Hostname: page-component-848d4c4894-2pzkn Total loading time: 0 Render date: 2024-05-24T07:38:03.474Z Has data issue: false hasContentIssue false

Does an inflammatory diet affect mental well-being in late childhood and mid-life? A cross-sectional study

Published online by Cambridge University Press:  17 May 2021

Kate M. Lycett*
Centre for Social & Early Emotional Development, School of Psychology, Deakin University, Burwood, Melbourne, VIC, Australia Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, The University of Melbourne, Parkville, Melbourne, VIC, Australia
Disna J. Wijayawickrama
Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia
Mengjiao Liu
Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, The University of Melbourne, Parkville, Melbourne, VIC, Australia
Anneke Grobler
Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, The University of Melbourne, Parkville, Melbourne, VIC, Australia
David P. Burgner
Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, The University of Melbourne, Parkville, Melbourne, VIC, Australia Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, Monash University, Clayton, Melbourne, VIC, Australia
Louise A. Baur
Discipline of Child and Adolescent Health, The University of Sydney, Sydney, NSW, Australia
Richard Liu
Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, The University of Melbourne, Parkville, Melbourne, VIC, Australia
Katherine Lange
Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, The University of Melbourne, Parkville, Melbourne, VIC, Australia
Melissa Wake
Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, The University of Melbourne, Parkville, Melbourne, VIC, Australia Liggins Institute, The University of Auckland, Grafton, Auckland, New Zealand
Jessica A. Kerr
Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Melbourne, VIC, Australia Department of Paediatrics, The University of Melbourne, Parkville, Melbourne, VIC, Australia
*Corresponding author: Kate Lycett, fax +61 3 9345 5900, email
Rights & Permissions [Opens in a new window]


Inflammatory diets are increasingly recognised as a modifiable determinant of mental illness. However, there is a dearth of studies in early life and across the full mental well-being spectrum (mental illness to positive well-being) at the population level. This is a critical gap given that inflammatory diet patterns and mental well-being trajectories typically establish by adolescence. We examined the associations of inflammatory diet scores with mental well-being in 11–12-year-olds and mid-life adults. Throughout Australia, 1759 11–12-year-olds (49 % girls) and 1812 parents (88 % mothers) contributed cross-sectional population-based data. Alternate inflammatory diet scores were calculated from a twenty-six-item FFQ, based on the prior literature and prediction of inflammatory markers. Participants reported negatively and positively framed mental well-being via psychosocial health, quality of life and life satisfaction surveys. We used causal inference modelling techniques via generalised linear regression models (mean differences and risk ratios (RR)) to examine how inflammatory diets might influence mental well-being. In children and adults, respectively, a 1 sd higher literature-derived inflammatory diet score conferred between a 44 % (RR 95 % CI 1·2, 1·8) to 57 % (RR 95 % CI 1·3, 2·0) and 54 % (95 % CI 1·2, 2·0) to 86 % (RR 95 % CI 1·4, 2·4) higher risk of being in the worst mental well-being category (i.e. <16th percentile) across outcome measures. Results for inflammation-derived scores were similar. BMI mediated effects (21–39 %) in adults. Inflammatory diet patterns were cross-sectionally associated with mental well-being at age 11–12 years, with similar effects observed in mid-adulthood. Reducing inflammatory dietary components in childhood could improve population-level mental well-being across the life course.

Full Papers
© The Author(s), 2021. Published by Cambridge University Press on behalf of The Nutrition Society

Diet is increasingly recognised as a modifiable factor for mental health, with inflammatory pathways a key suggested underlying mechanism(Reference Phillips, Chen and Heude1). Randomised controlled trials in adults support this hypothesis, with anti-inflammatory dietary intervention (i.e. high in fruit, vegetables, low-fat dairy products, fish and wholegrains, and moderate in unsaturated fat) shown to decrease symptoms of depression(Reference Firth, Marx and Dash2). Similarly, longitudinal studies in adults and adolescents suggest that pro-inflammatory diets (i.e. high in sugar, saturated fats, refined carbohydrates and red meat) increase the risk of mental illness(Reference Phillips, Chen and Heude1). In reality, an individual’s habitual diet is rarely anti- or pro-inflammatory but is a composite of both. This overall inflammatory potential of diet has also been shown to affect mental illnesses, such as major depression and schizophrenia(Reference Phillips, Chen and Heude1,Reference Kheirouri and Alizadeh3) . However, it is unclear whether an individual’s overall dietary inflammatory potential affects the full mental well-being spectrum (i.e. from diagnosed mental illness to general well-being/happiness) at the population level and, if so, when in the life course effects begin to emerge.

The Dietary Inflammatory Index (DII) is the only standardised dietary score of overall inflammatory potential(Reference Shivappa, Steck and Hurley4). It assigns specific values to both pro- and anti-inflammatory foods using FFQ with a minimum of forty-five items, derived from the literature and validated against inflammatory biomarkers (e.g. C-reactive protein)(Reference Shivappa, Steck and Hurley4). In a recent systematic review, eleven out of twelve studies showed that a higher DII (i.e. more inflammatory potential) was associated with a higher risk of depression in adults(Reference Kheirouri and Alizadeh3). These findings were recently extended beyond depression in a large sample of mid-life Irish adults (51 % female; 50–69 years) to anxiety symptoms and positive well-being (i.e. validated, survey-assessed, positively framed items such as ‘I have felt cheerful and in good spirits’)(Reference Phillips, Shivappa and Hebert5). Women with a higher DII score (tertile 1 v. 3) had at least doubled odds of experiencing elevated depression and anxiety symptoms, and a lower likelihood of reporting positive well-being; however, there was little evidence of associations in men.

Fewer studies looking at inflammatory diet and mental health have been conducted in children and adolescents. In a study of adolescent Iranian girls, those with high DII scores (tertile 1 v. 3) were at least three times more likely to have moderate stress scores(Reference Shivappa, Hebert and Rashidkhani6). Similarly, in an Australian cohort, a pro-inflammatory ‘Western’ dietary pattern at 14 years of age was indirectly associated with mental health problems 3 years later, with effects mediated via adiposity and inflammatory pathways(Reference Oddy, Allen and Trapp7). In addition, in Spanish children and adolescents, adhering to an anti-inflammatory Mediterranean diet was cross-sectionally associated with higher levels of well-being (health-related quality of life and positive and negative effects), but not associated with well-being 2 years later(Reference Esteban-Gonzalo, Turner and Torres8).

These recent studies suggest that the inflammatory potential of diet may affect mental well-being from childhood onwards, but the scarcity of studies in younger children and the lack of positive well-being measures warrant further investigation. Such knowledge would inform public health campaigns to target the most appropriate age group/s and also elucidate the aetiology of mental well-being. This is particularly important in childhood when dietary patterns become established(Reference Mikkila, Rasanen and Raitakari9) and given that over half of lifetime mental health disorders develop by early adolescence(Reference Kessler, Berglund and Demler10). In addition, more recently described inflammatory markers (e.g. glycoprotein acetyls (GlycA)) may better reflect cumulative inflammation and more diverse inflammatory pathways than acute phase reactants such as C-reactive protein(Reference Ballout and Remaley11) and therefore may be more informative regarding diet-related inflammation.

Thus, we used causal modelling techniques(Reference Krieger and Davey Smith12) to account for underlying confounding structures to best examine potential pathways between inflammatory diet scores and mental well-being in two generations (11–12-year-olds and mid-life adults) in a population-based study. We used both a literature- and GlycA-derived inflammatory diet score.


Study design

The Child Health CheckPoint (CheckPoint) was a physical/biomarkers module nested within the population-based Longitudinal Study of Australian Children (LSAC)(Reference Clifford, Davies and Wake13). The cross-sectional CheckPoint data were collected across Australia in 2015–2016 between LSAC’s 6th and 7th waves, when children were aged 11–12 years. The Royal Children’s Hospital (HREC33225) and The Australian Institute of Family Studies (AIFS4-05) Ethics Committees approved the project. Written informed consent from a parent/guardian was provided for their child, as well as their own participation.


Details of the CheckPoint methods are described elsewhere(Reference Clifford, Davies and Wake13) and are summarised below. In 2004, LSAC randomly recruited a nationally representative sample of 5107 infants (age 0–1 years) into its Birth (B) cohort. LSAC has since followed these families biennially, with six waves of data collection complete in 2014 (retention rate 75 %). The CheckPoint was offered to all B-cohort families that took part in wave 6 (Appendix Figure 1).


Most participants attended a 3·5-h main or 2·5-h Mini Assessment Center in one of Australia’s capital cities or large regional towns. Participants rotated through a series of 15-min physical assessment and biospecimen collection stations at which semi-fasting venous blood samples were collected, as detailed elsewhere(Reference Calder, Ahluwalia and Brouns14). Serum aliquots were frozen at −80˚C and then shipped to Finland for metabolomics analysis via a high-throughput proton NMR metabolomics platform (Nightingale Health), generating the inflammatory biomarker GlycA. Participants unable to attend the Center were offered a 90-min home visit. On an iPad, parents and children separately self-reported standardised survey measures regarding their diet and mental well-being at all three types of assessments, but those having home visits did not contribute venous blood.


Despite the cross-sectional nature of the study, we considered the exposure variables to be the inflammatory diet scores and the outcome variables to be the mental well-being measures.

Inflammatory diet scores

On an iPad, children and adults separately self-reported their usual intake of various foods and drinks via the National Secondary Students’ Diet and Activity (NaSSDA) survey(Reference Rutishauser, Webb and Abraham15). The twenty-six-item survey fell short of the minimum forty-five items needed to calculate the DII(Reference Shivappa, Steck and Hurley4). Therefore, we used each relevant NaSSDA item (twenty-three of the twenty-six items) to derive two inflammatory diet scores based on: (1) published literature (‘literature-derived’) and (2) the statistical correlations with levels of GlycA (‘GlycA-derived’). Procedures used to calculate inflammatory diet scores can be found elsewhere(Reference Davis, Liu and Kerr16) and are summarised below. Higher scores indicate a more pro-inflammatory diet. NaSSDA items allow direct comparison with the Australian Dietary Guidelines and demonstrate good validity(Reference Rutishauser, Webb and Abraham15) and expected gradients with socio-economic position(Reference Morley, Scully and Niven17). Appendix Table 1 details the scoring for each NaSSDA item for both inflammatory diet scores.

The literature-derived inflammatory diet score was generated using two highly cited reviews(Reference Calder, Ahluwalia and Brouns14,Reference Barbaresko, Koch and Schulze18) that used C-reactive protein to establish the ‘inflammatory potential’ of commonly consumed different food and beverage components. Based on this information, we classified each NaSSDA item as either anti-inflammatory (e.g. fish consumption) or pro-inflammatory (e.g. red meat consumption). We then assigned each item’s response options a value from −2 (anti-inflammatory) to +2 (pro-inflammatory) and summed all items to calculate an overall literature-derived inflammatory diet score for each participant.

The GlycA-derived inflammatory diet score was based on parents’ measured inflammatory GlycA, because GlycA as a marker of chronic inflammation at the time the score was derived was based on adult populations(Reference Ballout and Remaley11); we then applied this score to the child NaSSDA data. GlycA values were highly positively skewed; therefore, we naturally log-transformed the values. Following this, adult NaSSDA items were individually regressed against log-transformed GlycA values (univariable models) and items that reached a statistical significance level of P < 0·20 and were entered into Multivariable Model 1 to ascertain their combined association with GlycA. Items that remained associated in Model 1 were then entered into a final Multivariable Model 2 (Appendix Table 1). The coefficients from this final model were then used to generate an inflammatory diet score for each adult and child with the following formula:

sum(model regression coefficient for item multiplied by participants’ NaSSDA item response value) + ‘model constant’.

Mental well-being measures

Table 1 details each mental well-being measure. All cumulative scores are on a 0–1 or 0–100 scale; higher scores indicate better mental well-being.

Table 1. Mental well-being measures

For children, measures tapping into mental well-being included two negatively framed measures: overall health-related quality of life (QoL) assessed via the child version of the Child Health Utility-9D(Reference Ratcliffe, Chen and Stevens19); and Psychosocial QoL assessed via the psychosocial summary score of the Pediatric QoL Inventory(Reference Varni, Burwinkle and Seid20). We also included two positively framed measures: General well-being, assessed via General Wellbeing Scale(Reference Varni, Burwinkle and Seid20); and Life satisfaction, assessed via the International Scale of Child Wellbeing’s Brief Multidimensional Students’ Life Satisfaction sub-scale(Reference Seligson, Huebner and Valois21).

For adults, the mental well-being measures included Psychosocial QoL assessed via the psychosocial health domain of the Assessment of Quality of Life 8D (AQoL-8D-PS)(Reference Maxwell, Ozmen and Iezzi22); and health-related QoL assessed via the adult version of the Child Health Utility-9D(Reference Ratcliffe, Chen and Stevens19).

Potential confounders

Potential confounders included age, sex and a range of measures to tap into the socio-economic background (socio-economic position (SEP), neighbourhood disadvantage and education level), given that lower socio-economic background is associated with unhealthier diets(Reference Gasser, Mensah and Kerr23) and worse mental health(Reference Allen, Balfour and Bell24). SEP was calculated from the most recently available parent-reported education, income and occupation data at LSAC’s wave 6. Scores were internally standardised (mean: 0; standard deviation (sd): 1), where higher scores represent higher SEP(Reference Blakemore and Gibbings25). Neighbourhood disadvantage was calculated based on family postcode of residence at the CheckPoint using the census-derived Index of Relative Socioeconomic Disadvantage (SEIFA; national mean: 1000; sd: 100, where higher values = less disadvantage)(26). Using all three socio-economic variables was deemed appropriate as they were not highly correlated (<0·30).

Potential mediator

Given that an inflammatory diet is directly implicated in higher BMI, and that BMI is known to influence mental well-being(Reference Phillips, Chen and Heude1) particularly in adults, we considered whether BMI was a mediator of the effect of a pro-inflammatory diet on mental well-being. Child and adult BMI (kg/m2) was calculated from researcher-measured height and weight. For children, BMI was converted to age- and sex-specific BMI z-scores (CDC reference values)(Reference Wake, Lycett and Clifford27).

Statistical analyses

Each of the inflammatory diet scores was dichotomised as above the 75th percentile compared with the rest. We dichotomised the dietary exposure given that it was unlikely that the continuous version would meet the assumption of linearity (i.e. that a one-unit change is the same at all levels of this exposure). Each mental well-being score was considered continuously and dichotomously (<1 sd below the mean (i.e. <16th percentile)) to identify those at the highest risk of poor mental well-being. Child and adult analyses were considered separately.

To examine whether a diet high in pro-inflammatory potential affects mental well-being, we used two causal modelling approaches to account for underlying confounding structures in order to estimate the same causal effects: First, a classical regression approach which makes an assumption about constant effects within confounder strata and second, a more flexible approach that does not require those assumptions, which is implemented by extending those regressions to include interactions between the exposure and confounders (SEP and SEIFA in this instance), and then averaging the causal effects within strata using the margins command.

We used ordinary linear regression to estimate mean difference and log-binomial regression to estimate risk ratios in models adjusted for age, sex, SEP, SEIFA and education (parent education for children). A mediation analysis was also used to examine BMI as a mediator of the association between a pro-inflammatory diet and mental well-being. The mediation analysis estimated the total causal effect of inflammatory diet (‘exposure’) on mental well-being (‘outcome’) occurring via an intermediate variable (‘mediator’; BMI in this case) using the ‘paramed command’. All analyses were conducted in 2020 using Stata/IC(15.1).


Sample characteristics

Of the 3513 families retained at LSAC’s wave 6, 1874 (50 %) parent–child dyads took part in the CheckPoint (see Appendix Figure 1). The analytic sample included 1812 adults (mean age 43·7 years (sd 5·2)) and 1759 11–12-year-olds who had at least one inflammatory diet score and one mental well-being measure.

Sex was evenly distributed in children (49 % girls), but adults were mostly mothers (88 %; see Table 2). Similar to the Australian population(28), our sample comprised 24 % of children and 57 % of adults with overweight/obesity. The average family SEP (0·18) suggested a slightly more advantaged sample than the wave 6 LSAC cohort (SEP mean 0 (sd 1))(Reference Blakemore and Gibbings25). Children’s and adults’ average literature-derived diet scores were 2·51 (sd: 3·05; range: –5 to 14) and 0·76 (sd: 2·46; range: –5 to 13), and their GlycA-derived scores were 0·06 (sd: 0·06; range: –0·14 to 0·42) and 0·03 (sd: 0·06; range –0·15 to 0·35), respectively. Mental well-being measures were in line with population norms for adults and children of this age.

Table 2. Characteristics of analytic sample

(Mean values and standard deviations)

HRQL, Health-related Quality of Life; QoL, Quality of Life.

* Unless otherwise specified.

SEP was drawn from LSAC Wave 6 assessments, conducted approximately 1 year prior to CheckPoint Neighborhood disadvantage: national mean 1000 (sd 100).

Inflammatory diet and mental well-being

Estimated causal effects calculated using the flexible approach were almost identical to the classical regression results. Thus, we not only focus on reporting classical regression models below but also report the more flexible effects in Table 3.

Table 3. Causal linear regression analyses comparing diet score above 75th percentile with others

HRQL, Health-related Quality of Life; QoL, Quality of Life; SEIFA, neighbourhood disadvantage; SEP, socio-economic position.

All models adjusted for age, sex, SEP, SEIFA and parental education.

Higher inflammatory diet scores were associated with worse mental well-being in both age groups, with the literature-derived and GlycA-derived diet scores showing similar associations. For example, for negatively framed mental well-being in children, each sd increment in the literature-derived inflammatory diet score was associated with −0·19 (95 % CI −0·30, −0·08) to −0·27 (95 % CI −0·37, −0·18) lower mental well-being scores, with similar results for positively framed mental well-being (−0·15 (95 % CI −0·25, −0·05) to −0·24 (95 % CI −0·34, −0·13)). In adults, estimated effects for negatively framed mental well-being were similar to children, although associations were slightly stronger for the literature-derived inflammatory diet score compared with the GlycA-derived score.

Binary outcomes showed similar results (Table 3), with inflammatory diet scores associated with a higher relative risk of poor mental well-being (i.e. lowest 16th percentile) across all measures in both age groups. Associations were larger in adults, compared with children, and in adults, they were also slightly stronger for the literature-derived, compared with the GlycA-derived inflammatory diet score. For example, in adults, each sd increment in the literature-derived inflammatory diet score was associated with a 1·48 (95 % CI 1·10, 1·98) to 1·86 (95 % CI 1·42, 2·43) higher risk of poor mental well-being.

Among children, there was little evidence to suggest that BMI mediated effects of a pro-inflammatory diet on mental well-being (Table 4). However, among adults, the percentage mediation by BMI ranged from 19 % to 21 % for the literature-derived score, and 39–40 % for the GlycA-derived score.

Table 4. Direct/indirect effects of pro-inflammatory diet (>75th percentile v. Others) on mental well-being through BMI

(Risk ratios (RR) and 95 % confidence intervals)

HRQL, Health-related Quality of Life; QoL, Quality of Life; SEP, socio-economic position.

Estimates adjusted for age, sex, SEP, neighbourhood disadvantage and parental education.

*RR represents effect estimates for direct, indirect, and total effect derived from causal mediation analysis, compared with the reference group. The proportion of the total effect mediated by BMI was calculated as RRDE(RRIE–1)/(RRDERRIE–1); if direct (DE) and indirect effect (IE) have opposite direction of association, the proportion will not be calculated and marked as NA.

Boldface indicates statistical significance (P < 0·01, P < 0·001).


Principal findings

In both children and mid-life adults, a diet with greater overall inflammatory potential was associated with worse mental well-being across the spectrum. The consistency of these associations across all mental well-being measures, with both the inflammatory diet scores, and within both age groups, provides confidence in the results that an inflammatory diet worsens mental well-being.

Estimated effects were largest for the psychosocial QoL measures in both children and adults. This may reflect that these measures have a more specific focus on mental health constructs, rather than on general well-being. It may also be a reflection of greater detail gained from more items (i.e. twenty-five items included for adults and fifteen items for children) compared with the other briefer measures (i.e. 5–9 items).

BMI did not mediate effects between an inflammatory diet and mental well-being in children and accounted for between about 20 % (literature-derived) and about 39 % (GlycA-derived) of the direct effects for adults. Therefore, regardless of the pathway between diet, BMI and mental health, these mediation results suggest that diet may more directly affect mental health. That is, the associations between diet and mental health might be the direct result of inflammation or additional factors may lie on the pathway linking inflammatory diet and poor mental well-being. For example, there are various molecular mechanisms that link dietary factors with brain health(Reference Godos, Currenti and Angelino29), such as changes in circadian rhythm, hormonal homoeostasis and neuronal plasticity. In addition, high intake of inflammatory foods may affect the brain by negatively affecting the gut microbiota composition(Reference Berding, Vlckova and Marx30), including loss of microbial diversity and function that can impact physical and mental health.

Interpretation in light of previous research

Our results align with the majority of previous research conducted with the DII, showing that a diet with greater inflammatory potential is associated with worse mental health among adults(Reference Phillips, Chen and Heude1,Reference Kheirouri and Alizadeh3,Reference Wang, Zhou and Chen31) and with stress in adolescents(Reference Shivappa, Hebert and Rashidkhani6). We extend these findings to a cross-sectional study of a population-derived cohort of both children and adults, using two somewhat cruder measures of an inflammatory diet, but broader measures of mental well-being, which included positively framed measures in children. Our results in children are congruent with Phillips and colleagues' findings in adults(Reference Phillips, Shivappa and Hebert5), in that we showed a diet with higher inflammatory potential is associated with lower positively framed well-being in children. Because our study was cross-sectional, we cannot directly interpret the results alongside previous longitudinal studies(Reference Kheirouri and Alizadeh3) or randomised controlled trials in adults(Reference Firth, Marx and Dash2), nor can we conduct comparable mediation analysis(Reference Oddy, Allen and Trapp7). However, we note that longitudinal mediation analysis will be possible as the LSAC waves progress in future years and continue to assess adolescent and parental BMI and mental well-being.

Strengths and limitations

Strengths of our study include its large, population-based sample of both children and mid-life adults, enhancing generalisability. Combining both pro- and anti-inflammatory dietary items into one inflammatory score gives a more accurate representation of dietary habits by better capturing overall inflammatory potential than looking at anti-inflammatory or pro-inflammatory diet alone. Widely used and validated measures of mental well-being also provide confidence that the measures are indicative of the designed concepts and provide scope for comparison with future studies. Another strength included the incorporation of measures that reflect the positive spectrum of mental well-being by assessing positively framed questions in children – unfortunately, we were unable to look at this in adults.

Our results in adults may not generalise to men and to mid-life adults without children, given that most were parents who attended the sessions with their child and 88 % were women. Similarly, our findings may not generalise to those from highly disadvantaged backgrounds whose diets and mental health may be different to the CheckPoint cohort. Habitual diet is inherently difficult to measure(Reference Ioannidis32) and like the majority of self-report diet assessments, the NaSSDA does not capture portion sizes or cooking methods and is subject to social desirability and recall biases(Reference Ioannidis32). Furthermore, because we only measured the intake of twenty-three dietary items, we were unable to generate the well-used DII(Reference Shivappa, Steck and Hurley4). However, our approach illustrates a low-burden approach (using brief FFQ) to potentially investigate similar questions in large cohorts. Another limitation is that children’s data-derived inflammatory diet algorithm relied on adult data, which assumes foods and drinks have the same inflammatory potential for both age groups. In addition, given that we examined inflammatory diet scores dichotomously, we cannot rule out misclassification of the exposure. However, the use of a binary exposure simplifies the translation of our findings to public health and policy statements.

Furthermore, the study’s cross-sectional nature limits identifying temporal relationships between inflammatory diet patterns and mental well-being. However, given that randomised controlled trials show that reducing inflammatory diets can reduce mental illness, such as depression(Reference Firth, Marx and Dash2), we have assumed directionality from diet to mental well-being and used the best possible causal modelling approaches to assess this. Even though we wanted to examine the causal relationship between an inflammatory diet and mental well-being and used causal modelling approaches, one can never be sure that the associations found are indeed causal and all interpretations are open to alternative reasons for these associations.


Findings have important implications for understanding the origins of mental well-being and for policymakers designing strategies to tackle poor diet and mental well-being. However, shifting overconsumption of inflammatory diets is no easy task. Individual-level interventions to optimise diet are rarely sustained, and evidence-based societal level policies (e.g. sugar taxes)(Reference Teng, Jones and Mizdrak33) fail to be implemented(Reference Swinburn, Kraak and Allender34).

The overwhelming lack of success translating dietary intervention to policy has led to novel ideas, such as tackling diets high in inflammatory foods (e.g. red meat) from the perspective of climate change/sustainability. To improve dietary choices and their adverse impacts on mental well-being and physical health clearly requires political will(Reference Swinburn, Kraak and Allender34), and far greater investments in preventative efforts(Reference Moodie, Tolhurst and Martin35) early in the life course. However, the benefits of doing so could extend well beyond mental well-being, to all aspects of human health and the natural systems on which it depends(Reference Swinburn, Kraak and Allender34).


Inflammatory diet patterns were cross-sectionally associated with mental well-being at age 11–12 years, with similar effects observed in mid-adulthood. Findings highlight the benefits associated with a low inflammatory diet beyond physical health, which begin at 11–12 years of age, and may emerge earlier. This highlights yet another compelling reason to urgently address inflammatory diets early in life.


We thank the wider Child Health CheckPoint team who helped collect and prepare these data. This paper uses unit record data from the Longitudinal Study of Australian Children. The study is conducted in partnership between the Department of Social Services (DSS), the Australian Institute of Family Studies (AIFS) and the Australian Bureau of Statistics (ABS). The findings and views reported in this paper are those of the authors and should not be attributed to DSS, AIFS or the ABS.

This work has been supported to date by the National Health and Medical Research Council of Australia (NHMRC; 1041352 and 1109355), The Royal Children’s Hospital Foundation (2014-241), Murdoch Children’s Research Institute, The University of Melbourne, National Heart Foundation of Australia (NHF; 100660), Financial Markets Foundation for Children (2014-055 and 2016-310) and Victorian Deaf Education Institute. The NHRMC supported K. L. (Early Career Fellowship 1091124, also Honorary NHF Postdoctoral Fellowship 101239), R. L. (Postgraduate Scholarship 1114567), D. B. (Senior Research Fellowship 1064629, also NHF Honorary Future Leader Fellowship 100369) and M. W. (Principal Research Fellowship 1160906). Research at the Murdoch Children’s Research Institute is supported by the Victorian Government’s Operational Infrastructure Program. The MCRI administered the research grants for the study and provided infrastructural support (IT and biospecimen management) to its staff and the study, but played no role in the conduct or analysis of the trial. The Department of Social Services played a role in study design; however, no other funding bodies had a role in the study design and conduct; data collection, management, analysis, and interpretation; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

K. L. led this study as the primary supervisor of D. J. W., she contributed to the study design, analysis of the data, interpretation of data and revised the article critically for important intellectual content. D. J. W. initiated this work as part of a student project, made substantial contributions to the conception and design of the study, and revised it critically. M. L. conducted the data analysis, contributed to the interpretation of data and revised the article critically for important intellectual content. A. G. oversaw the analysis of data, contributed to the interpretation of data and revised the article critically for important intellectual content. D. P. B. and L. A. B. are Investigators on the Child Health CheckPoint study; they made contributions to the study design, interpretation of data and revised the article critically for important intellectual content. R. L. and K. L. made contributions to the study design, contributed to the interpretation of data and revised the article critically for important intellectual content. M. W. and J. A. K. are Investigators on the Child Health CheckPoint study, they made major contributions to the study design, co-supervised D. J. W.’s student project, contributed to the interpretation of data and revised the article critically for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

The authors have no conflict of interest.

Supplementary material

To view supplementary material for this article, please visit


These senior authors contributed equally to this work.


Phillips, CM, Chen, LW, Heude, B, et al. (2019) Dietary inflammatory index and non-communicable disease risk: a narrative review. Nutrients 11, 1873.CrossRefGoogle ScholarPubMed
Firth, J, Marx, W, Dash, S, et al. (2019) The effects of dietary improvement on symptoms of depression and anxiety: a meta-analysis of randomized controlled trials. Psychosom Med 81, 265280.CrossRefGoogle ScholarPubMed
Kheirouri, S & Alizadeh, M (2019) Dietary inflammatory potential and the risk of incident depression in adults: a systematic review. Adv Nutr 10, 918.CrossRefGoogle ScholarPubMed
Shivappa, N, Steck, SE, Hurley, TG, et al. (2014) Designing and developing a literature-derived, population-based dietary inflammatory index. Public Health Nutr 17, 16891696.CrossRefGoogle ScholarPubMed
Phillips, CM, Shivappa, N, Hebert, JR, et al. (2018) Dietary inflammatory index and mental health: a cross-sectional analysis of the relationship with depressive symptoms, anxiety and well-being in adults. Clin Nutr 37, 14851491.CrossRefGoogle ScholarPubMed
Shivappa, N, Hebert, JR & Rashidkhani, B (2017) Association between inflammatory potential of diet and stress levels in adolescent women in Iran. Arch Iran Med 20, 108112.Google ScholarPubMed
Oddy, WH, Allen, KL, Trapp, GSA, et al. (2018) Dietary patterns, body mass index and inflammation: pathways to depression and mental health problems in adolescents. Brain Behav Immun 69, 428439.CrossRefGoogle ScholarPubMed
Esteban-Gonzalo, L, Turner, AI, Torres, SJ, et al. (2019) Diet quality and well-being in children and adolescents: the UP&DOWN longitudinal study. Br J Nutr 121, 221231.CrossRefGoogle ScholarPubMed
Mikkila, V, Rasanen, L, Raitakari, OT, et al. (2005) Consistent dietary patterns identified from childhood to adulthood: the cardiovascular risk in Young Finns Study. Br J Nutr 93, 923931.CrossRefGoogle ScholarPubMed
Kessler, RC, Berglund, P, Demler, O, et al. (2005) Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatr 62, 593602.CrossRefGoogle ScholarPubMed
Ballout, RA & Remaley, AT (2020) GlycA: a new biomarker for systemic inflammation and cardiovascular disease (CVD) risk assessment. J Lab Precis Med 5, 17.CrossRefGoogle ScholarPubMed
Krieger, N & Davey Smith, G (2016) The tale wagged by the DAG: broadening the scope of causal inference and explanation for epidemiology. Int J Epidemiol 45, 17871808.Google ScholarPubMed
Clifford, SA, Davies, S & Wake, M (2019) Child Health CheckPoint: cohort summary and methodology of a physical health and biospecimen module for the Longitudinal Study of Australian Children. BMJ Open 9, 322.CrossRefGoogle ScholarPubMed
Calder, PC, Ahluwalia, N, Brouns, F, et al. (2011) Dietary factors and low-grade inflammation in relation to overweight and obesity. Br J Nutr 3, S5S78.CrossRefGoogle Scholar
Rutishauser, IHE, Webb, K, Abraham, B, et al. (2001) Evaluation of Short Dietary Questions from the 1995 National Nutrition Survey. Australian Food and Nutrition Monitoring Unit. Herston: The University of Queensland.Google Scholar
Davis, A, Liu, R, Kerr, JA, et al. (2019) Inflammatory diet and preclinical cardiovascular phenotypes in 11–12 year-olds and mid-life adults: a cross-sectional population-based study. Atherosclerosis 285, 93101.CrossRefGoogle ScholarPubMed
Morley, B, Scully, M, Niven, P, et al. (2012) Prevalence and socio-demographic distribution of eating, physical activity and sedentary behaviours among Australian adolescents. Health Promot J Austr 23, 213218.CrossRefGoogle ScholarPubMed
Barbaresko, J, Koch, M, Schulze, MB, et al. (2013) Dietary pattern analysis and biomarkers of low-grade inflammation: a systematic literature review. Nutr Rev 71, 511527.CrossRefGoogle ScholarPubMed
Ratcliffe, J, Chen, G, Stevens, K, et al. (2016) Valuing the Child Health Utility 9D: using profile case best worst scaling methods to develop a new adolescent specific scoring algorithm. Soc Sci Med 157, 4859.CrossRefGoogle ScholarPubMed
Varni, JW, Burwinkle, TM & Seid, M (2006) The PedsQL 4.0 as a school population health measure: feasibility, reliability, and validity. Qual Life Res 15, 203215.CrossRefGoogle ScholarPubMed
Seligson, JL, Huebner, ES & Valois, RF (2003) Preliminary Validation of the Brief Multidimensional Students’ Life Satisfaction Scale (BMSLSS). Soc Indic Res 61, 121145.CrossRefGoogle Scholar
Maxwell, A, Ozmen, M, Iezzi, A, et al. (2016) Deriving population norms for the AQoL-6D and AQoL-8D multi-attribute utility instruments from web-based data. Qual Life Res 25, 32093219.CrossRefGoogle ScholarPubMed
Gasser, CE, Mensah, FK, Kerr, JA, et al. (2017) Early life socioeconomic determinants of dietary score and pattern trajectories across six waves of the Longitudinal Study of Australian Children. J Epidemiol Community Health 71, 11521160.Google ScholarPubMed
Allen, J, Balfour, R, Bell, R, et al. (2014) Social determinants of mental health. Int Rev Psychiatr 26, 392407.CrossRefGoogle ScholarPubMed
Blakemore, T & Gibbings, J (2009) Measuring family socioeconomic position. Austr Soc Policy 8, 121168.Google Scholar
Australian Bureau of Statistics (2011) Census of population and housing: Socio-Economic Indexes for Areas (SEIFA) 2011. (accessed June 2017).Google Scholar
Wake, M, Lycett, K, Clifford, SA, et al. (2013) Shared care obesity management in 3–10 year old children: 12 month outcomes of HopSCOTCH randomised trial. BMJ 112.Google ScholarPubMed
Australian Institute of Health and Welfare (2019) Overweight and obesity: an interactive insight. (accessed December 2020).Google Scholar
Godos, J, Currenti, W, Angelino, D, et al. (2020) Diet and mental health: review of the recent updates on molecular mechanisms. Antioxidants 9, 346.CrossRefGoogle ScholarPubMed
Berding, K, Vlckova, K, Marx, W, et al. (2021) Diet and the microbiota-gut-brain axis: sowing the seeds of good mental health. Adv Nutr nmaa181.Google ScholarPubMed
Wang, J, Zhou, Y, Chen, K, et al. (2018) Dietary inflammatory index and depression: a meta-analysis. Public Health Nutr 17.Google ScholarPubMed
Ioannidis, JPA (2013) Implausible results in human nutrition research. BMJ 347, f6698.CrossRefGoogle ScholarPubMed
Teng, AM, Jones, AC, Mizdrak, A, et al. (2019) Impact of sugar-sweetened beverage taxes on purchases and dietary intake: systematic review and meta-analysis. Obes Rev 20, 11871204.CrossRefGoogle ScholarPubMed
Swinburn, BA, Kraak, VI, Allender, S, et al. (2019) The Global Syndemic of Obesity, Undernutrition, and Climate Change: the Lancet Commission report. Lancet 393, 791846.CrossRefGoogle ScholarPubMed
Moodie, AR, Tolhurst, P & Martin, JE (2016) Australia’s health: being accountable for prevention. Med J Aust 204, 223225.CrossRefGoogle ScholarPubMed
Huebner, ES, Seligson, JL, Valois, RF, et al. (2006) A review of the brief multidimensional students’ life satisfaction scale. Soc Indic Res 79, 477484.CrossRefGoogle Scholar
Varni, JW, Seid, M & Kurtin, PS (2001) PedsQL 4.0: reliability and validity of the Pediatric Quality of Life Inventory version 4.0 generic core scales in healthy and patient populations. Med Care 39, 800812.CrossRefGoogle ScholarPubMed
Furber, G & Segal, L (2015) The validity of the Child Health utility instrument (CHU9D) as a routine outcome measure for use in child and adolescent mental health services. Health Qual Life Out 13, 114.CrossRefGoogle ScholarPubMed
Richardson, J, Iezzi, A, Khan, MA, et al. (2014) Validity and reliability of the Assessment of Quality of Life (AQoL)-8D multi-attribute utility instrument. Patient 7, 8596.CrossRefGoogle ScholarPubMed
Figure 0

Table 1. Mental well-being measures

Figure 1

Table 2. Characteristics of analytic sample(Mean values and standard deviations)

Figure 2

Table 3. Causal linear regression analyses comparing diet score above 75th percentile with others

Figure 3

Table 4. Direct/indirect effects of pro-inflammatory diet (>75th percentile v. Others) on mental well-being through BMI(Risk ratios (RR) and 95 % confidence intervals)

Supplementary material: File

Lycett et al. supplementary material


Download Lycett et al. supplementary material(File)
File 366.2 KB