Journal of Alzheimer’s Disease 59 (2017) 815–849DOI 10.3233/JAD-170248IOS Press

815

Relationships of Dietary Patterns,Foods, and Micro- and Macronutrientswith Alzheimer’s Disease and Late-LifeCognitive Disorders: A Systematic Review

Vincenzo Solfrizzia,1,∗, Carlo Custoderoa, Madia Lozuponeb, Bruno P. Imbimboc,Vincenzo Valiania, Pasquale Agostia, Andrea Schilardia, Alessia D’Intronoa,Maddalena La Montagnad, Mariapaola Calvanib, Vito Guerrae, Rodolfo Sardonee,Daniela I. Abbresciae, Antonello Bellomod, Antonio Grecof , Antonio Danieleg,Davide Seripaf , Giancarlo Logroscinob,h, Carlo Sabbaa

and Francesco Panzab,f,h,1,∗aGeriatric Medicine-Memory Unit and Rare Disease Centre, University of Bari Aldo Moro, Bari, ItalybNeurodegenerative Disease Unit, Department of Basic Medicine, Neuroscience, and Sense Organs,University of Bari Aldo Moro, Bari, ItalycDepartment of Research and Development, Chiesi Farmaceutici, Parma, ItalydDepartment of Clinical and Experimental Medicine, Psychiatric Unit, University of Foggia, Foggia, ItalyeNational Institute for Digestive Diseases, IRCCS “Saverio de Bellis”, Castellana, Bari, Italyf Geriatric Unit and Laboratory of Gerontology and Geriatrics, Department of Medical Sciences,IRCCS “Casa Sollievo della Sofferenza”, San Giovanni Rotondo, Foggia, ItalygInstitute of Neurology, Catholic University of Sacred Heart, Rome, ItalyhDepartment of Clinical Research in Neurology, University of Bari Aldo Moro, “Pia FondazioneCardinale G. Panico”, Tricase, Lecce, Italy

Accepted 23 May 2017

Abstract. In the last decade, the association between diet and cognitive function or dementia has been largely investigated. Inthe present article, we systematically reviewed observational studies published in the last three years (2014–2016) on the rela-tionship among dietary factors and late-life cognitive disorders at different levels of investigation (i.e., dietary patterns, foodsand food-groups, and dietary micro- and macronutrients), and possible underlying mechanisms of the proposed associations.From the reviewed evidence, the National Institute on Aging–Alzheimer’s Association guidelines for Alzheimer’s disease(AD) and cognitive decline due to AD pathology introduced some evidence suggesting a direct relation between diet andchanges in the brain structure and activity. There was also accumulating evidence that combinations of foods and nutrientsinto certain patterns may act synergistically to provide stronger health effects than those conferred by their individual dietarycomponents. In particular, higher adherence to a Mediterranean-type diet was associated with decreased cognitive decline.

1These authors contributed equally to this work.∗Correspondence to: Vincenzo Solfrizzi, MD, PhD, Depart-

ment of Geriatric Medicine, Memory Unit and Rare DiseaseCenter, University of Bari Aldo Moro - Policlinico, Piazza GiulioCesare, 11, 70124 Bari, Italy. Tel.: +39 080 5473685; Fax: +39080 5478633; E-mails: vincenzo.solfrizzi@uniba.it and FrancescoPanza, MD, PhD, Geriatric Unit and Laboratory of Gerontology

and Geriatrics, Department of Medical Sciences, IRCCS “CasaSollievo della Sofferenza”, San Giovanni Rotondo, Foggia,Italy, and Neurodegenerative Disease Unit, Department of BasicMedicine, Neuroscience, and Sense Organs, University of Bari“Aldo Moro”, Bari, Italy. Tel./Fax: +39 0882 410271; E-mails:geriat.dot@uniba.it, f panza@hotmail.com

ISSN 1387-2877/17/$35.00 © 2017 – IOS Press and the authors. All rights reserved

816 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

Moreover, also other emerging healthy dietary patterns such as the Dietary Approach to Stop Hypertension (DASH) andthe Mediterranean-DASH diet Intervention for Neurodegenerative Delay (MIND) diets were associated with slower ratesof cognitive decline and significant reduction of AD rate. Furthermore, some foods or food groups traditionally consideredharmful such as eggs and red meat have been partially rehabilitated, while there is still a negative correlation of cognitivefunctions with saturated fatty acids and a protective effect against cognitive decline of elevated fish consumption, high intakeof monounsaturated fatty acids and polyunsaturated fatty acids (PUFA), particularly n-3 PUFA.

Keywords: Alzheimer’s disease, dementia, dietary pattern, food groups, foods, healthy diet, Mediterranean diet, macronutri-ents, micronutrients, mild cognitive impairment

INTRODUCTION

Alzheimer’s disease (AD) is the most commonform of dementia. Over 5.4 million people are esti-mated to be affected in the Unites States [1]. In thelast 20 years, most of the therapeutic approaches weredirected against the production and accumulation ofamyloid-� (A�), one of the neuropathological hall-marks of AD, and, up to now, have failed to showclinical efficacy [2].

Currently, available drugs for the treatment ofAD have only symptomatic effects [2], and thereis an unmet need of preventing AD onset anddelaying or slowing disease progression from mildcognitive impairment (MCI) in the absence ofdisease-modifying therapies. Therefore, at present,the management of potential risk factors is believedto be the most effective way of preventing dementia,AD, and MCI. In the last decade, one of the mostintriguing and appealing links hypothesized was theassociation between lifestyle factors such as diet anddietary habits and the occurrence of AD and dementia[3–7]. Dietary factors may affect the risk of cardio-vascular disease (CVD), also influencing the risk ofAD and dementia [8–10]. A growing body of evi-dence suggests that certain diets are associated witha lower incidence of AD, so maintaining a healthydiet may have an impact on many of these possiblerisk factors for cognitive decline, and the model tofollow seems to be the Mediterranean diet (MeDi)[4, 9]. Therefore, elevated dietary monounsaturatedfatty acids (MUFA) and n-3 polyunsaturated fattyacids (n-3 PUFA) and high fish consumption [10, 11],alongside high levels of antioxidants from fruit andvegetables [12, 13], and moderate alcohol and coffeeconsumption [14–16] may have a beneficial effecton the risk of dementia. Nonetheless, diet should beconsidered as a “whole”, consisting of a complexof nutritional principles, foods, micronutrients andmacronutrients interacting with each other. In fact,the combinations of foods and nutrients into certain

patterns may act synergistically to provide strongerhealth effects than those conferred by their individualdietary components. This hypothesis has addressedthe interest in the role of dietary patterns rather thanof the individual components of the diet against CVD,neurodegenerative diseases, and all-cause mortality[17, 18]. Some recent systematic reviews and meta-analyses of pooled studies on this issue found that ahigher adherence to the MeDi was associated with areduced risk of cognitive impairment, MCI and AD,as well as the transition from MCI to AD [19–22].However, in the last few years, some changes haveemerged in approaching the relationship between dietand cognitive impairment. In fact, the National Insti-tute on Aging–Alzheimer’s Association (NIA-AA)guidelines for AD and cognitive decline due to ADpathology [23] introduced some evidence suggest-ing a direct relation between diet and changes in thebrain structure and activity, opening the era of brainimaging biomarkers in nutritional epidemiology. Theaim of the present systematic review article was toshed light on the relationship among dietary patterns,foods, food groups, micro-, and macronutrients andlate-life cognitive disorders considering the resultsof observational studies published in the last threeyears (2014–2016), as well as possible underlyingmechanisms of the proposed associations.

METHODS

In the present systematic review article, we fol-lowed the Preferred Reporting Items for Systematicreviews and Meta-Analyses (PRISMA) guidelines,adhering to the PRISMA 27-item checklist [24].This systematic review was based upon searches ofUS National Library of Medicine (PubMed), OvidMEDLINE, EMBASE, Google Scholar, Web of Sci-ence, and Scopus databases, picking the followingterms to identify the risk exposure (dietary patternsOR foods OR food groups OR micronutrients ORmacronutrients) combined with terms to determine

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 817

the outcomes of interest [cognitive AND (impair-ment OR decline OR disorders) OR cognition ORAlzheimer’s disease OR dementia OR mild cognitiveimpairment OR vascular dementia]. There were nolanguage restrictions on the search. To be includedin this systematic review, studies were limited toobservational studies (including both cross-sectionaland longitudinal population-based studies) publishedbetween January 1, 2014 and December 31, 2016. Wechoose these time limits on the basis of a pre-searchwithout time limits that included a very high num-ber of identified articles to review (36104 articles).Studies were further required to meet the followinginclusion criteria: (1) studies conducted in humansaged 45 years or older; (2) studies that provideda description of the tools used for collecting theadherence to the different dietary patterns and theintake of foods, food groups, micro- and macronutri-ents (e.g., validated semi-quantitative food frequencyquestionnaires, 7-day dietary records, or 24-h dietaryrecall) or that evaluated nutrient consumption fromthe values of biochemical markers (e.g., serum con-centration or red blood cells levels); (3) studiesassessing cognitive functions according to the diag-nostic criteria for the diagnoses of MCI [Petersencriteria and their revision/modifications, Interna-tional Working Group on Mild Cognitive Impairmentcriteria, European Alzheimer’s Disease Consortium(EADC) criteria, NIA-AA criteria for MCI due toAD, and DSM-5 criteria for Mild NeurocognitiveDisorder), AD [National Institute of Neurologicaland Communicative Disorders and Stroke and theAlzheimer’s Disease and Related Disorders Associ-ation (NINCDS-ADRDA) criteria, NIA-AA criteriafor dementia due to AD, International Working Group(IWG)-1 criteria for AD, and IWG-2 criteria forAD], vascular dementia (VaD) (NINCDS-AIREN),unspecified dementia [Diagnostic and StatisticalManual for Mental Disorders (DSM)-III-R criteria,DSM-IV criteria, DSM-IV-TR criteria, DSM-5 crite-ria for Major Neurocognitive Disorder, InternationalClassification of Diseases (ICD), 9th Revision, Clin-ical Modification (CM), and ICD-10-CM] or thatprovided the neuropsychological tools used for defin-ing late-life cognitive impairment/decline also innon-demented older subjects. The studies includedhad to present original data.

Figure 1 shows the stages in obtaining studiesfor inclusion in the present report (PRISMA Four-phase Flow Diagram). From 9,074 articles identifiedwith multiple electronic searches, we screened titlesand abstracts of the citations downloaded from the

searches and identified 1,877 potential relevant arti-cles chosen for a closer review. We excluded 1,798articles not meeting inclusion criteria and obtainedfull copies of the 79 potentially suitable reports forfurther assessment. After inclusion of 4 articles ofinterest from the reference lists of the selected arti-cles and exclusion of another 36 articles, 47 studiesmet study eligibility criteria and were finally includedin the overall systematic review (Tables 1–3) [25–71].We did not use formal methods of assessment of thequality of the studies included in the present sys-tematic review, given the lack of randomized clinicaltrials, but we described in depth study design andsample together with dietary and cognitive assess-ment and principal results of the included studiesin Tables 1–3 [25–71]. Finally, we used a narrativesynthesis to summarize the findings of the includedstudies, subdividing the articles for the four prin-cipal diet-based approaches (dietary patterns, foodsand food groups, micronutrients, and macronutri-ents), specifying sample size and study design andthe cognitive outcomes of the included studies (late-life cognitive impairment/decline or dementia, AD,and non-AD dementias), when possible.

RESULTS

Dietary patterns and late-life cognition

Table 1 shows selected observational studies pub-lished in the last three years that evaluate theassociation of dietary patterns with late-life cogni-tive disorders [25–41]. Among these diets, MeDi isa typical dietary pattern of Mediterranean countries,characterized by high consumption of fruits, vegeta-bles, legumes and cereals, olive oil as the main addedlipid, moderate consumption of alcohol (mainly wineand during meals) and low consumption of red meatand dairy products. It is doubtless the most analyzeddietary pattern, and accumulating evidence supporta potential protective role against cognitive declineand dementia, although there are still inconsisten-cies in the reported data. In particular, the findingsfrom prospective studies and very recent systematicreviews and meta-analyses suggested that adherenceto the MeDi fulfilling the whole-diet approach mayaffect not only the risk of AD, but also of predemen-tia syndromes and their progression to overt dementia[19]. A cross-sectional analysis from the US NationalHealth and Nutrition Survey (NHANES) 1999–2002and from the Israeli National Health and NutritionSurvey (MABAT ZAHAV) 2005–2006 suggested

818 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

Fig. 1. PRISMA Four-phase Flow Diagram of retrieved and selected articles showing associations among dietary patterns, foods, foodgroups, micro- and macronutrients, and late-life cognitive disorders considering the results of observational studies published in the lastthree years (2014–2016).

that cognitive and physical functions were signifi-cantly better among older subjects with the higheradherence to the MeDi [25] (Table 1). In the Spanishprospective cohort of the SUN project, in an 8-yearfollow-up, a higher cognitive decline was observedamong participants with low or moderate baselineadherence to the MeDi than among those with betteradherence [26] (Table 1). In the European Prospec-tive Investigation into Cancer and Nutrition (EPIC)study, in a cohort of Greek elderly population thatstill adhered to the traditional MeDi, it was demon-strated that closer adherence to MeDi was associatedwith less decline in Mini-Mental State Examination(MMSE) performance over a period of about 7 years,especially in individuals aged 75 years or older [27](Table 1). The association between dietary patternsand cognitive decline was also recently investigated

in non-Western populations [28, 29] (Table 1). Infact, in a prospective Chinese cohort study on 1,650Chinese adults ≥55 years of age, the subjects inthe highest tertile of a modified MeDi score or awheat-based diverse diet with similar features of theadapted MeDi had a slower rate of cognitive decline[28]. Moreover, in the population-based AustralianImaging, Biomarkers and Lifestyle Study of Age-ing, higher adherence to the Australian-style MeDiwas associated with better performance in the exec-utive function, while higher Western diet adherencewas associated with greater decline in the visuospatialcognitive domain [29].

Other emerging dietary patterns are the DietaryApproach to Stop Hypertension (DASH) and theMediterranean-DASH diet Intervention for Neurode-generative Delay (MIND) diets. The DASH diet is

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 819Ta

ble

1O

bser

vatio

nals

tudi

eson

the

rela

tions

hip

amon

gdi

etar

ypa

ttern

san

dla

te-l

ife

cogn

itive

diso

rder

s(2

014–

2016

)

Ref

eren

ces

Stud

yD

esig

nSa

mpl

eO

utco

me

Cog

nitiv

ean

dPr

inci

palR

esul

tsN

utri

tiona

lAss

essm

ent

Die

tary

Patt

erns

Zbe

ida

etal

.,[2

5]C

ross

-sec

tiona

lstu

dyn

=4,

577

subj

ects

,2,7

91fr

omN

HA

NE

San

d1,

786

from

MA

BA

TZ

AH

AV

(mea

nag

e71

.2an

d74

.9y,

resp

ectiv

ely)

Ass

ocia

tion

betw

een

adhe

renc

eto

the

MeD

iand

cogn

itive

and

phys

ical

func

tion.

Cog

nitiv

ean

dfu

nctio

nal

asse

ssm

ent;

MD

S.C

ogni

tive

and

phys

ical

func

tions

sign

ifica

ntly

bette

ram

ong

the

high

est

MD

S.In

part

icul

ar,M

DS

(hig

hve

rsus

low

)w

asas

soci

ated

with

few

erdi

sabi

litie

s(O

R:0

.51,

95%

Cl:0

.28–

0.93

).G

albe

teet

al.,

[26]

Pros

pect

ive

coho

rtst

udy

n=

823

part

icip

ants

(mea

nag

e62

y)A

ssoc

iatio

nbe

twee

nad

here

nce

toth

eM

eDia

ndco

gniti

vefu

nctio

n.

FFQ

(136

item

s);M

DS;

TIC

S-m

.H

ighe

rco

gniti

vede

clin

eam

ong

part

icip

ants

with

low

orm

oder

ate

base

line

adhe

renc

eto

the

MeD

i(a

djus

ted

diff

eren

ce=

–0.5

6po

ints

inT

ICS-

m,9

5%C

I:–0

.99

to–0

.13)

.

Follo

w-u

p:8

y

Tri

chop

oulo

uet

al.,

[27]

Pros

pect

ive

coho

rtst

udy

n=

401

olde

rsu

bjec

tsfr

omE

PIC

-Gre

ece

coho

rt(m

ean

age

74y)

Ass

ocia

tion

ofM

eDio

ran

ypa

rtic

ular

MeD

icom

pone

ntw

ithco

gniti

vede

clin

e.

FFQ

(150

item

s);M

DS;

MM

SE;M

MSE

chan

ge(c

MM

SE).

Dec

line

inM

MSE

perf

orm

ance

inve

rsel

yas

soci

ated

with

adhe

renc

eto

trad

ition

alM

eDi.

For

subs

tant

ialv

ersu

sno

decl

ine

(hig

hve

rsus

low

MD

S),O

R:0

.34,

95%

CI:

0.13

–0.8

9).O

nly

vege

tabl

eco

nsum

ptio

n,si

gnifi

cant

inve

rse

asso

ciat

ion

with

cogn

itive

decl

ine.

Follo

w-u

p:av

erag

e6.

6y

Qin

etal

.,[2

8]Pr

ospe

ctiv

eco

hort

stud

yn

=16

50ad

ults

≥55

yA

ssoc

iatio

nbe

twee

nad

apte

dM

DS

orfa

ctor

anal

ysis

deri

ved

diet

ary

patte

rns

with

cogn

itive

chan

ges.

Glo

bala

ndco

mpo

site

cogn

itive

scor

es.B

asel

ine

diet

by24

hre

calls

over

3da

ys.A

dapt

edM

DS

and

two

diet

ary

patte

rnsc

ores

.

An

adap

ted

MeD

ior

aw

heat

-bas

eddi

vers

edi

etw

ithsi

mila

rco

mpo

nent

sof

the

MeD

imay

redu

ceth

era

teof

cogn

itive

decl

ine

amon

gad

ults

≥65

y(�

:0.

042

SU/y

,95%

CI:

0.00

2–0.

081

and

�:0

.069

SU/y

,95%

CI:

0.02

3–0.

114)

Follo

w-u

p:5

y

(Con

tinu

ed)

820 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older AgeTa

ble

1(C

onti

nued

)

Ref

eren

ces

Stud

yD

esig

nSa

mpl

eO

utco

me

Cog

nitiv

ean

dPr

inci

palR

esul

tsN

utri

tiona

lAss

essm

ent

Gar

dene

ret

al.,

[29]

Pros

pect

ive

coho

rtst

udy

n=

527

heal

thy

part

icip

ants

Ass

ocia

tion

ofth

ree

wel

l-re

cogn

ized

diet

ary

patte

rns

with

cogn

itive

chan

ge.

Can

cer

Cou

ncil

ofV

icto

ria

FFQ

atba

selin

e.C

ompo

site

scor

esfo

rsi

xco

gniti

vedo

mai

nsan

da

glob

alco

gniti

vesc

ore;

Aus

MeD

isc

ore;

Wes

tern

diet

scor

e;Pr

uden

tdie

tsco

re.

Hig

her

adhe

renc

eto

the

Aus

MeD

iass

ocia

ted

with

bette

rpe

rfor

man

cein

the

exec

utiv

efu

nctio

n(8

.1%

ofth

eva

rian

ceex

plai

ned)

.H

ighe

rW

este

rndi

etad

here

nce

asso

ciat

edw

ithgr

eate

rde

clin

ein

the

visu

ospa

tialc

ogni

tive

dom

ain

(3.6

%of

the

vari

ance

expl

aine

d).

Follo

w-u

p:3

y

Exc

lude

dA

Dpa

tient

san

dM

CI

subj

ects

(dia

gnos

edw

ithPe

ters

encr

iteri

a).

Tang

ney

etal

.,[3

0]Pr

ospe

ctiv

eco

hort

stud

yn

=82

6ol

der

pers

ons

(mea

nag

e81

.5y)

Ass

ocia

tion

betw

een

DA

SHdi

etor

Med

Die

tSco

rean

dco

gniti

vede

clin

e.

MA

PFF

Q(1

44ite

m);

DA

SHdi

et;M

edD

ietS

core

;Glo

bal

com

posi

tesc

ore

of19

cogn

itive

test

s.

Asl

ower

rate

ofgl

obal

cogn

itive

decl

ine

was

asso

ciat

edw

ithD

ASH

(by

0.00

7SU

)an

dM

eDi(

by0.

002

SU)

patte

rns.

Follo

w-u

p:m

ean

of4.

1y

Mor

ris

etal

.,[3

1]Pr

ospe

ctiv

eco

hort

stud

yn

=92

3pa

rtic

ipan

ts,a

ges

58to

98y

Ass

ocia

tion

ofM

IND

diet

,D

ASH

diet

and

Med

Die

tSco

rew

ithin

cide

ntA

D.

AD

diag

nosi

sat

each

annu

alev

alua

tion.

Hig

had

here

nce

toal

lthr

eedi

ets

may

redu

ceA

Dri

sk(M

IND

=H

R:0

.47,

95%

CI:

0.26

–0.7

6;D

ASH

=H

R:0

.61,

95%

CI:

0.38

–0.9

7;M

eDi=

HR

:0.

46,9

5%C

I:0.

26–0

.79)

.M

oder

ate

adhe

renc

eto

the

MIN

Ddi

etm

ayal

sode

crea

seA

Dri

sk(H

R:

0.65

,95%

CI:

0.44

–0.9

8)

Follo

w-u

p:m

ean

of4.

5y

FFQ

(144

item

s);M

IND

diet

scor

e;D

ASH

diet

scor

e;M

edD

ietS

core

.

Mor

ris

etal

.,[3

2]Pr

ospe

ctiv

eco

hort

stud

yn

=96

0pa

rtic

ipan

ts(m

ean

age

81.4

y)A

ssoc

iatio

nof

MIN

Ddi

etsc

ore

with

cogn

itive

decl

ine.

Ann

ualc

ogni

tive

asse

ssm

ents

,glo

bala

ndco

mpo

site

scor

esof

5do

mai

ns.M

CI

subj

ects

(dia

gnos

ispa

rtly

base

don

Pete

rsen

crite

ria)

not

excl

uded

,exc

epti

nse

cond

ary

anal

yses

.FFQ

(144

item

s)at

each

annu

alcl

inic

alev

alua

tion;

MIN

Ddi

etsc

ore;

DA

SHdi

etsc

ore;

Med

Die

tSco

re.

The

MIN

Dsc

ore

posi

tivel

yas

soci

ated

with

slow

erde

clin

ein

glob

alco

gniti

vesc

ore

and

with

each

offiv

eco

gniti

vedo

mai

ns.T

heM

IND

diet

scor

e(�

coef

ficie

nt:4

.39)

mor

epr

edic

tive

ofco

gniti

vede

clin

eth

anei

ther

ofth

eD

ASH

diet

scor

e(�

coef

ficie

nt:2

.60)

and

the

Med

Die

tSco

re(�

coef

ficie

nt:2

.46)

.

Follo

w-u

p:m

ean

of4.

7y

Com

pare

dth

ees

timat

edef

fect

sof

MIN

Ddi

etto

thos

eof

the

Med

Die

tSco

rean

dD

ASH

diet

.

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 821M

osco

niet

al.,

[33]

Cro

ss-s

ectio

nals

tudy

n=

52cl

inic

ally

and

cogn

itive

lyno

rmal

subj

ects

(mea

nag

e54

y)

Ass

ocia

tions

betw

een

adhe

renc

eto

aM

eDia

ndst

ruct

ural

MR

I-ba

sed

brai

nat

roph

yin

key

regi

ons

for

AD

.

Sem

iqua

ntita

tive

FFQ

(61-

item

);C

DR

=0;

MD

S;M

RI

CT

mea

sure

sfo

r5

RO

Is.

Subj

ects

with

high

erM

eDi

adhe

renc

esh

owed

grea

ter

CT

ofA

D-v

ulne

rabl

eR

OIs

asco

mpa

red

tosu

bjec

tsw

ithlo

wer

MeD

iad

here

nce.

Ber

tiet

al.,

[34]

Cro

ss-s

ectio

nals

tudy

n=

52cl

inic

ally

and

cogn

itive

lyno

rmal

subj

ects

(mea

nag

e54

y)

Ass

ocia

tions

ofnu

trie

ntpa

ttern

s(N

P)w

ithm

ajor

brai

nA

Dbi

omar

kers

.Fiv

eN

PSw

ere

iden

tified

.

Sem

iqua

ntita

tive

FFQ

(116

-ite

m);

CD

R=

0;Pi

B-

and

FD-

PET

scan

sM

RI

mea

sure

s.

NP4

scor

es(v

itam

ins

B12

,D,

and

zinc

)w

ere

posi

tivel

yas

soci

ated

with

ME

Tgl

can

dT

GM

V,a

ndne

gativ

ely

asso

ciat

edw

ithPi

Bre

tent

ion

inA

D-v

ulne

rabl

ere

gion

s.M

ET

glc

and

TG

MV

wer

epo

sitiv

ely

asso

ciat

edw

ithN

P2sc

ores

(vita

min

E.M

UFA

,and

PUFA

),an

dne

gativ

ely

asso

ciat

edw

ithN

P5sc

ores

(SFA

,tra

nsun

satu

rate

dfa

ttyac

ids,

chol

este

rola

ndso

dium

),an

dM

ET

glc

was

posi

tivel

yas

soci

ated

with

high

erN

P3sc

ores

(vita

min

A,v

itam

inC

,car

oten

oids

,an

ddi

etar

yfib

ers)

.St

aubo

etal

.,[3

5]C

ross

-sec

tiona

lstu

dyn

=67

2co

gniti

vely

norm

alpa

rtic

ipan

ts(m

ean

age:

79.8

y)

Ass

ocia

tion

ofM

DS

and

MeD

icom

pone

nts

with

MR

Im

easu

res

ofC

Tfo

rth

efo

urlo

bes

sepa

rate

lyan

dav

erag

ed.

FFQ

(128

item

s);M

DS;

MR

IC

Tm

easu

res.

Exc

lude

dpa

tient

sw

ithde

men

tiaan

dM

CI

subj

ects

(dia

gnos

edw

ithPe

ters

encr

iteri

a).

Hig

her

MD

Sas

soci

ated

with

larg

erC

T.H

ighe

rle

gum

e,fis

h,ve

geta

bles

,who

legr

ains

orce

real

sin

take

sw

ere

asso

ciat

edw

ithla

rger

CT.

Gu

etal

.,[3

6]C

ross

-sec

tiona

lstu

dyn

=67

4el

derl

yad

ults

with

out

dem

entia

(mea

nag

e80

.1y)

Ass

ocia

tion

betw

een

high

erad

here

nce

toM

eDiw

ithla

rger

MR

Im

easu

red

brai

nvo

lum

eor

CT.

FFQ

;MD

S;M

RI

scan

sfo

rT

BV

,TG

MV

,TW

MV

and

mC

T.

Hig

her

MeD

iadh

eren

ceas

soci

ated

with

13.1

1,5.

00,

and

6.41

mm

larg

erT

BV

,T

GM

Van

dT

WM

V,

resp

ectiv

ely.

Hig

her

fish

inta

keas

soci

ated

with

larg

erT

GM

Van

dm

CT.

Low

erm

eati

ntak

eas

soci

ated

with

larg

erT

GM

Van

dT

BV

.

Exc

lude

dpa

tient

sw

ithde

men

tiaan

dM

CI

subj

ects

(dia

gnos

edw

ithPe

ters

encr

iteri

a).

(Con

tinu

ed)

822 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

Tabl

e1

(Con

tinu

ed)

Ref

eren

ces

Stud

yD

esig

nSa

mpl

eO

utco

me

Cog

nitiv

ean

dPr

inci

palR

esul

tsN

utri

tiona

lAss

essm

ent

Jack

aet

al.,

[37]

Pros

pect

ive

coho

rtst

udy

n=

255

olde

rad

ults

(mea

nag

e62

.6y)

Ass

ocia

tion

betw

een

diet

ary

patte

rns

and

hipp

ocam

pal

volu

me.

Ass

ocia

tion

betw

een

diet

and

diff

eren

tialr

ates

ofhi

ppoc

ampa

latr

ophy

over

time.

FFQ

;“Pr

uden

t”(h

ealth

y)di

etan

d“W

este

rn”

(unh

ealth

y)di

et;T

wo

MR

Isc

ans.

Eve

ryon

eSD

incr

ease

ina

“pru

dent

”di

etar

ypa

ttern

was

asso

ciat

edw

itha

45.7

mm

3(S

E22

.9m

m3)

larg

erle

fthi

ppoc

ampa

lvo

lum

e,w

hile

high

erco

nsum

ptio

nof

a“W

este

rn”

diet

ary

patte

rnw

asas

soci

ated

with

a52

.6m

m3

(SE

26.6

mm

3)

smal

ler

left

hipp

ocam

pal

volu

me.

No

evid

ence

that

diet

ary

patte

rns

influ

ence

dhi

ppoc

ampa

lvol

ume

decl

ine.

Follo

w-u

p:4

y

Pelle

tier

etal

.,[3

8]Pr

ospe

ctiv

eco

hort

stud

yn

=14

6no

n-de

men

ted

part

icip

ants

(mea

nag

e73

y)

Ass

ocia

tion

betw

een

high

erad

here

nce

toth

eM

eDia

ndpr

eser

ved

brai

nG

Mvo

lum

ean

dW

Mm

icro

stru

ctur

e.

FFQ

(148

item

s);M

DS;

MR

I;B

rain

GM

and

WM

volu

mes

,and

WM

mic

rost

ruct

ure;

Cog

nitiv

eas

sess

men

t.

Adh

eren

ceto

the

MeD

isi

gnifi

cant

lyas

soci

ated

with

pres

erve

dW

Mm

icro

stru

ctur

ein

exte

nsiv

ear

eas,

aga

inin

stru

ctur

alco

nnec

tivity

rela

ted

tost

rong

cogn

itive

bene

fits.

Follo

w-u

p:M

RI

perf

orm

edm

ean

of8.

9y

afte

rdi

etar

yas

sess

men

t

Exc

lude

dpa

tient

sw

ithde

men

tia.

Shak

ersa

inet

al.,

[39]

Popu

latio

n-ba

sed

long

itudi

nals

tudy

n=

2,22

3de

men

tia-f

ree

olde

rad

ults

(mea

nag

e70

.6y)

Impa

ctof

diet

ary

patte

rns

onco

gniti

vede

clin

e.M

MSE

;Sem

i-qu

antit

ativ

eFF

Q(9

8ite

ms)

;Tw

odi

etar

ypa

ttern

s:1)

the

“Wes

tern

”,2)

the

“pru

dent

”;fa

ctor

scor

esfo

rea

chdi

etar

ypa

ttern

cate

gori

zed

into

quin

tiles

.E

xclu

ded

patie

nts

with

dem

entia

.

Hig

hest

adhe

renc

eto

prud

ent

patte

rnre

late

dto

less

MM

SEde

clin

e(�

:0.1

06),

whe

reas

the

high

est

adhe

renc

eto

Wes

tern

patte

rnw

asas

soci

ated

with

mor

eM

MSE

decl

ine

(�:0

.156

).D

eclin

eas

soci

ated

with

Wes

tern

diet

,atte

nuat

edby

high

adhe

renc

eto

prud

ent

patte

rn.

Follo

w-u

p:6

y

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 823

Smyt

het

al.,

[40]

Pros

pect

ive

coho

rtst

udy

n=

27,8

60pa

tient

s(m

ean

age

66.2

y)A

ssoc

iatio

nof

diet

ary

fact

ors

and

cogn

itive

decl

ine

ina

popu

latio

nat

high

risk

ofca

rdio

vasc

ular

dise

ase.

MM

SE;F

FQ(2

0ite

ms)

;m

AH

EI.

Hig

hest

quin

tile

ofm

AH

EI

(hea

lthie

stdi

et)

asso

ciat

edw

itha

redu

ced

risk

ofco

gniti

vede

clin

e(H

R:

0.76

,95%

CI:

0.66

–0.8

6).

Follo

w-u

p:56

mon

ths

Oza

wa

etal

.,[4

1]Pr

ospe

ctiv

eco

hort

stud

yn

=5,

083

patie

nts

(mea

nag

e56

y)In

vest

igat

edi

etar

ypa

ttern

sas

soci

ated

with

infla

mm

atio

n.

Alic

eH

eim

4-I,

shor

t-te

rmve

rbal

mem

ory,

phon

emic

and

sem

antic

fluen

cy,

MM

SE.F

FQ(1

27ite

m).

Seru

mIL

-6.I

DP.

Die

tary

patte

rnw

ithhi

gher

inta

keof

red

and

proc

esse

dm

eat,

peas

,leg

umes

and

frie

dfo

od,a

ndlo

wer

inta

keof

who

legr

ains

asso

ciat

edw

ithhi

gher

infla

mm

ator

ym

arke

rsan

dac

cele

rate

dco

gniti

vede

clin

e.In

part

icul

ar,t

here

was

agr

eate

rde

clin

ein

reas

onin

gin

part

icip

ants

inth

ehi

ghes

tter

tile

ofad

here

nce

toth

eID

P(–

0.37

SD,9

5%C

I:–0

.40,

–0.3

4)co

mpa

red

toth

ose

inth

elo

wes

tter

tile

(–0.

31SD

;95%

CI:

–0.3

4,–0

.28)

afte

rad

just

men

tfor

cova

riat

es.

Follo

w-u

p:10

y

Ass

ocia

tion

ofsu

chdi

etw

ithco

gniti

vede

clin

e.

NH

AN

ES,

Nat

iona

lHea

lthan

dN

utri

tion

Surv

ey;M

AB

AT

ZA

HA

V,I

srae

liN

atio

nalH

ealth

and

Nut

ritio

nSu

rvey

;MeD

i,M

edite

rran

ean

diet

;MD

S,M

edite

rran

ean

diet

scor

e;O

R,o

dds

ratio

;CI,

confi

denc

ein

terv

al;

FFQ

,Foo

dFr

eque

ncy

Que

stio

nnai

re;

TIC

S-m

,Tel

epho

neIn

terv

iew

ofC

ogni

tive

Stat

us-m

odifi

ed;

MM

SE,M

ini-

Men

tal

Stat

eE

xam

inat

ion;

SU,s

tand

ard

units

;A

usM

eDi,

Aus

tral

ian-

styl

eM

edite

rran

ean

diet

;DA

SH,D

ieta

ryA

ppro

ach

toSt

opH

yper

tens

ion;

MA

P,M

emor

yan

dA

ging

Proj

ect;

AD

,Alz

heim

er’s

dise

ase;

MIN

D,M

edite

rran

ean-

DA

SHdi

etIn

terv

entio

nfo

rN

euro

dege

nera

tive

Del

ay;C

DR

,Clin

ical

Dem

entia

Rat

ing

scal

e;H

R,h

azar

dra

tio;M

RI,

mag

netic

reso

nanc

eim

agin

g;C

T,co

rtic

alth

ickn

ess;

RO

I,re

gion

ofin

tere

st;P

ET,

posi

tron

emis

sion

tom

ogra

phy;

PiB

,11

C-P

ittsb

urgh

com

poun

d-B

;FD

G,

18F-

fluor

odeo

xygl

ucos

e;M

ET

glc,

gluc

ose

met

abol

ism

;T

GM

V,

tota

lgr

aym

atte

rvo

lum

e;M

UFA

,m

onou

nsat

urat

edfa

ttyac

ids;

PUFA

,po

lyun

satu

rate

dfa

ttyac

ids;

SFA

,sat

urat

edfa

ttyac

ids;

TB

V,t

otal

brai

nvo

lum

e;T

WM

V,t

otal

whi

tem

atte

rvo

lum

e;m

CT,

mea

nco

rtic

alth

ickn

ess;

SD,s

tand

ard

devi

atio

n;SE

,sta

ndar

der

ror;

GM

,gra

ym

atte

r;W

M,w

hite

mat

ter;

mA

HE

I,m

odifi

edA

ltern

ativ

eH

ealth

yE

atin

gIn

dex;

IL-6

,int

erle

ukin

-6;I

DP,

infla

mm

ator

ydi

etar

ypa

ttern

.

824 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

characterized by low consumption of saturated fat andcommercial pastries and sweets, and higher intake ofdairy than in the MeDi. In the last three years, in theMemory and Aging Project (MAP) study, a prospec-tive study on older adults with 4 years of follow-up,the DASH pattern was associated with slower ratesof cognitive decline. In particular, a 1-unit-higherDASH score, was equivalent of being at least 4.4years younger [30] (Table 1). These results were inline with those of Morris and colleagues, in the sameMAP study, in which higher adherence to DASH dietwas related with greater reduction of incident ADrather than higher adherence to MeDi (54% and 39%reduction, respectively) [31] (Table 1).

The MIND diet was based on the dietary compo-nents of the MeDi and DASH diet with modificationsthat highlight the foods and nutrients shown to beassociated with dementia prevention. Among theMIND diet components, there are 10 brain healthyfood groups (green leafy vegetables, other vegetables,nuts, berries, beans, whole grains, seafood, poultry,olive oil, and wine) and five unhealthy food groups(red meats, butter and stick margarine, cheese, pas-tries and sweets, and fried/fast food). Hence, MINDdiet uniquely specifies consumption of berries andgreen leafy vegetables and does not specify high fruitconsumption (both DASH and MeDi), high dairy(DASH), high potato consumption, or >1 fish meal perweek (MeDi). Other recent findings from the MAPstudy suggested that higher MIND diet score wasassociated with slower decline in cognitive abilities[32] (Table 1). The rate reduction for persons in thehighest tertile of diet scores compared with the lowesttertile was the equivalent of being 7.5 years younger.MIND diet score was also more predictive of cogni-tive decline than either of the other diet scores (DASHand MeDi) [32]. Furthermore, in a follow-up of 4.5years of the MAP study, participants with higher andmoderate adherence to MIND diet had statisticallysignificant reduction in AD rate compared with thosewith lower adherence (53% and 35% respectively)[31] (Table 1). Instead, only the highest tertiles of theDASH and MeDi scores were significantly associatedwith incident AD reduction [31].

Despite the promising results of these two diets,to date, we have brain imaging data only on thecorrelation with the MeDi (Table 1). The fewcross-sectional studies carried out on cognitively nor-mal people showed that higher adherence to MeDiwas related to greater magnetic resonance imaging(MRI)-based cortical thickness in AD-vulnerableregions and larger brain volumes. MeDi effects on

MRI biomarkers were significant in the left, but notin the right hemisphere, and were most pronouncedin entorhinal cortex, orbito-frontal cortex and pos-terior cingulate cortex [33] (Table 1). In the samesample of cognitively normal people, 5 differentnutrients patterns (NPs) were identified and associ-ated with major brain AD biomarkers. NP4 scores(vitamins B12, D, and zinc) were positively asso-ciated with glucose metabolism investigated with18F-fluorodeoxyglucose positron emission tomog-raphy (PET) scans and MRI-based total graymatter volume, and negatively associated with PET-11CPittsburgh compound-B (PiB) retention in AD-vulnerable regions. Glucose metabolism and totalgray matter volume were also positively associatedwith NP2 scores (vitamin E, MUFA, and PUFA),and negatively associated with NP5 scores [satu-rated fatty acids (SFA), trans unsaturated fatty acids,cholesterol and sodium], and glucose metabolism waspositively associated with higher NP3 scores (vita-min A, vitamin C, carotenoids, and dietary fibers)[34] (Table 1). Higher adherence to a Mediterraneandietary pattern was associated with larger MRI mea-sures of cortical thickness and with several individualregion of interests (ROIs) that undergo age-relatedor AD-related neurodegeneration, was marginallyassociated with temporal and AD signature corticalthickness, and was not associated with hippocampalvolume [35] (Table 1). This finding may be explainedwith the observation from the Alzheimer’s DiseaseNeuroimaging Initiative in which presymptomaticindividuals had significantly reduced cortical thick-ness in AD-vulnerable regions compared to controlsbut did not differ in regard to hippocampal volume[72]. In the Washington Heights–Inwood CommunityAging Project (WHICAP), higher MeDi adherencewas associated with less brain atrophy (larger totalbrain volume, total gray matter volume, total whitematter volume), with an effect similar to 5 years ofaging [36] (Table 1).

In the last three years, only one prospectiveimaging-diet study on older adults was conducted[37] (Table 1), confirming other results coming fromcross-sectional studies. In fact, Jacka and colleagues,in the Personality and Total Health Through LifeStudy, found that a healthy “prudent” dietary pat-tern characterized by intake of fresh vegetables,salad, fruit and grilled fish was associated with alarger left hippocampal volume on MRI over 4 yearsof follow-up [37] (Table 1). In particular, every onestandard deviation (SD) increase in healthy “pru-dent” dietary pattern was associated with a 45.7 mm3

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 825

larger left hippocampal volume [37]. While higherconsumption of an unhealthy Western dietary pat-tern characterized by intake roast meat, sausages,hamburgers, steak, chips, crisps, and soft drinks wasindependently associated with a 52.6 mm3 smallerleft hippocampal volume [37]. The difference in hip-pocampal volume between those classified with ahealthy and or unhealthy diet was 203 mm3, a dif-ference which corresponds to 62% of the averagedecline in left hippocampal volume observed overthe 4-year period. It was found no interaction betweenright hippocampus volumes and the two dietary factorscores [37]. Other studies suggested a strong impactof healthy diets on structural connectivity in oldersubjects, rather than gray and white matter volumes.In fact, through diffusion tensor imaging (DTI) atMRI examination was seen that higher adherenceto the MeDi was associated with preserved whitematter microstructure in multiple brain areas andappeared to delay cognitive aging by up to 10 years[38] (Table 1). None of the individual componentswas strongly associated with DTI parameters, sup-porting the hypothesis that overall diet quality maybe more important to preserve brain structure than anysingle food. These results suggested the involvementof vascular pathways rather than neurodegenerativemechanisms in the link between the MeDi and lowerrisks of cognitive decline and related diseases [38].

The importance of components of prudent dietarypattern (vegetables, fruit, cooking/dressing oil, cere-als and legumes, whole grains, rice/pasta, fish, low-fatdairy, poultry and water) was confirmed by theobservation that the MMSE decline associated withWestern diet may be attenuated by high adherenceto prudent pattern [39] (Table 1). In fact, the declinebecame less pronounced (53.5%) and non-significantamong people who had a high adherence to both theprudent and Western patterns. Furthermore, West-ern dietary pattern score was significantly associatedwith all-cause mortality in the older age cohorts [39].Instead, people followed healthiest diet was slightlyolder, more active, less likely to smoke, had a lowerbody mass index, normal serum creatinine, and hadhigher MMSE score [40] (Table 1). The healthi-est diet was associated with a reduction of about24% in risk of cognitive decline and in particularwas shown a significant association between higherdiet quality and reduced risk of decline in 4 com-ponents of the MMSE including copying, attentionand calculation, registration, and writing [40]. Thebrain damage related to an unhealthy diet may bebased on a pro-inflammatory mechanism. Ozawa and

colleagues detected an inflammatory dietary patterncharacterized by higher intake of red meat, processedmeat, peas and legumes, and fried food, and lowerintake of whole grains which correlated with elevatedinterleukin (IL)-6 [41] (Table 1). It was related withgreater decline in reasoning and in global cognitionand, in a cross-sectional analysis at baseline, a twotimes greater risk of having a decline of 3 points ormore in MMSE [41].

Foods, food groups, and late-life cognitivedisorders

Fish and seafoodTable 2 shows selected observational studies pub-

lished in the last three years and evaluating therelationship among foods, food-groups, and late-lifecognitive disorders [42–54]. In particular, the emerg-ing data from the last studies on the correlationbetween fish and seafood consumption and cogni-tive decline are conflicting. Significant correlationswere found in some particular population subgroups[ ≥65 years and apolipoprotein E (APOE) �4 car-riers]. Age significantly modified the associationbetween fish consumption and cognitive change [42](Table 2). In fact, no association was observed amongadults aged 55–64 years. Conversely, adults aged ≥65years, that consuming ≥1 servings/week fish (i.e.,100 g) had a reduction of cognitive decline rate[42]. Compared with individuals who consumed <1serving/week fish, the mean annual rate of global cog-nitive decline was reduced by 0.35-point equivalent tothe disparity associated with 1.6 years of age. Remov-ing shellfish and/or preserved fish from the total fishdid not appreciably alter the results [42].

Interestingly, Morris and colleagues showed that,in APOE �4 carriers, seafood consumption≥β1meals/week was correlated with lesser burden ofbrain AD neuropathology, including lower densityof neuritic plaques, less severe and widespread neu-rofibrillary tangles, and lower neuropathologicallydefined AD [43] (Table 2). Furthermore, some studiesdemonstrated an association between fish consump-tion and MRI biomarkers [34, 35] (Table 1). Inthe Mayo Clinic Study of Aging, higher fish intakewas associated with larger cortical thickness sum-mary measures for parietal and average lobar corticalthickness and marginally associated with AD signa-ture cortical thickness, temporal and frontal corticalthickness, and also associated with several individ-ual cortical thickness measures: precuneus, superiorparietal, posterior cingulate, supramarginal, middle

826 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

Tabl

e2

Obs

erva

tiona

lstu

dies

onth

ere

latio

nshi

pam

ong

food

s,fo

od-g

roup

s,an

dla

te-l

ife

cogn

itive

diso

rder

s(2

014–

2016

)

Ref

eren

ces

Stud

yD

esig

nSa

mpl

eO

utco

me

Cog

nitiv

ean

dN

utri

tiona

lA

sses

smen

tPr

inci

palR

esul

ts

Food

san

dFo

odG

roup

sQ

inet

al.,

[42]

Pros

pect

ive

coho

rtst

udy

n=

1,56

6co

mm

unity

-dw

ellin

gad

ults

(mea

nag

e63

y).

Ass

ocia

tion

offis

hco

nsum

ptio

nw

ithde

clin

ein

cogn

itive

func

tion.

Die

tmea

sure

dby

3-d

24-h

reca

lls.T

ICS-

m:g

loba

land

com

posi

teco

gniti

vesc

ores

.

Age

sign

ifica

ntly

mod

ified

the

asso

ciat

ion

betw

een

fish

cons

umpt

ion

and

cogn

itive

chan

ge.

Am

ong

adul

tsag

ed≥6

5y,

atle

ast

1se

rvin

g/w

eek

fish

pred

icte

dsl

ower

cogn

itive

decl

ine

(mea

nan

nual

rate

ofgl

obal

cogn

itive

decl

ine

was

redu

ced

by0.

35po

int,

95%

CI:

0.13

–0.5

8).

Follo

w-u

p:m

ean

5.3

y

Mor

ris

etal

.,[4

3]C

ross

-sec

tiona

lstu

dyn

=28

6au

tops

ied

brai

ns(m

ean

age

atde

ath

89.9

y)R

elat

ion

ofse

afoo

dco

nsum

ptio

nw

ithbr

ain

mer

cury

leve

ls.A

ssoc

iatio

nof

seaf

ood

cons

umpt

ion

orbr

ain

mer

cury

leve

lsw

ithbr

ain

neur

opat

holo

gies

.

Bra

inau

topt

ical

asse

ssm

ent.

Mer

cury

and

sele

nium

brai

ntis

sue

conc

entr

atio

ns.

FFQ

for

cons

umpt

ion

ofse

afoo

dan

dn-

3fa

ttyac

ids

inth

e4.

5y

befo

rede

ath.

Seaf

ood

cons

umpt

ion

[>1

mea

l(s)

/wee

k]si

gnifi

cant

lyco

rrel

ated

with

less

AD

path

olog

y(�

:–0.

69sc

ore

units

,95%

CI:

–1.3

4,–0

.04)

.Sea

food

cons

umpt

ion

corr

elat

edw

ithhi

gher

brai

nle

vels

ofm

ercu

ry,

thes

ele

vels

notc

orre

late

dw

ithbr

ain

neur

opat

holo

gy.

Dan

thiir

etal

.,[4

4]C

ross

-sec

tiona

lstu

dyn

=39

0co

mm

unity

-dw

ellin

gco

gniti

vely

norm

alol

der

adul

ts(m

ean

age

73.1

y)

Ass

ocia

tions

betw

een

mul

tiple

dom

ains

ofco

gniti

onan

der

ythr

ocyt

em

embr

ane

n-3

PUFA

prop

ortio

nsan

dhi

stor

ical

and

cont

empo

rary

fish

inta

kein

olde

rad

ults

.

n-3

PUFA

anal

ysis

iner

ythr

ocyt

em

embr

anes

No

evid

ence

that

high

erpr

opor

tions

oflo

ng-c

hain

n-3

fatty

acid

sor

fish

inta

kebe

nefit

sco

gniti

vepe

rfor

man

ce.S

mal

lneg

ativ

eef

fect

offis

hin

take

inch

ildho

odan

dol

der

age

onla

te-l

ife

cogn

itive

func

tion.

Fish

cons

umpt

ion

(cur

rent

:FF

Q,h

isto

rica

l:L

DQ

).C

ogni

tive

test

s.

Don

get

al.,

[45]

Cro

ss-s

ectio

nals

tudy

n=

894

Chi

nese

adul

tsw

ithno

rmal

cogn

ition

orw

ithM

CI

(mea

nag

e62

.9y)

Ass

ocia

tion

betw

een

nuts

,ve

geta

bles

and

frui

t-ri

chdi

etan

dth

eri

skof

cogn

ition

impa

irm

ent.

MoC

A;F

FQof

13fo

odgr

oups

tota

ling

41ite

ms;

MC

Iop

erat

iona

lized

with

MoC

Asc

ores

(cut

-off

poin

ts:1

3/14

for

nofo

rmal

educ

atio

n,19

/20

for

1–6

yof

educ

atio

n,an

d24

/25

for

7or

mor

ey

ofed

ucat

ion)

.

The

nuts

and

cook

ing

oili

ntak

eof

MC

Ipa

tient

sw

asle

ssth

anth

eno

rmal

subj

ects

(7.1

4ve

rsus

10.7

1nu

ts/d

aily

and

29.7

6ve

rsus

35.2

0m

illili

ter

cook

ing

oilp

erda

y).

Frui

tand

vege

tabl

ein

take

impr

oved

orie

ntat

ion,

nam

ean

dat

tent

ion

abili

ty.F

ruit

and

vege

tabl

eju

ice

drin

king

impr

oved

abst

ract

ion

abili

ty.

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 827Pa

stor

-Val

ero

etal

.,[4

6]C

ross

-sec

tiona

lstu

dyn

=18

49lo

w-i

ncom

eol

der

subj

ects

with

(n=

147,

mea

nag

e77

.5y)

and

with

outc

ogni

tive

impa

irm

ent(

n=

1702

,m

ean

age

71.5

y)

Ass

ocia

tion

betw

een

frui

tand

vege

tabl

ein

take

and

cogn

itive

impa

irm

ent.

CSI

-D;F

FQ:1

0ve

geta

bles

item

s,an

d17

frui

tand

natu

ralj

uice

site

ms;

Mon

thly

cons

umpt

ion

offis

h.

Dai

lyin

take

offr

uita

ndve

geta

ble≥4

00gr

ams/

day

asso

ciat

edw

ithde

crea

sed

prev

alen

ceof

cogn

itive

impa

irm

ent(

OR

:0.5

3,95

%C

I:0.

31–0

.89)

.Fis

hco

nsum

ptio

nw

asno

tas

soci

ated

with

cogn

itive

impa

irm

ent.

Zha

oet

al.,

[47]

Cro

ss-s

ectio

nals

tudy

n=

404

patie

nts,

aged

60y

orab

ove

with

orw

ithou

tMC

IA

ssoc

iatio

nof

diet

ary

and

lifes

tyle

patte

rns

with

MC

I.M

oCA

;FFQ

;MC

Iop

erat

iona

lized

with

MoC

Asc

ores

(cut

-off

poin

ts:1

3/14

for

nofo

rmal

educ

atio

n,19

/20

for

1–6

yof

educ

atio

n,an

d24

/25

for

7or

mor

ey

ofed

ucat

ion)

.

Hig

her

daily

inta

keof

eggs

(OR

:0.9

75,9

5%C

I:0.

959–

0.99

2)an

dm

arin

epr

oduc

ts(O

R:0

.975

,95%

CI:

0.95

9–0.

992)

sign

ifica

ntly

decr

ease

dod

dsof

suff

erin

gfr

omM

CI.

Xu

etal

.,[4

8]C

ross

-sec

tiona

lstu

dyn

=51

7C

hine

seel

derl

yw

ithpo

ssib

lede

men

tia(2

2.1%

,m

ean

age

73.8

y)an

dw

ithou

tcog

nitiv

eim

pair

men

t(77

.9%

mea

nag

e65

.7y)

Eff

ecto

fw

eekl

yto

fuin

take

onco

gniti

vepe

rfor

man

ce.

HV

LTIR

;FFQ

;Dia

gnos

isof

dem

entia

and

MC

I(r

evis

edPe

ters

encr

iteri

a).

Am

ong

olde

rsu

bjec

ts(≥

68y

ofag

e),h

igh

tofu

inta

kein

crea

sed

the

risk

(of

alm

ost3

0%)

ofco

gniti

veim

pair

men

tind

icat

ive

ofde

men

tia(O

R:1

.27,

95%

CI:

0.99

–1.6

4),a

fter

adju

stin

gfo

ral

lcov

aria

tes.

Con

sum

ptio

nof

mea

tand

gree

nve

geta

bles

inde

pend

ently

asso

ciat

edw

ithbe

tter

mem

ory

func

tion.

O’B

rien

etal

.,[4

9]Po

pula

tion-

base

dpr

ospe

ctiv

eco

hort

stud

y.n

=16

010

wom

enw

ithou

tahi

stor

yof

stro

ke(m

ean

age

74y)

;fina

lsam

ple

n=

1546

7

Ass

ocia

tion

oflo

ng-t

erm

inta

keof

nuts

with

cogn

ition

.

FFQ

;TIC

S,im

med

iate

and

dela

yed

reca

lls,c

ateg

ory

fluen

cy,d

elay

edre

call

ofth

eT

ICS

10-w

ord

lista

ndth

edi

gits

pan

back

war

dste

st.

Incr

easi

ngly

high

erto

taln

utin

take

(≥5

nuts

/wee

kve

rsus

neve

r<1/

mon

th)

rela

ted

toin

crea

sing

lybe

tter

over

allc

ogni

tion

atol

der

ages

(TIC

S=

0.21

SU,9

5%C

I:–0

.10–

0.52

;gl

obal

com

posi

tesc

ore

=0.

08SU

,95%

CI:

0.00

4–0.

15;a

ndth

eve

rbal

com

posi

tesc

ore

=0.

09SU

,95

%C

I:0

.01–

0.17

)

Follo

w-u

p:6

y

(Con

tinu

ed)

828 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

Tabl

e2

(Con

tinu

ed)

Ref

eren

ces

Stud

yD

esig

nSa

mpl

eO

utco

me

Cog

nitiv

ean

dN

utri

tiona

lA

sses

smen

tPr

inci

palR

esul

ts

Solf

rizz

ieta

l.,[5

0]Po

pula

tion-

base

dpr

ospe

ctiv

eco

hort

stud

yn

=5,

632

subj

ects

,age

d65

–84

y;fin

alsa

mpl

en

=14

45

Ass

ocia

tion

betw

een

chan

geor

cons

tant

habi

tsin

coff

eeco

nsum

ptio

nan

dth

ein

cide

nce

ofM

CI.

FFQ

;MC

Idi

agno

sis

with

slig

htly

mod

ified

Pete

rsen

crite

ria.

Cog

nitiv

ely

norm

alol

der

indi

vidu

als

who

incr

ease

dth

eir

coff

eeco

nsum

ptio

nha

da

high

erra

teof

deve

lopi

ngM

CI

(HR

:1.8

0,95

%C

I:1.

11–2

.92,

whi

lea

cons

tant

intim

em

oder

ate

coff

eeco

nsum

ptio

nw

asas

soci

ated

toa

redu

ced

rate

ofth

ein

cide

nce

ofM

CI

(1cu

p/da

y=

HR

:0.4

7,95

%

Follo

w-u

p:3.

5y

CI:

0.21

1–1.

02or

1-2

cups

/day

=H

R:

0.31

,95%

CI:

0.13

–0.7

5).

Ara

ujo

etal

.,[5

1]C

ross

-sec

tiona

lstu

dyn

=14

,563

publ

icse

rvic

ew

orke

rs(m

ean

age

51.9

y)R

elat

ion

ofco

ffee

cons

umpt

ion

tope

rfor

man

ceon

spec

ific

dom

ains

ofco

gniti

on.

Cog

nitiv

ete

sts

from

CE

RA

Dba

ttery

;FFQ

;C

offe

eco

nsum

ptio

nas

soci

ated

with

bette

rco

gniti

vepe

rfor

man

ceon

mem

ory

(AM

R:1

.03,

95%

CI:

1.00

–1.0

7)an

def

ficie

ncy

ofse

arch

ing

inlo

ng-t

erm

mem

ory

(dif

fere

nce

ofth

em

ean:

1.23

,95%

CI:

0.16

–2.2

9)on

lyin

elde

rly,

but

with

outa

dose

resp

onse

rela

tions

hip.

Type

ofco

ffee

,caf

fein

eco

nten

t,ad

ditio

nali

tem

sad

ded.

Bey

doun

etal

.,[5

2]Pr

ospe

ctiv

eco

hort

stud

yn

=62

8–1,

305

subj

ects

free

ofde

men

tia(m

ean

age

66.8

y)A

ssoc

iatio

nof

caff

eine

and

alco

holi

ntak

ew

ithco

gniti

vepe

rfor

man

ce.

MM

SE,B

VR

T,C

VLT

,V

FT-L

,VFT

-C,T

MT

A,

Stra

tum

-spe

cific

asso

ciat

ions

byse

xan

dba

selin

eag

e,be

twee

nca

ffei

nean

dal

coho

lint

ake

and

cogn

ition

.Fo

llow

-up:

∼2vi

sits

/per

son

each

∼2y

inte

rval

sT

MT

B,D

S-F,

DS-

BPu

tativ

ebe

nefic

iale

ffec

tsof

caff

eine

and

NA

Son

glob

alco

gniti

on,

verb

alm

emor

y,an

dat

tent

ion,

and

mix

edef

fect

sof

alco

holo

nle

tter

fluen

cy,a

ttent

ion,

and

wor

king

mem

ory.

7-d

diet

ary

reco

rds

for

caff

eine

and

alco

hol

inta

kes;

NA

S.

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 829

Tra

vass

oset

al.,

[53]

Cro

ss-s

ectio

nalm

ultic

ente

rst

udy

n=

88pa

tient

sw

ithA

D(5

8%)

orM

CI

(42%

)(m

ean

age

66.3

y)

Ass

ocia

tion

ofca

ffei

neco

nsum

ptio

nw

ithth

eC

SFbi

omar

kers

,par

ticul

arly

A�

.

FFQ

;Caf

fein

ean

dm

ain

activ

em

etab

olite

sin

the

CSF

and

plas

ma

A�

1-42

,to

talt

auan

dph

osph

oryl

ated

tau

and

othe

rA

�sp

ecie

sin

the

CSF

;Dia

gnos

isof

AD

and

MC

I(E

AD

Ccr

iteri

a).

Caf

fein

eco

nsum

ptio

ndi

dno

tmod

ify

the

leve

lsof

CSF

biom

arke

rs.

The

obro

min

eas

soci

ated

with

afa

vora

ble

A�

profi

lein

the

CSF

(A�

1-42

=r:

0.43

3;A

�X

-42

=r:

0.47

7).

Kes

se-G

uyot

etal

.,[5

4]Pr

ospe

ctiv

eco

hort

stud

yn

=2,

983

mid

dle-

aged

adul

tsfr

omth

eSU

.VI.

MA

Xst

udy

(mea

nag

eat

cogn

itive

eval

uatio

n65

.5y)

Ass

ocia

tion

betw

een

CD

Pan

dco

gniti

vepe

rfor

man

ce.

Plas

ma

conc

entr

atio

nsof

caro

teno

ids;

24h

diet

ary

reco

rdev

ery

2m

onth

s;R

RR

stat

istic

alm

etho

d;6

neur

opsy

chol

ogic

alte

sts

Posi

tive

corr

elat

ion

betw

een

CD

Pan

dco

nsum

ptio

nof

oran

ge-

and

gree

n-co

lore

dfr

uits

and

vege

tabl

es,v

eget

able

oils

and

soup

Follo

w-u

p:m

ean

13.6

y

CD

Pas

soci

ated

with

ahi

gher

com

posi

teco

gniti

vesc

ore

(mea

ndi

ffer

ence

1.04

,95

%C

I:0.

20–1

.87)

,aft

erad

just

men

tfor

soci

odem

ogra

phic

,lif

esty

lean

dhe

alth

fact

ors.

Posi

tive

asso

ciat

ion

betw

een

CD

Pan

dco

gniti

vefu

nctio

n(e

xecu

tive

func

tioni

ngan

dep

isod

icm

emor

y).

TIC

S-m

,Tel

epho

neIn

terv

iew

ofC

ogni

tive

Stat

usm

odifi

ed;

CI,

confi

denc

ein

terv

al;

FFQ

,Foo

dFr

eque

ncy

Que

stio

nnai

re;

AD

,Alz

heim

er’s

dise

ase;

PUFA

,pol

yuns

atur

ated

fatty

acid

s;L

DQ

,L

ifet

ime

Die

tQ

uest

ionn

aire

;M

CI,

mild

cogn

itive

impa

irm

ent;

MoC

A,

Mon

trea

lC

ogni

tive

Ass

essm

ent;

CSI

-D,

Com

mun

itySc

reen

ing

Inst

rum

ent

for

Dem

entia

;O

R,

odds

ratio

;H

VLT

IR,

Hop

kins

Ver

balL

earn

ing

Test

imm

edia

tere

call;

SU,s

tand

ard

units

;HR

,haz

ard

ratio

;CE

RA

D,C

onso

rtiu

mto

Est

ablis

hA

Reg

istr

yfo

rAlz

heim

er’s

Dis

ease

;AM

R,a

rith

met

icm

ean

ratio

;MM

SE,

Min

i-M

enta

lSt

ate

Exa

min

atio

n;B

VR

T,B

ento

nV

isua

lR

eten

tion

Test

;C

VLT

,Cal

ifor

nia

Ver

bal

Lea

rnin

gTe

st;

VFT

-L,V

erba

lFl

uenc

yTe

st-

Let

ter;

VFT

-C,V

erba

lFl

uenc

yTe

st-

Cat

egor

ical

;T

MT

A,T

rail

Mak

ing

Test

Part

A;T

MT

B,T

rail

Mak

ing

Test

Part

B;D

S-F,

Dig

itSp

anFo

rwar

dte

st;D

S-B

,Dig

itSp

anB

ackw

ard

test

;NA

S,N

utri

enta

dequ

acy

scor

e;C

SF,c

ereb

rosp

inal

fluid

;A

�,a

myl

oid-

�;E

AD

C,E

urop

ean

Alz

heim

er’s

Dis

ease

Con

sort

ium

;SU

.VI.

MA

X,S

uppl

emen

tatio

nen

Vita

min

eset

Min

erau

xA

ntio

xyda

nts;

CD

P,ca

rote

noid

-ric

hdi

etar

ypa

ttern

;RR

R,r

educ

edra

nkre

gres

sion

.

830 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

temporal, and inferior parietal, and marginally associ-ated with fusiform CT [34]. Higher fish consumptionwas also related with larger total gray matter vol-ume [35]. Fish consumption was also associatedwith a slower decline in composite and verbal mem-ory scores [42] (Table 2). Of note, Danthiir andcolleagues demonstrated that more frequent con-sumption of total fish (oily and white) was associatedwith slower cognitive speed for the constructs ofinhibition, simple/choice reaction time, reasoningspeed, and memory scanning [44] (Table 2). Morefrequent consumption of oily fish was significantlyassociated with worse inhibitory processes; similarly,consumption of white fish significantly and nega-tively predicted simple/choice reaction time [44].Danthiir and colleagues hypothesized that the nega-tive trends observed between cognitive performanceand fish consumption were due to neurotoxic contam-inants in fish, such as methylmercury [44]. However,as seen above, Morris and colleagues found thathigher brain levels of mercury were not correlatedwith brain neuropathology [43] (Table 2). Otherstudies did not suggest evidence that higher fishintake may impact positively cognitive performancein older adults with cognitive impairment [45, 46]or in those cognitively normal [46, 47] (Table 2).However, Dong and colleagues found that cognitivelynormal Chinese older subjects consumed more fishthan MCI subjects [45], and Zhao and colleaguessuggested that higher consumption of marine productwas associated with a significantly decreased odd ofsuffering from MCI [47], although in these two Chi-nese studies, MCI was diagnosed only on the basisof a screening cognitive test (the Montreal CognitiveAssessment - MoCA) and not with clinical criteria(Table 2).

Fruits and vegetablesIn Greece, in the EPIC study, among the compo-

nents of MeDi, only vegetable consumption exhibiteda significant inverse association with cognitivedecline [27] (Table 1). The diet low in fruit and veg-etable might increase the risk of cognitive functiondecline in older adults [45] (Table 2). In fact, adher-ence to WHO recommendations for daily intakes offruit and vegetable, that are eating 5 or more por-tions of fruit and/or vegetables a day (≥400 g/day),were significantly associated with a 47% decreasedprevalence of cognitive impairment [45]. In contrastto these findings, Xu and colleagues found that amongolder adults (≥68 y) being vegetarian (not eatingmeat), the risk for cognitive impairment increased

almost 4-fold [48] (Table 2). Imaging data in oldercohort showed that higher intake of total vegeta-bles was associated with larger dorsolateral prefrontaland superior parietal cortical thickness, while veg-etables without legumes were associated with largermiddle temporal, superior parietal, and dorsolateralprefrontal cortical thickness [35] (Table 1). In con-trast, fruit consumption was negatively associatedwith inferior parietal, supramarginal, superior pari-etal, parietal, and precuneus cortical thickness [35].These findings are in keeping with result of anotherstudy in which higher fruit intake was associatedwith lower temporal and hippocampal volumes [36](Table 1). This is probably due to high content ofsimple sugars and a high glycemic index of severalfruits and so the effects of carbohydrate componenton increased risk of MCI [73]. In older adults, fruitintake would benefit name ability and attention level,while vegetables intake would benefit orientationability [45] (Table 2). Finally, consumption of greenvegetables was independently associated with bettermemory function and among older elderly (≥68 y) itreduced the risk for cognitive impairment by almost20% [48] (Table 2).

NutsNuts are rich in PUFA (n-3 and n-6) and MUFA,

and also contain a significant amount of minerals,such as phosphorus, potassium, magnesium, calcium,iron, and sulfur, and vitamins, such as B1, B2, B6,and E. It was found the nut intake of MCI patientswas less than that of cognitively normal subjects[45] (Table 2). In fact, a study performed on olderwomen found that higher total nut intake (i.e.,≥5/week) over the long term was associated withmodestly better cognitive performance [49](Table 2).Increasingly higher total nut intake was related toincreasingly better overall cognition at older age.Considering that one year of age was associated witha mean decline of 0.04 standard units on both theglobal and verbal composite scores, the mean dif-ferences comparing the highest to lowest categoriesof nut intake were equivalent to approximately twoyears of cognitive aging [49]. In the same study, itwas found a suggestion that those who consumedwalnuts 1 to 3 times per month had better cogni-tion than those who consumed walnuts less than onceper month, but there was no overall trend of increas-ingly better cognitive performance with increasingwalnut intake [49]. Dong and colleagues also showedthat nut intake benefited delayed memory [45](Table 2).

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 831

Coffee and caffeine intakeA recent systematic review suggested that several

cross-sectional and longitudinal population-basedstudies suggested a protective effect of coffee, tea,and caffeine use against late-life cognitive impair-ment/decline, although the association was not foundin all cognitive domains investigated and there wasa lack of a distinct dose-response association, witha stronger effect among women than men [74]. Thefindings on the association of coffee, tea, and caffeineconsumption or plasma caffeine levels with incidentMCI and its progression to dementia were too limitedto draw any conclusion [74]. Furthermore, for demen-tia and AD prevention, some studies with baselineexamination in midlife pointed to a lack of associ-ation, although other case-control and longitudinalpopulation-based studies with briefer follow-up peri-ods supported favorable effects of coffee, tea, andcaffeine consumption against AD [74]. Recent find-ings from the Italian Longitudinal Study on Aging(ILSA) suggested that cognitively normal older indi-viduals who increased their coffee consumption hada higher rate of developing MCI, while a constant intime moderate coffee consumption was associated toa reduced rate of the incidence of MCI [50] (Table 2).Among older adults in Brazil, coffee consumptionwas associated with better cognitive performance onmemory and efficiency of searching in long-termmemory (drinking 2–3 cups of coffee per day wasassociated with about a 3% increase in the mean num-ber of words remembered on the learning, recall, andword recognition tests) [51] (Table 2). Also, drink-ing ≥3 cups/day of coffee was associated with anincrease of about 1.23 words in the mean number ofwords pronounced in the semantic verbal fluency test[51]. However, in this Brazilian study, Araujio andcolleagues did not find indication of a dose responserelationship in these associations [51]. In the Chi-nese study on cognitively normal and MCI adultsof Dong and colleagues, no significant associationwas detected between drinking of coffee and cogni-tive function [45] (Table 2). Another aspect of coffeeassumption is the role of its component such as thecaffeine. Coffee is a rich source of caffeine, whichacts as a psychoactive stimulant. In a cross sectionalanalysis, Beydoun and colleagues found that caffeineintake was associated with better global cognitivefunction (MMSE) at baseline for patients ≥70 years[52] (Table 2). However, in a study that evaluated theassociation of caffeine consumption with the cere-brospinal fluid (CSF) biomarkers, particularly A�,in AD and MCI patients, no significant difference

was found in daily consumption of caffeine betweenMCI and AD patients, with no correlation betweencaffeine consumption and A�1-42 in the CSF [53](Table 2). In the same study, theobromine, xanthineformed upon caffeine metabolism and also directlyingested from chocolate products, was associatedwith a favorable A� profile in the CSF [53]. Inter-estingly, theobromine in the CSF did not correlatewith caffeine consumption, theobromine consump-tion, or the levels of caffeine and other xanthines inthe plasma, but instead it correlated with levels ofcaffeine, theophylline, and paraxanthine in the CSF,suggesting that it may be formed by central metabolicpathways [53].

EggsEggs have a high content of proteins and lipids, in

particular cholesterol. For this reason, they are tra-ditionally considered an unhealthy food. However,eggs have also a significant amount of vitamins A,B6, B12, riboflavin, folic acid, choline, iron, calcium,phosphorus and potassium. In a recent study, higherdaily intake of eggs reduced of about 3% the oddsof suffering from MCI [47] (Table 2). Instead, in theChinese study of Dong and colleagues, no signifi-cant association was detected between intake of eggswith cognitive function in normal and MCI adults[45] (Table 2).

TofuTofu is a common food in most of the Far East. It

is obtained from curdling of the juice extracted fromsoybeans. It has a high protein and PUFA content.Higher weekly intake of tofu was associated withworse memory performance, furthermore amongolder elderly (≥68 y), high tofu intake increases therisk (of almost 30%) of cognitive impairment indica-tive of dementia [48] (Table 2).

MeatRed meat is a classical element of Western diet that,

as mentioned previously, was associated with worsecognitive performance in several imaging studies [37,39, 41] (Table 1). Consistent with these findings, anegative association of red meat with inferior andsuperior parietal cortical thickness was found [35](Table 1). However, this concept should be partiallyreviewed. In fact, in the last years, eating meat (notbeing vegetarian) was independently associated withbetter memory function and in older age (≥68 y)with a four-fold decrease in risk of possible dementia[48] (Table 2). Furthermore, Staubo and colleagues

832 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

also observed that higher red meat intake was asso-ciated with larger entorhinal cortical thickness [35](Table 1). This effect could also be related to somebeneficial components of lean red meat (iron, protein,MUFA, PUFA, cobalamine) and beneficial effects inincreasing satiety and reducing weight gain. In theChinese study of Dong and colleagues, no significantassociation was detected between intake of light orred meat with cognitive function in normal and MCIadults [45] (Table 2).

OilDong and colleagues, in their Chinese cohort,

found that oil intake of MCI patients was lessthan the normal subjects (29.76 versus 35.20 mLcooking oil per day), and, in particular, wouldhave a positive impact on visual-spatial ability [45](Table 2). Vegetable oils are rich in carotenoids, andin the Supplementation en Vitamines et MinerauxAntioxydants (SU.VI.MAX) study carotenoids wereassociated with higher cognitive performance [54](Table 2). Extra-virgin olive oil (EVOO) is one ofthe main elements of MeDi, and clinical trials andpopulation studies indicated that this dietary patternand its main lipid component EVOO could have aprotective role against AD [75].

LegumesDong and colleagues, in their Chinese cohort,

showed that normal subjects consumed more legumesand legume products than MCI subjects, demonstrat-ing that intake of legumes and legume products wouldbenefit overall cognition level [45] (Table 2). Thesedata were confirmed by imaging biomarkers; in fact,Staubo and colleagues also found that higher intakeof legumes was associated with larger parietal andoccipital cortical thickness, and with larger thicknessin ROIs for superior parietal, inferior parietal, pre-cuneus, and lingual cortical thickness [35] (Table 1).

GrainsIn their imaging biomarker study, Staubo and col-

leagues also showed that intake of whole grains orcereals was associated with larger temporal pole andsuperior temporal cortical thickness [35] (Table 1).Conversely, lower intake of whole grains was associ-ated with higher inflammatory markers (IL-6) andaccelerated cognitive decline at older age in theWhitehall II prospective cohort study [41] (Table 1).However, in the Chinese cohort of Dong and col-leagues, no significant association was detected

between intake of whole grains and cognitive func-tion in normal and MCI adults [45] (Table 2).

AlcoholRecent findings from the Baltimore Longitudinal

Study of Aging suggested that alcohol intake wasassociated with slower improvement on letter flu-ency and global cognition among those aged <70years at baseline [52] (Table 2). Conversely, alco-hol intake was associated with better attention andworking memory performance, particularly amongmen and individuals ≥70 years at baseline [52]. Com-pared with moderate consumption (14 to 28 g/d),individuals with higher alcohol intake (>28 g/d) hadfaster decline or slower improvement on the MMSE,particularly among women and in the older group.Overall, among men, and for those aged ≥70 years,lower alcohol intake (<14 g/d) compared with mod-erate consumption (14 to 28 g/d) was associated withpoorer performance in working memory [52]. In theyounger group, consuming <14 g/day was associatedwith slower decline or faster improvement in the let-ter fluency compared with a moderate intake of 14 to28 g/day. Similar pattern was showed also for atten-tion and executive functioning [52].

Relation of micro- and macronutrients withcognition in older age

Vitamin ETable 3 shows selected observational studies pub-

lished in the last three years and evaluating theassociation of micro- and macronutrients with late-life cognitive disorders [55–71]. Vitamin E is a potentantioxidant and seems to slow the effects of aging.It is found mainly in nuts, seeds, and oils. VitaminE is composed by four tocopherols (�-, �, δ, and�-tocopherols) and four corresponding tocotrienols,and it is known that �- and �-tocopherols foundin foods were linked to slower rate of cognitivedecline over a 6-year period [76]. More recently, inthe HANDLS study, Beydoun and colleagues inves-tigated the effects on cognitive domains of only oneof its isoform, �-tocopherol, showing that vitamin Ewas positively associated with performance in verbalmemory and fluency in the total population, partic-ularly among the younger age group (<48 y) andpsychomotor speed among women [55] (Table 3).

Vitamin DVitamin D deficiency is a common condition in

older adults. Furthermore, the detection of hydrox-

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 833

Tabl

e3

Obs

erva

tiona

lstu

dies

onth

eas

soci

atio

nof

mic

ro-

and

mac

ronu

trie

nts

with

late

-lif

eco

gniti

vedi

sord

ers

(201

4–20

16).

Ref

eren

ces

Stud

yD

esig

nSa

mpl

eO

utco

me

Cog

nitiv

ean

dN

utri

tiona

lPr

inci

palr

esul

tsA

sses

smen

t

Mic

ronu

trie

nts

Bey

doun

etal

.,[5

5]C

ross

-sec

tiona

lstu

dyn

=12

74ad

ults

(mea

nag

e47

.5y)

Ass

ocia

tion

ofdi

etar

yan

tioxi

dant

sw

ithco

gniti

vefu

nctio

n.

MM

SE,C

VLT

-Lis

tA,

CV

LT-D

FR,D

S-B

,DS-

F,B

VR

T,A

F,B

TA,T

MT

A,

TM

TB

,CD

T,C

R,I

P;

Vita

min

Ein

take

posi

tivel

yas

soci

ated

with

perf

orm

ance

indo

mai

nsof

verb

alm

emor

y(�

:+0.

64),

verb

alflu

ency

(�:+

0.53

),an

dps

ycho

mot

orsp

eed

(wom

en)

(�:+

0.68

).T

hevi

tam

inE

-ver

balm

emor

yas

soci

atio

nw

aspa

rtia

llym

edia

ted

byde

pres

sive

sym

ptom

s.

Two

24-h

reca

llses

timat

ing

daily

cons

umpt

ion

ofto

tal

caro

teno

ids

and

vita

min

sA

,C,a

ndE

per

1000

kcal

;

The

20-i

tem

CE

S-D

.

Hoo

shm

and

etal

.,[5

6]C

ross

-sec

tiona

lstu

dyn

=75

part

icip

ants

,29

SCI,

28M

CI,

and

18A

Dpa

tient

s(m

ean

age

61.6

y)

Ass

ocia

tion

betw

een

25(O

H)D

and

cogn

itive

impa

irm

ent,

CSF

biom

arke

rsof

AD

and

brai

nvo

lum

es.

Plas

ma

leve

lsof

25(O

H)D

Posi

tive

asso

ciat

ion

of25

(OH

)Dw

ithco

gniti

vest

atus

,CSF

A�

1-42

leve

lsan

dbr

ain

volu

mes

.In

part

icul

ar,f

orco

gniti

veim

pair

men

tthe

rew

asan

OR

:0.9

69,9

5%C

I:0.

948–

0.99

0pe

rin

crea

seof

1nm

ol/L

of25

(OH

)Dan

dan

OR

:4.1

9,95

%C

I:1.

30–1

3.52

for

25(O

H)D

valu

esle

ssth

an50

nmol

/Lco

mpa

red

with

valu

esgr

eate

rth

anor

equa

lto

50nm

ol/L

.Adj

ustin

gfo

rC

SFA

�1–

42at

tenu

ated

the

25(O

H)D

-cog

nitio

nlin

k.

CSF

leve

lsof

A�

1-42

,to

tal-

tau,

and

phos

phor

ylat

edta

uM

RI

brai

nvo

lum

es.

Dia

gnos

isof

SCI,

AD

,and

MC

I(I

WG

-MC

Icr

iteri

a).

Mos

coni

etal

.,[5

7]C

ross

-sec

tiona

lstu

dyn

=49

cogn

itive

lyno

rmal

indi

vidu

als

(mea

nag

e54

y)A

ssoc

iatio

nbe

twee

ndi

etar

ynu

trie

nts

and

brai

nbi

omar

kers

ofA

D.

Neu

roco

gniti

vete

stba

ttery

Hig

her

inta

keof

vita

min

B12

,vi

tam

inD

,and

n-3

PUFA

asso

ciat

edw

ithlo

wer

A�

load

inA

Dre

gion

son

PiB

-PE

T.H

ighe

rin

take

of�

-car

oten

ean

dfo

late

asso

ciat

edw

ithhi

gher

gluc

ose

met

abol

ism

onFD

G-P

ET.

PET

with

PiB

and

FDG

Sem

iqua

ntita

tive

FFQ

(116

item

s).

(Con

tinu

ed)

834 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

Tabl

e3

(Con

tinu

ed)

Ref

eren

ces

Stud

yD

esig

nSa

mpl

eO

utco

me

Cog

nitiv

ean

dN

utri

tiona

lPr

inci

palr

esul

tsA

sses

smen

t

Wils

onet

al.,

[58]

Pros

pect

ive

coho

rtst

udy

n=

2,77

7w

ell-

func

tioni

ngad

ults

(mea

nag

e73

.6y)

Ass

ocia

tion

betw

een

25(O

H)D

leve

lsan

dco

gniti

vepe

rfor

man

ce.

Seru

m25

(OH

)Dle

vels

;3M

S;D

SST.

Low

er25

(OH

)Dle

vels

asso

ciat

edw

ithlo

wer

base

line

cogn

itive

scor

eson

the

3MS

and

the

DSS

Tan

dw

ithgr

eate

rde

clin

esin

3MS

scor

esov

er4

y(l

east

squa

rem

ean

chan

ge:

–1.0

,95%

CI:

–1.5

to–0

.6fo

r<20

.0ng

/mL

;lea

stsq

uare

mea

nch

ange

:–0.

8,95

%C

I:–1

.2to

–0.3

for

20.0

–29.

9ng

/mL

;lea

stsq

uare

mea

nch

ange

:–0.

2,95

%C

I:–0

.7to

0.2

for≥

30.0

ng/m

L).

Follo

w-u

p:4

y

Litt

lejo

hns

etal

.,[5

9]Pr

ospe

ctiv

eco

hort

stud

yn

=1,

658

elde

rly

ambu

lato

ryad

ults

free

from

dem

entia

,ca

rdio

vasc

ular

dise

ase,

and

stro

ke

Ass

ocia

tion

betw

een

25(O

H)D

leve

lsan

dal

l-ca

use

dem

entia

and

AD

.

Seru

m25

(OH

)Dle

vels

For

inci

dent

all-

caus

ede

men

tia,

part

icip

ants

seve

rely

25(O

H)D

defic

ient

(<25

nmol

/L)

and

defic

ient

(25–

50nm

ol/L

)ha

dH

R:2

.25,

95%

CI:

1.23

–4.1

3an

dH

R:1

.53,

95%

CI:

1.06

–2.2

1)co

mpa

red

topa

rtic

ipan

tsw

ithsu

ffici

ent

conc

entr

atio

ns(>

50nm

ol/L

).Fo

rin

cide

ntA

D,p

artic

ipan

tsse

vere

ly25

(OH

)Dde

ficie

ntan

dde

ficie

ntco

mpa

red

topa

rtic

ipan

tsw

ithsu

ffici

ent

conc

entr

atio

nsha

dH

R:2

.22,

95%

CI:

1.02

–4.8

3an

dH

R;1

.69,

95%

CI:

1.06

–2.6

9,re

spec

tivel

y.

Follo

w-u

p:5.

6y

Dia

gnos

isof

all-

caus

ede

men

tiaan

dA

D.

Mill

eret

al.,

[60]

Pros

pect

ive

coho

rtst

udy

n=

382

olde

rad

ults

(mea

nag

e75

.5y)

Ass

ocia

tion

betw

een

25(O

H)D

stat

usan

dtr

ajec

tori

esof

cogn

itive

decl

ine.

Seru

m25

-OH

Dle

vels

with

vita

min

Dst

atus

;Spa

nish

and

Eng

lish;

Neu

rops

ycho

logi

cal

Ass

essm

entS

cale

s.

Low

vita

min

Dst

atus

asso

ciat

edw

ithac

cele

rate

dde

clin

ein

epis

odic

mem

ory

(vita

min

d-de

ficie

nt=

�:–

0.04

,SE

:0.

02),

and

exec

utiv

efu

nctio

n(v

itam

inD

-defi

cien

t=�

:–0.

05,

SE:0

.02)

.

Follo

w-u

p:m

ean

4.8

y

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 835

Gra

nic

etal

.,[6

1]Pr

ospe

ctiv

eco

hort

stud

yn

=77

5ol

der

adul

tsag

edov

er85

yA

ssoc

iatio

nbe

twee

n25

(OH

)Dan

dco

gniti

on.

MM

SE;A

ttent

ion

batte

ryof

the

CD

Rte

st;S

erum

25(O

H)D

seas

on-s

peci

ficqu

artil

es.

Low

and

high

seas

on-s

peci

fic25

(OH

)Dqu

artil

esas

soci

ated

with

high

erod

dsof

prev

alen

tco

gniti

veim

pair

men

t(O

R:

1.66

,95%

CI:

1.06

–2.6

0an

dO

R:1

.62,

95%

CI:

1.02

–2.5

9,re

spec

tivel

y)an

dpo

orer

atte

ntio

n,bu

tnot

with

odds

ofgl

obal

inci

dent

cogn

itive

impa

irm

ento

rde

clin

e.

Follo

w-u

p:3

y

Ass

man

net

al.,

[62]

Pros

pect

ive

coho

rtst

udy

n=

1009

part

icip

ants

(mea

nag

e66

.6y)

Ass

ocia

tion

betw

een

mid

life

plas

ma

25(O

H)D

conc

entr

atio

nsan

dsu

bseq

uent

cogn

itive

perf

orm

ance

.

Plas

ma

25(O

H)D

conc

entr

atio

ns;P

hone

mic

and

sem

antic

fluen

cyta

sks,

RI-

48,T

MT,

DS-

F,D

S-B

.

Am

ong

indi

vidu

als

with

low

educ

atio

n,po

sitiv

eas

soci

atio

nbe

twee

nhi

gher

25(O

H)D

conc

entr

atio

nsan

dbe

tter

shor

t-te

rm(w

orki

ngm

emor

y(�

:0.1

4,95

%C

I:0.

01–0

.24)

.

Follo

w-u

p:13

.3y

Moo

reet

al.,

[63]

Pros

pect

ive

coho

rtst

udy

n=

1354

part

icip

ants

(mea

nag

e66

.6y)

Ass

ocia

tion

ofse

rum

vita

min

B12

,red

cell

fola

te,a

ndco

gniti

veim

pair

men

t.

Seru

mvi

tam

inB

12;R

edce

llfo

late

;MM

SE.

Part

icip

ants

with

low

seru

mvi

tam

inB

12an

dhi

ghre

dce

llfo

late

leve

lsw

ere

mor

elik

ely

toha

veim

pair

edco

gniti

vepe

rfor

man

ce(a

djus

ted

OR

:3.

45,9

5%C

I:1.

60–7

.43)

.Pa

rtic

ipan

tsw

ithhi

ghfo

late

leve

ls,b

utno

rmal

seru

mvi

tam

inB

12,w

ere

also

mor

elik

ely

toha

veim

pair

edco

gniti

vepe

rfor

man

ce(a

djus

ted

OR

:1.7

4,95

%C

I:1.

03–2

.95)

.

Follo

w-u

p:10

y

Doe

tset

al.,

[64]

Pros

pect

ive

popu

latio

n–ba

sed

stud

yn

=22

03ol

der

subj

ects

(mea

nag

e72

.5y)

Ass

ocia

tion

betw

een

plas

ma

conc

entr

atio

nsof

fola

tean

dvi

tam

inB

12m

arke

rsin

rela

tion

toco

gniti

vepe

rfor

man

ce.

6co

gniti

vete

sts:

mod

ified

MM

SE,m

odifi

edD

ST,

shor

tfor

mof

Blo

ckD

esig

n;th

eK

OLT

,abr

idge

dve

rsio

nof

CO

WA

T,th

eT

MT-

A

Plas

ma

vita

min

B12

conc

entr

atio

nsin

the

low

est

quar

tile

com

bine

dw

ithpl

asm

afo

late

conc

entr

atio

nsin

the

high

estq

uart

ilew

ere

asso

ciat

edw

itha

redu

ced

risk

ofco

gniti

veim

pair

men

tco

mpa

red

with

plas

ma

conc

entr

atio

nsin

the

mid

dle

quar

tiles

ofbo

thvi

tam

ins

(OR

:0.

22,9

5%

CI:

0.05

–0.9

2).

Follo

w-u

p:Fo

late

and

vita

min

B12

mar

kers

in19

92–1

993

and

agai

nin

1997

–199

9,co

gniti

vete

stin

1997

–199

9

Fola

teco

ncen

trat

ions

and

vita

min

B12

mar

kers

.

(Con

tinu

ed)

836 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

Tabl

e3

(Con

tinu

ed)

Ref

eren

ces

Stud

yD

esig

nSa

mpl

eO

utco

me

Cog

nitiv

ean

dN

utri

tiona

lPr

inci

palr

esul

tsA

sses

smen

t

Kim

etal

.,[6

5]C

ross

-sec

tiona

lstu

dyn

=10

0ad

ults

with

MC

I,10

0w

ithA

D,a

nd12

1no

rmal

subj

ects

(mea

nag

e74

.8y)

Ass

ocia

tion

betw

een

Bvi

tam

ins

inta

kean

dco

gniti

vefu

nctio

n.

A24

-hdi

etar

yre

call;

Plas

ma

fola

te,v

itam

inB

12an

dho

moc

yste

ine

conc

entr

atio

ns;

MM

SE-K

C,B

N,W

F,W

LM

,CR

,WL

R,a

ndC

PT;D

iagn

osis

ofA

Dan

dM

CI

(Pet

erse

ncr

iteri

a).

Tota

lBvi

tam

inin

take

asso

ciat

edw

ithco

gniti

vefu

nctio

nin

cogn

itive

lyim

pair

edA

Dan

dM

CI

elde

rly,

and

the

asso

ciat

ion

was

stro

nger

inA

Dpa

tient

s.

Agn

ew-B

lais

etal

.,[6

6]Pr

ospe

ctiv

eco

hort

stud

yn

=70

30po

stm

enop

ausa

lw

omen

free

ofM

CI/

prob

able

dem

entia

(≥63

y)

Ass

ocia

tion

betw

een

fola

te,

vita

min

B6,

and/

orvi

tam

inB

12in

take

,alo

neor

inco

mbi

natio

n,w

ithin

cide

ntM

CI/

prob

able

dem

entia

.

FFQ

(122

item

s);

Low

erfo

late

inta

keas

soci

ated

with

incr

ease

dri

skof

inci

dent

MC

I/pr

obab

lede

men

tia(H

R:

2.0,

95%

CI:

1.3–

2.9)

.No

sign

ifica

ntas

soci

atio

nsbe

twee

nvi

tam

ins

B6

orB

12an

dM

CI/

prob

able

dem

entia

.

Follo

w-u

p:m

ean

5.0

yD

iagn

osis

ofpr

obab

lede

men

tiaan

dM

CI

(Pet

erse

ncr

iteri

a).

Rab

assa

etal

.,[6

7]Pr

ospe

ctiv

eco

hort

stud

yn

=65

2pa

tient

sw

ithou

tde

men

tia(m

ean

age

73.4

y)A

ssoc

iatio

nbe

twee

nto

tal

urin

ary

poly

phen

ols

(TU

Ps)

and

tota

ldie

tary

poly

phen

ols

(TD

Ps)

and

cogn

itive

decl

ine.

TU

P(F

olin

-Cio

calte

uas

say)

and

TD

P(F

FQ);

MM

SE;

TM

T-A

(atte

ntio

n)an

dT

MT-

B(e

xecu

tive

func

tion)

Hig

her

TU

Ple

vels

asso

ciat

edw

ithlo

wer

risk

ofsu

bsta

ntia

lco

gniti

vede

clin

eon

the

MM

SE(O

R:0

.53,

95%

CI:

0.34

–0.8

5)an

don

the

TM

T-A

(OR

:0.5

2,95

%C

I:0.

28–0

.96)

,but

noto

nT

MT-

B.

Follo

w-u

p:3

y

Fear

teta

l.,[6

8]Pr

ospe

ctiv

eco

hort

stud

yn

=10

92no

n-de

men

ted

com

mun

itydw

ellin

gol

der

subj

ects

(mea

nag

e74

.4y)

Ass

ocia

tion

plas

ma

caro

teno

ids

and

the

risk

ofde

men

tiaan

dA

D.

Bas

elin

eco

ncen

trat

ion

ofpl

asm

aca

rote

noid

s;M

MSE

,IST

,and

BV

RT;

Dia

gnos

isof

dem

entia

and

AD

Onl

yhi

gher

lute

inco

ncen

trat

ion

sign

ifica

ntly

asso

ciat

edw

itha

decr

ease

dri

skof

dem

entia

and

AD

(HR

:0.8

08,

95%

CI:

0.67

1–0.

973

and

HR

:0.7

59,9

5%C

I:0.

600–

0.96

0,re

spec

tivel

y).

Follo

w-u

p:10

y

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 837

Liu

etal

.,[6

9]C

ross

-sec

tiona

lstu

dyn

=33

8pa

tient

sag

ed40

yor

olde

rfr

omge

nera

lpo

pula

tion

Ass

ocia

tion

betw

een

caps

aici

nin

take

and

cogn

ition

and

bloo

dA

Dm

arke

rs.

FFQ

for

aca

psai

cin-

rich

diet

(chi

lian

dsp

icin

ess)

;C

hine

seve

rsio

nof

MM

SE;

Seru

mA

�1-

40an

dA

�1-

42le

vels

(EL

ISA

kits

).

Cap

saic

in-r

ich

diet

posi

tivel

yas

soci

ated

with

MM

SEsc

ores

and

inve

rsel

yas

soci

ated

with

AD

bloo

dbi

omar

kers

(ser

umA

�1-

40le

vels

and

tota

lser

umA

�le

vels

).M

acro

nutr

ient

sB

aier

leet

al.,

[70]

Cro

ss-s

ectio

nals

tudy

n=

45el

derl

ysu

bjec

tsin

two

grou

ps:w

ith(n

=12

,mea

nag

e78

.6y)

and

with

out

cogn

itive

impa

irm

ent

(n=

33,m

ean

age

73.3

y)

Ass

ocia

tion

offa

ttyac

idst

atus

with

cogn

itive

func

tion

and

card

iova

scul

arri

sk.

MM

SE;A

dapt

edC

ER

AD

batte

ry;S

erum

fatty

acid

com

posi

tion,

hom

ocys

tein

e,hs

-CR

P,lip

idpr

ofile

.

Cog

nitiv

efu

nctio

npo

sitiv

ely

asso

ciat

edw

ithth

ene

rvon

icac

id,D

HA

and

tota

ln-3

PUFA

,w

hile

myr

istic

acid

,pal

miti

cac

id,p

alm

itole

icac

id,t

hen-

6/n-

3ra

tioan

dho

moc

yste

ine

wer

ein

vers

ely

asso

ciat

ed.

Potta

laet

al.,

[71]

Cro

ss-s

ectio

nals

tudy

n=

1111

post

men

opau

sal

wom

enfr

eeof

dem

entia

(mea

nag

eat

the

time

ofM

RI

78y)

Ass

ocia

tion

betw

een

red

bloo

dce

ll(R

BC

)le

vels

ofm

arin

eom

ega-

3fa

ttyac

ids

and

MR

Ibr

ain

volu

mes

.

MR

Isc

anof

isch

emic

tissu

ean

d13

gray

and

whi

tem

atte

ran

atom

ical

regi

ons;

Tota

land

norm

albr

ain

volu

me

dire

ctly

rela

ted

toth

eom

ega-

3in

dex

(a1-

SDin

crea

sew

asas

soci

ated

with

2.1

cm3,9

5%C

I:0.

0–4.

3cm

3).

Lev

els

ofom

ega-

3in

dex

corr

elat

edw

ithla

rger

hipp

ocam

palv

olum

es(p

er-S

Dba

sis=

50m

m3,9

5%C

I:3–

97m

m3).

RB

Cfa

ttyac

idan

alys

isn-

3in

dex.

MM

SE,M

ini-

Men

talS

tate

Exa

min

atio

n;C

LVT-

Lis

tA,C

alif

orni

aV

erba

lLea

rnin

gTe

stim

med

iate

free

reca

llL

istA

;CV

LT-D

FR,C

alif

orni

aV

erba

lLea

rnin

gTe

stim

med

iate

Del

ayed

Free

Rec

all;

DS-

F,D

igit

Span

Forw

ard;

DS-

B,D

igit

Span

Bac

kwar

d;B

VR

T,B

ento

nV

isua

lR

eten

tion

Test

;C

VLT

,Cal

ifor

nia

Ver

bal

Lea

rnin

gTe

st;

AF,

Ani

mal

Flue

ncy

test

;B

TA,B

rief

Test

ofA

ttent

ion;

TM

TA

,Tra

ilM

akin

gTe

stPa

rtA

;TM

TB

,Tra

ilM

akin

gTe

stPa

rtB

;CD

T,C

lock

Dra

win

gTe

st;C

R,C

ard

Rot

atio

ns;I

P,Id

entic

alPi

ctur

es;C

ES-

D,C

ente

rfo

rE

pide

mio

logi

cSt

udie

sD

epre

ssio

nSc

ale;

SCI,

subj

ectiv

eco

gniti

veim

pair

men

t;M

CI,

mild

cogn

itive

impa

irm

ent;

AD

,Alz

heim

er’s

dise

ase;

25(O

H)D

,25-

hydr

oxyv

itam

inD

;CSF

,cer

ebro

spin

alflu

id;A

�,a

myl

oid-

�;I

WG

-MC

I,In

tern

atio

nalW

orki

ngG

roup

onM

ildC

ogni

tive

Impa

irm

ent;

OR

,odd

sra

tio;C

I,co

nfide

nce

inte

rval

;MR

I,m

agne

ticre

sona

nce

imag

ing;

PET,

posi

tron

emis

sion

tom

ogra

phy;

PiB

,(11

)C-P

ittsb

urgh

Com

poun

d-B

;FD

G,

(18)

F-flu

orod

eoxy

gluc

ose;

FFQ

,Fo

odFr

eque

ncy

Que

stio

nnai

re;

PUFA

,po

lyun

satu

rate

dfa

ttyac

ids;

3MS,

mod

ified

MM

SE;

DSS

T,D

igit

Sym

bol

Subs

titut

ion

Test

;H

R,

haza

rdra

tio;S

E,s

tand

ard

erro

r;C

DR

,Cog

nitiv

eD

rug

Res

earc

h;R

I-48

,rap

peli

ndic

e-48

item

s;D

ST,D

igit

Sym

bolT

est;

KO

LT,K

endr

ick

Obj

ectL

earn

ing

Test

;CO

WA

T,C

ontr

olle

dO

ralW

ord

Ass

ocia

tion

Test

;M

MSE

-KC

,Kor

ean

vers

ion

ofM

MSE

inth

eK

orea

nve

rsio

nof

the

Con

sort

ium

toE

stab

lish

aR

egis

try

for

Alz

heim

er’s

dise

ase

Ass

essm

ent

Pack

et(C

ER

AD

-K);

BN

,Bos

ton

Nam

ing;

WF,

Wor

dFl

uenc

y;W

LM

,Wor

dL

istM

emor

y;C

R,C

onst

ruct

iona

lRec

all;

WL

R,W

ord

Lis

tRec

ogni

tion;

CPT

,Con

stru

ctio

nalP

raxi

sTe

sts;

TU

P,to

talu

rina

rypo

lyph

enol

s;T

DP,

tota

ldi

etar

ypo

lyph

enol

s;IS

T,Is

aac’

sse

ttes

t;C

ER

AD

,Con

sort

ium

toE

stab

lish

aR

egis

try

forA

lzhe

imer

’sdi

seas

e;hs

-CR

P,hi

gh-s

ensi

tivity

Cre

activ

epr

otei

n;D

HA

,doc

osah

exae

noic

acid

;RB

C,r

edbl

ood

cell;

SD,s

tand

ard

devi

atio

n.

838 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

ylases for vitamin D activation and vitamin Dreceptors in neurons and glia suggested that vitaminD might have a role in cognition. In fact, serum 25-hydroxyvitamin D [25(OH)D] was cross-sectionallyassociated with reduced A�1-42 concentrations inCSF [56] (Table 3). These findings were on the sameline of Mosconi and colleagues showing an asso-ciation of higher intake of vitamin D with lowerA� load in AD regions on PET with 11C-PittsburghCompound-B (PiB) [57] (Table 3). Vitamin D defi-ciency was also associated with greater decline inglobal cognitive function over 4 years of follow-up [58] (Table 3) and with a substantially increasedrisk of all-cause dementia and AD over a meanfollow-up of 5.6 years [59] (Table 3). In particular,vitamin D deficiency was related with significantlyfaster declines in both episodic memory and execu-tive function performance, the two cognitive domainsstrongly associated with AD dementia [60] (Table 3).However, Granic and colleagues found a U-shapedrelationship between 25(OH)D and global cognitivefunction and attention in the very old [61] (Table 3).In fact, both low and high season-specific quar-tiles of 25(OH)D were associated with higher oddsof prevalent cognitive impairment, poorer attentionreaction times/processing speed and focused atten-tion/concentration, and greater attention fluctuation[61]. Therefore, it could be hypothesized that theneuroprotective effects of vitamin D may be attainedonly at moderate but not at low or high 25(OH)Dconcentrations. On the other hand, very old sub-jects taking vitamin D supplements could be probablyonly recently advised by their physician to take vita-min D, perhaps due to an underlying vitamin Ddeficiency condition [77]. Finally, findings from theSU.VI.MAX study suggested a positive associationbetween higher 25(OH)D concentrations and bettershort-term and working memory only among individ-uals with low education [62] (Table 3).

Vitamins B and folateVitamin B12 deficiency has been associated with

neurological and psychiatric symptoms includingcognitive impairment [78]. Additionally, vitaminB12, along with folate and vitamin B6, are key actorsin the homocysteine cycle, and the disruption of thiscycle can lead to elevated homocysteine, which hasbeen found among individuals with dementia [79]and has been shown to predict dementia prospec-tively [80]. Despite potential biological mechanismsthrough which B vitamins could play a role in theonset of dementia, results from observational and

intervention studies have been inconsistent. In thelast three years, in imaging studies, higher intake offolate was associated with higher glucose metabolismon (18)F-fludeoxyglucose (FDG)-PET [57] (Table 3)and greater superior temporal cortical thickness [35](Table 1). Positive associations were also foundbetween vitamins B1 and B2 and superior tempo-ral and temporal pole cortical thickness, and vitaminB6 and superior temporal and middle temporal corti-cal thickness [35]. Higher intake of vitamin B12 wasalso associated with lower A� load in AD regionson PiB-PET. The association of vitamin B12 withPiB retention was independent of gender, APOE, andfamily history [57].

Data collected on 1,354 subjects from three Aus-tralian cohorts suggested that participants with lowserum vitamin B12 (<250 pmol/L) and high redcell folate (>1,594 nmol/L) levels were more likelyto have impaired cognitive performance when com-pared to participants with biochemical measurementsthat were within the normal ranges [63] (Table 3). Inthe same cohorts, participants with high folate lev-els, but normal serum vitamin B12, were also morelikely to have impaired cognitive performance [63].Furthermore, in the Hordaland Health Study, plasmavitamin B12 concentrations in the lowest quartile(<274 pmol/l) combined with plasma folate concen-trations in the highest quartile (>18·5 nmol/l) wereassociated with a reduced risk of cognitive impair-ment compared with plasma concentrations in themiddle quartiles of both vitamins [64] (Table 3).

In a cross-sectional study on older Koreans, inthe AD group, positive associations were observedbetween vitamin B2 intake and global cognitivestatus, language, verbal fluency, and verbal andpraxis memory; between vitamin B6 intake andglobal cognitive status, language, verbal fluency, ver-bal and praxis memory, and constructional praxis;between vitamin B12 intake and global cognitivestatus, language, praxis memory, and constructionalpraxis; and between folate intake and praxis mem-ory [65] (Table 3). In the MCI group, vitaminB2 intake was positively associated with globalcognitive status and language, vitamin B6 intakewas positively associated with language, and folateintake was positively associated with global cog-nitive status and verbal memory. No associationswere observed in the normal group [65]. Finally, inthe Women’s Health Initiative Memory Study, folateintake below the recommended daily allowance atstudy baseline was associated with increased riskof incident MCI/probable dementia, after controlling

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 839

for multiple confounders [66] (Table 3). There wereno significant associations between vitamins B6 orB12 and MCI/probable dementia, nor any evidenceof an interaction between these vitamins and folateintake [66].

PolyphenolsPolyphenols are natural bioactive compounds that

have been identified in foods and beverages. They areassociated with neuroprotective activity probably dueto their antioxidant and anti-inflammatory properties.In the inCHIANTI study, in older adults with-out dementia, high concentrations of total urinarypolyphenols were associated with an approximately47% lower risk of substantial cognitive decline inglobal cognitive function and an approximately 48%lower risk of substantial cognitive decline in attentionover a 3-year period, but not with executive func-tion. No significant association was found betweentotal dietary polyphenols and any cognitive test [67](Table 3).

CarotenoidsCarotenoids are ubiquitous colored pigments,

found mainly in fruits and vegetables such as kiwi,carrots, and tomatoes or in green leafy vegetables.Among carotenoids, positive association was foundbetween higher intake of �-carotene and dorsolat-eral prefrontal and temporal pole cortical thickness[35] (Table 1) and with higher glucose metabolismon FDG-PET [57] (Table 3). However, the �-carotenelevels were related with reduced glucose metabolismfor women, APOE �4-carriers, and participants withpositive AD family history [57].

Among carotenoids, findings from the Three-City-Bordeaux cohort suggested that higher luteinconcentration, considered as a function of plasmalipids, was consistently significantly associated witha decreased risk of all-cause dementia and AD[68] (Table 3). Moreover, in the SU.VI.MAX study,a carotenoid-rich dietary pattern (CDP) stronglycorrelated with the plasma status of �-carotene, �-carotene, �-cryptoxanthin, and lutein, was positivelycorrelated with the consumption of green-coloredfruits and vegetables, vegetable oils, orange-coloredfruits and vegetables and soup, and was negativelycorrelated with that of beer-cider and wine. CDPscore was negatively associated with alcohol andprotein intake and positively associated with carbohy-drate intake [54] (Table 2). Furthermore, CDP scorewas positively associated with the composite cogni-tive performance score assessed 13 years later. More

specifically, higher CDP scores were related to betterepisodic memory, semantic fluency, working mem-ory, and executive functioning [54].

CapsaicinCapsaicin is a micronutrient acting as an active

component of chili peppers. In recent studies, ithas been shown to have favorable effects on vari-ous diseases including atherosclerosis, CVD, cancer,obesity, gastrointestinal diseases, hypertension, andstroke, and also reducing total and cause-specificmortality [81, 82]. Liu and colleagues, in a cross-sectional study collecting the dietary habits ofparticipants in 12 months, found that capsaicin-richdiet was associated with MMSE and serum A�40 lev-els, but not with serum A�42 levels [69] (Table 3).A possible explanation may be that neurons producemore A�40 than A�42 and serum A�42 levels havehigher cerebral tropism with increased capacity ofaggregation in the brain, so it is less prone to effluxfrom the brain to blood.

Macronutrients: Fatty acidsMany epidemiological studies have demonstrated

that dietary fatty acids may play a key role in severalpathological conditions. In particular, an increas-ing body of epidemiological evidence suggested thatelevated SFA could have negative effects on MCI[11], while a clear reduction of risk for cognitivedecline has been found in population samples withelevated fish consumption, high intake of MUFA andPUFA, particularly n-3 PUFA [11]. In fact, it is ofextreme nutritional importance the intake of essen-tial fatty acids from both the n-3 and n-6 series ofPUFA. In the n-3 series, eicosapentaenoic acid (EPA)and docosahexaenoic acid (DHA) are considered themost important. In particular, DHA might affect theneuronal survival. In agreement with this hypoth-esis, Baierle and colleagues showed that cognitivefunction was positively associated with DHA, ner-vonic acid (MUFA), and total n-3 PUFA, while wasinversely related to myristic acid (SFA), palmitic acid(SFA), palmitoleic acid (MUFA), the n-6 PUFA/n-3PUFA ratio, and homocysteine levels [70] (Table 3).In this study, elevated levels of homocysteine werealso inversely associated with total n-3 PUFA whichmay lead to vascular disease and, thus, may beinvolved in the risk of cognitive decline and otherbrain dysfunctions. Therefore, balanced intake ofboth n-3 and n-6 fatty acids may be essential forhealth by avoiding a harmful permanent proinflam-matory state. These data are supported also by recent

840 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

imaging studies. In fact, positive associations (higherintake with larger cortical thickness) were observedfor n-3 PUFA and linolenic acid (n-3 PUFA) withsuperior temporal cortical thickness and for linoleicacid (n-6 PUFA) with superior temporal, precuneus,middle temporal, and parietal cortical thickness [35](Table 1). Furthermore, in postmenopausal womenstudy, n-3 index [red blood cell (RBC) EPA+ DHAlevels] was directly associated with the total brainand hippocampal volumes. In particular, a 2 stan-dard deviation (SD) (3.2% absolute) higher n-3 indexwas associated with a 0.48% (4.2 cm3) larger totalbrain volume measured 8 years later [71] (Table 3).A marginally significant association between a 1-SDincrease in DHA with total and normal brain vol-ume has also been found, instead no association wasfound for EPA [71]. Furthermore, Pottala and col-leagues showed that higher levels of RBC EPA+ DHAwere associated with 50 mm3 larger hippocampal vol-umes [71]. Furthermore, higher intake of n-3 PUFAwas associated with lower A� load in AD regionson PiB-PET, independently of gender, APOE, andfamily history [57] (Table 3). On the contrary, inanother study, in normal older adults, higher propor-tions of long-chain n-3 PUFA did not give benefitson cognitive performances [44] (Table 2). Higherconcentrations of EPA were associated with poorerperceptual and reasoning speed in females, but notin males, before controlling for current fish intake[44]. Therefore, it is plausible that a physiologicallydifferent effect of long-chain n-3 PUFA in relationto gender may lead to differential associations withcognition [44].

Mechanisms underlying the relationship amongdietary factors and late-life cognitive disorders

Dietary patterns and late-life cognitivedisorders: Possible mechanism models

We need to pursue studies which will improve theknowledge of the biochemical mechanisms under-lying the pathophysiological processes and, in apublic health perspective, will examine dietary pat-terns. These mechanisms could be searched at thedifferent levels of investigation of dietary factorslinked to late-life cognitive disorders used in thepresent review article (i.e., dietary patterns, foodsand food-groups, and dietary micro- and macronu-trients). Consumption of the western diet contributesto the development of chronic conditions, such asatherosclerosis, diabetes, obesity, hypertension, andhyperlipidemia, which in turn results in metabolic,

inflammatory, and microvascular changes that induceinjury to the white matter of the brain [83, 84].Subsequently, white matter disease promotes thedevelopment and progression of cognitive declineand dementia [85].

There is converging evidence that compositedietary patterns, such as the MeDi, are relatedto lower risk for CVD, dyslipidemia, hyperten-sion and coronary artery disease, several formsof cancer, cognitive decline, and overall mortality[86, 87]. Several mechanisms have been proposedunderlying the suggested protective role of theMeDi against age-related changes in cognitive func-tion, predementia syndromes and dementia: vascularvariables and non-vascular biological mechanismssuch as metabolic, oxidative, and inflammatory pro-cesses [88]. Therefore, MeDi may act on cognitionthrough cerebrovascular mechanisms. Alternatively,AD almost certainly has its pathological origins inmidlife, and it is possible that components of theMeDi could affect the metabolism of A� or tauprotein [89]. Moreover, the model of “gene environ-ment interaction” may be also considered, in light ofthe recently proposed Latent Early Life AssociatedRegulation model, in which environmental agents,such as heavy metals and toxins, and dietary factors,could perturb gene regulation in a long-term fashion,beginning at early developmental stages, but theseperturbations do not have pathological results untilsignificantly later in life [90].

The DASH is a diet that has been shown by ran-domized clinical trials in the US to protect againstmany cardiovascular risk factors of dementia [91].Some of the protective effects of DASH includedlower blood pressure [92] and blood low-densitylipoprotein (LDL) cholesterol levels [93], weightreduction [94], reduced oxidative stress and inflam-mation [95], improved insulin sensitivity and reducedincidence of diabetes mellitus [96]. DASH diet andthe MeDi have many similar components, such asemphasis on natural plant-based foods and limitedconsumption of red meat, but there are differences.The MeDi is a cultural diet that uniquely empha-sizes daily intake of olive oil (MUFA), fish, andpotatoes, and moderate wine consumption, whereasDASH uniquely emphasizes dairy consumption andrestricted intake of sodium, commercial sweets, andsaturated fat.

The MIND diet is a refinement of the exten-sively studied cardiovascular diets, the MeDi andDASH diets, with modifications based on exist-ing studies relevant to nutrition and associated with

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 841

dementia prevention. MIND diet is not specific tothe underlying AD pathology, but perhaps betteroverall functioning and protection of the brain. Anumber of prospective cohort studies found thathigher consumption of vegetables was associatedwith slower cognitive decline with the strongest rela-tions observed for green leafy vegetables [12]. Greenleafy vegetables of MIND diet are sources of folate,vitamin E, carotenoids, and flavonoids, micronutri-ents that have been related to lower risk of dementiaand cognitive decline [31, 32].

Foods, food groups, and late-life cognitivedisorders: Possible mechanism models

Fish consumption has been associated with lowerrisk of AD because it could have a preventive effectagainst dementia through its anti-thrombotic andanti-inflammatory properties in addition to its spe-cific effect on neural functions, an effect thought tobe mediated through altering gene expression [97].Furthermore, the favorable effects of fish oil on cere-brovascular function are known, in particular for theimpact of DHA on endothelial nitric oxide synthase(eNOS) expression and a lower risk for ischemicstroke [98]. It was hypothesized that some nega-tive epidemiological findings observed for the impactof fish consumption on cognitive performance maybe due to neurotoxic contaminants in fish, such asmethylmercury [44]. However, as seen in one ofthe reviewed studies, higher brain levels of mer-cury were not correlated with brain neuropathology[43]. Diet may affect trace metal concentrations inthe brain of subjects with normal cognition [99] orAD [100–102], and transition metals, though cru-cial to many biochemical neuronal processes, wereabnormally aggregated and distributed in AD [103,104]. Dyshomeostasis in aluminum, silicon, lead,mercury, zinc, iron, and copper have been reportedin AD brains, where the abnormal accumulation anddistribution of the latter three may elicit oxidativestress and macromolecular damage, impeding cellu-lar function [104–107]. However, considering recentadvances in neurobiology of metals, the complex cor-relation between biometal metabolism, genetics, andthe environmental and the pathophysiology of manyneurodegenerative diseases [106, 107] warrants fur-ther investigation.

In recent years, numerous preclinical studiesinvolving fruits, nuts, and vegetables have identifiedpotential beneficial effects to brain health and cog-nition resulting from their nutrient and non-nutrientphytochemical content. These phytochemicals have

the ability to alter cellular function by modulat-ing transcription factors and altering the expressionof genes, cellular metabolism, and cellular signal-ing [108]. In vitro and in vivo studies have firmlyestablished the mechanistic effects of these phyto-chemicals on cognitive function, with many studiesdemonstrating their ability to counter oxidative stress[109], inflammation [110], as well as promote neu-ronal signaling and regulate transcription [111]. Inthe past five years, there has been an increasinginterest in the role that gut microbiota may play inthe modulation of the bidirectional signaling under-lying the gut–brain axis [112, 113]. In particular,in an Irish study, there was an association betweenindicators of general health, including frailty andMMSE, and microbiota composition only in long-stay subjects [114]. It has been also suggested thatdiet and specific nutrients can affect the compositionof the gut microbiota [115] that might influence theproduction or aggregation of A� protein in AD patho-genesis [116]. Bacteria populating the microbiomehave been shown to produce amyloids, lipopolysac-charides, and other immunogenic compounds [117,118] that might contribute to the regulation of sig-naling pathways implicated in neuroinflammation,brain A� deposition, and AD pathogenesis [116].Plant-derived nutrients and phytocompounds, byameliorating gut inflammation and dysbiosis, mightstimulate a positive modulation of the gut–brainaxis, reduce neuroinflammation, and retard or regresscognitive impairments associated with AD [119].Additionally, high intake of plant foods and spe-cific plant-derived compounds have been shown toreduce A� aggregation and, for this reason, are cur-rently considered suitable for the prevention of AD,as showed in some selected studies of the presentsystematic review [34, 57].

Caffeine, as an antioxidant substance, was revealedto reduce oxidative stress [120] and also to pro-tect against the disruptions of the blood-brain barrier[121]. Most of the biological effects of caffeine,including those on the brain and the central ner-vous system, are mediated through antagonism ofthe adenosine receptors, specifically the A1 andA2A receptors [122]. Other complementary mech-anisms of caffeine action that could contribute toits cognitive benefits in AD mice are the abil-ity of caffeine to decrease hippocampal levels ofpro-inflammatory cytokines [e.g., tumor necrosisfactor-�, interleukin (IL)-12(p70), and interferon-�)[123], caffeine’s beneficial effects on signal trans-duction factors is involved in neuronal plasticity and

842 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

survival [124], and the ability of caffeine to enhancebrain mitochondrial function [125].

Micronutrients, macronutrients, and late-lifecognitive disorders: Possible mechanism models

Optimal micronutrient status would moderate thedeterioration in brain integrity. Autophagy is proba-bly a crucial and a major process in the preservationof brain integrity. Micronutrients (vitamins, trace-elements, and also antioxidants) most likely affectbrain integrity by normalizing efficient autophagy[126]. Multiple lines of evidence indicate that oxida-tive stress not only strongly participates in an earlystage of AD prior to cytopathology, but it plays animportant role in inducing and activating multiplecell signaling pathways that contribute to the lesionformations of toxic substances and then promote thedevelopment of AD. Exogenous anti-oxidants (suchas vitamin E and C, carotenoids, flavonoids) decreasefree radical mediated damage caused by toxic chainreactions in neuronal cells and reduce the toxicity ofA� in in vitro studies of brains of patients with AD[127].

There is a vast literature demonstrating neuro-protection of the brain by vitamin E, rich sourcesof which are vegetable oils, nuts, and whole grains[128]. In particular, vitamin E might exert a regu-latory effect on cell proliferation and a beneficialeffect in improving glucose transport and insulinsensitivity [129]. Most of the known actions offlavonoids are related to their antioxidant propertiesthat can include suppression and scavenging of reac-tive oxygen species, and upregulation of antioxidantdefenses, the control of neuronal survival, death, anddifferentiation, long-term potentiation, and memoryas well as the effect on gene expression and interac-tions with mitochondria [130]. Therefore, flavonoids,lead to increased NO bioavailability in the vascula-ture of the hippocampus and subsequent angiogenesisand neurogenesis.

It has now been proven that vitamin B6, vita-min B12, and folate deficiency, as co-factors in themethylation of homocysteine, are associated withincreased homocysteine concentrations. Supraphys-iological concentrations of homocysteine or deficitsin folate and vitamin B12 should promote A� andtau protein accumulation and neuronal death, andalso have a direct effect on cognitive decline. Thepotential mechanisms whereby homocysteine mightmediate cognitive decline and dementia include:neurotoxicity induced by activation of N-methl-D-aspartate receptors, promotion of apoptosis, vas-

cular injury from promotion of atherogenesis makingvascular endothelium prothrombotic and prolifera-tion of smooth muscle cells, platelet activation, andincreased burden of ischemic strokes and white mat-ter lesions [131].

Epidemiologic investigations have revealed abeneficial role of vitamin D in muscle function, car-diovascular health, diabetes, and cancer prevention.Low levels of serum 25(OH)D are also associatedwith increased odds of prevalent cognitive dysfunc-tion, AD and all cause dementia in a number ofstudies. Vitamin D contributes to neuroprotectionby modulating the production of nerve growth, neu-rotrophin, glial cell derived neurotrophic factor, nitricoxide synthase and choline acetyl transferase [132],and neuroprotective mechanisms including vasopro-tection and A� phagocytosis and clearance [133,134].

Several mechanisms have been postulated for thepossible protective role of n-3 PUFA in demen-tia [77]. Supplementation with n-3 PUFA mayhave cognitive benefits based on animal studiesshowing neuroprotection through anti-oxidant, anti-inflammatory and anti-amyloid effects. DHA, thepredominant n-3 PUFA in the brain, is a key com-ponent of membrane phospholipids in the brain andadequate n-3 PUFA status, coming from dietaryconsumption, may help maintain integrity and neu-ronal function [135]. Secondly, n-3 PUFAs appearto have effects on dopaminergic and serotonergicsystems [136]. The oxidative products of PUFA actas key cellular mediators of inflammation, allergyand immunity, oxidative stress, bronchial constric-tion, vascular response, and thrombosis, and maythereby influence risk especially for vascular demen-tia. DHA may be directly involved in enhancingneuronal health in the aging brain through a range ofpotential mechanisms. In particular, it increases cellviability via neuroprotective and antiapoptotic mech-anisms, while also promoting dendritic arborizationand synaptogenesis [137]. DHA may modify theexpression of genes that regulate a variety of bio-logical functions potentially important for cognitivehealth, including neurogenesis and neuronal function[138]. Long-chain n-3 PUFA have been shown toreduce A� formation and oxidative damage [139].Higher intakes of saturated and trans fat and lowerintakes of mono- and polyunsaturated fat can con-tribute to insulin resistance and an atherogenic lipidprofile [140]. Moreover, insulin resistance, highinsulin levels, and cholesterol are all implicatedin A� accumulation in the brain, the pathologic

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 843

hallmark of AD [141]. In a very recent neuropatho-logical metabolomic-based study, in the brains ofpatients with varying degrees of AD pathology(healthy brains, brains with AD pathology but with-out cognitive symptoms, and AD brains), significantdifferences in the abundance of six unsaturated fattyacids were identified in three brain regions bothvulnerable and resistant to classical AD pathology,with gradations in these metabolites being relatedto both severity of neuropathology at death as wellas domain-specific cognitive performance during life[142]. These findings suggested that dysregulation ofunsaturated fatty acid metabolism may play a rolein driving AD pathology, providing with further evi-dence for the metabolic basis of AD pathogenesis.

DISCUSSION

In the last decade, the association between dietand cognitive function or dementia has been largelyinvestigated. In the present article, we systematicallyreviewed observational studies published in the lastthree years (2014–2016) on the relationship amongdietary factors and late-life cognitive disorders usingdifferent levels of investigation (i.e., dietary pat-terns, foods and food-groups, and dietary micro- andmacronutrients). In the present systematic review,considering separately the different cognitive out-comes in older age, we included studies focusing ondementia, AD, MCI, and different models of late-life cognitive impairment/decline, but we did notfind studies focusing in particular on VaD. From thereviewed studies, it appeared to be apparent that theNIA–AA guidelines for AD and cognitive declinedue to AD pathology introduced some evidence sug-gesting a direct relation between diet and changesin the brain structure and activity. Several studiesfocused on the role of the dietary patterns on late-lifecognition, with accumulating evidence that higheradherence to a Mediterranean-type diet was associ-ated with decreased cognitive decline, although theMeDi combines several foods, micronutrients, andmacronutrients already separately proposed as poten-tial protective factors against dementia and MCI.Moreover, also other emerging healthy dietary pat-terns, such as the DASH and the MIND diets, wereassociated with slower rates of cognitive decline andsignificant reduction in AD rate.

Furthermore, some foods or food groups tradi-tionally considered harmful such as eggs and redmeat have been partially rehabilitated, while there

is still a negative correlation of cognitive functionswith saturated fatty acids and a protective effectagainst cognitive decline of elevated fish consump-tion, high intake of MUFA and PUFA, particularlyn-3 PUFA. In fact, while a Western diet was asso-ciated with worse cognitive performance in severalimaging studies selected in the present systematicreview [37, 39, 41]; however, for red meat, a classicalelement of this dietary pattern, the epidemiologicalfindings were controversial [35, 48], also includingstudies outside the time limits of the present sys-tematic reviews [143, 144]. Multi-country ecologicalstudies of dietary meat and increased risk of cancer[145] have been strongly supported by trends in can-cer risk in Asian countries undergoing the nutritiontransition [146, 147], as well as in observational stud-ies in Uruguay [148]. Thus, it would be expected thatthey would also be reliable for AD. In 2014, an eco-logical study linked the nutrition transition in Japan tothe greatly increased risk of AD [149], with increasedmeat consumption as an important factor. Interest-ingly, rates of VaD have been relatively unchangedfor many years [149]. Recent observational and eco-logical findings added support for dietary advancedglycation end products as an important risk factorfor AD [150], with meat cooked at high tempera-tures in a dry environment as an important source.Another recent ecological study compared dietaryfactors with prevalence rates of AD in ten countrieswith recent prevalence determinations, showing thatdietary supply of meat or animal products less milk5 years before AD prevalence had the highest corre-lations with AD rates [151]. The three countries withhighest rates of AD and dietary meat, Brazil, Mongo-lia, and the U.S., are considerably different in manyrespects, adding confidence that the finding was notdue to confounding factors [151]. Noteworthy, MeDifalls in the middle of the regression line for meatand AD prevalence [151]. In these ecological stud-ies, countries with high consumption of rice and lowconsumption of meat had the lowest rates of AD, asdid Japan prior to the nutrition transition [149]. Also,eggs and fish added to meat gave similar results asmeat [151]. Fish that does not have PUFA n-3 andvitamin D may act like meat and eggs in terms of ADrisk.

However, some limitations should be reported forthis systematic review article. Heterogeneity existsin the quantification of individual items as wellamong the different diet backgrounds of the popu-lations investigated, especially in view of differentgeographical areas, setting of dietary patterns, such

844 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

as the Mediterranean countries in which a largesegment of the population still adhered to MeDi.Heterogeneity in time between the two assessments,among studies using paired assessments (or a sin-gle assessment) several years after study populationenrollment. Nevertheless, these data represent a brickin the construction of the building of the causal linkbetween dietary habits and cognitive impairment.

The absence of causal etiological therapies againstAD leads to seek multimodal alternative strategies,increasing the interest in the potential for preventionof dementia by targeting modifiable risk factors. Itis now evident that dietary habits influence diversecardiometabolic risk factors, including not only obe-sity and low-density lipoprotein cholesterol, butalso blood pressure, glucose-insulin homeostasis,lipoprotein concentrations and function, oxidativestress, inflammation, endothelial health, hepatic func-tion, adipocyte metabolism, pathways of weightregulation, visceral adiposity, and the gut micro-biome. Whereas decades of dietary recommendationsfocused on dietary fat and single vascular risk fac-tors (e.g., hypertension, blood cholesterol, etc.) andcurrent dietary discussions are often worried abouttotal calories and obesity, the full health impact ofdiet extends far beyond these pathways. Consideringstrategies of prevention of AD could be complicatedand take to negative results. A second key lesson is theimportance of pointing out specific foods and over-all diet patterns, rather than single isolated nutrients,for cognitive impairment. No single micronutrientor macronutrient, no single antioxidant or vitamin,especially in their synthetic forms, may be able toshow a dramatic effect on dementia prevention, dueto the complexity of biochemistry. Even a mixture offixed beneficial compounds applied to everybody willnot show dramatic effects against late-life cognitivedisorders due to interindividual and intraindividualresponse. Natural nutrition may exert better effectsthan single nutrients due to bioavailability and bio-chemical synergisms. However, also the healthiestnatural nutrition alone, such as MeDi, could not exertdramatic effects against cognitive disorders in olderage if not part of diet intended as lifestyle. A food-based approach also better facilitates public guidanceand minimizes industry manipulation. Nevertheless,the complexity of the stake, the correction of mod-ifiable risk factors to expect “the compression ofcognitive morbidity” still remains a desirable goalof public health. Healthy diets and lifestyles could benot very powerful if applied in fixed amounts regard-less of individualized multidimensional needs and

resources. Larger observational studies with longerfollow-up periods should be encouraged, address-ing other potential bias and confounding sources, sohopefully opening new ways for diet-related preven-tion of dementia and AD.

DISCLOSURE STATEMENT

Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/17-0248r1).

REFERENCES

[1] Alzheimer’s, Association (2016) 2016 Alzheimer’s dis-ease facts and figures. Alzheimers Dement 12, 459-509.

[2] Panza F, Seripa D, Solfrizzi V, Imbimbo BP, LozuponeM, Leo A, Sardone R, Gagliardi G, Lofano L, Cre-anza BC, Bisceglia P, Daniele A, Bellomo A, Greco A,Logroscino G (2016) Emerging drugs to reduce abnormal�-amyloid protein in Alzheimer’s disease patients. ExpertOpin Emerg Drugs 21, 377-391.

[3] Grant WB (1999) Dietary links to Alzheimer’s disease:1999 Update. J Alzheimers Dis 1, 197-201.

[4] Solfrizzi V, Panza F, Capurso A (2003) The role of diet incognitive decline. J Neural Transm 110, 95-110.

[5] Luchsinger JA, Mayeux R (2004) Dietary factors andAlzheimer’s disease. Lancet Neurol 3, 579-587.

[6] Tangney CC (2014) DASH and Mediterranean-typedietary patterns to maintain cognitive health. Curr NutrRep 3, 51-61.

[7] Xu W, Tan L, Wang HF, Jiang T, Tan MS, Tan L, Zhao QF,Li JQ, Wang J, Yu JT (2015) Meta-analysis of modifiablerisk factors for Alzheimer’s disease. J Neurol NeurosurgPsychiatry 86, 1299-1306.

[8] Luchsinger JA, Noble JM, Scarmeas N (2007) Diet andAlzheimer’s disease. Curr Neurol Neurosci Rep 7, 366-372.

[9] Sofi F, Macchi C, Abbate R, Gensini GF, Casini A (2010)Effectiveness of the Mediterranean diet: Can it help delayor prevent Alzheimer’s disease? J Alzheimers Dis 20, 795-801.

[10] Cunnane SC, Plourde M, Pifferi F, Begin M, Feart C,Barberger-Gateau P (2009) Fish, docosahexaenoic acidand Alzheimer’s disease. Prog Lipid Res 48, 239-256.

[11] Solfrizzi V, Frisardi V, Capurso C, D’Introno A, ColaciccoAM, Vendemiale G, Capurso A, Panza F (2010) Dietaryfatty acids in dementia and predementia syndromes:Epidemiological evidence and possible underlying mech-anisms. Ageing Res Rev 9, 184-199.

[12] Loef M, Walach H (2012) Fruit, vegetables and preventionof cognitive decline or dementia: A systematic review ofcohort studies. J Nutr Health Aging 16, 626-630.

[13] Parrott MD, Greenwood CE (2007) Dietary influenceson cognitive function with aging. From high-fat diets tohealthful eating. Ann N Y Acad Sci 1114, 389-397.

[14] Peters R, Peters J, Warner J, Beckett N, Bulpitt C (2008)Alcohol, dementia and cognitive decline in the elderly. Asystematic review. Age Ageing 37, 505-512.

[15] Panza F, Frisardi V, Seripa D, Logroscino G, SantamatoA, Imbimbo BP, Scafato E, Pilotto A, Solfrizzi V (2012)

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 845

Alcohol consumption in mild cognitive impairment anddementia: Harmful or neuroprotective? Int J Geriatr Psy-chiatry 27, 1218-1238.

[16] Panza F, Solfrizzi V, Barulli MR, Bonfiglio C, Guerra V,Osella A, Seripa D, Sabba C, Pilotto A, Logroscino G(2015) Coffee, tea, and caffeine consumption and pre-vention of late-life cognitive decline and dementia: Asystematic review. J Nutr Health Aging 19, 313-328.

[17] Solfrizzi V, Frisardi V, Capurso C, D’Introno A, Colaci-cco AM, Chiloiro R, Dellegrazie F, Di Palo A, Capurso A,Panza F (2006) Whole-diet approach: Working on a cri-terion validity for age-related cognitive decline and mildcognitive impairment. J Am Geriatr Soc 57, 1944-1946.

[18] Gu Y, Scarmeas N (2011) Dietary patterns in Alzheimer’sdisease and cognitive aging. Curr Alzheimer Res 8, 510-519.

[19] Solfrizzi V, Panza F (2014) Mediterranean diet and cogni-tive decline. A lesson from the whole-diet approach: Whatchallenges lie ahead? J Alzheimers Dis 39, 283-286.

[20] Lourida I, Soni M, Thompson-Coon J, Purandare N, LangIA, Ukoumunne OC, Llewellyn DJ (2013) Mediterraneandiet, cognitive function, and dementia: A systematicreview. Epidemiology 24, 479-489.

[21] Psaltopoulou T, Sergentanis TN, Panagiotakos DB, Ser-gentanis IN, Kosti R, Scarmeas N (2013) Mediterraneandiet and stroke, cognitive impairment, depression: A meta-analysis. Ann Neurol 74, 580-591.

[22] Singh B, Parsaik AK, Mielke MM, Erwin PJ, Knop-man DS, Petersen RC, Roberts RO (2014) Associationof Mediterranean diet with mild cognitive impairmentand Alzheimer’s disease: A systematic review and meta-analysis. J Alzheimers Dis 39, 271-282.

[23] McKhann GM, Knopman DS, Chertkow H, Hyman BT,Jack CR Jr, Kawas CH, Klunk WE, Koroshetz WJ, ManlyJJ, Mayeux R, Mohs RC, Morris JC, Rossor MN, Schel-tens P, Carrillo MC, Thies B, Weintraub S, Phelps CH(2011) The diagnosis of dementia due to Alzheimer’s dis-ease: Recommendations from the National Institute onAging-Alzheimer’s Association workgroups on diagnos-tic guidelines for Alzheimer’s disease. Alzheimers Dement7, 263-269.

[24] Moher D, Liberati A, Tetzlaff J, Altman DG, GroupPRISMA (2009) Preferred reporting items for systematicreviews and meta-analyses: The PRISMA statement. PLoSMed 6, e1000097.

[25] Zbeida M, Goldsmith R, Shimony T, Vardi H, NagganL, Shahar DR (2014) Mediterranean diet and functionalindicators among older adults in non-Mediterranean andMediterranean countries. J Nutr Health Aging 18, 411-418.

[26] Galbete C1, Toledo E, Toledo JB, Bes-Rastrollo M, Buil-Cosiales P, Marti A, Guillen-Grima F, Martınez-GonzalezMA (2015) Mediterranean diet and cognitive function: TheSUN project. J Nutr Health Aging 19, 305-212.

[27] Trichopoulou A, Kyrozis A, Rossi M, Katsoulis M,Trichopoulos D, La Vecchia C, Lagiou P (2015) Mediter-ranean diet and cognitive decline over time in an elderlyMediterranean population. Eur J Nutr 54, 1311-1321.

[28] Qin B, Adair LS, Plassman BL, Batis C, Edwards LJ,Popkin BM, Mendez MA (2015) Dietary patterns and cog-nitive decline among Chinese older adults. Epidemiology26, 758-768.

[29] Gardener SL, Rainey-Smith SR, Barnes MB, Sohrabi HR,Weinborn M, Lim YY, Harrington K, Taddei K, Gu Y,Rembach A, Szoeke C, Ellis KA, Masters CL, Macaulay

SL, Rowe CC, Ames D, Keogh JB, Scarmeas N, MartinsRN (2015) Dietary patterns and cognitive decline in anAustralian study of ageing. Mol Psychiatry 20, 860-866.

[30] Tangney CC, Li H, Wang Y, Barnes L, Schneider JA,Bennett DA, Morris MC (2014) Relation of DASH- andMediterranean-like dietary patterns to cognitive decline inolder persons. Neurology 83, 1410-1416.

[31] Morris MC, Tangney CC, Wang Y, Sacks FM, Bennett DA,Aggarwal NT (2015) MIND diet associated with reducedincidence of Alzheimer’s disease. Alzheimers Dement 11,1007-1014.

[32] Morris MC, Tangney CC, Wang Y, Sacks FM4, BarnesLL5, Bennett DA6, Aggarwal NT (2015) MIND dietslows cognitive decline with aging. Alzheimers Dement11, 1015-1022.

[33] Mosconi L, Murray J, Tsui WH, Li Y, Davies M, WilliamsS, Pirraglia E, Spector N, Osorio RS, Glodzik L, McHughP, de Leon MJ (2014) Mediterranean diet and magneticresonance imaging-assessed brain atrophy in cognitivelynormal individuals at risk for Alzheimer’s disease. J PrevAlzheimers Dis 1, 23-32.

[34] Berti V, Murray J, Davies M, Spector N, Tsui WH, LiY, Williams S, Pirraglia E, Vallabhajosula S, McHugh P,Pupi A, de Leon MJ, Mosconi L (2015) Nutrient patternsand brain biomarkers of Alzheimer’s disease in cognitivelynormal individuals. J Nutr Health Aging 19, 413-423.

[35] Staubo SC, Aakre JA, Vemuri P, Syrjanen JA, Mielke MM,Geda YE, Kremers WK, Machulda MM, Knopman DS,Petersen RC, Jack CR Jr, Roberts RO (2017) Mediter-ranean diet, micronutrients and macronutrients, and MRImeasures of cortical thickness. Alzheimers Dement 13,168-177.

[36] Gu Y, Brickman AM, Stern Y, Habeck CG, Razlighi QR,Luchsinger JA, Manly JJ, Schupf N, Mayeux R, ScarmeasN (2015) Mediterranean diet and brain structure in a mul-tiethnic elderly cohort. Neurology 85, 1744-1751.

[37] Jacka FN, Cherbuin N, Anstey KJ, Sachdev P, ButterworthP (2015) Western diet is associated with a smaller hip-pocampus: A longitudinal investigation. BMC Medicine13, 215.

[38] Pelletier A, Barul C, Feart C, Helmer C, Bernard C, PeriotO, Dilharreguy B, Dartigues JF, Allard M, Barberger-Gateau P, Catheline G, Samieri C (2015) Mediterraneandiet and preserved brain structural connectivity in oldersubjects. Alzheimers Dement 11, 1023-1031.

[39] Shakersain B, Santoni G, Larsson SC, Faxen-Irving G,Fastbom J, Fratiglioni L, Xu W (2016) Prudent diet mayattenuate the adverse effects of Western diet on cognitivedecline. Alzheimers Dement 12, 100-109.

[40] Smyth A, Dehghan M, O’Donnell M, Anderson C, TeoK, Gao P, Sleight P, Dagenais G, Probstfield JL, MenteA, Yusuf S, ONTARGET and TRANSCEND Investiga-tors (2015) Healthy eating and reduced risk of cognitivedecline: A cohort from 40 countries. Neurology 84, 2258-2265.

[41] Ozawa M, Shipley M, Kivimaki M, Singh-Manoux A,Brunner EJ (2017) Dietary pattern, inflammation and cog-nitive decline: The Whitehall II prospective cohort study.Clin Nutr 36, 506-512.

[42] Qin B, Plassman BL, Edwards LJ, Popkin BM, AdairLS, Mendez MA (2014) Fish intake is associated withslower cognitive decline in Chinese older adults. J Nutr144, 1579-1585.

[43] Morris MC, Brockman J, Schneider JA, Wang Y, BennettDA, Tangney CC, van de Rest O (2016) Association of

846 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

seafood consumption, brain mercury level, and APOE�4status with brain neuropathology in older adults. JAMA315, 489-497.

[44] Danthiir V, Hosking D, Burns NR, Wilson C, NettelbeckT, Calvaresi E, Clifton P, Wittert GA (2014) Cognitiveperformance in older adults is inversely associated withfish consumption but not erythrocyte membrane n–3 fattyacids. J Nutr 144, 311-320.

[45] Dong L, Xiao R, Cai C, Xu Z, Wang S, Pan L, Yuan L(2016) Diet, lifestyle and cognitive function in old Chineseadults. Arch Gerontol Geriatr 63, 36-42.

[46] Pastor-Valero M, Furlan-Viebig R, Rossi Menezes P, daSilva SA, Vallada H, Scazufca M (2014) Education andWHO recommendations for fruit and vegetable intake areassociated with better cognitive function in a disadvan-taged Brazilian elderly population: A population-basedcross-sectional study. PLoS One 9, e94042.

[47] Zhao X, Yuan L, Feng L, Xi Y, Yu H, Ma W, Zhang D,Xiao R (2015) Association of dietary intake and lifestylepattern with mild cognitive impairment in the elderly.J Nutr Health Aging 19, 164-168.

[48] Xu X, Xiao S, Rahardjo TB, Hogervorst E (2015)Tofu intake is associated with poor cognitive perfor-mance among community-dwelling elderly in China.J Alzheimers Dis 43, 669-675.

[49] O’Brien J, Okereke O, Devore E, Rosner B, Breteler M,Grodstein F (2014) Long-term intake of nuts in relation tocognitive function in older women. J Nutr Health Aging18, 496-502.

[50] Solfrizzi V, Panza F, Imbimbo BP, D’Introno A, Gal-luzzo L, Gandin C, Misciagna G, Guerra V, Osella A,Baldereschi M, Di Carlo A, Inzitari D, Seripa D, PilottoA, Sabba C, Logroscino G, Scafato E; Italian LongitudinalStudy on Aging Working Group (2015) Coffee consump-tion habits and the risk of mild cognitive impairment: TheItalian Longitudinal Study on Aging. J Alzheimers Dis 47,889-899.

[51] Araujo LF, Giatti L, Padilha dos Reis RC, Goulart AC,Schmidt MI, Duncan BB, Ikram MA, Barreto SM (2015)Inconsistency of association between coffee consumptionand cognitive function in adults and elderly in a cross-sectional study (ELSA-Brasil). Nutrients 7, 9590-9601.

[52] Beydoun MA, Gamaldo AA, Beydoun HA, Tanaka T,Tucker KL, Talegawkar SA, Ferrucci L, Zonderman AB(2014) Caffeine and alcohol intakes and overall nutrientadequacy are associated with longitudinal cognitive per-formance among U.S. adults. J Nutr 144, 890-901.

[53] Travassos M, Santana I, Baldeiras I, Tsolaki M, Gkatz-ima O, Sermin G, Yener GG, Simonsen A, HasselbalchSG, Kapaki E, Mara B, Cunha RA, Agostinho P, BlennowK, Zetterberg H, Mendes VM, Manadas B, de Mendon A(2015) Does caffeine consumption modify cerebrospinalfluid amyloid-� levels in patients with Alzheimer’s dis-ease? J Alzheimers Dis 47, 1069-1078.

[54] Kesse-Guyot E, Andreeva VA, Ducros V, Jeandel C, JuliaC, Hercberg S, Galan P (2014) Carotenoid-rich dietarypatterns during midlife and subsequent cognitive function.Br J Nutr 111, 915-923.

[55] Beydoun MA, Fanelli-Kuczmarski MT, Kitner-TrioloMH, Beydoun HA, Kaufman JS, Mason MA, Evans MK,Zonderman AB (2015) Dietary antioxidant intake and itsassociation with cognitive function in an ethnically diversesample of US adults. Psychosom Med 77, 68-82.

[56] Hooshmand B, Lokk J, Solomon A, Mangialasche F,Miralbell J, Spulber G, Annerbo S, Andreasen N, Winblad

B, Cedazo-Minguez A, Wahlund LO, Kivipelto M (2014)Vitamin D in relation to cognitive impairment, cere-brospinal fluid biomarkers, and brain volumes. J GerontolA Biol Sci Med Sci 69, 1132-1138.

[57] Mosconi L, Murray J, Davies M, Williams S, PirragliaE, Spector N, Tsui WH, Li Y, Butler T, Osorio RS,Glodzik L, Vallabhajosula S, McHugh P, Marmar CR,de Leon MJ (2014) Nutrient intake and brain biomark-ers of Alzheimer’s disease in at-risk cognitively normalindividuals: A cross-sectional neuroimaging pilot study.BMJ Open 4, e004850.

[58] Wilson VK, Houston DK, Kilpatrick L, Lovato J, YaffeK, Cauley JA, Harris TB, Simonsick EM, AyonayonHN, Kritchevsky SB, Sink KM (2014) Health, Agingand Body Composition Study, Relationship between 25-hydroxyvitamin D and cognitive function in older adults:The Health, Aging and Body Composition Study. J AmGeriatr Soc 62, 636-641.

[59] Littlejohns TJ, Henley WE, Lang IA, Annweiler C,Beauchet O, Chaves PH, Fried L, Kestenbaum BR, KullerLH, Langa KM, Lopez OL, Kos K, Soni M, Llewellyn DJ(2014) Vitamin D and the risk of dementia and Alzheimerdisease. Neurology 83, 920-928.

[60] Miller JW, Harvey DJ, Beckett LA, Green R, FariasST, Reed BR, Olichney JM, Mungas DM, DeCarli C(2015) Vitamin D status and rates of cognitive declinein a multi-ethnic cohort of older adults. JAMA Neurol 72,1295-1303.

[61] Granic A, Hill TR, Kirkwood TB, Davies K, CollertonJ, Martin-Ruiz C, von Zglinicki T, Saxby BK, WesnesKA, Collerton D, Mathers JC, Jagger C (2015) Serum 25-hydroxyvitamin D and cognitive decline in the very old:The Newcastle 85+Study. Eur J Neurol 22, 106-115.

[62] Assmann KE, Touvier M, Andreeva VA, Deschasaux M,Constans T, Hercberg S, Galan P, Kesse-Guyot E (2015)Midlife plasma vitamin D concentrations and performancein different cognitive domains assessed 13 years later.Br J Nutr 113, 1628-1637.

[63] Moore EM, Ames D, Mander AG, Carne RP, Brodaty H,Woodward MC, Boundy K, Ellis KA, Bush AI, Faux NG,Martins RN, Masters CL, Rowe CC, Szoeke C, Watters DA(2014) Among vitamin B12 deficient older people, highfolate levels are associated with worse cognitive function:Combined data from three cohorts. J Alzheimers Dis 39,661-668.

[64] Doets EL, Ueland PM, Tell GS, Vollset SE, Nygard OK,Van’t Veer P, de Groot LC, Nurk E, Refsum H, Smith AD,Eussen SJ (2014) Interactions between plasma concentra-tions of folate and markers of vitamin B(12) status withcognitive performance in elderly people not exposed tofolic acid fortification: The Hordaland Health Study. Br JNutr 111, 1085-1095.

[65] Kim H, Kim G, Jang W, Kim SY, Chang N (2014) Associ-ation between intake of B vitamins and cognitive functionin elderly Koreans with cognitive impairment. Nutr J 13,118.

[66] Agnew-Blais JC, Wassertheil-Smoller S, Kang JH, HoganPE, Coker LH, Snetselaar LG, Smoller JW (2015) Folate,vitamin B-6, and vitamin B-12 intake and mild cogni-tive impairment and probable dementia in the Women’sHealth Initiative Memory Study. J Acad Nutr Diet 115,231-241.

[67] Rabassa M, Cherubini A, Zamora-Ros R, Urpi-Sarda M,Bandinelli S, Ferrucci L, Andres-Lacueva C (2015) Lowlevels of a urinary biomarker of dietary polyphenol are

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 847

associated with substantial cognitive decline over a 3-yearperiod in older adults: The Invecchiare in Chianti Study.J Am Geriatr Soc 63, 938-946.

[68] Feart C, Letenneur L, Helmer C, Samieri C, Schalch W,Etheve S, Delcourt C, Dartigues JF, Barberger-Gateau P(2016) Plasma carotenoids are inversely associated withdementia risk in an elderly French cohort. J Gerontol ABiol Sci Med Sci 71, 683-688.

[69] Liu CH, Bu XL, Wang J, Zhang T, Xiang Y, Shen LL,Wang QH, Deng B, Wang X, Zhu C, Yao XQ, Zhang M,Zhou HD, Wang YJ (2016) The associations between acapsaicin-rich diet and blood amyloid-� levels and cogni-tive function. J Alzheimers Dis 52, 1081-1088.

[70] Baierle M, Vencato PH, Oldenburg L, Bordignon S,Zibetti M, Trentini CM, Duarte MM, Veit JC, SomacalS, Emanuelli T, Grune T, Breusing N, Garcia SC (2014)Fatty acid status and its relationship to cognitive declineand homocysteine levels in the elderly. Nutrients 6,3624-3640.

[71] Pottala JV, Yaffe K, Robinson JG, Espeland MA, WallaceR, Harris WS (2014) Higher RBC EPA+ DHA corre-sponds with larger total brain and hippocampal volumes:WHIMS-MRI study. Neurology 82, 435-442.

[72] Sabuncu MR, Desikan RS, Sepulcre J, Yeo BT, LiuH, Schmansky NJ, Reuter M, Weiner MW, BucknerRL, Sperling RA, Fischl B; Alzheimer’s Disease Neu-roimaging Initiative (2011) The dynamics of cortical andhippocampal atrophy in Alzheimer disease. Arch Neurol68, 1040-1048.

[73] Roberts RO, Roberts LA, Geda YE, Cha RH, PankratzVS, O’Connor HM, Knopman DS, Petersen RC (2012)Relative intake of macronutrients impacts risk of mildcognitive impairment or dementia. J Alzheimers Dis 32,329-339.

[74] Panza F, Solfrizzi V, Barulli MR, Bonfiglio C, Guerra V,Osella A, Seripa D, Sabba C, Pilotto A, Logroscino G(2015) Coffee, tea, and caffeine consumption and pre-vention of late-life cognitive decline and dementia: Asystematic review. J Nutr Health Aging 19, 313-328.

[75] Casamenti F, Stefani M (2017) Olive polyphenols: Newpromising agents to combat aging-associated neurodegen-eration. Expert Rev Neurother 17, 345-358.

[76] Morris MC, Evans DA, Tangney CC, Bienias JL, Wil-son RS, Aggarwal NT, Scherr PA (2005) Relation of thetocopherol forms to incident Alzheimer disease and tocognitive change. Am J Clin Nutr 81, 508-514.

[77] Grant WB, Karras SN, Bischoff-Ferrari HA, AnnweilerC, Boucher BJ, Juzeniene A, Garland CF, Holick MF(2016) Do studies reporting ‘U’-shaped serum 25-hydroxyvitamin D-health outcome relationships reflectadverse effects? Dermatoendocrinol 8, e1187349.

[78] Morris M, Selhub J, Jacques P (2012) Vitamin B-12 andfolate status in relation to decline in scores on the Mini-Mental State Examination in the Framingham CohortStudy. J Am Geriatr Soc 60, 1457-1464.

[79] Clarke R, Jobst KA, Refsum H, Sutton L, Ueland PM(1998) Folate, vitamin B12, and serum total homocysteinelevels in confirmed Alzheimer disease. Arch Neurol 55,1449-1455.

[80] Seshadri S, Beiser A, Selhub J, Jacques PF, Rosen-berg IH, D’Agostino RB, Wilson PW, Wolf PA (2002)Plasma homocysteine as a risk factor for dementia andAlzheimer’s disease. NEJM 289, 2651-2662.

[81] Sharma SK, Vij AS, Sharma M (2013) Mechanisms andclinical uses of capsaicin. Eur J Pharmacol 720, 55-62.

[82] Fattori V, Hohmann MS, Rossaneis AC, Pinho-Ribeiro FA,Verri WA (2016) Capsaicin: Current understanding of itsmechanisms and therapy of pain and other pre-clinical andclinical uses. Molecules 21, pii: E844.

[83] Dearborn JL, Knopman D, Sharrett AR, Schneider AL,Jack CR Jr, Coker LH, Alonso A, Selvin E, MosleyTH, Wagenknecht LE, Windham BG, Gottesman RF(2014) The meta-bolic syndrome and cognitive decline inthe Atherosclerosis Risk in Communities study (ARIC).Dement Geriatr Cogn Disord 38, 337-346.

[84] Wang M, Norman JE, Srinivasan VJ, Rutledge JC (2016)Metabolic, inflammatory, and microvascular determinantsof white matter disease and cognitive decline. Am J Neu-rodegener Dis 5, 171-177.

[85] Smallwood A, Oulhaj A, Joachim C, Christie S, Sloan C,Smith AD, Esiri M (2012) Cerebral subcortical small ves-sel disease and its relation to cognition in elderly subjects:A pathological study in the Oxford Project to InvestigateMemory and Ageing (OPTIMA) cohort. Neuropathol ApplNeurobiol 38, 337-343.

[86] Gu Y, Nieves JW, Stern Y, Luchsinger JA, Scarmeas N(2010) Food combination and Alzheimer disease risk: Aprotective diet. Arch Neurol 67, 699-706.

[87] Frisardi V, Panza F, Seripa D, Imbimbo BP, Vendemiale G,Pilotto A, Solfrizzi V (2010) Nutraceutical properties ofMediterranean diet and cognitive decline: Possible under-lying mechanisms. J Alzheimers Dis 22, 715-740.

[88] Solfrizzi V, Panza F, Frisardi V, Seripa D, LogroscinoG, Imbimbo BP, Pilotto A (2011) Diet and Alzheimer’sdisease risk factors or prevention: The current evidence.Expert Rev Neurother 11, 677-708.

[89] Whitmer RA, Sidney S, Selby J, Johnston SC, Yaffe K(2005) Midlife cardiovascular risk factors and risk ofdementia in late life. Neurology 64, 277-281.

[90] Lahiri DK (2006) Where the actions of environment (nutri-tion), gene and protein meet: Beneficial role of fruitand vegetable juices in potentially delaying the onset ofAlzheimer’s disease. J Alzheimers Dis 10, 359-361.

[91] Smith PJ, Blumenthal JA, Babyak MA, Craighead L,Welsh-Bohmer KA, Browndyke JN, Strauman TA, Sher-wood A (2010) Effects of the dietary approaches to stophypertension diet, exercise, and caloric restriction on neu-rocognition in overweight adults with high blood pressure.Hypertension 55, 1331-1338.

[92] Appel LJ, Moore TJ, Obarzanek E, Vollmer WM, SvetkeyLP, Sacks FM, Bray GA, Vogt TM, Cutler JA, Wind-hauser MM, Lin PH, Karanja N (1997) A clinical trialof the effects of dietary patterns on blood pressure. DASHCollaborative Research Group. N Engl J Med 336, 1117-1124.

[93] Blumenthal JA, Babyak MA, Sherwood A, CraigheadL, Lin PH, Johnson J, Watkins LL, Wang JT, Kuhn C,Feinglos M, Hinderliter A (2010) Effects of the dietaryapproaches to stop hypertension diet alone and in com-bination with exercise and caloric restriction on insulinsensitivity and lipids. Hypertension 55, 1199-1205.

[94] Moore TJ, Alsabeeh N, Apovian CM, Murphy MC, Coff-man GA, Cullum-Dugan D, Jenkins M, Cabral H (2008)Weight, blood pressure, and dietary benefits after 12months of a Web-based Nutrition Education Program(DASH for health): Longitudinal observational study. JMed Internet Res 10, e52.

[95] Azadbakht L, Surkan PJ, Esmaillzadeh A, Willett WC(2011) The Dietary Approaches to Stop Hypertensioneating plan affects C-reactive protein, coagulation abnor-

848 V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age

malities, and hepatic function tests among type 2 diabeticpatients. J Nutr 141, 1083-1088.

[96] Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF,Lachin JM, Walker EA, Nathan DM, Diabetes Preven-tion Program Research, Group (2002) Reduction in theincidence of type 2 diabetes with lifestyle intervention ormetformin. N Engl J Med 346, 393-403.

[97] Weaver KL, Ivester P, Seeds M, Case LD, Arm JP, ChiltonFH (2009) Effect of dietary fatty acids on inflammatorygene expression in healthy humans. J Biol Chem 284,15400-15407.

[98] Fotuhi M, Mohassel P, Yaffe K (2009) Fish consump-tion, long-chain omega-3 fatty acids and risk of cognitivedecline or Alzheimer disease: A complex association. NatClin Pract Neurol 5, 140-152.

[99] Hunt JR (2003) Bioavailability of iron, zinc, and othertrace minerals from vegetarian diets. Am J Clin Nutr 78(3Suppl), 633S-639S.

[100] Okubo H, Sasaki S, Murakami K, Takahashi Y; Freshmenin Dietetic Courses Study II group (2011) The ratio of fishto meat in the diet is positively associated with favorableintake of food groups and nutrients among young Japanesewomen. Nutr Res 31, 169-177.

[101] Loef M, Walach H (2012) Copper and iron in Alzheimer’sdisease: A systematic review and its dietary implications.Br J Nutr 107, 7-19.

[102] Brewer GJ (2015) Copper-2 ingestion, plus increased meateating leading to increased copper absorption, are majorfactors behind the current epidemic of Alzheimer’s dis-ease. Nutrients 7, 10053-10064.

[103] Grant WB (1997) Dietary links to Alzheimer’s disease.Alz Dis Rev 2, 42-55.

[104] Bonda DJ, Lee HG, Blair JA, Zhu X, Perry G, SmithMA (2011) Role of metal dyshomeostasis in Alzheimer’sdisease. Metallomics 3, 267-270.

[105] Wong BX, Hung YH, Bush AI, Duce JA (2014) Met-als and cholesterol: Two sides of the same coin inAlzheimer’s disease pathology. Front Aging Neurosci6, 91.

[106] Xu H, Perreau VM, Dent KA, Bush AI, Finkelstein DI,Adlard PA (2016) Iron regulates apolipoprotein E expres-sion and secretion in neurons and astrocytes. J AlzheimersDis 51, 471-487.

[107] Prakash A, Dhaliwal GK, Kumar P, Majeed AB (2017)Brain biometals and Alzheimer’s disease - boon or bane?Int J Neurosci 127, 99-108.

[108] Tsao R (2010) Chemistry and biochemistry of dietarypolyphenols. Nutrients 2, 1231-1246.

[109] Thangthaeng N, Poulose SM, Miller MG, Shukitt-HaleB (2016) Preserving brain function in aging: The anti-glycative potential of berry fruit. Neuromolecular Med 18,465-473.

[110] Heim KC, Angers P, Leonhart S, Ritz BW (2012) Anti-inflammatory and neuroactive properties of selected fruitextracts. J Med Food 15, 851-854.

[111] Rendeiro C, Rhodes JS, Spencer JP (2015) The mecha-nisms of action of flavonoids in the brain: Direct versusindirect effects. Neurochem Int 89, 126-139.

[112] Collins SM1, Surette M, Bercik P (2012) The interplaybetween the intestinal microbiota and the brain. Nat RevMicrobiol 10, 735-742.

[113] Pistollato F, Sumalla Cano S, Elio I, Masias VergaraM, Giampieri F, Battino (2016) Role of gut microbiotaand nutrients in amyloid formation and pathogenesis ofAlzheimer disease. Nutr Rev 74, 624-634.

[114] Claesson MJ, Jeffery IB, Conde S, Power SE, O’ConnorEM, Cusack S, Harris HM, Coakley M, Lakshmi-narayanan B, O’Sullivan O, Fitzgerald GF, Deane J,O’Connor M, Harnedy N, O’Connor K, O’Mahony D, vanSinderen D, Wallace M, Brennan L, Stanton C, MarchesiJR, Fitzgerald AP, Shanahan F, Hill C, Ross RP, O’ToolePW (2012) Gut microbiota composition correlates withdiet and health in the elderly. Nature 488, 178-184.

[115] Scott KP, Gratz SW, Sheridan PO, Flint HJ, Duncan SH(2013) The influence of diet on the gut microbiota. Phar-macol Res 69, 52-60.

[116] Friedland RP (2015) Mechanisms of molecular mimicryinvolving the microbiota in neurodegeneration. JAlzheimers Dis 45, 349-362.

[117] Bhattacharjee S, Lukiw WJ (2015) Alzheimer’s diseaseand the microbiome. Front Cell Neurosci 7, 153.

[118] Schwartz K, Boles BR (2013) Microbial amyloids–functions and interactions within the host. Curr OpinMicrobiol 16, 93-99.

[119] Daulatzai MA (2014) Chronic functional bowel syndromeenhances gut-brain axis dysfunction, neuroinflammation,cognitive impairment, and vulnerability to dementia. Neu-rochem Res 39, 624-644.

[120] Prasanthi JR, Dasari B, Marwarha G, Larson T, ChenX, Geiger JD, Ghribi O (2010) Caffeine protects againstoxidative stress and Alzheimer’s disease-like pathology inrabbit hippocampus induced by cholesterol-enriched diet.Free Radic Biol Med 49, 1212-1220.

[121] Chen X, Ghribi O, Geiger JD (2010) Caffeine protectsagainst disruptions of the blood-brain barrier in ani-mal models of Alzheimer’s and Parkinson’s diseases. JAlzheimers Dis 20, S127-S141.

[122] Fredholm BB, Battig K, Holmen J, Nehlig A, ZvartauEE (1999) Actions of caffeine in the brain with specialreference to factors that contribute to its widespread use.Pharmacol Rev 51, 83-133.

[123] Cao C, Cirrito JR, Lin X, Verges DK, Dickson A, Mam-carz M, Zhang C, Mori T, Arendash GW, Holtzman DM,Potter H (2009) Caffeine suppresses amyloid-beta levels inplasma and brain of Alzheimer’s disease transgenic mice.J Alzheimers Dis 17, 681-697.

[124] Zeitlin R, Patel S, Burgess S, Arendash GW, EcheverriaV (2011) Caffeine induces beneficial changes in PKA sig-naling and JNK and ERK activities in the striatum andcortex of Alzheimer’s transgenic mice. Brain Res 1417,127-136.

[125] Dragicevic N, Delic V, Cao C, Copes N, Lin X, MamcarzM, Wang L, Arendash GW, Bradshaw PC (2012) Caffeineincreases mitochondrial function and blocks melatoninsignaling to mitochondria in Alzheimer’s mice and cells.Neuropharmacology 63, 1368-1379.

[126] Dror Y, Stern F, Gomori MJ (2014) Vitamins in the preven-tion or delay of cognitive disability of aging. Curr AgingSci 7, 187-213.

[127] Feng Y, Wang X (2012) Antioxidant therapies forAlzheimer’s disease. Oxid Med Cell Longev 2012, 472932.

[128] Yamada K, Tanaka T, Han D, Senzaki K, Kameyama T,Nabeshima T (1999) Protective effects of idebenone andalpha-tocopherol on beta-amyloid-(1–42)-induced learn-ing and memory deficits in rats: Implication of oxidativestress in beta-amyloid-induced neurotoxicity in vivo. EurJ Neurosci 11, 83-90.

[129] Yu BP, Kang CM, Han JS, Kim DS (1998) Can antioxi-dant supplementation slow the aging process? Biofactors7, 93101.

V. Solfrizzi et al. / Diet and Cognitive Disorders in Older Age 849

[130] Spencer JP, Vauzour D, Rendeiro C (2009) Flavonoids andcognition: The molecular mechanisms underlying theirbehavioural effects. Arch Biochem Biophys 492, 19.

[131] Garcia A, Zanibbi K (2004) Homocysteine and cognitivefunction in elderly people. CMAJ 171, 897-904.

[132] Balion C, Griffith LE, Strifler L, Henderson M, PattersonC, Heckman G, Llewellyn DJ, Raina P (2012) Vitamin D,cognition, and dementia: A systematic review and meta-analysis. Neurology 79, 1397-405.

[133] Dickens AP, Lang IA, Langa KM, Kos K, Llewellyn DJ(2011) Vitamin D, cognitive dysfunction and dementia inolder adults. CNS Drugs 25, 629-639.

[134] Lu’o’ng KV, Nguyen LT (2011) The beneficial role ofvitamin D in Alzheimer’s disease. Am J Alzheimers DisOther Demen 26, 511-520.

[135] Schubert R, Kitz R, Beermann C, Rose MA, Baer PC,Zielen S, Boehles H (2007) Influence of low-dose polyun-saturated fatty acids supplementation on the inflammatoryresponse of healthy adults. Nutrition 23, 724-730.

[136] Chalon S (2006) Omega-3 fatty acids and monoamine neu-rotransmission. Prostaglandins Leukot Essent Fatty Acids75, 259-269.

[137] Lukiw WJ, Bazan NG (2008) Docosahexaenoic acid andthe aging brain. J Nutr 138, 2510-2514.

[138] Sydenham E, Dangour AD, Lim W-S (2012) Omega 3 fattyacid for the prevention of cognitive decline and dementia.Cochrane Database Syst Rev 6, CD005379.

[139] Lim GP, Calon F, Morihara T, Yang F, Teter B, Ubeda O,Salem N Jr, Frautschy SA, Cole GM (2005) A diet enrichedwith the omega-3 fatty acid docosahexaenoic acid reducesamyloid burden in an aged Alzheimer mouse model. JNeurosci 25, 3032-3040.

[140] Lichtenstein AH, Schwab US (2000) Relationship ofdietary fat to glucose metabolism. Atherosclerosis 150,227-243.

[141] Luchsinger JA, Tang MX, Shea S, Mayeux R (2004)Hyperinsulinemia and risk of Alzheimer disease. Neurol-ogy 63, 1187-1192.

[142] Snowden SG, Ebshiana AA, Hye A, An Y, Pletnikova O,O’Brien R, Troncoso J, Legido-Quigley C, Thambisetty M(2017) Association between fatty acid metabolism in thebrain and Alzheimer disease neuropathology and cogni-tive performance: A nontargeted metabolomic study. PLoSMed 14, e1002266.

[143] Giem P, Beeson WL, Fraser GE (1993) The incidence ofdementia and intake of animal products: Preliminary find-ings from the Adventist Health Study. Neuroepidemiology12, 28-36.

[144] Barberger-Gateau P, Letenneur L, Deschamps V, Peres K,Dartigues JF, Renaud S (2002) Fish, meat, and risk ofdementia: Cohort study. BMJ 325, 932-933.

[145] Armstrong B, Doll R (1975) Environmental factors andcancer incidence and mortality in different countries, withspecial reference to dietary practices. Int J Cancer 15,617-631.

[146] Kono S (2004) Secular trend of colon cancer incidenceand mortality in relation to fat and meat intake in Japan.Eur J Cancer Prev 13, 127-132.

[147] Zhang J, Dhakal IB, Zhao Z, Li L (2012) Trends in mortal-ity from cancers of the breast, colon, prostate, esophagus,and stomach in East Asia: Role of nutrition transition. EurJ Cancer Prev 21, 480-489.

[148] Aune D, De Stefani E, Ronco A, Boffetta P, Deneo-Pellegrini H, Acosta G, Mendilaharsu M (2009) Meatconsumption and cancer risk: A case-control study inUruguay. Asian Pac J Cancer Prev 10, 429-436.

[149] Grant WB (2014) Trends in diet and Alzheimer’s diseaseduring the nutrition transition in Japan and developingcountries. J Alzheimers Dis 38, 611-620.

[150] Perrone L, Grant WB (2015) Observational and ecolog-ical studies of dietary advanced glycation end productsin national diets and Alzheimer’s disease incidence andprevalence. J Alzheimers Dis 45, 965-979.

[151] Grant WB (2016) Using multicountry ecological andobservational studies to determine dietary risk factors forAlzheimer’s disease. J Am Coll Nutr 35, 476-489.