ORIGINAL CONTRIBUTION

Effects of Comprehensive Lifestyle

Modification on Blood Pressure Control
Main Results of the PREMIER Clinical Trial
Writing Group of the PREMIER Context Weight loss, sodium reduction, increased physical activity, and limited alcohol
Collaborative Research Group
intake are established recommendations that reduce blood pressure (BP). The Dietary Ap-
proaches to Stop Hypertension (DASH) diet also lowers BP. To date, no trial has evalu-

H
IGH BLOOD PRESSURE (BP) IS ated the effects of simultaneously implementing these lifestyle recommendations.
a common, powerful, and Objective To determine the effect on BP of 2 multicomponent, behavioral inter-
independent risk factor for ventions.
cardiovascular disease Design, Setting, and Participants Randomized trial with enrollment at 4 clinical
(CVD). Almost 50 million US adults, centers ( January 2000-June 2001) among 810 adults (mean [SD] age, 50 [8.9] years;
or approximately 25% of the US adult 62% women; 34% African American) with above-optimal BP, including stage 1 hy-
population, have hypertension, de- pertension (120-159 mm Hg systolic and 80-95 mm Hg diastolic), and who were not
fined as BP of 140/90 mm Hg or taking antihypertensive medications.
higher and/or current use of antihy- Intervention Participants were randomized to one of 3 intervention groups: (1) “es-
pertensive medication.1 The preva- tablished,” a behavioral intervention that implemented established recommendations
lence of hypertension increases pro- (n=268); (2) “established plus DASH,”which also implemented the DASH diet (n=269);
gressively with age, so that more than and (3) an “advice only” comparison group (n=273).
half of all individuals aged 60 years or Main Outcome Measures Blood pressure measurement and hypertension status
older in the United States have hyper- at 6 months.
tension.2 The estimated lifetime risk of Results Both behavioral interventions significantly reduced weight, improved fit-
developing hypertension is 90%.3 ness, and lowered sodium intake. The established plus DASH intervention also in-
Above-optimal BP that is not in the creased fruit, vegetable, and dairy intake. Across the groups, gradients in BP and hy-
hypertensive range also confers ex- pertensive status were evident. After subtracting change in advice only, the mean net
cess CVD risk.4 In fact, almost a third reduction in systolic BP was 3.7 mm Hg (P⬍.001) in the established group and 4.3
of BP-related deaths from coronary mm Hg (P⬍.001) in the established plus DASH group; the systolic BP difference be-
tween the established and established plus DASH groups was 0.6 mm Hg (P=.43).
heart disease are estimated to occur in
Compared with the baseline hypertension prevalence of 38%, the prevalence at 6 months
nonhypertensive individuals with a sys- was 26% in the advice only group, 17% in the established group (P=.01 compared
tolic BP of 120 to 139 mm Hg or dia- with the advice only group), and 12% in the established plus DASH group (P⬍.001
stolic BP of 80 to 89 mm Hg.5 There- compared with the advice only group; P=.12 compared with the established group).
fore, reduction of BP to optimal levels, The prevalence of optimal BP (⬍120 mm Hg systolic and ⬍80 mm Hg diastolic) was
control of hypertension, and preven- 19% in the advice only group, 30% in the established group (P=.005 compared with
tion of the age-related increase in BP re- the advice only group), and 35% in the established plus DASH group (P⬍.001 com-
main major public health priorities. pared with the advice only group; P=.24 compared with the established group).
Current national recommendations Conclusion Individuals with above-optimal BP, including stage 1 hypertension, can make
for the prevention and treatment of high multiple lifestyle changes that lower BP and reduce their cardiovascular disease risk.
BP emphasize nonpharmacological JAMA. 2003;289:2083-2093 www.jama.com
therapy, also termed “lifestyle modifi-
cation.”6,7 Lifestyle modifications that tion, and the Dietary Approaches to Author Affiliations: PREMIER Authors and Group
effectively lower BP are weight loss, re- Stop Hypertension (DASH) diet.6-8 The Members are listed at the end of this article.
Corresponding Author and Reprints: Lawrence J. Ap-
duced sodium intake, increased physi- DASH diet emphasizes consumption of pel, MD, MPH, Departments of Medicine; Epidemi-
cal activity, limited alcohol consump- fruits, vegetables, and low-fat dairy ology, and International Health (Human Nutrition),
Johns Hopkins Medical Institutions, 2024 E Monu-
products; includes whole grains, poul- ment St, Suite 2-645, Baltimore, MD 21205-2223
For editorial comment see p 2131.
try, fish, and nuts; and is reduced in fats, (e-mail: [email protected]).

©2003 American Medical Association. All rights reserved. (Reprinted) JAMA, April 23/30, 2003—Vol 289, No. 16 2083

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

red meat, sweets, and sugar-contain- pating institutions included the Na- and diastolic BP of 80 to 95 mm Hg,
ing beverages. As such, the DASH diet tional Heart, Lung, and Blood Institute based on the mean BP across 3 screen-
has reduced levels of total fat, satu- Project Office (Bethesda, Md), the co- ing visits. Nonhypertensive individu-
rated fat, and cholesterol and in- ordinating center (Kaiser Permanente als with above optimal BP (120-139
creased levels of potassium, calcium, Center for Health Research in Port- mm Hg systolic and/or 80-89 mm Hg
magnesium, fiber, and protein.9 land, Ore), and 4 clinical centers (Johns diastolic) were included because of the
These lifestyle modifications are rec- Hopkins University, Baltimore, Md; potential for preventing hypertension
ommended in nonhypertensive indi- Pennington Biomedical Research Cen- and the excess CVD risk associated with
viduals with above-optimal BP. Life- ter, Baton Rouge, La; Duke University BP in this range.4 Individuals with stage
style modification is also recommended Medical Center, Durham, NC; and Kai- 1 hypertension (140-159 mm Hg sys-
as initial therapy in stage 1 hyperten- ser Permanente Center for Health Re- tolic and/or 90-95 mm Hg diastolic) were
sion (for up to 12 months in those with- search, Portland, Ore). Institutional re- included because of the potential for
out other risk factors [risk class A] or view boards at each center and an nonpharmacological control of hyper-
for up to 6 months in those with other external protocol review committee ap- tension.
risk factors [risk class B]).7 For indi- proved the protocol. Each participant Other inclusion criteria were 25 years
viduals taking BP medication, lifestyle provided written consent. of age or older and body mass index
modification is recommended as ad- (BMI) of 18.5 to 45.0 (measured as
junctive therapy to lower BP. Al- Study Participants weight in kilograms divided by the
though lifestyle therapies are gener- The target population consisted of gen- height in meters squared). Major exclu-
ally recommended as a group, no erally healthy adults with above opti- sion criteria were regular use of drugs
previous trial has evaluated the effects mal BP including individuals with stage that affect BP, JNC-VI risk category C
of implementing these recommenda- 1 hypertension who met Joint National (target organ damage and/or diabetes),
tions simultaneously, and no trial has Committee on Detection, Evaluation, use of weight-loss medications, prior car-
tested the feasibility of implementing and Treatment of High Blood Pressure diovascular event, congestive heart fail-
the DASH diet in free-living persons. (JNC-VI) criteria for at least a 6-month ure, angina, cancer diagnosis or treat-
trial of nonpharmacological therapy.7 ment in past 2 years, consumption of
METHODS Persons were eligible if they were not more than 21 alcoholic drinks per week,
The rationale for the PREMIER clini- taking antihypertensive medication and and pregnancy, planned pregnancy, or
cal trial10 has been published. Partici- had a systolic BP of 120 to 159 mm Hg lactation. Although individuals with dia-
betes were excluded, persons with other
cardiovascular risk factors (ie, ciga-
Figure 1. Participant Flow in the PREMIER Clinical Trial
rette smoking and dyslipidemia) could
3964 Individuals Screened
enroll. Vitamin and mineral supple-
ment use was not an exclusion.
3154 Ineligible
194 BP Too High Trial Conduct
2103 BP Too Low
857 Other Exclusions Participants were recruited using mass
mailings, community-based screen-
810 Randomized
ing, and mass-media announcements.
Enrollment began in January 2000 and
273 Assigned to Advice Only 268 Assigned to Established 269 Assigned to Established +
ended in June 2001. For logistical pur-
Group Intervention Group DASH Intervention Group poses, participants were enrolled in 3
or 4 cohorts at each center. Baseline data
3-Month Assessment 3-Month Assessment 3-Month Assessment were collected during 3 screening vis-
250 Had BP Measurement 250 Had BP Measurement 252 Had BP Measurement
its and a randomization visit, each
6-Month Assessment 6-Month Assessment 6-Month Assessment
scheduled at least 7 days apart. Fol-
259 BP Measurement 253 BP Measurement 253 BP Measurement low-up data were collected at 1 visit 3
256 Weight 251 Weight 249 Weight
219 24-h Urine 220 24-h Urine 223 24-h Urine months after randomization and at 3
236 Fitness Test 229 Fitness Test 238 Fitness Test visits 6 months after randomization.
243 Dietary Recall 233 Dietary Recall 236 Dietary Recall
242 Physical Activity 233 Physical Activity 239 Physical Activity Participant flow during the trial is
253 Blood Sample 242 Blood Sample 249 Blood Sample
shown in FIGURE 1.
273 Included in Primary Analyses 268 Included in Primary Analyses 269 Included in Primary Analyses
of Blood Pressure of Blood Pressure of Blood Pressure Randomization
Randomization assignments were made
BP indicates blood pressure; DASH, Dietary Approaches to Stop Hypertension. centrally by a computer program. Clini-
2084 JAMA, April 23/30, 2003—Vol 289, No. 16 (Reprinted) ©2003 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

cal center staff then notified partici- the DASH diet. In this intervention, par- assessment [2 BP measurements at 1
pants of their assigned group. Assign- ticipant goals designed to accomplish visit], and 6-month assessment [6 BP
ments were stratified by clinic and the DASH diet were increased con- measurements across 3 visits]).
hypertension status; the randomiza- sumption of fruits and vegetables (9-12 Weight was measured using a cali-
tion block size was 24. Eligible partici- servings/d) and low-fat dairy products brated scale, and height was measured
pants were randomly assigned to 1 of (2-3 servings/d), and reduced intake of using a wall-mounted stadiometer. Other
3 groups: (1) an “advice only” com- saturated fat (ⱕ7% of energy) and total data included the Rose Angina question-
parison group; (2) a behavioral inter- fat (ⱕ25% of energy). The established naire14; a medication questionnaire; a
vention, termed “established” that intervention did not have goals for fruit, symptoms/adverse effects question-
implemented traditional lifestyle rec- vegetable, or dairy intake; the goal for naire; 24-hour urine collections for so-
ommendations,11 ie, weight loss among saturated fat was 10% of energy or less, dium, potassium, phosphorus, and urea
those who were overweight, reduced so- and the goal for total fat was 30% of en- nitrogen; submaximal treadmill tests;
dium intake, increased physical activ- ergy or less. To achieve weight loss, waist circumference; 24-hour dietary re-
ity, and limited alcohol intake among both interventions emphasized in- calls; fasting blood analysis; and 7-day
those who drank alcohol; or (3) a be- creased physical activity and reduced physical activity recalls. Each of these
havioral intervention, termed “estab- total energy intake; in addition to these measurements was obtained at baseline
lished plus DASH” that implemented strategies, the established plus DASH and 6 months after randomization.
the same traditional recommenda- intervention also emphasized substi- Intake of nutrients and food groups
tions plus the DASH diet.7 tution of fruits and vegetables for high- was assessed from unannounced 24-
fat, high-calorie foods. hour dietary recalls conducted by tele-
Advice Only Group The format and contact pattern of the phone interviewers.15 Two recalls (one
An interventionist, typically a regis- established and established plus DASH obtained on a weekday and the other
tered dietitian, discussed nonpharma- interventions were identical. During the on a weekend day) were obtained at
cological factors that affect BP (weight, initial 6 months, there were 18 face-to- baseline and 6 months by the Diet As-
sodium intake, physical activity, and the face intervention contacts (14 group sessment Center of Pennsylvania State
DASH diet) and provided printed edu- meetings and 4 individual counseling University. Nutrient and food group in-
cational materials. This advice was pro- sessions). Participants in both inter- takes were then calculated using the
vided in a single 30-minute individual ventions kept food diaries, recorded Nutrition Data System Version NDS-R
session immediately following random- physical activity, and monitored calo- 1998 (University of Minnesota). Bio-
ization. Counseling on behavior change rie and sodium intake. Participants in markers of dietary intake were 24-
was not provided. No further contact the established plus DASH group also hour urinary excretion of sodium, po-
with the interventionist occurred un- monitored intake of fruits, vegetables, tassium (reflecting fruit and vegetable
til after completion of the data collec- and dairy products and monitored their intake), phosphorus (reflecting dairy
tion visits at 6 months. intake of fat. intake), and urea nitrogen (reflecting
protein intake). Alcohol intake was ob-
Behavioral Interventions Measurements tained from questionnaire.
Participant goals for both the estab- Staff who were masked to randomiza- Cardiorespiratory fitness was as-
lished and established plus DASH in- tion assignment collected measure- sessed using a submaximal treadmill ex-
terventions were as follows: (1) weight ments. Blood pressure measurements ercise test developed for the PREMIER
loss of at least 15 lb (6.8 kg) at 6 months were obtained by trained, certified indi- trial. This 2-stage, 10-minute protocol
for those with a BMI of at least 25, (2) viduals who used a random zero sphyg- was designed to achieve an age- and sex-
at least 180 min/wk of moderate- momanometer. The BP measurement specific effort of moderate intensity.16
intensity physical activity, (3) daily in- protocol was similar to protocols used The first stage achieved a light-
take of no more than 100 mEq of di- in prior studies.12,13 After the partici- intensity effort (approximately 40% es-
etary sodium, and (4) daily intake of pant sat quietly for 5 minutes, the ob- timated maximal metabolic equiva-
1 oz or less of alcohol (2 drinks) for server measured BP in the right arm with lents [METs]), followed by a second
men and 1⁄2 oz of alcohol (1 drink) for an appropriately sized cuff. At each visit, stage of moderate intensity (approxi-
women. 2 BP measurements separated by at least mately 60% estimated maximal METs).
The established and the established 30 seconds were obtained. Systolic BP The main fitness outcome was heart rate
plus DASH interventions differed from was the appearance of the first Korot- at the end of stage 2 or the last avail-
each other with respect to certain di- koff sound, and diastolic BP was the dis- able heart rate from stage 1 for partici-
etary goals and the strategies to achieve appearance of Korotkoff sounds. At each pants who did not complete stage 2. A
weight loss. Only the participants in the assessment point, BP was the mean of all 7-day physical activity recall was used
established plus DASH intervention re- available measurements (baseline [8 BP to assess physical activity.17 Partici-
ceived instruction and counseling on measurements across 4 visits], 3-month pants who reported 35 kcal/kg or less
©2003 American Medical Association. All rights reserved. (Reprinted) JAMA, April 23/30, 2003—Vol 289, No. 16 2085

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

daily of physical activity were classi- tions, indicators of clinical center and Based on a planned sample size of 800
fied as sedentary.18 cohort, and baseline BP. In prespeci- (267 per group), the study had 90%
fied subgroup analyses (hypertensive power to detect pairwise between-
Specific Aims and Outcomes and nonhypertensive), the models also group differences in systolic BP of 1.6
The specific aims of the trial were to test included a main effect for the sub- to 1.8 mm Hg in the whole sample, 3.2
the effects of the established interven- group indicator and interactions be- to 3.6 mm Hg among hypertensive
tion compared with the advice only in- tween this indicator and the 2 treat- participants (assuming that 30% of
tervention; the effects of the estab- ment group indicators. The effects of the individuals in the sample were
lished plus DASH intervention compared the interventions did not differ by clini- hypertensive), and 1.7 to 1.9 mm Hg
with the advice only intervention; and cal center (P=.62 for center ⫻ treatment among nonhypertensive participants.
the effects of the established plus DASH interaction for systolic BP and P=.54 All analyses were performed using SAS
intervention compared with the estab- for diastolic BP). version 8 (SAS Institute Inc, Cary, NC).
lished intervention. The primary out- Primary analyses of BP change are Nominal P values are presented. For a
come was change in systolic BP from based on intention to treat. For indi- given outcome, we only considered the
baseline to 6 months. Hypertension sta- viduals without BP at the 6-month as- pairwise contrasts vs the advice only in-
tus and change in diastolic BP at 6 sessment and for those who had been tervention to be significantly different if
months were secondary outcomes. Blood taking antihypertensive medication, at least one of them achieved a P value
pressure measurements were censored 3-month BP measurements were car- of ⬍.025; in that case, the other con-
if the participant reported taking any an- ried forward; if a 3-month BP measure- trast with advice only and the contrast
tihypertensive medication or other medi- ment was unavailable, values were im- between the established and estab-
cations known to have major BP effects puted using a “hot deck” procedure that lished plus DASH interventions were
(eg, oral steroids). Hypertension was de- drew values from participants in the ad- evaluated at the .05 level.21
fined as a mean BP of 140/90 mm Hg or vice only group.19 We also conducted
higher or use of antihypertensive medi- post-hoc “on treatment” analyses lim- RESULTS
cation. ited to those participants in the estab- A total of 810 participants were en-
Although the intervention pro- lished and established plus DASH rolled in the trial (Figure 1). Baseline
grams lasted 18 months, the protocol- groups who attended at least 15 inter- characteristics were similar in the ran-
specified, primary outcome assess- vention sessions. domized groups (TABLE 1). The mean
ment occurred at 6 months because For all other variables, including hy- (SD) age was 50.0 (8.9) years, 62% were
national guidelines recommend that in- pertension status, we used available data women, and 34% were African Ameri-
dividuals with persistent BP of 140/90 and did not impute values for missing cans. Of the 279 African Americans,
mm Hg or higher after a period of life- data. To analyze continuous indica- 74% were women. The participants
style modification be referred for medi- tors of intervention effects, such as were generally overweight and seden-
cation treatment.7 During the design of change in body weight, we used a simi- tary. Mean (SD) systolic and diastolic
the trial, we anticipated that approxi- lar analytic model. We used the Mantel- BP were 134.9 (9.6) and 84.8 (4.2)
mately 30% of participants would have Haenszel ␹2 test for 2⫻2 tables to com- mm Hg. Among the 38% of partici-
stage 1 hypertension at baseline. 10 pare the proportion of individuals pants with hypertension, mean (SD)
Hence, we expected that a large num- meeting intervention targets at 6 systolic and diastolic BP were 143.9
ber of individuals would need to be re- months.20 Because the focus of these (7.6) and 87.5 (4.3) mm Hg; corre-
ferred for medication treatment at 6 analyses was the proportion actually sponding BP measurements in the non-
months, requiring censoring of their BP meeting target at 6 months in each hypertensive participants were 129.5
data, and that medication treatment group and not necessarily the change (5.8) and 83.2 (3.1) mm Hg. Six months
would occur differentially across the 3 from baseline status, these analyses did after randomization, 94% of partici-
randomized groups. Defining the pri- not condition on initial status. pants had their BP measured at 1 or
mary outcome at 6 months reduced the Hypertension status at 6 months was more visits; 87% attended all 3 visits.
risk of bias and ensured that we would assessed separately for those who were
have a maximum number of BP mea- and were not hypertensive at baseline, Intervention Attendance
surements for analysis. reflecting persistent and incident hy- and Effects
pertension, respectively. We also com- Of the 18 intervention sessions of-
Data Analysis pared the prevalence of hypertension fered during the intial 6 months, 70%
Blood pressure data were analyzed us- in all participants. Pairwise differ- of participants in the established group
ing a linear regression model in which ences in the incidence, persistence, and attended at least 15 sessions; just 8%
change in BP (mean 6-month value − overall prevalence of hypertension be- attended 5 sessions or less. In the es-
mean baseline) was regressed on indi- tween treatment groups were also as- tablished plus DASH group, corre-
cators of the 2 behavioral interven- sessed using the Mantel-Haenszel test.20 sponding data were 78% and 7%. Mean
2086 JAMA, April 23/30, 2003—Vol 289, No. 16 (Reprinted) ©2003 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

(SD) attendance was 14.5 (4.5) and 15.4

Table 1. Baseline Characteristics by Randomized Group*
(4.4) sessions, respectively.
Advice Only Established Established + DASH
Differences in weight, physical fit- Characteristic (n = 273) (n = 268) (n = 269)
ness, and diet among the randomized Age, mean (SD), y 49.5 (8.8) 50.2 (8.6) 50.2 (9.3)
groups were achieved. TABLE 2 dis- Female 172 (63.0) 174 (64.9) 154 (57.2)
plays intervention outcomes, and Race or ethnicity
TABLE 3 lists the number of individu- African American 100 (36.6) 100 (37.3) 79 (29.4)
als who reached the intervention goals. Non-Hispanic white 167 (61.2) 163 (60.8) 181 (67.3)
Weight loss occurred in each group, in- All others 6 (2.2) 5 (1.9) 9 (3.3)
cluding the advice only group. While BMI, mean (SD)† 32.9 (5.6) 33.0 (5.5) 33.3 (6.3)
changes in physical activity did not dif- Weight classification
Nonoverweight (BMI, ⬍25) 15 (5.5) 13 (4.9) 16 (6.0)
fer among the groups, fitness signifi-
Overweight (BMI, 25-29.9) 76 (27.8) 80 (29.9) 82 (30.5)
cantly improved in both behavioral in- Obese (BMI, ⱖ30) 182 (66.7) 175 (65.3) 171 (63.6)
terventions. Alcohol intake was low and Alcohol, mean (SD), drinks/d 0.21 (0.41) 0.24 (0.47) 0.29 (0.52)
did not change in any group. Sedentary (kcal/kg/d ⱕ35) 223 (81.7) 217 (81.0) 224 (83.6)
Mean reductions in urinary sodium Annual household income
excretion occurred in both behavioral ⬍$30 000 31 (11.4) 26 (9.7) 27 (10.0)
interventions, but only the reduction $30 000-$60 000 91 (33.3) 83 (31.0) 82 (30.5)
in the established group differed sig- ⬎$60 000 142 (52.0) 151 (56.3) 148 (55.0)
nificantly from that of advice only Unknown (no answer) 9 (3.3) 8 (3.0) 12 (4.5)
group. However, in both behavioral in- Education
High school or less 21 (7.7) 20 (7.5) 33 (12.3)
tervention groups, the percentage of in-
Some college 175 (64.1) 157 (58.6) 144 (53.5)
dividuals who achieved the trial goal of
Some graduate school 77 (28.2) 91 (34.0) 92 (34.2)
less than 100 mEq/d differed signifi-
Current cigarette smokers 14 (5.1) 18 (6.7) 7 (2.6)
cantly from the advice only group
Dyslipidemia‡ 59 (21.6) 68 (25.4) 64 (23.8)
(Table 3). Also, based on 24-hour di-
Blood pressure, mean (SD), mm Hg
etary recall data, both behavioral inter- Systolic 134.2 (10.1) 135.5 (9.2) 134.9 (9.4)
ventions significantly reduced so- Diastolic 84.8 (4.3) 85.0 (4.1) 84.6 (4.0)
dium intake in comparison with the Hypertensive 104 (38.1) 100 (37.3) 100 (37.2)
advice only group (data not shown). Abbreviations: BMI, body mass index; DASH, Dietary Approaches to Stop Hypertension.
In the established plus DASH group, *Data are presented as No. (%) unless otherwise indicated.
†Body mass index is calculated as weight in kilograms divided by the square of the height in meters.
fruit and vegetable intake increased sig- ‡Total cholesterol ⱖ240 mg/dL (6.21 mmol/L) and/or use of lipid-lowering medication.
nificantly compared with the other 2
groups; parallel changes in urinary po-
tassium excretion occurred. One third of From baseline to 6 months, mean (SD) TABLE 4 displays pairwise differ-
participants in the established plus DASH reductions in systolic BP were 6.6 (9.2) ences in BP. In all participants, nonhy-
group, but only 6% of participants in the mm Hg in the advice only group, pertensive participants, and hyperten-
other 2 groups, consumed the goal of 9 10.5 (10.1) mm Hg in the established sive participants, the established and
or more servings of fruits and veg- group, and 11.1 (9.9) mm Hg in the established plus DASH interventions
etables per day at 6 months. Compared established plus DASH diet group. significantly reduced systolic and di-
with the advice only and established Corresponding diastolic BP reduc- astolic BP in comparison with the ad-
groups, consumption of dairy products tions were 3.8 (6.3), 5.5 (6.7), and 6.4 vice only group. Although BP change
increased significantly in the estab- (6.8) mm Hg, respectively. In hyper- in the established plus DASH group was
lished plus DASH group as did dietary tensive participants, mean (SD) reduc- consistently greater than correspond-
calcium intake and net urinary phos- tions in systolic BP were 7.8 (10.3), 12.5 ing BP change in the established group,
phorus excretion. The percentage of es- (11.5), and 14.2 (10.1) mm Hg, and none of the pairwise differences was sta-
tablished plus DASH participants who mean (SD) reductions in diastolic BP tistically significant.
consumed 2 or more dairy servings was were 3.8 (7.1), 5.8 (7.0), and 7.4 (7.1) The pattern of results was similar in
59%. Saturated and total fat consump- mm Hg, respectively. In nonhyperten- the “on treatment” analyses, which in-
tion significantly decreased in both in- sive participants, mean (SD) reduc- cluded those individuals in the estab-
tervention groups. tions in systolic BP were 5.8 (8.4), 9.4 lished and established plus DASH
(9.1), and 9.2 (9.3) mm Hg, and mean groups (70% and 78% of participants,
Blood Pressure Effects (SD) reductions in diastolic BP were 3.8 respectively) who attended more than
Blood pressure declined progressively (5.8), 5.3 (6.5), and 5.8 (6.6) mm Hg, 15 intervention sessions. For con-
over time in each group (FIGURE 2). respectively. trasts with the advice only group, BP
©2003 American Medical Association. All rights reserved. (Reprinted) JAMA, April 23/30, 2003—Vol 289, No. 16 2087

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

Table 2. Intervention Outcomes at Baseline and at 6 Months by Randomized Group

P Value*
Mean (SD)
Established Established
Established Established vs + DASH vs + DASH vs
Intervention Outcome Advice Only Established + DASH Advice Only Advice Only Established
Weight, kg† n = 242 n = 238 n = 233
Baseline 95.8 (17.0) 96.2 (17.8) 98.8 (19.3)
6 Months 94.7 (17.2) 91.3 (18.2) 93.0 (19.0)
Change −1.1 (3.2) −4.9 (5.5) −5.8 (5.8) ⬍.001 ⬍.001 .07
Physical activity (estimated energy expenditure, kcal/kg/d) n = 239 n = 229 n = 233
Baseline 33.7 (2.6) 33.8 (2.6) 33.8 (3.5)
6 Months 34.0 (2.4) 34.2 (2.2) 34.4 (3.2)
Change 0.3 (2.9) .4 (2.9) 0.6 (2.4) .66 .10 .23
Fitness (heart rate at stage 2 of exercise test, beats/min) n = 235 n = 226 n = 232
Baseline 130.3 (14.7) 130.6 (14.2) 130.1 (14.6)
6 Months 125.0 (15.6) 122.6 (15.7) 121.1 (15.8)
Change −5.3 (9.7) −8.0 (11.1) −9.0 (10.7) .005 ⬍.001 .28
Alcohol intake, drinks/d n = 228 n = 226 n = 229
Baseline 0.2 (0.4) 0.2 (0.5) 0.3 (0.5)
6 Months 0.2 (0.4) 0.2 (0.4) 0.3 (0.5)
Change 0.0 (0.3) 0.0 (0.3) 0.0 (0.4) .53 .87 .42
Urine collections n = 215 n = 212 n = 211
Sodium, mEq/24 h
Baseline 173.3 (66.7) 167.8 (70.0) 178.2 (78.9)
6 Months 152.8 (66.3) 136.2 (64.6) 145.6 (71.6)
Change −20.6 (71.6) −31.6 (74.7) −32.6 (78.1) .01 .12 .36
Potassium, mEq/24 h
Baseline 66.4 (28.7) 66.6 (23.8) 67.9 (26.1)
6 Months 65.1 (27.5) 67.5 (24.9) 87.3 (36.0)
Change −1.3 (28.7) 0.9 (22.3) 19.3 (32.1) .35 ⬍.001 ⬍.001
Urea nitrogen, mg/24 h
Baseline 11 325.6 (4237.8) 11 415.1 (3819.1) 11 933.6 (3987.5)
6 Months 10 651.2 (3805.3) 11 042.5 (3632.8) 12 085.3 (3991.9)
Change −674.4 (4246.6) −372.6 (3457.0) 151.7 (3507.2) .28 ⬍.001 .01
Phosphorus, mg/24 h n = 216 n = 213 n = 212
Baseline 919.4 (433.0) 903.8 (333.5) 976.4 (368.4)
6 Months 827.3 (341.3) 833.3 (331.0) 978.8 (372.7)
Change −92.1 (441.0) −70.4 (333.2) 2.4 (350.4) .75 ⬍.001 ⬍.001
Dietary recalls n = 232 n = 227 n = 230
Fruits and vegetables, servings/d
Baseline 4.4 (2.3) 4.6 (2.4) 4.8 (2.5)
6 Months 4.9 (2.7) 5.1 (2.5) 7.8 (3.2)
Change 0.5 (2.8) 0.5 (2.6) 3.0 (3.6) .79 ⬍.001 ⬍.001
Dairy, servings/d
Baseline 1.6 (1.2) 1.7 (1.3) 1.8 (1.3)
6 Months 1.7 (1.4) 1.5 (1.1) 2.3 (1.2)
Change 0.1 (1.6) −0.2 (1.5) 0.5 (1.6) .02 ⬍.001 ⬍.001
Dietary calcium, mg/24 h‡
Baseline 732.3 (350.2) 728.8 (336.5) 763.0 (365.7)
6 Months 699.2 (402.4) 683.9 (350.7) 940.2 (403.5)
Change −33.1 (418.1) −44.9 (327.6) 177.2 (439.7) .55 ⬍.001 ⬍.001
Total fat, % kcal
Baseline 32.9 (7.1) 33.4 (8.0) 33.3 (7.8)
6 Months 31.9 (7.6) 29.4 (8.4) 23.8 (8.6)
Change −1.0 (7.9) −3.9 (9.8) −9.5 (9.5) ⬍.001 ⬍.001 ⬍.001
Saturated fat, % kcal
Baseline 11.0 (3.3) 10.8 (3.2) 11.0 (3.1)
6 Months 10.6 (3.5) 9.4 (3.5) 7.7 (3.2)
Change −0.4 (3.9) −1.5 (4.0) −3.3 (3.9) ⬍.001 ⬍.001 ⬍.001
Abbreviation: DASH, Dietary Approaches to Stop Hypertension.
*Corresponding to pairwise differences in change.
†Among overweight or obese participants (body mass index ⱖ25).
‡Not including supplemental calcium.

2088 JAMA, April 23/30, 2003—Vol 289, No. 16 (Reprinted) ©2003 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

reductions in the on treatment analy- prevalence of hypertension (12%) oc- individuals who achieved an optimal
ses were approximately 20% to 40% curred in the established plus DASH BP (⬍120 mm Hg systolic and ⬍80
greater than corresponding reduc- group; this prevalence corresponds to mm Hg diastolic; FIGURE 4).
tions in the intention to treat analyses. a 53% risk reduction (ie, 1 – relative
FIGURE 3 displays the percentage of risk) compared with the advice only Other Effects
nonhypertensive participants who be- group. By 6 months, antihypertensive A serious musculoskeletal injury oc-
came hypertensive, the percentage of medication had been started in 19 par- curred in 20 participants in the advice
hypertensive participants who re- ticipants in the advice only group, 2 par- only group, 17 in the established group,
mained hypertensive, and the percent- ticipants in the established group, and and 16 in the established plus DASH
age of all participants who were hyper- 5 participants in the established plus group. One stroke, 1 transient ische-
tensive at 6 months. In each instance, DASH group. The established and es- mic attack, and 1 myocardial infarc-
there was a gradient in hypertensive sta- tablished plus DASH groups signifi- tion occurred in the advice only group.
tus across the 3 groups. The lowest cantly increased the percentage of No cardiovascular event occurred in the

Table 3. Participants Meeting Intervention Goals at Baseline and at 6 Months by Randomized Group
No./Total (%) P Value*

Established Established
Established Established vs + DASH vs + DASH vs
Intervention Goal Advice Only Established + DASH Advice Only Advice Only Established
Weight loss ⬎15 lb (6.8 kg)†
6 Months 15/242 (6.2) 68/238 (28.6) 80/233 (34.3) ⬍.001 ⬍.001 .18
Urinary sodium ⱕ100 mmol/d
Baseline 25/215 (11.6) 38/212 (17.9) 32/211 (15.2)
6 Months 42/215 (19.5) 80/212 (37.7) 59/211 (28.0)
Change 17 (7.9) 42 (19.8) 27 (12.8) ⬍.001 .04 .03
Alcoholic drinks, ⱕ2 drinks/d (men),
ⱕ1 drink/d (women)
Baseline 225/228 (98.7) 218/226 (96.5) 221/229 (96.5)
6 Months 223/228 (97.8) 221/226 (97.8) 224/229 (97.8)
Change 2 (−0.9) 3 (1.3) 3 (1.3) .99 .99 .98
Fruits and vegetables, ⱖ9 servings/d
Baseline 7/232 (3.0) 11/227 (4.9) 16/230 (7.0)
6 Months 15/232 (6.5) 14/227 (6.2) 78/230 (33.9)
Change 8 (3.5) 3 (1.3) 62 (26.9) .90 ⬍.001 ⬍.001
Dairy products, ⱖ2 servings/d
Baseline 72/232 (31.0) 72/227 (31.7) 83/230 (36.1)
6 Months 78/232 (33.6) 64/227 (28.2) 136/230 (59.1)
Change 6 (2.6) 8 (−3.5) 53 (23.0) .21 ⬍.001 ⬍.001
Fat ⱕ30% kcal/d‡
Baseline 83/232 (35.8) 74/227 (32.6) 75/230 (32.6)
6 Months 87/232 (37.5) 123/227 (54.2) 180/230 (78.3)
Change 4 (1.7) 49 (21.6) 105 (45.7) ⬍.001 ⬍.001 ⬍.001
Fat ⱕ25% kcal/d§
Baseline 33/232 (14.2) 33/227 (14.5) 38/230 (16.5)
6 Months 41/232 (17.7) 66/227 (29.1) 135/230 (58.7)
Change 8 (3.5) 33 (14.6) 97 (42.2) .004 ⬍.001 ⬍.001
Saturated fat ⱕ10% kcal/d‡
Baseline 92/232 (39.7) 95/227 (41.9) 94/230 (40.9)
6 Months 103/232 (44.4) 139/227 (61.2) 182/230 (79.1)
Change 11 (4.7) 44 (19.3) 88 (38.2) ⬍.001 ⬍.001 ⬍.001
Saturated fat ⱕ7% kcal/d§
Baseline 27/232 (11.6) 29/227 (12.8) 22/230 (9.6)
6 Months 36/232 (15.5) 60/227 (26.4) 107/230 (46.5)
Change 9 (3.9) 31 (13.6) 85 (36.9) .004 ⬍.001 ⬍.001
Abbreviation: DASH, Dietary Approaches to Stop Hypertension.
*Corresponding to pairwise differences at 6 months from logistic regression analysis.
†Among overweight or obese participants (body mass index ⱖ25).
‡Dietary fat goals for the established group.
§Dietary fat goals for the established + DASH group.

©2003 American Medical Association. All rights reserved. (Reprinted) JAMA, April 23/30, 2003—Vol 289, No. 16 2089

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

established group, and 1 myocardial in- educational attainment and house-

Figure 2. Mean Systolic and Diastolic Blood
Pressure (BP) Over Time by Randomized farction occurred in the established plus hold income were broad, but slightly
Group DASH group. skewed toward persons with higher
education and income. The BP inclu-
Advice Only Established Established
COMMENT sion criteria of PREMIER would in-
+ DASH
The PREMIER trial documented that in- clude approximately 50% of US adults.
dividuals with above-optimal BP, includ- These aspects of the study suggest that
Mean Systolic BP
136 ing stage 1 hypertension, can make mul- trial results should be applicable to a
tiple lifestyle changes that lower BP and large portion of the US population.
132
control hypertension. Both of the Markers of adherence, including sev-
mm Hg

128 PREMIER behavioral interventions ac- eral objective measurements, con-

complished substantial weight loss, re- firmed that participants in the behav-
124 duced sodium intake, and increased ioral interventions made lifestyle
120
physical fitness. Individuals assigned to changes. Each behavioral interven-
the established plus DASH intervention tion led to substantial weight loss as well
also made dietary changes consistent as increased fitness. The reduced heart
86
Mean Diastolic BP with the DASH diet, ie, increased their rate on the treadmill test, an objective
intake of fruits, vegetables, and dairy measure of improved fitness, suggests
82 products. In aggregate, these lifestyle that participants increased their physi-
changes should substantially lower the cal activity, even though self-reported
mm Hg

78
risk of CVD as well as the risk of other physical activity as measured by 7-day
74 chronic diseases, including diabetes, os- physical activity recalls did not change
teoporosis, and perhaps cancer. significantly. Both behavioral interven-
70
Baseline 3 mo 6 mo Trial participants were demographi- tions significantly reduced sodium in-
cally heterogeneous. More than 50% take, although not to the same extent
DASH indicates Dietary Approaches to Stop Hyper- were women, and more than 30% were as behavioral interventions that fo-
tension. African American. The distributions of cused exclusively on this factor.12,13,22

Table 4. Mean Between-Group Differences in Blood Pressure (BP) Change in All Participants, Nonhypertensive Participants, and Hypertensive
Participants
Change in Established Change in Established + DASH Change in Established + DASH
Minus Change in Advice Only* Minus Change in Advice Only Minus Change in Established

Mean (95% CI) P Value Mean (95% CI) P Value Mean (95% CI) P Value
Intention-to-Treat Analyses
Systolic BP
All −3.7 (−5.3 to −2.1) ⬍.001 −4.3 (−5.9 to −2.8) ⬍.001 −0.6 (−2.2 to 0.9) .43
Nonhypertensive −3.1 (−5.1 to −1.1) .003 −3.1 (−5.1 to −1.1) .002 0.0 (−2.0 to 2.0) .97
Hypertensive −4.6 (−7.2 to −2.1) ⬍.001 −6.3 (−8.9 to −3.8) ⬍.001 −1.7 (−4.3 to 0.9) .20
Diastolic BP
All −1.7 (−2.8 to −0.6) .002 −2.6 (−3.7 to −1.5) ⬍.001 −0.9 (−2.0 to 0.2) .11
Nonhypertensive −1.6 (−2.9 to −0.2) .027 −2.0 (−3.4 to −0.6) .005 −0.4 (−1.8 to 0.9) .53
Hypertensive −2.0 (−3.8 to −0.3) .025 −3.6 (−5.4 to −1.9) ⬍.001 −1.6 (−3.4 to 0.2) .08
On Treatment Analyses†
Systolic BP
All −4.9 (−6.6 to −3.3) ⬍.001 −5.7 (−7.2 to −4.1) ⬍.001 −0.7 (−2.5 to 1.0) .41
Nonhypertensive −4.3 (−6.4 to −2.2) ⬍.001 −4.7 (−6.7 to −2.7) ⬍.001 −0.4 (−2.6 to 1.8) .69
Hypertensive −5.9 (−8.5 to −3.2) ⬍.001 −7.1 (−9.7 to −4.5) ⬍.001 −1.3 (−4.1 to 1.6) .38
Diastolic BP
All −2.5 (−3.7 to −1.3) ⬍.001 −3.2 (−4.3 to −2.0) ⬍.001 −0.7 (−1.9 to 0.6) .29
Nonhypertensive −2.2 (−3.7 to −0.7) .003 −2.5 (−3.9 to −1.1) .001 −0.3 (−1.8 to 1.3) .72
Hypertensive −3.0 (−4.9 to −1.1) .002 −4.2 (−6.1 to −2.4) ⬍.001 −1.3 (−3.3 to 0.7) .22
Abbreviation: DASH, Dietary Approaches to Stop Hypertension.
*Change is 6-month BP minus baseline BP.
†Analyses include all advice only participants (n = 273) and those persons in the established group (n = 188, 71% of randomized participants) and established + DASH group
(n = 210, 78% of randomized participants) who completed 15 or more of the 18 possible intervention sessions.

2090 JAMA, April 23/30, 2003—Vol 289, No. 16 (Reprinted) ©2003 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

The established plus DASH interven- tion, the latter is below what was pro-
Figure 3. Percentage of Participants With
tion also increased consumption of vided in the DASH feeding studies, Hypertension at 6 Months by Randomized
fruits, vegetables, and dairy products; namely, 9.6 servings per day. Group Among Nonhypertensive,
analyses of urinary potassium and phos- Another plausible reason is subad- Hypertensive, and All Participants at Baseline
phorus were consistent with data from ditivity of intervention effects. Specifi-
Advice Only Established Established +
the 24-hour dietary recalls. cally, the net BP effect of the DASH diet DASH
Across the 3 groups, gradients in BP in the PREMIER trial likely underesti-
100
and hypertensive status were evident. mates the BP effects of the DASH diet P = .12 P<.001 P<.001
The smallest BP reduction occurred in if it were implemented alone.8,25 It has
80

Hypertensive at 6 mo, %
the advice only group, while the great- been well documented that the com-
est BP reduction occurred in the estab- bined effect of an intervention that 60
P = .01 P = .10

lished plus DASH group. Hyperten- implements 2 or more BP-reducing 52

sion control was most successful in the components is less than the sum of BP 40 P = .01 P = .12
34
established plus DASH group, in which reductions from interventions that 26
23
77% of individuals with stage 1 hyper- implement each component alone. Sub- 20
P = .42 P = .43
17
12
tension at baseline had a systolic BP of additivity can occur from reduced ad- 11
8 6
less than 140 mm Hg and a diastolic BP herence13 because of the effort and com- 0
of less than 90 mm Hg at 6 months. In plexity of making more than 1 lifestyle No. 160 162 156 97 89 97 257 251 253
Nonhypertensive Hypertensive All
the established group, the correspond- change. For example, participants in the Hypertension Status at Baseline
ing figure was 66%. These rates com- established plus DASH intervention
pare favorably with survey data2 and were advised to increase their intake of DASH indicates Dietary Approaches to Stop Hyper-
trial data,23 in which drug therapy con- dairy products while reducing total ca- tension.

trols BP in approximately half of hy- loric intake. Still, even in the setting of
pertensive individuals. Hence, these be- high adherence, such as feeding stud- Figure 4. Percentage of Participants With
havioral interventions should be viable ies, subadditivity occurs.24 Optimal Blood Pressure at 6 Months by
treatment options, at least among those The advice only comparison group Randomized Group Among Nonhypertensive
hypertensive individuals who are mo- in the PREMIER trial likely accom- Participants at Baseline, Hypertensive
Participants at Baseline, and All Participants
tivated to make lifestyle changes. plished some lifestyle modifications.
The established and established plus Such behavior changes might have re- Advice Only Established Established +
DASH interventions also reduced BP in sulted from recruitment of motivated DASH
nonhypertensive individuals with participants; secular trends (eg, grow- 100
above-optimal BP. Specifically, an op- ing awareness of the obesity epi- P<.001 P = .009 P<.001
Optimal Blood Pressure at 6 mo, %

timal BP level (⬍120 mm Hg systolic demic); the 30-minute intervention ses- 80

and ⬍80 mm Hg diastolic) was sion postrandomization; and the
achieved in 40% and 48% of partici- multiple, regular data collection visits 60 P = .04 P = .12
pants assigned to these 2 interven- and contacts (4 visits and 2 dietary re- 48
P = .005 P = .24
tions, respectively. Overall, this pat- calls at baseline; 1 visit at 3 months and 40
40 35
30
tern of findings suggests that even in 3 visits at 6 months, along with tele- 29
P = .02 P = .83
the context of other effective lifestyle phone calls and reminders). Although 20
19
12 13
modifications, adoption of the DASH before-after changes must be inter- 3
diet further improves BP control. preted cautiously, the weight loss of 1.1 0
No. 160 162 156 97 89 97 257 251 253
Still, the BP effects attributed to the kg at 6 months in this group exceeded Nonhypertensive Hypertensive All
DASH diet, specifically the BP differ- what has occurred in other trials, which Hypertension Status at Baseline
ences between the established plus have often reported weight gain in the
DASH and the established interven- control group.13 Also, the 20-mEq re- DASH indicates Dietary Approaches to Stop
Hypertension.
tions, were less than previously found duction in sodium excretion may re-
in the DASH feeding studies,8,24 and flect efforts to lower salt intake. Such
none of the contrasts in hypertension modest changes in behavior might have reduction of 3.8 mm Hg) greatly ex-
status was statistically significant. One reduced BP in this group, thereby at- ceed what was observed in the control
potential reason is that participants re- tenuating the pairwise differences in BP groups of other studies.8,12,13,22
ceived an inadequate “dose” of the between the advice only group and the In the PREMIER trial, the primary
DASH diet. For instance, even though 2 behavioral interventions. In fact, the outcome variables were collected at 6
mean fruit and vegetable intake in- within-group BP reductions observed months postrandomization. The inclu-
creased from 4.8 to 7.8 servings per day in the advice only group (systolic BP re- sion of hypertensive individuals in the
in the established plus DASH interven- duction of 6.6 mm Hg and diastolic BP PREMIER trial precluded use of BP
©2003 American Medical Association. All rights reserved. (Reprinted) JAMA, April 23/30, 2003—Vol 289, No. 16 2091

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

measured at a later follow-up visit as tensive individuals who are not receiv- search Center, Baton Rouge, La: Philip Brantley, Al-
lison Worthen, Betty Kennedy, Emily Griffin, Erma Levy,
the primary outcome because na- ing medication therapy. Although we Terri Keller, Shantell Jones, and Katherine Lastor; Johns
tional guidelines recommend initia- did not study individuals receiving drug Hopkins Medical Center, Baltimore, Md: Barbara Bai-
ley, MS, RD, Jeanne Charleston, RN, MSN, Sharrone
tion of drug therapy for individuals with therapy, available data indicate that Cypress, Arlene Dalcin, MS, RD, Maura Deeley, Chara-
Class B hypertension who remain hy- nonpharmacological interventions also lett Diggs, RN, Thomas P. Erlinger, MD, MPH, Ann
pertensive after a 6-month period of reduce BP in these individuals.35,36 Ul- Fouts, RN, Angela Hall, Charles Harris, Tara Harri-
son, Megan Jehn, Shirley Kritt, Estelle Levitas, Phyllis
nondrug therapy.7 Still, evidence from timately, population-wide adoption of McCarron, MS, RD, Edgar R. Miller III, MD, PhD, Pau-
clinical trials suggests that as long as ad- healthy lifestyles as promoted in the line Patrick, LD, Joy Peterson, Charles Powell, Thomas
Shields, LeeLana Thomas, MS, RD, Letitia Thomas,
herence is sustained, BP effects per- PREMIER interventions should sub- Bobbie Weiss, and Deborah Young, PhD; National
sist.26,27 Also, in longitudinal observa- stantially reduce the societal burden of Heart, Lung, and Blood Institute, Bethesda, Md: Jef-
frey A. Cutler, MD, Michael Proschan, PhD, and Denise
tional studies, healthy dietary patterns CVD and other chronic diseases. Simons-Morton, MD, PhD; National Health and Nu-
indicative of long-term habits are as- trition Examination Survey and Global Micronutri-
Authors/Writing Group of the PREMIER Collabora- ent Laboratory, Centers for Disease Control and Pre-
sociated with reduced CVD28,29 and tive Research Group: Lawrence J. Appel, MD (chair), vention, Atlanta, Ga: Christine Pfeiffer, PhD; Division
mortality.30 Departments of Medicine, Epidemiology, and Inter- of Laboratory Sciences, Centers for Disease Control
national Health (Human Nutrition), Johns Hopkins and Prevention: Elaine W. Gunter, MT; Core Labo-
In addition to hypertension, high- Medical Institutions, Baltimore, Md; Catherine M. ratory for Clinical Studies, Washington University, St
normal BP is also associated with ex- Champagne, PhD, and David W. Harsha, PhD, Pen- Louis, Mo: Thomas G. Cole, PhD; Pennsylvania State
cess CVD risk.4 Of the general popu- nington Biomedical Research Center, Baton Rouge, La; University Diet Assessment Center, University Park:
Lawton S. Cooper, MD, and Eva Obarzanek, PhD, Na- Helen S. Wright, PhD, and Diane C. Mitchell; Center
lation, about 34% have BP in the tional Heart, Lung, and Blood Institute, Bethesda, Md; for Clinical Epidemiology and Biostatistics, Univer-
nonhypertensive yet above optimal BP Patricia J. Elmer, PhD, Victor J. Stevens, PhD, and Wil- sity of Pennsylvania School of Medicine, Philadel-
liam M. Vollmer, PhD, Kaiser Permanente Center for phia: Shiriki Kumanyika, PhD, MPH; Howard Univer-
range and another 14% have stage 1 hy- Health Research, Portland, Ore; Pao-Hwa Lin, PhD, sity, Washington, DC: Jerome Williams, PhD; Stanford
pertension.2 Furthermore, the major- and Laura P. Svetkey, MD, Duke Hypertension Cen- Center for Research in Disease Prevention, Stanford
ter and Sarah W. Stedman Center for Nutritional Stud- University School of Medicine, Palo Alto, Calif:
ity of BP-related events occur in the ies, Duke University Medical Center, Durham, NC; and Leslie Pruitt, PhD, and Abby King, PhD.
range of BP studied in the PREMIER Deborah R. Young, PhD, Department of Kinesiol- Data and Safety Monitoring Board: Jerome D. Co-
ogy, University of Maryland, College Park.
trial.5 In this setting, it is reasonable to Author Contributions: Study concept and design:
hen (chair), Nancy R. Cook, ScD, Patricia Dubbert, PhD,
Keith C. Ferdinand, MD, Jim Raczynski, PhD, and Linda
speculate that widespread implemen- Appel, Elmer, Harsha, Obarzanek, Stevens, Svetkey, Van Horn, PhD.
tation of the PREMIER behavioral in- Vollmer, Champagne, Lin, Young. Funding/Support: This work was supported by Na-
Acquisition of data: Appel, Elmer, Harsha, Svetkey, tional Institutes of Health grants UO1 HL60570, UO1
terventions, particularly the estab- Champagne, Lin, Young. HL60571, UO1 HL60573, UO1 HL60574, UO1
lished plus DASH intervention, should Analysis and interpretation of data: Appel, Cooper, HL62828, and MO1 RR00052.
Elmer, Harsha, Obarzanek, Stevens, Svetkey, Vollmer. Acknowledgment: We are especially grateful to our
decrease CVD risk through reduced BP Drafting of the manuscript: Appel, Elmer, Harsha, participants for their sustained commitment to the trial.
in nonhypertensive individuals and in- Stevens, Svetkey, Vollmer. We also thank members of the trial’s data and safety
Critical revision of the manuscript for important in- monitoring board.
creased BP control in hypertensive in- tellectual content: Appel, Cooper, Elmer, Harsha,
dividuals. Obarzanek, Stevens, Svetkey, Vollmer, Champagne,
Yet most health care insurers do not Lin, Young.
REFERENCES
Statistical expertise: Vollmer.
cover behavioral interventions for the Obtained funding: Appel, Elmer, Harsha, Stevens, 1. Muntner P, He J, Roccella EJ, Whelton PW. The
prevention and treatment of hyperten- Svetkey, Vollmer, Young. impact of JNC-VI guidelines on treatment recommen-
Administrative, technical, or material support: Appel, dations in the US population. Hypertension. 2002;
sion.31 Given the substantial health ben- Cooper, Elmer, Harsha, Obarzanek, Stevens, Cham- 39:897-902.
efits of these programs in improving BP pagne, Lin. 2. Burt VL, Whelton P, Roccella EJ, et al. Prevalence
Study supervision: Appel, Elmer, Harsha, Stevens, of hypertension in the US adult population: results from
control, preventing diabetes,32,33 and Svetkey, Vollmer, Champagne, Young. the Third National Health and Nutrition Examination
controlling dyslipidemia,34 it is time to PREMIER Collaborative Research Group, Participat- Survey, 1988-1991. Hypertension. 1995;25:305-
ing Sites: Coordinating Center, Center for Health Re- 313.
consider how such programs might be search, Portland, Ore: Mikel Aickin, PhD, Jack Hollis, 3. Vasan RS, Beiser A, Seshadri S, Larson MG, Kan-
implemented, particularly for those pa- PhD, Njeri Karanja, PhD, Fran Heinith, BSN, Kristy Funk, nel WB, D’Agostino RB, Levy D. Residual life-time risk
tients at elevated CVD risk. The costs MS, RD, Michael Allison, BS, Chuhe Chen, PhD, Clifton for developing hypertension in middle-aged women
Hindmarsh, MS, Terry Kimes, MS, Wan-Ru Li, MBA, and men: the Framingham Heart Study. JAMA. 2002;
of such programs should be balanced Gayle Meltesen, MS, Carrie Meeks, Nadia Redmond, 287:1003-1010.
against the benefits of preventing hy- MSPH, and Rina Smith, BA, PgDipLGA; Clinical Cen- 4. Vasan RS, Larson MG, Leip EP, et al. Impact of high-
ter, Center for Health Research, Portland, Ore: Adri- normal blood pressure on the risk of cardiovascular
pertension, diabetes, heart disease, and anne C. Feldstein, MD, MS, Daniel S. Laferriere, RN, disease. N Engl J Med. 2001;345:1291-1297.
other conditions, thereby preventing the MSN, Shirley R. Craddick, RD, MHA, Dana R. Lar- 5. Stamler J, Stamler R, Neaton JD. Blood pressure,
son, RD, MS, Diane J. Cook, RD, MPH, Carol L. Young, systolic and diastolic, and cardiovascular risks: US popu-
need for medical treatments that may Susan D. Arnold, RN, BSN, Donna L. Clark, Stanley lation data. Arch Intern Med. 1993;153:598-615.
be costly. B. Postlethwaite, Titza Suvalcu-Constantin, Carol M. 6. Whelton PK, He J, Appel LJ et al. Primary preven-
In summary, our trial results dem- Maul, Donna M. Gleason, Cheryl A. Johnson, EdM, tion of hypertension: clinical and public health advi-
Pamela G. McNeal, and Debra D. Burch; Duke Uni- sory from the National High Blood Pressure Educa-
onstrate the feasibility of comprehen- versity Medical Center, Durham, NC: Colleen tion Program. JAMA. 2002;288:1882-1888.
sive behavioral interventions and their McBride, PhD, Jamy Ard, MD, Kathleen Aicher, 7. Joint National Committee on Prevention, Detec-
Blondeaner Brown, Denise Ernst, Jeanne Gresko, tion, Evaluation, and Treatment of High Blood Pres-
beneficial effects on BP and hyperten- Madhuri Kesari, Femke Lamers, LaTonya Nealon, Tori sure. The Sixth Report of the Joint National Commit-
sion control. Benefits extend to both Phelps, LaVerne Pruden, LaChanda Reams, Patrice tee on the Prevention, Detection, Evaluation, and
Reams, Benjamin Reese, PsyD, Fran Rukenbrod, So- Treatment of High Blood Pressure ( JNC VI). Arch In-
nonhypertensive individuals at risk for nia Steele, Natalie Thorpe, Olaunda Williams, and tern Med. 1997;157:2413-2446.
developing hypertension and hyper- Chenghua (Cherry) Yang; Pennington Biomedical Re- 8. Appel LJ, Moore TJ, Obarzanek E, et al. A clinical

2092 JAMA, April 23/30, 2003—Vol 289, No. 16 (Reprinted) ©2003 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015

 EFFECTS OF LIFESTYLE MODIFICATION ON BLOOD PRESSURE CONTROL

trial of the effects of dietary patterns on blood pres- habitual physical activity methodology by a seven- of Hypertension Prevention, phase I. Hypertension.
sure. N Engl J Med. 1997;336:1117-1124. day recall in a community survey and controlled ex- 1993;22:502-512.
9. Karanja NM, Obarzanek E, Lin PH, et al. Descrip- periments. Am J Epidemiol. 1985;122:794-804. 27. Stevens VJ, Obarzanek E, Cook NR, et al. Long-
tive characteristics of the dietary patterns used in the 18. Writing Group for the Activity Counseling Trial (ACT) term weight loss and changes in blood pressure: re-
Dietary Approaches to Stop Hypertension trial. J Am Research Group. Effects of physical activity counseling sults of the Trials of Hypertension Prevention, phase
Diet Assoc. 1999;99(suppl):S19-S27. in primary care: the Activity Counseling Trial: a ran- II. Ann Intern Med. 2001;134:1-11.
10. Svetkey LP, Harsha DW, Vollmer WM, et al. domized controlled trial. JAMA. 2001;286:677-687. 28. McCullough ML, Feskanich D, Rimm EB, et al. Ad-
PREMIER: a clinical trial of comprehensive lifestyle modi- 19. Schafer JL. Multiple imputation: a primer. Stat herence to the Dietary Guidelines for Americans and
fication for blood pressure control: rationale, design and Methods Med Res. 1999;8:3-15. risk of major chronic disease in men. Am J Clin Nutr.
baseline characteristics. Ann Epidemiol. In press. Avail- 20. Agresti A. An Introduction to Categorical Data 2000;72:1223-1231.
able at: http://www.sciencedirect.com/science/journal Analysis. New York, NY: John Wiley & Sons Inc; 1996. 29. Fung TT, Willett WC, Stampfer MJ, Manson JE, Hu
/10472797. Accessibility verified March 31, 2003. 21. Aickin M, Gensler H. Adjusting for multiple test- FB. Dietary patterns and risk of coronary heart disease
11. Joint National Committee on Detection, Evalua- ing when reporting research results: Bonferroni vs Holm in women. Arch Intern Med. 2001;161:1857-1862.
tion, and Treatment of High Blood Pressure. The Fifth methods. Am J Public Health. 1996;86:726-728. 30. Kant AK, Schatzkin A, Graubard BI, Schairer C. A
Report of the Joint National Committee on the De- 22. Whelton PK, Appel LJ, Espeland MA, et al, for the prospective study of diet quality and mortality in
tection, Evaluation, and Treatment of High Blood Pres- TONE Collaborative Research Group. Efficacy of so- women. JAMA. 2000;283:2109-2115.
sure ( JNC V). Arch Intern Med. 1993;153:154-183. dium reduction and weight loss in the treatment of 31. The Role of Nutrition in Maintaining Health in
12. The Trials of Hypertension Prevention Collabo- hypertension in older persons: main results of the ran- the Nation’s Elderly: Evaluating Coverage of Nutri-
rative Research Group. The effects of nonpharmaco- domized, controlled trial of nonpharmacologic inter- tion Services for the Medicare Population. Washing-
logic interventions on blood pressure of persons with ventions in the elderly (TONE). JAMA. 1998;279:839- ton, DC: Institute of Medicine, National Academy of
high normal levels. JAMA. 1992;267:1213-1220. 846. Sciences; 2000.
13. The Trials of Hypertension Prevention Collabo- 23. Cushman WC, Ford CE, Cutler JA, et al. Success 32. Tuomilehto J, Lindstrom J, Eriksson JG, et al. Pre-
rative Research Group. Effects of weight loss and so- and predictors of blood pressure control in diverse vention of type 2 diabetes mellitus by changes in life-
dium reduction intervention on blood pressure and hy- North American settings: the Antihypertensive and style among subjects with impaired glucose toler-
pertension incidence in overweight people with high- Lipid-Lowering Treatment to Prevent Heart Attack Trial ance. N Engl J Med. 2001;344:1343-1350.
normal blood pressure: the trials of hypertension (ALLHAT). J Clin Hypertens. 2002;4:393-405. 33. Knowler WC, Barrett-Connor E, Fowler SE, et al.
prevention, phase II. Arch Intern Med. 1997;157:657- 24. Sacks FM, Svetkey LP, Vollmer WM, et al, for Reduction in the incidence of type 2 diabetes with life-
667. the DASH-Sodium Collaborative Research Group. A style intervention or metformin. N Engl J Med. 2002;
14. Rose G, McCartney P, Reid DD. Self- clinical trial of the effects on blood pressure of 346:393-403.
administration of a questionnaire on chest pain and reduced dietary sodium and the DASH dietary pat- 34. Dattilo AM, Kris-Etherton PM. Effects of weight
intermittent claudication. Br J Prev Soc Med. 1977; tern (the DASH-Sodium Trial). N Engl J Med. 2001; reduction on blood lipids and lipoproteins: a meta-
31:42-48. 344:3-10. analysis. Am J Clin Nutr. 1992;56:320-328.
15. Fox TA, Heimendinger J, Block G. Telephone sur- 25. John JH, Ziebland S, Yudkin P, Roe LS, Neil HA. 35. Langford HG, Davis BR, Blaufox D, et al, for the
veys as a method for obtaining dietary information: a Effects of fruit and vegetable consumption on plasma TAIM Research Group. Effect of drug and diet treat-
review. J Am Diet Assoc. 1992;92:729-732. anti-oxidant concentrations and blood pressure: a ran- ment of mild hypertension on diastolic blood pres-
16. American College of Sports Medicine. ACM’s domised controlled trial. Lancet. 2002;359:1969- sure. Hypertension. 1991;17:210-217.
Guidelines for Exercise Testing and Prescription. 5th 1974. 36. Miller ER, Erlinger EP, Young DR, et al. Results
ed. Media, Pa: Williams & Wilkins; 1995. 26. Kumanyika SK, Hebert PR, Cutler JA, et al. Fea- of the Diet, Exercise and Weight-loss Intervention Trial
17. Blair SN, Haskell WL, Po H, et al. Assessment of sibility and efficacy of sodium reduction in the Trials (DEW-IT). Hypertension. 2002;40:612-618.

If I had a device, it would be the True, the True only,

leaving the Beautiful and the Good to settle matters
afterwards as best they could.
—Charles Augustin Sainte-Beuve (1804-1869)

©2003 American Medical Association. All rights reserved. (Reprinted) JAMA, April 23/30, 2003—Vol 289, No. 16 2093

Downloaded From: http://jama.jamanetwork.com/ by a University of Georgia User on 05/29/2015