Content uploaded by Kathryn M Rexrode
Author content
All content in this area was uploaded by Kathryn M Rexrode
Content may be subject to copyright.
ORIGINAL INVESTIGATION
Adherence to a DASH-Style Diet and Risk
of Coronary Heart Disease and Stroke in Women
Teresa T. Fung, ScD; Stephanie E. Chiuve, ScD; Marjorie L. McCullough, ScD;
Kathryn M. Rexrode, MD; Giancarlo Logroscino, MD, PhD; Frank B. Hu, MD, PhD
Background: The Dietary Approaches to Stop Hyper-
tension (DASH) diet has been shown to lower blood pres-
sure, but little is known about its long-term effect on car-
diovascular end points. Our objective was to assess the
association between a DASH-style diet adherence score
and risk of coronary heart disease (CHD) and stroke in
women.
Methods: In this prospective cohort study, diet was as-
sessed 7 times during 24 years of follow-up (1980-
2004) with validated food frequency questionnaires. A
DASH score based on 8 food and nutrient components
(fruits, vegetables, whole grains, nuts and legumes, low-
fat dairy, red and processed meats, sweetened bever-
ages, and sodium) was calculated. Lifestyle and medical
information was collected biennially with a question-
naire. The Cox proportional hazard model was used to
adjust for potential confounders. The study population
comprised 88 517 female nurses aged 34 to 59 years with-
out a history of cardiovascular disease or diabetes in 1980.
The main outcome measures were the numbers of con-
firmed incident cases of nonfatal myocardial infarction,
CHD death, and stroke.
Results: We documented 2129 cases of incident non-
fatal myocardial infarction, 976 CHD deaths, and 3105
cases of stroke. After adjustment for age, smoking, and
other cardiovascular risk factors, the relative risks of CHD
across quintiles of the DASH score were 1.0, 0.99, 0.86,
0.87, and 0.76 (95% confidence interval, 0.67-0.85)
(P⬍ .001 for trend). The magnitude of risk difference was
similar for nonfatal myocardial infarction and fatal CHD.
The DASH score was also significantly associated with
lower risk of stroke (multivariate relative risks across quin-
tiles of the DASH score were 1.0, 0.92, 0.91, 0.89, and
0.82) (P=.002 for trend). Cross-sectional analysis in a
subgroup of women with blood samples showed that the
DASH score was significantly associated with lower plasma
levels of C-reactive protein (P=.008 for trend) and in-
terleukin 6 (P= .04 for trend).
Conclusion: Adherence to the DASH-style diet is asso-
ciated with a lower risk of CHD and stroke among middle-
aged women during 24 years of follow-up.
Arch Intern Med. 2008;168(7):713-720
T
HE DIETARY APPROACHES TO
Stop Hypertension (DASH)
diet, which is high in fruits
and vegetables, moderate in
low-fat dairy products, and
low in animal protein but with substan-
tial amount of plant protein from le-
gumes and nuts, substantially reduces both
systolic and diastolic blood pressure among
hypertensive and normotensive individu-
als.
1
In fact, this diet pattern showed greater
improvement in blood pressure than one
high in fruits and vegetables only. Subse-
quently, the DASH-Sodium trial found that
additional sodium restriction resulted in
even greater blood pressure reduction.
2
The DASH diet has also been shown to re-
duce low-density lipoprotein cholesterol
levels.
3
The DASH diet is now widely pro-
moted by the National Heart, Lung, and
Blood Institute for the prevention and
treatment of hypertension
4
and is in-
cluded as an example of a healthy eating
pattern in the 2005 Dietary Guidelines for
Americans.
5
Since hypertension and plasma lipid
levels are strong risk factors for cardio-
vascular disease (CVD), the DASH diet
may lower the risk of coronary heart dis-
ease (CHD) and stroke. However, among
members of the PREMIER (Prospective
Registry Evaluating Outcomes After Myo-
cardial Infarction: Events and Recovery)
trial, a dietary index with 3 components
(dairy, saturated fat, and fruits and veg-
etables) developed to measure adherence
CME available online at
www.jamaarchivescme.com
and questions on page 681
Author Affiliations:
Department of Nutrition,
Simmons College, Boston,
Massachusetts (Dr Fung);
Departments of Nutrition
(Drs Fung and Hu) and
Epidemiology (Drs Logroscino
and Hu), Harvard School of
Public Health, Boston;
Channing Laboratory,
Department of Medicine,
Brigham and Women’s Hospital
and Harvard Medical School,
Boston (Drs Chiuve, Rexrode,
and Hu); and Epidemiology and
Surveillance Research,
American Cancer Society,
Atlanta, Georgia
(Dr McCullough).
(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 7), APR 14, 2008 WWW.ARCHINTERNMED.COM
713
©2008 American Medical Association. All rights reserved.
on April 24, 2008 www.archinternmed.comDownloaded from
to the DASH diet was not significantly associated with
blood pressure after accounting for weight loss.
6
It would
be difficult to conduct a long-term randomized clinical
trial with CVD as an end point, especially for primary
prevention. Data from observational studies with indi-
viduals following a diet similar to the DASH pattern may
be helpful in elucidating the long-term effectiveness of
the DASH diet. Therefore, we developed a score that re-
flects adherence to the DASH-style diet. Since materials
are already available for individual diet planning and hence
adherence assessment,
4
our goal in the construct of the
score is to highlight key components of the DASH diet,
including the sodium component, and rank individuals
by their consumption of foods emphasized in the DASH
diet. In this prospective analysis, we assessed the asso-
ciation between adherence to a DASH-style diet and in-
cidence of CHD and stroke during 24 years of follow-up
in the Nurses’ Health Study cohort. We also assessed the
cross-sectional association between the DASH score and
biomarkers of CVD in a subsample of the cohort.
METHODS
STUDY POPULATION
The Nurses’ Health Study (NHS) began in 1976, when 121 700 fe-
male nurses aged 30 to 55 years living in 11 US states responded
to a questionnaire regarding medical, lifestyle, and other health-
related information.
7
Beginning in 1976, questionnaires have been
sent biennially to update this information. Follow-up was com-
plete for more than 95% of the potential person-time up to 2004.
In 1980, the participants completed a 61-item food frequency ques-
tionnaire (FFQ). In 1984, the FFQ was expanded to 116 items. Simi-
lar FFQs were sent in 1986, 1990, 1994, 1998, and 2002.
For this analysis, we included women who completed the 1980
FFQ with fewer than 70 missing items and total energy intake
(as calculated from the FFQ) between 500 and 3500 kcal/d. At
baseline, we excluded those with a history of CHD, stroke, or dia-
betes because diagnoses of these conditions may lead to changes
in diet. After these exclusions, 88 517 women with follow-up from
1980 through 2004 were included in the analyses. This study was
approved by the institutional review board of the Brigham and
Women’s Hospital, Boston, Massachusetts.
DIETARY ASSESSMENT
The FFQs were designed to assess average food intake over the
preceding year. A standard portion size and 9 possible frequency-
of-consumption responses, ranging from “never or less than once
per month” to “6 or more times per day” was given for each
food item. Total energy and nutrient intake was calculated by
summing up energy or nutrients from all foods. Previous vali-
dation studies among members of the NHS revealed good cor-
relations between nutrients assessed by the FFQ and multiple
weeks of food records completed over the preceding year.
8
For
example, correlation coefficients between 1986 FFQ and diet
records obtained in 1986 were 0.68 for saturated fat, 0.48 for
polyunsaturated fat, and 0.78 for crude fiber. The mean cor-
relation coefficient between frequencies of intake of 55 foods
assessed by 2 FFQ 12 months apart was 0.57.
9
For example,
correlation coefficients between FFQs and diet records were
0.69 for broccoli, 0.17 for spinach, and 0.80 for apples.
We constructed the DASH score based on food and nutrients
emphasized or minimized in the DASH diet,
4
focusing on 8 com-
ponents: high intake of fruits, vegetables, nuts and legumes, low-
fat dairy products, and whole grains and low intake of sodium,
sweetened beverages, and red and processed meats. The first 4
components were directly targeted in the DASH diet, which also
included lower consumption of red and processed meats and
greater consumption of whole grains.
10
The original DASH diet
limits saturated fat intake, which would be in part reflected by a
deemphasis on red and processed meats in our score.
We calculated a DASH score for each FFQ. For each of the
components, we classified women into quintiles according to their
intake ranking. Component score for fruits, vegetables, nuts and
legumes, low-fat dairy products, and whole grains is the wom-
en’s quintile ranking. For example, quintile 1 is assigned 1 point
and quintile 5, 5 points. For sodium, red and processed meats,
and sweetened beverages, low intake was desired. Therefore, the
lowest quintile was given a score of 5 points and the highest quin-
tile, 1 point (
Table 1). We then summed up the component
scores to obtain an overall DASH score ranging from 8 to 40. Be-
cause sodium was not measured well with an FFQ, scoring by
quintiles would be least prone to misclassification.
END POINT ASCERTAINMENT
For this analysis, we ascertained incident cases of CHD (nonfa-
tal myocardial infarction [MI] or fatal CHD) and stroke that oc-
curred after the return of the 1980 questionnaire in women but
Table 1. Scoring Criteria for the DASH-Style Diet and Mean
a
Intake for Q1 (Low Consumption) and Q5 (High Consumption)
in the Cohort
Component Foods Scoring Criteria Q1, Servings/d Q5, Servings/d
Fruits All fruits and fruit juices Q1 = 1 point
Q2 = 2 points
Q3 = 3 points
Q4 = 4 points
Q5 = 5 points
0.7 4.1
Vegetables All vegetables except potatoes and legumes 1.1 4.6
Nuts and legumes Nuts and peanut butter, dried beans, peas, tofu 0.3 1.5
Whole grains Brown rice, dark breads, cooked cereal, whole grain cereal,
other grains, popcorn, wheat germ, bran
0.1 2.4
Low-fat dairy Skim milk, yogurt, cottage cheese 0.1 2.3
Sodium
b
Sum of sodium content of all foods in FFQ Reverse scoring:
Q1 = 5 points
Q2 = 4 points
Q3 = 3 points
Q4 = 4 points
Q5 = 1 point
1041 mg 2676 mg
Red and processed meats
b
Beef, pork, lamb, deli meats, organ meats, hot dogs, bacon 0.4 1.8
Sweetened beverages
b
Carbonated and noncarbonated sweetened beverages 0 1.2
Abbreviations: DASH, Dietary Approaches to Stop Hypertension; FFQ, food frequency questionnaire; Q, quintile.
a
Mean of 5 FFQs.
b
Higher quintiles represent higher intake; however, in constructing the DASH score, high intake and high quintiles received lower scores.
(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 7), APR 14, 2008 WWW.ARCHINTERNMED.COM
714
©2008 American Medical Association. All rights reserved.
on April 24, 2008 www.archinternmed.comDownloaded from
before June 1, 2004. We requested permission to review medi-
cal records from women who reported having a nonfatal MI or
stroke on each biennial questionnaire. Physicians unaware of the
self-reported risk factor status systematically reviewed the rec-
ords. For MI, we noted whether it was fatal or nonfatal, and MI
was classified as confirmed if the criteria of the World Health
Organization were met, specifically, symptoms and either elec-
trocardiographic changes or elevated cardiac enzyme levels.
11
We
included confirmed and probable cases for the analyses. Fatal
CHD was confirmed by hospital records or an autopsy or by CHD
listed as the cause of death on the death certificate, if it was listed
as an underlying and the most plausible cause of death and if
evidence of previous CHD was available. Strokes were con-
firmed by medical record review using National Survey of Stroke
criteria,
12
which require a constellation of neurologic deficits, sud-
den or rapid in onset, and duration of at least 24 hours or until
death. We noted whether stroke was ischemic (embolic or throm-
botic) or hemorrhagic (subarachnoid or intracerebral) accord-
ing to Perth Community Stroke Study criteria and based on find-
ings from computed tomography, magnetic resonance imaging,
or autopsy. Deaths were identified from state vital statistics rec-
ords and the National Death Index or reported by the families
and the postal system.
COLLECTION AND ANALYSIS
OF BIOMARKER LEVELS
Blood was collected in 1989-1990. Each willing participant was
sent a blood collection kit containing instructions and needed
supplies (eg, blood tubes and needles). Each participant made
arrangements for blood to be drawn, packaged the sample in an
enclosed cool pack, and sent it to the laboratory by overnight
courier. Almost all the samples arrived within 26 hours of the
blood draw. On their arrival at the laboratory, the whole-blood
samples were centrifuged and aliquotted and stored at tempera-
tures no higher than −80°C. The lifestyles and dietary intakes of
women who returned a blood sample were in general similar to
those who did not provide a blood sample. All biomarkers were
measured in the Clinical Chemistry Laboratory at the Chil-
dren’s Hospital in Boston. High-sensitivity C-reactive protein
(CRP) levels were measured by a latex-enhanced turbidmetric
assay on a Hitachi 911 (Denka Seiken, Tokyo, Japan). Interleu-
kin 6 (IL-6) levels were measured by an ultrasensitive enzyme-
linked immunosorbent assay (R&D Systems, Minneapolis, Min-
nesota). Interassay coefficients of variation were 3.4% to 3.8%
for CRP and 5.8% to 8.2% for IL-6. C-reactive protein measure-
ment was available for 1176 participants and IL-6 measurement
for 1160 participants. Total cholesterol was measured enzymati-
cally,
13
with a coefficient of variation lower than 1.7%. Concen-
trations of fasting triglycerides and high-density lipoprotein cho-
lesterol were analyzed simultaneously on the Hitachi 911, with
a coefficient of variation of 1.75% and 2.5%, respectively. Low-
density lipoprotein cholesterol was determined by a homoge-
neous direct method (Genzyme Corp, Cambridge, Massachu-
setts), with a coefficient of variation lower than 3.1%. We had
490 samples for high-density lipoprotein cholesterol and total
cholesterol, 478 for low-density lipoprotein cholesterol, and 436
for fasting triglycerides.
ASSESSMENT OF COVARIATES
Body mass index (BMI) (calculated as weight in kilograms di-
vided by height in meters squared) was derived from weight re-
ported on each biennial questionnaire and height reported in 1976.
In each biennial questionnaire, we also assessed smoking status
(including number of cigarettes), frequency and number of as-
pirin tablets used, multivitamin intake, and menopausal status
and use of postmenopausal hormones. Leisure-time physical ac-
tivity was measured biennially (beginning in 1986) with a vali-
dated questionnaire asking about the average time spent on 10
common activities. The information is then summed and calcu-
lated as metabolic equivalent hours.
14
In 1980 and 1982, we mea-
sured hours of moderate to vigorous physical activity.
STATISTICAL ANALYSIS
We used Cox proportional hazard modeling to assess the as-
sociation between the DASH score and risk of CHD and stroke,
including separate models for fatal and nonfatal CHD and is-
chemic and hemorrhagic stroke, since the causes of these stroke
types may differ. To reduce random within-person variation
and to best represent long-term dietary intake, we calculated
cumulative means of the DASH score from our repeated FFQ
administrations.
15
For example, the DASH score in 1980 was
used to predict CHD and stroke occurrence from 1980 to 1984,
and the mean score from 1980 and 1984 was used to predict
CHD and stroke risk from 1984 to 1986. We adjusted for the
following potential confounders, which were updated at each
2-year cycle: age (continuous), smoking (never, past, or cur-
rent with cigarette use of 1-14/d, 15-24/d, ⱖ25/d, or missing
[current smokers but unknown quantity]), BMI (quintiles),
menopausal status and postmenopausal hormone use (pre-
menopausal and never, past, or current hormone use), energy
intake (quintiles), multivitamin intake (yes/no), alcohol in-
take (0 g/d, up to 5 g/d, 5-15 g/d, or ⬎15 g/d), family history
of CHD (yes/no), physical activity (quintiles), and aspirin use
(⬍1/wk, 1-2/wk, 3-6/wk, 7-14/wk, or ⱖ15/wk). Statistical analy-
sis was conducted with SAS version 9 statistical software (SAS
Institute Inc, Cary, NC). In secondary regression analysis, we
additionally adjusted for intake of trans fat and omega-3 fatty
acids to explore the influence of the DASH score on CHD and
stroke beyond the score’s association with these 2 fatty acids.
We also stratified the analysis by major risk factors at baseline
to evaluate potential interactions between these factors and the
DASH score in relation to CHD and stroke risk.
RESULTS
During 24 years of follow-up, we documented 2317 cases
of CHD, of which 2129 were nonfatal and 976 fatal, and
2317 cases of stroke, of which 1242 were ischemic and
440 hemorrhagic (the remainder could not be clearly clas-
sified). Women with higher DASH scores tended to use
multivitamins, exercise more, and consume more fiber
and omega-3 fatty acids but less saturated fat, trans fat,
and total energy. They were also less likely to be current
smokers and more likely to report a history of hyperten-
sion (
Table 2). Because the 1980 FFQ has only 66 items,
we present 1990 dietary intake data, which is the mid
point of follow-up.
We observed an inverse association between the DASH
score and risk of both CHD and stroke. After we adjusted
for potential confounders, women in the top quintile of
the DASH score, compared with those in the bottom
quintile, had a relative risk (RR) of 0.76 (95% confidence
interval [CI], 0.67-0.85; P⬍.001 for trend) for CHD
(
Table 3). Additional adjustment for trans fat and omega-3
fatty acids did not materially alter the results. Because his-
tory of hypertension is considered an intermediate vari-
able, we did not adjust for it in the primary analysis. How-
ever, adjustment for history of hypertension or high
cholesterol level did not change the results. This risk re-
duction was significant for both fatal and nonfatal CHD.
(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 7), APR 14, 2008 WWW.ARCHINTERNMED.COM
715
©2008 American Medical Association. All rights reserved.
on April 24, 2008 www.archinternmed.comDownloaded from
Table 2. Age-Adjusted Health and Lifestyle Characteristics in 1980 According to Quintiles of 1980 DASH Score and 1990 Dietary
Intake According to Quintiles of 1990 DASH Score
Characteristic
1980 DASH Score Range
8-20 (Q1)
(n=18 497)
21-22 (Q2)
(n=19 531)
23-25 (Q3)
(n=15 052)
26-28 (Q4)
(n=19 908)
29-38 (Q5)
(n=15 427)
BMI 24.3 24.3 24.3 24.2 24.1
Current smoker, % 38 31 28 24 19
Family history of CHD, % 15 15 16 16 16
Physical activity, METs/wk 10 12 13 16 19
History of hypertension, % 13 14 15 16 16
History of high cholesterol, % 3 4 5 6 7
Multivitamin use, % 27 31 33 37 42
1990 intake
Alcohol intake, g 6 6 5 5 4
Total energy intake, kcal 1631 1671 1740 1810 1910
Glycemic load
a
99 103 105 108 113
Total dietary fiber, g
a
14 16 18 20 22
Saturated fat, g
a
22.0 20.1 18.8 17.6 15.8
Polyunsaturated fat, g
a
10.8 10.7 10.5 10.4 10.1
Monounsaturated fat, g
a
24.4 22.6 21.3 20.2 18.4
Trans fat, g
a
3.3 2.9 2.7 2.4 2.0
Omega-3 fatty acids, g
a
1.1 1.2 1.2 1.2 1.2
Components of DASH score
Fruit, servings/d 1.3 1.9 2.3 2.8 3.5
Vegetables, servings/d 2.2 2.7 3.1 3.6 4.4
Nuts, servings/d 0.5 0.6 0.7 0.8 1.0
Sodium, mg/d
b
2124 2048 1996 1936 1926
Low-fat dairy, servings/d 0.6 0.9 1.1 1.4 1.7
Red and processed meats, servings/d
b
1.0 0.8 0.7 0.6 0.5
Whole grains, servings/d 0.7 1.0 1.3 1.6 2.1
Sweetened beverages, servings/d
b
0.3 0.3 0.2 0.2 0.2
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CHD, coronary heart disease; DASH, Dietary
Approaches to Stop Hypertension; METs, metabolic equivalent tasks; Q, quintile.
a
Energy adjusted.
b
In constructing the DASH score, higher intakes receive lower scores.
Table 3. Relative Risks of Cumulative DASH Score for CHD
Variable
Relative Risk (95% Confidence Interval)
P Value
for TrendQ1 Q2 Q3 Q4 Q5
Total CHD (3105 cases)
No. of cases 686 657 612 599 551
Crude incidence rate, per 100 000 person-years 165 174 144 153 140
DASH score (age and energy intake adjusted) 1 [Reference] 0.87 (0.78-0.97) 0.68 (0.61-0.76) 0.66 (0.59-0.73) 0.53 (0.47-0.60) ⬍.001
DASH score (multivariate adjusted)
a
1 [Reference] 0.99 (0.89-1.11) 0.86 (0.76-0.96) 0.87 (0.78-0.98) 0.76 (0.67-0.85) ⬍.001
DASH score (multivariate adjusted)
a
⫹ omega-3
and trans fat
1 [Reference] 0.98 (0.88-1.10) 0.84 (0.75-0.94) 0.85 (0.76-0.96) 0.73 (0.64-0.84) ⬍.001
Nonfatal CHD (2129 cases) 479 437 397 434 382
Crude incidence rate, per 100 000 person-years 115 116 93 111 97
DASH score (age and energy intake adjusted) 1 [Reference] 0.84 (0.74-0.96) 0.65 (0.57-0.75) 0.71 (0.62-0.81) 0.56 (0.49-0.64) ⬍.001
DASH score (multivariate adjusted)
a
1 [Reference] 0.94 (0.82-1.07) 0.79 (0.69-0.90) 0.91 (0.79-1.04) 0.78 (0.67-0.90) .002
DASH score (multivariate adjusted)
a
⫹ omega-3
and trans fat
1 [Reference] 0.94 (0.82-1.07) 0.79 (0.68-0.90) 0.90 (0.78-1.04) 0.78 (0.66-0.91) .005
Fatal CHD (976 cases) 207 220 215 165 169
Crude incidence rate, per 100 000 person-years 50 58 51 42 43
DASH score (age and energy intake adjusted) 1 [Reference] 0.93 (0.77-1.12) 0.75 (0.62-0.91) 0.55 (0.45-0.67) 0.48 (0.39-0.59) ⬍.001
DASH score (multivariate adjusted)
a
1 [Reference] 1.11 (0.92-1.35) 1.02 (0.84-1.24) 0.80 (0.65-0.99) 0.71 (0.58-0.89) ⬍.001
DASH score (multivariate adjusted)
a
⫹ omega-3
and trans fat
1 [Reference] 1.09 (0.90-1.32) 0.98 (0.80-1.20) 0.75 (0.60-0.94) 0.66 (0.52-0.83) ⬍.001
Abbreviations: CHD, coronary heart disease; DASH, Dietary Approaches to Stop Hypertension; Q, quintile.
a
Adjusted for age, smoking, body mass index, menopausal status and postmenopausal hormone use, energy intake, multivitamin intake, alcohol intake, family
history, physical activity, and aspirin use.
(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 7), APR 14, 2008 WWW.ARCHINTERNMED.COM
716
©2008 American Medical Association. All rights reserved.
on April 24, 2008 www.archinternmed.comDownloaded from
When we stratified our analysis by major risk factors
at baseline for CHD, there are some hints that associa-
tions with the DASH score appeared stronger among nor-
mal weight women than among overweight women
(P= .10 for interaction) and for current smokers than for
nonsmokers (P=.07 for interaction). Although tests for
interactions were not statistically significant, the statis-
tical power to detect interaction is lower in such sub-
group analyses (
Figure 1). However, the association was
similar among women with and without a history of hy-
pertension and active and sedentary women.
For total stroke, the RR comparing the top to bottom
quintiles of the DASH score was 0.82 (95% CI, 0.71-
0.94; P=.002 for trend) (
Table 4). Additional adjust-
ment for history of hypertension and incident diabetes
did not materially change any association (data not
shown). When we stratified the analysis by selected risk
factor status at baseline, the inverse association be-
tween the DASH score may be stronger among smokers
(P= .09 for interaction) and women with hypertension
(P= .05 for interaction) for total stroke (
Figure 2).
In a subsample of the cohort, we analyzed the cross-
sectional relationship between DASH score and several
lipids and biomarkers of inflammation. The DASH score
was not significantly associated with blood lipid levels,
but it was associated with lower CRP and IL-6 levels. Af-
ter adjustment for age, smoking, and other cardiovascu-
lar risk factors, the geometric mean values of CRP across
quintiles of the DASH score were 1.86 mg/L, 1.79 mg/L,
1.85 mg/L, 1.69 mg/L, and 1.48 mg/L (to convert CRP
to nanomoles per liter, multiply by 9.524) (P=.008 for
trend). For IL-6, geometric mean values across quin-
tiles were 1.95 ng/L, 1.82 ng/L, 1.81 ng/L, 2.08 ng/L, and
1.69 ng/L (P= .04 for trend).
COMMENT
In this large prospective cohort study, women with a high
DASH score had lower risks of CHD and stroke. These re-
sults are consistent with data from short-term clinical trials
in which individuals assigned to the DASH diet had lower
systolic and diastolic blood pressures compared with con-
trols.
1,2
Because a randomized clinical trial of the DASH
diet on cardiovascular end points may not be feasible, our
study provides, to our knowledge, the strongest evidence
to date on the long-term benefits of the DASH diet in the
primary prevention of CVD among healthy subjects.
1.1
0.6
0.7
0.8
0.9
1.2
1.3
1.4
1.2
1.3
1.4
0.5
1
[Ref] [Ref] [Ref] [Ref]
[Ref] [Ref][Ref] [Ref]
1
Q1
1.09
0.90
Q2
0.95
0.77
Q3
0.95
0.80
Q4
0.89
0.64
Q5
Quintiles of DASH Score
Relative Risk
11
Q1
1.01
0.95
Q2
0.83
0.87
Q3
0.79
0.91
Q4
0.65
0.80
Q5
Quintiles of DASH Score
1.1
0.6
0.7
0.8
0.9
0.5
11
Q1
0.82
1.06
Q2
0.72
0.91
Q3
0.79
0.88
Q4
0.68
0.76
Q5
Quintiles of DASH Score
Relative Risk
11
Q1
0.91
1.06
Q2
0.81
0.85
Q3
0.88
0.84
Q4
0.65
0.83
Q5
Quintiles of DASH Score
BMI < 25 (P
<.001 for trend)
BMI ≥ 25 (P
=
.06 for trend)
No history of hypertension (P
<.001 for trend)
With history of hypertension (P
=
.001 for trend)
Nonsmokers (P
<.001 for trend)
Smokers (P
<.001 for trend)
Physical activity >median (P
<.09 for trend)
Physical activity ≤median (P
<.001 for trend)
A B
C D
1.0
1.0
Figure 1. Multivariate relative risks of quintiles of Dietary Approaches to Stop Hypertension (DASH) score for coronary heart disease (CHD) according to risk
factor status at baseline (multivariate adjusted for the same factors listed in Table 3, except for the respective stratifying factor). A, Body mass index lower than 25
vs 25 or higher (calculated as weight in kilograms divided by height in meters squared) (P=.10 for interaction); B, nonsmokers vs smokers (P=.07 for
interaction); C, no hypertension vs with hypertension (P=.84 for interaction); and D, physical activity ⬎median vs ⱕmedian (P=.40 for interaction).
Ref indicates reference.
(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 7), APR 14, 2008 WWW.ARCHINTERNMED.COM
717
©2008 American Medical Association. All rights reserved.
on April 24, 2008 www.archinternmed.comDownloaded from
Several components of our DASH score have been linked
to lower blood pressure. In particular, fruits and veg-
etables—or plant foods in general—are associated with
lower blood pressure in both cross-sectional
16
and pro-
spective
17-19
observational studies in US and other popu-
lations. In a 6-month randomized trial of fruit and veg-
etable intake in healthy individuals, a mean increase of 1.4
servings per day resulted in a reduction of 4.0 mm Hg in
systolic and 1.5 mm Hg in diastolic blood pressure, with
greater reduction in the intervention group than in the con-
trol group.
20
On the other hand, meat intake
19
or a dietary
pattern characterized by high red meat and refined grain
intake
21
has been associated with increased blood pressure.
The DASH diet may influence blood lipid levels as well.
Adherence to the original DASH diet has been linked to lower
low-density lipoprotein cholesterol levels, although an un-
desirable reduction in high-density lipoprotein levels was
also observed.
3
Because the original DASH diet targeted 27%
of calories from fat, with a polyunsaturated fat to saturated
fat ratio of 1:1, it is relatively high in carbohydrates. The Op-
timal Macronutrient Intake Trial to Prevent Heart Disease
(OmniHeart) trial, which altered the macronutrient content
of the DASH diet, found that substituting part of the carbo-
hydrate content with either plant protein or unsaturated fat
conferred an additional reduction of blood pressure and im-
provement of lipid profile beyond a diet similar to the DASH
diet.
22
Our DASH score was not associated with serum lipid
levels, even though our scoring criteria emphasized high
whole grain, nut, and legume consumption. However, our
scoring criteria did not reward high intake of unsaturated
fats or plant protein. Thus, although women with a high
score consumed more fiber, intake of polyunsaturated and
monounsaturated fat actually were lower. Giving all com-
ponents equal weight may dilute the effect of certain com-
ponents. In addition, the small sample size in the cross-
sectional analysis may have limited statistical power.
Inflammation is now established as a pathway for ath-
erosclerosis and CVD development.
23
Components of our
DASH score may also mediate cardiovascular pathophysi-
ologic processes through inflammation. Fruits and veg-
etables have been linked to lower CRP levels.
24
Data from
our cohort also showed that dietary patterns high in fruits,
vegetables, legumes, and whole grains were associated
with lower CRP levels.
25,26
In our cross-sectional analy-
ses, higher DASH score was associated with lower plasma
levels of CRP and IL-6.
Dietary patterns that emphasize foods and nutrients
similar to those emphasized in our DASH score, such as
fruits and vegetables and whole grains, have been shown
to reduce the risk of CHD and stroke in our cohort.
27-29
Several components of our DASH score have also been
associated with the risk of CHD or stroke in our cohort
including an inverse association between (1) fruits, veg-
etables, and whole grains and CHD and ischemic stroke
30-33
and (2) nuts and CHD.
34
High intake of red and pro-
cessed meats, which are deemphasized in our DASH score,
have been positively associated with stroke.
27
To our knowledge, there has only been 1 other scor-
ing system to evaluate adherence to the DASH diet.
6
It
comprised dairy, fruits and vegetables, and saturated fat
only and was not associated with blood pressure after ac-
counting for weight loss, perhaps because of the lack of
inclusion of other beneficial food groups such as whole
grains or the limited range of score distribution in the
sample. Our study was the first to our knowledge to ex-
Table 4. Relative Risks of Cumulative DASH Score for Stroke
Variable
Relative Risk (95% Confidence Interval)
P Value
for TrendQ1 Q2 Q3 Q4 Q5
Total stroke
No. of cases (2317 cases) 486 440 485 459 447
Crude incidence rate, per 100 000 person, y 119 119 116 120 116
DASH score (age and energy intake adjusted) 1 [Reference] 0.83 (0.73-0.95) 0.75 (0.66-0.86) 0.70 (0.62-0.80) 0.60 (0.52-0.68) ⬍.001
DASH score (multivariate adjusted)
a
1 [Reference] 0.92 (0.81-1.05) 0.91 (0.80-1.03) 0.89 (0.78-1.02) 0.82 (0.71-0.94) .002
DASH score (multivariate adjusted)
a
⫹ omega-3
and trans fat
1 [Reference] 0.93 (0.81-1.06) 0.91 (0.80-1.04) 0.90 (0.79-1.03) 0.83 (0.71-0.96) .007
Ischemic stroke (1242 cases)
No. of cases 247 229 263 247 256
Crude incidence rate, per 100 000 person-years 61 62 63 65 66
DASH score (age and energy intake adjusted) 1 [Reference] 0.84 (0.70-1.01) 0.78 (0.66-0.93) 0.72 (0.60-0.86) 0.64 (0.53-0.77) ⬍.001
DASH score (multivariate adjusted)
a
1 [Reference] 0.92 (0.76-1.10) 0.92 (0.77-1.10) 0.90 (0.74-1.08) 0.89 (0.73-1.07) .13
DASH score (multivariate adjusted)
a
⫹ omega-3
and trans fat
1 [Reference] 0.93 (0.77-1.12) 0.94 (0.78-1.13) 0.92 (0.76-1.12) 0.92 (0.75-1.12) .30
Hemorrhagic stroke (440 cases)
No. of cases 90 86 106 83 75
Crude incidence rate, per 100 000 person-years 22 23 25 22 19
DASH score (age and energy intake adjusted) 1 [Reference] 0.92 (0.69-1.24) 0.96 (0.73-1.28) 0.77 (0.71-1.05) 0.63 (0.46-0.86) .004
DASH score (multivariate adjusted)
a
1 [Reference] 1.04 (0.77-1.40) 1.18 (0.88-1.57) 1.01 (0.74-1.38) 0.86 (0.62-1.18) .56
DASH score (multivariate adjusted)
a
⫹ omega-3
and trans fat
1 [Reference] 1.03 (0.76-1.39) 1.16 (0.86-1.56) 0.99 (0.72-1.36) 0.83 (0.59-1.16) .45
Abbreviations: DASH, Dietary Approaches to Stop Hypertension; Q, quintile.
a
Adjusted for age, smoking, body mass index, menopausal status and postmenopausal hormone use, energy intake, multivitamin intake, alcohol intake, family
history, physical activity, and aspirin use.
(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 7), APR 14, 2008 WWW.ARCHINTERNMED.COM
718
©2008 American Medical Association. All rights reserved.
on April 24, 2008 www.archinternmed.comDownloaded from
amine the association between a DASH-style diet and clini-
cal CVD outcomes.
Construction of an adherence score for the DASH diet
can take several approaches. The approach that most closely
resembles the recommendations of National Heart, Lung,
and Blood Institute, and the way the DASH diet should
be followed is to determine specific intake levels for each
food group according to energy requirement and then
award points for adhering to the recommended level of
intake. We chose not to use this approach because of the
difficulty in estimating energy needs owing to incom-
plete physical activity information; however, the highest
quintile intakes for dairy, nuts, and fruits and vegetables
were within the recommended targets for DASH. A less
precise approach in constructing a DASH adherence score
is to choose an energy level that best approximates aver-
age energy needs (eg, 2000 kcal/d) of our cohort and as-
signs points for adhering to the recommended level of in-
take for each food group at that energy level. However,
the assignment of points would still be imprecise for in-
dividuals whose energy requirement is far from the level
chosen. The DASH results were published in the mid-
1990s, so we do not expect a substantial portion of our
cohort to follow the DASH diet. Therefore, scoring crite-
ria that use actual food group intake levels were intended
to assess the adherence to the DASH-style pattern rather
than the DASH diet defined by the clinical trials.
In addition, our FFQ is not designed to provide pre-
cise measurements of some dietary components, such as
sodium. Therefore, meaningful cutoffs cannot be deter-
mined for those components of the DASH score. So-
dium, in particular, is difficult to estimate accurately with
an FFQ. However, it is a hallmark of the DASH-Sodium
trial; thus, using the quintile approach would probably
be least prone to misclassification. In constructing the
score, we have tried the actual cutoff approach and found
our results much weakened, likely owing to greater mis-
classification. Hence, we chose to focus on food groups
and nutrients targeted by the DASH diet and award points
for adhering to what it emphasizes rather than precise
compliance to DASH recommendation in the number of
servings from each targeted food group. We believe this
approach addresses the focus of the DASH diet and at the
same time allows for adequate variation in adherence to
examine any potential association with CVD. As it is dif-
ficult to specify the contribution of each food group to
CVD risk, we gave each food group equal weight in the
overall score.
The long follow-up in this analysis allowed us to as-
sess long-term associations between the DASH score and
1.1
0.6
0.7
0.8
0.9
1.2
0.5
11
Q1
0.93
0.91
Q2
0.90
0.91
Q3
0.91
0.88
Q4
0.82
0.79
Q5
Quintiles of DASH Score
Relative Risk
11
Q1
0.89
0.97
Q2
0.78
1.00
Q3
0.77
0.98
Q4
0.67
0.90
Q5
Quintiles of DASH Score
1.1
0.6
0.7
0.8
0.9
1.2
0.5
11
Q1
0.77
0.97
Q2
0.75
0.95
Q3
0.68
0.96
Q4
0.64
0.87
Q5
Quintiles of DASH Score
Relative Risk
11
Q1
0.93
0.84
Q2
1.00
0.71
Q3
0.83
0.84
Q4
0.88
0.70
Q5
Quintiles of DASH Score
BMI <25 (P
=
.009 for trend)
BMI ≥25 (P
=
.06 for trend)
No history of hypertension (P
=
.15 for trend)
With history of hypertension (P
<.001 for trend)
Nonsmokers (P
=
.13 for trend)
Smokers (P
=
.001 for trend)
Physical activity >median (P
=
.001 for trend)
Physical activity ≤median (P
=
.27 for trend)
A B
C D
1.0
1.0
[Ref] [Ref] [Ref] [Ref]
[Ref] [Ref] [Ref] [Ref]
Figure 2. Multivariate relative risks of quintiles of Dietary Approaches to Stop Hypertension (DASH) score for stroke according to risk factor status at baseline
(multivariate adjusted for the same factors listed in Table 3, except for the respective stratifying factor). A, Body mass index lower than 25 vs 25 or higher
(calculated as weight in kilograms divided by height in meters squared) (P = .44 for interaction); B, nonsmokers vs smokers (P = .09 for interaction); C, no
hypertension vs with hypertension (P = .05 for interaction); and D, physical activity ⬎median vs ⱕmedian (P=.15 for interaction). Ref indicates reference.
(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 7), APR 14, 2008 WWW.ARCHINTERNMED.COM
719
©2008 American Medical Association. All rights reserved.
on April 24, 2008 www.archinternmed.comDownloaded from
CVD. We used repeated measurements of diet to obtain
a better assessment of long-term overall diet and to re-
duce measurement error. The prospective nature of this
analysis reduces the probability of recall bias and selec-
tion bias. In addition, a high rate of follow-up reduced
potential bias due to loss to follow-up. Nevertheless, our
results need to be replicated in other populations. Be-
cause other diet indexes such as the Alternate Healthy
Eating Index
29
and Mediterranean Diet Score
35
have also
been strongly associated with lower risk of CVD, the value
of the DASH score as a predictor of CHD and stroke should
be measured against other indexes. Although members
of this cohort are middle-aged nurses, we expect the iden-
tified associations should be generalizable to middle-
aged American women because the biological effects of
dietary patterns should be the same for them.
In conclusion, a diet that resembles the DASH eating
plan, with high intake of fruits, vegetables, and whole
grains, moderate intake of legumes, nuts, and low-fat dairy
products, and low intake of red and processed meats and
sodium, was significantly associated with lower risk of
CHD and stroke in women.
Accepted for Publication: October 22, 2007.
Correspondence: Teresa T. Fung, ScD, Department of
Nutrition, Simmons College, 300 The Fenway, Boston,
MA 02115 (fung@simmons.edu).
Author Contributions: Dr Hu had full access to all the
data in the study and takes responsibility for the integ-
rity of the data and the accuracy of the data analysis. Study
concept and design: Fung and Hu. Acquisition of data:
Rexrode, Logroscino, and Hu. Analysis and interpreta-
tion of data: Fung, Chiuve, McCullough, Rexrode, Logro-
scino, and Hu. Drafting of the manuscript: Fung and
Logroscino. Critical revision of the manuscript for impor-
tant intellectual content: Fung, Chiuve, McCullough,
Rexrode, and Hu. Statistical analysis: Fung, Chiuve,
Rexrode, and Hu. Obtained funding: Rexrode and Hu. Ad-
ministrative, technical, and material support: Rexrode and
Hu. Study supervision: Hu.
Financial Disclosure: None reported.
Funding/Support: This study was funded by grants
CA87969 and HL60712 from the National Institutes of
Health.
REFERENCES
1. Appel LJ, Moore TJ, Obarzanek E, et al; DASH Collaborative Research Group.
A clinical trial of the effects of dietary patterns on blood pressure. N Engl J Med.
1997;336(16):1117-1124.
2. Sacks FM, Svetkey LP, Vollmer WM, et al; DASH-Sodium Collaborative Research
Group. Effects on blood pressure of reduced dietary sodium and the Dietary Ap-
proaches to Stop Hypertension (DASH) diet. N Engl J Med. 2001;344(1):3-10.
3. Obarzanek E, Sacks FM, Vollmer WM, et al. Effects on blood lipids of a blood
pressure-lowering diet: the Dietary Approaches to Stop Hypertension (DASH)
Trial. Am J Clin Nutr. 2001;74(1):80-89.
4. United States Department of Health and Human Services, National Institutes of
Health, National Heart Lung, and Blood Institute. Your Guide to Lowering Your
Blood Pressure With DASH . http://www.nhlbi.nih.gov/health/public/heart/hbp
/dash/. Accessed December 5, 2006.
5. US Department of Health and Human Services, US Department of Agriculture.
Dietary Guidelines for Americans 2005. Washington, DC: US Government Print-
ing Office; 2005.
6. Obarzanek E, Vollmer WM, Appel LJ, et al. Effects of individual components of
multiple behavior changes: the PREMIER Trial. Am J Health Behav. 2007;31
(5):545-560.
7. Colditz GA, Martin P, Stampfer MJ, et al. Validation of questionnaire information
on risk factors and disease outcomes in a prospective cohort of women. Am J
Epidemiol. 1986;123(5):894-900.
8. Willett WC. Nutritional Epidemiology. New York, NY: Oxford University Press; 1998.
9. Salvini S, Hunter DJ, Sampson L, et al. Food-based validation of a dietary ques-
tionnaire: the effects of week-to-week variation in food consumption. Int J
Epidemiol. 1989;18(4):858-867.
10. Karanja NM, Obarzanek E, Lin PH, et al; DASH Collaborative Research Group.
Descriptive characteristics of the dietary patterns used in the Dietary Approaches
to Stop Hypertension Trial. J Am Diet Assoc. 1999;99(8)(suppl):S19-S27.
11. Rose GA, Blackburn H, Gillum R, Prineas R. Cardiovascular Survey Methods.
Geneva, Switzerland: World Health Organization; 1982.
12. Walker AE, Robins M, Weinfeld FD. The National Survey of Stroke: clinical findings.
Stroke. 1981;12(2, pt 2)(suppl 1):I13-I44.
13. Allain CC, Poon LS, Chan CSG, Richmond W, Fu PC. Enzymatic determination of
total serum cholesterol. Clin Chem. 1974;20(4):470-475.
14. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities:
an update of activity codes and MET intensities. Med Sci Sports Exerc. 2000;
32(9)(suppl):S498-S504.
15. Hu FB, Stampfer MJ, Rimm E, et al. Dietary fat and coronary heart disease: a
comparison of approaches for adjusting for total energy intake and modeling re-
peated dietary measurements. Am J Epidemiol. 1999;149(6):531-540.
16. Alonso A, de la Fuente C, Martin-Arnau AM, de Irala J, Martinez JA, Martı´nez-
Gonza´lez MA. Fruit and vegetable consumption is inversely associated with blood
pressure in a Mediterranean population with a high vegetable-fat intake: the
Seguimiento Universidad de Navarra (SUN) Study. Br J Nutr. 2004;92(2):311-
319.
17. Alonso A, Beunza JJ, Bes-Rastrollo M, Pajares RM, Martı´nez-Gonza´lez MA.
Vegetable protein and fiber from cereal are inversely associated with the risk of
hypertension in a Spanish cohort. Arch Med Res. 2006;37(6):778-786.
18. Miura K, Greenland P, Stamler J, Liu K, Daviglus ML, Nakagawa H. Relation of
vegetable, fruit, and meat intake to 7-year blood pressure change in middle-
aged men: the Chicago Western Electric Study. Am J Epidemiol. 2004;159
(6):572-580.
19. Steffen LM, Kroenke CH, Yu X, et al. Associations of plant food, dairy product,
and meat intakes with 15-y incidence of elevated blood pressure in young black
and white adults: the Coronary Artery Risk Development in Young Adults (CARDIA)
Study. Am J Clin Nutr. 2005;82(6):1169-1177.
20. John JH, Ziebland S, Yudkin P, Roe LS, Neil HA; Oxford Food and Vegetable Study
Group. Effects of fruit and vegetable consumption on plasma antioxidant con-
centrations and blood pressure: a randomised controlled trial. Lancet. 2002;
359(9322):1969-1974.
21. van Dam RM, Grievink L, Ocke MC, Feskens EJ. Patterns of food consumption
and risk factors for cardiovascular disease in the general Dutch population. Am
J Clin Nutr. 2003;77(5):1156-1163.
22. Appel LJ, Sacks FM, Carey VJ, et al; OmniHeart Collaborative Research Group.
Effects of protein, monounsaturated fat, and carbohydrate intake on blood pres-
sure and serum lipids: results of the OmniHeart randomized trial. JAMA. 2005;
294(19):2455-2464.
23. Libby P, Ridker PM, Maseri A. Inflammation and atherosclerosis. Circulation. 2002;
105(9):1135-1143.
24. Esmaillzadeh A, Kimiagar M, Mahrabi Y, Azadbakht L, Hu FB, Willett WC. Fruit
and vegetable intakes, C-reactive protein, and the metabolic syndrome. Am J Clin
Nutr. 2006;84(6):1489-1497.
25. Fung TT, McCullough ML, Newby PK, et al. Diet-quality scores and plasma con-
centrations of markers of inflammation and endothelial dysfunction. AmJClin
Nutr. 2005;82(1):163-173.
26. Lopez-Garcia E, Schulze MB, Fung TT, et al. Major dietary patterns are related to
plasma concentrations of markers of inflammation and endothelial dysfunction.
Am J Clin Nutr. 2004;80(4):1029-1035.
27. Fung TT, Stampfer MJ, Manson JE, Rexrode KM, Willett WC, Hu FB. Prospec-
tive study of major dietary patterns and stroke risk in women. Stroke. 2004;
35(9):2014-2019.
28. Fung TT, Willett WC, Stampfer MJ, Manson JE, Hu FB. Dietary patterns and the risk
of coronary heart disease in women. Arch Intern Med. 2001;161(15):1857-1862.
29. McCullough ML, Feskanich D, Stampfer MJ, et al. Diet quality and major chronic
disease risk in men and women: moving toward improved dietary guidance. Am
J Clin Nutr. 2002;76(6):1261-1271.
30. Joshipura KJ, Ascherio A, Manson JE, et al. Fruit and vegetable intake in relation
to risk of ischemic stroke. JAMA. 1999;282(13):1233-1239.
31. Joshipura KJ, Hu FB, Manson JE, et al. The effect of fruit and vegetable intake on
risk for coronary heart disease. Ann Intern Med. 2001;134(12):1106-1114.
32. Liu S, Manson JE, Stampfer MJ, et al. Whole grain consumption and risk of ische-
mic stroke in women: a prospective study. JAMA. 2000;284(12):1534-1540.
33. Liu S, Stampfer MJ, Hu FB, et al. Whole-grain consumption and risk of coronary
heart disease: results from the Nurses’ Health Study. Am J Clin Nutr. 1999;
70(3):412-419.
34. Hu FB, Stampfer MJ, Manson JE, et al. Frequent nut consumption and risk of
coronary heart disease in women: prospective cohort study. BMJ. 1998;317
(7169):1341-1345.
35. Martı´nez-Gonza´lez MA, Fernandez-Jarne E, Serrano-Martinez M, Marti A, Mar-
tinez JA, Martin-Moreno JM. Mediterranean diet and reduction in the risk of a
first acute myocardial infarction: an operational healthy dietary score. Eur J Nutr.
2002;41(4):153-160.
(REPRINTED) ARCH INTERN MED/ VOL 168 (NO. 7), APR 14, 2008 WWW.ARCHINTERNMED.COM
720
©2008 American Medical Association. All rights reserved.
on April 24, 2008 www.archinternmed.comDownloaded from
