18). A total of 39 876 female US

health professionals, who were aged 45 y in 1992, were
postmenopausal or not intending to become pregnant, and had
no history of myocardial infarction (MI), stroke, transient ischemic attack, or cancer (except for nonmelanoma skin cancer)
were enrolled in the study. Fasting baseline blood samples
were collected from 28 345 (71%) participants and stored in
liquid nitrogen until analyzed.
A prospective, nested, case-control design was used to identify 483 case-control pairs of WHS participants. Cases included
women who provided baseline blood samples and subsequently
experienced a cardiovascular event, defined as CVD death,
nonfatal MI, nonfatal stroke, percutaneous transluminal coronary angioplasty, coronary artery bypass graft, or angina pectoris. Study physicians conducted blinded reviews of all cases.
CVD death was documented on the basis of convincing evidence of a cardiovascular mechanism from death certificates
and medical records. The diagnosis of MI was confirmed on
the basis of World Health Organization criteria (19). A stroke
was defined as a typical neurologic deficit, sudden or rapid in
onset, that lasted 24 h. Revascularization procedures were
confirmed by hospital records. Angina was self-reported on
questionnaires and confirmed in a small proportion of cases
through hospital records.
Each case of CVD was matched with a control subject
according to age (1 y), smoking status (never, former, and
current), and follow-up time (6 mo). In addition, each woman
selected as a control must have provided a baseline blood
sample and remained free of CVD during the follow-up period.
The Institutional Review Board at Brigham and Womens
Hospital approved all procedures, and written informed consent was obtained from all participants.
Blood assays and covariates
All investigators and laboratory personnel were blinded to
the subjects case-control status. Blood samples from all subjects were handled identically and blindly through all stages of
the blood collection, storage, retrieval, and analytic processes.
The storage time for plasma in liquid nitrogen freezers was 7
y until the assays were conducted. Baseline plasma blood
samples from cases and controls were thawed and assayed for
total lycopene, other carotenoids, and retinol at Our Lady of
Mercy Medical Center, Bronx, NY. All analytes were quantitated by reversed-phase HPLC after extraction with the use of
standard methods (20). Internal standards (echinenone for carotenoids and retinyl proprionate for retinol) were used to correct
for recoveries of all samples that were analyzed, and the
laboratory has participated in the US Quality Assurance Program to ensure consistent methodology with other laboratories.
Plasma total cholesterol also was assayed by enzymatic, endpoint spectroscopy with the use of commercially available
diagnostic kits (Sigma-Aldrich Chemical Co, St Louis) and
conventional methods (21, 22). Plasma total cholesterol was
assayed because plasma lipoproteins are nonspecific carriers
for all the carotenoids in plasma and, to date, total cholesterol
appears to be the best way to control for confounding effects
due to differences in lipoprotein concentrations between subjects (23). The laboratory maintains its own sets of internally
and externally prepared control specimens that are assayed in
every run. From these control specimens, the laboratory accu-

racy was within 7% for each measured carotenoid and retinol,

whereas the day-to-day and within-day precision (CV) for
these analytes was 5%.
On the WHS baseline questionnaire, women also provided
data on age (in y), weight and height (converted to body mass
index; in kg/m2), smoking status (categorized as never, former,
or current), alcohol use (categorized as rarely or never, 13
drinks/mo, 16 drinks/wk, and 1 drink/d), frequency of exercise (categorized as rarely or never, 1 time/wk, 13 times/
wk, and 4 times/wk), parental history of MI at 60 y (no,
yes), history of hypertension (no, yes), history of diabetes (no,
yes), history of hypercholesterolemia (no, yes), and postmenopausal hormone use (categorized as never, former, or current).
Women also completed a 131-item validated semiquantitative
food-frequency questionnaire developed by Willett et al (24),
from which the dietary intake of specific nutrients included
lycopene (g/d), total fiber (g/d), folate (g/d), and saturated
fat intake (g/d), which were each adjusted for total energy
intake with the use of the residual method (25). The total fruit
and vegetable intake was summed and expressed as servings/d.
Data analyses
Women were first compared according to case-control status
by using mean values or proportions of baseline coronary
disease risk factors and biochemical markers. Measurements of
plasma lycopene concentrations were divided into fourths on
the basis of overall distribution in the 483 controls. Coronary
disease risk factors were also compared according to quartiles
of plasma lycopene among the control population to assess
potential confounding with the use of analysis of variance for
continuous variables, Cochran-Armitage trend tests for dichotomous variables, or chi-square tests for categorical variables.
In this same group of women, age-adjusted Spearman correlation coefficients were used to assess the associations between
plasma lycopene, its primary food sources (tomatoes, tomato
juice, tomato sauce, and pizza; all expressed as servings/d), and
dietary lycopene.
We used conditional logistic regression analyses to compute
RRs and 95% CIs for future CVD risk for increasing quartiles
of plasma lycopene concentrations, with the lowest quartile as
the referent. Models were first adjusted for randomized treatment assignments and plasma total cholesterol concentrations;
the next model added body mass index, exercise, parental
history of MI before the age of 60 y, hypertension, diabetes,
postmenopausal hormone use, and hypercholesterolemia. Another multivariate model added dietary factors, including intakes of fruit and vegetables, alcohol, fiber, folate, and saturated fat. A fourth model included dietary lycopene intake to
examine whether any association between plasma lycopene and
CVD may be independent of its dietary component.
Linear trend tests across quartiles of plasma lycopene concentrations were tested by using the median concentrations for
each quartile as an ordinal variable. Post hoc tests comparing
quartiles 3 and 4 versus quartiles 1 and 2 for CVD, and
quartiles 24 versus quartile 1 for important vascular events,
were also conducted in light of an apparent threshold of plasma
lycopene above which there was a possible reduced risk of
CVD. Because of concern that discrepant results may exist for
angina pectoris, we repeated the models with 339 case-control
pairs after excluding 144 cases of angina pectoris. We also
examined whether other plasma carotenoids besides lycopene



or retinol added individually to a multivariate model attenuated

the association between plasma lycopene and the risk of CVD.
We then compared multivariate models for plasma lycopene
with those for retinol, -cryptoxanthin, and lutein/zeaxanthin,
with each biochemical marker divided into fourths among the
controls and entered into a separate model. All analyses were
conducted with the use of SAS (version 8; SAS Institute Inc,
Cary, NC).

Of the 483 cases identified with CVD, there were 109 cases
of MI, 112 cases of stroke (86 ischemic, 24 hemorrhagic, and
2 other), 85 cases of revascularization, 33 cases of CVD death,
and 144 cases of angina pectoris. The mean (SD) follow-up
time was 4.8 2.4 y for cases and controls matched for
follow-up time. Baseline lifestyle, clinical, and dietary risk
factors for cases and controls at the start of the study are
compared in Table 1. As expected, women who developed
CVD tended to weigh more, and more of these women had a
history of hypertension, diabetes mellitus, and hypercholesterolemia than did those women who remained free of CVD.
Women who developed CVD also had higher total cholesterol
concentrations and were more likely to have a parent with a
history of MI before the age of 60 y. Age and smoking status
were identical in the cases and controls because of our matching criteria. Small differences in dietary factors between the
cases and controls were observed.
The mean plasma lycopene concentration in women who
developed CVD was 16.2 7.6 g/dL and in the controls was
16.9 8.2 g/dL. Plasma lycopene concentrations ranged
from 2.3 to 81.4 g/dL in the 483 controls free of CVD and
were significantly correlated with dietary lycopene concentrations (r 0.14, P 0.004), although the magnitude of this
age-adjusted Spearman correlation coefficient was low and suggested little or no correlation. Of the major lycopene food sources,
tomato sauce was the only food source with a significant ageadjusted Spearman correlation coefficient (r 0.16, P 0.001)
with plasma lycopene in the controls. Intake of other dietary
factorscheese, refined grains, sodium, and red meatwere not
correlated with plasma lycopene concentrations.
Baseline characteristics of the subjects according to quartiles
of plasma lycopene in the 483 women who remained free of
CVD are shown in Table 2. Women with higher plasma
lycopene concentrations tended to be younger, consume more
alcohol, and weigh less. In contrast, only women in the highest
quartile of plasma lycopene were more likely to be former
smokers. Total cholesterol increased with increasing quartiles
of plasma lycopene (P 0.001), as did the proportion of
women with a history of hypercholesterolemia (P 0.04). No
appreciable differences in dietary factors were observed, other
than lycopene intake (P 0.03), which was higher in higher
quartiles of plasma lycopene. The proportion of women randomly assigned to active treatment were equally distributed
within each quartile of plasma lycopene (all P 0.05).
In the analyses matched for age and smoking and with
adjustment for randomized treatment assignments and plasma
total cholesterol concentrations, the risk of total CVD appeared
to decrease with higher concentrations of plasma lycopene
(Table 3). The RRs of total CVD for women in the lowest to
highest quartiles of plasma lycopene were 1.00 (referent), 0.78,

Baseline characteristics of 483 women who subsequently developed
cardiovascular disease (cases) and an equal number of women who
remained free of cardiovascular disease (controls)

Age (y)1
Smoking status (%)1
BMI (kg/m2)
History of hypertension (%)
History of diabetes mellitus (%)
History of hypercholesterolemia (%)
Plasma total cholesterol (mg/dL)
Parental history of myocardial infarction
at 60 y of age (%)
Exercise (%)
Rarely or never
1 time/wk
13 times/wk
4 times/wk
Alcohol consumption (%)
Rarely or never
13 drinks/mo
16 drinks/wk
1 drink/d
Postmenopausal hormone use (%)
Fruit and vegetable intake (servings/d)
Other dietary factors6
Lycopene intake (g/d)
Total fiber intake (g/d)
Folate intake (g/d)
Saturated fat intake (g/d)
Plasma lycopene (g/dL)
Cholesterol-adjusted plasma lycopene
Plasma -cryptoxanthin (g/dL)
Plasma lutein/zeaxanthin (g/dL)
Plasma retinol (g/dL)

(n 483)

(n 483)

58.8 8.42

58.8 8.4

27.3 5.2
207 40

25.8 4.73
199 364







6.0 3.4

6.1 3.5

8705 5874
19.0 6.0
435 234
20.0 5.2
16.2 7.6

9078 5564
19.2 5.8
414 214
20.0 4.9
16.9 8.2

16.0 7.3
10.6 7.9
16.4 7.0
56.2 16.7

17.1 8.05
11.2 7.7
17.6 9.55
54.1 15.45

Matching variable.
x SD.
Significantly different from cases: 3P 0.0001, 4P 0.001, 5P


Energy adjusted by using the residual method (25).

0.56, and 0.62. Exclusive of cases of angina, the risk reductions

for CVD were slightly stronger, with a 52% risk reduction
among women in the second through fourth quartiles. For total
CVD, despite a borderline significant multivariate linear trend
(P 0.05), the pattern in RRs compelled us to do a post hoc
comparison of women in the upper compared with the lower
half of plasma lycopene concentrations. We found that the
women in the upper half of plasma lycopene had a significant
34% reduction in the risk of total CVD. Similarly, for women
with CVD (exclusive of angina), an L-shaped association was
apparent with a consistent reduction in risk starting at the
second quartile. In fact, women in the upper 3 quartiles of
plasma lycopene had a significant 50% reduction in the risk of



Comparison of baseline characteristics of 483 controls according to quartile of plasma lycopene
Quartile of plasma lycopene

Age (y)
Smoking status (%)
BMI (kg/m2)
History of hypertension (%)
History of diabetes mellitus (%)
History of hypercholesterolemia (%)
Plasma total cholesterol (mg/dL)
Parental history of myocardial infarction
at 60 y of age (%)
Exercise (%)
Rarely or never
1 time/wk
13 times/wk
4 times/wk
Alcohol consumption (%)
Rarely or never
13 drinks/mo
16 drinks/wk
1 drinks/d
Postmenopausal hormone use (%)
Fruit and vegetable intake (servings/d)
Other dietary factors3
Lycopene intake (g/d)
Total fiber intake (g/d)
Folate intake (g/d)
Saturated fat intake (g/d)

1 (n 121)
11.7 g/dL

2 (n 121)
11.716.4 g/dL

3 (n 121)
16.4 21.0 g/dL

4 (n 120)
21.0 g/dL

60.0 8.92

59.0 8.0

58.7 8.7

57.5 7.9

26.4 5.1
188 34

25.5 4.3
197 38

26.0 4.9
203 34

25.2 4.2
209 35
















6.3 4.0

6.3 3.8

5.8 3.0

6.0 3.1


8117 5654
18.9 5.5
395 183
20.1 5.3

8563 4652
19.8 6.6
447 263
19.1 4.6

9683 5440
19.1 5.4
392 194
20.2 4.8

9925 6233
19.1 5.8
423 207
20.6 4.9


ANOVA was used for continuous variables, Cochran-Armitage trend tests were used for dichotomous variables, and chi-square tests were used for
categorical variables.
x SD.
Energy adjusted by using the residual method (25).

CVD (exclusive of angina) compared with those in the lowest

quartile of plasma lycopene. Multivariate models that included
dietary lycopene did not appreciably confound the RRs for
plasma lycopene.
Plasma lycopene was most strongly correlated with plasma
-carotene (Spearman r 0.34), followed by -cryptoxanthin
(r 0.33), lutein/zeaxanthin (r 0.29), -carotene (r 0.28),
and retinol (r 0.07). We also individually and separately added
other plasma carotenoids and retinol to the multivariate model and
found no differences in the RRs, which suggested that these
correlated nutritional biomarkers did not explain the reductions in
total CVD risk for plasma lycopene shown in Table 3.
We then compared multivariate models for plasma lycopene
with retinol and other carotenoids for the risk of CVD (Table
4). Increasing quartiles of plasma retinol, -cryptoxanthin, and
lutein/zeaxanthin were not associated with the risk of CVD.
Results for other plasma markers related to the WHS randomized treatments (-carotene and -carotene) further showed a
distinct inverse association limited to lycopene on CVD among
the carotenoids investigated (data not shown).

Other potential coronary disease risk factors collected in the

WHS that may be potential confounders (blood pressure, multivitamin use, and other dietary factors) did not significantly
affect our findings. Because of the concern that plasma lycopene may have differential effects on the risk of stroke and
nonstroke, we ran separate models for these 2 endpoints. The
multivariate RRs were similar, which suggested similar effects
for plasma lycopene for stroke and nonstroke CVD endpoints.
All tests for interaction between plasma lycopene and potential
modifiers of interest (baseline age, smoking status, dietary
lycopene intake, and each randomized treatment) were not
significant (all P 0.05), because the RRs were not different
within each stratum (data not shown).

Our prospective data indicate an apparent L-shaped association between increasing quartiles of plasma lycopene and a
lower risk of CVD in middle-aged and elderly women that
appears to be independent of major lifestyle, clinical, and



Relative risks (RRs) and 95% CIs of cardiovascular disease (CVD) and important vascular events (myocardial infarction, stroke, and CVD death,
exclusive of angina) according to quartiles of plasma lycopene
Quartile of plasma lycopene
11.7 g/dL
Median plasma lycopene (g/dL)
Total CVD2
Multivariate-adjusted RR4
Multivariate-adjusted RR5,6
Plasma -cryptoxanthin
Plasma lutein/zeaxanthin
Plasma -carotene
Plasma -carotene
Plasma retinol
CVD, exclusive of angina7
Multivariate-adjusted RR4
Multivariate-adjusted RR5


11.716.4 g/dL

16.4 21.0 g/dL



21.0 g/dL

P (linear



1.00 (referent)
1.00 (referent)
1.00 (referent)
1.00 (referent)
1.00 (referent)
1.00 (referent)
1.00 (referent)
1.00 (referent)

0.78 (0.55, 1.11)

0.89 (0.58, 1.37)
0.94 (0.60, 1.49)
0.94 (0.59, 1.50)
0.98 (0.62, 1.56)
0.95 (0.60, 1.50)
0.98 (0.61, 1.55)
0.93 (0.58, 1.47)

0.56 (0.39, 0.82)

0.64 (0.40, 1.00)
0.62 (0.39, 1.00)
0.62 (0.38, 1.01)
0.64 (0.40, 1.05)
0.63 (0.39, 1.02)
0.66 (0.41, 1.08)
0.60 (0.37, 0.97)

0.62 (0.43, 0.90)

0.80 (0.50, 1.26)
0.67 (0.41, 1.11)
0.67 (0.40, 1.13)
0.71 (0.43, 1.18)
0.69 (0.42, 1.15)
0.75 (0.44, 1.28)
0.65 (0.39, 1.07)


0.66 (0.51, 0.87)

0.75 (0.54, 1.05)
0.66 (0.47, 0.95)

1.00 (referent)
1.00 (referent)
1.00 (referent)

0.48 (0.29, 0.80)

0.44 (0.22, 0.85)
0.44 (0.22, 0.88)

0.48 (0.28, 0.82)

0.57 (0.28, 1.16)
0.54 (0.25, 1.16)

0.48 (0.28, 0.82)

0.73 (0.36, 1.47)
0.58 (0.27, 1.24)


0.48 (0.32, 0.73)

0.55 (0.32, 0.94)
0.50 (0.28, 0.90)

For total CVD, post hoc tests compared quartiles 3 and 4 versus quartiles 1 and 2 of plasma lycopene; for important vascular events, post hoc tests
compared quartiles 2 4 versus quartile 1 of plasma lycopene.
n values for cases and controls, respectively, were as follows: quartiles 1 (151 and 121), 2 (124 and 121), 3 (97 and 121), and 4 (111 and 120).
Matched for age and smoking status and adjusted for randomized aspirin treatment, randomized vitamin E treatment, randomized -carotene
treatment, and plasma cholesterol concentration.
Adjusted for the covariates in footnote 3 plus BMI, exercise, postmenopausal hormone use, parental history of MI at 60 y of age, diabetes,
hypertension, and high cholesterol.
Adjusted for the covariates in footnote 4 plus alcohol, fiber, folate, saturated fat, and fruit and vegetable intakes.
Other carotenoids and retinol were added separately into the model.
n values for cases and controls, respectively, were as follows: quartiles 1 (109 and 79), 2 (85 and 97), 3 (69 and 77), and 4 (76 and 86).

dietary risk factors. Women with concentrations of plasma

lycopene greater than the median (16.5 g/dL) had a possible
34% reduction in total CVD compared with those in the lowest
quartile after multivariate adjustment. The magnitude of effect
was stronger after women with angina were excluded. Additional
adjustment for other plasma carotenoids and retinol did not explain the association between plasma lycopene and CVD. Individual results for other plasma carotenoids and retinol showed
that plasma lycopene might have a distinct inverse association
with the risk of CVD that warrants further investigation.

Although most research on lycopene has focused on a potential reduction in the risk of prostate cancer (26), limited
epidemiologic data support a preventive effect of lycopene
against CVD. The broad range of plasma lycopene concentrations in the WHS is at the low end of the range of lycopene
measurements reported in previous observational studies (9,
10, 2735). This may reflect differences in assay methods, diet,
and baseline disease status or between serum and plasma
concentrations of lycopene across study populations. This
raises questions about the absence of any specific evidence of

Multivariate relative risks (RRs) and 95% CIs of cardiovascular disease (CVD) in a comparison of quartiles of plasma lycopene with other plasma
carotenoids and retinol in separate models
Quartile of plasma biomarker
Plasma lycopene (g/dL)
Multivariate-adjusted RR2
Plasma -cryptoxanthin (g/dL)
Multivariate-adjusted RR2
Plasma lutein/zeaxanthin (g/dL)
Multivariate-adjusted RR2
Plasma retinol (g/dL)
Multivariate-adjusted RR2

1.00 (referent)
1.00 (referent)
1.00 (referent)
1.00 (referent)

0.94 (0.60, 1.49)
0.67 (0.41, 1.07)
0.82 (0.51, 1.33)
1.36 (0.83, 2.24)

0.62 (0.39, 1.00)
0.95 (0.58, 1.56)
0.89 (0.55, 1.44)
1.13 (0.65, 1.94)

0.67 (0.41, 1.11)
0.70 (0.41, 1.18)
0.81 (0.47, 1.41)
1.22 (0.69, 2.14)

P (linear trend)

Median value for each plasma biomarker for that quartile.

Matched for age and smoking status and adjusted for randomized aspirin treatment, randomized vitamin E treatment, randomized -carotene
treatment, plasma cholesterol concentration, BMI, exercise, postmenopausal hormone use, parental history of myocardial infarction at 60 y of age,
diabetes, hypertension, high cholesterol, and alcohol, fiber, folate, saturated fat, and fruit and vegetable intakes.



a possible threshold or L-shaped effect for plasma lycopene in

other studies. Two reports suggest that intimal wall thickness
and the risk of MI are lower in persons with higher adipose
tissue concentrations of lycopene (7, 8). Another study of
Finnish men and women noted similar significant decreases in
intimal wall thickness in men but not in women (34). Other
studies report possible inverse associations between serum
lycopene concentrations and a reduced risk of MI (9, 35), CVD
(11, 12, 35), carotid atherosclerosis (10, 13), and aortic atherosclerosis (36).
Despite early promise from initial epidemiologic studies,
little is known regarding the specific biological mechanisms
responsible for the lycopene-associated reduction in CVD.
Lycopene is one of the most potent singlet oxygen quenchers,
which suggests that it may have comparatively stronger antioxidant properties than other major plasma carotenoids (37).
Lycopene may also contribute to the regulation of cholesterol
metabolism in cell cultures because the addition of lycopene to
macrophage cell lines has been shown to augment the activity
of macrophage LDL receptors (3). In 2 intervention studies,
lycopene reduced LDL oxidation (4, 5). However, lycopene
supplementation over several weeks did not reduce LDL cholesterol (4). Lycopene and other carotenoids may also be associated with an acute phase response in atherosclerosis. Of the
carotenoids, only lycopene modulated adhesion molecule expression in human aortic endothelial cell cultures, which suggests a role in mitigating atherogenesis (38). In the crosssectional third National Health and Nutrition Examination
Survey, nonsmoking subjects in the upper 15% of the C-reactive protein distribution had significantly lower serum lycopene
concentrations (28). Another study of elderly nuns also found
that elevated C-reactive protein concentrations were associated
with lower plasma lycopene concentrations (39).
Dietary, lifestyle, and other factors that influence plasma
lycopene concentrations must be considered given the possible
inverse association with CVD. In our study, dietary and plasma
lycopene concentrations were weakly but significantly correlated, whereas other studies reported stronger correlations (29).
Some small-scale dietary intervention studies indicate that lycopene supplements or tomato products may increase plasma
lycopene concentrations in healthy subjects (46). Correlations
among tomato products with plasma lycopene were highest for
tomato sauce, which is consistent with the lipophilic properties
of lycopene that require the simultaneous ingestion of fats for
optimal absorption (40, 41). No other dietary factors were
strongly associated with plasma lycopene. Therefore, beyond
modest effects for tomato sauce in our data, no clear dietary
pattern emerged that greatly affected plasma lycopene concentrations. Other factors related to lycopene absorption and metabolism may also confound the association between plasma
and dietary lycopene. For example, small, dense LDLa potential independent coronary risk factor (42)carries proportionally less lycopene than do LDL particles of other sizes (43).
Some potential limitations of the present study also warrant
discussion. First, we relied on a single baseline measurement of
plasma lycopene, which raises the possibility of regression to
the mean that may bias our RRs toward the null hypothesis,
thus underestimating the observed risk reductions. Second, we
have not directly assessed the long-term stability of plasma
lycopene stored at 140 C in liquid nitrogenchilled freezers
since 1993. However, studies that have compared the long-term

stability of plasma lycopene and other carotenoid measurements support the stability of the biochemical markers used in
this study (44, 45). Third, because we observed a potential
threshold effect in reducing the risk of CVD, we performed
post hoc tests that compared the upper with the lower half of
plasma lycopene concentrations for the risk of total CVD and
the upper 3 quartiles with the lowest quartile for CVD (exclusive of angina). Although such tests yielded results that support
the possibility of a threshold effect, they must be interpreted
with caution because we had not anticipated the observed
pattern of RRs that emerged in our results.
Next, we relied on a composite endpoint for total CVD that
included MI, stroke, revascularization, CVD death, and angina.
Heterogeneity in the results by specific CVD endpoints cannot
be ruled out because we had limited power to assess any single
outcome. The exclusion of angina case-control pairs strengthened the inverse association, which suggested that plasma
lycopene may prevent more serious CVD outcomes. Another
potential limitation was that we lacked information on whether
-carotene or vitamin E treatment affected carotenoid concentrations. However, because the randomized treatment status
was evenly distributed within each quartile of plasma lycopene,
potential confounding was minimized. Finally, residual confounding may also be of concern, as it would in any epidemiologic investigation. However, in addition to matching for age,
smoking, and follow-up time, we comprehensively controlled
for other coronary disease risk factors.
In conclusion, higher plasma lycopene concentrations may
be associated with a lower risk of CVD in middle-aged and
elderly women. Alternatively, plasma lycopene may simply be
a marker for another dietary or nutritional component that
travels with lycopene in foods. It remains unclear whether the
inverse association between plasma lycopene and CVD risk
follows a strict linear dose-response trend or is L-shaped, as
observed in the present study. Further research on the potential
mechanistic actions of lycopene in the prevention of CVD and
other chronic diseases is warranted.
We acknowledge the crucial contributions of the entire staff of the
WHS, including the leadership of David Gordon, Susan Burt, Mary Breen,
Marilyn Chown, Lisa Fields-Johnson, Georgina Friedenberg, Inge Judge,
Jean MacFadyen, Geneva McNair, David Potter, Claire Ridge, and Harriet
Samuelson. We are also indebted to the 39 876 dedicated and committed
participants of the WHS.
HDS helped conceive and design the study, analyze and interpret the
data, draft the article, and obtain funding. JEB helped critically revise the
article, collect and assemble the data, and obtain funding. EPN helped
critically revise the article and collect and assemble the data. JMG helped
conceive and design the study, interpret the data, critically revise the
article, and obtain funding. The authors had no financial or personal
interests related to this research.

1. Arab L, Steck S. Lycopene and cardiovascular disease. Am J Clin Nutr
2. Clinton SK. Lycopene: chemistry, biology, and implications for human health and disease. Nutr Rev 1998;56:3551.
3. Fuhrman B, Elis A, Aviram M. Hypocholesterolemic effect of lycopene and beta-carotene is related to suppression of cholesterol synthesis and augmentation of LDL receptor activity in macrophages. Biochem Biophys Res Commun 1997;233:658 62.
4. Agarwal S, Rao AV. Tomato lycopene and low density lipoprotein
oxidation: a human dietary intervention study. Lipids 1998;33:981 4.
5. Bub A, Watzl B, Abrahamse L, et al. Moderate intervention with












carotenoid-rich vegetable products reduces lipid peroxidation in men.

J Nutr 2000;130:2200 6.
Dugas TR, Morel DW, Harrison EH. Dietary supplementation with
beta-carotene, but not with lycopene, inhibits endothelial cell-mediated oxidation of low-density lipoprotein. Free Radic Biol Med 1999;
26:1238 44.
Kohlmeier L, Kark JD, Gomez-Gracia E, et al. Lycopene and myocardial infarction risk in the EURAMIC Study. Am J Epidemiol
1997;146:618 26.
Gomez-Aracena J, Sloots J, Garcia-Rodriguez A, et al. Antioxidants in
adipose tissue and myocardial infarction in a Mediterranean area. Nutr
Metab Cardiovasc Dis 1997;7:376 82.
Street DA, Comstock GW, Salkeld RM, Schuep W, Klag MJ. Serum
antioxidants and myocardial infarction. Are low levels of carotenoids
and alpha-tocopherol risk factors for myocardial infarction? Circulation 1994;90:1154 61.
Iribarren C, Folsom AR, Jacobs DR Jr, Gross MD, Belcher JD,
Eckfeldt JH. Association of serum vitamin levels, LDL susceptibility
to oxidation, and autoantibodies against MDA-LDL with carotid atherosclerosis. A case-control study. The ARIC Study Investigators
Atherosclerosis Risk in Communities Arterioscler Thromb Vasc Biol
Kristenson M, Zieden B, Kucinskiene Z, et al. Antioxidant state and
mortality from coronary heart disease in Lithuanian and Swedish men:
concomitant cross sectional study of men aged 50. BMJ 1997;314:
629 33.
Schmidt R, Fazekas F, Hayn M, et al. Risk factors for microangiopathy-related cerebral damage in the Austrian stroke prevention study.
J Neurol Sci 1997;152:1521.
Rissanen TH, Voutilainen S, Nyyssonen K, Salonen R, Kaplan GA,
Salonen JT. Serum lycopene concentrations and carotid atherosclerosis: the Kuopio Ischaemic Heart Disease Risk Factor Study. Am J Clin
Nutr 2003;77:133 8.
Dwyer JH, Navab M, Dwyer KM, et al. Oxygenated carotenoid lutein
and progression of early atherosclerosis: the Los Angeles atherosclerosis study. Circulation 2001;103:29227.
Morris DL, Kritchevsky SB, Davis CE. Serum carotenoids and coronary heart disease. The Lipid Research Clinics Coronary Primary
Prevention Trial and Follow-up Study. JAMA 1994;272:1439 41.
Buring JE, Hennekens CH, for the Womens Health Study Research
Group. The Womens Health Study: summary of the study design. J
Myocardial Ischemia 1992;4:279.
Lee IM, Cook NR, Manson JE, Buring JE, Hennekens CH. Betacarotene supplementation and incidence of cancer and cardiovascular
disease: the Womens Health Study. J Natl Cancer Inst 1999;91:
2102 6.
Hennekens CH, Buring JE, Manson JE, et al. Lack of effect of
long-term supplementation with beta carotene on the incidence of
malignant neoplasms and cardiovascular disease. N Engl J Med 1996;
World Health Organization. Ischaemic heart disease registers: report
of the Fifth Working Group, including a second revision of the
operating protocol: Copenhagen, 26-29 April 1971. Copenhagen: Regional Office for Europe, World Health Organization, 1971.
Sowell AL, Huff DL, Yeager PR, Caudill SP, Gunter EW. Retinol,
alpha-tocopherol, lutein/zeaxanthin, beta-cryptoxanthin, lycopene, alpha-carotene, trans-beta-carotene, and four retinyl esters in serum
determined simultaneously by reversed-phase HPLC with multiwavelength detection. Clin Chem 1994;40:411 6.
Allain CC, Poon LS, Chan CS, Richmond W, Fu PC. Enzymatic
determination of total serum cholesterol. Clin Chem 1974;20:470 5.
Roeschlau P, Bernt E, Gruber W. Enzymatic determination of total
cholesterol in serum. Z Klin Chem Klin Biochem 1974;12:226.
Thurnham DI, Davies JA, Crump BJ, Situnayake RD, Davis M. The
use of different lipids to express serum tocopherol: lipid ratios for the
measurement of vitamin E status. Ann Clin Biochem 1986;23:514 20.
Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and
validity of a semiquantitative food frequency questionnaire. Am J
Epidemiol 1985;122:51 65.
Willett WC. Nutritional epidemiology. 2nd ed. New York: Oxford
University Press, 1998.


26. Giovannucci E, Rimm EB, Liu Y, Stampfer MJ, Willett WC. A

prospective study of tomato products, lycopene, and prostate cancer
risk. J Natl Cancer Inst 2002;94:391 8.
27. Ruiz Rejon F, Martin-Pena G, Granado F, Ruiz-Galiana J, Blanco I,
Olmedilla B. Plasma status of retinol, alpha- and gamma-tocopherols,
and main carotenoids to first myocardial infarction: case control and
follow-up study. Nutrition 2002;18:26 31.
28. Kritchevsky SB, Bush AJ, Pahor M, Gross MD. Serum carotenoids
and markers of inflammation in nonsmokers. Am J Epidemiol 2000;
29. Mayne ST, Cartmel B, Silva F, et al. Plasma lycopene concentrations
in humans are determined by lycopene intake, plasma cholesterol
concentrations and selected demographic factors. J Nutr 1999;129:
849 54.
30. Vogel S, Contois JH, Tucker KL, Wilson PW, Schaefer EJ, LammiKeefe CJ. Plasma retinol and plasma and lipoprotein tocopherol and
carotenoid concentrations in healthy elderly participants of the Framingham Heart Study. Am J Clin Nutr 1997;66:950 8.
31. DOdorico A, Martines D, Kiechl S, et al. High plasma levels of alphaand beta-carotene are associated with a lower risk of atherosclerosis:
results from the Bruneck study. Atherosclerosis 2000;153:2319.
32. McQuillan BM, Hung J, Beilby JP, Nidorf M, Thompson PL. Antioxidant vitamins and the risk of carotid atherosclerosis. The Perth
Carotid Ultrasound Disease Assessment study (CUDAS). J Am Coll
Cardiol 2001;38:1788 94.
33. Su LC, Bui M, Kardinaal A, et al. Differences between plasma and
adipose tissue biomarkers of carotenoids and tocopherols. Cancer
Epidemiol Biomarkers Prev 1998;7:1043 8.
34. Rissanen T, Voutilainen S, Nyyssonen K, Salonen R, Salonen JT. Low
plasma lycopene concentration is associated with increased intimamedia thickness of the carotid artery wall. Arterioscler Thromb Vasc
Biol 2000;20:2677 81.
35. Rissanen TH, Voutilainen S, Nyyssonen K, et al. Low serum lycopene
concentration is associated with an excess incidence of acute coronary
events and stroke: the Kuopio Ischaemic Heart Disease Risk Factor
Study. Br J Nutr 2001;85:749 54.
36. Klipstein-Grobusch K, Launer LJ, Geleijnse JM, Boeing H, Hofman
A, Witteman JC. Serum carotenoids and atherosclerosis. The Rotterdam Study. Atherosclerosis 2000;148:49 56.
37. Di Mascio P, Kaiser S, Sies H. Lycopene as the most efficient
biological carotenoid singlet oxygen quencher. Arch Biochem Biophys
1989;274:532 8.
38. Martin KR, Wu D, Meydani M. The effect of carotenoids on the expression of cell surface adhesion molecules and binding of monocytes to
human aortic endothelial cells. Atherosclerosis 2000;150:26574.
39. Boosalis MG, Snowdon DA, Tully CL, Gross MD. Acute phase
response and plasma carotenoid concentrations in older women: findings from the nun study. Nutrition 1996;12:475 8.
40. Giovannucci E, Ascherio A, Rimm EB, Stampfer MJ, Colditz GA,
Willett WC. Intake of carotenoids and retinol in relation to risk of
prostate cancer. J Natl Cancer Inst 1995;87:176776.
41. Shi J, Le Maguer M. Lycopene in tomatoes: chemical and physical
properties affected by food processing. Crit Rev Food Sci Nutr 2000;
40:1 42.
42. Lamarche B, Lemieux I, Despres JP. The small, dense LDL phenotype
and the risk of coronary heart disease: epidemiology, patho-physiology
and therapeutic aspects. Diabetes Metab 1999;25:199 211.
43. Goulinet S, Chapman MJ. Plasma LDL and HDL subspecies are heterogenous in particle content of tocopherols and oxygenated and hydrocarbon carotenoids. Relevance to oxidative resistance and atherogenesis.
Arterioscler Thromb Vasc Biol 1997;17:78696.
44. Comstock GW, Burke AE, Hoffman SC, Norkus EP, Gross M, Helzlsouer KJ. The repeatability of serum carotenoid, retinoid, and tocopherol concentrations in specimens of blood collected 15 years
apart. Cancer Epidemiol Biomarkers Prev 2001;10:65 8.
45. Comstock GW, Burke AE, Norkus EP, Gordon GB, Hoffman SC,
Helzlsouer KJ. Effects of repeated freeze-thaw cycles on concentrations of cholesterol, micronutrients, and hormones in human plasma
and serum. Clin Chem 2001;47:139 42.