While high body mass index is associated with an increased risk of depression and anxiety, cumulative evidence indicates that it is a protective factor for suicide. The associations from conventional observational studies of body mass index with mental health outcomes are likely to be influenced by reverse causality or confounding by ill-health. In the present study, we investigated the associations between offspring body mass index and parental anxiety, depression and suicide in order to avoid problems with reverse causality and confounding by ill-health.
We used data from 32,457 mother-offspring and 27,753 father-offspring pairs from the Norwegian HUNT-study. Anxiety and depression were assessed using the Hospital Anxiety and Depression Scale and suicide death from national registers. Associations between offspring and own body mass index and symptoms of anxiety and depression and suicide mortality were estimated using logistic and Cox regression. Causal effect estimates were estimated with a two sample instrument variable approach using offspring body mass index as an instrument for parental body mass index.
Both own and offspring body mass index were positively associated with depression, while the results did not indicate any substantial association between body mass index and anxiety. Although precision was low, suicide mortality was inversely associated with own body mass index and the results from the analysis using offspring body mass index supported these results. Adjusted odds ratios per standard deviation body mass index from the instrumental variable analysis were 1.22 (95% CI: 1.05, 1.43) for depression, 1.10 (95% CI: 0.95, 1.27) for anxiety, and the instrumental variable estimated hazard ratios for suicide was 0.69 (95% CI: 0.30, 1.63).
Citation: Bjørngaard JH, Carslake D, Lund Nilsen TI, Linthorst ACE, Davey Smith G, Gunnell D, et al. (2015) Association of Body Mass Index with Depression, Anxiety and Suicide—An Instrumental Variable Analysis of the HUNT Study. PLoS ONE 10(7): e0131708. https://doi.org/10.1371/journal.pone.0131708
Editor: Andreas Stengel, Charité-Universitätsmedizin Berlin, Campus Benjamin Franklin, GERMANY
Received: February 16, 2015; Accepted: June 4, 2015; Published: July 13, 2015
Copyright: © 2015 Bjørngaard et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: Data are available from HUNT Research Centre (hunt.no) and Statistics Norway (ssb.no), dependent on approval from the Ethics Committee, for researchers who meet the criteria for access to confidential data.
Funding: David Carslake works in a unit that receives funds from the UK Medical Research Council and the University of Bristol. David Gunnell is a National Institute for Health Research Senior Investigator. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors state that there are no conflicts of interest.
Competing interests: The authors have declared that no competing interests exist.
Obesity is associated with an increased risk of depression as well as other mental illnesses[1–3]. In contrast, prospective studies from US and Scandinavia [1, 5] have indicated that suicide risk declines as body mass index increases. This is a paradoxical result since depression and other mental illnesses are the strongest risk factors for suicide. Prospective evidence of the link between obesity and anxiety disorders is, however, scarce and mixed. Disentangling the causal nature of these associations is important as obesity, suicide and common mental illness such as anxiety and depression constitute major public health challenges.
The associations from observational studies of body mass index with mental health outcomes are likely to be influenced by reverse causality or confounding by ill-health. The mechanisms linking body mass index and depression are likely to be convoluted, as several studies have indicated that each may induce or worsen the other. High body mass index could influence the risk of depression via biological mechanisms such as inflammation, dysregulation of hormonal stress systems and the risk of developing physical diseases that may have secondary effects on mental health–while being obese could also have negative psychological effects on, for instance, self-image. Depression could also influence the risk of high body mass index via unhealthy lifestyles as mental illness is associated with physical inactivity and unhealthy diets [10–13]. Likewise, medicines prescribed for some mental illnesses might lead to weight gain. The association between obesity and depression could also be confounded by common causes of both being obese and mental illness such as poor socioeconomic position.
The causal mechanisms underlying the association between body weight and suicide are not well understood. Different possible explanations have been suggested. High body mass index might influence suicidal risk directly–for instance through decreased impulsive aggressive behaviours possibly due to the influence of body mass index on several hormones and neurotransmitters like testosterone, leptin, and serotonin levels [6, 16–18]. However, associations could also be influenced by pre-existing substance misuse, severe somatic or mental illness that are associated with low body mass index [6, 18].
The use of instrumental variables has previously proven to be a fruitful way to assess the causal nature of potentially confounded associations, or associations driven by reverse causality. In particular, genetic variants (single nucleotide polymorphisms, SNPs) associated with known functional phenotypes have been used in Mendelian randomization studies to assess causality[20, 21]. The results from such studies on obesity and mental health are however, scarce and mixed. For example, results from a UK study using the polymorphism of the FTO gene as an instrumental variable for obesity, provided evidence that obesity is a risk factor for the development of common mental disorders. On the other hand, results from a Danish study using the polymorphisms of the FTO and MC4R genes, and from a collaborative Canadian study using the FTO gene polymorphism as instruments for obesity, suggested a possible protective effect of obesity on psychological distress and depression. Finally, the results from a recent cross national study using different polymorphisms as instruments for increased body mass index did not support a causal relationship between obesity and major depression . However, the use of genetic polymorphisms in such studies is often restricted by lack of precision due to weak associations between the variants and the phenotypic traits. This is particularly a problem when studying suicide, since suicide is relatively rare in most populations.
In the present study we use an offspring’s body mass index as an instrument for their parent’s body mass index. This has previously been used as an instrument in studies of cause specific mortality and medical costs. Several twin and adoption studies have found a strong genetic component in both adult and childhood body mass index, with heritability estimates ranging from 0.6 to 0.9 in different studies, and that this association is mainly due to common genes rather than shared family environment [29–32]. Further, offspring body mass index is not likely to be a consequence of parents’ mental illness or suicide risk, and hence, reverse causality should be avoided.
In the present study we have assessed the association of body mass index with symptoms of anxiety and depression and suicide death in a large population study, using an offspring’s body mass index as an instrumental variable for own body mass index.
Study population and data linkage
The present study’s results are based partly on a re-analysis of previously published data . The HUNT study is a large population-based health study conducted in Nord-Trøndelag, a rural county in central Norway with a population of about 130,000. At each of three waves (1984–1986, 1995–1997 and 2006–2008), every resident aged 20 years or over was invited to participate. In addition, 13 to 19 year olds were recruited in three waves (the “YoungHUNT” study). The first and third of these coincided with HUNT2 and HUNT3, respectively and the second took place in 2000–2001. As the whole community was invited to participate in the study, many families were represented in the dataset by members from two or more generations. Full details of the study are available on the HUNT website (http://www.ntnu.no/hunt). Briefly, participants in each wave attended a clinic where their height and weight were measured. Body mass index (kg per metre squared) was calculated from these measures. Participants also completed questionnaires, which in HUNT2 and HUNT3 included the Hospital Anxiety and Depression Scale (HADS) .
The questionnaire also provided information on subjects' smoking, alcohol use, exercise, education, employment and marital status. Statistics Norway's Death Registry was used to identify all deaths among HUNT participants between the beginning of HUNT1 (1st January 1984) and 31st December 2009. The international classification of diseases (ICD) code accompanying each death was used to identify deaths by suicide (codes E950-959 and E980-E989 in ICD-9; and codes X600-X849, Y100-Y349, Y870 and Y872 in ICD-10). The present study was approved by the Regional Committee for Medical Research Ethics central Norway—2011/1455/REK midt. Each participant and the parents/legal guardians of the participants younger than 16 years old gave their written consent to participate.
We initially extracted data for study members who had at least one participating parent (66,262 offspring, with 62,414 participating mothers and 53,145 participating fathers). Individuals were removed from each analysis if their body mass index data, or the body mass index data of both parents, were missing, and the data were restricted to one randomly chosen offspring per parent. This left 32,457 mother-offspring and 27,753 father-offspring pairs for the main analysis of suicide mortality. The analyses of depression and anxiety additionally required the appropriate parental HADS data, recorded later than the parent's body mass index. This left 21,977 mother-offspring pairs and 17,260 father-offspring pairs in the analysis of depression and 21,659 mother-offspring pairs and 17,153 father-offspring pairs in the analysis of anxiety. The data selection procedure is shown in Fig 1. For those individuals who participated in more than one wave of the study, we used body mass index data from their earliest HUNT (i.e. adult) participation. This minimised the risk of weight loss due to age-related illness, maximised the period at risk for the suicide analysis and maximised the chances of parents' HADS data post-dating their own body mass index data. For offspring only, we used the latest available YoungHUNT body mass index data if no HUNT body mass index data were available. For the analyses of depression and anxiety, we used HADS data from the earliest HUNT wave post-dating the subject's own body mass index measurement. Subjects without HADS data from a later wave than their own body mass index data were not included, but we did not require HADS data to post-date offspring body mass index data (since there is no assertion of causality in this association).
Numbers are listed as mother-offspring pairs / father-offspring pairs.
The HADS data consist of seven depression-related and seven anxiety-related questions, each answered on a scale of 0 to 3, giving a total symptom score of 0–21. A value of ≥8 was set to define a case of depression or anxiety . For each outcome, participants were excluded from analysis if the answers to four or more of the seven questions were missing. If 1–3 answers were missing, cases were defined from the non-missing data, which were weighted according to each question's mean contribution to the total score.
Each measure of body mass index was assigned to a quintile among participants of the same sex and similar age, measured in the same HUNT wave. Within these quintiles, we summarized offspring and parents' health-related characteristics and characterised their association with body mass index using linear or logistic regression, as appropriate.
Before all further analyses, body mass index in all participants was adjusted for age, sex and year of measurement by taking residuals from sex-specific full factorial regression models against HUNT wave (categorical, with contemporary HUNT and YoungHUNT combined) and a cubic spline of age. These residuals were then divided by the residual standard deviation of the model to give sex-specific Z-scores of body mass index. Cubic splines used five knots, because the size of the data set permitted a reasonably high precision in the fitting of the relationship. The knots were placed at percentiles of 5, 27.5, 50, 72.5, 95, following the recommendations of Harrell et al. .
Statistical analysis of depression and anxiety
Symptoms of depression and anxiety were analysed separately, using logistic regression models to estimate the odds ratio (OR) for depression/anxiety per sex-specific standard deviation of (i) a subject's own body mass index, and (ii) their offspring's body mass index. All models were adjusted for parental date of birth and age using cubic splines with 3 knots at percentiles of 10, 50, 90. Three knots were chosen to avoid over-adjusting in the analysis of relatively rare binary outcomes, and their positions were those recommended by Harrell et al. . Models were run with and without adjustment for parental education, coded as missing (14% of mothers, 18% of fathers), <10, 10–12, or >12 years and recorded at the same time as the parent's body mass index, as a representation of socioeconomic position. The linearity of associations was assessed by plotting OR by quartiles of adjusted body mass index. Models were run separately for mothers and fathers, but also for both parents together. These combined models were additionally adjusted for parental sex, and used robust standard errors clustered by offspring identity to account for the non-independence of an offspring's two parents. The difference between maternal and paternal OR was tested by adding an interaction between parental sex and the exposure (i.e. own or offspring body mass index) and examining the p-value from a Z-test of the resulting coefficient.
Two-sample instrumental variable IV analysis was used to estimate parental OR for depression/anxiety per sex-specific standard deviation of parental body mass index, with offspring body mass index as an instrument. This method avoids some of the sources of confounding/reverse causality between parental depression/anxiety and body mass index [35, 36]. First, parental body mass index (mothers, fathers and combined parents in turn) was regressed on offspring body mass index, with adjustment as described above for the corresponding logistic regression. Two-sample IV estimates of the OR were made by dividing the natural logarithm of the OR per SD of offspring body mass index (with or without adjustment for education) by the comparably adjusted regression coefficient for parental body mass index against offspring body mass index, and exponentiating. Confidence intervals were calculated using Taylor series expansions . OR calculated by two-sample IV were compared with conventional OR per SD of the parent's own body mass index using 1,000 bootstrap resamples.
Statistical analysis of suicide mortality
Suicide mortality was analysed similarly to depression and anxiety, but with the use of Cox regression in place of logistic regression. Parental age was used as the time axis, so adjustment for parental age was not necessary. Observations were censored to restrict them to the period when a parent's death would have been recorded and used in the study. Left-censoring was thus applied at the latest of the offspring's date of birth and the parent's body mass index measurement, and right-censoring applied at the earliest of the parent's date of death or emigration, or 31st December 2009 (the end of follow-up). We did not left-censor on the date of the offspring's body mass index measurement, because offspring body mass index was used as an instrument for parental body mass index during life, regardless of whether the parent was alive when the offspring's body mass index was measured.
The health-related characteristics of parents and their offspring by quintiles of offspring body mass index are presented in Table 1. The parents of offspring with higher body mass index, and the offspring themselves, had higher blood pressure and were less educated and less physically active. The parents of high-body mass index offspring had higher body mass index, were slightly younger at the time of the child's birth, and were also more likely to smoke. Offspring in the lowest quintile of body mass index, and their fathers, were slightly more likely to drink frequently, but this pattern was not repeated in mothers, and there was no clear trend over the other four quintiles. The standard deviation of body mass index was 3.4 kg m-2 for women and 4.3 kg m-2 for men.
For the analysis of depression, there were 2,640 maternal and 2,159 paternal cases of depression among the sample of 21,977 mothers and 17,260 fathers. For the analysis of anxiety, there were 3,902 maternal and 1,899 paternal cases of anxiety among the sample of 21,659 mothers and 17,153 fathers. For the analysis of suicide, there were 34 maternal and 101 paternal suicides among the sample of 32,457 mothers and 27,761 fathers.
Association between offspring and parental body mass index
Regression coefficients for parental body mass index against offspring body mass index, used in the calculation of two-sample IV estimates, are shown in Table 2. Values were calculated separately for each outcome because slightly different subsets of the data were used. In the data subset for the analysis of depression, for each standard deviation increase in combined parental body mass index, there was a 0.21 (95% CI: 0.20, 0.22) fully adjusted standard deviation increase in the offspring body mass index. Similar values were obtained regardless of the parent in question, the adjustment used, and the data subset.
Body mass index and depression symptoms
Mothers and fathers showed similar positive associations between body mass index and depression (Table 3, Fig 2). Full adjustment attenuated the odds ratios a little for own body mass index and for offspring body mass index. The wide confidence intervals made it difficult to infer the exact shape of the association, but there was no evidence of non-linearity (Fig 2). OR from two-sample IV analyses were rather higher than those from conventional analyses using subjects' own body mass index, but with wide confidence intervals, and bootstrap comparisons provided only some evidence for a difference in odds ratios between the analyses using a subject's own body mass index and the IV analyses (with minimal adjustment: p = 0.06 for mothers and p = 0.09 for fathers and with full adjustment: p = 0.13 for mothers and p = 0.27 for fathers).
Body mass index was pre-adjusted for age and HUNT wave and converted to sex-specific Z scores before classification into quartiles. It is plotted as the mean body mass index for each quartile, converted back to kg m-2 for men (upper X-axis) and women (lower X-axis). All models were adjusted for parental age, date of birth and education. Note the different scale on the y-axis for suicide.
Body mass index and Anxiety symptoms
There was no evidence of an association between anxiety and a person's own body mass index (Table 4, Fig 2). When an offspring's body mass index was used as an instrument for their parents' body mass index, there was some evidence of a positive association, but this was somewhat attenuated by full adjustment, and 95% confidence intervals before and after adjustment included the null. We found only weak evidence for a difference in odds ratios between the analyses using a subject's own body mass index and the IV analyses with minimal adjustment (bootstrap comparisons; p = 0.09 for mothers and p = 0.14 for fathers),or with full adjustment (bootstrap comparisons; p = 0.21 for mothers and p = 0.25 for fathers). Odds ratios were similar for mothers and fathers.
Body mass index and suicide
Suicide mortality was inversely associated with body mass index, whether a subject's own body mass index, their offspring's, or an instrumental variables approach was used (Table 5, Fig 2). However, the wide confidence intervals around these hazard ratios, particularly for women, mean that this result should be interpreted cautiously. We found no statistical evidence that hazard ratios for suicide mortality differed between mothers and fathers (Table 4), or that hazard ratios calculated conventionally from own body mass index differed from those calculated by IV (fully adjusted analysis of 1000 bootstrap resamples comparing IV hazard ratios with conventional ones; p = 0.80 in mothers and 0.85 in fathers). Adjustment did not substantially alter the hazard ratios.
In this large population study, we studied the associations between body mass index and depression, anxiety and suicide, using an offspring’s body mass index as an instrumental variable for parental body mass index. Our results indicated that body mass index and depression were positively associated, while the results did not indicate any substantial association between body mass index and anxiety. The results indicated that suicide mortality was inversely associated with body mass index, but the wide confidence intervals around the estimates made by instrumental variable methods suggest that this result should be interpreted with care and leaves chance variation as an alternative explanation. These findings are in keeping with our earlier analysis of HUNT participants' own body mass index in relation to depression, anxiety and suicide. Effect estimates calculated by two-sample instrumental variable methods, particularly those for depression, were further from the null than the equivalent figures calculated by conventional methods using the parent's own body mass index. This suggests that conventional analyses might underestimate the strength of the causal effect of body mass index on the outcomes, due to some form of confounding associating them in the opposite direction although chance variation could also be an alternative explanation.
Strengths and limitations
Our study was based on data from a large population based study, with clinical measures of the weight and height of both parents and their children. Despite the large sample size, the low incidence of suicide (135 suicides over the follow-up period) left us with limited power to investigate associations of offspring body mass index with suicide risk.
Information on suicide was collected from the National Cause of Death Registry where the positive predictive value of suicide is likely to be high, although suicide might be underreported. Possible underreporting of suicide is however, not likely to bias our results given the prospective approach of the study. We used validated assessments of anxiety and depression, and although not feasible in a study of this size, a structured psychiatric diagnostic interview would have given more reliable diagnostic information than the fourteen-item HADS questionnaire. Particularly, our ability to differentiate between anxiety and depression symptoms could be limited given the considerable overlap between symptoms of anxiety and depression. Given the objective measurement of body mass index and the instrumental variable approach, the risk of any differential misclassification should be avoided. However, in the case of non-differential misclassification of anxiety and depression symptoms, we would expect an attenuation of the associations. Hence, the lack of an association between body mass index and anxiety symptoms could also be influenced by measurement error.
Our study included only families with information from mother-offspring, father-offspring or both parents-offspring pairs. Hence, we cannot rule out possible selection effects from excluded participants without participating parents or offspring.
Instrumental variable assumptions
By using an instrument variable approach, we could arguably provide better estimates of the body mass index-depression,-anxiety and-suicide associations than we could have done using parental body mass index only. Nevertheless, instrumental variable analyses are based on the strength of the association between the instrument and the exposure, and on several key assumptions. There was a strong association between offspring and parental body mass index in our data, but the essential assumption that the instrument should only influence the outcome through the exposure of interest is difficult to evaluate . There is evidence to suggest that the association between parental and offspring body mass index can be largely attributed to common genes and not shared environment [29–32]. The same pattern has been suggested for mental illnesses for which shared family environment seems to play a minor etiologic role in the development of depression. Hence, any association between offspring and parental mental illness would not likely be caused by common environmental features of the family. Furthermore, high offspring body mass index is not likely to be caused by parental ill health or parental poor mental health. Hence, the typical problem with reverse causality and confounding by ill-health in observational studies of consequences of obesity should be avoided in our study using a first degree relative’s body mass index as an instrument for own body mass index. Nevertheless, offspring body mass index might to some extent be related to shared confounding at the family level, and in the present study we had information from a range of possible confounding variables that could have influenced the observed associations in our analyses.
Comparison with other studies
To our knowledge, this is the first study to use offspring body mass index as an instrument for parental body mass index to study the association of body mass index with depression, anxiety and suicide. The positive association between offspring body mass index and parental depression is in line with former evidence linking obesity to depression[1, 9, 22], although results in the opposite direction have been reported[23, 24]. A recent review indicated positive cross sectional associations between obesity and anxiety disorders and symptoms, while prospective evidence for an effect of obesity on anxiety disorders is scarce and mixed. We did not find substantial evidence of an association between body mass index and anxiety symptoms–a result that might indicate that body mass index is not a cause of anxiety. However, since there is a considerable overlap between symptoms of anxiety and depression, one should be cautious interpreting separate effects of either. Our results give further support to a protective link between high body mass index and suicide risk, a result which has been reported in several other studies, although not in a recent study from Taiwan  and a study from UK. Further, although there is some evidence that symptoms of depression decrease after bariatric surgery, risk of suicide may increase –but the latter observation is based on a relatively small number of events.
The paradoxical finding that high body mass index is associated with an increased risk of depression but seems to be a protective factor for suicide, was not altered using our analytical approach, although the precision was low for the body mass index-suicide association. Several possible mechanisms have been suggested for the association between obesity and suicide[44–47]. For instance high body mass index might decrease impulsive suicide attempts via its association with several hormones and neurotransmitters [6, 16–18]. However, a Swedish study of attempted suicide found limited evidence that associations differed according to method of self-harm. Confounding by pre-existing substance misuse or severe somatic or mental illness associated with low body mass index is another possible explanation [6, 18]. A large population-based US study investigated the associations between body mass index and several potential risk factors for suicide, but did not find that these risk factors were consistently inversely associated with greater body mass index. Finally, most studies have investigated associations between adult body mass index and suicide risk. Hence, at the start of follow up, mental health status and also treatment might themselves have influenced body weight. A major advantage of the present study is that reverse causality and confounding by ill-health should be avoided since high offspring body mass index is not likely to be caused by poor parental mental health. Hence, although we do not have information on the possible mediating factors, our study indicates that body mass index influences suicidal risk via other causal pathways than mental illness.
Although depression, anxiety and suicide are associated with body mass index and share common risk factors, their underlying neurobiology is likely to be different. This notion is underscored by the diversity of psychiatric illnesses which may culminate in suicide and include depression, bipolar disorder, schizophrenia, PTSD and substance abuse. A large number of studies have shown an important role for serotonin  and glucocorticoid hormones in the aetiology of depression. There are further reports suggesting that these systems are malfunctioning also in suicidal behaviour and suicide completion. In this respect it is of interest to note that both body weight control and food intake, and consequently body mass index, are regulated by serotonin  and glucocorticoids  at the level of the brain. Low levels of brain serotonin result in hyperphagia and an increase in food intake. It is therefore of interest that body mass index seems to be negatively associated with the serotonin transporter binding (SERT) in several parts of the brain including the midbrain. Thus a low body mass index is associated with higher SERT levels in the brain which may result in decreased extracellular levels of serotonin, which could be interpreted as a compensatory drive to stimulate food intake.
Alterations in glucocorticoid hormone physiology in individuals with symptoms of depression and/or anxiety and in suicide victims may involve various levels of the hypothalamic-pituitary-adrenal (HPA) axis. Elevated plasma concentrations of cortisol and abnormal neuroendocrine challenge tests point to a hyperactive HPA axis in depressed patients. Cortisol non-suppression in the Dexamethasone Suppression Test (DST) has been linked to an increased risk of suicide attempts and suicide completion. However, whether changes in HPA axis activity are involved in the negative relationship between body mass index and suicide is presently difficult to confirm.
Currently it is impossible to draw a conclusive and comprehensive picture of the neurobiological mechanisms linking body mass index with depression and suicide completion. However, it is most likely that the exact make-up of changes in serotonergic neurotransmission and glucocorticoid hormone physiology, together with changes in other systems such as GABAergic neurotransmission and also leptin, will determine both body mass index and psychopathology, including affective disturbances and suicide.
Separately and together, anxiety, depression, suicide and obesity constitute substantial challenges on the global burden of disease. Our results give support to a positive association between body mass index and depression, but not anxiety–while the risk of suicide on the other hand, seems to be reduced with increased body mass index. As obesity is an increasingly prevalent disease[56, 57], clinicians should take into account possible depression symptoms in clinical settings with obese patients. Furthermore, the present study’s results indicate that the body mass index–suicide association is not mediated by depression. This calls for further studies investigating the biologic mechanisms underlying suicide. Furthermore, the underlying mechanisms in the associations between obesity and risk of mental illness and suicide are complex and intricate, involving different bio-psycho-social causal pathways. Improved understanding of these causal mechanisms could lead to better personalized medicine for individuals with obesity as well as mental problems like depression.
Nord-Trøndelag Health Study (The HUNT Study) is collaboration between HUNT Research Centre (Faculty of Medicine, Norwegian University of Science and Technology NTNU), Nord-Trøndelag County Council and the Norwegian Institute of Public Health.
David Carslake works in a unit that receives funds from the UK Medical Research Council and the University of Bristol. David Gunnell is a National Institute for Health Research Senior Investigator.
Conceived and designed the experiments: JHB GDS DG PRR. Analyzed the data: DC. Wrote the paper: JHB DC TILN ACEL GDS DG PRR.
- 1. Bjerkeset O, Romundstad P, Evans J, Gunnell D. Association of adult body mass index and height with anxiety, depression, and suicide in the general population: the HUNT study. Am J Epidemiol. 2008;167(2):193–202. Epub 2007/11/06. doi: kwm280 [pii] pmid:17981889.
- 2. Kivimäki M, Batty GD, Singh-Manoux A, Nabi H, Sabia S, Tabak AG, et al. Association between common mental disorder and obesity over the adult life course. The British Journal of Psychiatry. 2009;195(2):149–55. pmid:19648547
- 3. Simon G, von Korff M, Saunders K, Miglioretti D, Crane P, van Belle G, et al. Association between obesity and psychiatric disorders in the us adult population. Archives of General Psychiatry. 2006;63(7):824–30. pmid:16818872
- 4. Mukamal KJ, Rimm EB, Kawachi I, O'Reilly EJ, Calle EE, Miller M. Body Mass Index and Risk of Suicide Among One Million US Adults. Epidemiology. 2010;21(1):82–6 pmid:19907331
- 5. Batty GD, Whitley E, Kivimäki M, Tynelius P, Rasmussen F. Body Mass Index and Attempted Suicide: Cohort Study of 1,133,019 Swedish Men. American Journal of Epidemiology. 2010;172(8):890–9. pmid:20829269
- 6. Klinitzke G, Steinig J, Blüher M, Kersting A, Wagner B. Obesity and suicide risk in adults—A systematic review. Journal of Affective Disorders. 2013;145(3):277–84. pmid:22871535
- 7. Gunnell D, Lewis G. Studying suicide from the life course perspective: implications for prevention. Br J Psychiatry. 2005;187:206–8. Epub 2005/09/02. doi: 187/3/206 [pii] pmid:16135856.
- 8. Gariepy G, Nitka D, Schmitz N. The association between obesity and anxiety disorders in the population: a systematic review and meta-analysis. Int J Obes. 2010;34(3):407–19.
- 9. Luppino FS, de Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx BW, et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry. 2010;67(3):220–9. Epub 2010/03/03. doi: 67/3/220 [pii] pmid:20194822.
- 10. Lasser K, Boyd JW, Woolhandler S, Himmelstein DU, McCormick D, Bor DH. Smoking and Mental Illness. JAMA: The Journal of the American Medical Association. 2000;284(20):2606–10.
- 11. Daumit GL, Goldberg RW, Anthony C, Dickerson F, Brown CH, Kreyenbuhl J, et al. Physical Activity Patterns in Adults With Severe Mental Illness. The Journal of Nervous and Mental Disease. 2005;193(10):641–6. pmid:16208158
- 12. Regier DA, Farmer ME, Rae DS, Locke BZ, Keith SJ, Judd LL, et al. Comorbidity of mental disorders with alcohol and other drug abuse: Results from the epidemiologic catchment area (eca) study. JAMA: The Journal of the American Medical Association. 1990;264(19):2511–8.
- 13. Osborn D, Nazareth I, King M. Physical activity, dietary habits and Coronary Heart Disease risk factor knowledge amongst people with severe mental illness. Social Psychiatry and Psychiatric Epidemiology. 2007;42(10):787–93. pmid:17721669
- 14. Dent R, Blackmore A, Peterson J, Habib R, Kay GP, Gervais A, et al. Changes in Body Weight and Psychotropic Drugs: A Systematic Synthesis of the Literature. PLoS One. 2012;7(6):e36889. pmid:22719834
- 15. Moore ME, Stunkard A, Srole L. Obesity, Social Class, and Mental Illness. Obesity Research. 1997;5(5):503–8. pmid:9385629
- 16. Laaksonen DE, Niskanen L, Punnonen K, Nyyssonen K, Tuomainen TP, Valkonen VP, et al. Testosterone and sex hormone-binding globulin predict the metabolic syndrome and diabetes in middle-aged men. Diabetes care. 2004;27(5):1036–41. Epub 2004/04/28. pmid:15111517.
- 17. Pelusi C, Pasquali R. The Significance of Low Testosterone Levels in Obese Men. Curr Obes Rep. 2012;1(4):181–90.
Gunnell D, Thomas K. Body weight, obesity and suicide across the life course. In: O'Connor R, Platt S, Gordon J, editors. International handbook of suicide prevention: research, policy and practice. Chichester: Wiley-Blackwell; 2011.
- 19. Greenland S. An introduction to instrumental variables for epidemiologists. International Journal of Epidemiology. 2000;29(4):722–9. pmid:10922351
- 20. Davey Smith G, Ebrahim S. Mendelian randomization: prospects, potentials, and limitations. International Journal of Epidemiology. 2004;33(1):30–42. pmid:15075143
- 21. Davey Smith G, Lawlor DA, Harbord R, Timpson N, Day I, Ebrahim S. Clustered Environments and Randomized Genes: A Fundamental Distinction between Conventional and Genetic Epidemiology. PLoS Medicine. 2007;4(12):e352. pmid:18076282
- 22. Kivimaki M, Jokela M, Hamer M, Geddes J, Ebmeier K, Kumari M, et al. Examining overweight and obesity as risk factors for common mental disorders using fat mass and obesity-associated (FTO) genotype-instrumented analysis: The Whitehall II Study, 1985–2004. Am J Epidemiol. 2011;173(4):421–9. Epub 2011/01/21. doi: kwq444 [pii] pmid:21248310; PubMed Central PMCID: PMC3032807.
- 23. Lawlor DA, Harbord RM, Tybjaerg-Hansen A, Palmer TM, Zacho J, Benn M, et al. Using genetic loci to understand the relationship between adiposity and psychological distress: a Mendelian Randomization study in the Copenhagen General Population Study of 53 221 adults. Journal of Internal Medicine. 2011;269(5):525–37. pmid:21210875
Samaan Z, Anand S, Zhang X, Desai D, Rivera M, Pare G, et al. The protective effect of the obesity-associated rs9939609 A variant in fat mass- and obesity-associated gene on depression. Mol Psychiatry. 2012. doi: https://doi.org/http://www.nature.com/mp/journal/vaop/ncurrent/suppinfo/mp2012160s1.html.
Hung C-F, Rivera M, Craddock N, Owen MJ, Gill M, Korszun A, et al. Relationship between obesity and the risk of clinically significant depression: Mendelian randomisation study2014 2014-07-01 00:00:00. 24–8 p.
- 26. Nock MK, Borges G, Bromet EJ, Cha CB, Kessler RC, Lee S. Suicide and Suicidal Behavior. Epidemiologic Reviews. 2008;30(1):133–54.
- 27. Davey Smith G, Sterne JA, Fraser A, Tynelius P, Lawlor DA, Rasmussen F. The association between BMI and mortality using offspring BMI as an indicator of own BMI: large intergenerational mortality study. BMJ. 2009;339:b5043. Epub 2009/12/24. bmj.b5043 [pii]. pmid:20028778; PubMed Central PMCID: PMC2797052.
- 28. Cawley J, Meyerhoefer C. The medical care costs of obesity: an instrumental variables approach. J Health Econ. 2012;31(1):219–30. Epub 2011/11/19. doi: S0167-6296(11)00136-6 [pii] pmid:22094013.
- 29. O'Rahilly S, Farooqi IS. Genetics of obesity. Philosophical Transactions of the Royal Society B: Biological Sciences. 2006;361(1471):1095–105.
- 30. Silventoinen K, Kaprio J. Genetics of tracking of body mass index from birth to late middle age: evidence from twin and family studies. Obesity facts. 2009;2(3):196–202. pmid:20054225
- 31. Silventoinen K, Rokholm B, Kaprio J, Sorensen TIA. The genetic and environmental influences on childhood obesity: a systematic review of twin and adoption studies. Int J Obes. 2009;34(1):29–40.
- 32. Stunkard AJ, Sørensen TIA, Hanis C, Teasdale TW, Chakraborty R, Schull WJ, et al. An Adoption Study of Human Obesity. New England Journal of Medicine. 1986;314(4):193–8. pmid:3941707
- 33. Bjelland I, Dahl AA, Haug TT, Neckelmann D. The validity of the Hospital Anxiety and Depression Scale: An updated literature review. Journal of Psychosomatic Research. 2002;52(2):69–77. pmid:11832252
Harrell FE. Regression modeling strategies: with applications to linear models, logistic regression, and survival analysis. New York: Springer; 2001. XXII, 568 s., ill. p.
- 35. Angrist JD, Krueger AB. The effect of age at school entry on educational attainment: an application of instrumental variables with moments from two samples. Journal of the American Statistical Association. 1992;87(418):328–36. pmid:WOS:A1992HW62200008.
- 36. Didelez V, Meng S, Sheehan NA. Assumptions of IV methods for observational epidemiology. Statistical Science. 2010;25(1):22–40. pmid:WOS:000281096800002.
- 37. Thomas DC, Lawlor DA, Thompson JR. Re: Estimation of bias in nongenetic observational studies using "Mendelian triangulation" by Bautista et al. Ann Epidemiol. 2007;17(7):511–3. pmid:WOS:000247888900004.
- 38. Tyrer P. From The Editor's Desk. The British Journal of Psychiatry. 2008;192(3):242. pmid:18399019
- 39. Kendler KS. Twin Studies of Psychiatric Illness: An Update. Arch Gen Psychiatry. 2001;58(11):1005–14. pmid:11695946
- 40. Chang S-S, Wen CP, Tsai MK, Lawlor DA, Yang YC, Gunnell D. Adiposity, Its Related Biologic Risk Factors, and Suicide: A Cohort Study of 542,088 Taiwanese Adults. American Journal of Epidemiology. 2012;175(8):804–15. pmid:22427611
- 41. Elovainio M, Shipley MJ, Ferrie JE, Gimeno D, Vahtera J, Marmot MG, et al. Obesity, unexplained weight loss and suicide: The original Whitehall study. Journal of Affective Disorders. 2009;116(3):218–21. pmid:19097646
- 42. Herpertz S, Kielmann R, Wolf AM, Langkafel M, Senf W, Hebebrand J. Does obesity surgery improve psychosocial functioning? A systematic review. International Journal of Obesity. 2003;27(11):1300–14. pmid:WOS:000186111900004.
- 43. Peterhänsel C, Petroff D, Klinitzke G, Kersting A, Wagner B. Risk of completed suicide after bariatric surgery: a systematic review. Obesity Reviews. 2013;14(5):369–82. pmid:23297762
Gunnell D, Thomas K. Body weight, obesity, and suicide across life course. In: O'Connor RC, Platt S, Gordon J, editors. International handbook of suicide prevention: John Wiley & Sons Ltd.; 2011.
- 45. Mukamal KJ, Miller M. BMI and Risk Factors for Suicide: Why Is BMI Inversely Related to Suicide? Obesity. 2009;17(3):532–8. pmid:19219065
Mukamal KJ, Miller M. Invited Commentary: Body Mass Index and Suicide—Untangling an Unlikely Association. American Journal of Epidemiology. 2010. https://doi.org/10.1093/aje/kwq278
- 47. Zhang J, Yan F, Li Y, McKeown RE. Body mass index and suicidal behaviors: A critical review of epidemiological evidence. Journal of Affective Disorders. 2013;148(2–3):147–60. pmid:22999892
- 48. Ressler KJ, Nemeroff CB. Role of serotonergic and noradrenergic systems in the pathophysiology of depression and anxiety disorders. Depression and anxiety. 2000;12 Suppl 1:2–19. pmid:11098410
- 49. Steckler T, Holsboer F, Reul JMHM. Glucocorticoids and depression. Best Practice & Research Clinical Endocrinology & Metabolism. 1999;13(4):597–614.
Costanza A, D’Orta I, Perroud N, Burkhardt S, Malafosse A, Mangin P, et al. Neurobiology of suicide: do biomarkers exist? Int J Legal Med. 2013:1–10. https://doi.org/10.1007/s00414-013-0835-6
- 51. Lam DD, Garfield AS, Marston OJ, Shaw J, Heisler LK. Brain serotonin system in the coordination of food intake and body weight. Pharmacology Biochemistry and Behavior. 2010;97(1):84–91.
- 52. Spencer S, Tilbrook A. The glucocorticoid contribution to obesity. Stress. 2011;14(3):233–46. pmid:21294656.
- 53. Erritzoe D, Frokjaer VG, Haahr MT, Kalbitzer J, Svarer C, Holst KK, et al. Cerebral serotonin transporter binding is inversely related to body mass index. NeuroImage. 2010;52(1):284–9. pmid:20382236
- 54. Pompili M, Serafini G, Innamorati M, Möller-Leimkühler A, Giupponi G, Girardi P, et al. The hypothalamic-pituitary-adrenal axis and serotonin abnormalities: a selective overview for the implications of suicide prevention. Eur Arch Psychiatry Clin Neurosci. 2010;260(8):583–600. pmid:20174927
- 55. Yamada N, Katsuura G, Ochi Y, Ebihara K, Kusakabe T, Hosoda K, et al. Impaired CNS Leptin Action Is Implicated in Depression Associated with Obesity. Endocrinology. 2011;152(7):2634–43. pmid:21521746
- 56. Haslam DW, James WPT. Obesity. The Lancet. 2006;366(9492):1197–209.
- 57. Parikh NI, Pencina MJ, Wang TJ, Lanier KJ, Fox CS, D’Agostino RB, et al. Increasing Trends in Incidence of Overweight and Obesity over 5 Decades. The American Journal of Medicine. 2007;120(3):242–50.e2. pmid:17349447