Effect of the Mediterranean diet on blood

pressure in the PREDIMED trial: results from a

randomized controlled trial
Toledo et al.

Toledo et al. BMC Medicine 2013, 11:207

[Link]
Toledo et al. BMC Medicine 2013, 11:207
[Link]

RESEARCH ARTICLE Open Access

Effect of the Mediterranean diet on blood

pressure in the PREDIMED trial: results from a
randomized controlled trial
Estefania Toledo1,2,3, Frank B Hu2,4, Ramon Estruch3,5, Pilar Buil-Cosiales1,3, Dolores Corella3,6, Jordi Salas-Salvadó3,7,
M Isabel Covas3,8, Fernando Arós3,9, Enrique Gómez-Gracia3,10, Miquel Fiol3,11, Jose Lapetra3,12, Luis Serra-Majem3,13,
Xavier Pinto3,14, Rosa M Lamuela-Raventós3,15, Guillermo Saez3,16,17, Mònica Bulló3,7, Valentina Ruiz-Gutiérrez3,18,
Emilio Ros3,19, José V Sorli3,6,20 and Miguel Angel Martinez-Gonzalez1,3*

Abstract
Background: Hypertension can be prevented by adopting healthy dietary patterns. Our aim was to assess the
4-year effect on blood pressure (BP) control of a randomized feeding trial promoting the traditional Mediterranean
dietary pattern.
Methods: The PREDIMED primary prevention trial is a randomized, single-blinded, controlled trial conducted in
Spanish primary healthcare centers. We recruited 7,447 men (aged 55 to 80 years) and women (aged 60 to 80
years) who had high risk for cardiovascular disease. Participants were assigned to a control group or to one of two
Mediterranean diets. The control group received education on following a low-fat diet, while the groups on
Mediterranean diets received nutritional education and also free foods; either extra virgin olive oil, or nuts. Trained
personnel measured participants’ BP at baseline and once yearly during a 4-year follow-up. We used generalized
estimating equations to assess the differences between groups during the follow-up.
Results: The percentage of participants with controlled BP increased in all three intervention groups (P-value for
within-group changes: P<0.001). Participants allocated to either of the two Mediterranean diet groups had
significantly lower diastolic BP than the participants in the control group (−1.53 mmHg (95% confidence interval
(CI) −2.01 to −1.04) for the Mediterranean diet supplemented with extra virgin olive oil, and −0.65 mmHg (95%
CI -1.15 to −0.15) mmHg for the Mediterranean diet supplemented with nuts). No between-group differences in
changes of systolic BP were seen.
Conclusions: Both the traditional Mediterranean diet and a low-fat diet exerted beneficial effects on BP and could
be part of advice to patients for controlling BP. However, we found lower values of diastolic BP in the two groups
promoting the Mediterranean diet with extra virgin olive oil or with nuts than in the control group.
Trial registration: Current Controlled Trials ISRCTN35739639
Keywords: Mediterranean diet, Low-fat diet, Systolic blood pressure, Diastolic blood pressure, Controlled trial,
PREDIMED trial, Monounsaturated fat, Dietary patterns, Olive oil, Nuts

* Correspondence: mamartinez@[Link]
1
Department of Preventive Medicine and Public Health, University of Navarra,
Pamplona, Navarra, Spain
3
CIBER Fisiopatología de la Obesidad y Nutrición (CIBER OBN), Instituto de
Salud Carlos III (ISCIII), Spanish Government, Madrid, Spain
Full list of author information is available at the end of the article

© 2013 Toledo et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License ([Link] which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
Toledo et al. BMC Medicine 2013, 11:207 Page 2 of 9
[Link]

Background Methods
In 2003, the Joint National Committee on Prevention, Design overview
Detection, Evaluation, and Treatment of High Blood The PREDIMED (Prevención con Dieta Mediterránea)
Pressure estimated that hypertension affects approxi- study is a multicenter, randomized, parallel-group trial
mately 1 billion people worldwide [1]. This condition is conducted in Spain ([Link]). A detailed
a major risk factor for stroke, ischemic heart disease, description of the methods and objectives of the
and other chronic cardiovascular diseases (CVDs) [2]. In PREDIMED trial can be found elsewhere [10]. Briefly,
fact, the relationship between blood pressure (BP) and this trial was designed to assess the effects of the trad-
risk of CVD events is continuous, consistent, and inde- itional MD on the primary prevention of CVD. The
pendent of other risk factors [1]. Because of its high main outcome was an aggregate of non-fatal myocardial
prevalence and its related conditions, hypertension is the infarction, non-fatal stroke, or cardiovascular death. The
leading individual risk factor for mortality, and is re- trial was stopped because of early benefit by December
sponsible for 7.6 million deaths per year [3]. Therefore, 1, 2010 after a median follow-up time of 4.8 years [4].
from a public heath perspective, approaches to tackle The current work ascertains the long-term effect of
this condition are needed urgently. the dietary interventions on changes in BP during 4
Adopting a healthy lifestyle is a cornerstone of hyper- years of follow-up.
tension prevention and treatment, and a healthy diet rep-
resents a major lifestyle modification for BP control [1,4]. Ethics approval
High-quality overall dietary patterns, such as the Dietary The protocol was written in accordance with the princi-
Approaches to Stop Hypertension (DASH) diet, can be of ples of the Declaration of Helsinki, was approved by
utmost importance in the prevention and treatment of the Institutional Review Boards at all study sites (for
hypertension [5]. Of these high-quality dietary patterns, more detailed information, please check Additional file 1),
one in particular has received much recent attention be- and was registered at [Link]
cause of the growing evidence for its role in cardiovascular ISRCTN35739639. Written informed consent was pro-
protection, namely, the traditional Mediterranean diet vided by all study participants.
(MD) [6]. The MD is a traditional food pattern present in
the olive oil-producing areas of the Mediterranean basin. Setting and participants
Like the DASH diet, the traditional MD is rich in fruits Eligible participants were men (aged 55 to 80 years) and
and vegetables, but it also includes an abundance of le- women (aged 60 to 80 years) who were free of CVD at
gumes, a moderate intake of fish, dairy products, and study inception but at high cardiovascular risk because
wine, small portions of meat and poultry, and little con- of the presence of either type 2 diabetes (T2D) or at
sumption of candies (sweets) [7]. A key characteristic of least three major CVD risk factors, including current
this diet is the low amount of animal and trans fatty acids. smoking, hypertension, high levels of low-density lipo-
Extra virgin olive oil (EVOO), the primary source of fat in protein cholesterol, low levels of high-density lipoprotein
the MD, along with plant foods and nuts, makes this diet cholesterol, overweight/obesity, or family history of pre-
ideal for health because these fresh foods undergo min- mature coronary heart disease (CHD). Further details of
imal processing, so they are rich in fiber, antioxidant poly- the inclusion and exclusion criteria can be found in our
phenols, and essential micronutrients and macronutrients. previously published report [10]. Study candidates were
Recently, the PREDIMED primary prevention trial showed selected from databases of primary care facilities. Of
that a dietary intervention designed to foster adherence to those who met entry requirements, 89% agreed to
the traditional MD significantly reduced the risk of CVD participate.
clinical end-points [8]. The reported reduction in CVD At baseline, participants completed a general med-
was most evident for stroke, a condition known to be ical questionnaire, a 137-item previously validated
highly dependent on BP. Therefore, one mechanism by food-frequency questionnaire [11,12], the Minnesota
which the traditional MD may exert its beneficial effect is Leisure-Time Physical Activity Questionnaire [13,14],
in the control of BP. In fact, a recent meta-analysis and a 14-item screening questionnaire of adherence to
reporting results from clinical studies supported a protect- the traditional MD [15].
ive role for the MD on both systolic and diastolic BP.
However, only two studies had a follow-up time beyond Randomization and interventions
2 years, and the largest of the two had a sample size of During the period October 2003 to June 2009, 7,447 par-
605 subjects [9]. ticipants were enrolled in the study, and randomly allo-
The aim of this study was to assess the long-term ef- cated in a [Link] ratio by means of a computer-generated
fect on BP of a dietary intervention to improve adher- random-number sequence to one of the three intervention
ence to the traditional MD. groups: MD supplemented with EVOO (MD+EVOO),
Toledo et al. BMC Medicine 2013, 11:207 Page 3 of 9
[Link]

MD supplemented with mixed nuts (MD+nuts: wal- or >260 mmHg, diastolic BP <40 mmHg or >135 mmHg,
nuts, almonds, and hazelnuts), or the control diet (low-fat systolic BP changes >40 mmHg at 1 year, and diastolic BP
diet). The coordinating center constructed a computer- changes >25 mmHg at 1 year.
generated randomization table, with allocation conceal- For the present analysis, we have included information
ment by opaque, sequentially numbered, sealed envelopes. for 4 years of follow-up (median follow-up time 3.8
At baseline and quarterly thereafter, dieticians ran in- years) because the recruitment ended in 2009, and BP
dividual and group sessions, with no more than 20 par- measurements were not available for a substantial num-
ticipants, separately for each of the three groups. Group ber of participants beyond 4 years.
sessions were specific for each intervention group so
that participants were assessed only for adherence to the
Statistical analysis
diet to which they had been allocated. In the appropriate
Baseline characteristics are presented according to the
individual sessions, a 14-item dietary screening question-
intervention group, as mean (SD) for quantitative traits
naire was used to check for adherence to either of the
and n (%) for categorical variables.
MDs, and a 9-item dietary screening questionnaire
All analyses were performed in accordance with an
was used to check for adherence to the control low-fat
intention-to-treat approach. First, we assessed differ-
diet [15]. The questionnaire responses were used to
ences in changes in BP between groups during follow-up
personalize the intervention for each participant, and
[18]. For participants with missing values of BP in the
to negotiate dietary changes to upgrade adherence to
year 4 visit, we used the most recent available BP infor-
either the MD or the low-fat diet. Participants in the
mation. Second, we used generalized estimating equa-
two intervention groups were given supplementary
tions to calculate mean differences in systolic and
foods at no cost: either EVOO (1 liter/week for the
diastolic BP changes between the groups allocated to the
participant and their families) or mixed nuts (30 g/day:
MD+EVOO or MD+nuts versus the control group in
15 g walnuts, 7.5 g hazelnuts, and 7.5 g almonds)
crude analyses, and after adjustment for center, age, sex
according to their randomization group. Supplementation
and baseline T2D, and, additionally for baseline number
of these foods was intended to ensure high consumption
of anti-hypertensive drugs and baseline systolic or dia-
of these key elements of the traditional MD, and to pro-
stolic BP. We assumed an unstructured correlation
mote a better overall adherence to the target overall diet-
matrix and calculated robust variance estimates. Third,
ary pattern.
we also used generalized estimating equations to calculate
The control group received usual care and dietary
the percentage of participants with controlled BP levels
counseling (including group sessions) aimed to increase
(systolic BP <140 mmHg and diastolic BP <90 mmHg)
their adherence to a lower-fat diet. The control group
during follow-up. Fourth, we used generalized estimating
received non-food items as incentives throughout the
equations to ascertain the number of anti-hypertensive
study.
drugs that were prescribed during follow-up. Analyses
Energy restriction was not specifically advised nor was
were performed using STATA software (version 11.0;
physical activity promoted in any of the three groups,
StataCorp, College Station, TX, USA).
and the interventions did not target sodium intake or
sleep habits. Drugs were prescribed during regular med-
ical care of the participants and were not influenced by Results
the intervention. Of the 7,447 participants recruited to the PREDIMED
trial, 289 were excluded either because there was no infor-
Outcomes and follow-up mation on their baseline BP or they had extreme BP
At baseline and once yearly thereafter, trained personnel values. Thus, our effective sample size was 7,158 (Figure 1).
measured participants’ BP in each arm with a validated On average, participants had 3.8 visits with available BP
semiautomatic oscillometer (Omron HEM-705CP, Hoofd- information during follow-up. Specifically, 2,345 partici-
dorp, the Netherlands) at three time points, separated by pants in the MD+EVOO group, 2178 participants in the
2 minutes, while the participant was in a seated position MD+nuts group, and 2,064 participants in the control
after 5 minutes of rest. Arm circumference determined group had BP measurements during follow-up.
the cuff size, and BP was measured in the forearm at heart In the PREDIMED trial, slightly more women than
level. The average of the second and third measurement men were recruited. Mean age was 67 years. Participants
was recorded in the data collection form. were at high cardiovascular risk as per protocol, and
The mean of the systolic and diastolic BP measure- they had an average baseline MD score of 8.6 in the 14-
ments was also calculated [16,17]. The following point score of adherence to the MD. All three groups
values were considered extreme and were not taken were well balanced in their baseline characteristics, in-
into account for the analyses: systolic BP <70 mmHg cluding their dietary and non-dietary traits (Table 1).
Toledo et al. BMC Medicine 2013, 11:207 Page 4 of 9
[Link]

Figure 1 Flowchart of participants in the PREDIMED trial.

Significant reductions in both systolic and diastolic BP adjusting for center, age, sex, and baseline T2D. No sig-
were apparent for the three randomized groups during nificant between-group differences were seen.
follow-up (P<0.001 for within-group changes during The average number of anti-hypertensive drugs pre-
follow-up time, adjusted for center, sex, age, and baseline scribed for the PREDIMED participants significantly in-
T2D) (Figure 2). There were no significant between- creased during follow-up in the three intervention
group for systolic (P = 0.51 for MD+EVOO versus con- groups after adjustment for center, sex, age and baseline
trol and P>0.99 for MD+nuts versus control) or diastolic T2D. We found no significant between-group differ-
(P = 0.39 for MD+EVOO versus control and P = 0.09 ences. At the end of follow-up, the average number of
for MD+nuts versus control) BP. These latter compari- BP-lowering drugs in the PREDIMED participants was
sons were adjusted for center, sex, age, and baseline 1.41 (95% CI 1.36 to 1.46) in the MD+EVOO group,
T2D. 1.39 (95% CI 1.33 to 1.44) in the MD+nuts group, and
We found a greater reduction in average systolic BP in 1.39 (95% CI 1.33 to 1.45) in the control group (P>0.99).
the MD+nuts group than in the control group. However,
between-group differences in systolic BP versus control Discussion
with up to 4 years of follow-up were apparent only in In this large randomized controlled trial, participants in
crude analyses, and they became non-significant after all three groups showed improvement in their BP levels,
multivariate adjustment. No differences in systolic BP and, consequently, the percentage of participants with
were found between the MD+EVOO and the control controlled BP also increased in all three groups. How-
group (Table 2). However, compared with the control ever, a greater reduction in diastolic BP was obtained
group, greater reductions in diastolic BP were seen for with the MD interventions than with control interven-
both MDs. These differences remained significant in tion (advice to follow a low-fat diet). This could partly
multivariate-adjusted analyses, with adjusted differences explain the recently reported benefit of the MD inter-
of −1.53 mmHg (95% CI −2.01 to −1.04) for MD+EVOO vention on clinical disease end-points [8], especially the
versus control and −0.65 (95% CI −1.15 to −0.15) mmHg reduction in incidence of stroke, a cardiovascular event
for MD+nuts versus control. clearly related to high BP. However, other mechanisms
Improvements in BP control were apparent for all apart from BP also need to be taken into account [19].
three groups. The percentage of participants who BP tends to increase with age. Thus, had our partici-
attained appropriate control of BP levels significantly in- pants not experienced any intervention, they would be
creased during follow-up time in all the three interven- expected to show an increase in their BP levels during
tion groups (P<0.001 for time) (Table 3). This beneficial the follow-up period [20]; however, they actually had a
significant within-group change was maintained after decrease in their BP levels during the intervention. It
Toledo et al. BMC Medicine 2013, 11:207 Page 5 of 9
[Link]

Table 1 Baseline characteristics of the PREDIMED trial

participants according to intervention group
Characteristic MD+EVOO, MD+nuts, Control diet,
n = 2441 n = 2367 n = 2350
Female sex, n (%) 1424 (58.3) 1275 (53.9) 1402 (59.7)
Age, years 66.9 ± 6.2 66.6 ± 6.1 67.3 ± 6.3
Smoking habit, n (%)
Never-smoker 1505 (61.7) 1414 (59.7) 1462 (62.2)
Former smoker 599 (24.5) 613 (25.9) 561 (23.9)
Current smoker 337 (13.8) 340 (14.4) 327 (13.9)
Body mass indexa 29.9 ± 3.7 29.7 ± 3.8 30.2 ± 4.0
Waist circumference, cm 100 ± 10 100 ± 11 101 ± 11
Hypertension, n (%)b 1999 (81.9) 1951 (82.4) 1972 (83.9)
Systolic BP, mmHg 148 ± 19 149 ± 18 149 ± 19
Diastolic BP, mmHg 83 ± 10 83 ± 10 82 ± 10
Anti-hypertensive therapy, n (%) 1660 (68.0) 1648 (68.4) 1666 (70.9)
Type 2 diabetes, n (%)c 1224 (50.1) 1092 (46.1) 1127 (48.0)
Treatment for type 2 diabetes, 801 (32.8) 715 (30.2) 786 (33.5)
n (%)
Dyslipidemia, n (%)d 1755 (71.9) 1741 (73.6) 1697 (72.2)
Lipid-lowering therapy, n (%) 1095 (44.9) 1041 (44.0) 1032 (43.9)
Family history of premature 553 (22.7) 514 (21.7) 538 (22.9)
CHD, n (%)e
MD adherence scoref 8.7 ± 2.0 8.8 ± 2.0 8.4 ± 2.0
Dietary sodium intake, g/day 2.4 ± 0.9 2.4 ± 0.9 2.3 ± 0.9
Dietary potassium intake, g/day 4.4 ± 1.2 4.4 ± 1.1 4.2 ± 1.1
Dietary calcium intake, g/day 1.1 ± 0.4 1.1 ± 0.4 1.0 ± 0.4
CHD, coronary heart disease; MD+EVOO, Mediterranean diet plus extra virgin
olive oil; MD+nuts, Mediterranean diet plus nuts.
Values are mean ± SD unless otherwise specified.
a
Body mass index is the weight in kilograms divided by the square of the
height in meters.
b
Hypertension was defined as systolic BP ≥140 mm Hg, diastolic BP ≥90 mm
Hg, or use of anti-hypertensive therapy.
c
Diabetes was defined as fasting blood glucose ≥126 mg/dl (7.0 mmol/l) on
two occasions, or 2-hour plasma glucose ≥ 200 mg/dl (11.1 mmol/l) after a
Figure 2 Adjusted mean systolic and diastolic blood pressure
75 g oral glucose load, or use of anti-diabetic medication. at baseline and yearly visits according to intervention group.
d
Dyslipidemia was defined as low-density lipoprotein cholesterol >160 mg/dl, Values are adjusted for center, sex, age, type 2 diabetes and baseline
high-density lipoprotein cholesterol ≤40 mg/dl in men or ≤50 mg/dl in blood pressure.
women, or use of lipid-lowering therapy.
e
A family history of premature CHD was defined as diagnosis of the disease in
a male first-degree relative before the age of 55 years or in a female first-
degree relative before the age of 65 years.
f
MedDiet adherence score (minimum adherence = 0 points; maximum already attending medical consultations regularly prior
adherence = 14 points). to study entry. In fact, the number of outpatient contacts
in 2004 in Spain was higher than the average in the
European Union [21]. In addition, our participants were
could be argued that participants in clinical trials already under medical treatment for their cardiovascular
undergo more evaluations, and thus their doctors are risk factors.
more likely to prescribe a better adjustment of their Current guidelines for the prevention and treatment of
anti-hypertensive medication. However, it should be high BP recommend adhering to the DASH diet [22],
noted that the intervention was based only on dietary which is a healthy eating plan low in saturated fat, chol-
changes, and no adjustments in the participants’ regular esterol, and in total fat. This diet emphasizes the con-
prescriptions were part of the intervention. In addition, sumption of fruits, vegetables, and fat-free or low-fat
participants in the PREDIMED trial were recruited from milk and dairy products [5]. The basic recommendations
their primary healthcare centers, therefore, they were provided to our participants in the control group had
Toledo et al. BMC Medicine 2013, 11:207 Page 6 of 9
[Link]

Table 2 Mean differences in BP changes (mmHg) in the two intervention groups versus the control group after 4 years
of follow-up (median follow-up 3.8 years)
MD+EVOO versus control P value MD+nuts versus control P value
Systolic BP
Crude model 0.42 (−0.46 to 1.30) 0.35 −0.90 (−1.77 to −0.03) 0.04
Multivariate-adjusted 1 a
0.40 (−0.46 to 1.27) 0.36 −0.73 (−1.58 to 0.13) 0.10
Multivariate-adjusted 2b 0.41 (−0.46 to 1.28) 0.35 −0.72 (−1.57 to 0.13) 0.10
Multivariate-adjusted 3 c
0.39 (−0.48 to 1.26) 0.38 −0.72 (−1.58 to 0.13) 0.10
Diastolic BP
Crude model −1.41 (−1.92 to −0.91) <0.001 −0.61 (−1.12 to −0.09) 0.02
Multivariate-adjusted 1a −1.49 (−1.98 to −1.00) <0.001 −0.65 (−1.15 to −0.14) 0.01
Multivariate-adjusted 2 b
−1.49 (−1.97 to −1.00) <0.001 −0.64 (−1.15 to −0.14) 0.01
Multivariate-adjusted 4d −1.53 (−2.01 to −1.04) <0.001 −0.65 (−1.15 to −0.15) 0.01
BP, blood pressure; MD+EVOO, Mediterranean diet plus extra virgin olive oil; MD+nuts, Mediterranean diet plus nuts.
a
Adjusted for center, age, sex and diabetes.
b
Multivariate-adjusted 1 + additional adjustment for number of baseline anti-hypertensive drugs.
c
Multivariate-adjusted 2 + additional adjustment for baseline systolic BP.
d
Multivariate-adjusted 2 + additional adjustment for baseline diastolic BP.

many aspects in common with the DASH eating plan, syndrome found beneficial effects on average systolic
and was not similar to a conventional placebo. There- and diastolic BP levels [9]. Similarly, we found a signifi-
fore, it was foreseeable that participants in the control cant decrease in systolic and diastolic BP in both MD
group would also improve their BP levels if they groups. Even though the intervention did not target so-
followed this advice. In fact, a cohort study with healthy dium intake, participants in the PREDIMED trial on the
young participants also conducted in Spain previously whole did experience a significant reduction in their aver-
reported an inverse association between adherence to age sodium intake, as measured by the semi-quantitative
the DASH diet and incident hypertension [23]. Thus, food-frequency questionnaire. In addition, we found
had we had a ‘true’ control group (for example, with a between-group significant differences (P<0.001) in sodium
typical Western dietary pattern, or with no intervention reductions favoring the two MD groups. Specifically, par-
at all) the between-group differences both in stroke and ticipants in both intervention (MD) groups experienced
BP would have been greater. Regarding the intervention greater sodium reductions than did participants in the
groups, the traditional MD is also rich in fruits and veg- control group. However, these differences are unlikely to
etables, has low content of saturated fat and dietary explain the observed results, as 1-year changes in sodium
cholesterol, and is rich in magnesium and potassium intake were not significantly associated with 1-year
[24], thus, in spite of its high total fat content, the MD changes in BP after adjustment for major confounders, in-
could enhance BP control. Even though greater adher- cluding the allocation group (data not shown). In addition,
ence to the MD has shown no association with incident changes in potassium or calcium were also significantly
hypertension in some large cohorts [25], a meta-analysis associated with changes in BP in the multivariate ana-
of trials with the MD on the components of metabolic lyses (data not shown). When we compared the two

Table 3 Percentage of participants with controlled BP levels (systolic BP <140 mmHg and diastolic BP <90 mmHg)
during follow-up in the PREDIMED triala
Intervention group
MedDiet+EVOO P value b
MedDiet+nuts P valueb Control P valueb P valuec
Baseline 33.6 (31.7 to 35.5) – 31.1 (29.3 to 33.0) – 31.1 (29.2 to 33.0) – NA
Year 1 36.2 (34.2 to 38.2) – 36.9 (34.8 to 39.0) – 37.2 (34.9 to 39.4) – >0.99
Year 2 38.6 (36.5 to 40.7) – 40.4 (38.2 to 42.6) – 41.4 (38.9 to 43.9) – 0.35
Year 3 37.8 (35.7 to 40.0) – 39.2 (36.8 to 41.5) – 39.0 (36.4 to 41.5) – >0.99
Year 4 39.9 (37.4 to 42.3) <0.001 41.5 (38.8 to 44.3) <0.001 42.6 (39.5 to 45.7) <0.001 0.69
BP, blood pressure; MD+EVOO, Mediterranean diet plus extra virgin olive oil; MD+nuts, Mediterranean diet plus nuts; NA, not applicable.
a
Adjusted for center, age, sex and diabetes.
b
P-value for within-group changes.
c
P-value for between-group changes, after adjustment for multiple comparisons with the Bonferroni method.
Toledo et al. BMC Medicine 2013, 11:207 Page 7 of 9
[Link]

intervention groups with the control group, we found a similar to the trial Mediterranean diet. In addition, even
significantly larger decrease in diastolic BP in both MD though participants in the control group received advice
groups than in the control (low-fat) group. These results to reduce fat intake, changes in total fat were small, and
suggest that the MD may have a greater effect on diastolic the largest differences at the end of the trial were in the
BP control than a low-fat diet. Even though the between- distribution of fat subtypes. The good quality of the diet
group differences may seem small, it has been estimated in the control group may have impaired our ability to
that small differences in BP may have a large influence on find large between-group differences in BP changes. Not-
cardiovascular and total mortality [26]. This influence withstanding, a significantly better adherence to the pre-
needs to be considered within the context of the popula- scribed diet was found in the two MD groups than in
tion strategy for preventive medicine [27]. Considering the the control group, and after the first follow-up year,
strong association between diastolic BP and vascular mor- mean scores of adherence to the prescribed diet were
tality [28], these results have important clinical relevance significantly higher in the two MD groups than in the
but need to be taken in consideration when explaining the control diet group (P<0.001 for all yearly comparisons
mechanisms of CVD risk reduction of the MD. from years 1 to 4 of follow-up). After 3 years of follow-
Our results may not seem in perfect agreement with up we found significantly better scores in both MD
our previously published results in a small subsample of groups than in the control group for 12 of the 14 items
PREDIMED participants (our pilot study) after only a 3- included in the MD adherence screening questionnaire.
month follow-up. Greater reductions in systolic and dia- Therefore, a modest change in many aspects of the over-
stolic BP were then seen in both MD groups compared all dietary pattern, and not only in the provided supple-
with the control group [29]. There may be several expla- mental foods, was achieved with our intervention. Third,
nations for these differences. First, only the first partici- information on BP during follow-up was not available
pants recruited for the trial were included in the pilot for a subset of participants, especially in the control
study. Second, the current work is based on a longer group. As has been already published, participants for
follow-up than the pilot study, therefore, a different and whom this information was not available during follow-
longer induction period is assumed. Third, in 2006 the up had a worse cardiovascular profile at study inception
protocol was reviewed; prior to 2006, no active nutri- than participants who were retained, suggesting a bias
tional education was given to the control group to foster toward a benefit in the control group [8]. Fourth, our
their adherence to the low-fat diet, and they received participants lived in a Mediterranean country, had a high
only an information brochure. After the protocol review, cardiovascular risk, and were mainly hypertensive sub-
an educational intervention was also devised and jects; all these characteristics may limit the generalizability
implemented for participants in the control group to of our findings. Fifth, information on anti-hypertensive
promote the adherence to a low-fat diet with similar drugs dosage was not available, and this precluded a de-
methodology to that of the two MD groups. Fourth, a tailed analysis on anti-hypertensive drug usage. However,
higher rate of loss to follow-up occurred in the control because all participants usually attend consultations with
group than in the two MD groups. It is possible that the their primary healthcare providers, it is unlikely that par-
participants retained in the control group had a healthier ticipants in one or the other group would be differentially
profile, as suggested by their baseline information [8]; treated.
this would selectively bias the results in the control The strengths of the study include the randomized de-
group towards better BP levels had all participants in sign, the long duration of the intervention, the high
this group been followed up. compliance of the participants allocated to the MD with
The present study has several limitations. First, the intended intervention, the large study size, and the
changes in BP were a secondary end-point, not the pri- uniformity of study implementation across the different
mary end-point of the PREDIMED trial. Nevertheless, study sites.
changes in BP were included in the protocol as a sec-
ondary specific aim from the very beginning of the trial
design. Second, at baseline, a initial fair level of adher- Conclusions
ence to the MD was present in all participants, regard- In conclusion, our randomized trial conducted in pa-
less of their allocated group, and participants in the tients at high risk of CVD supports the traditional MD
control group also maintained their relatively high scores supplemented with either EVOO or nuts and the control
of adherence to the traditional MD during the study [8]. diet as having beneficial effects on BP. After 4 years of
Therefore, the magnitude of attained between-group dif- follow-up, lower values of diastolic BP were seen in the
ferences in adherence to the MD during follow-up was two groups that received an intervention with a trad-
not large. These modest differences can be explained be- itional MD supplemented with either EVOO or with
cause for most participants their baseline diet was nuts than in the control group.
Toledo et al. BMC Medicine 2013, 11:207 Page 8 of 9
[Link]

Additional file (walnuts), Borges S.A. (almonds), and La Morella Nuts (hazelnuts), both from
Reus, Spain. CIBERobn and RTIC RD 06/0045 are initiatives of ISCIII, Spain. The
funding sources played no role in the design, collection, analysis, or
Additional file 1: Institutional Review Boards that approved the
interpretation of the data, or in the decision to submit the manuscript for
PREDIMED trial protocol.
publication.

Abbreviations Financial support

BP: Blood pressure; CHD: Coronary heart disease; CI: Confidence interval; This work was supported by the Official Funding Agency for Biomedical
CVD: Cardiovascular disease; MD: Mediterranean diet; PREDIMED: Prevencion Research of the Spanish Government, Instituto de Salud Carlos III (ISCIII),
con Dieta Mediterranea. through grants provided to research networks specifically developed for the
trial (RTIC G03/140, to Dr Estruch; RTIC RD 06/0045, to Dr Martínez-González
Competing interests and through Centro de Investigación Biomédica en Red de Fisiopatología de
RE has served on the board of and received lecture fees from the Research la Obesidad y Nutrición [CIBERobn]), and by grants from Centro Nacional de
Foundation on Wine and Nutrition (FIVIN); served on the boards of the Beer Investigaciones Cardiovasculares (CNIC 06/2007), Fondo de Investigación
and Health Foundation and the European Foundation for Alcohol Research Sanitaria-Fondo Europeo de Desarrollo Regional (PI04-2239, PI 05/2584,
(ERAB); received lecture fees from Cerveceros de España and Sanofi-Aventis; CP06/00100, PI07/0240, PI07/1138, PI07/0954, PI 07/0473, PI10/00802, PI10/
and received grant support through his institution from Novartis. ER has 01407, PI10/02658, PI11/01647, and P11/02505), Ministerio de Ciencia e
served on the board of and received personal travel support as well as grant Innovación (AGL-2009- 13906-C02 and AGL2010-22319-C03), Fundación
support through his institution, from the California Walnut Commission; Mapfre 2010, Agencia Canaria de Investigación, Innovación y Sociedad de la
served on the board of the Flora Foundation (Unilever); served on the board Información-EU FEDER (PI 2007/050), Consejería de Salud de la Junta de
of and received lecture fees from Roche; served on the board of and Andalucía (PI0105/2007), Public Health Division of the Department of Health
received grant support through his institution from Amgen; received of the Autonomous Government of Catalonia, Generalitat Valenciana (ACOM/
consulting fees from Damm and Abbott; received consulting fees and 2012/238, ACOMP06109, GVACOMP2010-181, GVACOMP2011-151, CS2010-
lecture fees, as well as grant support through his institution, from Merck; AP-111, CS2011-AP-042 ACOMP/2013/159, and ACOMP/213/165), and
received lectures fees from Danone, Pace, AstraZeneca, and Rottapharm; Regional Government of Navarra (P27/2011). ET is supported by a Rio
received lecture fees and payment for the development of educational Hortega post-residency fellowship of the Instituto de Salud Carlos III, Ministry
presentations, as well as grant support through his institution, from Ferrer; of Economy and Competitiveness, Spanish Government and by the
received payment for the development of educational presentations from Fundación Mutua Madrileña (Spain).
Ricordati; and received grant support through his institution from Sanofi-
Aventis, Takeda, Daiichi Sankyo, Nutrexpa, Feiraco, Unilever, and Karo Bio. Author details
1
JS-S has served on the board of and received grant support through his Department of Preventive Medicine and Public Health, University of Navarra,
institution from the International Nut and Dried Fruit Council; received Pamplona, Navarra, Spain. 2Department of Nutrition, Harvard School of Public
consulting fees from Danone; and received grant support through his Health, Boston, MA, USA. 3CIBER Fisiopatología de la Obesidad y Nutrición
institution from Eroski and Nestlé. FA has received payment for the (CIBER OBN), Instituto de Salud Carlos III (ISCIII), Spanish Government, Madrid,
development of educational presentations from Menarini and AstraZeneca. Spain. 4Harvard Medical School and Channing Lab, Brigham Women’s
RM L-R has served on the board of and received lecture fees from FIVIN; Hospital, Boston, MA, USA. 5Department of Internal Medicine, Institut
received lecture fees from Cerveceros de España; and received lecture fees d’Investigacions Biomèdiques August Pi Sunyer (IDIBAPS), Hospital Clinic,
and travel support from PepsiCo. LS-M has served on the boards of the University of Barcelona, Barcelona, Spain. 6Department of Preventive
Mediterranean Diet Foundation and the Beer and Health Foundation. SP has Medicine and Public Health, University of Valencia, Valencia, Spain. 7Human
served on the board of and received grant support through his institution Nutrition Department, IISPV, Universitat Rovira i Virgili, Reus, Spain.
8
from the Residual Risk Reduction Initiative (R3i) Foundation; served on the Cardiovascular and Nutrition Research Group, Institut de Recerca Hospital
board of Omega Fort; served on the board of and received payment for the del Mar, Barcelona, Spain. 9Department of Cardiology, University Hospital of
development of educational presentations, as well as grant support through Alava, Vitoria, Spain. 10Department of Preventive Medicine, University of
his institution, from Ferrer; received consulting fees from Abbott; received Malaga, Malaga, Spain. 11Institute of Health Sciences IUNICS, University of the
lecture fees, as well as grant support through his institution, from Merck and Balearic Islands, and Hospital Son Espases, Palma de Mallorca, Spain.
12
Roche; received lecture fees from Danone and Esteve; received payment for Department of Family Medicine, Primary Care Division of Sevilla, San Pablo
the development of educational presentations from Menarini; and received Health Center, Sevilla, Spain. 13Department of Clinical Sciences, University of
grant support through his institution from Sanofi-Aventis, Kowa, Unilever, Las Palmas de Gran Canaria, Las Palmas, Spain. 14Lipids and Vascular Risk
Boehringer Ingelheim, and Karo Bio. No other potential conflict of interest Unit, Internal Medicine, Hospital Universitario de Bellvitge, Hospitalet de
relevant to this article is reported. Llobregat, Barcelona, Spain. 15Department of Nutrition and Food Science,
School of Pharmacy, XaRTA, INSA, University of Barcelona, Barcelona, Spain.
16
Authors’ contributions Department of Biochemistry and Molecular Biology, School of Medicine,
ET analysed and interpreted the data and was primarily responsible for University of Valencia, Valencia, Spain. 17Department of Clinical Analyses,
writing the manuscript. FBH made substantial contributions to the study University Hospital of Valencia, Valencia, Spain. 18Instituto de la Grasa,
design and interpretation of data and critically reviewed the intellectual Consejo Superior de Investigaciones Cientificas, Sevilla, Spain. 19Lipid Clinic,
content. RE, PB-C, JS-S, and LS-M made substantial contributions to the Department of Endocrinology and Nutrition, Institut d’Investigacions
acquisition of data and their interpretation, and to the trial implementation. Biomèdiques August Pi Sunyer (IDIBAPS), Hospital Clinic, University of
DC, MIC, FA, EG-G, MF, JL, XP, and JVS substantially contributed to the Barcelona, Barcelona, Spain. 20Primary Care Division, Valencia Institute of
acquisition of data and trial implementation. RML-R, GS, and VR-G Health, Valencia, Spain.
substantially contributed to the study conception. MB revised the manuscript
for important intellectual content. ER made substantial contributions to study Received: 11 June 2013 Accepted: 23 August 2013
design and interpretation of data. MAM-G substantially contributed to the Published: 19 September 2013
study conception and design, trial implementation, amd acquisition and
interpretation of data, and critically revised the manuscript for important References
intellectual content. All authors read and approved the final manuscript. 1. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, Jones
DW, Materson BJ, Oparil S, Wright JT Jr, et al: The Seventh Report of the
Acknowledgements Joint National Committee on Prevention, Detection, Evaluation, and
We thank the participants in the trial for their enthusiastic and maintained Treatment of High Blood Pressure: the JNC 7 report. JAMA 2003,
collaboration. The supplemental foods used in the study were generously 289(19):2560–2572.
donated by Patrimonio Comunal Olivarero and Hojiblanca from Spain (extra 2. Lawes CM, Vander Hoorn S, Rodgers A: Global burden of blood-pressure-related
virgin olive oil), the California Walnut Commission from Sacramento, CA disease, 2001. Lancet 2008, 371(9623):1513–1518.
Toledo et al. BMC Medicine 2013, 11:207 Page 9 of 9
[Link]

3. Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJ: Global and of hypertension: 6-year follow-up of the SUN (Seguimiento Universidad
regional burden of disease and risk factors, 2001: systematic analysis of de Navarra) prospective cohort. Public Health Nutr 2010, 13(3):338–349.
population health data. Lancet 2006, 367(9524):1747–1757. 24. Trichopoulou A, Costacou T, Bamia C, Trichopoulos D: Adherence to a
4. U.S. Department of Agriculture and U.S. Department of Health and Human Mediterranean diet and survival in a Greek population. New Engl J Med
Services: Dietary Guidelines for Americans. Washington, DC: US Government 2003, 348(26):2599–2608.
Printing Office; 2010. 25. Nunez-Cordoba JM, Valencia-Serrano F, Toledo E, Alonso A, Martinez-
5. Appel LJ, Moore TJ, Obarzanek E, Vollmer WM, Svetkey LP, Sacks FM, Bray Gonzalez MA: The Mediterranean diet and incidence of hypertension: the
GA, Vogt TM, Cutler JA, Windhauser MM, et al: A clinical trial of the effects Seguimiento Universidad de Navarra (SUN) Study. Am J Epidemiol 2009,
of dietary patterns on blood pressure. DASH Collaborative Research 169(3):339–346.
Group. New Engl J Med 1997, 336(16):1117–1124. 26. Appel LJ, Brands MW, Daniels SR, Karanja N, Elmer PJ, Sacks FM: Dietary
6. Sofi F, Abbate R, Gensini GF, Casini A: Accruing evidence on benefits of approaches to prevent and treat hypertension: a scientific statement
adherence to the Mediterranean diet on health: an updated systematic from the American Heart Association. Hypertension 2006, 47(2):296–308.
review and meta-analysis. Am J Clin Nutr 2010, 92(5):1189–1196. 27. Rose G, Khaw KT, Marmot M: Rose’s Strategy of Preventive Medicine, Updated
7. Willett WC, Sacks F, Trichopoulou A, Drescher G, Ferro-Luzzi A, Helsing E, edition. USA: Oxford University Press; 2008.
Trichopoulos D: Mediterranean diet pyramid: a cultural model for healthy 28. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R: Age-specific relevance
eating. Am J Clin Nutr 1995, 61(6 Suppl):1402S–1406S. of usual blood pressure to vascular mortality: a meta-analysis of
8. Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, Aros F, Gomez-Gracia individual data for one million adults in 61 prospective studies.
E, Ruiz-Gutierrez V, Fiol M, Lapetra J, et al: Primary prevention of Lancet 2002, 360(9349):1903–1913.
cardiovascular disease with a Mediterranean diet. Am J Clin Nutr 2013, 29. Estruch R, Martinez-Gonzalez MA, Corella D, Salas-Salvado J, Ruiz-Gutierrez V,
368(14):1279–1290. Covas MI, Fiol M, Gomez-Gracia E, Lopez-Sabater MC, Vinyoles E, et al:
9. Kastorini CM, Milionis HJ, Esposito K, Giugliano D, Goudevenos JA, Effects of a Mediterranean-style diet on cardiovascular risk factors: a
Panagiotakos DB: The effect of Mediterranean diet on metabolic randomized trial. Ann Intern Med 2006, 145(1):1–11.
syndrome and its components: a meta-analysis of 50 studies and
534,906 individuals. J Am Coll Cardiol 2011, 57(11):1299–1313. doi:10.1186/1741-7015-11-207
10. Martinez-Gonzalez MA, Corella D, Salas-Salvado J, Ros E, Covas MI, Fiol M, Cite this article as: Toledo et al.: Effect of the Mediterranean diet on
Warnberg J, Aros F, Ruiz-Gutierrez V, Lamuela-Raventos RM, et al: Cohort blood pressure in the PREDIMED trial: results from a randomized
profile: design and methods of the PREDIMED study. Int J Epidemiol 2012, controlled trial. BMC Medicine 2013 11:207.
41(2):377–385.
11. Martin-Moreno JM, Boyle P, Gorgojo L, Maisonneuve P, Fernandez-
Rodriguez JC, Salvini S, Willett WC: Development and validation of a food
frequency questionnaire in Spain. Int J Epidemiol 1993, 22(3):512–519.
12. Fernandez-Ballart JD, Pinol JL, Zazpe I, Corella D, Carrasco P, Toledo E, Perez-
Bauer M, Martinez-Gonzalez MA, Salas-Salvado J, Martin-Moreno JM:
Relative validity of a semi-quantitative food-frequency questionnaire in
an elderly Mediterranean population of Spain. Brit J Nutr 2010, 103
(12):1808–1816.
13. Elosua R, Garcia M, Aguilar A, Molina L, Covas MI, Marrugat J: Validation of
the Minnesota Leisure Time Physical Activity Questionnaire in Spanish
women. Investigators of the MARATDON Group. Med Sci Sport Exer 2000,
32(8):1431–1437.
14. Elosua R, Marrugat J, Molina L, Pons S, Pujol E: Validation of the Minnesota
Leisure Time Physical Activity Questionnaire in Spanish men. The
MARATHOM Investigators. Am J Epidemiol 1994, 139(12):1197–1209.
15. Schroder H, Fito M, Estruch R, Martinez-Gonzalez MA, Corella D, Salas-
Salvado J, Lamuela-Raventos R, Ros E, Salaverria I, Fiol M, et al: A short
screener is valid for assessing Mediterranean diet adherence among
older Spanish men and women. J Nutr 2011, 141(6):1140–1145.
16. O’Brien E, Mee F, Atkins N, Thomas M: Evaluation of three devices for self-
measurement of blood pressure according to the revised British
Hypertension Society Protocol: the Omron HEM-705CP, Philips HP5332,
and Nissei DS-175. Blood Press Monit 1996, 1(1):55–61.
17. Iglesias Bonilla P, Mayoral Sanchez E, Lapetra Peralta J, Iborra Oquendo M,
Villalba Alcala F, Cayuela Dominguez A: [Validation of two systems of self-
measurement of blood pressure, the OMRON HEM-705 CP and OMRON
M1 (HEM 422C2-E) models]. Atencion primaria / Sociedad Espanola de
Medicina de Familia y Comunitaria 2002, 30(1):22–28.
18. Twisk JWR: Applied Longitudinal Data Analysis for Epidemiology. Cambridge:
Cambridge University Press; 2003. Submit your next manuscript to BioMed Central
19. Libby P: Mechanisms of acute coronary syndromes and their implications and take full advantage of:
for therapy. New Engl J Med 2013, 368(21):2004–2013.
20. Singh GM, Danaei G, Pelizzari PM, Lin JK, Cowan MJ, Stevens GA, Farzadfar
• Convenient online submission
F, Khang YH, Lu Y, Riley LM, et al: The age associations of blood pressure,
cholesterol, and glucose: analysis of health examination surveys from • Thorough peer review
international populations. Circulation 2012, 125(18):2204–2211. • No space constraints or color figure charges
21. Martin-Moreno JM, Alonso P, Claveria A, Gorgojo L, Peiro S: Spain: a
decentralised health system in constant flux. BMJ 2009, 338:b1170. • Immediate publication on acceptance
22. US Department of Agriculture and US Department of Health and Human • Inclusion in PubMed, CAS, Scopus and Google Scholar
Services: DASH Eating Plan. Washington, DC: USDA; 2006. • Research which is freely available for redistribution
23. Toledo E, de A Carmona-Torre F, Alonso A, Puchau B, Zulet MA, Martinez JA,
Martinez-Gonzalez MA: Hypothesis-oriented food patterns and incidence
Submit your manuscript at
[Link]/submit