You are on page 1of 13

SYSTEMATIC REVIEW

Relationship Between Diet and Mental Health in Children


and Adolescents: A Systematic Review
We systematically re- Adrienne O’Neil, BA(Psych/Soc)(Hons), PhD, Shae E. Quirk, BAppSci(Psych), GradDipPsych, Siobhan Housden, MA
viewed 12 epidemiological (Hons), Sharon L. Brennan, BA(Hons), PhD, GCALL, Lana J. Williams, BPsych, GradDipAppPsych, PhD, Assoc MAPS,
studies to determine whether Julie A. Pasco, BSc(Hons), Dip Ed, PhD, MEpi, Michael Berk, MBBCh, PhD, and Felice N. Jacka, PgDipSci, PhD
an association exists between
diet quality and patterns and
THE ROLE OF HABITUAL DIET IN depression and anxiety, examina- 19 years or younger on enrolment
mental health in children and
the development of depressive tion of these same mental health (as per the United Nations13 defini-
adolescents; 9 explored the
disorders and symptoms has be- parameters in children and ado- tion); and (4) used nonclinical study
relationship using diet as
the exposure, and 3 used come a recent research focus over lescents is needed. In terms of samples that were population based
mental health as the expo- the past decade. Data from adult what evidence is available to date, rather than from acute or institu-
sure. populations have indicated that findings remain inconsistent. For tional settings. We included studies
We found evidence of a better-quality diet is associated example, although some studies defining dietary patterns using ei-
significant, cross-sectional re- with better mental health out- have observed a dose---response ther a priori or a posteriori ap-
lationship between unhealthy comes.1---5 In fact, new meta- relationship between diet quality
proaches, as well as studies using
dietary patterns and poorer analyses have confirmed the in- and mental health in young ado-
mental health in children and proxy measures of diet quality (e.g.,
verse association between healthy lescents,10 others have shown no
adolescents. We observed a studies that derived a diet quality
diets and depression.4,5 A habitu- significant association.11 The evi-
consistent trend for the re- score from food frequency data).
ally poor diet (e.g., increased con- dence is even less comprehensive
lationship between good- In the absence of a standard
quality diet and better men- sumption of Western processed for the relationship between dietary
definition of diet quality or diet
tal health and some evidence foods) is also independently asso- intake and anxiety symptoms.12
patterns, we defined them as the
for the reverse. When in- ciated with a greater likelihood of To our knowledge, no system-
quality of overall habitual dietary
cluding only the 7 studies or risk for depression1,6,7 and atic reviews to date have specifi-
deemed to be of high meth- anxiety.1 Although stress and de- cally investigated the association intake or the pattern of overall
odological quality, all but 1 pression can promote unhealthy between diet, measured using diet habitual dietary intake, as previ-
of these trends remained. eating, recent longitudinal studies quality scores, dietary pattern ously reported.1,14,15 Although
Findings highlight the po- have suggested that reverse cau- varying in composition according
analysis, or both and internalizing
tential importance of the re- to the country of origin, healthy
sality is a less likely explanation for behaviors that characterize low or
lationship between dietary
long-term associations.8 depressive mood and anxiety or prudent dietary patterns are
patterns or quality and men-
tal health early in the life However, our understanding of symptoms in child and adolescent characterized by a higher intake of
span. (Am J Public Health. these associations earlier in the life populations. nutrient-dense foods, including
2014;104:e31–e42. doi:10. span remains unclear. To date, vegetables, salads, fruits, fish, and
2105/AJPH.2014.302110) much of the research around this METHODS other foods groups known to be
relationship has focused on die- healthful. Conversely, unhealthy
tary intake and externalizing be- Studies considered for inclusion patterns are characterized by
haviors (particularly hyperactiv- in this review (1) were full-text a higher intake of foods with
ity). For example, poor nutritional articles; (2) consisted of epidemio-
increased saturated fat, refined
quality is independently associ- logical cohort, case-control, and
carbohydrates, and processed
ated with symptoms of attention- cross-sectional study designs; (3)
food products. We excluded
deficit hyperactivity disorder.9 examined associations between diet
studies that
However, the relationship be- quality or patterns and internalizing
tween dietary intake in childhood disorders that encompassed de- 1. examined individual nutri-
and adolescence and internalizing pression, low mood, depressive ents or supplements,
behaviors, which represent de- symptoms, emotional problems, 2. examined the effects of pre-
pressive symptoms, low mood, or and anxiety (as distinct from exter- servatives on mental health,
anxiety, has received comparably nalizing disorders), assessed via 3. examined emotional or binge
less attention. Given that the pre- self- or informant report, medical eating,
vious literature in adults regarding records, or the application of di- 4. examined dietary restraint
diet and mental health has focused agnostic measurement tools in chil- or restriction (i.e., the restric-
on the common mental disorders, dren or adolescents who were aged tion of calories or food

October 2014, Vol 104, No. 10 | American Journal of Public Health O’Neil et al. | Peer Reviewed | Systematic Review | e31
SYSTEMATIC REVIEW

consumption for the purpose then another author (A. O.) repli-
of weight loss), cated the search. The abstracts or Criteria list for the assessment of study quality
5. used trait-based (assessing full-text articles of those studies (modified from Lievense et al.17)
personality) or measures of deemed potentially relevant were
Study population
stress (as distinct from inter- obtained. We (A. O. and S. H.)
1. Selection at uniform point C/CC/CS
nalizing behaviors), conferred to finalize the articles to
2. Cases and controls drawn from the same population CC
6. presented only univariate be included in the review accord-
3. Participation rate > 80% for cases/cohort C/CC
analyses, and ing to the predetermined inclusion
4. Participation rate > 80% for controls CC
7. were published in languages and exclusion criteria; where con-
Assessment of risk factor
other than English. sensus was not reached, the senior
5. Exposure assessment blinded C/CC/CS
author (F. N. J.) was consulted.
6. Exposure measured identically for cases and controls CC
We also excluded studies that
7. Exposure assessed according to validated measures C/CC/CS
used samples in which the age range Assessment of
Assessment of outcome
overlapped between adolescence Methodological Quality and
8. Outcome assessed identically in studied population C/CC/CS
and adulthood and studies that Best-Evidence Synthesis
9. Outcome reproducibly C/CC/CS
presented mental health data only The heterogeneity between
10. Outcome assessed according to validated measures C/CC/CS
as a composite measure (e.g., overall study definitions of dietary and
Study design
behavior scores as distinct from in- internalizing symptomatology
11. Prospective design used C/CC
ternalizing behavior scores) or as variables precluded formal meta-
12. Follow-up time > 12 months C
a comorbidity only. Furthermore, analysis. As such, we determined
13. Withdrawals < 20% C
given that the purpose of this review a priori that included studies
Analysis and data presentation
was to assess whether a relationship would be analyzed by assessing
14. Appropriate analysis techniques used C/CC/CS
exists between diet quality or di- their methodological quality and
15. Adjusted for at least age, and gender C/CC/CS
etary patterns and mental health performing a best-evidence syn-
(rather than examining dietary in- thesis of those studies meeting Note. C = applicable to cohort studies; CC = applicable to case-control studies; CS =
tervention effects), we also excluded quality standards. applicable to cross-sectional studies. Studies were scored as positive (1), negative (0),
or unclear (?), and 100% represented the maximum possible score.
studies with a randomized con- To assess the methodological
trolled trial design. quality of the reviewed studies, we
used a scoring system based on
Search Strategy and Data that of Lievense et al.,16 which has subsequently deemed high quality evidence, ranging from strong
Extraction been used in musculoskeletal and were those with a score exceeding evidence, moderate evidence, lim-
We performed the search strat- obesity research and recently in the mean of all the total scores. The ited evidence, and conflicting evi-
egy using medical, health, psychiat- another review article examining mean score was 83.7% (range = dence to no evidence (Table 1).
ric, and social sciences databases the relationship between dietary 62.5%---100%). The scoring system The synthesis took into account
(PubMed, OVID, MEDLINE, variables and depression in is preferential to cohort studies, and the type of study design used
CINAHL, PsycINFO) to identify rel- adults.15,17,18 We (S. E. Q. and S. H.) this is reflected by cohort studies (i.e., strong evidence was defined
evant literature published through independently scored 14 items re- being eligible for a greater number by generally consistent findings in
August 30, 2012. We used the lating to the methodological quality of criteria (e.g., related to prospec- multiple high-quality cohort stud-
following search terms: (depression of studies in the following areas: tive study designs, participation and ies). This systematic review ad-
OR depressive disorder OR anxiety study population, assessment of risk attrition rates) than case-control hered to the guidelines outlined in
disorders OR affective symptoms factors, assessment of outcomes, and cross-sectional study designs. the 2009 Preferred Reporting
OR anxi* OR mood OR internali* study design, and data analysis (see In the case that we (S. E. Q. and Items for Systematic Reviews and
OR psychological symptoms OR the box on this page). If they en- S. H.) did not agree on the ratings, Meta-Analyses statement.19
psychological distress) AND (diet dorsed an item, a positive score of 1 a co-author experienced in best
OR food habits OR dietary OR was applied; if they did not endorse evidence synthesis (S. L. B.) pro- RESULTS
dietary patterns OR dietary quality the item, a score of 0 was applied. In vided the final judgment in 1 con-
OR western diet OR Mediterranean cases in which information was sensus meeting. Applying the initial search
diet) AND (youth OR adolescen* missing or was insufficiently docu- Our best-evidence synthesis in- strategy identified 1255 citations;
OR child OR infant). mented, the item was coded as cluded those studies that met we subsequently excluded 242
We also searched reference lists unclear (?), and a score of 0 was high-quality standards as we have because of duplication, leaving
of relevant reviews and studies. subsequently applied. We calcu- defined them. We performed the 1013 potentially relevant studies.
One author (S. H.) performed the lated the total score (percentage) for synthesis by ranking the findings Of those, we excluded 991 on the
electronic search strategy, and each of the studies, and those across all studies into 5 levels of basis of information available

e32 | Systematic Review | Peer Reviewed | O’Neil et al. American Journal of Public Health | October 2014, Vol 104, No. 10
SYSTEMATIC REVIEW

TABLE 1—Criteria for Ascertainment of Evidence Level for Identified citations from
Best-Evidence Synthesis
PubMed, PsycInfo, OVID,
Level of Evidence Criteria for Inclusion in Best Evidence Synthesis MEDLINE, and CINAHL (n = 1255)
Strong evidence Generally consistent findings in multiple high-quality
cohort studies Duplicates (n = 242)
Moderate evidence Generally consistent findings in 1 high-quality cohort
study and > 2 high-quality case-control studies
Titles and abstracts reviewed
Limited evidence Generally consistent findings in single cohort study, 1 or 2
case-control studies, or multiple cross-sectional studies (n = 1013)
Conflicting evidence Inconsistent findings in < 75% of the studies
No evidence No studies found
Excluded (not relevant)
Note. Adapted from Lievense et al.20 (n = 991)

from the abstract and title. We Population and Design


obtained the full texts of the We extracted the following key
Complete articles reviewed (n = 22)
remaining 22 articles to assess information from those articles
eligibility. Additionally, we manu- eligible for inclusion: author, coun-
ally examined reference lists and try, sample, diet measure, mental Additional articles reviewed
citations for further relevant stud- health measure, statistical presenta-
Identified from reference lists (n = 2)
ies, revealing 7 new full-text arti- tion of results (including exposure
cles (n = 29). On examination of variable), covariates, and key find-
Identified through citations (n = 5)
the full-text articles, 14 studies did ings. Key characteristics of included
not fulfill inclusion criteria and articles are displayed in Table 2.
were subsequently excluded. We Briefly, studies were from Aus- Did not meet criteria (n = 17)
(A. O. and S. H.) achieved a high tralia,10,22---24,27 the United
level of consensus (89%); decisions States,21,26 the United Kingdom,20
surrounding the remaining articles Germany,25 China,12 Canada,11 and
were reached in consultation with Norway.28
the senior author (F. N. J). As a re- Collectively, studies included
Articles included in review (n = 12)
sult, a further 3 articles were ex- 82 779 participants (Oddy et al.22
cluded (investigated stress or neu- [n = 1324] and Robinson et al.24
roticism or parental restrictions on [n = 1860] used the same data
foods), thus leaving 12 articles for set). Data were derived predomi-
inclusion in this review. Figure 1 nantly from cross-sectional studies FIGURE 1—Flowchart summary of search results.
displays a summary of the results of and 3 prospective cohort studies.
the systematic search. For the latter, follow-up assess- to measure mental health were sleeping, feeling unhappy, sad, de-
The most common reasons for ment periods ranged from 2 to 4 subscales of the Child Behavior pressed, hopeless, nervous or tense,
exclusion were (1) “diet” defined years. Age of participants ranged Checklist22,24,28 and the Strengths and worrying too much about
as skipping meals, caloric con- from 4.5 to 18 years in all studies and Difficulties Question- things26; the Pediatric Quality of
trol, or binge eating; (2) results (Table 2). naire.20,25,27 Other instruments in- Life Inventory (n =1)10; and a ques-
taken from participants enrolled cluded the Short Mood and Feel- tionnaire about frequency of feeling
in an intervention study, com- Measures of Mental Health ings Questionnaire (n = 1)23; the depressed (n =1).21
munity- or school-based pro- Table 2 displays the instru- Depression Self-rating Scale for
gram, or both; (3) study explored ments used to measure outcome Children (n = 1)12; physician diag- Measures of Diet Quality and
the role of stress or well-being as and exposure variables. We pres- nosis using health records (applying Patterns
opposed to psychological out- ent studies in which mental health the International Classification of Dietary intake was most com-
comes; or (4) study investigated was treated as the exposure vari- Diseases, Ninth Revision; n =1)11; monly measured using variations
the impact of parental practices able separately from studies ex- a 6-item checklist (n =1) consisting of a Food Frequency Question-
or attitudes related to provision ploring diet as the exposure. The of symptoms of depression includ- naire (FFQ),12,20,25 including
of food. most commonly used instruments ing feeling tired, having trouble the Harvard Youth/Adolescent

October 2014, Vol 104, No. 10 | American Journal of Public Health O’Neil et al. | Peer Reviewed | Systematic Review | e33
TABLE 2—Characteristics of 12 Included Studies, Grouped by Study Design, Author, and Exposure of Interest
No. Participants, Sex (%), Age at Recruitment, Years,
Author and Country and Follow-Up Period (If Applicable) Mean (SD) or Range Dietary Measure Mental Health Measure Questionnaire Respondent

Cohort, prospective studies


Wiles et al.,20 United Kingdom 4541,a 2.5 y 4.5b FFQ; principal components analysis Emotional Symptoms subscale of Mother
to identify junk food dietary SDQ
pattern
Jacka et al.,10 Australia 3040, 56% male, 2 y 11–18 Dietary questionnaire used to Emotional Functioning subscale of Adolescent
construct “healthy diet” score the PedsQL
based on National Healthy Eating
Guidelines, “unhealthy diet” score
constructed by summing
consumption of unhealthy foods

e34 | Systematic Review | Peer Reviewed | O’Neil et al.


McMartin et al.,11 Canada 3757, 48% male, 3 y 10–11 Harvard Youth/Adolescent FFQ, Health records of physician- Child
overall DQI-I score (0–100), diagnosed internalizing disorders
variety component score of the (ICD-9)
DQI-I
Cross-sectional studies
Brooks et al.,21 United States 2224, 52% male 16.2b Dichotomized question: “Do you eat Frequency of feeling depressed or Adolescents
a healthy diet?” distressed in the past 30 d,
dichotomized cutoff score ‡ 10
Oddy et al.,22 Australia 1324, 51.2% male 14.1 (0.2) CSIRO FFQ, healthy and Western Internalizing subscale of CBCL Parent
dietary patterns based on factor
analysis
SYSTEMATIC REVIEW

Jacka et al.,23 Australia 7114, 47.2% male 11.6 (0.81) Dietary questionnaire adapted from SMFQ; dichotomized cutpoint score
Amherst Health and Activity Study of ‡ 8 defined as symptomatic
Adult Survey of Child Health; diet
quality score divided into quintiles
for unhealthy and healthy diets
Jacka et al.,10 Australia 3040, 56% male 11–18 Dietary questionnaire, healthy diet Emotional functioning subscale of Adolescent
and unhealthy diet scores the PedsQL
constructed on the basis of
National Healthy Eating Guidelines
Robinson et al.,24 Australia 1598,a 51.3% male ;14b CSIRO FFQ; food groups: cereals Internalizing subscale of CBCL Parent
and grains, fruit, dairy products,
meat and meat alternatives,
vegetables, and extras based on
Australian dietary
recommendations

Continued

American Journal of Public Health | October 2014, Vol 104, No. 10


TABLE 2—Continued

Weng et al.,12 China 5003, 52.09% male 13.21 (0.99) FFQ, 38 items, principal component DSRS (Chinese version), cutoff score Adolescent
analysis produced snack, animal, of 15 defined as screening
and traditional dietary patterns, positive for depressive symptoms;
divided into quintiles SCARED (Chinese version), cutoff
score of > 23 defined as screening
positive for an anxiety disorder
Kohlboeck et al.,25 Germany 3361, 62.8% male 11.15 (0.5) FFQ, 82 items, constructed diet Emotional Symptoms subscale of Parent
quality score based on the SDQ
concept of the German OMD for
children and adolescents, food
categories defined according to
the Codex General Standard for
Food Additives food category
system
Fulkerson et al.,26 United States 4734,b 50.2% male Boys: 14.9b YAQ, FFQ, 149 items Constructed 6-item scale of the Adolescent

October 2014, Vol 104, No. 10 | American Journal of Public Health


Girls: 14.7b following symptoms: feeling tired;
having trouble sleeping; feeling
unhappy, sad, depressed,
hopeless, nervous or tense,
worrying too much about things,
stratified depressive symptom
groups (low, moderate, high)
Renzaho et al.,27 Australia 3370, 51.8% male 4–12 Questionnaire about fruit and Emotional Symptoms subscale of Parent or caregiver
SYSTEMATIC REVIEW

vegetable consumption according SDQ


to Australian Guide to Healthy
eating
Vollrath et al.,28 Norway 42 451 ;1.5b Dietary questionnaire, Internalizing subscale of CBCL, 14 Mother
questionnaire consisting of 4 items; EAS, 11 items; mean score
items based on consumption of derived from 25 items of
cakes, waffles, or sweet cookies; combined scales
desserts or ice cream; chocolate;
and other sweets, jellybeans, or
confectionary; summary score
dichotomized at 85th percentile

Note. CBCL = Child Behavior Checklist; CSIRO = Commonwealth Scientific and Industrial Research Organization; DQI-I = Diet Quality Index—International; DSRS = Depression Self-Rating Scale for Children; EAS = Emotionality Activity and
Sociability Questionnaire; FFQ = Food Frequency Questionnaire; ICD-9 = International Classification of Diseases, 9th Revision; OMD = optimized mixed diet; PedsQL = Pediatric Quality of Life Inventory; SCARED = Screen for Child Anxiety
Related Emotional Disorders; SDQ = Strengths and Difficulties Questionnaire; SMFQ = Short Mood and Feelings Questionnaire; YAQ = Youth and Adolescent Food Frequency Questionnaire.
a
Number of participants for which data are available on both mental health measure and dietary measure,
b
Data not provided.

O’Neil et al. | Peer Reviewed | Systematic Review | e35


SYSTEMATIC REVIEW

Questionnaire (n = 2)11,26 and the demonstrated a significant rela- DISCUSSION affect. Indeed, there are numerous
Commonwealth Scientific and In- tionship between higher diet potential biological pathways by
dustrial Research Organization quality (i.e., higher intakes of Our aim was to review and which diet quality may have an
FFQ (n = 2).22,24 Other measures healthy, nutrient-dense foods) synthesize the existing literature to impact on mental health in chil-
included a 14-item questionnaire and better mental health (Brooks determine whether an association dren and adolescents. First, a
based on the Amherst Health and et al.21 for females only). Of the exists between diet quality and poor quality diet that is lacking
Activity Study Adult Survey of 3 studies that looked specifically mental health in childhood or nutrient-dense foods may lead to
Child Health Habits (n = 1),23 the at the association between adolescence, with a focus on in- nutrient deficiencies that have
German optimized mixed diet lower diet quality (i.e., higher in- ternalizing disorders including been associated with mental
concept for children and adoles- take of unhealthy foods) and depression, low mood, and anxi- health issues. For example, the
cents (n = 1),25 the Australian poorer mental health outcomes, ety. We observed consistent dietary intake of folate, zinc, and
Guide to Healthy Eating by Ques- 2 also reported significant cross-sectional associations be- magnesium is inversely associated
tionnaire (n = 2),1,27 4 questions relationships. tween unhealthy dietary patterns with depressive disorders,30
about frequency of sweet con- A paucity of available prospec- and worse mental health in child- whereas dietary long-chain
sumption (n = 1),28 a self-report tive studies (n = 3) investigated hood or adolescence. In contrast, omega-3 fatty acids are inversely
questionnaire about nutrition the association between dietary we found inconsistent trends for related to anxiety disorders.31
(n = 1),10 and the question “Do patterns or quality and mental Dietary intake may also have
the relationships between healthy
you eat a healthy diet?” (n = 1).21 health1,10,11,20; where evidence was a direct impact on various biolog-
diet patterns or quality and better
available, findings were conflicting ical systems and mechanisms that
mental health. We also found in-
Key Findings Including Data (Table 3). underpin depression, including
consistent trends for unhealthy
From All Studies When the relationship be- oxidative processes, the function-
diet quality and worse mental
Key results of the 12 studies tween mental health and diet ing of the immune system, and
health. Overall, best-evidence cri-
reviewed are provided in Table 3; quality was explored using mental levels of salient brain proteins. For
teria confirmed that this area had
cross-sectional and prospective health as the exposure variable, example, in patients with depres-
a limited level of evidence, largely
analyses performed within a study data were also limited. However, sion, markers of systemic inflam-
attributable to a dearth of pro-
are presented separately. Of the 2 of 3 of these studies (66%) mation are often significantly
spective and case-control data,
12 studies, 9 explored the rela- demonstrated that children and greater than in controls, which is
which thereby precludes us from
tionship between diet quality and adolescents with worse mental indicative of immune system dys-
inferring causal associations about
mental health using diet as the health reported significantly regulation.32 Studies have indi-
these relationships.
exposure variable1,10---13, 20---25; of poorer dietary patterns.27,28 No cated that markers of inflamma-
These findings add to the exist-
these 9 studies with diet as the data were available on the rela- tion are positively correlated with
ing literature that has attempted to
exposure, 5 explored the relation- tionship between mental health as components of a poor diet, and
ship between dietary patterns and elucidate the relationship between
an exposure and its relationship a healthy diet is associated with
mental health and 3 explored the to measures of healthy dietary diet quality and mental health in reduced inflammation.33 The
relationship between diet quality habits. adult populations.6,7,15,29 To our available evidence also suggests
and mental health; 1 explored both knowledge, this is the first review that high-fat, high-sugar diets can
diet quality, dietary patterns and Best-Evidence Synthesis of its kind in this area to focus affect proteins that are important
mental health.25 Of those exam- When we applied criteria for specifically on children and ado- in brain development, such as the
ining dietary patterns as the ex- the best-evidence synthesis, the lescents. Although this review signaling molecule brain-derived
posure, the majority of studies mean score was 83.7% (range = generated insufficient evidence to neurotrophic factor.34 Brain-
(n = 4) consistently demonstrated 62.5%---100%, where 100% is elucidate the directionality of the derived neurotrophic factor is of-
significant relationships between the maximum obtainable score). relationship, several potential ex- ten reduced in patients with de-
unhealthy dietary patterns and Methodological quality ratings of planations exist for the relation- pression,35 and when its synthesis
poorer mental health. Evidence of each study are displayed in Table ship between diet and mental is increased, symptoms of depres-
an association between healthy 3. The 7 studies exceeding the health in this population. sion can improve.36 It is important
dietary patterns and better mental mean were subsequently in- It may be the case that children to note that consistent evidence
health was less consistent, with cluded in a best-evidence syn- and adolescents with internalizing has shown that higher quality diets
significant positive associations thesis. When we applied the disorders or symptoms eat more (i.e., those higher in nutrient-dense
observed in only half of the 6 criteria for ascertainment of evi- poorly as a form of self-medication. foods) and diets high in saturated
studies. Of the 5 studies exploring dence level for best-evidence However, it is equally as conceiv- fats and refined carbohydrates are
the association between diet qual- synthesis, we deemed the level able that the influence of early each independently related to de-
ity, measured using diet quality of evidence for all of these eating habits and nutritional in- pression, suggesting the possibility
scores, and mental health, all associations as limited. take has an important impact on of different operant pathways. The

e36 | Systematic Review | Peer Reviewed | O’Neil et al. American Journal of Public Health | October 2014, Vol 104, No. 10
TABLE 3—Key Results of Included Articles, Including Summary, Covariates, and Methodological Quality Score
Results
Author Key Results Summary Covariates Quality Score, %

Cohort, Prospective Studies


Wiles et al.20 OR = 1.01 (CI = 0.94, 1.09) No association between junk food dietary Sex, SDQ total difficulties or subscale 75
pattern at age 4.5 and emotional score at age 4.5 y, maternal smoking,
problems at age 7 (dietary pattern) maternal age at birth of child, number of
siblings, socioeconomic markers, birth
weight and gestational age, maternal
depression and anxiety, maternal
enjoyment score, and single-parent
household
Jacka et al.10 C2: b = 0.11* (CI = 0.01, 0.21) Dose–response prospective association Age, sex, area-level SES, dieting behaviors, 75
C3: b = 0.14* (CI = 0.02,0.27) between higher baseline healthy diet BMI, physical activity, baseline PedsQL
C2: b = 0.05 (CI = –0.04,0.27) quality scores and higher PedsQL scores scores
C3: b = –0.07 (CI = –0.18, 0.03) at 2-y follow-up (dietary quality)

October 2014, Vol 104, No. 10 | American Journal of Public Health


No prospective association between higher
unhealthy diet quality scores and lower
PedsQL (dietary quality) after controlling
for baseline mental health
McMartin et al.11 IRR = 1.09 (CI = 0.73, 1.63) No association between “overall” DQI-I Sex, energy intake, household income, 91.7
diet quality scores and rates of parental marital status and education,
internalizing disorders (dietary quality) body weight status, physical activity level,
IRR = 0.45* (CI = 0.25–0.82) Greater “variety” component DQI-I diet geographic area
SYSTEMATIC REVIEW

quality scores associated with lower rates


of internalizing disorders over 3-y follow-
up period (dietary quality)
Cross-Sectional Studies
Brooks et al.21 Males: OR = 1.07 (CI = 0.98,1.18); No association between a healthy diet and Age, race 62.5
Females: OR = 0.89* (CI = 0.83,0.96) feeling depressed or stressed for males
Eating a healthy diet was associated with
reduced odds of feeling depressed or
stressed for females
Oddy et al.22 b = 1.25* (CI = 0.15, 2.35) A Western dietary pattern was associated Sex, total energy intake, BMI category, 100
b = 0.17 (CI = –0.54, 0.88) with higher CBCL internalizing scores physical activity, screen use, family
(dietary pattern) structure, family income, family
No association between healthy dietary functioning at age 14 and maternal
pattern and CBCL internalizing scores education at pregnancy
(dietary pattern)

Continued

O’Neil et al. | Peer Reviewed | Systematic Review | e37


TABLE 3—Continued

Jacka et al.23 Age, sex, physical activity, parental work 75


and educational status, household level,
SES, dieting behaviors, family conflict
and poor family management, BMI,
smoking, and physical activity
Unhealthy diet Q2: OR = 1.03 (CI = 0.87, 1.22) Dose–response association between higher
Q3: OR = 1.22 (CI = 1.03, 1.44) unhealthy diet scores and higher odds of
Q4: OR = 1.29 (CI = 1.12, 1.50) being depressed (SMFQ score; dietary
Q5: OR = 1.79 (CI = 1.52, 2.11) quality)
Healthy diet Q2: OR = 0.61 (CI = 0.45, 0.84) Dose–response association between higher
Q3: OR = 0.58 (CI = 0.43, 0.79) scores on healthy diet scores and lower
Q4: OR = 0.47 (CI = 0.35, 0.64) odds of being depressed (SMFQ score;

e38 | Systematic Review | Peer Reviewed | O’Neil et al.


Q5: OR = 0.55 (CI = 0.40, 0.77) dietary quality)

Jacka et al.10 Age, sex, area level, SES, dieting 75


behaviors, BMI, and physical activity
Healthy diet score (least healthy) C2: b = 0.31* (CI = 0.22, 0.39) Dose–response associations between
C3: b = 0.42* (0.31, 0.53) higher healthy diet quality scores and
higher PedsQL scores (dietary quality)
Unhealthy diet score C2: b = –0.14* (CI = –0.23, –0.06) Dose–response associations between
C3: b = –0.29* (CI = –0.38, –0.20) higher unhealthy diet quality scores and
lower PedsQL (dietary quality)
Robinson et al.24 b = 0.32* (CI = 0.03, 0.60) “Extras” food group associated with higher Sex, sociodemographic factors (family 87.5
SYSTEMATIC REVIEW

internalizing CBCL scores (dietary income, father not at home, and


pattern) maternal employment), family
functioning, physical activity, screen use,
smoking, alcohol, marijuana use, and
early sexual activity
Vegetable food group b = 0.14 (CI = –0.41, 0.69) No association between the vegetable,
Fruit food group b = 0.09 (CI = –0.39, 0.21) fruit, dairy, cereal, or meat and meat
Dairy food group b = 0.11 (CI = –0.36, 0.59) alternatives food groups and internalizing
Cereal food group b = –0.09 (CI = –0.57, 0.39) CBCL scores (dietary pattern)
Meat and meat alternatives food group b = 0.01(CI = –0.62, 0.63)
Weng et al.12 Age, sex, maternal education, paternal 87.5
education, family income, BMI, physical
activity

Continued

American Journal of Public Health | October 2014, Vol 104, No. 10


TABLE 3—Continued

Snack dietary pattern and pure Pure depression: Highest tertile of snack dietary pattern
depression, pure anxiety, and T2: OR = 0.98 (CI = 0.77, 1.25) associated with higher odds of pure
coexisting depression and anxiety T3: OR = 1.64* (CI = 1.30, 2.06) depression, anxiety, and coexisting
Pure anxiety: depression and anxiety (dietary pattern)
T2: OR = 1.38* (CI = 1.08, 1.65)
T3: OR = 1.87* (CI = 1.51, 2.31)
Coexisting depression and anxiety:
T2: OR = 1.27 (CI = 1.00, 2.43)
T3: OR = 1.93* (CI = 1.54, 2.43)
Animal dietary pattern and pure Pure depression: Highest tertile of animal dietary pattern
depression, pure anxiety, and T2: OR = 1.08 (CI = 0.86, 1.37) associated with higher odds of pure
coexisting depression and anxiety T3: OR = 1.21 (CI = 0.95, 1.53) anxiety and coexisting depression and
Pure anxiety: anxiety, but not pure depression (dietary
T2: OR = 1.34* (CI = 1.08, 1.65) pattern)
T3: OR = 1.87* (CI = 1.51, 2.32)

October 2014, Vol 104, No. 10 | American Journal of Public Health


Coexisting depression and anxiety:
T2: OR = 1.10 (CI = 0.88, 1.39)
T3: OR = 1.71* (CI = 1.37, 2.15)
Traditional dietary pattern and pure Pure depression: Highest tertile of traditional dietary pattern
depression, pure anxiety, and T2: OR = 0.61* (CI = 0.49, 0.79) associated with reduced odds of pure
coexisting depression and anxiety T3: OR = 0.38* (CI = 0.30, 0.49) depression and coexisting depression
Pure anxiety: and anxiety, but not pure anxiety (dietary
T2: OR = 0.98 (CI = 0.79, 1.23) pattern)
SYSTEMATIC REVIEW

T3: OR = 0.85 (CI = 0.69, 1.04)


Coexisting depression and anxiety:
T2: OR = 0.74* (CI = 0.60, 0.92)
T3: OR = 1.50* (CI = 0.39, 0.63)
25
Kohlboeck et al. OR = 0.89* (CI = 0.80, 0.98) Higher diet quality score associated with Sex, study center, total energy intake, 87.5
OR = 1.19* (CI = 1.08,1.32) lower emotional SDQ scores (dietary parental background (education, income,
quality) single-parent family), BMI, physical
Increased confectionary food group activity, TV viewing or video game use
associated with higher emotional SDQ
scores (dietary pattern)
No association between dairy, fats and
oils, fruits and vegetables, confectionary,
cereals, bakery wares, meat products, egg
products, or ready-to-eat savories and
emotional SDQ scores (dietary pattern)

Continued

O’Neil et al. | Peer Reviewed | Systematic Review | e39


SYSTEMATIC REVIEW

correlation between healthy and


unhealthy dietary patterns is also

Note. BMI = body mass index; C = category; CBCL = Child Behavior Checklist; CI = confidence interval; IRR = incident rate ratio; OR = odds ratio; PedsQL = Pediatric Quality of Life Inventory; Q = quartile; SES = socioeconomic status;
weak (e.g., Jacka et al.37). How-

87.5
100

75
ever, we acknowledge that these
interpretations remain speculative
in view of the limited evidence for
causality currently available.
Race, grade level; all analyses stratified by

temperament, and weight for length at 1 y


affectivity, completed education in years,
age and sex, parental educational level,

stress, food security, and social support


Household income, family structure, child
Limitations

duration of breastfeeding), child sex,


language spoken at home, financial
The studies acquired through

Maternal characteristics (negative


the systematic search have various
methodological strengths and
weaknesses that might have influ-
enced the outcomes reported. A
strength of this review was the
inclusion of a range of studies
sex

conducted across several coun-


tries and settings. However, the
way in which diet was measured
varied greatly between the studies.
Internalizing scores associated with higher
lower vegetable consumption for females
vegetables or fruits for either sex (dietary

lower fruit consumption for males and


Emotional SDQ scores associated with

Emotional SDQ scores associated with

consumption of sweet foods (dietary


Many used FFQs, a common vali-
No association between depressive
symptoms and daily servings of

but not males (dietary pattern)

dated tool, to assess dietary qual-


ity. The FFQs were completed by
females (dietary pattern)

the child or adolescent in some


cases26 and by the primary care-
giver in others.28 The Youth and
Adolescent Food Frequency
pattern)

pattern)

Questionnaire, a validated mea-


sure, was also used,26 and some
SDQ = Strengths and Difficulties Questionnaire; SMFQ = Short Mood and Feelings Questionnaire; T = tertile.

studies did not use a validated


tool.21 A number of these articles
Males: b = –0.054* (CI = –0.095, –0.012)
Males: 1.8 (low), 1.9, (moderate), and 1.9

Males: 2.3 (low), 2.2 (moderate), and 2.3

were secondary analyses from


Females: 2.0 (low), 1.9 (moderate), and

Females: 2.5 (low), 2.3 (moderate), and

Males: b = –0.014 (CI = –0.054, 0.025)

larger health studies (e.g., Robinson


Females: b = –0.064* (CI = –0.112,

Females: b = –0.050* (CI = –0.088,

et al.24). In such studies, statis-


OR = 1.47* (CI = 1.32, 1.65)

tical techniques were used to score


the available data to create diet-
quality scores. Techniques of this
kind are deemed an appropriate
2.3 (high)

2.3 (high)

–0.015)

–0.012)

method to rank individuals in


(high)

(high)

terms of their diet quality when


available data are limited.10,23
Owing to the length and complex-
ity of FFQs, they are not always
used in health studies.
Aside from the quality of the
tools used to assess diet, other
TABLE 3—Continued

Vegetable consumption

issues surrounding reporting may


Daily vegetable intake

Fruit consumption

have influenced the results


*Significant results.
Daily fruit intake
Fulkerson et al.26

Renzaho et al.27

obtained for habitual food intake


Vollrath et al.28

in the reviewed articles. Studies rely


on accurate reporting; however,
reporting biases can occur for many
reasons, including recall ability and

e40 | Systematic Review | Peer Reviewed | O’Neil et al. American Journal of Public Health | October 2014, Vol 104, No. 10
SYSTEMATIC REVIEW

social desirability biases,38 in which the association between diet qual- Correspondence should be sent to Adrienne received grant and research support from
O’Neil, Innovation in Mental and Physical Eli Lilly, Pfizer, The University of Mel-
respondents are more likely to re- ity and patterns and mental health
Health and Clinical Treatment (IMPACT) bourne, Deakin University, and NHMRC.
port healthier food intake as a result in children and adolescents. Find- Strategic Research Centre, School of
of knowledge about healthy eating ings from the cross-sectional stud- Medicine, Deakin University, Kitchener
House, Ryrie St, Geelong, VIC 3226 Human Participant Protection
guidelines. Reporting biases may ies included in this review high- Institutional review board approval was
Australia (e-mail: a.oneil@deakin.edu.au).
differ between children and adults, light the potential importance of Reprints can be ordered at http://www.ajph. not needed because data were obtained
org by clicking the “Reprints” link. from secondary sources.
making the comparison of results the relationship between dietary
This article was accepted May 31,
additionally challenging because patterns or quality and the mental
2014. References
FFQs were completed by the adults, health of young people; however,
1. Jacka FN, Pasco JA, Mykletun A, et al.
or children, or both. Moreover, dif- we acknowledge that further Association of Western and traditional
Contributors
ferential reporting might also occur studies that contribute to a stron- A. O’Neil and F. N. Jacka conceptualized diets with depression and anxiety in
in those with depressive symptoms ger level of evidence are required. the article. A. O’Neil and S. Housden women. Am J Psychiatry. 2010;167
developed the search strategy and (3):305---311.
in comparison with those without. We recommend that the relation-
conducted the review. S. E. Quirk and 2. Akbaraly TN, Brunner EJ, Ferrie JE,
Moreover, we acknowledge the po- ships between both poor diet and S. Housden performed the quality Marmot MG, Kivimaki M, Singh-Manoux
tential influence of the file-drawer poorer mental health and good assessment and assisted A. O’Neil in the A. Dietary pattern and depressive symp-
development of the 1st draft of the article toms in middle age. Br J Psychiatry.
effect, in which positive studies are diet and better mental health be
and subsequent revisions. S. L. Brennan 2009;195(5):408---413.
overrepresented in the literature examined using longitudinal study oversaw the methodology. F. N. Jacka
3. Sánchez-Villegas A, Delgado-
because studies containing null designs. Moreover, the need for oversaw the writing of the article. L. J.
Rodríguez M, Alonso A, et al. Association
Williams, J. A. Pasco, and M. Berk
findings remain unpublished.39 studies explicating specific biolog- of the Mediterranean dietary pattern with
contributed to drafts of the article.
Finally, it is important to note ical mechanisms of action, as well the incidence of depression: The Segui-
miento Universidad de Navarra/Univer-
that although many studies have as data from well-designed ran-
Acknowledgments sity of Navarra follow-up (sun) cohort.
taken into account many of the domized controlled trials to sup- A. O’Neil, S. L. Brennan, and F. N. Jacka Arch Gen Psychiatry. 2009;66(10):
common confounding variables, port or refute direct biologically are recipients of National Health and 1090---1098.
Medical Research Council (NHMRC)
other studies have not. For exam- causal relationships, is clear.42 4. Lai JS, Hiles S, Bisquera A, Hure AJ,
Fellowships (1052865, 1012472, and
McEvoy M, Attia J. A systematic review
ple, socioeconomic status is con- 628912). L. J. Williams is the recipient of
and meta-analysis of dietary patterns and
sidered an important potential Conclusions an NHMRC Career Development Award
depression in community-dwelling adults.
and was supported by an Alfred Deakin
confounder that must be consid- Despite a paucity of data, our Am J Clin Nutr. 2014; 99(1):181---197.
Postdoctoral Fellowship while undertak-
ered because it is associated with findings highlight the potential im- ing this work. M. Berk is supported by 5. Psaltopoulou T, Sergentanis TN,
both diet40 and mental health.41 portance of the relationship be- an NHMRC Senior Principal Research Panagiotakos DB, Sergentanis IN, Kosti R,
Fellowship (1059660). Scarmeas N. Mediterranean diet, stroke,
Several studies in this review did tween dietary patterns or quality cognitive impairment, and depression:
A. O’Neil, M. Berk, and F. N. Jacka have
not take socioeconomic status into and mental health early in the life received project funding from Meat and a meta-analysis. Ann Neurol. 2013;74
account; results from such studies span. Prospective and intervention Livestock Australia. M. Berk has received (4):580---591.
grant and research support from the 6. Sánchez-Villegas A, Verberne L, De
should be viewed with caution studies are now required to im-
National Institutes of Health, Cooperative Irala J, et al. Dietary fat intake and the risk
because the relationships ob- prove the level of evidence. Given Research Centre, Simons Autism Foun- of depression: the SUN Project. PLoS One.
served may possibly be explained that the average age of onset for dation, Cancer Council of Victoria, Stan- 2011;6(1):e16268.
ley Medical Research Foundation,
by socioeconomic status or other anxiety and mood disorders is 6 7. Sánchez-Villegas A, Toledo E, De
Medical Benefits Fund, NHMRC, Beyond
related variables rather than by years and 13 years, respectively,43 Blue, Rotary Health, Geelong Medical Irala J, et al. Fast-food and commercial
baked goods consumption and the risk of
diet or mental health alone. Simi- the potential for early intervention Research Foundation, Bristol Myers
Squibb, Eli Lilly, Glaxo SmithKline, depression. Public Health Nutr. 2012;15
larly, several studies did not ac- using strategies targeted at improv- (3):424---432.
Organon, Novartis, Mayne Pharma, and
count for physical activity. Un- ing dietary intake at a population Servier; has been a speaker for Astra 8. Le Port A, Gueguen A, Kesse-Guyot
avoidable residual confounding level may be of substantial public Zeneca, Bristol Myers Squibb, Eli Lilly, E, et al. Association between dietary
may also explain the associations health benefit. However, this would Glaxo SmithKline, Janssen Cilag, Lund- patterns and depressive symptoms over
beck, Merck, Pfizer, Sanofi Synthelabo, time: a 10-year follow-up study of the
observed. For example, certain require policy action to improve the Servier, Solvay, and Wyeth; and served as GAZEL cohort. PLoS One. 2012;7(12):
complex variables, such as socio- global food environment.44 j a consultant to Astra Zeneca, Bristol e51593.
economic status, are difficult to Myers Squibb, Eli Lilly, Glaxo SmithKline,
9. van Egmond-Fröhlich AW,
Janssen Cilag, Lundbeck Merck, and
measure and are therefore likely Weghuber D, de Zwaan M. Association of
Servier. F. N. Jacka has received grant and
symptoms of attention-deficit/hyperactiv-
to result in residual confounding. About the Authors research support from the Brain and
ity disorder with physical activity, media
This type of confounding cannot Adrienne O’Neil, Shae E. Quirk, Siobhan Behaviour Research Institute, NHMRC,
time, and food intake in children and
Housden, Sharon L. Brennan, Lana J. Australian Rotary Health, Geelong Medi-
be ruled out as an explanatory Williams, Julie A. Pasco, Michael Berk, and cal Research Foundation, Ian Potter
adolescents. PLoS ONE. 2012;7(11):
e49781.
factor in the associations observed Felice N. Jacka are with the Innovation in Foundation, Eli Lilly, and The University
Mental and Physical Health and Clinical of Melbourne and has been a paid 10. Jacka FN, Kremer PJ, Berk M, et al. A
in the studies reviewed.
Treatment (IMPACT) Strategic Research speaker for Sanofi-Synthelabo, Janssen prospective study of diet quality and
To our knowledge, this is the Centre, School of Medicine, Deakin Cilag, Servier, Pfizer, Health Ed, Network mental health in adolescents. PLoS One.
first review of its kind to explore University, Geelong, Victoria, Australia. Nutrition, and Eli Lilly. L. J. Williams has 2011;6(9):e24805.

October 2014, Vol 104, No. 10 | American Journal of Public Health O’Neil et al. | Peer Reviewed | Systematic Review | e41
SYSTEMATIC REVIEW

11. McMartin S, Kuhle S, Kirk S, of poor mental health in early adoles- 37. Jacka FN, Cherbuin J, Anstey K,
Veugelers P, Colman I. The association cence. J Paediatr Child Health. 2011;47(1- Butterworth P. Dietary patterns and de-
between diet quality and internalizing 2):54---61. pressive symptoms over time: examining
disorders in children. Am J Epidemiol. 25. Kohlboeck G, Sausenthaler S, Standl the relationships with socioeconomic po-
2011;173:S289---S289. M, et al. Food intake, diet quality and sition, health behaviours and cardiovas-
behavioral problems in children: results cular risk. PLoS One. 2014;9(1):e87657.
12. Weng TT, Hao JH, Qian QW, et al. Is
there any relationship between dietary from the GINI-plus/LISA-plus studies. 38. Hebert JR, Clemow L, Pbert L,
patterns and depression and anxiety in Ann Nutr Metab. 2012;60(4):247---256. Ockene IS, Ockene JK. Social desirability
Chinese adolescents? Public Health Nutr. 26. Fulkerson JA, Sherwood NE, Perry bias in dietary self-report may compro-
2012;15(4):673---682. CL, Neumark-Sztainer D, Story M. De- mise the validity of dietary intake mea-
sures. Int J Epidemiol. 1995;24(2):389---
13. UNICEF. The State of the World’s pressive symptoms and adolescent eating
398.
Children 2011. New York, NY: UNICEF and health behaviors: a multifaceted view
USA; 2011. in a population-based sample. Prev Med. 39. Song F, Parekh S, Hooper L, et al.
2004;38(6):865---875. Dissemination and publication of re-
14. Chatzi L, Melaki V, Sarri K, et al.
search findings: an updated review of
Dietary patterns during pregnancy and the 27. Renzaho AMN, Kumanyika S,
related biases. Health Technol Assess.
risk of postpartum depression: the mother- Tucker KL. Family functioning, parental
2010;14(8):iii, ix---xi, 1---193.
child “Rhea” cohort in Crete, Greece. Public psychological distress, child behavioural
Health Nutr. 2011;14(9):1663---1670. problems, socio-economic disadvantage 40. Brennan SL, Henry MJ, Nicholson
and fruit and vegetable consumption GC, Kotowicz MA, Pasco JA. Socioeco-
15. Quirk SE, Williams LJ, O’Neil A, et al.
among 4-12 year-old Victorians, Australia. nomic status, obesity and lifestyle in men:
The association between diet quality, di- the Geelong Osteoporosis Study. J Mens
Health Promot Int. 2011;26(3):263---275.
etary patterns and depression in adults. Health. 2010;7(1):31---41.
BMC Psychiatry. 2013;13:175. doi: . 28. Vollrath ME, Tonstad S, Rothbart
MK, Hampson SE. Infant temperament is 41. Williams LJ, Brennan SL, Henry MJ,
16. Lievense A, Bierma-Zeinstra S, et al. Area-based socioeconomic status
associated with potentially obesogenic
Verhagen A, Verhaar J, Koes B. Influence and mood disorders: cross-sectional evi-
diet at 18 months. Int J Pediatr Obes. 2011;
of work on the development of osteoar- dence from a randomly selected cohort
6(2-2):e408---e414.
thritis of the hip: a systematic review. of adult women. Maturitas. 2011;69(2):
J Rheumatol. 2001;28(11):2520---2528. 29. Sanhueza C, Ryan L, Foxcroft DR.
173---178.
Diet and the risk of unipolar depression in
17. Lievense AM, Bierma-Zeinstra SM, 42. Jacka FN, Mykletun A, Berk M.
adults: systematic review of cohort stud-
Vergahen AP, van Baar ME, Verhaar JAN, ies. J Hum Nutr Diet. 2013;26(1):56---70. Moving towards a population health ap-
Koes BW. Influence of obesity on the proach to the primary prevention of
development of osteoarthritis of the hip: 30. Jacka FN, Maes M, Pasco JA,
common mental disorders. BMC Med.
a systematic review. Rheumatology (Oxford). Williams LJ, Berk M. Nutrient intakes and
2012;10(1):149.
2002;41(10):1155---1162. the common mental disorders in women.
J Affect Disord. 2012;141(1):79---85. 43. Merikangas KR, He JP, Burstein M,
18. Brennan SL, Pasco JA, Urquhart DM, et al. Lifetime prevalence of mental
Oldenburg B, Wang Y, Wluka AE. Associa- 31. Jacka FN, Pasco JA, Williams LJ,
disorders in US adolescents: results
tion between socioeconomic status and bone Meyer BJ, Digger R, Berk M. Dietary
from the National Comorbidity Survey
mineral density in adults: a systematic re- intake of fish and PUFA, and clinical
Replication---Adolescent Supplement
depressive and anxiety disorders in women.
view. Osteoporos Int. 2011;22(2):517-- 527. (NCS-A). J Am Acad Child Adolesc Psychi-
Br J Nutr. 2013;109(11):2059---2066.
19. Moher D, Liberati A, Tetzlaff J, atry. 2010;49(10):980---989.
32. Irwin MR, Miller AH. Depressive
Altman DG. Preferred reporting items 44. Jacka FN, Sacks G, Berk M, Allender
disorders and immunity: 20 years of
for systematic reviews and meta-analyses: S. Food policies for physical and mental
progress and discovery. Brain Behav
the PRISMA statement. PLoS Med. 2009; health. BMC Psychiatry. 2014;14:132.
Immun. 2007;21(4):374---383.
6(7):e1000097.
33. Esposito K, Marfella R, Ciotola M,
20. Wiles NJ, Northstone K, Emmett P,
et al. Effect of a Mediterranean-style diet
Lewis G. Junk food diet and childhood
on endothelial dysfunction and markers
behavioural problems: results from the
of vascular inflammation in the metabolic
ALSPAC cohort. Eur J Clin Nutr. 2009;
syndrome: a randomized trial. JAMA.
63(4):491---498.
2004;292(12):1440---1446.
21. Brooks TL, Harris SK, Thrall JS,
34. Molteni R, Barnard R, Ying Z,
Woods ER. Association of adolescent risk
Roberts C, Gomez-Pinilla F. A high-fat,
behaviors with mental health symptoms
refined sugar diet reduces hippocampal
in high school students. J Adolesc Health.
brain-derived neurotrophic factor, neu-
2002;31(3):240---246.
ronal plasticity, and learning. Neurosci-
22. Oddy WH, Robinson M, Ambrosini ence. 2002;112(4):803---814.
GL, et al. The association between dietary
35. Bocchio-Chiavetto L, Bagnardi V,
patterns and mental health in early ado-
Zanardini R, et al. Serum and plasma
lescence. Prev Med. 2009;49(1):39---44.
BDNF levels in major depression: a repli-
23. Jacka FN, Kremer PJ, Leslie ER, et al. cation study and meta-analyses. World J
Associations between diet quality and Biol Psychiatry. 2010;11(6):763---773.
depressed mood in adolescents: results
36. Brunoni AR, Lopes M, Fregni F. A
from the Australian Healthy Neighbour- systematic review and meta-analysis of
hoods Study. Aust N Z J Psychiatry. clinical studies on major depression and
2010;44(5):435---442. BDNF levels: implications for the role of
24. Robinson M, Kendall GE, Jacoby P, neuroplasticity in depression. Int J Neuro-
et al. Lifestyle and demographic correlates psychopharmacol. 2008;11(8):1169---1180.

e42 | Systematic Review | Peer Reviewed | O’Neil et al. American Journal of Public Health | October 2014, Vol 104, No. 10
Copyright of American Journal of Public Health is the property of American Public Health
Association and its content may not be copied or emailed to multiple sites or posted to a
listserv without the copyright holder's express written permission. However, users may print,
download, or email articles for individual use.