Published by Oxford University Press on behalf of the International Epidemiological Association International Journal of Epidemiology 2012;41:377–385

ß The Author 2010; all rights reserved. Advance Access publication 20 December 2010 doi:10.1093/ije/dyq250

COHORT PROFILE

Cohort Profile: Design and methods of the

PREDIMED study
Miguel Ángel Martı́nez-González,1*y Dolores Corella,2,3 Jordi Salas-Salvadó,3,4 Emilio Ros,3,5
Marı́a Isabel Covas,3,6 Miquel Fiol,3,7 Julia Wärnberg,1,8 Fernando Arós,9 Valentina Ruı́z-Gutiérrez,10
Rosa Marı́a Lamuela-Raventós,11 Jose Lapetra,3,12 Miguel Ángel Muñoz,13 José Alfredo Martı́nez,3,14
Guillermo Sáez,15 Lluis Serra-Majem,16 Xavier Pintó,17 Marı́a Teresa Mitjavila,18 Josep Antoni Tur,19
Marı́a del Puy Portillo20 and Ramón Estruch3,21y, for the PREDIMED Study Investigators
1
Department of Preventive Medicine and Public Health, University of Navarra, Pamplona, Spain, 2Department of Preventive
Medicine, University of Valencia, Valencia, Spain, 3CIBER Fisiopatologı́a de la Obesidad y Nutrición (CIBERobn), Instituto de Salud
Carlos III (ISCIII), Spain, 4Human Nutrition Unit, IISPV, Universitat Rovira i Virgili, Reus, Spain, 5Lipid Clinic, Department of
Endocrinology and Nutrition, Institut d’Investigacions Biomèdiques August Pi Sunyer (IDIBAPS), Hospital Clı́nic, Barcelona, Spain,
6
Lipids and Cardiovascular Epidemiology Research Unit, Institut Municipal d’Investigació Mèdica (IMIM), Barcelona, Spain,
7
Institute of Health Sciences (IUNICS), University of Balearic Islands, Palma de Mallorca, Spain, 8Department of Preventive
Medicine, University of Málaga, Málaga, Spain, 9Department of Cardiology, University Hospital Txagorritxu, Vitoria, Spain,
10
Instituto de la Grasa, Consejo Superior de Investigaciones Cientı́ficas, Sevilla, Spain, 11Nutrition and Food Science Department–
XaRTA, INSA, University of Barcelona, Barcelona, Spain, 12Department of Family Medicine, Primary Care Division of Sevilla, Centro
de Salud San Pablo, Sevilla, Spain, 13Primary Care Division, Catalan Institute of Health, Barcelona, Spain, 14Department of
Nutrition and Food Sciences, Physiology and Toxicology, University of Navarra, Pamplona, Spain, 15Department of Biochemistry
and Molecular Biology, CDB-HGUV, University of Valencia, Valencia, Spain, 16Department of Clinical Sciences, University of Las
Palmas de Gran Canaria, Las Palmas, Spain, 17Lipids and Vascular Risk Unit, Internal Medicine, Hospital Universitario de Bellvitge,
Hospitalet de Llobregat, Barcelona, Spain, 18Department of Physiology, University of Barcelona, Spain, 19Department of
Fundamental Biology and Health Sciences, University of Balearic Islands, Palma de Mallorca, Spain, 20Department of Nutrition and
Food Science, University of Basque Country, Vitoria, Spain and 21Department of Internal Medicine, IDIBAPS, Hospital Clinic,
University of Barcelona, Barcelona, Spain
*Corresponding author. Department of Preventive Medicine and Public Health, University of Navarra, C/Irunlarrea, 1.
31080-Pamplona (Navarra), Spain. E-mail: mamartinez@unav.es
y
These authors contributed equally to this work

Accepted 29 November 2010

How did the PREDIMED study forms and data entry/management systems. The
needed personnel (a minimum of a dietician and a
come about? nurse for each of the 11 field centres, FCs) were
A call for grants was issued in 2002 by the Spanish then hired, trained and certified. Each FC contacted
Government (Instituto de Salud Carlos III). This call approximately 20 primary care practices (PCPs) to
was specifically designed to initiate networking re- recruit participants. The recruitment of participants
search among Spanish biomedical investigators. started in October 2003. The name PREDIMED (in
During 2002, Ramón Estruch—the leader of our ini- Spanish: PREvención con DIeta MEDiterránea) was
tiative—contacted different Spanish investigators (the proposed by Dolores Corella. This name is applied to
rest of us) working in nutrition from different per- both the cohort study and the networking group.
spectives. We applied together for a grant to start a Despite being an interventional study, the
large randomized trial to test the effectiveness of a PREDIMED study provides a unique opportunity
Mediterranean diet (MeDiet) on the ‘primary’ preven- for conducting the long-term follow-up (after the
tion of cardiovascular disease (CVD) and to continue completion of the trial) of a large observational
the study as an observational cohort of high-risk par- cohort of high cardiovascular risk subjects in a
ticipants to be followed-up in the long term. On Mediterranean setting.
6 January 2003 our project was funded. The MeDiet represents the dietary exposure in clo-
From January to June 2003, we developed the proto- sest agreement with the Bradford Hill criteria for a
col: logistics, manual of operations, instruments, potential causal protection against coronary heart

377
378 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

disease (CHD) according to a recent systematic higher palatability and acceptance of MeDiets in
review.1 This conclusion is mainly supported by ob- comparison with low-fat diets17,18 lend support to
servational cohort studies. A recent meta-analysis of our hypothesis.
these cohorts showed that adherence to the MeDiet The results of some observational studies have been
was associated with reductions in total mortality and subsequently refuted by evidence from clinical trials
CHD mortality.2 Subsequently, similar evidences have (i.e. the presumptive cardioprotective effects of
been collected for ‘non-fatal’ CVD.3–5 An increasing postmenopausal hormone therapy19 or antioxidant
body of evidence is supporting also a benefit of the supplements20). This highlights the need to obtain
MeDiet against major cancers and neurodegenerative first-level evidence before considering any global
diseases.6–10 public health strategy. Dietary guidelines can be
Our hypothesis was that two traditional MeDiets, safely issued when consistency is found between
one enriched with virgin olive oil (VOO) and another observational and experimental studies.
enriched with nuts, both high in total fat and unsat-
urated fat, would be superior to the usually recom-
mended low-fat diet for the primary prevention of
CVD in a high-risk population. This fit well into the
What does the PREDIMED study
paradigm of focusing on dietary patterns instead of cover?
isolated foods or nutrients. Overall patterns better A large cohort has been assembled for long-term
represent dietary practices found in free-living popu- follow-up. This cohort includes 7447 high-risk partici-
lations and provide useful epidemiological informa- pants. The last participant was recruited on 30 June
tion with a high potential for acceptability, 2009. Trial closeout will take place by 31 December
palatability and future compliance. 2011. Subsequent follow-up will continue as an
However, no randomized controlled trial has ever observational multi-purpose cohort to explore other
been conducted to test the MeDiet in the ‘primary hypotheses (i.e. the roles of different types of alco-
prevention’ of major chronic diseases. The only avail- holic beverages in cancer or CVD prevention) and to
able clinical trial supporting a cardioprotective role of develop nested case–control analyses for studies of
the MeDiet is the Lyon Diet Heart Study.11 It was an biomarkers and gene–nutrient interactions.
important step to support the benefits of the MeDiet, The ‘primary aim’ of the trial is to assess the effects
but it included only myocardial infarction survivors of two MeDiets on a composite endpoint of cardio-
(i.e. it was a ‘secondary’ prevention trial). However, vascular death, myocardial infarction and stroke (‘pri-
it showed a remarkable 50–70% reduction in CHD mary endpoint’) in comparison with a low-fat control
event rates and mortality with a ‘MeDiet’ (enriched diet. ‘Secondary endpoints’ are death of any cause,
with a-linolenic acid, but not with olive oil). These incidence of heart failure, diabetes mellitus, dementia
results were criticized because no special consider- or other neurodegenerative disorders and major can-
ation was given to olive oil, which is the major cers (colorectal, breast, lung, stomach and prostate).
source of dietary fat in Mediterranean countries.12 To better understand how dietary changes may
Another problem was that dietary assessments at modify the risk of clinical events, we also evaluate
baseline and at the end of the study were reported intermediate outcomes, including changes in blood
for only 30% of the control group and 50% of the pressure (BP), weight gain, fasting blood glucose,
experimental group and no biochemical markers of blood lipids and markers of inflammation.
adherence were obtained. Finally, concerns have
been raised regarding the low number of observed
endpoints (44 in the control group vs 14 in the treat-
ment group), and the improbable contrast between
Who are the participants in
the large reduction in risk and the lack of changes the PREDIMED study?
in most classical risk factors. The PREDIMED study Participants are men (55–80 years old) or women
attempts to overcome previous limitations and to pro- (60–80 years old) who were free of CVD at baseline.
vide the best quality of evidence to answer the ques- Inclusion criteria were to have either type 2 diabetes
tion of whether the MeDiet, compared with the or 53 major cardiovascular risk factors, out of the
previously tested model of advice on a low-fat diet, following: current smoking (41 cig/day during the
provides relatively higher protection against chronic last month); hypertension (systolic BP 5140 mmHg
disease. In the face of the increasing global burden or diastolic BP 590 mmHg or antihypertensive medi-
of CVD and cancer, the answer to this question is a cation); LDL cholesterol 5160 mg/dl or lipid-lowering
major public health priority. The long tradition of therapy; HDL cholesterol 440 mg/dl in men or
adherence to this food pattern in Mediterranean 450 mg/dl in women; body mass index 525 kg/m2;
countries, where CHD incidence is low despite high and family history of premature CHD.
levels of cardiovascular risk factors;13 the diversity of Exclusion criteria were the previous history of CVD
mechanisms supporting the beneficial effects on car- (i.e. a previous medical diagnosis of CHD, stroke or
diovascular health of olive oil14,15 or nuts;16 and the peripheral arterial disease), any severe chronic illness,
 DESIGN AND METHODS OF THE PREDIMED STUDY 379

8713 assessed for eligibility

1266 excluded
973 refused to participate
108 did not meet inclusion criteria
51 had prior cardiovascular disease
45 had difficulties with follow-up
39 were unwilling to change dietary habits
20 had excessive alcohol intake
12 had food allergies
10 had other chronic diseases
8 had other exclusion criteria

7447 randomized

2543 assigned to 2454 assigned to 2450 assigned

Mediterranean Mediterranean to low-fat diet
diet with virgin diet with mixed (AHA
olive oil nuts guidelines)

2-year retention ratea 2-year retention ratea 2-year retention ratea

96.2% 92.1% 82.7%

aCalculated only among participants recruited during 2003–05

Figure 1 Flow chart of participants in the PREDIMED study

immunodeficiency or human immunodeficiency virus The Antihypertensive and Lipid-Lowering Treatment

(HIV) positive status, illegal drug or alcohol misuse, to Prevent Heart Attack Trial (ALLHAT) included
history of allergy to olive oil or nuts and low predicted similar participants and observed an 8.9% cumulative
likelihood of changing dietary habits according to the rate for the primary outcome (fatal CHD þ non-fatal
Prochaska and DiClemente stages of change model.21 myocardial infarction) after 4.9 years of follow-up.22
Figure 1 provides further details on the selection Adapting this figure to a 6-year follow-up and includ-
procedure. ing also stroke in the endpoint definition, an 11%
The selection process started by extracting names absolute risk in the control group could be conserva-
of potential participants from the records of the tively assumed in our study. We expect a 25% relative
PCPs. Most PCPs participating in the study have risk reduction in both MeDiet groups. Under these
computer-based records of patients, making the selec- assumptions, the total number of participants
tion relatively simple. The clinical records of these (3 equally-sized groups) required was 5631 (1877
persons were then individually reviewed to exclude per group) for b ¼ 0.2 and two-tailed a ¼ 0.05. We
those who did not meet eligibility criteria. Potential included 7447 subjects to allow for both 10% losses
participants were approached by PCPs by a telephone during follow-up and a lower incidence than ex-
call or during their clinical visits. If candidates were pected. Figure 1 shows the flow of participants.
interested in participating, a face-to-face interview Study participants were randomized to three equally
was scheduled. During this interview, the purpose sized groups. Tables of random allocation were cen-
and characteristics of the study were explained, and trally elaborated. The study nurses in charge of the
signed informed consent was obtained from willing random allocation were independent of the nursing
participants. A brief explanation of the study, includ- staff of the PCP. At baseline, general practitioners
ing the possibility that they might receive free allow- (GPs) were not informed of the allocation of partici-
ances of VOO or nuts for the duration of the trial, was pants. This is consistent with CONSORT guidelines for
given at this first visit. Most (470%) candidates ap- randomized trials to prevent selection biases.23
proached in this way agreed to return for the screen- Characteristics of participants according to group
ing visit. allocation are shown in Table 1. The Institutional
380 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

Table 1 Description of participants in the PREDIMED study at baseline according to intervention group

MeDiet þ VOO MeDiet þ nuts Control (low-fat)

Characteristics at baseline (n ¼ 2543) (n ¼ 2454) (n ¼ 2450)
Age (mean years, SD) 67 (6) 67 (6) 67 (6)
Gender (women) (%) 58.7 54.0 59.7
Diabetes (%) 50.2 46.5 48.4
Hypertension (%) 82.1 82.4 83.7
Current smokers (%) 13.9 14.5 13.8
Former smokers (%) 24.3 25.8 23.8
High blood total cholesterol (%) 71.6 73.3 71.9
Family history of CHDa (%) 22.7 21.7 22.8
BMI (mean, SD) 30.0 (3.7) 29.7 (3.8) 30.2 (4.0)
Waist circumference (mean cm, SD) 100 (10) 100 (11) 101 (11)
Adherence to MeDietb (mean, SD) 8.7 (2.0) 8.7 (2.0) 8.4 (2.1)
Percentage of participants from each region
North (Navarra and Basque country) 23.4 22.9 22.2
North-East (Catalonia) 31.3 33.2 31.5
East (Valencia and Balearic Islands) 23.0 24.2 23.1
South (Andalusia and Canary Islands) 22.3 19.7 23.1
SD: standard deviation.
a
Definite myocardial infarction or sudden death before 55 years in male first-degree relatives or before 65 years in female
first-degree relatives.
b
14-point score of adherence to MeDiet.

Review Board (IRB) of Hospital Clinic (Barcelona, (ROO).14 ROO lose polyphenols and other elements in
Spain) approved the study protocol on July 2002. the refining process, although fatty acid composition
This IRB is accredited by the US Department of is similar to that of VOO.24 When compared with
Health and Human Services (DHHS). Later, the IRBs ROO, VOO increases HDL cholesterol, total plasma
of all other centres also approved the protocol. The antioxidant capacity and LDL resistance to oxidation.
trial is registered (http://www.controlled-trials.com/ In vivo markers of lipid and LDL oxidation decrease in
ISRCTN35739639). a dose-dependent manner with the phenolic content
of the olive oil.25 Most nuts are rich in MUFA (mostly
oleic acid), whereas walnuts are high in polyunsatur-
ated fatty acids (PUFAs, i.e. linoleic and a-linolenic
What are the interventions in acids). The dietary fibre content in nuts is also high.
the PREDIMED study? Nuts are good sources of arginine, potassium, vitamin
Participants were randomly assigned to three inter- E and other bioactive compounds. This may help
ventions: MeDiet with VOO, MeDiet with mixed explain their beneficial health effects.16 The rationale
nuts or control group (low-fat diet). The two groups for the free provision of these food items (VOO and
allocated MeDiets receive intensive education to nuts) is that they may contribute to a higher compli-
follow the MeDiet and supplemental foods at no ance with the overall MeDiet food pattern.
cost. VOO (1 l/week) is provided to the first group The PREDIMED dieticians are directly responsible
and 30 g/day of mixed nuts (15 g walnuts, 7.5 g hazel- for the dietary intervention. After two screening
nuts and 7.5 g almonds) to the second group. In the visits, participants randomized to each one of the
control group, participants do not receive education three treatment arms had a face-to-face interview
on the MeDiet, but are given advice to follow a with the dietician and a group session (less than 20
low-fat diet. subjects). A 14-point score of adherence to the MeDiet
Besides being an excellent source of monounsatu- is a main tool to change dietary habits (Table 2).26–28
rated fat (MUFA), VOO also contains significant A similar 9-point score is used for the low-fat control
amounts of phenolic antioxidants and other phyto- group. For total fat intake, the recommendations
chemicals (tocopherols, polyphenols) because it is ob- given to participants in the low-fat diet group are op-
tained from the first pressing of the ripe fruit (i.e. it is posite to those given to participants in the two
an olive juice). In contrast, these phytochemicals are MeDiet groups. The focus can be shifted from chan-
present to a lower extent in common refined olive oils ging portion sizes, frequency of intake or cooking
 DESIGN AND METHODS OF THE PREDIMED STUDY 381

Table 2 Short questionnaire to assess adherence to the MeDiet

Criteria for
Questions 1 point
1. Do you use olive oil as main culinary fat? Yes
2. How much olive oil do you consume in a given day (including oil used for frying, salads, 54 tbsp
out-of-house meals, etc.)?
3. How many vegetable servings do you consume per day? [1 serving: 200 g (consider side dishes 52 (51 portion
as half a serving)] raw or as a salad)
4. How many fruit units (including natural fruit juices) do you consume per day? 53
5. How many servings of red meat, hamburger or meat products (ham, sausage, etc.) do you <1
consume per day?
6. How many servings of butter, margarine, or cream do you consume per day? (1 serving: 12 g) <1
7. How many sweetened and/or carbonated beverages do you drink per day? < 1
8. How much wine do you drink per week? 57 glasses
9. How many servings of legumes do you consume per week? (1 serving: 150 g) 53
10. How many servings of fish or shellfish do you consume per week? (1 serving 100–150 g of fish 53
or 4–5 units or 200 g of shellfish)
11. How many times per week do you consume commercial sweets or pastries (not homemade), <3
such as cakes, cookies, biscuits or custard?
12. How many servings of nuts (including peanuts) do you consume per week? (1 serving 30 g) 51
13. Do you preferentially consume chicken, turkey or rabbit meat instead of veal, pork, hamburger Yes
or sausage?
14. How many times per week do you consume vegetables, pasta, rice or other dishes seasoned 52
with sofrito (sauce made with tomato and onion, leek or garlic and simmered with olive oil)?

methods. We have reported an adequate effectiveness every 3 months with the same contents, except that
of the intervention after 1 year of follow-up.28 shopping lists and recipes vary with the season of the
Because unsaturated fats like those contained in year. Each visit includes three steps: assessment,
olive oil and nuts are still wrongly perceived as fat- intervention and future directions. Once a year, gen-
tening, it has been particularly important to allay the eral medical and food frequency questionnaires (FFQ)
fear of an eventual weight gain. Tactful exposition are obtained, an electrocardiogram (ECG) is per-
of recent scientific evidence,18,29–31 together with formed and blood and urine samples are collected
the fact that body weight did not change after (Table 3).
3 months of MeDiet intervention in the pilot phase After the trial formally terminates (December 2011),
of the PREDIMED study,26 have been instrumental in we will follow our cohort for occurrence of clinical
achieving this aim. events, ECG and measurement of weight and BP.
The PREDIMED group sessions are organized separ- We will continue to ascertain participants’ vital
ately for each of the three intervention groups. status through yearly personal interviews by
Participants are provided with written material (see: PREDIMED personnel, close contact with GPs who
http://www.predimed.org and http://www.predimed care for them and reviews of medical records. On a
.es) including descriptions of seasonal foods, shopping yearly basis, the Spanish official mortality index
lists, weekly meal plans and cooking recipes. Olive (Indice Nacional de Defunciones) is also reviewed.
oil and nut industry companies are committed to
supplying for free the food supplements used in the
study until December 2011. None of the investigators
has any commercial interest with these food What variables are measured?
companies. Table 3 shows the variables collected in the
PREDIMED study. The yearly administered FFQ pro-
vide information about compliance with food and nu-
How often are cohort volunteers trient targets. This FFQ was previously validated in
Spain.32 We have performed a new validation of the
contacted? FFQ with high-risk persons similar to PREDIMED
Table 3 shows the frequency of contacts with partici- participants33 and have confirmed its reproducibil-
pants. The individual and group visits are repeated ity.34 Biological markers of compliance (plasma oleic
382 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

Table 3 Measurements in the PREDIMED study

Description of
measurements Number of repeated measurements
Measurements
Number Year Year Year Year Year Year Year
of items Content Baseline 1 2 3 4 5 6 7a
Eligibility 33 Socio-demographic 1
questionnaire inclusion and exclusion
criteria, smoking
General 77 Marital status, job, BP and 1
questionnaireb anthropometry, medical
conditions, medications,
CAGE
14-item MeDiet 14 MeDiet adherence 1 2–5 6–9 10–13 14–17 18–21 22–25 26–29
questionnaire (intervention tool)
Food frequency 137 Previously validated32–34 1 2 3 4 5 6 7 8
questionnaire
Physical activity 67 Validated Minnesota 1 2 3 4 5 6 7 8
questionnaire questionnaire35,36
Follow-up 75 Risk factors, symptoms 1 2 3 4 5 6 7
questionnaireb and conditions, job,
BP, anthropometry,
medication
Tolerance 6 Potential adverse events 1 2 3 4 5 6 7
questionnaire
Abandonment 21 Reasons for terminating 1 2 3 4 5 6 7
questionnairec study
ECG 1 2 3 4 5 6 7 8
Blood chemistry 14 Lipids, glucose, renal 1 2 3 4 5 6 7 8
function, transaminases,
blood count and others
Blood sample 9 tubes (38.5 ml) 40 1 2 (o) 3 (o) 4 5 6
aliquots (–808C)
Urine sample 16 aliquots (–808C) 1 2 (o) 3 (o) 4 5 6
Toenail sample A clip of each toenail 1 2
SF36d 36 Quality of life 1 2 3 4 5
CAGE, 4-item screening test for alcohol dependence; SF-36, short-form 36, 36-item questionnaire for quality of life; (o), optional
collection.
a
Only for participants recruited before 2005.
b
Includes direct measurements of weight, height, waist circumference, BP and ankle-brachial blood pressure index.
c
Only if applicable.
d
Only for participants recruited after 2007.

and a-linolenic acid proportions and urinary concen- Blood and urine samples are collected at baseline
trations of tyrosol and hydroxytyrosol, resveratrol and and Years 1, 3, 5 and 6 (or final visit). Tubes for
ethanol) are measured in random subsets of partici- EDTA plasma, citrate plasma, buffy coat and serum
pants from the three arms of the trial.28 Clinical are collected and aliquots are kept frozen (–808C).
evaluations are limited to yearly follow-up visits The Short-Form 36 to assess quality of life is com-
that include the same examinations performed at pleted by all participants recruited after 2007.
baseline, with the exception of the general question- Toenails are collected at baseline and the final visit.
naire, which is substituted by a follow-up question- In two centres (Barcelona-North and Pamplona), ca-
naire, and a tolerance/adverse events questionnaire. rotid intima–media thickness has been measured in
Although no intervention on physical activity is per- subsets of participants.37,38 Outcomes are ascertained
formed, the Minnesota physical activity questionnaire on a yearly basis by a Clinical Events Committee
(validated Spanish version)35,36 is completed each whose members are blinded to the intervention
year. group.
 DESIGN AND METHODS OF THE PREDIMED STUDY 383

What is the attrition? change long-established dietary habits and increase

adherence to a low-fat diet in participants allocated
A high retention rate is a major methodological re- the control group.
quirement in follow-up studies. The attrition rate after From a public health perspective, a behavioural
2 years’ follow-up for participants recruited before intervention coupled with an easy (free) access to rep-
2006 (n ¼ 4.381) was 9.3%. The highest retention resentative healthy foods is a realistic test of the ef-
rates occurred in the MeDiet with VOO group and fectiveness to be attained with official policies and
the lowest retention rate was observed in the control health promotion activities. The PREDIMED trial at-
(low-fat diet) group (Figure 1). The highest retention tempts to obtain relevant information for public
rate in the two groups allocated to MeDiets can be health use, because the nutritional intervention is
partly attributed to the free provision of food items undertaken in free-living persons who receive informa-
(VOO and nuts). However, in the PREDIMED trial, we tion, motivation, support and empowerment to modify
will eventually be able to obtain a nearly complete their food habits in a real-life context, i.e. they con-
follow-up for the main outcomes because participants tinue to buy their foods and cook their meals. Such an
represent a stable and well-defined population regu- intervention provides a real-life scenario that may be
larly attending their GPs. In addition, a comprehen- easily applied to public health policies.
sive search for events is performed yearly through
review of the medical records of participants in all
the hospitals of the city where the respective FC is
located. Where can I find out more?
A list of original publications and other information
can be found at www.predimed.es (or www
What has been found so far? .predimed.org). Collaboration with national and inter-
national studies is welcome and can be proposed to:
The pilot study of the PREDIMED trial (n ¼ 772) sug- restruch@clinic.ub.es.
gested that a MeDiet was a safe strategy to reduce the
levels of major cardiovascular risk factors after a
3-month follow-up.26 Inverse baseline associations
with inflammatory markers for cereals, fruits, nuts Funding
and VOO were found.39 In the first 3204 participants, The Spanish Ministry of Health—Instituto de Salud
the 14-point score was able to predict the prevalence Carlos III (ISCIII) funded the project for the period
of diabetes, hypertension and obesity or the joint 2003–05 (RTIC G03/140). In 2006 a new funding
presence of metabolic conditions.40 In a 3-month lon- modality was established by ISCIII through the
gitudinal study, we found a favourable effect of the CIBER (Centros de Investigación Biomédica En Red)
MeDiet interventions on LDL oxidation41 and cellular Fisiopatologı́a de la Obesidad y Nutrición (CIBERobn)
and serum inflammatory biomarkers related to ath- which is providing funding for 7 of the original re-
erosclerosis.42 After a 12-month follow-up of the search groups, whereas the other 12 were funded by a
first 1224 participants, the prevalence of metabolic new research network (RTIC RD 06/0045). Other of-
syndrome was reduced in all groups, but it was ficial funds from Spanish government agencies have
more marked in the MeDiet groups, especially in the been obtained for subprojects related to intermediate
MeDiet þ nuts group.43 Other studies have as- outcomes (lipoproteins, inflammatory markers, vascu-
sessed dietary associations with hypertension44–46 lar imaging, genomic and proteomic studies, etc.).
and gene–nutrient interactions in obesity and weight Obviously, the donation by food companies of all
gain.47–49 In 2009, more than 30 papers derived from the VOO and mixed nuts needed throughout the dur-
the PREDIMED study have been either published or ation of the study is a substantial contribution. None
are accepted for publication in peer-reviewed journals. of these companies (Patrimonio Comunal Olivarero,
In 2010, a substantial effect of both MeDiets in the California Walnut Commission, Borges, La Morella
reduction of type 2 diabetes risk after a median Nuts and Hojiblanca) played or will play any role in
follow-up of 4.0 years was reported from a nested the design, collection, analysis or interpretation of the
analysis conducted in one of the centres.50 data or in the decision to submit manuscripts for
publication.
Conflict of interest: None declared.
What are the main strengths
and weaknesses?
The strengths of the study are the randomized design, References
the large sample size, the storing of abundant bio- 1
Mente A, Koning L, Shannon HS, Anand SS. A systematic
logical samples, the objective assessment of compli- review of the evidence supporting a causal link between
ance with biomarkers and the close monitoring of dietary factors and coronary heart disease. Arch Intern Med
participants. The main weakness is the difficulty to 2009;169:659–69.
384 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

2
Sofi F, Abbate R, Gensini GF, Casini A. Accruing evidence disease: multiple potential mechanisms. J Nutr 2008;
on benefits of adherence to the Mediterranean diet on 138:1746S–51S.
health: an updated systematic review and meta-analysis. 17
McManus K, Antinoro L, Sacks F. A randomized con-
Am J Clin Nutr 2010;92:1189–96. trolled trial of a moderate-fat, low-energy diet compared
3
Fung TT, Rexrode KM, Mantzoros CS, Manson JE, with a low fat, low-energy diet for weight loss in over-
Willett WC, Hu FB. Mediterranean diet and incidence weight adults. Int J Obes Relat Metab Disord 2001;25:
of and mortality from coronary heart disease and stroke 1503–11.
in women. Circulation 2009;119:1093–100. 18
Shai I, Schwarzfuchs D, Henkin Y et al. Weight loss with
4
Buckland G, Gonzalez CA, Vilardell M et al. Adherence to low-carbohydrate, Mediterranean and low-fat diet. N Engl
the Mediterranean diet and risk of coronary heart disease J Med 2008;359:229–41.
in the Spanish EPIC cohort study. Am J Epidemiol 2009; 19
Manson JE, Hsia J, Johnson KC et al. Estrogen plus pro-
170:1518–29. gestin and the risk of coronary heart disease. N Engl J
5
Martı́nez-González MA, Garcı́a-López M, Bes-Rastrollo M Med 2003;349:523–34.
et al. Mediterranean diet and the incidence of cardiovas- 20
Bjelakovic G, Nikolova D, Gluud LL, Simonetti RG,
cular disease: A Spanish cohort. Nutr Metabol Cardiovasc Gludd C. Mortality in randomized trials of antioxidant
Dis 2010; doi:10.1016/j.numecd.2009.10.005 [Epub 20 supplements for primary and secondary prevention.
January 2010]. Systematic review and meta-analysis. JAMA 2007;297:
6
Benetou V, Trichopoulou A, Orfanos P et al. Conformity to 842–57.
traditional Mediterranean diet and cancer incidence: the 21
Nigg CR, Burbank PM, Padula C et al. Stages of change
Greek EPIC cohort. Br J Cancer 2008;99:191–5. across ten health risk behaviors for older adults.
7
Buckland G, Agudo A, Luján L et al. Adherence to a Gerontologist 1999;39:473–82.
Mediterranean diet and risk of gastric adenocarcinoma 22
ALLHAT Officers and Coordinators for the ALLHAT
within the European Prospective Investigation into
Collaborative Research Group. The Antihypertensive and
Cancer and Nutrition (EPIC) cohort study. Am J Clin
Lipid-Lowering Treatment Prevent Heart Attack Trial.
Nutr 2009;91:381–90.
8 JAMA 2002;288:2998–3007.
Scarmeas N, Stern Y, Tang MX, Mayeux R, 23
Altman DG, Schulz KF, Moher D et al. The revised
Luchsinger JA. Mediterranean diet and risk for
CONSORT statement for reporting randomized trials
Alzheimer’s disease. Ann Neurol 2006;59:912–21.
9 explanation and elaboration. Ann Intern Med 2001;134:
Féart C, Samieri C, Rondeau V et al. Adherence to 663–94.
a Mediterranean diet, cognitive decline, and risk of 24
dementia. JAMA 2009;302:638–48, Erratum in: JAMA Owen RW, Mier W, Giacosa A, Hull WE, Spiegelhalder B,
2009;302:2436. Bartsch H. Phenolic compounds and squalene in olive
10 oils: the concentration and antioxidant potential of total
Sánchez-Villegas A, Delgado-Rodrı́guez M, Alonso A et al. phenols, simple phenols, secoiridoids, lignans and squa-
Association of the Mediterranean dietary pattern with the lene. Food Chem Toxicol 2000;38:647–59.
incidence of depression: the Seguimiento Universidad de 25
Navarra/University of Navarra follow-up (SUN) cohort. Covas MI, Nyyssönen K, Poulsen HE et al. The effect of
Arch Gen Psychiatry 2009;66:1090–98. polyphenols in olive oil on heart disease risk factors: a
11 randomized trial. Ann Intern Med 2006;145:333–41.
de Lorgeril M, Salen P, Martin JL, Monjaud I, Delaye J, 26
Mamelle N. Mediterranean diet, traditional risk factors, Estruch R, Martı́nez-González MA, Corella D et al.
and the rate of cardiovascular complications after myo- Effects of a Mediterranean-style diet on cardiovascular
cardial infarction: final report of the Lyon Diet Heart risk factors. A randomized trial. Ann Intern Med 2006;
Study. Circulation 1999;99:779–85. 145:1–11.
27
12
Martinez-Gonzalez MA, Sanchez-Villegas A. The emer- Martinez-Gonzalez MA, Fernandez-Jarne E, Serrano-
ging role of Mediterranean diets in cardiovascular epi- Martinez M, Wright M, Gomez-Gracia E. Development
demiology: monounsaturated fats, olive oil, red wine or of a short dietary intake questionnaire for the quantita-
the whole pattern? Eur J Epidemiol 2004;19:9–13. tive estimation of adherence to a cardioprotective
13
Gabriel R, Alonso M, Segura A et al. Prevalence, geo- Mediterranean diet. Eur J Clin Nutr 2004;58:1550–52.
28
graphic distribution, and geographic variability of major Zazpe I, Sanchez-Tainta A, Estruch R et al. A large ran-
cardiovascular risk factors in Spain. Pooled analysis of domized individual and group intervention conducted by
data from population-based epidemiological studies: the dieticians increased the adherence to Mediterranean-type
ERICE Study. Rev Esp Cardiol 2008;61:1030–40. diets: The PREDIMED study. J Am Diet Assoc 2008;108:
14
López-Miranda J, Pérez-Jiménez F, Ros E et al. Olive oil 1134–44.
29
and health: Summary of the II International Conference Sacks FM, Bray GA, Carey VJ et al. Comparison of
on Olive Oil and Health consensus report, Jaén and weight-loss diets with different compositions of fat, pro-
Córdoba (Spain) (2008). Nutr Metab Cardiovasc Dis 2010; tein, and carbohydrates. N Engl J Med 2009;360:859–73.
30
20:284–94. Bes-Rastrollo M, Sánchez-Villegas A, de la Fuente C, de
15
Fernández-Jarne E, Martı́nez-Losa E, Prado-Santamarı́a M, Irala J, Martinez JA, Martı́nez-González MA. Olive oil
Brugarolas-Brufau C, Serrano-Martı́nez M, Martı́nez- consumption and weight change: the SUN prospective
González MA. Risk of first non-fatal myocardial infarction cohort study. Lipids 2006;41:249–56.
negatively associated with olive oil consumption: a 31
Bes-Rastrollo M, Wedick NM, Martinez-Gonzalez MA,
case-control study in Spain. Int J Epidemiol 2002;31:474–80. Li TY, Sampson L, Hu FB. Prospective study of nut con-
16
Kris-Etherton PM, Hu F, Ros E, Sabaté J. The role of tree sumption, long-term weight change, and obesity risk in
nuts and peanuts in the prevention of coronary heart women. Am J Clin Nutr 2009;89:1913–19.
 DESIGN AND METHODS OF THE PREDIMED STUDY 385

32 42
Martin-Moreno JM, Boyle P, Gorgojo L et al. Development Mena MP, Sacanella E, Vazquez-Agell M et al. Inhibition
and validation of a food frequency questionnaire in of circulating immune cell activation: a molecular
Spain. Int J Epidemiol 1993;22:512–19. anti-inflammatory effect of the Mediterranean diet. Am
33 J Clin Nutr 2009;89:248–56.
Fernández-Ballart JD, Piñol JL, Zazpe I et al. Relative val-
43
idity of a semi-quantitative food-frequency questionnaire Salas-Salvadó J, Fernández-Ballart J, Ros E et al. Effect of
in an elderly Mediterranean population of Spain. Br J a Mediterranean diet supplemented with nuts on meta-
Nutr 2009;103:1808–16. bolic syndrome status. Arch Intern Med 2008;168:2449–58.
34 44
de la Fuente-Arrillaga C, Vázquez Z, Bes-Rastrollo M, Guxens M, Fitó M, Martı́nez-González MA et al.
Sampson L, Martı́nez-González MA. Reproducibility of a Hypertensive status and lipoprotein oxidation in an eld-
Food Frequency Questionnaire (FFQ) validated in Spain. erly population at high cardiovascular risk. Am J Hypertens
Public Health Nutr 2009;13:1364–72. 2009;22:68–73.
35
Elosua R, Marrugat J, Molina L, Pons S, Pujol E. 45
Toledo E, Delgado-Rodrı́guez M, Estruch R et al. Low-fat
Validation of the Minnesota Leisure Time Physical dairy products and blood pressure: follow-up of 2290
Activity Questionnaire in Spanish men. The older persons at high cardiovascular risk participating in
MARATHOM Investigators. Am J Epidemiol 1994;139: the PREDIMED study. Br J Nutr 2009;101:59–67.
1197–209. 46
36
Barceló F, Perona JS, Prades J et al. Mediterranean-style
Elosua R, Garcia M, Aguilar A, Molina L, Covas MI, diet effect on the structural properties of the erythrocyte
Marrugat J. Validation of the Minnesota Leisure Time cell membrane of hypertensive patients: the Prevencion
Physical Activity Questionnaire In Spanish Women. con Dieta Mediterranea Study. Hypertension 2009;54:
Investigators of the MARATDOM Group. Med Sci Sports 1143–50.
Exerc 2000;32:1431–37. 47
37 Razquin C, Martinez JA, Martinez-Gonzalez MA,
Buil-Cosiales P, Irimia P, Berrade N et al. Carotid Bes-Rastrollo M, Fernández-Crehuet J, Marti A. A
intima-media thickness is inversely associated with olive 3-year intervention with a Mediterranean diet modified
oil consumption. Atherosclerosis 2008;196:742–48. the association between the rs9939609 gene variant in
38
Buil-Cosiales P, Irimia P, Ros E et al. Dietary fibre intake FTO and body weight changes. Int J Obes 2010;34:266–72.
is inversely associated with carotid intima-media thick- 48
Razquin C, Martinez JA, Martinez-Gonzalez MA,
ness: a cross-sectional assessment in the PREDIMED
Mitjavila MT, Estruch R, Marti A. A 3 years follow-up
study. Eur J Clin Nutr 2009;63:1213–19.
39 of a Mediterranean diet rich in virgin olive oil is asso-
Salas-Salvado J, Garcia-Arellano A, Estruch R et al. ciated with high plasma antioxidant capacity and reduced
Components of the Mediterranean-type food pattern body weight gain. Eur J Clin Nutr 2009;63:1387–93.
and serum inflammatory markers among patients at 49
high risk for cardiovascular disease. Eur J Clin Nutr Razquin C, Martinez JA, Martinez-Gonzalez MA,
2008;62:651–59. Corella D, Santos JM, Marti A. The Mediterranean diet
40 protects against waist circumference enlargement in
Sanchez-Tainta A, Estruch R, Bullo M et al. Adherence to 12Ala carriers for the PPARgamma gene: 2 years’
a Mediterranean-type diet and reduced prevalence of
follow-up of 774 subjects at high cardiovascular risk.
clustered cardiovascular risk factors in a cohort of 3204
Br J Nutr 2009;102:672–79.
high-risk patients. Eur J Cardiov Prev Rehab 2008;15: 50
589–93. Salas-Salvadó J, Bulló M, Babio N et al. Reduction in the
41 Incidence of Type 2-Diabetes with the Mediterranean
Fitó M, Guxens M, Corella D et al. Effect of a traditional
Mediterranean diet on lipoprotein oxidation: a Diet: Results of the PREDIMED-Reus Nutrition
randomized controlled trial. Arch Intern Med 2007;167: Intervention Randomized Trial. Diabetes Care 2010;
1195–203. doi:10.2337/dc10-1288 [Epub 13 October 2010].