Polyphenol intake and mortality riska re-analysis of the PREDIMED trial

The Harvard community has made thisarticle openly available Please share howthis access benefits you Your story matters

Citation Tresserra-Rimbau A E B Rimm A Medina-RemoacutenM A Martiacutenez-Gonzaacutelez M C Loacutepez-Sabater M ICovas D Corella et al 2014 ldquoPolyphenol intake andmortality risk a re-analysis of the PREDIMED trialrdquo BMCMedicine 12 (1) 77 doi1011861741-7015-12-77 httpdxdoiorg1011861741-7015-12-77

Published Version doi1011861741-7015-12-77

Citable link httpnrsharvardeduurn-3HULInstRepos12717399

Terms of Use This article was downloaded from Harvard Universityrsquos DASHrepository and is made available under the terms and conditionsapplicable to Other Posted Material as set forth at httpnrsharvardeduurn-3HULInstReposdashcurrentterms-of-useLAA

RESEARCH ARTICLE Open Access

Polyphenol intake and mortality risk a re-analysisof the PREDIMED trialAnna Tresserra-Rimbau12 Eric B Rimm3 Alexander Medina-Remoacuten217 Miguel A Martiacutenez-Gonzaacutelez24M Carmen Loacutepez-Sabater12 Mariacutea I Covas25 Dolores Corella26 Jordi Salas-Salvadoacute27 Enrique Goacutemez-Gracia28Joseacute Lapetra29 Fernando Aroacutes210 Miquel Fiol211 Emili Ros212 Lluis Serra-Majem213 Xavier Pintoacute214Miguel A Muntildeoz215 Alfredo Gea24 Valentina Ruiz-Gutieacuterrez216 Ramoacuten Estruch217 Rosa M Lamuela-Raventoacutes12

and on behalf of the PREDIMED Study Investigators

Abstract

Background Polyphenols may lower the risk of cardiovascular disease (CVD) and other chronic diseases due totheir antioxidant and anti-inflammatory properties as well as their beneficial effects on blood pressure lipids andinsulin resistance However no previous epidemiological studies have evaluated the relationship between the intakeof total polyphenols intake and polyphenol subclasses with overall mortality Our aim was to evaluate whetherpolyphenol intake is associated with all-cause mortality in subjects at high cardiovascular risk

Methods We used data from the PREDIMED study a 7447-participant parallel-group randomized multicentercontrolled five-year feeding trial aimed at assessing the effects of the Mediterranean Diet in primary prevention ofcardiovascular disease Polyphenol intake was calculated by matching food consumption data from repeated foodfrequency questionnaires (FFQ) with the Phenol-Explorer database on the polyphenol content of each reportedfood Hazard ratios (HR) and 95 confidence intervals (CI) between polyphenol intake and mortality were estimatedusing time-dependent Cox proportional hazard models

Results Over an average of 48 years of follow-up we observed 327 deaths After multivariate adjustment wefound a 37 relative reduction in all-cause mortality comparing the highest versus the lowest quintiles of totalpolyphenol intake (hazard ratio (HR) = 063 95 CI 041 to 097 P for trend = 012) Among the polyphenol subclassesstilbenes and lignans were significantly associated with reduced all-cause mortality (HR =048 95 CI 025 to 091 P fortrend = 004 and HR = 060 95 CI 037 to 097 P for trend = 003 respectively) with no significant associations apparentin the rest (flavonoids or phenolic acids)

Conclusions Among high-risk subjects those who reported a high polyphenol intake especially of stilbenes andlignans showed a reduced risk of overall mortality compared to those with lower intakes These results may be usefulto determine optimal polyphenol intake or specific food sources of polyphenols that may reduce the risk of all-causemortality

Clinical trial registration ISRCTN35739639

Keywords Polyphenol intake All-cause mortality PREDIMED Mediterranean diet Stilbenes Lignans

Correspondence lamuelaubedu1Nutrition and Food Science Department XaRTA INSA Pharmacy SchoolUniversity of Barcelona Barcelona Spain2CIBER CB0603 Fisiopatologiacutea de la Obesidad y la Nutricioacuten (CIBERObn)Institute of Health ldquoCarlos IIIrdquo Government of Spain Madrid SpainFull list of author information is available at the end of the article

copy Tresserra-Rimbau et al licensee BioMed Central Ltd This is an Open Access article distributed under the terms of theCreative Commons Attribution License (httpcreativecommonsorglicensesby20) which permits unrestricted usedistribution and reproduction in any medium provided the original work is properly credited The Creative Commons PublicDomain Dedication waiver (httpcreativecommonsorgpublicdomainzero10) applies to the data made available in thisarticle unless otherwise stated

Tresserra-Rimbau et al BMC Medicine

2014

2014 1277httpwwwbiomedcentralcom1741-70151277

BackgroundDiet and lifestyle are crucial in the prevention of chronicillnesses and therefore substantially lower all-cause mortal-ity in most westernized countries There is evidence thatthe Mediterranean diet (MedDiet) a well characterizeddietary pattern is associated with longevity and improvedquality of life by reducing the risk of the most frequentchronic diseases such as cardiovascular diseases (CVD)metabolic syndrome age-related cognitive impairment type2 diabetes mellitus (T2DM) cancer and also all-cause mor-tality [12] The MedDiet is rich in fruits and vegetablesolive oil nuts legumes whole-wheat bread and fish andwine is consumed in moderate amounts during meals [2]With respect to nutrients the MedDiet is very rich inmono- and polyunsaturated fatty acids [3] and also inpolyphenols which are bioactive compounds mainlyfound in plant foods and plant-derived beverages suchas coffee tea and red wineSeveral studies have examined the association between

intake of certain polyphenol subgroups and their sourcesand the incidence of chronic degenerative diseases [4] aswell as their effects on blood pressure lipid profile andendothelial and platelet function [5-7] If polyphenolintake does protect against the development of chronicdiseases such as CVD cancer or T2DM we hypothe-sized that a greater consumption of polyphenols wouldcontribute to lower the risk of all-cause mortality andprovide a greater life expectancyTo date the association between specific groups of

polyphenols and mortality has been described [8] but toour knowledge neither total polyphenol intake nor thatof the different polyphenol subgroups have been associ-ated with all-cause mortality We therefore embarked ona study to evaluate the association between the intake oftotal polyphenols and polyphenol subgroups and the riskof overall mortality using the Phenol-Explorer database[9] to estimate the polyphenol intake recorded by thefood frequency questionnaires (FFQ) administered yearlyin the PREDIMED (Prevencioacuten con Dieta Mediterraacutenea)trial These results may be useful to determine optimalpolyphenol intake or specific food sources of polyphenolsthat may reduce the risk of all-cause mortality amongsubjects at high cardiovascular risk

MethodsThe PREDIMED studyThe PREDIMED study was a parallel-group randomizedmulticenter controlled feeding trial aimed at assessing theeffects of the MedDiet in the primary prevention of car-diovascular disease Details of the recruitment methodand study design have been described elsewhere [10] Theeligible participants were 7447 community-dwelling men(55 to 80 years) and women (60 to 80 years) from Spainwho had no cardiovascular disease at enrollment but were

at high risk they had either T2DM or at least three ofthe following major risk factors smoking hypertensiondyslipidemia overweight or obesity or a family historyof premature coronary heart disease Starting on 1 October2003 the eligible participants were randomized in a111 ratio to one of three dietary intervention groups1) MedDiet supplemented with extra-virgin olive oil(EVOO) 2) MedDiet supplemented with mixed nuts or3) control diet (low-fat diet) The trial was stopped aftera median follow-up of 48 years due to the benefit of theMedDiets with respect to major cardiovascular eventsmyocardial infarction stroke or death from cardiovas-cular causes (analysis performed by the Drug and SafetyMonitoring Board of the trial) compared to a controllow-fat group [2] All participants provided written in-formed consent and the study protocol was approvedby the Institutional Review Boards of the participatingcenters (Hospital Cliacutenic of Barcelona (coordinating centre)Universities of Barcelona Valencia Rovira-Virgili Maacutelagaand Las Palmas Municipal Institute for Medical ResearchPrimary Care Division of Barcelona and Sevilla Instituteof Research in Health Sciences (IUNICS) at Palma deMallorca Hospital Txangorritxu of Vitoria and UniversityHospital of Bellvitge) and registered [11]

Study population and characteristicsThe present study was conducted as a re-analysis of anintervention feeding study using polyphenol intake asthe exposure Data came from all participants of thePREDIMED trial but we excluded 247 individuals withan inadequate FFQ at baseline and 28 with a total energyintake out of the predefined limits (that is daily energyintake lt500 or gt3500 for women and lt800 or gt4000 kcaldfor men n = 28) [12] Therefore data from 7172 partici-pants were available for this analysisParticipants filled out the following questionnaires at

baseline and yearly thereafter a validated 14-point scorequestionnaire on adherence to the traditional MedDiet[13] a validated 137-item FFQ [14] and a general ques-tionnaire which included data on lifestyle habits concur-rent diseases and medication used

Polyphenol intake and dietary assessmentAt baseline and yearly thereafter trained dietitians com-pleted the validated 137-item FFQ [14] in a face-to-faceinterview with the participant Energy and nutrient in-take were estimated from the FFQ by multiplying thefrequency of consumption by the average portion sizeusing Spanish food composition tablesIn a previous study conducted by our group total poly-

phenol excreted in spot urine samples was validated asa biomarker of total polyphenol intake from FFQ in aclinical trial (r = 048 P lt001) and in a cross-sectionalstudy (r = 026 P= 004) [15] The Phenol-Explorer database

Tresserra-Rimbau et al BMC Medicine Page 2 of 112014 1277httpwwwbiomedcentralcom1741-70151277

[9] was used to obtain information about polyphenol con-tent in foods This database included 516 polyphenols con-tained in 456 foods [16] at the time of our analysis beingthe most complete database currently available for polyphe-nol content Correspondence between food items in theFFQ and the Phenol-Explorer database has been describedpreviously [17] Individual polyphenol intake was calculatedby multiplying the content of each polyphenol in a particularfood item (mgg) by the daily consumption of this food item(gday) and then summing the product across all food itemsTotal polyphenol intake was the sum of all individual poly-phenol intakesPolyphenol and other nutrient intakes were adjusted

for total energy intake because it is associated with diseaserisk and is usually proportional to most nutrient intake[18] To conduct the analyses we also used weighted cu-mulative averages that is the polyphenol intake of a givenyear was the average between the intake of that year andthe average of the previous years

Ascertainment of the outcomeInformation on mortality was updated yearly by the end-point adjudication committee whose members wereunaware of dietary intakes or intervention assignmentsThe sources of information were the following yearlyquestionnaires and examinations from all participantsfamily physicians yearly review of medical records andlinkage to the National Death Index All outcomes werereported between 1 October 2003 and 1 December 2010

Statistical analysesWe calculated the weighted cumulative average of poly-phenol intake at each yearly visit to represent long-termpolyphenol intake Polyphenols and other food and nu-trient intake were adjusted for total calories using theresidual method Non-dietary covariates such as smokingbody mass index (BMI) physical activity and medicationuse were updated yearlyThe baseline characteristics of the 7172 participants

were distributed by quintiles of total polyphenol intakeData were presented as means (plusmnSD) for continuousvariables and frequencies and percentages for categor-ical variables We used one-factor ANOVA or Pearsonchi-squared tests to compare the quantitative or cat-egorical baseline characteristics of the study participantsacross quintiles of baseline polyphenol intake Person-time for each participant was calculated as the timebetween randomization and the date of death the datewhen completing the last interview 1 December 2010or date at death whichever came first To assess the riskof total mortality by quintiles of polyphenol intakewe ran time-dependent Cox proportional hazard regres-sions with updated diet and covariates The referentgroup was the lowest quintile of polyphenol intake

Results are expressed as hazard ratios (HRs) with 95confidence intervals (CIs) To show the crude differ-ences in death rates by groups of polyphenol intake weperformed a Nelson Aalen survival function a non-parametric estimator of the survival function for cen-sored dataMoreover we used likelihood ratio tests of interaction

in stratified analyses to study the possible interactionsamong the main risk factors and as sensitivity analyseswe estimated the fully adjusted HR excluding participantswith less than one or two years of follow-up

CovariatesTo take into account the potential differences in risk fac-tors all Cox proportional hazard analyses were carried outwith stratification for recruitment center sex and interven-tion group In model 2 we adjusted for sex age (lt60 60 to649 65 to 699 70 to 749 gt=75 years) smoking status(never past and current cigarettes (lt5 5 to 19 gt20 perday) or cigars and pipes (lt3 3 to 6 gt6 per day)) BMI(lt25 25 to 299 or gt=30 Kgm2) baseline diabetes alco-hol consumption (0 01 to 149 15 to 299 gt=30 gday)total energy intake (continuous variable) physical activity(continuous variable) family history of CVD andorcancer aspirin use antihypertensive drug use use ofcardiovascular medication use of oral hypoglycemicagents insulin and other medication In model 3 weadditionally adjusted for intake of protein saturated fattyacids polyunsaturated fatty acids monounsaturated fattyacids and cholesterol We did not include in the modelother variables that did not change the HR by 10 ormoreStatistical analyses were conducted using SAS soft-

ware version 93 (SAS Institute Inc Cary NC USA)All t tests were two-sided and P-values below 005 wereconsidered significant

ResultsThe baseline characteristics of participants are shown byquintiles of energy-adjusted total polyphenol intake inTable 1 Participants with a greater intake of total polyphe-nols had a closer adherence to the traditional MedDietThey also tended to be more physically active consumemore alcoholic beverages (mostly wine and beer) and tohave less hypertension On the contrary the prevalence ofhypercholesterolemia was higher in those who consumedmore polyphenols at baseline and they were more likely tobe smokers The groups did not differ in terms of diabetesstatus use of medication and distribution into the threearms of the trialDuring a mean of 48 years of follow-up among 31068

person-years the total number of observed deaths was327 Of these 131 were due to cancer 81 were cardio-vascular and 115 were for other causes The Nelson

Tresserra-Rimbau et al BMC Medicine Page 3 of 112014 1277httpwwwbiomedcentralcom1741-70151277

Table 1 Baseline characteristics according to quintiles of total polyphenol intake at baseline (energy-adjusted)

Q1 Q2 Q3 Q4 Q5 P-value

(n = 1434) (n = 1435) (n = 1434) (n = 1435) (n = 1434)

Polyphenol intake mean (cutoff values) mgd 483 (lt642) 674 (642 to 749) 794 (750 to 852) 937 (853 to 995) 1235 (gt995)

Sex women 836 (583) 924 (644) 712 (608) 803 (560) 648 (452) lt00001

Age mean (SD) y 676 (62) 674 (61) 674 (59) 669 (60) 662 (61) lt00001

BMI mean (SD) Kgm2 300 (37) 303 (37) 297 (35) 297 (37) 296 (35) lt00001

Current smoker 217 (151) 210 (146) 194 (135) 265 (185) 317 (221) lt00001

Former smoker 273 (190) 263 (183) 317 (221) 319 (222) 413 (288)

Sportsexercise mean (SD) MET-hd 337 (356) 362 (383 377 (366) 405 (425) 459 (454) lt00001

Diabetes 706 (492) 680 (474) 712 (496) 704 (491) 668 (466) 040

Hypertension 1230 (858) 1224 (853) 1192 (831) 1166 (813) 1117 (779) lt00001

Hypercholesterolemia 983 (686) 1018 (709) 1053 (734) 1065 (742) 1069 (746) 0001

Hypolipidemic drug use 660 (461) 670 (467) 712 (497) 716 (501) 706 (495) 009

Antihypertensive drug use 1071 (747) 1095 (764) 1027 (717) 1030 (720) 994 (697) 00004

Cardiovascular drugs use 118 (85) 114 (82) 120 (86) 110 (79) 109 (79) 094

Insulin use 90 (63) 87 (61) 115 (80) 95 (66) 99 (69) 026

Anti-diabetes drug use other than insulin 463 (323) 454 (317) 478 (334) 465 (325) 439 (308) 065

Aspirin use 302 (211) 326 (228) 337 (235) 318 (222) 324 (227) 063

Int Group MedDiet-EVOO 489 (341) 506 (353) 477 (336) 473 (330) 517 (361) 0001

Int Group MedDiet-nuts 444 (310) 467 (325) 454 (317) 491 (342) 519 (362)

Mean daily intake

Total energy intake mean (SD) Kcald 2397 (642) 2180 (589) 2161 (540) 2229 (563) 2369 (577) lt00001

Carbohydrates mean (SD) gd 240 (45) 237 (39) 235 (37) 234 (41) 236 (45) 0006

Protein mean (SD) gd 919 (151) 924 (138) 924 (132) 915 (136) 906 (149) 0004

SFA mean (SD) gd 261 (67) 254 (57) 251 (53) 249 (55) 235 (58) lt00001

MUFA mean (SD) gd 490 (122) 488 (106) 488 (107) 487 (113) 466 (112) lt00001

PUFA mean (SD) gd 156 (58) 159 (51) 158 (50) 158 (52) 150 (52) lt00001

Fiber mean (SD) gd 215 (61) 239 (64) 255 (67) 266 (74) 294 (89) lt00001

Total cholesterol mean (SD) mgd 372 (121) 367 (103) 368 (107) 360 (94) 354 (122) lt00001

Alcohol mean (SD) gd 410 (109) 63 (101) 76 (105) 93 (128) 146 (189) lt00001

Vegetables mean (SD) gd 296 (140) 319 (127) 338 (139) 351 (142) 369 (169) lt00001

Fruits mean (SD) gd 240 (133) 319 (145) 364 (157) 404 (172) 521 (245) lt00001

Legumes mean (SD) gd 205 (153) 207 (152) 203 (109) 206 (124) 206 (130) 093

Dairy products mean (SD) gd 398 (226) 391 (216) 389 (208) 380 (219) 353 (217) lt00001

Cereals mean (SD) gd 247 (98) 233 (81) 227 (78) 219 (79) 209 (80) lt00001

Meat or meat products mean (SD) gd 135 (60) 132 (54) 132 (50) 130 (50) 129 (55) 003

Fish mean (SD) gd 943 (533) 999 (468) 101 (515) 996 (450) 102 (492) 00005

Sugar-sweetened soft drinks mean (SD) gd 250 (843) 197 (633) 178 (558) 154 (561) 126 (463) lt00001

Coffee mean (SD) gd 258 (363) 436 (401) 552 (429) 703 (492) 901 (638) lt00001

14-points MedDiet questionnaire score mean (SD) 82 (19) 85 (19) 87 (19) 87 (19) 92 (18) lt00001

Risk factors

Waist-to-height ratio mean (SD) 064 (006) 063 (007) 063 (006) 062 (006) 062 (006) lt00001

Systolic BP mean (SD) mmHg 150 (19) 151 (19) 149 (19) 148 (18) 148 (18) 001

Diastolic BP mean (SD) mmHg 83 (10) 84 (98) 82 (96) 82 (98) 83 (96) 0003

Hearth rate mean (SD) beatsmin 717 (110) 712 (109) 707 (111) 700 (105) 705 (105) 002

Tresserra-Rimbau et al BMC Medicine Page 4 of 112014 1277httpwwwbiomedcentralcom1741-70151277

Aalen survival function (Figure 1) shows the crude dif-ferences in death rates by groups of polyphenol intakelow (lt600 mgd) medium (600 to 750 mgd) and high(gt750 mgd)Table 2 shows Cox Proportional HRs and 95 CI for

total mortality according to quintiles of cumulative in-take of total polyphenols (according to yearly updatedassessments) After adjusting for all potential confoundersand stratifying by sex recruitment center and interventiongroup the HR for the highest versus the lowest quintilewas 060 (95 CI 039 to 091 P-trend = 007) After fur-ther adjustment for other dietary confounders the associ-ation was not substantially attenuated (HR 063 95 CI041 to 097 P-trend = 012) We did not see a strong in-verse linear trend for total polyphenols instead the resultssuggest a modest threshold above the first quintile ofintakeIn some cases follow-ups were too short to assess a

mortality endpoint because the ill-health conditionsleading to death may influence diet Therefore as sen-sitivity analyses we estimated the fully adjusted HR forthe category of the highest total polyphenol intake vs

the lowest excluding participants with less than one (31excluded) or two years of follow-up (75 excluded) In bothcases the association was robust and remained statisticallysignificant HR 057 95 CI 036 to 090 P-trend = 007and HR 049 95 CI 030 to 082 P-trend = 003respectivelyWe also conducted stratified analyses (Table 3) by the

other strong predictors of mortality In multivariablemodels the inverse association between total polyphenolintake and risk of death comparing the extreme quintileswas stronger among women (HR 042 95 CI 018 to098 P-trend = 024) than men (HR 076 95 CI 046 to126 P-trend = 023) although the interaction for sex wasnot significant (P-interaction = 039) We also observed nosignificant differences by strata of age (lt70 vs gt=70 years)However we noted that those who did not drink alcoholhad a stronger inverse association with total polyphenolintake (HR 039 95 CI 017 to 090 P-trend = 004) thandrinkers (HR 099 95 CI 059 to 165 P-trend = 091)but the interaction was not significant (P-interaction =016) In other stratified analyses we observed that theinverse association did not change substantially among

Table 1 Baseline characteristics according to quintiles of total polyphenol intake at baseline (energy-adjusted)(Continued)

Glucose (n = 4311) mean (SD) mgdL 118 (41) 116 (39) 122 (42) 123 (43) 123 (43) 00007

Cholesterol (n = 4286) mean (SD) mgdL 202 (36) 206 (38) 207 (39) 208 (38) 207 (36) 0003

HDL (n = 4236) mean (SD) mgdL 50 (11) 51 (11) 51 (11) 52 (12) 52 (11) 0007

Triglycerides (n = 4291) mean (SD) mgdL 130 (67) 133 (74) 137 (79) 130 (63) 138 (80) 006

BMI Body Mass Index BP Blood pressure MedDiet-EVOO Mediterranean Diet supplemented with extra virgin olive oil MedDiet-nuts Mediterranean Diet supplementedwith nuts SFA Saturated fatty acids MUFA Monounsaturated fatty acids PUFA Polyunsaturated fatty acids HDL High density lipoproteinData are expressed as No () unless otherwise indicatedP-values calculated by analysis of variance or χ2 tests

Figure 1 Nelson Aalen estimates of the incidence of death by groups of polyphenol intake

Tresserra-Rimbau et al BMC Medicine Page 5 of 112014 1277httpwwwbiomedcentralcom1741-70151277

smokers and non-smokers in those who were physicallyactive or inactive or in those with or without T2DM orhypertension and none of these interactions were signifi-cant Finally we conducted stratified analyses by interven-tion groups and found a slightly stronger associationbetween total polyphenol intake and death in the controlarm of the trial (HR 048 CI 023 to 098 P-trend = 001)than in the MedDiet + EVOO arm (HR 067 CI 031 to146 P-trend = 068) and the MedDiet + nuts arm (HR068 CI 034 to 135 P-trend = 081) However the inter-action (P = 071) was not statistically significant suggestingno apparent effect modificationWe further investigated the possible effects of the in-

take of the main polyphenol groups on mortality by anycause (Table 4) Although no significant associationswere found for flavonoids or phenolic acids we observeda 46 reduction in risk of death in participants whoconsumed more stilbenes (HR 048 CI 025 to 091P-trend = 004) and lignans (HR 060 CI 037 to 095P-trend = 003) For ldquoother polyphenolsrdquo such as tyro-sols alkylphenols hydroxybenzaldehydes furanocouma-rins and hydroxycoumarins the association was attenuatedafter adjustment for other nutrientsExploratory analyses (Figure 2) were done for flavo-

noids (see Additional file 1) and phenolic acid subclasses(see Additional file 2) We found a strong trend towardsa reduction in death risk with a higher intake of isofla-vones (HR 049 CI 028 to 084 P-trend = 0009) Dihydro-flavonols were also inversely associated with the risk ofdeath after multivariable adjustment (HR 053 CI 028 to099 P-trend = 005) and the inverse trend was statisticallysignificant after additional adjustment (P-trend = 004) Noother subclasses were associated with mortality by anycause

DiscussionIn this reanalysis of the data of the PREDIMED trial weobserved a 37 reduction of mortality when comparing

extreme quintiles of total polyphenol intake The dose-response trend for the association between total polyphe-nol intake and all-cause mortality suggested an L-shapedrelationship with an apparent threshold after the firstquintile of polyphenol intake instead of an inverse lineardose-response relationship Within the polyphenol sub-classes stilbenes and lignans were inversely associatedwith total mortalityIn stratified analyses we found a stronger association

between total polyphenol intake and mortality risk forwomen and for those who did not drink alcohol Althoughthe interaction terms were not significant the observedtrend was suggestive especially for non-drinkers The re-lationship between alcohol intake and polyphenols shouldbe the main focus of future studiesTo our knowledge though previous studies have in-

vestigated the association between intake of specificgroups of polyphenols and mortality this is the firststudy to investigate the association between total poly-phenol intake as well as that of all polyphenol sub-groups with all-cause mortality In addition we shouldacknowledge that the effect of polyphenols and polyphenol-rich foods on chronic degenerative diseases and clinicalbiomarkers has been broadly studied [19-24] Previousstudies have analyzed the association between polyphenolsfrom wine tea chocolate berries soy and olive oil withseveral chronic degenerative disease risk or mortalityrisk [625-29] The reported inverse association specif-ically for olive oil and red wine is consistent with theinverse association we found for stilbenes and lignans[29-31] The suggestion of an inverse association thatwe found for several flavonoid compounds is also con-sistent with previous studies of berries dark chocolateand soy [62526] In many of these previously studiedpopulations intake of any one polyphenol-rich food wasnot great enough to reduce mortality but in our studytotal polyphenol intake was a wider range coming fromseveral food sources

Table 2 Cox proportional hazard ratios for total mortality according to quintiles of cumulative total polyphenol intake

Quintiles of cumulative intake of total polyphenols mgd

Q1 (535) Q2 (700) Q3 (800) Q4 (917) Q5 (1170) P-trend

No of deaths 88 62 52 63 62

No of person-years 5505 6599 6767 6559 5638

Age- and sex-adjusted HR (95 CI) 100 065 (044 to 095) 055 (037 to 082) 073 (050 to 106) 066 (044 to 098) 012

Multivariable-adjusted HR (95 CI)dagger 100 068 (046 to 101) 060 (039 to 090) 075 (051 to 112) 060 (039 to 091) 007

Additionally adjusted HR (95 CI)Dagger 100 071 (048 to 105) 062 (041 to 095) 079 (053 to 117) 063 (041 to 097) 012

HR Hazard ratio CI Confidence intervalAnalyses were stratified by sex recruitment center and intervention groupdaggerThe multivariable HR has been additionally adjusted for age (lt60 60 to 649 65 to 699 70 to 749 gt=75 years) smoking (never past and current cigarettes(lt5 5 to 19 gt20 per day) or cigars and pipes (lt3 3 to 6 gt6 per day)) BMI (lt25 25 to 299 or gt=30 Kgm2) baseline diabetes alcohol (0 01 to 149 15 to299 gt=30 gday) total energy intake (continuous variable) physical activity (continuous variable) family history of CVD or cancer aspirin use antihypertensivedrug use use of cardiovascular medication use of oral hypoglycaemic agents insulin other medicationDaggerThis model has been additionally adjusted for intake of protein saturated fatty acids polyunsaturated fatty acids monounsaturated fatty acids and cholesterol(all as continuous variables)

Tresserra-Rimbau et al BMC Medicine Page 6 of 112014 1277httpwwwbiomedcentralcom1741-70151277

Kuriyama et al conducted a prospective cohort studyamong 40530 healthy Japanese adults and reported thatgreen tea consumption a polyphenol-rich beveragewas inversely associated with cardiovascular diseasesand all-cause mortality but not with mortality due tocancer [27] Other studies have also found an inverseassociation between polyphenol consumption and CVDand CVD-related mortality [20252632] Indeed it hasbeen demonstrated that some polyphenols and theirmetabolites exert anti-atherosclerotic effects improveendothelial function and antioxidant status increase ni-tric oxide release and modulate inflammation and lipidmetabolism [5212533-35]Polyphenols can also act as chemopreventive agents

For example resveratrol is a well-known stilbene mostly

found in red wine and grapes with several health benefitsincluding inhibition of tumorgenesis [83637] In vitro andin vivo studies have shown that epigallocatechin-3-gallatethe major polyphenol of green tea has anti-carcinogeniceffects such as inhibition of growth proliferation in-duction of apoptosis and phase II detoxifying enzymesand reduction of oxidative damage to DNA [36-38]Xanthohumol quercetin curcumin and genistein areother examples of polyphenols that have shown anti-carcinogenic properties due to their capacity to inhibittumor growth [8223738]Available evidence supports that dietary modifications

are able to reduce the risk of T2DM another highlyprevalent chronic disease Wedick et al found that antho-cyanins were inversely associated with the risk of T2DM

Table 3 HR for total mortality according to quintiles of total polyphenol intake (stratified by risk factors)

Risk factor No of deaths No of person-years Multivariable-adjusted HR(95 CI) Quintile 5 vs 1

P-trend P-interaction

Sex

Men 203 13317 076 (046 to 126) 023 039

Women 124 17751 042 (018 to 098) 024

Age y

lt70 142 21483 058 (031 to 108) 021 073

ge70 185 9585 070 (039 to 124) 034

Alcohol intake

Nondrinkers 133 12510 039 (017 to 090) 004 016

Drinkers 194 18558 099 (059 to 165) 091

Smoking

Never 144 19520 064 (031 to 132) 047 093

Former 111 7465 052 (025 to 107) 029

Current 72 4083 071 (029 to 175) 021

Physical activity

Less than median 203 16224 057 (032 to 102) 017 043

More than median 124 14844 077 (041 to 144) 073

Hypertension

Yes 184 12080 063 (036 to 110) 024 021

No 134 17721 082 (044 to 155) 076

Diabetes mellitus

Yes 205 15345 079 (047 to 133) 092 052

No 122 15723 060 (031 to 117) 009

Intervention group

MedDiet-EVOO 113 11478 067 (031 to 146) 068 071

MedDiet-Nuts 108 10134 068 (034 to 135) 081

Control Diet 106 9456 048 (023 to 098) 001

HR Hazard ratio CI Confidence interval MedDiet-EVOO Mediterranean Diet supplemented with extra virgin olive oil MedDiet-nuts Mediterranean Dietsupplemented with nutsThe multivariable HR has been additionally adjusted for age (lt60 60to 49 65 to 699 70 to 749 gt=75 years) smoking (never past and current cigarettes(lt5 5 to 19 gt20 per day) or cigars and pipes (lt3 3 to 6 gt6 per day)) BMI (lt25 25 to 299 or gt=30 Kgm2) baseline diabetes alcohol (0 01 to 149 15 to299 gt=30 gday) total energy intake (continuous variable) physical activity (continuous variable) family history of CVD or cancer aspirin use antihypertensivedrug use use of cardiovascular medication use of oral hypoglycemic agents insulin other medication Analyses were stratified by sex recruitment center andintervention group

Tresserra-Rimbau et al BMC Medicine Page 7 of 112014 1277httpwwwbiomedcentralcom1741-70151277

using data from three US prospective cohorts and Murakiet al found similar associations for blueberries grapesand apples [3940] Finally polyphenols have been pro-posed as promising phytochemicals for the treatmentand prevention of neurogenerative diseases such asAlzheimerrsquos disease Parkinsonrsquos disease and other neuro-logical disorders [2941]All of this evidence from chronic disease studies sup-

ports the hypothesis that greater polyphenol intake andthe many polyphenol subclasses this represents serves

to extend the life span through multifactorial etiologicalpathwaysOur study has some limitations First we controlled

for several confounders in multivariate models but otherunknown or unmeasured confounders may exist How-ever if this were the case we would expect relative risksfor all subclasses to be equally over or underestimated andthat was not the case Second the number of cases ofcause-specific deaths was too low to estimate individualrelative risks Others have found the benefits of specific

Table 4 Relationship between mortality and intake of the main polyphenol groups (in quintiles)

Main groups Q1 Q2 Q3 Q4 Q5 P-trend

Flavonoids (mgd) 273 362 431 512 670

No of deaths 76 73 42 69 67

No of person-years 4890 6599 6755 6867 5957

Age- and sex-adjusted HR (95 CI) 100 076 (052 to 110) 054 (036 to 081) 072 (049 to 105) 070 (047 to 105) 023

Multivariable-adjusted HR (95 CI)dagger 100 092 (062 to 134) 069 (045 to 107) 092 (062 to 136) 083 (055 to 127) 070

Additionally adjusted HR (95 CI)Dagger 100 096 (065 to 141) 075 (048 to 116) 099 (066 to 147) 089 (058 to 136) 095

Phenolic acids (mgd) 159 229 279 345 453

No of deaths 80 58 62 69 58

No of person-years 5928 6662 6716 6615 5147

Age- and sex-adjusted HR (95 CI) 100 095 (065 to 139) 078 (053 to 116) 101 (070 to 147) 095 (063 to 142) 064

Multivariable-adjusted HR (95 CI)dagger 100 094 (064 to 139) 082 (055 to 123) 107 (072 to 158) 079 (051 to 122) 025

Additionally adjusted HR (95 CI)Dagger 100 089 (060 to 131) 077 (052 to 116) 101 (068 to 150) 075 (049 to 116) 020

Stilbenes (mgd) 0 048 104 204 575

No of deaths 69 64 47 74 73

No of person-years 5191 6547 6840 6527 5963

Age- and sex-adjusted HR (95 CI) 100 071 (047 to 105) 066 (044 to 098) 081 (056 to 118) 073 (056 to 118) 044

Multivariable-adjusted HR (95 CI)dagger 100 061 (033 to 111) 053 (028 to 099) 068 (038 to 122) 042 (022 to 081) 004

Additionally adjusted HR (95 CI)Dagger 100 069 (038 to 127) 062 (033 to 116) 078 (043 to 140) 048 (025 to 091) 004

Lignans (mgd) 044 057 067 077 094

No of deaths 76 72 57 55 67

No of person-years 4457 6002 6737 7146 6726

Age- and sex-adjusted HR (95 CI) 100 066 (046 to 096) 058 (039 to 085) 058 (039 to 087) 054 (035 to 082) 0002

Multivariable-adjusted HR (95 CI)dagger 100 065 (044 to 099) 056 (038 to 084) 056 (036 to 084) 051 (032 to 079) 0001

Additionally adjusted HR (95 CI)Dagger 100 068 (046 to 100) 060 (040 to 092) 062 (039 to 098) 060 (037 to 097) 003

Others (mgd) 37 53 66 82 113

No of deaths 77 65 72 60 53

No of person-years 4604 6442 7320 6777 5925

Age- and sex-adjusted HR (95 CI) 100 076 (052 to 111) 078 (054 to 113) 068 (046 to 101) 064 (042 to 096) 004

Multivariable-adjusted HR (95 CI)dagger 100 076 (051 to 113) 080 (054 to 118) 067 (045 to 102) 061 (040 to 093) 003

Additionally adjusted HR (95 CI)Dagger 100 082 (055 to 122) 086 (058 to 127) 076 (050 to 116) 070 (046 to 109) 013

HR Hazard Ratio CI confidence intervalAnalyses were stratified by sex recruitment center and intervention groupdaggerThe multivariable HR has been additionally adjusted for age (lt60 60 to 649 65 to 699 70 to 749 gt=75 years) smoking (never past and current cigarettes(lt5 5 to 19 gt20 per day) or cigars and pipes (lt3 3 to 6 gt6 per day)) BMI (lt25 25 to 299 or gt=30 Kgm2) baseline diabetes alcohol (0 01 to 149 15 to299 gt=30 gday) total energy intake (continuous variable) physical activity (continuous variable) family history of CVD or cancer aspirin use antihypertensivedrug use use of cardiovascular medication use of oral hypoglycaemic agents insulin other medicationDaggerThis model has been additionally adjusted for intake of protein saturated fatty acids polyunsaturated fatty acids monounsaturated fatty acids and cholesterol(all as continuous variables)

Tresserra-Rimbau et al BMC Medicine Page 8 of 112014 1277httpwwwbiomedcentralcom1741-70151277

foods are stronger for CVD mortality than cancer or re-spiratory disease Future work in this area should includelarger studies with estimates of total polyphenol intakeThird there were limitations with respect to the estima-tion of polyphenol intake because data were indirectlyderived from the FFQs Although urinary excretion ofpolyphenols was validated as a biomarker of total polyphe-nol from the FFQ in two different studies the values of rwere relatively low The absence of information aboutsome foods in the FFQ could lead to an underestimationof the intake Moreover the study did not take intoaccount the bioavailability of these molecules Finallythese results might be valid only for elderly people athigh cardiovascular risk and other studies are needed togeneralize the conclusions to other populationsOn the other hand the main strengths of the study are

the prospective design the large sample size with a rela-tively long-term follow-up and comprehensive data onrisk factors and confounders Very importantly our useof the cumulative average of polyphenol intake acrossyearly repeated measurements of diet is considered as thebest approach to reduce measurement error in nutritional

epidemiology [42] and allowed changes in the diet dueto the intervention or other secular trends in intake inSpain to be controlled We also used the most com-prehensive polyphenol database currently available(Phenol-explorer database) which allowed risk estima-tion related not only to intake of total polyphenol butalso all the polyphenol subgroups and subclasses Thiscomprehensive analysis differentiates our paper fromprevious related studies

ConclusionsWe found an apparent inverse association between totalpolyphenol intake and the risk of overall mortality whichwas independent of other dietary and non-dietary risk fac-tors This may be helpful in establishing future daily poly-phenol intake recommendations However more studiesare needed to definitively clarify the benefits deriving fromlong-term consumption of polyphenol-rich foods

Other PREDIMED InvestigatorsOther contributors list (Additional file 3)

Figure 2 Hazard ratios (95 CI) of total mortality for the highest vs lowest quintiles of polyphenol intake

Tresserra-Rimbau et al BMC Medicine Page 9 of 112014 1277httpwwwbiomedcentralcom1741-70151277

Additional files

Flavonoidsdoc

Phenolic acidsdoc

Other contributorsrsquo listdoc

AbbreviationsANOVA Analysis of Variance BMI Body Mass Index CVD Cardiovasculardiseases EVOO Extra Virgin Olive Oil FFQ Food Frequency QuestionnaireHR Hazard ratio MedDiet Mediterranean Diet PREDIMED Prevencioacuten conDieta Mediterraacutenea SD Standard deviation T2DM Type 2 diabetes mellitus95 CI 95 Confidence interval

Competing interestsThe authors declare that they have no competing interests

Authorsrsquo contributionsATR RMLR EBR RE and MAMG carried out the statistical analyses interpretedthe data and drafted the manuscript RMLR RE MAMG AMR MCLS MIC DCJSS EGG JL FA MF ER LSM XP MAM AG and VRG participated in the designof the study and the acquisition of data and contributed to the critical reviewof the paper All authors read and approved the final manuscript

AcknowledgementsWe would like to thank all the volunteers involved in the PREDIMED studyfor their valuable cooperation This study was supported in part by CICYT(AGL2010-22319-C03) from the Spanish Ministry of Science and Innovation(MICINN) and the Instituto de Salud Carlos III ISCIII (CIBERobn-CB0603 RD060045 PI1002658 and PI1001407) The CIBERobn is an initiative of the ISCIIISpain ATR received support from ISCIII (FI1000265)

Author details1Nutrition and Food Science Department XaRTA INSA Pharmacy SchoolUniversity of Barcelona Barcelona Spain 2CIBER CB0603 Fisiopatologiacutea de laObesidad y la Nutricioacuten (CIBERObn) Institute of Health ldquoCarlos IIIrdquo Governmentof Spain Madrid Spain 3Harvard Medical School and Harvard School of PublicHealth Boston MA USA 4Department of Preventive Medicine and PublicHealth School of Medicine University of Navarra Pamplona Spain5Cardiovascular Epidemiology Unit Municipal Institute for Medical Research(IMIM) Barcelona Spain 6Department of Epidemiology Preventive Medicineand Public Health School of Medicine University of Valencia Valencia Spain7Human Nutrition Unit School of Medicine IISPV University Rovira i Virgili ReusSpain 8Department of Epidemiology School of Medicine University of MalagaMaacutelaga Spain 9Department of Family Medicine Primary Care Division of SevillaSan Pablo Health Center Sevilla Spain 10Department of Cardiology HospitalTxangorritxu Vitoria Spain 11Institut Universitari dacuteInvestigacioacute en Ciegravencies de laSalut (IUNICS) Palma de Mallorca Spain 12Lipid Clinic Endocrinology andNutrition Service Institut drsquoInvestigacions Biomeacutediques August Pi i Sunyer(IDIBAPS) Hospital Clinic Barcelona Spain 13Department of Clinical SciencesUniversity of Las Palmas de Gran Canaria Palmas de Gran Canaria Spain 14LipidUnit Department of Internal Medicine IDIBELL-Hospital Universitari de BellvitgeLHospitalet de Llobregat FIPEC Barcelona Spain 15Primary Care DivisionCatalan Institute of Health Barcelona Spain 16Nutrition and Lipids MetabolismInstituto de la Grasa Consejo Superior de Investigaciones Cientificas SevillaSpain 17Department of Internal Medicine Hospital Cliacutenic IDIBAPS University ofBarcelona Barcelona Spain

Received 23 January 2014 Accepted 10 April 2014Published

References1 Sofi F Abbate R Gensini GF Casini A Accruing evidence on benefits of

adherence to the Mediterranean diet on health an updated systematicreview and meta-analysis Am J Clin Nutr 2010 921189ndash1196

2 Estruch R Ros E Salas-Salvadoacute J Covas M Corella D Aroacutes F Goacutemez-GraciaE Ruiz-Gutieacuterrez V Fiol M Lapetra J Lamuela-Raventos R Serra-Majem LPintoacute X Basora J Muntildeoz MA Sorliacute JV Martiacutenez JA Martiacutenez-Gonzaacutelez MAPrimary prevention of cardiovascular disease with a Mediterranean dietN Engl J Med 2013 3681279ndash1290

3 Pauwels EK The protective effect of the Mediterranean diet focus oncancer and cardiovascular risk Med Princ Pract 2011 20103ndash111

4 Sies H Polyphenols and health update and perspectives Arch BiochemBiophys 2010 5012ndash5

5 Andriantsitohaina R Auger C Chataigneau T Etienne-Selloum N Li H MartinezMC Schini-Kerth VB Laher I Molecular mechanisms of the cardiovascularprotective effects of polyphenols Br J Nutr 2012 1081ndash18

6 Erlund I Koli R Alfthan G Marniemi J Puukka P Mustonen P Mattila P JulaA Favorable effects of berry consumption on platelet function bloodpressure and HDL cholesterol Am J Clin Nutr 2008 87323ndash331

7 Medina-Remoacuten A Estruch R Tresserra-Rimbau A Vallverdu-Queralt ALamuela-Raventos RM The effect of polyphenol consumption on bloodpressure Mini Rev Med Chem 2013 131137ndash1149

8 Aggarwal BB Bhardwaj A Aggarwal RS Seeram NP Shishodia S Takada YRole of resveratrol in prevention and therapy of cancer preclinical andclinical studies Anticancer Res 2004 242783ndash2840

9 Phenol-Explorer Database on Polyphenol Content in Foods[wwwphenol-explorereu]

10 Martinez-Gonzalez MA Corella D Salas-Salvado J Ros E Covas MI Fiol MWarnberg J Aros F Ruiz-Gutierrez V Lamuela-Raventos RM Lapetra JMuntildeoz MA Martinez JA Saez G Serra-Majem L Pinto X Mitjavila MT Tur JAPortillo MD Estruch R Cohort Profile design and methods of thePREDIMED study Int J Epidemiol 2012 41377ndash385

11 Schroumlder H Fitoacute M Estruch R Martiacutenez-Gonzaacutelez MA Corella D Salas-Salvadoacute JLamuela-Raventoacutes R Ros E Salaverriacutea I Fiol M Lapetra J Vinyoles EGoacutemez-Gracia E Lahoz C Serra-Majem L Pintoacute X Ruiz-Gutierrez V Covas MI AShort Screener Is Valid for Assessing Mediterranean Diet Adherence amongOlder Spanish Men and Women J Nutr 2011 1411140ndash1145

12 Willett W Issues in analysis and presentation of dietary data In NutritionalEpidemiology 2nd edition Edited by Willett W New York Oxford UniversityPress 1998321ndash346

13 Schroumlder H Covas MI Marrugat J Vila J Pena A Alcaacutentara M Masiaacute R Useof a three-day estimated food record a 72-hour recall and a food-frequency questionnaire for dietary assessment in a MediterraneanSpanish population Clin Nutr 2001 20429ndash437

14 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 Relativevalidity of a semi-quantitative food-frequency questionnaire in an elderlyMediterranean population of Spain Br J Nutr 2010 1031808ndash1816

15 Medina-Remoacuten A Barrionuevo-Gonzaacutelez A Zamora-Ros R Andres-LacuevaC Estruch R Martiacutenez-Gonzaacutelez M Diez-Espino J Lamuela-Raventos RMRapid FolinndashCiocalteu method using microtiter 96-well plate cartridgesfor solid phase extraction to assess urinary total phenolic compounds asa biomarker of total polyphenols intake Anal Chim Acta 2009 63454ndash60

16 Perez-Jimenez J Fezeu L Touvier M Arnault N Manach C Hercberg SGalan P Scalbert A Dietary intake of 337 polyphenols in French adultsAm J Clin Nutr 2011 931220ndash1228

17 Tresserra-Rimbau A Medina-Remoacuten A Peacuterez-Jimeacutenez J Martiacutenez-GonzaacutelezMA Covas MI Corella D Salas-Salvadoacute J Goacutemez-Gracia E Lapetra J Aroacutes FFiol M Ros E Serra-Majem L Pintoacute X Muntildeoz MA Saez GT Ruiz-Gutieacuterrez VWarnberg J Estruch R Lamuela-Raventoacutes RM Dietary intake and major foodsources of polyphenols in a Spanish population at high cardiovascular riskthe PREDIMED study Nutr Metab Cardiovasc Dis 2013 23953ndash959

18 Willett WC Howe GR Kushi LH Adjustment for total energy intake inepidemiologic studies Am J Clin Nutr 1997 651220ndash1228

19 Adlercreutz H Lignans and human health Crit Rev Clin Lab Sci 200744483ndash525

20 Cassidy A Mukamal KJ Liu L Franz M Eliassen AH Rimm EB Highanthocyanin intake is associated with a reduced risk of myocardialinfarction in young and middle-aged women Circulation 2013 127188ndash196

21 Hooper L Kroon PA Rimm EB Cohn JS Harvey I Le Cornu KA Ryder JJ HallWL Cassidy A Flavonoids flavonoid-rich foods and cardiovascular risk ameta-analysis of randomized controlled trials Am J Clin Nutr 2008 8838ndash50

22 Spagnuolo C Russo M Bilotto S Tedesco I Laratta B Russo GL Dietarypolyphenols in cancer prevention the example of the flavonoidquercetin in leukemia Ann N Y Acad Sci 2012 125995ndash103

23 Williamson G Manach C Bioavailability and bioefficacy of polyphenols inhumans II Review of 93 intervention studies Am J Clin Nutr 200581243ndash255

24 Quintildeones M Miguel M Aleixandre A Beneficial effects of polyphenols oncardiovascular disease Pharmacol Res 2013 68125ndash131

Tresserra-Rimbau et al BMC Medicine Page 10 of 11

Additional file 1

Additional file 2

Additional file 3

13 May 2014

2014 1277httpwwwbiomedcentralcom1741-70151277

25 Hooper L Kay C Abdelhamid A Kroon PA Cohn JS Rimm EB Cassidy AEffects of chocolate cocoa and flavan-3-ols on cardiovascular health asystematic review and meta-analysis of randomized trials Am J Clin Nutr2012 95740ndash751

26 Kokubo Y Iso H Ishihara J Okada K Inoue M Tsugane S Japan PublicHealth Center Study Group Association of dietary intake of soy beansand isoflavones with risk of cerebral and myocardial infarctions inJapanese populations the Japan Public Health Center-based (JPHC)study cohort I Circulation 2007 1162553ndash2562

27 Kuriyama S Shimazu T Ohmori K Kikuchi N Nakaya N Nishino Y TsubonoY Tsuji I Green tea consumption and mortality due to cardiovasculardisease cancer and all causes in Japan the Ohsaki study JAMA 20062961255ndash1265

28 Sasazuki S Inoue M Miura T Iwasaki M Tsugane S Plasma tea polyphenolsand gastric cancer risk a casendashcontrol study nested in a largepopulation-based prospective study in Japan Cancer Epidemiol BiomarkersPrev 2008 17343ndash351

29 Valls-Pedret C Lamuela-Raventos RM Medina-Remoacuten A Quintana M Corella DPinto X Martiacutenez-Gonzaacutelez MA Estruch R Ros E Polyphenol-rich foods in theMediterranean diet are associated with better cognitive function in elderlysubjects at high cardiovascular risk J Alzheimers Dis 2012 29773ndash782

30 Arranz S Chiva-Blanch G Valderas-Martiacutenez P Medina-Remoacuten ALamuela-Raventos RM Estruch R Wine beer alcohol and polyphenols oncardiovascular disease and cancer Nutrients 2012 4759ndash781

31 Covas MI Nyyssonen K Poulsen HE Kaikkonen J Zunft HJ Kiesewetter HGaddi A la TR D Mursu J Baumler H Nascetti S Salonen JT Fito MVirtanen J Marrugat J The effect of polyphenols in olive oil on heartdisease risk factors a randomized trial Ann Intern Med 2006 145333ndash341

32 Mink PJ Scrafford CG Barraj LM Harnack L Hong CP Nettleton JA JacobsDR Jr Flavonoid intake and cardiovascular disease mortality aprospective study in postmenopausal women Am J Clin Nutr 200785895ndash909

33 Weseler AR Ruijters EJ Drittij-Reijnders MJ Reesink KD Haenen GR Bast APleiotropic benefit of monomeric and oligomeric flavanols on vascularhealth ndash a randomized controlled clinical pilot study PLoS One 20116e28460

34 Jennings A Welch AA Fairweather-Tait SJ Kay C Minihane AM ChowienczykP Jiang B Cecelja M Spector T Macgregor A Cassidy A Higher anthocyaninintake is associated with lower arterial stiffness and central blood pressurein women Am J Clin Nutr 2012 96781ndash788

35 Chuang CC Martinez K Xie G Kennedy A Bumrungpert A Overman A Jia WMcIntosh MK Quercetin is equally or more effective than resveratrol inattenuating tumor necrosis factor-alpha-mediated inflammation and insulinresistance in primary human adipocytes Am J Clin Nutr 2010 921511ndash1521

36 Cimino S Sortino G Favilla V Castelli T Madonia M Sansalone S Russo GIMorgia G Polyphenols key issues involved in chemoprevention ofprostate cancer Oxid Med Cell Longev 2012 2012632959

37 Stagos D Amoutzias GD Matakos A Spyrou A Tsatsakis AM Kouretas DChemoprevention of liver cancer by plant polyphenols Food ChemToxicol 2012 502155ndash2170

38 Lambert JD Hong J Yang GY Liao J Yang CS Inhibition of carcinogenesisby polyphenols evidence from laboratory investigations Am J Clin Nutr2005 81284ndash291

39 Muraki I Imamura F Manson JE Hu FB Willett WC van Dam RM Sun QFruit consumption and risk of type 2 diabetes results from threeprospective longitudinal cohort studies BMJ 2013 347f5001

40 Wedick NM Pan A Cassidy A Rimm EB Sampson L Rosner B Willett W HuFB Sun Q van Dam RM Dietary flavonoid intakes and risk of type 2diabetes in US men and women Am J Clin Nutr 2012 95925ndash933

41 Markus MA Morris BJ Resveratrol in prevention and treatment ofcommon clinical conditions of aging Clin Interv Aging 2008 3331ndash339

42 Hu FB Stampfer MJ Rimm E Ascherio A Rosner BA Spiegelman D WillettWC Dietary fat and coronary heart disease a comparison of approachesfor adjusting for total energy intake and modeling repeated dietarymeasurements Am J Epidemiol 1999 149531ndash540

Cite this article as Tresserra-Rimbau et al Polyphenol intake and mortalityrisk a re-analysis of the PREDIMED trial BMC Medicine

Submit your next manuscript to BioMed Centraland take full advantage of

bull Convenient online submission

bull Thorough peer review

bull No space constraints or color figure charges

bull Immediate publication on acceptance

bull Inclusion in PubMed CAS Scopus and Google Scholar

bull Research which is freely available for redistribution

Submit your manuscript at wwwbiomedcentralcomsubmit

Tresserra-Rimbau et al BMC Medicine Page 11 of 11

1011861741-7015-12-77

2014 1277

2014 1277httpwwwbiomedcentralcom1741-70151277