|

Primary prevention of cardiovascular disease with a mediterranean diet

The new england journal of medicine established in 1812 Primary Prevention of Cardiovascular Disease with a Mediterranean Diet Ramón Estruch, M.D., Ph.D., Emilio Ros, M.D., Ph.D., Jordi Salas-Salvadó, M.D., Ph.D., Maria-Isabel Covas, D.Pharm., Ph.D., Dolores Corella, D.Pharm., Ph.D., Fernando Arós, M.D., Ph.D., Enrique Gómez-Gracia, M.D., Ph.D., Valentina Ruiz-Gutiérrez, Ph.D., Miquel Fiol, M.D., Ph.D., José Lapetra, M.D., Ph.D., Rosa Maria Lamuela-Raventos, D.Pharm., Ph.D., Lluís Serra-Majem, M.D., Ph.D., Xavier Pintó, M.D., Ph.D., Josep Basora, M.D., Ph.D., Miguel Angel Muñoz, M.D., Ph.D., José V. Sorlí, M.D., Ph.D., José Alfredo Martínez, D.Pharm, M.D., Ph.D., and Miguel Angel Martínez-González, M.D., Ph.D., for the PREDIMED Study Investigators* Background
Observational cohort studies and a secondary prevention trial have shown an in- The authors' affiliations are listed in the
Appendix. Address reprint requests to verse association between adherence to the Mediterranean diet and cardiovascular Dr. Estruch at the Department of Internal risk. We conducted a randomized trial of this diet pattern for the primary preven- Medicine, Hospital Clinic, Villarroel 170, tion of cardiovascular events.
08036 Barcelona, Spain, or at [email protected], or to Dr. Martínez-González at the Department of Preventive Medi-cine and Public Health, Facultad de Me- In a multicenter trial in Spain, we randomly assigned participants who were at high dicina–Clínica Universidad de Navarra, cardiovascular risk, but with no cardiovascular disease at enrollment, to one of Irunlarrea 1, 31008 Pamplona, Spain, or three diets: a Mediterranean diet supplemented with extra-virgin olive oil, a Medi- at [email protected].
terranean diet supplemented with mixed nuts, or a control diet (advice to reduce * The PREDIMED (Prevención con Dieta dietary fat). Participants received quarterly individual and group educational ses- Mediterránea) study investigators are sions and, depending on group assignment, free provision of extra-virgin olive oil, listed in the Supplementary Appendix, available at NEJM.org.
mixed nuts, or small nonfood gifts. The primary end point was the rate of major cardiovascular events (myocardial infarction, stroke, or death from cardiovascular Drs. Estruch and Martínez-González con- causes). On the basis of the results of an interim analysis, the trial was stopped tributed equally to this article.
after a median follow-up of 4.8 years.
This article was published on February 25, 2013, at NEJM.org.
Results
A total of 7447 persons were enrolled (age range, 55 to 80 years); 57% were women. N Engl J Med 2013;368:1279-90.
The two Mediterranean-diet groups had good adherence to the intervention, ac- Copyright 2013 Massachusetts Medical Society. cording to self-reported intake and biomarker analyses. A primary end-point event occurred in 288 participants. The multivariable-adjusted hazard ratios were 0.70 (95% confidence interval [CI], 0.54 to 0.92) and 0.72 (95% CI, 0.54 to 0.96) for the group assigned to a Mediterranean diet with extra-virgin olive oil (96 events) and the group assigned to a Mediterranean diet with nuts (83 events), respectively, ver- sus the control group (109 events). No diet-related adverse effects were reported.
Conclusions
Among persons at high cardiovascular risk, a Mediterranean diet supplemented
with extra-virgin olive oil or nuts reduced the incidence of major cardiovascular events. (Funded by the Spanish government's Instituto de Salud Carlos III and oth- ers; Controlled-Trials.com number, ISRCTN35739639.) n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. The traditional Mediterranean diet cholesterol levels, low high-density lipoprotein is characterized by a high intake of olive cholesterol levels, overweight or obesity, or a oil, fruit, nuts, vegetables, and cereals; a family history of premature coronary heart dis- moderate intake of fish and poultry; a low intake ease. Detailed enrollment criteria are provided of dairy products, red meat, processed meats, in the Supplementary Appendix, available at NEJM and sweets; and wine in moderation, consumed .org. All participants provided written informed with meals.1 In observational cohort studies2,3 consent.
and a secondary prevention trial (the Lyon Diet Beginning on October 1, 2003, participants Heart Study),4 increasing adherence to the Medi- were randomly assigned, in a 1:1:1 ratio, to one terranean diet has been consistently beneficial of three dietary intervention groups: a Mediter- with respect to cardiovascular risk.2-4 A system- ranean diet supplemented with extra-virgin olive atic review ranked the Mediterranean diet as the oil, a Mediterranean diet supplemented with most likely dietary model to provide protection nuts, or a control diet. Randomization was per- against coronary heart disease.5 Small clinical formed centrally by means of a computer-gener- trials have uncovered plausible biologic mecha- ated random-number sequence.
nisms to explain the salutary effects of this food pattern.6-9 We designed a randomized trial to Interventions and Measurements
test the efficacy of two Mediterranean diets (one The dietary intervention8,10-13 is detailed in the supplemented with extra-virgin olive oil and an- Supplementary Appendix. The specific recom- other with nuts), as compared with a control diet mended diets are summarized in Table 1. Par- (advice on a low-fat diet), on primary cardiovas- ticipants in the two Mediterranean-diet groups cular prevention.
received either extra-virgin olive oil (approxi- mately 1 liter per week) or 30 g of mixed nuts per day (15 g of walnuts, 7.5 g of hazelnuts, and 7.5 g of almonds) at no cost, and those in the Study design
control group received small nonfood gifts. No The PREDIMED trial (Prevención con Dieta Med- total calorie restriction was advised, nor was iterránea) was a parallel-group, multicenter, ran- physical activity promoted.
domized trial. Details of the trial design are pro- For participants in the two Mediterranean- vided elsewhere.10-12 The trial was designed and diet groups, dietitians ran individual and group conducted by the authors, and the protocol was dietary-training sessions at the baseline visit and approved by the institutional review boards at all quarterly thereafter. In each session, a 14-item study locations. The authors vouch for the accu- dietary screener was used to assess adherence to racy and completeness of the data and all analy- the Mediterranean diet8,14 (Table S1 in the Sup- ses and for the fidelity of this report to the pro- plementary Appendix) so that personalized ad- tocol, which is available with the full text of this vice could be provided to the study participants article at NEJM.org.
in these groups.
Supplemental foods were donated, including Participants in the control group also re- extra-virgin olive oil (by Hojiblanca and Patrimo- ceived dietary training at the baseline visit and nio Comunal Olivarero, both in Spain), walnuts completed the 14-item dietary screener used to (by the California Walnut Commission), al- assess baseline adherence to the Mediterranean monds (by Borges, in Spain), and hazelnuts (by diet. Thereafter, during the first 3 years of the La Morella Nuts, in Spain). None of the sponsors trial, they received a leaflet explaining the low- had any role in the trial design, data analysis, or fat diet (Table S2 in the Supplementary Appen- reporting of the results.
dix) on a yearly basis. However, the realization that the more infrequent visit schedule and less Participant Selection and Randomization
intense support for the control group might be Eligible participants were men (55 to 80 years of limitations of the trial prompted us to amend age) and women (60 to 80 years of age) with no the protocol in October 2006. Thereafter, par- cardiovascular disease at enrollment, who had ticipants assigned to the control diet received either type 2 diabetes mellitus or at least three personalized advice and were invited to group of the following major risk factors: smoking, sessions with the same frequency and intensity hypertension, elevated low-density lipoprotein as those in the Mediterranean-diet groups, with n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. Mediterranean Diet and Cardiovascular Events the use of a separate 9-item dietary screener Table 1. Summary of Dietary Recommendations to Participants in the
(Table S3 in the Supplementary Appendix).
Mediterranean-Diet Groups and the Control-Diet Group.
A general medical questionnaire, a 137-item validated food-frequency questionnaire,15 and the Minnesota Leisure-Time Physical Activity Questionnaire were administered on a yearly basis.10 Information from the food-frequency questionnaire was used to calculate intake of Tree nuts and peanuts† energy and nutrients. Weight, height, and waist ≥3 servings/day circumference were directly measured.16 Bio- markers of compliance, including urinary hy- ≥2 servings/day droxytyrosol levels (to confirm compliance in Fish (especially fatty fish), seafood the group receiving extra-virgin olive oil) and plasma alpha-linolenic acid levels (to confirm compliance in the group receiving mixed nuts), Instead of red meat were measured in random subsamples of par- Wine with meals (optionally, only for habitual ticipants at 1, 3, and 5 years (see the Supplemen- tary Appendix).
End Points
Commercial bakery goods, sweets, and pastries§ <3 servings/wk The primary end point was a composite of myo- <1 serving/day cardial infarction, stroke, and death from cardio- Red and processed meats <1 serving/day vascular causes. Secondary end points were stroke, myocardial infarction, death from cardio- Low-fat diet (control)
vascular causes, and death from any cause. We used four sources of information to identify end Low-fat dairy products ≥3 servings/day points: repeated contacts with participants, con- Bread, potatoes, pasta, rice ≥3 servings/day tacts with family physicians, a yearly review of ≥3 servings/day medical records, and consultation of the Nation- al Death Index. All medical records related to Lean fish and seafood end points were examined by the end-point adju- dication committee, whose members were un- Vegetable oils (including olive oil) aware of the study-group assignments. Only end points that were confirmed by the adjudication Commercial bakery goods, sweets, and pastries§ committee and that occurred between October 1, Nuts and fried snacks 2003, and December 1, 2010, were included in Red and processed fatty meats the analyses. The criteria for adjudicating pri- Visible fat in meats and soups¶ mary and secondary end points are detailed in Fatty fish, seafood canned in oil the Supplementary Appendix.
Statistical Analysis
We initially estimated that a sample of 9000 par- * The amount of olive oil includes oil used for cooking and salads and oil con-
ticipants would be required to provide statistical sumed in meals eaten outside the home. In the group assigned to the Medi- terranean diet with extra-virgin olive oil, the goal was to consume 50 g (ap- power of 80% to detect a relative risk reduction proximately 4 tbsp) or more per day of the polyphenol-rich olive oil supplied, of 20% in each Mediterranean-diet group versus instead of the ordinary refined variety, which is low in polyphenols.
the control-diet group during a 4-year follow-up † For participants assigned to the Mediterranean diet with nuts, the recommend- ed consumption was one daily serving (30 g, composed of 15 g of walnuts, period, assuming an event rate of 12% in the 7.5 g of almonds, and 7.5 g of hazelnuts).
control group.10,17 In April 2008, on the advice of ‡ Sofrito is a sauce made with tomato and onion, often including garlic and aro- the data and safety monitoring board and on the matic herbs, and slowly simmered with olive oil.
§ Commercial bakery goods, sweets, and pastries (not homemade) included basis of lower-than-expected rates of end-point cakes, cookies, biscuits, and custard.
events, the sample size was recalculated as 7400 ¶ Participants were advised to remove the visible fat (or the skin) of chicken, participants, with the assumption of a 6-year duck, pork, lamb, or veal before cooking and the fat of soups, broths, and cooked meat dishes before consumption.
follow-up period and underlying event rates of n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. 8.8% and 6.6% in the control and intervention with nuts) to obtain two hazard ratios for the groups, respectively. Power curves under several comparison with the control group. To account assumptions can be found in Figure S1 in the for small imbalances in risk factors at baseline among the groups, Cox regression models were Yearly interim analyses began after a median used to adjust for sex, age, and baseline risk fac- of 2 years of follow-up. With the use of O'Brien– tors. We tested the proportionality of hazards Fleming stopping boundaries, the P values for with the use of time-varying covariates. All stopping the trial at each yearly interim analysis analyses were stratified according to center. Pre- were 5×10−6, 0.001, 0.009, and 0.02 for benefit specified subgroup analyses were conducted ac- and 9×10−5, 0.005, 0.02, and 0.05 for adverse ef- cording to sex, age, body-mass index (BMI), fects.18 The stopping boundary for the benefit of cardiovascular-risk-factor status, and baseline the Mediterranean diets with respect to the pri- adherence to the Mediterranean diet. Sensitivity mary end point was crossed at the fourth inter- analyses were conducted under several assump- im evaluation; on July 22, 2011, the data and tions, including imputation of data for missing safety monitoring board recommended stopping values and participants who dropped out (see the the trial on the basis of end points documented Supplementary Appendix).
through December 1, 2010.
All primary analyses were performed on an intention-to-treat basis by two independent ana- lysts. Time-to-event data were analyzed with the Baseline Characteristics of the Study
use of Cox models with two dummy variables Participants
(one for the Mediterranean diet with extra-virgin From October 2003 through June 2009, a total of olive oil and another for the Mediterranean diet 8713 candidates were screened for eligibility, and Table 2. Baseline Characteristics of the Participants According to Study Group.*
Diet with EVOO
Diet with Nuts
Control Diet
(N = 2543)
(N = 2454)
(N = 2450)
Female sex — no. (%)† Race or ethnic group — no. (%) White, from Europe Hispanic, from Central or South America Smoking status — no. (%) Body-mass index†‡ <25 — no. (%) 25–30 — no. (%) >30 — no. (%) Waist circumference — cm Hypertension — no. (%)¶ Type 2 diabetes — no. (%)†‖ Dyslipidemia — no. (%)** Family history of premature CHD — no. (%)†† n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. Mediterranean Diet and Cardiovascular Events Table 2. (Continued.)
Diet with EVOO
Diet with Nuts
Control Diet
(N = 2543)
(N = 2454)
(N = 2450)
Medication use — no. (%) Other antihypertensive agents Other lipid-lowering agents Oral hypoglycemic agents† Antiplatelet therapy Score for adherence to Med diet§§ * Plus–minus values are means ±SD. ACE denotes angiotensin-converting enzyme, and EVOO extra-virgin olive oil.
† P<0.05 for comparisons between groups.
‡ The body-mass index is the weight in kilograms divided by the square of the height in meters.
§ The waist-to-height ratio (an index of central obesity) is the waist circumference divided by height.
¶ Hypertension was defined as a systolic blood pressure of 140 mm Hg or higher, a diastolic blood pressure of 90 mm Hg or higher, or the use of antihypertensive therapy.
‖ Diabetes was defined as a fasting blood glucose level of 126 mg per deciliter (7.0 mmol per liter) or higher on two occasions, a 2-hour plasma glucose level of 200 mg per deciliter (11 mmol per liter) or higher during a 75-g oral glu- cose-tolerance test, or the use of antidiabetic medication.
** Dyslipidemia was defined as a low-density lipoprotein cholesterol level higher than 160 mg per deciliter (4.1 mmol per liter), a high-density lipoprotein cholesterol level of 40 mg per deciliter (1.0 mmol per liter) or lower in men or 50 mg per deciliter (1.3 mmol per liter) or lower in women, or the use of lipid-lowering therapy.
†† A family history of premature coronary heart disease (CHD) was defined as a diagnosis of the disease in a male first- degree relative younger than 55 years of age or in a female first-degree relative younger than 65 years of age.
‡‡ The values for hormone-replacement therapy are for women only.
§§ The score for adherence to the Mediterranean diet is based on the 14-item dietary screener shown in Table S1 in the Supplementary Appendix (a score of 0 indicates minimum adherence, and a score of 14 indicates maximum adherence).
7447 were randomly assigned to one of the three were younger (by 1.4 years), had a higher BMI study groups (Fig. S2 in the Supplementary Ap- (the weight in kilograms divided by the square of pendix). Their baseline characteristics according the height in meters; by 0.4), a higher waist-to- to study group are shown in Table 2. Drug-treat- height ratio (by 0.01), and a lower score for ad- ment regimens were similar for participants in herence to the Mediterranean diet (by 1.0 points the three groups, and they continued to be bal- on the 14-item dietary screener) (P<0.05 for all anced during the follow-up period (Table S4 in comparisons).
the Supplementary Appendix).
Participants were followed for a median of Compliance with the Dietary Intervention
4.8 years (interquartile range, 2.8 to 5.8). After Participants in the three groups reported similar the initial assessment, 209 participants (2.8%) adherence to the Mediterranean diet at baseline chose not to attend subsequent visits, and their (Table 2, and Fig. S3 in the Supplementary Ap- follow-up was based on reviews of medical re- pendix) and similar food and nutrient intakes. cords. By December 2010, a total of 523 partici- During follow-up, scores on the 14-item Medi- pants (7.0%) had been lost to follow-up for 2 or terranean-diet screener increased for the par- more years. Dropout rates were higher in the con- ticipants in the two Mediterranean-diet groups trol group (11.3%) than in the Mediterranean- (Fig. S3 in the Supplementary Appendix). There diet groups (4.9%) (Fig. S2 in the Supplementary were significant differences between these groups Appendix). As compared with participants who and the control group in 12 of the 14 items at remained in the trial, those who dropped out 3 years (Table S5 in the Supplementary Appen- n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. dix). Changes in objective biomarkers also indi- nificantly increased their consumption of extra- cated good compliance with the dietary assign- virgin olive oil (to 50 and 32 g per day, respec- ments (Fig. S4 and S5 in the Supplementary tively) and nuts (to 0.9 and 6 servings per week, respectively). The main nutrient changes in the Participants in the two Mediterranean-diet Mediterranean-diet groups reflected the fat con- groups significantly increased weekly servings of tent and composition of the supplemental foods fish (by 0.3 servings) and legumes (by 0.4 serv- (Tables S7 and S8 in the Supplementary Appen- ings) in comparison with those in the control dix). No relevant diet-related adverse effects group (Table S6 in the Supplementary Appendix). were reported (see the Supplementary Appen- In addition, participants assigned to a Mediter- dix). We did not find any significant difference ranean diet with extra-virgin olive oil and those in changes in physical activity among the three assigned to a Mediterranean diet with nuts sig- groups.
Table 3. Outcomes According to Study Group.*
Diet with EVOO
Diet with Nuts
Control Diet
End Point
(N = 2543)
(N = 2454)
(N = 2450)
P Value†
Mediterranean Mediterranean Diet with EVOO Diet with Nuts vs. Control Diet vs. Control Diet Person-yr of follow-up Primary end point‡ Crude rate/1000 person-yr (95% CI) 11.2 (9.2–13.5) Secondary end points Crude rate/1000 person-yr (95% CI) Myocardial infarction Crude rate/1000 person-yr (95% CI) Death from cardiovascular causes Crude rate/1000 person-yr (95% CI) Death from any cause Crude rate/1000 person-yr (95% CI) 10.0 (8.2–11.9) 11.2 (9.3–13.4) 11.7 (9.6–14.0) Hazard ratio for each Mediterranean diet vs. control (95% CI) Primary end point 0.70 (0.53–0.91) 0.70 (0.53–0.94) 0.69 (0.53–0.91) 0.72 (0.54–0.97) 0.70 (0.54–0.92) 0.72 (0.54–0.96) Secondary end points‖ 0.67 (0.46–0.98) 0.54 (0.35–0.84) Myocardial infarction 0.80 (0.51–1.26) 0.74 (0.46–1.19) Death from cardiovascular causes 0.69 (0.41–1.16) 1.01 (0.61–1.66) Death from any cause 0.82 (0.64–1.07) 0.97 (0.74–1.26) n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. Mediterranean Diet and Cardiovascular Events Table 3. (Continued.)
Diet with EVOO
Diet with Nuts
Control Diet
End Point
(N = 2543)
(N = 2454)
(N = 2450)
P Value†
Mediterranean Mediterranean Diet with EVOO Diet with Nuts vs. Control Diet vs. Control Diet Hazard ratio for Mediterranean diets combined vs. control (95% CI) Primary end point 0.70 (0.55–0.89) 0.71 (0.56–0.90) 0.71 (0.56–0.90) Secondary end points‖ 0.61 (0.44–0.86) Myocardial infarction 0.77 (0.52–1.15) Death from cardiovascular causes 0.83 (0.54–1.29) Death from any cause 0.89 (0.71–1.12) * CI denotes confidence interval, and ref reference.
† All P values were calculated with the use of Cox proportional-hazards models with robust variance estimators and stratification according to recruiting center.
‡ The primary end point was a composite of myocardial infarction, stroke, and death from cardiovascular causes.
§ The primary end point was stratified according to recruiting center and adjusted for sex, age (continuous variable), family history of pre- mature coronary heart disease (yes or no), and smoking status (never smoked, former smoker, or current smoker).
¶ The primary end point was additionally adjusted for body-mass index (continuous variable), waist-to-height ratio (continuous variable), hypertension at baseline (yes or no), dyslipidemia at baseline (yes or no), and diabetes at baseline (yes or no).
‖ The secondary end points were stratified according to recruiting center and adjusted for sex, age (continuous variable), family history of premature coronary heart disease (yes or no), smoking status (never smoked, former smoker, or current smoker), body-mass index (con- tinuous variable), waist-to-height ratio (continuous variable), hypertension at baseline (yes or no), dyslipidemia at baseline (yes or no), and diabetes at baseline (yes or no).
End Points
the primary end point (Table 3). Regarding com- The median follow-up period was 4.8 years. A ponents of the primary end point, only the com- total of 288 primary-outcome events occurred: parisons of stroke risk reached statistical signifi- 96 in the group assigned to a Mediterranean diet cance (Table 3, and Fig. S6 in the Supplementary with extra-virgin olive oil (3.8%), 83 in the group Appendix). The Kaplan–Meier curves for the assigned to a Mediterranean diet with nuts primary end point diverged soon after the trial (3.4%), and 109 in the control group (4.4%). Tak- started, but no effect on all-cause mortality was ing into account the small differences in the ac- apparent (Fig. 1). The results of several sensitiv- crual of person-years among the three groups, ity analyses were also consistent with the find- the respective rates of the primary end point ings of the primary analysis (Table S9 in the were 8.1, 8.0, and 11.2 per 1000 person-years Supplementary Appendix).
(Table 3). The unadjusted hazard ratios were 0.70 (95% confidence interval [CI], 0.53 to 0.91) for a Subgroup Analyses
Mediterranean diet with extra-virgin olive oil and Reductions in disease risk in the two Mediterra- 0.70 (95% CI, 0.53 to 0.94) for a Mediterranean nean-diet groups as compared with the control diet with nuts (Fig. 1) as compared with the con- group were similar across the prespecified sub- trol diet (P = 0.015, by the likelihood ratio test, groups (Fig. 2, and Table S10 in the Supplemen- for the overall effect of the intervention).
tary Appendix). In addition, to account for the The results of multivariate analyses showed a protocol change in October 2006 whereby the similar protective effect of the two Mediterra- intensity of dietary intervention in the control nean diets versus the control diet with respect to group was increased, we compared hazard ratios n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. A Primary End Point (acute myocardial infarction, stroke, or death from cardiovascular causes)
Med diet, EVOO: hazard ratio, 0.70 (95% CI, 0.53–0.91); P=0.009 Med diet, nuts: hazard ratio, 0.70 (95% CI, 0.53–0.94); P=0.02 End Point
Incidence of Composite Cardiovascular
No. at Risk
Control diet
B Total Mortality
Med diet, EVOO: hazard ratio, 0.81 (95% CI, 0.63–1.05); P=0.11 Med diet, nuts: hazard ratio, 0.95 (95% CI, 0.73–1.23); P=0.68 No. at Risk
Control diet
Figure 1. Kaplan–Meier Estimates of the Incidence of Outcome Events in the Total Study Population.
Panel A shows the incidence of the primary end point (a composite of acute myocardial infarction, stroke, and death
from cardiovascular causes), and Panel B shows total mortality. Hazard ratios were stratified according to center
(Cox model with robust variance estimators). CI denotes confidence interval, EVOO extra-virgin olive oil, and Med
Mediterranean.
for the Mediterranean-diet groups (both groups merged vs. the control group) before and after this date. Adjusted hazard ratios were 0.77 (95% In this trial, an energy-unrestricted Mediterra- CI, 0.59 to 1.00) for participants recruited before nean diet supplemented with either extra-virgin October 2006 and 0.49 (95% CI, 0.26 to 0.92) for olive oil or nuts resulted in an absolute risk re- those recruited thereafter (P = 0.21 for interaction). duction of approximately 3 major cardiovascular n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. Mediterranean Diet and Cardiovascular Events P Value for
Hazard Ratio (95% CI)
no. of participants with primary end-point event/total no. of participants 0.69 (0.51–0.94) 0.73 (0.50–1.07) 0.73 (0.52–1.05) 0.71 (0.51–0.98) 0.67 (0.45–1.01) 0.71 (0.53–0.96) 1.25 (0.64–2.45) 0.65 (0.50–0.84) 0.95 (0.64–1.42) 0.60 (0.44–0.80) 0.67 (0.47–0.94) 0.75 (0.54–1.03) Family history of premature CHD 0.72 (0.55–0.94) 0.75 (0.43–1.29) 0.69 (0.29–1.67) 1.04 (0.71–1.54) 0.51 (0.37–0.71) 0.76 (0.53–1.08) 0.67 (0.48–0.93) Waist-to-height ratio 0.74 (0.52–1.06) 0.68 (0.50–0.94) Baseline score for adherence to Mediterranean diet 0.81 (0.58–1.12) 0.64 (0.45–0.92) End-point components 0.61 (0.44–0.86) Myocardial infarction 0.77 (0.52–1.15) Death from cardiovascular causes 0.83 (0.54–1.29) Mediterranean Diets Better Control Diet Better
Figure 2. Results of Subgroup Analyses.
Shown are adjusted hazard ratios for the primary end point within specific subgroups. Squares denote hazard ratios; horizontal lines
represent 95% confidence intervals. Hazard ratios indicate the relative risk in both intervention groups merged together (vs. the control
group) within each stratum. Hazard ratios were stratified according to recruiting center and were adjusted for sex, age (continuous vari-
able), family history of premature coronary heart disease (CHD) (yes or no), smoking (never smoked, former smoker, or current smok-
er), body-mass index (BMI) (continuous variable), waist-to-height ratio (continuous variable), hypertension at baseline (yes or no), dys-
lipidemia at baseline (yes or no), and diabetes at baseline (yes or no). Scores for adherence to the Mediterranean diet range from 0 to
14, with higher scores indicating greater adherence.
events per 1000 person-years, for a relative risk reduction. They are particularly relevant given the reduction of approximately 30%, among high- challenges of achieving and maintaining weight risk persons who were initially free of cardiovas- loss. The secondary prevention Lyon Diet Heart cular disease. These results support the benefits Study also showed a large reduction in rates of of the Mediterranean diet for cardiovascular risk coronary heart disease events with a modified n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. Mediterranean diet enriched with alpha-linolenic Our study has several limitations. First, the acid (a key constituent of walnuts). That result, protocol for the control group was changed half- however, was based on only a few major events.4,19,20 way through the trial. The lower intensity of There were small between-group differences dietary intervention for the control group during in some baseline characteristics in our trial, which the first few years might have caused a bias to- were not clinically meaningful but were statisti- ward a benefit in the two Mediterranean-diet cally significant, and we therefore adjusted for groups, since the participants in these two groups these variables. In fully adjusted analyses, we found received a more intensive intervention during significant results for the combined cardiovas- that time. However, we found no significant in- cular end point and for stroke, but not for myo- teraction between the period of trial enrollment cardial infarction alone. This could be due to (before vs. after the protocol change) and the stronger effects on specific risk factors for stroke benefit in the Mediterranean-diet groups. Sec- but also to a lower statistical power to identify ond, we had losses to follow-up, predominantly effects on myocardial infarction. Our findings in the control group, but the participants who are consistent with those of prior observational dropped out had a worse cardiovascular risk pro- studies of the cardiovascular protective effects file at baseline than those who remained in the of the Mediterranean diet,2,5 olive oil,21-23 and study, suggesting a bias toward a benefit in the nuts24,25; smaller trials assessing effects on tra- control group. Third, the generalizability of our ditional cardiovascular risk factors6-9 and novel findings is limited because all the study partici- risk factors, such as markers of oxidation, in- pants lived in a Mediterranean country and were flammation, and endothelial dysfunction6,8,26-28; at high cardiovascular risk; whether the results and studies of conditions associated with high can be generalized to persons at lower risk or to cardiovascular risk — namely, the metabolic other settings requires further research.
syndrome6,16,29 and diabetes.30-32 Thus, a causal As with many clinical trials, the observed rates role of the Mediterranean diet in cardiovascular of cardiovascular events were lower than antici- prevention has high biologic plausibility. The pated, with reduced statistical power to sepa- results of our trial might explain, in part, the rately assess components of the primary end lower cardiovascular mortality in Mediterranean point. However, favorable trends were seen for countries than in northern European countries both stroke and myocardial infarction. We ac- or the United States.33 knowledge that, even though participants in the The risk of stroke was reduced significantly control group received advice to reduce fat in- in the two Mediterranean-diet groups. This is take, changes in total fat were small and the consistent with epidemiologic studies that showed largest differences at the end of the trial were in an inverse association between the Mediterra- the distribution of fat subtypes. The interventions nean diet2,34 or olive-oil consumption22 and in- were intended to improve the overall dietary pat- cident stroke.
tern, but the major between-group differences Our results compare favorably with those of involved the supplemental items. Thus, extra- the Women's Health Initiative Dietary Modifica- virgin olive oil and nuts were probably respon- tion Trial, wherein a low-fat dietary approach sible for most of the observed benefits of the resulted in no cardiovascular benefit.35 Salient Mediterranean diets. Differences were also ob- components of the Mediterranean diet report- served for fish and legumes but not for other edly associated with better survival include mod- food groups. The small between-group differ- erate consumption of ethanol (mostly from wine), ences in the diets during the trial are probably low consumption of meat and meat products, due to the facts that for most trial participants and high consumption of vegetables, fruits, nuts, the baseline diet was similar to the trial Mediter- legumes, fish, and olive oil.36,37 Perhaps there is ranean diet and that the control group was given a synergy among the nutrient-rich foods includ- recommendations for a healthy diet, suggesting ed in the Mediterranean diet that fosters favor- a potentially greater benefit of the Mediterra- able changes in intermediate pathways of car- nean diet as compared with Western diets.
diometabolic risk, such as blood lipids, insulin In conclusion, in this primary prevention trial, sensitivity, resistance to oxidation, inflammation, we observed that an energy-unrestricted Medi- and vasoreactivity.38 terranean diet, supplemented with extra-virgin n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. Mediterranean Diet and Cardiovascular Events olive oil or nuts, resulted in a substantial reduc- Laboratories; receiving consulting fees and lecture fees, as well tion in the risk of major cardiovascular events as grant support through his institution, from Merck; receiving lecture fees from Danone, Pace, AstraZeneca, and Rottapharm; among high-risk persons. The results support receiving lecture fees and payment for the development of edu- the benefits of the Mediterranean diet for the cational presentations, as well as grant support through his in- primary prevention of cardiovascular disease.
stitution, from Ferrer; receiving payment for the development of educational presentations from Recordati; and receiving grant support through his institution from Sanofi-Aventis, Takeda, Supported by the official funding agency for biomedical re- Daiichi Sankyo, Nutrexpa, Feiraco, Unilever, and Karo Bio. Dr. search of the Spanish government, Instituto de Salud Carlos III Salas-Salvadó reports serving on the board of and receiving grant (ISCIII), through grants provided to research networks specifi- support through his institution from the International Nut and cally developed for the trial (RTIC G03/140, to Dr. Estruch; RTIC Dried Fruit Council; receiving consulting fees from Danone; RD 06/0045, to Dr. Martínez-González and through Centro de and receiving grant support through his institution from Eroski Investigación Biomédica en Red de Fisiopatología de la Obesi- and Nestlé. Dr. Arós reports receiving payment for the develop- dad y Nutrición [CIBERobn]), and by grants from Centro Nacio- ment of educational presentations from Menarini and AstraZeneca. nal de Investigaciones Cardiovasculares (CNIC 06/2007), Fondo Dr. Lamuela-Raventos reports serving on the board of and re- de Investigación Sanitaria–Fondo Europeo de Desarrollo Re- ceiving lecture fees from FIVIN; receiving lecture fees from gional (PI04-2239, PI 05/2584, CP06/00100, PI07/0240, PI07/1138, Cerveceros de España; and receiving lecture fees and travel sup- PI07/0954, PI 07/0473, PI10/01407, PI10/02658, PI11/01647, and port from PepsiCo. Dr. Serra-Majem reports serving on the P11/02505), Ministerio de Ciencia e Innovación (AGL-2009- boards of the Mediterranean Diet Foundation and the Beer and 13906-C02 and AGL2010-22319-C03), Fundación Mapfre 2010, Health Foundation. Dr. Pintó reports serving on the board of Consejería de Salud de la Junta de Andalucía (PI0105/2007), Pub- and receiving grant support through his institution from the lic Health Division of the Department of Health of the Auto- Residual Risk Reduction Initiative (R3i) Foundation; serving on nomous Government of Catalonia, Generalitat Valenciana the board of Omegafort; serving on the board of and receiving (ACOMP06109, GVACOMP2010-181, GVACOMP2011-151, CS2010- payment for the development of educational presentations, as AP-111, and CS2011-AP-042), and Regional Government of Na- well as grant support through his institution, from Ferrer; re- varra (P27/2011).
ceiving consulting fees from Abbott Laboratories; receiving lec- Dr. Estruch reports serving on the board of and receiving ture fees, as well as grant support through his institution, from lecture fees from the Research Foundation on Wine and Nutri- Merck and Roche; receiving lecture fees from Danone and Esteve; tion (FIVIN); serving on the boards of the Beer and Health receiving payment for the development of educational presenta- Foundation and the European Foundation for Alcohol Research tions from Menarini; and receiving grant support through his (ERAB); receiving lecture fees from Cerveceros de España and institution from Sanofi-Aventis, Kowa, Unilever, Boehringer Sanofi-Aventis; and receiving grant support through his institu- Ingelheim, and Karo Bio. No other potential conflict of interest tion from Novartis. Dr. Ros reports serving on the board of and relevant to this article was reported.
receiving travel support, as well as grant support through his Disclosure forms provided by the authors are available with institution, from the California Walnut Commission; serving on the full text of this article at NEJM.org.
the board of the Flora Foundation (Unilever); serving on the We thank the participants in the trial for their enthusiastic board of and receiving lecture fees from Roche; serving on the and sustained collaboration and Joan Vila from Institut Munici- board of and receiving grant support through his institution pal d'Investigació Mèdica, Barcelona, for expert assessment in from Amgen; receiving consulting fees from Damm and Abbott the statistical analyses.
The author's affiliations are as follows: Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición (R.E., E.R., J.S.-S., M.-I.C., D.C., M.F., J.L., R.M.L.-R., J.B., J.V.S., J.A.M.) and the PREDIMED (Prevención con Dieta Mediterránea) Network (RD 06/0045) (R.E., J.S.-S., F.A., E.G.-G., V.R.-G., R.M.L.-R., L.S.-M., X.P., J.B., J.V.S., J.A.M., M.A.M.-G.), Instituto de Salud Carlos III, Madrid; the Department of Internal Medicine (R.E.) and Lipid Clinic, Department of Endocrinology and Nutrition (E.R.), Institut d'Investigacions Biomèdiques August Pi I Sunyer, Hospital Clinic, University of Barcelona, Barcelona; Human Nutrition Department, Hospital Universitari Sant Joan, Institut d'Investigació Sanitaria Pere Virgili, Universitat Rovira i Virgili, Reus (J.S.-S.); Cardiovascular and Nutrition Research Group, Institut de Recerca Hospital del Mar, Barcelona (M.-I.C.); the Department of Preventive Medicine, Uni- versity of Valencia, Valencia (D.C.); the Department of Cardiology, University Hospital of Alava, Vitoria (F.A.); the Department of Preven- tive Medicine, University of Malaga, Malaga (E.G.-G.); Instituto de la Grasa, Consejo Superior de Investigaciones Cientificas, Seville (V.R.-G.); Institute of Health Sciences (IUNICS), University of Balearic Islands, and Hospital Son Espases, Palma de Mallorca (M.F.); the Department of Family Medicine, Primary Care Division of Seville, San Pablo Health Center, Seville (J.L.); the Department of Nutrition and Food Science, School of Pharmacy, Xarxa de Referència en Tecnologia dels Aliments, Instituto de Investigación en Nutrición y Se- guridad Alimentaria, University of Barcelona, Barcelona (R.M.L.-R.); the Department of Clinical Sciences, University of Las Palmas de Gran Canaria, Las Palmas (L.S.-M.); Lipids and Vascular Risk Unit, Internal Medicine, Hospital Universitario de Bellvitge, Hospitalet de Llobregat, Barcelona (X.P.); Primary Care Division, Catalan Institute of Health, Institut d'Investigació en Atenció Primària Jordi Gol, Tarragona-Reus (J.B.) and Barcelona (M.A.M.); Primary Care Division, Valencia Institute of Health, Valencia (J.V.S.); and the Depart- ments of Nutrition and Food Sciences, Physiology and Toxicology (J.A.M.) and Preventive Medicine and Public Health (M.A.M.-G.), University of Navarra, Pamplona — all in Spain. 1. Willett WC, Sacks F, Trichopoulou A, A. Accruing evidence on benefits of ad-
3. Serra-Majem L, Roman B, Estruch R.
et al. Mediterranean diet pyramid: a cul- herence to the Mediterranean diet on Scientific evidence of interventions using tural model for healthy eating. Am J Clin health: an updated systematic review and the Mediterranean diet: a systematic re- meta-analysis. Am J Clin Nutr 2010;92: view. Nutr Rev 2006;64:S27-S47.
2. Sofi F, Abbate R, Gensini GF, Casini 1189-96.
4. de Lorgeril M, Salen P, Martin JL,
n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved. Mediterranean Diet and Cardiovascular Events Monjaud I, Delaye J, Mamelle N. Mediter- Ros E, et al. Effect of a Mediterranean diet matory effect of the Mediterranean diet. ranean diet, traditional risk factors, and supplemented with nuts on metabolic Am J Clin Nutr 2009;89:248-56.
the rate of cardiovascular complications syndrome status: one-year results of the 28. Fuentes F, López-Miranda J, Sánchez
after myocardial infarction: final report PREDIMED randomized trial. Arch Intern E, et al. Mediterranean and low-fat diets of the Lyon Diet Heart Study. Circulation Med 2008;168:2449-58.
improve endothelial function in hyper- 17. ALLHAT Officers and Coordinators cholesterolemic men. Ann Intern Med
5. Mente A, de Koning L, Shannon HS, for the ALLHAT Collaborative Research 2001;134:1115-9.
Anand SS. A systematic review of the evi- Group. Major outcomes in moderately hy- 29. Kastorini CM, Milionis HJ, Esposito
dence supporting a causal link between percholesterolemic, hypertensive patients K, Giugliano D, Goudevenos JA, Panagio- dietary factors and coronary heart dis- randomized to pravastatin vs usual care: takos DB. The effect of Mediterranean ease. Arch Intern Med 2009;169:659-69.
the Antihypertensive and Lipid-Lowering diet on metabolic syndrome and its com- 6. Esposito K, Marfella R, Ciotola M, et al. Treatment to Prevent Heart Attack Trial ponents: a meta-analysis of 50 studies
Effect of a Mediterranean-style diet on en- (ALLHAT-LLT). JAMA 2002;288:2998-3007. and 534,906 individuals. J Am Coll Car- dothelial dysfunction and markers of vas- 18. O'Brien PC, Fleming TR. A multiple diol 2011;57:1299-313.
cular inflammation in the metabolic syn- testing procedure for clinical trials. Bio- 30. Esposito K, Maiorino MI, Ceriello A,
drome: a randomized trial. JAMA 2004; metrics 1979;35:549-56.
Giugliano D. Prevention and control of 19. Kris-Etherton P, Eckel RH, Howard type 2 diabetes by Mediterranean diet:
7. Vincent-Baudry S, Defoort C, Gerber BV, St Jeor S, Bazzarre TL. Lyon Diet Heart a systematic review. Diabetes Res Clin
M, et al. The Medi-RIVAGE study: reduc- Study: benefits of a Mediterranean-style, Pract 2010;89:97-102.
tion of cardiovascular disease risk factors National Cholesterol Education Program/ 31. Salas-Salvadó J, Bulló M, Babio N, et al.
after a 3-mo intervention with a Mediter- American Heart Association step I dietary Reduction in the incidence of type 2 dia- ranean-type diet or a low-fat diet. Am J pattern on cardiovascular disease. Circu- betes with the Mediterranean diet: results Clin Nutr 2005;82:964-71.
of the PREDIMED-Reus nutrition inter- 8. Estruch R, Martínez-González MA, 20. de Lorgeril M, Renaud S, Mamelle N, vention randomized trial. Diabetes Care
Corella D, et al. Effects of a Mediterra- et al. Mediterranean alpha-linolenic acid- 2011;34:14-9.
nean-style diet on cardiovascular risk fac- rich diet in secondary prevention of coro- 32. Martínez-González MA, de la Fuente-
tors: a randomized trial. Ann Intern Med nary heart disease. Lancet 1994;343:1454- Arrillaga C, Nuñez-Córdoba JM, et al. Ad- 9. [Erratum, Lancet 1995;345:738.] herence to Mediterranean diet and risk of 9. Shai I, Schwarzfuchs D, Henkin Y, et al. 21. Bendinelli B, Masala G, Saieva C, et al. developing diabetes: prospective cohort
Weight loss with a low-carbohydrate, Med- Fruit, vegetables, and olive oil and risk of study. BMJ 2008;336:1348-51.
iterranean, or low-fat diet. N Engl J Med coronary heart disease in Italian women: 33. Müller-Nordhorn J, Binting S, Roll S,
2008;359:229-41. [Erratum, N Engl J the EPICOR Study. Am J Clin Nutr 2011; Willich SN. An update on regional varia- Med 2009;361:2681.] tion in cardiovascular mortality within 10. Martínez-González MA, Corella D, 22. Samieri C, Féart C, Proust-Lima C, et al. Europe. Eur Heart J 2008;29:1316-26.
Salas-Salvadó J, et al. Cohort profile: de- Olive oil consumption, plasma oleic acid, 34. Kastorini CM, Milionis HJ, Ioannidi A,
sign and methods of the PREDIMED and stroke incidence: the Three-City Study. et al. Adherence to the Mediterranean diet study. Int J Epidemiol 2012;41:377-85.
in relation to acute coronary syndrome or 11. The PREDIMED Study (http://www 23. Buckland G, Travier N, Barricarte A, stroke nonfatal events: a comparative
et al. Olive oil intake and CHD in the Euro- analysis of a case/case-control study. Am 12. The PREDIMED network (http://www pean Prospective Investigation into Can-
Heart J 2011;162:717-24.
cer and Nutrition Spanish cohort. Br J Nutr 35. Howard BV, Van Horn L, Hsia J, et al.
13. Zazpe I, Sanchez-Tainta A, Estruch R, 2012;108:2075-82.
Low-fat dietary pattern and risk of cardio- et al. A large randomized individual and 24. Kris-Etherton PM, Hu FB, Ros E, Sa- vascular disease: the Women's Health Ini-
group intervention conducted by registered baté J. The role of tree nuts and peanuts in tiative Randomized Controlled Dietary dieticians increased the adherence to the prevention of coronary heart disease: Modification Trial. JAMA 2006;295:655-66.
Mediterranean-type diets: the PREDIMED multiple potential mechanisms. J Nutr 36. Trichopoulou A, Bamia C, Trichopou-
study. J Am Diet Assoc 2008;108:1134-44. 2008;138:1746S-1751S.
los D. Anatomy of health effects of Medi- 14. Schröder H, Fitó M, Estruch R, et al. 25. Ros E, Tapsell LC, Sabaté J. Nuts and terranean diet: Greek EPIC prospective
A short screener is valid for assessing berries for heart health. Curr Atheroscler cohort study. BMJ 2009;338:b2337.
Mediterranean diet adherence among old- Rep 2010;12:397-406.
37. Buckland G, Mayén AL, Agudo A, et al.
er Spanish men and women. J Nutr 2011; 26. Fitó M, Guxens M, Corella D, et al. Olive oil intake and mortality within the
Effect of a traditional Mediterranean diet Spanish population (EPIC-Spain). Am J 15. Fernández-Ballart JD, Piñol JL, Zazpe I, on lipoprotein oxidation: a randomized Clin Nutr 2012;96:142-9.
et al. Relative validity of a semi-quantita- controlled trial. Arch Intern Med 2007; 38. Jacobs DR Jr, Gross MD, Tapsell LC.
tive food-frequency questionnaire in an 167:1195-203.
Food synergy: an operational concept for elderly Mediterranean population of Spain. 27. Mena MP, Sacanella E, Vázquez-Agell understanding nutrition. Am J Clin Nutr
Br J Nutr 2010;103:1808-16.
M, et al. Inhibition of circulating immune 2009;89:1543S-1548S.
16. Salas-Salvadó J, Fernández-Ballart J, cell activation: a molecular antiinflam-
Copyright 2013 Massachusetts Medical Society. receive immediate notification when an article
is published online first
To be notified by e-mail when Journal articles are published Online First, sign up at NEJM.org.
n engl j med 368;14 nejm.org april 4, 2013 The New England Journal of Medicine Downloaded from nejm.org on January 27, 2014. For personal use only. No other uses without permission. Copyright 2013 Massachusetts Medical Society. All rights reserved.

Source: http://infoalimenta.es/uploads/_ciencia/id124/Estudio%20Predimed.pdf