Common mental disorder and obesity: insight from four repeat measures over 19 years: prospective Whitehall II cohort study
BMJ 2009; 339 doi: https://doi.org/10.1136/bmj.b3765 (Published 06 October 2009) Cite this as: BMJ 2009;339:b3765All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Dear Editor,
We are pleased by the interest in our paper on common mental
disorders and obesity (1) and thank Drs Ramlackhansingh and Pegorie for
their comments on the importance and potential policy relevance of our
findings. Our response to their request for a more detailed explanation of
the numbers of participants in the analyses follows.
To make comparisons between different models possible, we used data
from participants with no missing data for any of the covariates (so
called "complete case" analysis) when we examined the effect of
multivariable adjustments on the association between common mental
disorders and obesity. If maximum number of participants was used in each
model (instead of an analytic sample based on only study members with no
missing data) then models with more extensive adjustments would have been
based on a smaller study population than models, with no or only a small
number of covariables. Any differences in estimates between the models
could therefore arise from two sources: first, due to the adjustments and,
second, because of different people included in different models. The
advantage with the 'complete case' approach is that all variation in
effects between the differently adjusted models must be the consequence of
adjustments rather than a variation in the analytic sample size. This
makes it easier to interpret the results.
On the other hand, complete case analysis may be subject to increased
selective attrition, potentially biasing results. To examine this, we
repeated the main analysis in the total population, after imputing any
missing data with multivariate multiple imputations techniques. As
presented in the web-appendix, the main finding was fully replicated in
this whole cohort multivariate imputed analysis, suggesting that there was
no major bias in the results reported in the paper.
The authors of the Editorial that accompanied our paper claim that
obesity is associated with subsequent depressive disorders (contrary to
our findings)(2). Prof. Schmitz and colleagues also emphasise that the
time sequence in our data is not necessarily from common mental disorders
to obesity because obesity occurring after phase 1 could have lead to
accumulation of common mental disorders at phases 1 to 3 and thus underlie
the apparent dose-response association between greater exposure to mental
disorder at phases 1 to 3 and obesity at phase 4.
Whilst observational evidence can never completely prove causality in
either direction, our uniquely prospective study with repeat measurements
on both obesity and common mental disorders provided considerably stronger
evidence for the reverse, that is, the status of common mental disorders
as a risk factor for obesity. This finding is important since previous
studies, including those cited in the editorial, have not had the
advantage of the repeat measurements available in our study and that are
necessary to really unpick the direction of association. Indeed, if the
scenario suggested by Schmitz was true in our cohort, then we would have
expected to observe a robust association between greater exposure to
obesity and subsequent common mental disorder. We did not find such an
association in participants who were initially free from mental disorders
– i.e. in the scenario that provides the most robust test for the
suggested temporal sequence of obesity resulting in common mental
disorders. In contrast, we saw a robust dose-response association from
long-term mental disorders to subsequent obesity, even after excluding
those who were overweight or obese at baseline.
We acknowledge that if in reality the risk of common mental disorders
is attributable to morbid forms of obesity (body mass index >=40), then
our opportunity to detect a positive association was limited. This is
because our study cohort, reflecting a relatively healthy working
population, included less than 3% morbidly obese participants. However,
neither the editorial accompanying our paper nor the letter by Schmitz
claim that the association is restricted just to the morbidly obese. We
acknowledge that additional studies with repeat measurements are required
to replicate our findings before we can conclude with certainty the
direction of the association between obesity and common mental disorder.
Furthermore, long-term follow-up of randomised controlled trials of
effective treatments of obesity and common mental health disorder would be
valuable in determining the direction and causal nature of these
associations.
Competing interests: None declared.
References
1 Kivimäki M, Lawlor DA, Singh-Manoux A, Batty GD, Ferrie JE, Shipley
MJ, Nabi H, Sabia S, Marmot MG, Jokela M. Common mental disorder and
obesity: insight from four repeat measures over 19 years: prospective
Whitehall II cohort study. BMJ 2009; 339: b3765.
2 Atlantis E, Goldney RD, Wittert GA. Obesity and depression or
anxiety. BMJ 2009;339:b3868.
Competing interests:
None declared
Competing interests: No competing interests
It is commendable that that the authors have tried to elucidate1:
1. A causative relationship and
2. The direction of the association between common mental health disorders
and obesity.
From a Public Health perspective, if policy is to be changed; one needs as
solid evidence as possible which this paper tried to examine.
There are some issues which should be highlighted:
• This was the most feasible method to address the association of
long term mental health problems and obesity.
• The authors did not state what aspect of ‘common mental health’ was
defined as a case. The GHQ was used but the reader was unsure whether
common mental disorder referred to depression/anxiety or both.
• Further sensitivity analysis using BMI as a continuous variable to rule
out effect of arbitrary cut-off definition for obesity increased the
internal validity.
• There could be more explanation and transparency regarding the numbers
of participants: 4363 participants were included after the final
screening, but adjusted analyses (1, 2 and sensitivity ones) appear to
only have been performed on those participants that had all data available
(not only re: outcome/exposure, but also on potential confounders) n=
4154.
The finding of a significant association between the numbers of times
being a case of mental health disorder and obesity is important. It can be
used to inform local policy, and an opportunity for further research.
References:
1. Kivimaki M, Lawlor DA, Singh-Manoux A, Batty GD, Ferrie JE,
Shipley MJ, et al. Common mental disorder and obesity: insight from four
repeat measures over 19 years: prospective Whitehall II cohort study. BMJ
2009;339(oct06_2):b3765
Competing interests:
None declared
Competing interests: No competing interests
In their cohort study on mental disorder and obesity, Kivimaki et al
[1] reported a direct and cumulative association from common mental
disorders to obesity. To our knowledge, this is the first paper to
investigate the effect of recurrent or chronic common mental health
disorders on the risk of obesity, and also the first to test the effect of
chronic obesity on common mental health outcomes.
However, we suggest caution when interpreting General Health
Questionnaire (GHQ) scores as ‘common mental disorder’. The GHQ is a
general screening instrument designed to assess symptoms of psychological
distress. Because it identifies cases based on severity of symptoms rather
than on diagnosis, it may overestimate prevalence of mental disorders. For
instance, over a quarter (26.7%) of study participants was classified as
currently having a psychiatric disorder at baseline (‘GHQ caseness’ in
table 1). This estimate seems high compared to previous studies based on
standardized psychiatric interviews [2-4]. Patten et al [2] found a point
prevalence of major depressive disorders that barely reached 1.8% in the
nationally representative Canadian Community Health Survey. We worry that
referring to GHQ caseness as ‘common mental disorder’ might be misleading
and think that ‘serious mental distress’ or ‘psychiatric syndrome’ might
be more appropriate terms.
The authors used data from three phases (from 1985/88 to 1997/99) to
define exposure status. In analysis 1 (common mental disorder as a risk
factor for subsequent obesity), participants were classified into four
groups based on the number of times that they were identified as a ‘case’
of common mental disorder. Given the time frame used in the GHQ (past few
weeks) and the length of the ‘baseline’ period, misclassification in
exposure status (chronicity of common mental disorder) is likely. For
example, those classified as non-cases might have been ‘cases’ between
1985 and 1997 but were missed by the survey. Similarly, different levels
of chronicity of obesity could be misclassified. The misclassification in
exposure makes interpretation difficult.
Further, we suspect that the direction of causality is not as easily
interpretable as reported in the paper. Although analyses adjusted for, or
excluded, those with the outcome of interest at phase 1 (i.e. obesity or
mental disorder), there was no such control for phases 3 and 5.
Consequently, it is possible that the outcome of interest occurred after
phase 1 but before phase 7. This early occurrence of the outcome could
have then become a risk factor for the predictor of interest. For example,
in analysis 1, obesity incurred after phase 1 could have lead to a ‘common
mental disorder’ in a subsequent phase and obesity at phase 7. The same
argument applies to analysis 2. In this situation, it is difficult to
untangle whether it is past obesity or past mental disorder predicting
later obesity.
References:
1. Kivimaki M, Lawlor DA, Singh-Manoux A, Batty GD, Ferrie JE,
Shipley MJ, et al. Common mental disorder and obesity: insight from four
repeat measures over 19 years: prospective Whitehall II cohort study. BMJ
2009;339(oct06_2):b3765.
2. Patten SB, Wang JL, Williams JVA, Currie S, Beek CA, Maxwell CJ,
et al. Descriptive epidemiology of major depression in Canada. Can J
Psychiatry 2006;51(2):84-90.
3. Blazer D, Kessler R, McGonagle K, Swartz M. The prevalence and
distribution of major depression in a national community sample: the
National Comorbidity Survey. Am J Psychiatry 1994;151(7):979-86.
4. Grant BF, Stinson FS, Dawson DA, Chou SP, Dufour MC, Compton W, et
al. Prevalence and Co-occurrence of Substance Use Disorders and
Independent Mood and Anxiety Disorders: Results From the National
Epidemiologic Survey on Alcohol and Related Conditions. Arch Gen
Psychiatry 2004;61(8):807-16.
Competing interests:
None declared
Competing interests: No competing interests
Does overweight increase the risk of mental disorders or do mental
disorders increase the risk of overweight? : understanding the causal
relationship between overweight and psychopathology remains a challenge
for epidemiologists.
In a prospective cohort study, Mika Kivimaki and colleagues reported
that in British adults the direction of association between common mental
disorders and obesity was from common mental disorder to increased future
risk of obesity (1). In a recent cross-sectional study, we examined the
relationships between overweight, psychopathology and social functioning
in 6- to 11- years old children (2). Our results showed that child
overweight was strongly associated with poor social functioning (odds
ratio = 2.06; 95% confidence interval = 1.27–3.35). As the design of the
study did not allow concluding on the direction of the association, we
discussed the bidirectional risk between child overweight and poor social
functioning. Our first hypothesis was that overweight children may
experience elevated peer problems and peer rejection compared to non-
overweight children. To support this hypothesis, previous studies have
reported that stigmatization of obese children has increased over the last
40 years (3,4). Our second hypothesis was that peer relationship
difficulties may lead to excess weight through a number of health damaging
behaviors such as increased food intake and decreased energy expenditure.
In our discussion, we also evoked the possibility of a third factor. For
example, personality traits such as child emotional immaturity may lead to
overweight and peer difficulties by the way of oral compensation and
childish behaviors.
1. Kivimäki M, Lawlor DA, Singh-Manoux A, Batty GD, Ferrie JE,
Shipley MJ, et al. Common mental disorder and obesity insight from four
repeat measures over 19 years: prospective Whitehall II cohort study. BMJ
2009; 339:b3765
2. Pitrou I, Shojaei T, Wazana A, Gilbert F, Kovess-Masfety V. Child
Overweight, Associated Psychopathology, and Social Functioning: A French
School-based Survey in 6- to 11-year-old Children. Obesity (Silver Spring)
2009.
3. Latner JD, Stunkard AJ. Getting worse: the stigmatization of obese
children. Obes Res 2003; 11: 452–456.
4. Latner J, Stunkard A. School children, stigma, and obesity. Obes
Res 2001; 9: 94S.
Competing interests:
None declared
Competing interests: No competing interests
Nutrition and physical exercise are necessary but not sufficient for preventing obesity
The growing burden of chronic noncomunicable diseases was recognized
by The World Health Assemby in 2004. According to WHO, currently more than
1 billion adults are overweight and around 300 million of them are
clinically obese. Obesity has been suggested to increase the risk of
chronic noncomunicable diseases such as type-2 diabetes cardiovascular
diseases, type-2 diabetes and cancer. Obesity accounts for 2-6% of total
health care costs in several developed countries and not all obesity-
related conditions are included in the calculations. Childhood obesity is
already epidemic in some areas of the world and on the rise in some
countries such as in USA and Italy. An estimated 22 million children under
five are estimated to be overweight
worldwid[http://www.who.int/dietphysicalactivity]. It is believed that the
main causes behind the rising levels of childhood obesity is the shift in
diet towards increased intake of energy-dense foods and decreased levels
of physical activity. However, many of the intervention trials conducted
in children that included nutrition education and physical exercise have
not been effective in decreasing body weight [1]. There are many reasons
to explain the results such as underpowered, short time intervention and
short follow-up time. However, there might be other factors that can
influence weight gain. Physical exercise and nutrition education are
certainly necessary interventions for improving general health but they
may be not sufficient to decrease obesity. There has been growing interest
in the last years to investigate the causes of obesity to prevent it. Two
recent British cohort studies suggest that mental disorders, in particular
depressive symptoms, are associated with obesity [2,3]. The study by
KIvimaki and colleagues [2] showed a dose-response association between
mental disorder and obesity while the study of Ternouth and co-workers
[3] suggest that poor childhood self-esteem contributes to later risk for
high adult weight. These findings could be translated into new strategies
to reverse the trend of obesity, in particular among children since
childhood obesity often continues in adulthood.
1.Bond M, Wyatt K, Lloyd J, Welch K, Taylor R.Systematic review of
the effectiveness and cost-effectiveness of weight management schemes for
the under fives: a short report. Health Technol Assess. 2009;13:1-75.
2.Mika Kivimäki, Debbie A Lawlor, Archana Singh-Manoux, G David
Batty, Jane E Ferrie, Martin J Shipley, Hermann Nabi, Séverine Sabia,
Michael G Marmot, and Markus Jokela Common mental disorder and obesity:
insight from four repeat measures over 19 years: prospective Whitehall II
cohort study
3. Ternouth A, Collier D, Maughan B. Childhood emotional problems and
self-perceptions predict weight gain in a longitudinal regression model.
BMC Med. 2009 Sep 11;7:46.
BMJ 2009; 339: b3765
Competing interests:
None declared
Competing interests: No competing interests