Hostname: page-component-7c8c6479df-27gpq Total loading time: 0 Render date: 2024-03-28T14:05:57.395Z Has data issue: false hasContentIssue false

Insulin resistance and obesity, and their association with depression in relatively young people: findings from a large UK birth cohort

Published online by Cambridge University Press:  11 March 2019

B. I. Perry*
Affiliation:
Department of Psychiatry, University of Cambridge, Cambridge, England Cambridgeshire and Peterborough National Health Service Foundation Trust, Cambridge, England
G. M. Khandaker
Affiliation:
Department of Psychiatry, University of Cambridge, Cambridge, England Cambridgeshire and Peterborough National Health Service Foundation Trust, Cambridge, England National Institute for Health Research Cambridge Biomedical Research Centre, Cambridge, England
S. Marwaha
Affiliation:
Institute for Mental Health, University of Birmingham, Birmingham, England Birmingham and Solihull Mental Health Foundation NHS Trust, Birmingham, England
A. Thompson
Affiliation:
Coventry and Warwickshire Partnership NHS Trust, Coventry, England Unit of Mental Health and Wellbeing, University of Warwick, Coventry, England
S. Zammit
Affiliation:
Centre for Academic Mental Health, School of Social and Community Medicine, University of Bristol, Bristol, England Institute of Psychological Medicine and Clinical Neurosciences, Medical Research Council Centre for Neuropsychiatric Genetics and Genomics, Cardiff University, Cardiff, Wales
S. P. Singh
Affiliation:
Coventry and Warwickshire Partnership NHS Trust, Coventry, England Unit of Mental Health and Wellbeing, University of Warwick, Coventry, England
R. Upthegrove
Affiliation:
Institute for Mental Health, University of Birmingham, Birmingham, England Early Intervention Service, Birmingham Women's and Children's NHS Trust, Birmingham, UK
*
Author for correspondence: B. I. Perry, E-mail: bip20@medschl.cam.ac.uk
Rights & Permissions [Opens in a new window]

Abstract

Background

Depression frequently co-occurs with disorders of glucose and insulin homeostasis (DGIH) and obesity. Low-grade systemic inflammation and lifestyle factors in childhood may predispose to DGIH, obesity and depression. We aim to investigate the cross-sectional and longitudinal associations among DGIH, obesity and depression, and to examine the effect of demographics, lifestyle factors and antecedent low-grade inflammation on such associations in young people.

Methods

Using the Avon Longitudinal Study of Parents and Children birth cohort, we used regression analyses to examine: (1) cross-sectional and (2) longitudinal associations between measures of DGIH [insulin resistance (IR); impaired glucose tolerance] and body mass index (BMI) at ages 9 and 18 years, and depression (depressive symptoms and depressive episode) at age 18 years and (3) whether sociodemographics, lifestyle factors or inflammation [interleukin-6 (IL-6) at age 9 years] confounded any such associations.

Results

We included 3208 participants. At age 18 years, IR and BMI were positively associated with depression. These associations may be explained by sociodemographic and lifestyle factors. There were no longitudinal associations between DGIH/BMI and depression, and adjustment for IL-6 and C-reactive protein did not attenuate associations between IR/BMI and depression; however, the longitudinal analyses may have been underpowered.

Conclusions

Young people with depression show evidence of DGIH and raised BMI, which may be related to sociodemographic and lifestyle effects such as deprivation, smoking, ethnicity and gender. In future, studies with larger samples are required to confirm this. Preventative strategies for the poorer physical health outcomes associated with depression should focus on malleable lifestyle factors.

Type
Original Articles
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
Copyright © Cambridge University Press 2019

Introduction

Depression in 10–24 year olds is a leading cause of disease burden throughout the world (Gore et al., Reference Gore, Bloem, Patton, Ferguson, Joseph, Coffey, Sawyer and Mathers2011). An important aspect of this burden is the co-occurrence of disorders of glucose and insulin homeostasis (DGIH) [type-2 diabetes mellitus (T2DM) and prediabetes] and obesity (Roy and Lloyd, Reference Roy and Lloyd2012; Vancampfort et al., Reference Vancampfort, Correll, Galling, Probst, De Hert, Ward, Rosenbaum, Gaughran, Lally and Stubbs2016). This may be a consequence of disease-related factors such as symptomatology [e.g. appetite disturbance and decreased physical activity (Lysy et al., Reference Lysy, Da Costa and Dasgupta2008; Vancampfort et al., Reference Vancampfort, Firth, Schuch, Rosenbaum, Mugisha, Hallgren, Probst, Ward, Gaughran, De Hert, Carvalho and Stubbs2017)], increased rates of smoking (Katon et al., Reference Katon, von Korff, Ciechanowski, Russo, Lin, Simon, Ludman, Walker, Bush and Young2004), alcohol use (Tann et al., Reference Tann, Yabiku, Okamoto and Yanow2007), an unhealthy diet (Firth et al., Reference Firth, Stubbs, Teasdale, Ward, Veronese, Shivappa, Hebert, Berk, Yung and Sarris2018) and also sociodemographic risk factors such as either male (Nichols and Brown, Reference Nichols and Brown2003; Ali et al., Reference Ali, Stone, Peters, Davies and Khunti2006; Ding et al., Reference Ding, Song, Malik and Liu2006; Timonen et al., Reference Timonen, Rajala, Jokelainen, Keinanen-Kiukaanniemi, Meyer-Rochow and Rasanen2006; Perreault et al., Reference Perreault, Ma, Dagogo-Jack, Horton, Marrero, Crandall and Barrett-Connor2008; Menke et al., Reference Menke, Rust, Fradkin, Cheng and Cowie2014) or female (Anderson et al., Reference Anderson, Freedland, Clouse and Lustman2001; Blazer et al., Reference Blazer, Moody-Ayers, Craft-Morgan and Burchett2002; Nichols and Brown, Reference Nichols and Brown2003; Ali et al., Reference Ali, Stone, Peters, Davies and Khunti2006; Lloyd et al., Reference Lloyd, Nouwen, Sartorius, Ahmed, Alvarez, Bahendeka, Basangwa, Bobrov, Boden, Bulgari, Burti, Chaturvedi, Cimino, Gaebel, de Girolamo, Gondek, de Braude, Guntupalli, Heinze, Ji, Hong, Khan, Kiejna, Kokoszka, Kamala, Lalic, Lecic Tosevski, Mankovsky, Li, Musau, Mussig, Ndetei, Rabbani, Srikanta, Starostina, Shevchuk, Taj, Vukovic, Wolwer and Xin2018) sex, non-white European race/ethnic group (Blazer et al., Reference Blazer, Moody-Ayers, Craft-Morgan and Burchett2002; Li et al., Reference Li, Ford, Strine and Mokdad2008; Dagenais et al., Reference Dagenais, Gerstein, Zhang, McQueen, Lear, Lopez-Jaramillo, Mohan, Mony, Gupta, Kutty, Kumar, Rahman, Yusoff, Zatonska, Oguz, Rosengren, Kelishadi, Yusufali, Diaz, Avezum, Lanas, Kruger, Peer, Chifamba, Iqbal, Ismail, Xiulin, Jiankang, Wenqing, Gejie, Rangarajan, Teo and Yusuf2016; Mangurian et al., Reference Mangurian, Schillinger, Newcomer, Vittinghoff, Essock, Zhu, Dyer and Schmittdiel2018) and lower social class or adversity (Everson et al., Reference Everson, Maty, Lynch and Kaplan2002; Tamayo et al., Reference Tamayo, Christian and Rathmann2010; Pisto et al., Reference Pisto, Vaden, Sillanmaki and Mattila2014).

Another postulated mechanism is that depression, DGIH and obesity are intrinsically linked beyond the above via common antecedent inflammatory processes. Raised interleukin-6 (IL-6) and tumour necrosis factor alpha are antecedent to insulin resistance (IR) (Pickup, Reference Pickup2004; Belgardt et al., Reference Belgardt, Mauer, Wunderlich, Ernst, Pal, Spohn, Bronneke, Brodesser, Hampel, Schauss and Bruning2010), and subsequently T2DM and obesity (DeFronzo and Ferrannini, Reference DeFronzo and Ferrannini1991; Dandona et al., Reference Dandona, Aljada and Bandyopadhyay2004; Rader, Reference Rader2007). A recent genome-wide association study (Milaneschi et al., Reference Milaneschi, Lamers, Peyrot, Baune, Breen, Dehghan, Forstner, Grabe, Homuth, Kan, Lewis, Mullins, Nauck, Pistis, Preisig, Rivera, Rietschel, Streit, Strohmaier, Teumer, Van der Auwera, Wray, Boomsma, Penninx and Group2017) found that common genetic variants for body mass index (BMI) and C-reactive protein (CRP) show overlap with gene variants associated with depression. In addition, inflammation may be prospectively linked to depression in young people, with longitudinal cohort-based research finding raised levels of IL-6 during childhood to be associated with future depressive symptoms and diagnosis of depression at age 18 years, which persisted after controlling for BMI, social class, and childhood psychological and behavioural problems preceding IL-6 measurement (Khandaker et al., Reference Khandaker, Pearson, Zammit, Lewis and Jones2014; Khandaker et al., Reference Khandaker, Zammit, Burgess, Lewis and Jones2018a, Reference Khandaker, Stochl, Zammit, Goodyer, Lewis and Jones2018b). This may relate to antecedent stressful events (Slopen et al., Reference Slopen, Kubzansky, McLaughlin and Koenen2013). A recent large meta-analysis has added to these findings, with cytokines including IL-6 marked as part of a potential chemokine/cytokine profile associated with depression (Köhler et al., Reference Köhler, Freitas, Maes, de Andrade, Liu, Fernandes, Stubbs, Solmi, Veronese, Herrmann, Raison, Miller, Lanctôt and Carvalho2017).

The finding that low-grade systemic inflammation appears antecedent to DGIH, obesity and depression may be evidence of a common biological pathway that begins with an inflammatory response. Research examining the association among DGIH, obesity and depression, particularly longitudinally, in a sample of relatively young people who are less affected by years of illness, is scarce. It is nonetheless an important extension of the current literature as the findings may promote earlier and closer monitoring of metabolic and inflammatory function in young people with depression. It may in addition further our pathophysiological understanding of the multi-systemic nature of depression and suggest possible preventative therapeutic targets (Insel and Charney, Reference Insel and Charney2003).

Using longitudinal population-based data, we tested the hypothesis that even relatively young people with depression may display early signs of DGIH or obesity. We tested cross-sectional (age 18 years) and longitudinal (age 9 and 18 years) associations among DGIH, obesity and depression. We hypothesised that early signs of DGIH might be explained by shared inflammatory processes, or demographic/lifestyle factors. We tested this by assessing for any attenuation (confounding) effect of either demographic, lifestyle or inflammatory measures on associations among DGIH, obesity and depression.

Methods

Description of cohort and sample selection

The Avon Longitudinal Study of Parents and Children (ALSPAC) birth cohort (Boyd et al., Reference Boyd, Golding, Macleod, Lawlor, Fraser, Henderson, Molloy, Ness, Ring and Davey Smith2013; Fraser et al., Reference Fraser, Macdonald-Wallis, Tilling, Boyd, Golding, Davey Smith, Henderson, Macleod, Molloy, Ness, Ring, Nelson and Lawlor2013) comprises 14 062 live births from mothers residing in (former) Avon County, southwest England, with expected dates of delivery between April 1991 and December 1992. Please note that the study website contains details of all the data that are available through a fully searchable data dictionary and variable search tool (http://www.bristol.ac.uk/alspac/researchers/our-data/). Parents completed regular postal questionnaires about all aspects of their child's health and development from birth. From age 7 years, the children attended an annual assessment clinic during which they participated in various face-to-face interviews and physical tests. In an attempt to boost study numbers, further phases of recruitment took place after the age of 7 years, leading to an additional 713 participants recruited by age 18.

We first selected all participants with a measure for the outcome (n = 4563) and removed participants with CRP > 10 (Khandaker et al., Reference Khandaker, Pearson, Zammit, Lewis and Jones2014), to minimise the potential confounding effect of ongoing or recent inflammatory disease/infection on our results, leaving 3208 participants. For cross-sectional analyses, 2231 participants had complete data on all exposure and confounder variables. For longitudinal analyses, 331 participants had complete data on all exposure and confounder variables. The lower sample with complete data for longitudinal analyses is explained by age 9 glycaemic data being derived from a smaller ALSPAC sub-study (Ong et al., Reference Ong, Petry, Emmett, Sandhu, Kiess, Hales, Ness and Dunger2004). Complete-case analysis is presented in online Supplementary Tables 1 and 2. See statistical methods for our means of addressing missing data.

The study received ethics approval from the ALSPAC Ethics and Law Committee and local research ethics committees. All participants provided written informed consent.

Outcome measures

Depressive symptoms/episode at age 18 years

Depression was measured using the Clinical Interview Schedule-Revised (CIS-R), a widely used standardised self-assessment tool for measuring depression and anxiety in community samples (Lewis et al., Reference Lewis, Pelosi, Araya and Dunn1992). It includes symptoms of depression based on International Statistical Classification of Diseases, 10th Revision (ICD-10) criteria, and gives a total depression score of 0–21 comprising symptom scores for depression, depressive thoughts, fatigue, concentration and sleep problems. Our primary outcome measure was the continuous CIS-R depression score (Khandaker et al., Reference Khandaker, Pearson, Zammit, Lewis and Jones2014). As a secondary outcome, we created a binary variable ‘depressive episode’. This consisted of participants meeting ICD-10 criteria for a depressive episode (mild/moderate/severe) (F32.0/F32.1/F32.2), as has been used previously (Bowes et al., Reference Bowes, Carnegie, Pearson, Mars, Biddle, Maughan, Lewis, Fernyhough and Heron2015; Davies et al., Reference Davies, Pearson, Stapinski, Bould, Christmas, Button, Skapinakis, Lewis and Evans2016; Quarini et al., Reference Quarini, Pearson, Stein, Ramchandani, Lewis and Evans2016).

Exposures

Fasting plasma glucose, fasting insulin, glucose tolerance

We used the biochemical measurements of fasting plasma glucose (FPG) and fasting insulin (FI) (ages 9 and 18 years), and 2-h glucose tolerance (2hrGT) (age 9 years). The age 9 glycaemic data were derived from a smaller ‘Before Breakfast Study’ sub-study (Ong et al., Reference Ong, Petry, Emmett, Sandhu, Kiess, Hales, Ness and Dunger2004). At both ages 9 and 18 years, fasting samples were taken at 0900 after a 10-h fast (water only). The 2hrGT test was obtained following the above fasting procedure, with the addition of a 75 g oral bolus of sugary syrup at 0900, with blood being sampled 2 h later. Blood samples were immediately spun and frozen at −80 °C.

IR was calculated as a continuous measure from FPG and FI by using the computerised, updated version of the homoeostatic measurement for insulin resistance (HOMA2) (Levy et al., Reference Levy, Matthews and Hermans1998). The algorithm generates a relatively precise measurement of IR taking into account variations in hepatic and peripheral glucose resistance, increases in the insulin secretion curve for plasma glucose concentrations above 10 mmol/L (180 mg/dL) and the contribution of circulating proinsulin (Levy et al., Reference Levy, Matthews and Hermans1998). We did not include a binary measure for clinical IR due to the known variation in HOMA2 score between populations (Wallace et al., Reference Wallace, Levy and Matthews2004) thus ascertaining a clinical ‘cut-off’ may be problematic.

BMI

BMI was calculated from clinic data in the ALSPAC cohort, from measurements of height (m) and weight (kg). We used data collected at ages 9 and 18 years.

Demographic confounders

We adjusted for paternal social class at birth (questionnaire data, categorical) and maternal education (questionnaire data, categorical) as proxies of participant social class and potential adversity, sex (clinic data, categorical), ethnicity (questionnaire data, categorical) and maternal Edinburgh Post-Natal Depression Score (EPDS) at 8-week post-partum (questionnaire data, continuous).

Lifestyle confounders

For analyses on DGIH we adjusted for BMI (clinic data, continuous, ages 9 and 18 years as per the exposure), smoking (questionnaire data, categorical), cortisol levels [age 9 only, continuous, clinic data from BBS sub-study (Ong et al., Reference Ong, Petry, Emmett, Sandhu, Kiess, Hales, Ness and Dunger2004)], physical activity (questionnaire data on average frequency of physical activity/exercise per week in the last year, categorical) and alcohol use (questionnaire data on average frequency of use, categorical). For BMI analyses, we used the same adjustments, however with HOMA2 (as previously described, continuous) in place of BMI.

Inflammatory confounders

We adjusted for IL-6 (age 9 years) and CRP (age 18 years). IL-6 was not available at age 18 years. Blood samples were collected from non-fasting participants (age 9 years), and fasting participants (age 18 years) and were immediately spun and frozen at −80 °C. IL-6 was measured by enzyme-linked immunosorbent assay (R&D Systems), and high-sensitivity CRP (hs-CRP) was measured by automated particle-enhanced immunoturbidimetric assay (Roche). All inter-assay coefficients of variation were less than 5%.

Statistical analysis

Biomarker values that were non-normally distributed (all except for FPG) were natural log-transformed. Resultant variables, alongside the continuous outcome ‘total depression score’ were standardised (Z-transformed) so the statistical estimations represent the increase in risk of depressive symptoms per s.d. increase in exposure. We completed tests for multi-collinearity of exposures/confounders in a linear regression model. The variance inflation factor for all covariates was between 1.01 and 1.11, suggesting minimal multi-collinearity. Adjustments were added using the enter method of multiple regression. All statistical analysis was performed using IBM SPSS 24.0.

Aims 1 and 2: cross-sectional and longitudinal relationships between DGIH/BMI (ages 9 and 18 years) and depression (age 18 years)

We completed cross-sectional linear and logistic regression analyses which examined the relationship between markers of DGIH/BMI and depressive symptoms/depressive episode (age 18 years). We completed longitudinal linear and logistic regression analyses which examined the relationship between markers of DGIH/BMI (age 9 years) and depressive symptoms/episode (age 18 years). Regression coefficients and 95% confidence intervals (95% CIs) were calculated per s.d. increase in the continuous ‘total depression score’ outcome, per s.d. increase in exposure, using linear regression. Odds ratios (ORs) and 95% CIs for the categorical depressive episode outcome, per s.d. increase in exposure, were estimated using logistic regression. Quadratic terms were created separately for all exposures and entered into a logistic regression model to simulate curvilinear regression, to test the linearity of relationships between exposures and depression; these data are only shown where there was evidence of a non-linear relationship.

Aim 3: adjusting for demographic factors, lifestyle factors and inflammation

We performed adjustments using linear and logistic regression as described above. First, we adjusted for the demographic, lifestyle and inflammation factors listed above, separately. Second, we completed a total adjustment model including all potential confounders together.

Missing data

Missing data for exposures and confounders were present in 30% of cases for cross-sectional analyses, and 90% of cases for longitudinal analyses. Due to the substantial amount of missing data in longitudinal analyses [which may be related to the age 9 glycaemic data being derived from a smaller ALSPAC sub-study (Ong et al., Reference Ong, Petry, Emmett, Sandhu, Kiess, Hales, Ness and Dunger2004)], we longitudinally analysed only complete cases (Lee and Huber, Reference Lee and Huber2011). For our cross-sectional analyses, Little's Missing Completely at Random (MCAR) test (p = 0.008) indicated that the data were not MCAR. We then used the missing value analysis function of SPSS to perform separate variance independent t tests (continuous variables) and χ2 tests (categorical variables) to check the missing at random (MAR) assumption. Each variable returned significance (p < 0.05) with at least one other included variable, indicating that missingness was correlated with another variable in the model, suggesting the missing data met the MAR assumption.

We completed multiple imputation (MI) using the fully conditional Markov chain Monte Carlo method, for all exposure and confounder variables, plus axillary continuous variables that were indicators of missingness in the population. The selected axillary variables included age 9 biochemical data (high-density lipoprotein, low-density lipoprotein, triglycerides), as well as birthweight and gestational age. As missing data were present in 30% of cases, we used 30 imputations as recommended (White et al., Reference White, Royston and Wood2011). Complete case analysis for the cross-sectional analyses is presented in online Supplementary Tables 1 and 2.

Results

Following imputation for exposure and confounder variables, our total sample was 3208 participants. The mean depression score in the imputed sample was 3.09; range 0–21 (complete cases 3.10; range 0–21). The number of participants meeting criteria for a depressive episode at age 18 was n = 227 (7%) (complete cases n = 179; 8%). Table 1 shows the sample clinical and biomarker characteristics at age 18.

Table 1. Baseline characteristics of sample

a Categorical variables (sex, social class, ethnicity, smoking) were compared using the χ2 test, normally distributed continuous variables (FPG, birthweight, gestational age) were compared using the two tailed t test; non-normally distributed continuous variables (HOMA2, CRP, IL-6, BMI) were compared using the Mann–Whitney U test.

b Maternal EPDS score recorded at 8 week post-partum.

c Frequency participant has had drinks containing alcohol.

d Physical activity corresponded to frequency respondent engaged in going to gym, brisk walking or any sports activity during the past year

Aim 1: cross-sectional association between DGIH/BMI and depression (age 18 years)

In the unadjusted analyses; HOMA2, FI and BMI were positively associated with depressive symptoms at age 18 years [β = 0.04 (95% CI 0.03–0.30) p = 0.02; β = 0.05 (95% CI 0.03–0.33) p = 0.01; β = 0.03 (95% CI 0.01–0.08) p = 0.04 respectively]; FPG was negatively associated with depressive symptoms [β = −0.05 (95% CI −0.36 to −0.02) p = 0.01]. See Table 2. The results from our complete case analysis were broadly similar. See online Supplementary Table 1.

Table 2. Cross-sectional association between DGIH/BMI and depressive symptoms (age 18)

a Not adjusted for in BMI analysis.

b Not adjusted for in HOMA/FPG/FI analysis.

*Indicates p < 0.05.

In the unadjusted analyses, both HOMA2 and FI were associated with the categorical depressive episode [OR 1.14 (95% CI 1.01–1.31) p = 0.04 and OR 1.16 (95% CI 1.01–1.33) p = 0.03 respectively]. See Table 3. The results are similar to those for our complete case analysis. See online Supplementary Table 2.

Table 3. Cross-sectional associations between DGIH/BMI and depressive episode (age 18 years)

a Not adjusted for in BMI analysis.

b Not adjusted for in HOMA/FPG/FI analysis.

*Indicates p < 0.05.

Aim 2: longitudinal association between DGIH/BMI (age 9 years) and depressive symptoms/episode (age 18 years)

Our longitudinal analysis of DGIH included 399 participants, and for BMI 2571 participants. There were no evident longitudinal associations between DGIH/BMI at age 9 years and depressive symptoms/episode at age 18. See Tables 4 and 5.

Table 4. Longitudinal association between DGIH/BMI (age 9) and depressive symptoms (age 18 years)

a Not adjusted for in BMI analysis.

b Not adjusted for in HOMA/FPG/FI analysis.

*Indicates p < 0.05.

Table 5. Longitudinal association between DGIH/BMI (age 9) and depressive episode (age 18 years)

a Not adjusted for in BMI analysis.

b Not adjusted for in HOMA/FPG/FI analysis.

*Indicates p < 0.05.

Aim 3: adjusting for confounders

As shown in Tables 2–5, after adjustments for demographic and lifestyle factors the point estimates did not change considerably but the 95% CIs widened to include the null. Adjustment for immune markers did not significantly alter the unadjusted associations. Following adjustment for all confounder variables in one model, there were no significant associations. Results from the complete case analysis are broadly similar. See online Supplementary Tables 1 and 2.

Discussion

In this study, we first tested the cross-sectional associations between DGIH/BMI and depression in a sample of young people who may have been less affected by years of illness, before and after adjustments for potential demographic, lifestyle and immune confounders. We then used longitudinal analysis to test the direction of association between these factors. To our knowledge, this is one of the first analyses of detailed longitudinal associations among DGIH, BMI, inflammation and depression, in a relatively young sample, albeit the sample size for some of the analyses was relatively small. We present several findings of note.

We found that the broadest marker of glycaemic function, FPG, was negatively associated with depressive symptoms at age 18 years. In addition, more sensitive markers of pre-clinical glucose dysregulation (FI, HOMA2) were positively associated with depression at age 18 years. IR in combination with low FPG shows biological plausibility; FPG can present low-normal in early IR, in response to the IR phenotype of increased insulin secretion thus increased intracellular glucose uptake (Ensling et al., Reference Ensling, Steinmann and Whaley-Connell2011). These associations were not significant following demographic and lifestyle adjustments suggesting that the metabolic dysfunction present in depression may be attributable to sociodemographic and lifestyle factors. Interestingly however, whilst the CIs became larger (and included the null) after confounding adjustments, the point estimates did not change substantially. Research conducted on larger samples of depressed patients would therefore be appropriate to increase statistical power, to further test these findings.

Our findings differ from previous longitudinal research from the Northern Finland Birth Cohort (NFBC) (Timonen et al., Reference Timonen, Rajala, Jokelainen, Keinanen-Kiukaanniemi, Meyer-Rochow and Rasanen2006), which found IR to be cross-sectionally associated with depressive symptoms even after adjustments for similar confounders. However, in that study, a different mathematical method, the Qualitative Insulin Sensitivity Check Index (QUICKI) (Katz et al., Reference Katz, Nambi, Mather, Baron, Follmann, Sullivan and Quon2000) was used to measure IR. The QUICKI is limited in being blind to several important physiological aspects of glucose homoeostasis; and being calibrated to an aged insulin assay (Wallace et al., Reference Wallace, Levy and Matthews2004). We used the computerised, updated HOMA2 model that addresses the shortcomings of QUICKI and other early models. Additionally, the participants in the Finnish study were older (age 31 years) thus the potential for confounding by the potential chronic lifestyle factors of depression was increased. Another previous study using NFBC data found no association between depression and the wider metabolic syndrome as a whole, after controlling for similar adjustments at age 31 years (Herva et al., Reference Herva, Räsänen, Miettunen, Timonen, Läksy, Veijola, Laitinen, Ruokonen and Joukamaa2006). The metabolic syndrome classification may be less sensitive to early metabolic dysfunction than measures of IR. For example, a smaller cross-sectional study from Taiwan of 323 participants (mean age of 19.5 years) found no association between depression and the metabolic syndrome as a whole, but did find associations with specific elements of metabolic dysfunction such as BMI and hypertension, though the associations attenuated following adjustments (Lin et al., Reference Lin, Liang, Liao and Tsay2014). Whilst this latter study was relatively small, the results are in line with ours. Another study of young adults from the USA found depression to be associated with the metabolic syndrome at age 30 years (Kinder et al., Reference Kinder, Carnethon, Palaniappan, King and Fortmann2004).

We found no associations longitudinally between age 9 markers of DGIH/BMI and later development of depressive symptoms/episode by age 18 years. However, our longitudinal analyses of glycaemic function were susceptible to reduced statistical power due to the smaller sample that underwent glycaemic testing at age 9 years in the cohort. Therefore, since the number of depression events in the longitudinal analyses of glycaemic function was relatively small, the corresponding demographic and lifestyle adjustment models may have been susceptible to model overfit (Peduzzi et al., Reference Peduzzi, Concato, Kemper, Holford and Feinstein1996), limiting the generalisability of these results. Results for the longitudinal analyses should therefore be interpreted with caution.

Taken together, our results suggest that the sociodemographic and lifestyle features of the depressive syndrome such as gender, ethnicity, paternal social class, smoking, alcohol use and physical activity levels may be driving the known associations between T2DM, obesity and depression. Nonetheless, sensitive metabolic changes are apparent from a relatively early age. This is an important finding. Whilst the demographic confounders we adjusted for are fixed, the lifestyle confounders we adjusted for may be malleable. For that reason, our results demonstrate the crucial importance for even relatively young patients diagnosed with depression to receive a full and comprehensive assessment of metabolic function. Encouragement and importance should be placed on encouraging and incentivising positive lifestyle changes, such as smoking cessation and reducing alcohol intake. Interestingly, other relevant lifestyle changes such as encouraging a healthy diet and regular exercise show some evidence for having intrinsic mood-boosting properties (Jacka et al., Reference Jacka, Kremer, Leslie, Berk, Patton, Toumbourou and Williams2010; Cooney et al., Reference Cooney, Dwan and Mead2014).

Strengths and limitations

The ALSPAC cohort provided a relatively large sample size in which to conduct analyses and we were able to consider detailed potential confounders including current/recent inflammation, alcohol use, BMI (in IR analyses), IR (in BMI analyses), smoking, physical activity, maternal post-natal EPDS score, paternal social class, ethnicity and sex. We attempted to reduce bias and increase statistical power (Dong and Peng, Reference Dong and Peng2013) by using the MI method to account for missing data where possible.

However, there are several limitations that should be considered. Firstly, we have put emphasis on the effects of lifestyle on glycaemic and anthropometric parameters, however the lifestyle data we collected in our analyses was mostly collected via self-report questionnaires. Self-report questionnaire data on lifestyle factors can be limited in its validity and reliability, for reasons such as social desirability or recall bias (Sallis and Saelens, Reference Sallis and Saelens2000; Del Boca and Darkes, Reference Del Boca and Darkes2003; Shipton et al., Reference Shipton, Tappin, Vadiveloo, Crossley, Aitken and Chalmers2009). ALSPAC does have a quantitative measure for physical activity at age 18 years. Participants were invited to take part in a week-long wrist accelerometer study, however, the sample size that were able to provide full data on this study was much smaller than data available for the self-report measure; participants were asked to remove the accelerometer for certain types of exercise; and, we felt that data collected in this manner may be susceptible to the Hawthorne effect. Nonetheless, the limitations of using self-report data for lifestyle parameters should be taken into account when interpreting our findings.

Due to the significant amount of missing data in our longitudinal analyses, imputing such missing data may have led to both to selection bias and to type-II statistical error (Lee and Huber, Reference Lee and Huber2011). It is therefore likely our longitudinal analyses, particularly of glycaemic function, are underpowered. In such analyses, models including demographic or lifestyle adjustments are likely to be overfit, limiting their generalisability. Our longitudinal findings should therefore be interpreted with caution. Selection bias is also a possibility since not all ALSPAC participants attended voluntary CIS-R assessment at age 18. In addition, we have used paternal social class at birth as a proxy of social class and potential adversity of the participant and these suppositions may be open to challenge. Furthermore, whilst most biochemical tests were sampled in the fasting state, age 9 inflammatory markers were sampled in the non-fasting state, which may increase measurement error. Measurement error can introduce a bias towards the null, so the results for IL-6 may be underestimated. Also, we have examined an a priori hypothesis based upon potential biological plausibility involving immune dysfunction upstream of both DGIH and BMI. However, obesity is known to be a pro-inflammatory state itself (Jung and Choi, Reference Jung and Choi2014), thus reverse causality may be a possibility. Future research may seek to take this into account. Additionally, whilst we restricted our analyses to participants with CRP < 10 mg/L to account for chronic/acute infection/inflammatory illness, we were unable to ascertain whether included participants were in receipt of immune-modulating medications. Finally, whilst we included data for IL-6 and CRP, future analyses may seek to examine additional circulating markers of innate and adaptive immune response.

Implications and future directions

Our findings have implications both in the assessment and management of patients who present with symptoms of depression. We found that even at the relatively young age of 18 years, depression is associated with DGIH and raised BMI. That the metabolic associations with depression can occur at such an early phase of a potentially chronic course of depression is significant and underlines the need for swift and comprehensive assessment and management of metabolic risk factors in people that present with depression. Our findings may provide impetus for the monitoring of more sensitive measures of metabolic function in people first presenting with depression, since the elements that make up the ‘metabolic syndrome’, of which IR and BMI are a part, are by definition reversible (Alberti et al., Reference Alberti, Zimmet and Shaw2005), and therefore early intervention may help to attenuate the significant morbidity (Goldney et al., Reference Goldney, Phillips, Fisher and Wilson2004) and socioeconomic cost (Molosankwe et al., Reference Molosankwe, Patel, Jose Gagliardino, Knapp and McDaid2012) associated with comorbid depression and metabolic dysfunction. Taken together with previous research (Khandaker et al., Reference Khandaker, Pearson, Zammit, Lewis and Jones2014; Miller and Raison, Reference Miller and Raison2016; Khandaker et al., Reference Khandaker, Zammit, Burgess, Lewis and Jones2018a) which suggests that immune dysfunction could be a target for prevention and treatment of depression, our findings may suggest that other factors also play an important role in increasing the physical health burden associated with depression. Impetus should be placed on encouraging healthy lifestyles such as with a healthy diet and exercise, which have both shown to be beneficial in improving depression (Schuch et al., Reference Schuch, Deslandes, Stubbs, Gosmann, Silva and Fleck2016a, Reference Schuch, Vancampfort, Richards, Rosenbaum, Ward and Stubbs2016b; Teasdale et al., Reference Teasdale, Ward, Rosenbaum, Samaras and Stubbs2017).

Future research should seek to examine associations between young adults with depression and measures of dyslipidaemia, which may also be relevant (Parekh et al., Reference Parekh, Smeeth, Milner and Thure2017), and should seek to address whether improved recognition and interventions for modifiable lifestyle factors in the early treatment of depression may result in more favourable long-term physical health outcomes.

Supplementary material

The supplementary material for this article can be found at https://doi.org/10.1017/S0033291719000308

Author ORCIDs

B. I. Perry, 0000-0002-1533-026X

Acknowledgements

We are extremely grateful to all families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, including interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses. The authors would also like to extend thanks to Kay Wright (R&I manager, Coventry and Warwickshire Partnership NHS Trust) for her help in facilitating the funding for data access.

Financial support

The UK Medical Research Council and Wellcome Trust (Grant no. 102215/2/13/2) and the University of Bristol provide core support for ALSPAC. This publication is the work of the authors, who will serve as guarantors for the contents of this paper. Funding for data access was obtained from Coventry and Warwickshire Partnership NHS Trust Research Development Fund. This report is independent research supported by the National Institute for Health Research (NIHR Doctoral Research Fellowship, Dr Benjamin Ian Perry, DRF-2018-11-ST2-018). The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health and Social Care. Dr Khandaker acknowledges funding support from the Wellcome Trust (Intermediate Clinical Fellowship; grant no. 201486/Z/16/Z), MRC (MICA: Mental Health Data Pathfinder; grant no. MC_PC_17213) and MQ: Transforming Mental Health (Data Science Award; grant no. MQDS17/40). A comprehensive list of grants funding is available on the ALSPAC website (http://www.bristol.ac.uk/alspac/external/documents/grant-acknowledgements.pdf/); this research was specifically funded by The Wellcome Trust (Grant no. 08426812/Z/07/Z).

Conflict of interest

None.

References

Alberti, KG, Zimmet, P, Shaw, J and Group IDFETFC (2005) The metabolic syndrome – a new worldwide definition. Lancet 366, 10591062.CrossRefGoogle ScholarPubMed
Ali, S, Stone, MA, Peters, JL, Davies, MJ and Khunti, K (2006) The prevalence of co-morbid depression in adults with type 2 diabetes: a systematic review and meta-analysis. Diabetic Medicine 23, 11651173.CrossRefGoogle ScholarPubMed
Anderson, RJ, Freedland, KE, Clouse, RE and Lustman, PJ (2001) The prevalence of comorbid depression in adults with diabetes: a meta-analysis. Diabetes Care 24, 10691078.CrossRefGoogle ScholarPubMed
Belgardt, BF, Mauer, J, Wunderlich, FT, Ernst, MB, Pal, M, Spohn, G, Bronneke, HS, Brodesser, S, Hampel, B, Schauss, AC and Bruning, JC (2010) Hypothalamic and pituitary c-Jun N-terminal kinase 1 signaling coordinately regulates glucose metabolism. Proceedings of the National Academy of Sciences of the United States of America 107, 60286033.CrossRefGoogle ScholarPubMed
Blazer, DG, Moody-Ayers, S, Craft-Morgan, J and Burchett, B (2002) Depression in diabetes and obesity: racial/ethnic/gender issues in older adults. Journal of Psychosomatic Research 53, 913916.CrossRefGoogle ScholarPubMed
Bowes, L, Carnegie, R, Pearson, R, Mars, B, Biddle, L, Maughan, B, Lewis, G, Fernyhough, C and Heron, J (2015) Risk of depression and self-harm in teenagers identifying with goth subculture: a longitudinal cohort study. The Lancet. Psychiatry 2, 793800.CrossRefGoogle ScholarPubMed
Boyd, A, Golding, J, Macleod, J, Lawlor, DA, Fraser, A, Henderson, J, Molloy, L, Ness, A, Ring, S and Davey Smith, G (2013) Cohort profile: the ‘children of the 90s’ – the index offspring of the Avon Longitudinal Study of Parents and Children. International Journal of Epidemiology 42, 111127.CrossRefGoogle ScholarPubMed
Cooney, G, Dwan, K and Mead, G (2014) Exercise for depression. JAMA 311, 24322433.CrossRefGoogle ScholarPubMed
Dagenais, GR, Gerstein, HC, Zhang, X, McQueen, M, Lear, S, Lopez-Jaramillo, P, Mohan, V, Mony, P, Gupta, R, Kutty, VR, Kumar, R, Rahman, O, Yusoff, K, Zatonska, K, Oguz, A, Rosengren, A, Kelishadi, R, Yusufali, A, Diaz, R, Avezum, A, Lanas, F, Kruger, A, Peer, N, Chifamba, J, Iqbal, R, Ismail, N, Xiulin, B, Jiankang, L, Wenqing, D, Gejie, Y, Rangarajan, S, Teo, K and Yusuf, S (2016) Variations in diabetes prevalence in low-, middle-, and high-income countries: results from the prospective urban and rural epidemiological study. Diabetes Care 39, 780787.CrossRefGoogle Scholar
Dandona, P, Aljada, A and Bandyopadhyay, A (2004) Inflammation: the link between insulin resistance, obesity and diabetes. Trends in Immunology 25, 47.CrossRefGoogle ScholarPubMed
Davies, SJ, Pearson, RM, Stapinski, L, Bould, H, Christmas, DM, Button, KS, Skapinakis, P, Lewis, G and Evans, J (2016) Symptoms of generalized anxiety disorder but not panic disorder at age 15 years increase the risk of depression at 18 years in the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort study. Psychological Medicine 46, 7385.CrossRefGoogle Scholar
DeFronzo, RA and Ferrannini, E (1991) Insulin resistance. A multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease. Diabetes Care 14, 173194.CrossRefGoogle ScholarPubMed
Del Boca, FK and Darkes, J (2003) The validity of self-reports of alcohol consumption: state of the science and challenges for research. Addiction 98(suppl. 2), 112.CrossRefGoogle ScholarPubMed
Ding, EL, Song, Y, Malik, VS and Liu, S (2006) Sex differences of endogenous sex hormones and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA 295, 12881299.CrossRefGoogle ScholarPubMed
Dong, Y and Peng, CY (2013) Principled missing data methods for researchers. Springerplus 2, 222.CrossRefGoogle ScholarPubMed
Ensling, M, Steinmann, W and Whaley-Connell, A (2011) Hypoglycemia: a possible link between insulin resistance, metabolic dyslipidemia, and heart and kidney disease (the Cardiorenal Syndrome). Cardiorenal Medicine 1, 6774.CrossRefGoogle Scholar
Everson, SA, Maty, SC, Lynch, JW and Kaplan, GA (2002) Epidemiologic evidence for the relation between socioeconomic status and depression, obesity, and diabetes. Journal of Psychosomatic Research 53, 891895.CrossRefGoogle Scholar
Firth, J, Stubbs, B, Teasdale, SB, Ward, PB, Veronese, N, Shivappa, N, Hebert, JR, Berk, M, Yung, AR and Sarris, J (2018) Diet as a hot topic in psychiatry: a population-scale study of nutritional intake and inflammatory potential in severe mental illness. World Psychiatry 17, 365367.CrossRefGoogle ScholarPubMed
Fraser, A, Macdonald-Wallis, C, Tilling, K, Boyd, A, Golding, J, Davey Smith, G, Henderson, J, Macleod, J, Molloy, L, Ness, A, Ring, S, Nelson, SM and Lawlor, DA (2013) Cohort profile: the Avon Longitudinal Study of Parents and Children: ALSPAC mothers cohort. International Journal of Epidemiology 42, 97110.CrossRefGoogle ScholarPubMed
Goldney, RD, Phillips, PJ, Fisher, LJ and Wilson, DH (2004) Diabetes, depression, and quality of life: a population study. Diabetes Care 27, 10661070.CrossRefGoogle ScholarPubMed
Gore, FM, Bloem, PJ, Patton, GC, Ferguson, J, Joseph, V, Coffey, C, Sawyer, SM and Mathers, CD (2011) Global burden of disease in young people aged 10–24 years: a systematic analysis. Lancet 377, 20932102.CrossRefGoogle ScholarPubMed
Herva, A, Räsänen, P, Miettunen, J, Timonen, M, Läksy, K, Veijola, J, Laitinen, J, Ruokonen, A and Joukamaa, M (2006) Co-occurrence of metabolic syndrome with depression and anxiety in young adults: the Northern Finland 1966 Birth Cohort Study. Psychosomatic Medicine 68, 213216.CrossRefGoogle ScholarPubMed
Insel, TR and Charney, DS (2003) Research on major depression: strategies and priorities. JAMA 289, 31673168.CrossRefGoogle ScholarPubMed
Jacka, FN, Kremer, PJ, Leslie, ER, Berk, M, Patton, GC, Toumbourou, JW and Williams, JW (2010) Associations between diet quality and depressed mood in adolescents: results from the Australian Healthy Neighbourhoods Study. Australia and New Zealand Journal of Psychiatry 44, 435442.CrossRefGoogle ScholarPubMed
Jung, UJ and Choi, MS (2014) Obesity and its metabolic complications: the role of adipokines and the relationship between obesity, inflammation, insulin resistance, dyslipidemia and nonalcoholic fatty liver disease. International Journal of Molecular Science 15, 61846223.CrossRefGoogle ScholarPubMed
Katon, W, von Korff, M, Ciechanowski, P, Russo, J, Lin, E, Simon, G, Ludman, E, Walker, E, Bush, T and Young, B (2004) Behavioral and clinical factors associated with depression among individuals with diabetes. Diabetes Care 27, 914920.CrossRefGoogle ScholarPubMed
Katz, A, Nambi, SS, Mather, K, Baron, AD, Follmann, DA, Sullivan, G and Quon, MJ (2000) Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans. Journal of Clinical Endocrinology & Metabolism 85, 24022410.CrossRefGoogle ScholarPubMed
Khandaker, GM, Pearson, RM, Zammit, S, Lewis, G and Jones, PB (2014) Association of serum interleukin 6 and C-reactive protein in childhood with depression and psychosis in young adult life: a population-based longitudinal study. JAMA Psychiatry 71, 11211128.CrossRefGoogle ScholarPubMed
Khandaker, GM, Zammit, S, Burgess, S, Lewis, G and Jones, PB (2018 a) Association between a functional interleukin 6 receptor genetic variant and risk of depression and psychosis in a population-based birth cohort. Brain Behavior Immunology 69, 264272.CrossRefGoogle Scholar
Khandaker, GM, Stochl, J, Zammit, S, Goodyer, I, Lewis, G and Jones, PB (2018 b) Childhood inflammatory markers and intelligence as predictors of subsequent persistent depressive symptoms: a longitudinal cohort study. Psychological Medicine 48, 15141522.CrossRefGoogle ScholarPubMed
Kinder, LS, Carnethon, MR, Palaniappan, LP, King, AC and Fortmann, SP (2004) Depression and the metabolic syndrome in young adults: findings from the Third National Health and Nutrition Examination Survey. Psychosomatic Medicine 66, 316322.CrossRefGoogle ScholarPubMed
Köhler, CA, Freitas, TH, Maes, M, de Andrade, NQ, Liu, CS, Fernandes, BS, Stubbs, B, Solmi, M, Veronese, N, Herrmann, N, Raison, CL, Miller, BJ, Lanctôt, KL and Carvalho, AF (2017) Peripheral cytokine and chemokine alterations in depression: a meta-analysis of 82 studies. Acta Psychiatrica Scandinavica 135, 373387.CrossRefGoogle ScholarPubMed
Lee, HJ and Huber, J (2011) Multiple imputation with large proportions of missing data: How much is too much? In United Kingdom Stata Users’ Group Meetings 2011. Stata Users Group.Google Scholar
Levy, JC, Matthews, DR and Hermans, MP (1998) Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care 21, 21912192.CrossRefGoogle ScholarPubMed
Lewis, G, Pelosi, AJ, Araya, R and Dunn, G (1992) Measuring psychiatric disorder in the community: a standardized assessment for use by lay interviewers. Psychological Medicine 22, 465486.CrossRefGoogle ScholarPubMed
Li, C, Ford, ES, Strine, TW and Mokdad, AH (2008) Prevalence of depression among U.S. adults with diabetes: findings from the 2006 behavioral risk factor surveillance system. Diabetes Care 31, 105107.CrossRefGoogle ScholarPubMed
Lin, KP, Liang, TL, Liao, IC and Tsay, SL (2014) Associations among depression, obesity, and metabolic syndrome in young adult females. Biological Research for Nursing 16, 327334.CrossRefGoogle ScholarPubMed
Lloyd, CE, Nouwen, A, Sartorius, N, Ahmed, HU, Alvarez, A, Bahendeka, S, Basangwa, D, Bobrov, AE, Boden, S, Bulgari, V, Burti, L, Chaturvedi, SK, Cimino, LC, Gaebel, W, de Girolamo, G, Gondek, TM, de Braude, MG, Guntupalli, A, Heinze, MG, Ji, L, Hong, X, Khan, A, Kiejna, A, Kokoszka, A, Kamala, T, Lalic, NM, Lecic Tosevski, D, Mankovsky, B, Li, M, Musau, A, Mussig, K, Ndetei, D, Rabbani, G, Srikanta, SS, Starostina, EG, Shevchuk, M, Taj, R, Vukovic, O, Wolwer, W and Xin, Y (2018) Prevalence and correlates of depressive disorders in people with Type 2 diabetes: results from the International Prevalence and Treatment of Diabetes and Depression (INTERPRET-DD) study, a collaborative study carried out in 14 countries. Diabetic Medicine 35, 760769.CrossRefGoogle ScholarPubMed
Lysy, Z, Da Costa, D and Dasgupta, K (2008) The association of physical activity and depression in Type 2 diabetes. Diabetic Medicine 25, 11331141.CrossRefGoogle ScholarPubMed
Mangurian, CV, Schillinger, D, Newcomer, JW, Vittinghoff, E, Essock, SM, Zhu, Z, Dyer, WT and Schmittdiel, JA (2018) Diabetes and prediabetes prevalence by race and ethnicity among people with severe mental illness. Diabetes Care 41, e119e120.CrossRefGoogle ScholarPubMed
Menke, A, Rust, KF, Fradkin, J, Cheng, YJ and Cowie, CC (2014) Associations between trends in race/ethnicity, aging, and body mass index with diabetes prevalence in the United States: a series of cross-sectional studies. Annals of Internal Medicine 161, 328335.CrossRefGoogle Scholar
Milaneschi, Y, Lamers, F, Peyrot, WJ, Baune, BT, Breen, G, Dehghan, A, Forstner, AJ, Grabe, HJ, Homuth, G, Kan, C, Lewis, C, Mullins, N, Nauck, M, Pistis, G, Preisig, M, Rivera, M, Rietschel, M, Streit, F, Strohmaier, J, Teumer, A, Van der Auwera, S, Wray, NR, Boomsma, DI, Penninx, B, Group, CIW and the Major Depressive Disorder Working Group of the Psychiatric Genomics C (2017) Genetic association of major depression with atypical features and obesity-related immunometabolic dysregulations. JAMA Psychiatry 74, 12141225.CrossRefGoogle ScholarPubMed
Miller, AH and Raison, CL (2016) The role of inflammation in depression: from evolutionary imperative to modern treatment target. Nature Reviews Immunology 16, 2234.CrossRefGoogle ScholarPubMed
Molosankwe, I, Patel, A, Jose Gagliardino, J, Knapp, M and McDaid, D (2012) Economic aspects of the association between diabetes and depression: a systematic review. Journal of Affective Disorders 142(suppl), S42S55.CrossRefGoogle ScholarPubMed
Nichols, GA and Brown, JB (2003) Unadjusted and adjusted prevalence of diagnosed depression in type 2 diabetes. Diabetes Care 26, 744749.CrossRefGoogle ScholarPubMed
Ong, KK, Petry, CJ, Emmett, PM, Sandhu, MS, Kiess, W, Hales, CN, Ness, AR, Dunger, DB and team As (2004) Insulin sensitivity and secretion in normal children related to size at birth, postnatal growth, and plasma insulin-like growth factor-I levels. Diabetologia 47, 10641070.CrossRefGoogle ScholarPubMed
Parekh, A, Smeeth, D, Milner, Y and Thure, S (2017) The role of lipid biomarkers in major depression. Healthcare (Basel) 5(1), 5.CrossRefGoogle ScholarPubMed
Peduzzi, P, Concato, J, Kemper, E, Holford, TR and Feinstein, AR (1996) A simulation study of the number of events per variable in logistic regression analysis. Journal of Clinical Epidemiology 49, 13731379.CrossRefGoogle ScholarPubMed
Perreault, L, Ma, Y, Dagogo-Jack, S, Horton, E, Marrero, D, Crandall, J, Barrett-Connor, E and Diabetes Prevention P (2008) Sex differences in diabetes risk and the effect of intensive lifestyle modification in the Diabetes Prevention Program. Diabetes Care 31, 14161421.CrossRefGoogle ScholarPubMed
Pickup, JC (2004) Inflammation and activated innate immunity in the pathogenesis of type 2 diabetes. Diabetes Care 27, 813823.CrossRefGoogle ScholarPubMed
Pisto, L, Vaden, A, Sillanmaki, L and Mattila, K (2014) Childhood adversities are associated with diabetes management in working age in Finland. International Journal of Family Medicine 2014, 864572.CrossRefGoogle ScholarPubMed
Quarini, C, Pearson, RM, Stein, A, Ramchandani, PG, Lewis, G and Evans, J (2016) Are female children more vulnerable to the long-term effects of maternal depression during pregnancy? Journal of Affective Disorders 189, 329335.CrossRefGoogle ScholarPubMed
Rader, DJ (2007) Effect of insulin resistance, dyslipidemia, and intra-abdominal adiposity on the development of cardiovascular disease and diabetes mellitus. American Journal of Medicine 120, S12S18.CrossRefGoogle ScholarPubMed
Roy, T and Lloyd, CE (2012) Epidemiology of depression and diabetes: a systematic review. Journal of Affective Disorders 142(suppl), S821.CrossRefGoogle ScholarPubMed
Sallis, JF and Saelens, BE (2000). Assessment of physical activity by self-report: status, limitations, and future directions. Research Quarterly for Exercise & Sport 71(suppl. 2), 114.CrossRefGoogle ScholarPubMed
Schuch, FB, Deslandes, AC, Stubbs, B, Gosmann, NP, Silva, CT and Fleck, MP (2016 a) Neurobiological effects of exercise on major depressive disorder: a systematic review. Neuroscience and Biobehavioral Reviews 61, 111.CrossRefGoogle ScholarPubMed
Schuch, FB, Vancampfort, D, Richards, J, Rosenbaum, S, Ward, PB and Stubbs, B (2016 b) Exercise as a treatment for depression: a meta-analysis adjusting for publication bias. Journal of Psychiatric Research 77, 4251.CrossRefGoogle ScholarPubMed
Shipton, D, Tappin, DM, Vadiveloo, T, Crossley, JA, Aitken, DA and Chalmers, J (2009) Reliability of self reported smoking status by pregnant women for estimating smoking prevalence: a retrospective, cross sectional study. BMJ 339, b4347.CrossRefGoogle ScholarPubMed
Slopen, N, Kubzansky, LD, McLaughlin, KA and Koenen, KC (2013) Childhood adversity and inflammatory processes in youth: a prospective study. Psychoneuroendocrinology 38, 188200.CrossRefGoogle ScholarPubMed
Tamayo, T, Christian, H and Rathmann, W (2010) Impact of early psychosocial factors (childhood socioeconomic factors and adversities) on future risk of type 2 diabetes, metabolic disturbances and obesity: a systematic review. BMC Public Health 10, 525.CrossRefGoogle ScholarPubMed
Tann, SS, Yabiku, ST, Okamoto, SK and Yanow, J (2007) triADD: the risk for alcohol abuse, depression, and diabetes multimorbidity in the American Indian and Alaska Native populations. American Indian Alaskan Native Mental Health Research 14, 123.Google ScholarPubMed
Teasdale, SB, Ward, PB, Rosenbaum, S, Samaras, K and Stubbs, B (2017) Solving a weighty problem: systematic review and meta-analysis of nutrition interventions in severe mental illness. British Journal of Psychiatry 210, 110118.CrossRefGoogle ScholarPubMed
Timonen, M, Rajala, U, Jokelainen, J, Keinanen-Kiukaanniemi, S, Meyer-Rochow, VB and Rasanen, P (2006) Depressive symptoms and insulin resistance in young adult males: results from the Northern Finland 1966 birth cohort. Molecular Psychiatry 11, 929933.CrossRefGoogle ScholarPubMed
Vancampfort, D, Correll, CU, Galling, B, Probst, M, De Hert, M, Ward, PB, Rosenbaum, S, Gaughran, F, Lally, J and Stubbs, B (2016) Diabetes mellitus in people with schizophrenia, bipolar disorder and major depressive disorder: a systematic review and large scale meta-analysis. World Psychiatry 15, 166174.CrossRefGoogle ScholarPubMed
Vancampfort, D, Firth, J, Schuch, FB, Rosenbaum, S, Mugisha, J, Hallgren, M, Probst, M, Ward, PB, Gaughran, F, De Hert, M, Carvalho, AF and Stubbs, B (2017) Sedentary behavior and physical activity levels in people with schizophrenia, bipolar disorder and major depressive disorder: a global systematic review and meta-analysis. World Psychiatry 16, 308315.CrossRefGoogle ScholarPubMed
Wallace, TM, Levy, JC and Matthews, DR (2004) Use and abuse of HOMA modeling. Diabetes Care 27, 14871495.CrossRefGoogle ScholarPubMed
White, IR, Royston, P and Wood, AM (2011) Multiple imputation using chained equations: issues and guidance for practice. Statistical Medicine 30, 377399.CrossRefGoogle ScholarPubMed
Figure 0

Table 1. Baseline characteristics of sample

Figure 1

Table 2. Cross-sectional association between DGIH/BMI and depressive symptoms (age 18)

Figure 2

Table 3. Cross-sectional associations between DGIH/BMI and depressive episode (age 18 years)

Figure 3

Table 4. Longitudinal association between DGIH/BMI (age 9) and depressive symptoms (age 18 years)

Figure 4

Table 5. Longitudinal association between DGIH/BMI (age 9) and depressive episode (age 18 years)

Supplementary material: File

Perry et al. supplementary material

Perry et al. supplementary material 1

Download Perry et al. supplementary material(File)
File 21.1 KB