Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Obesity and psychopathology in women: a three decade prospective study

Abstract

Objective:

To evaluate prospective associations between elevations in body mass index (BMI) at average age 27 and generalized anxiety disorder (GAD) and major depressive disorder (MDD) at average age 59 in a community sample of women.

Design:

Three waves of data collected over three decades were drawn from mothers in the Children in the Community (CIC) Study. Binary logistic regression was used to estimate predictive effects of two BMI cutpoints (30 and 25) on GAD and MDD independent of other risks for psychopathology.

Subjects:

The 544 mothers who were interviewed in the original wave of the CIC Study in 1975 and in the first and most recent follow-up waves in 1983 and 2002–2005.

Measurements:

Information about height and weight was obtained by self-report in face-to-face interviews. GAD and MDD were assessed by structured interview covering DSM-IV diagnostic criteria. Other potential risk factors examined included age, race, education, prior depressive symptoms and marital status, chronic disease, social support and financial strain concurrent with GAD and MDD.

Results:

A baseline BMI 30 significantly increased the odds for subsequent GAD and MDD by 6.27 and 5.25 times, respectively, after adjusting for other significant risk factors. Odds of GAD also increased significantly given a baseline BMI 25 (by 2.44 times); however this association was not independent of other significant risk factors. Predictive associations between a baseline BMI 30 and MDD were not attenuated by attained BMI assessed at outcome.

Conclusion:

Findings extend existing evidence of the mental health consequences of obesity in a representative sample of mothers, and suggest that obesity may have long-term implications for mental distress in women at a clinical level over the adult years.

Introduction

Evidence to the contrary notwithstanding,1, 2, 3, 4, 5 an increasing number of reports based on population samples in the United States (US) and elsewhere suggests that, among women, obesity raises the risk of psychopathology at both the sub-clinical6, 7, 8, 9, 10 and clinical11, 12, 13 levels. Because it is commonly assumed that obesity engenders mental distress, conclusions drawn from such cross-sectional findings typically attribute higher depression/anxiety symptom levels or disorder rates to elevated relative body weight (weight adjusted for height or body mass index (BMI)). Nonetheless, few studies actually have employed prospective data to test that hypothesis. The current research extends study findings that focus on comorbid associations between obesity and psychopathologic features in women, particularly at the clinical level, by investigating whether that link is longitudinal in nature.

In the largest cross-sectional study, based on the 1992 National Longitudinal Alcohol Epidemiologic Survey, odds of past year DSM-IV major depression increased significantly by 37% among obese women (BMI 30) compared to women with a BMI between 20.77 and 29.99, and by 22% with each 10-unit rise in continuous BMI, after controlling for age, race, education, income and disease history.11 National Comorbidity Survey Replication data were used to compare obese (BMI30) and nonobese (BMI<30) women on lifetime prevalence of DSM-IV anxiety disorders (generalized anxiety disorder, panic disorder or agoraphobia) and mood disorders (major depression or bipolar disorder): odds increased significantly in obese women by 34 and 29%, respectively, and were not explained by confounding due to age, smoking or comorbid psychiatric disorder.12 In findings from the Third National Health and Nutrition Examination Survey, unadjusted odds of past month DSM-III major depression increased significantly by 82% in obese women (BMI 30) and by 3.78 times in severely obese women (BMI 40) compared to women with a BMI between 18.5 and 24.99; tabulated adjusted odds, however, were not gender-specific.13

It is only recently that findings attributing increased risk of depression to elevations in BMI among women have been based on long-term prospective data, much of it being drawn from adolescent samples. Girls aged 14 in the Northern Finland 1966 Birth Cohort Study with a BMI 23.81, the gender- and age-specific 95th percentile, were at significantly greater odds (by 64%) of having elevated scores on the Hopkins Symptom Checklist (HSCL)-25-Depression Scale 17 years later after adjustment for parental marital status and social class, and disease, tobacco use, diet, physical activity and marital status at outcome.14 We recently evaluated those associations among offspring of the women studied here: girls in the Children in the Community (CIC) Study who were overweight between their 12th and 18th birthdays were at nearly four times the risk of DSM-IV major depression and at three times the risk of a DSM-IV anxiety disorder an average of 16 years later after adjustment for age, race, family socioeconomic status and tobacco use.15 On the other hand, in the Great Smoky Mountain Study, chronically obese girls assessed over an 8-year period were at increased risk of oppositional defiant disorder but not anxiety or depressive disorder; however, outcomes were assessed in adolescence.16

Evidence that obesity may be implicated in the etiology of clinical depression in women is based on prospective data drawn from adults over age 50 in the Alameda County Study:17 baseline obesity (BMI 30) significantly increased the odds of incident DSM-IV major depression by 79% 5 years later after adjustment for potential confounders that included measures of status attributes, personal and social resources and stressors. Because that association did not differ significantly across gender, however, separate odds were not given for men and women. Additional support for a causal pathway from obesity to psychiatric disturbance, most notably depression, comes from research on change in psychopathologic features among obese individuals as a function of weight loss after behavioral, dietary18, 19 or surgical20, 21 intervention. For example, Dixon et al.20 reported a significant and sustained drop in scores on the Beck Depression Inventory from 17.7±9.5 at baseline to 7.8±6.5 and 9.6±7.7 1 and 4 years later, respectively, following gastric-restrictive weight loss surgery. Nonetheless, obese individuals who enter weight reduction programs are reported to be more depressed or more anxious than obese individuals in the general population, and thus may be more likely to experience a decline in symptoms with weight loss.22

Prevalence of overweight and obesity in the US population now exceeds 60%23 and is expected to continue to rise;24 in addition, the rapid escalation in obesity rates has been implicated in a concomitant increase in incident depression.25, 26 Despite that overlap, in contrast to substantial documentation of the adverse impact of rising BMI levels on medical conditions and longevity,27, 28, 29, 30, 31 prospective evidence pertaining to the impact of BMI on mental health status has been much less prominent in the literature. Consequently, recognition of obesity as a risk factor for psychopathology, particularly at the clinical level, remains tentative. Longitudinal research on this issue is useful in that it will expand our knowledge of antecedent factors that may contribute to increased risk of affective disorder, especially among adult women, who are more prone to obesity21 and to anxiety and depression.32 Moreover, studies of mental health consequences will broaden our understanding of the detrimental influence of obesity on overall well-being and at the same time inform more accurate estimates of the social, medical and public health costs often associated with this condition.

Accordingly, given the dearth of empirical work on this issue, we examine prospective associations between obesity and psychopathologic features assessed at a clinical level, namely DSM-IV generalized anxiety disorder (GAD) and major depressive disorder (MDD), in a community-based sample of women. Obesity is defined here as a BMI 30, the recommended cutpoint for obesity in adults based on the National Heart, Blood, and Lung Institute (NHBLI)33 guidelines. Evidence of a prospective association between elevated relative body weight and anxiety disorder is especially limited; consequently, what BMI cutpoint is most useful to determine a potential link remains unclear. Thus, we also examine the predictive utility of a BMI 25, the recommended NHBLI cutpoint that includes both overweight and obesity.

Longitudinal data were drawn from mothers in the CIC Study, an ongoing investigation of child development and behavior in a randomly selected sample of families, to analyze the predictive effects of a BMI 30 and 25 (at average age 27) on increased risk of GAD and MDD (at average age 59). The other covariates considered here, including age, education level, prior depressive symptoms, marital status, chronic medical disease, social support and financial strain, are widely held to be determinants of elevated risk for depressive disorder and other mental distress in women.34, 35, 36, 37 Moreover, others have employed measures of similar constructs to adjust associations between obesity and psychopathologic features for those putative risk factors.17, 38, 39 Race also has been implicated in differential rates of depression among women;40 thus, it too is examined. In addition, we investigate whether prospective associations between BMI and GAD or MDD are mediated by attained BMI more proximal to psychopathology.

Hypotheses

The literature upon which hypotheses of long-term prospective associations of obesity and psychopathology may be based is quite sparse; consequently, our expectations lean heavily on limited prior findings. There is some support, however, for a causal connection between obesity and depression and related indicators of mental distress after adjusting for known covariates implicated in depressive disorder.17, 38, 39 Accordingly, we expected that obese women would be at greater odds of subsequent GAD and MDD compared to nonobese women independent of comparable covariates considered here. In addition, we expected that association to hold with the less severe cutpoint that included both overweight and obese women (BMI 25); however, we also assumed it would be weaker, suggesting a dose–response relationship. Given the absence of prior empirical or theoretical work to direct hypotheses on the mediational effect of more proximal BMI on prospective associations between baseline BMI and disorder, we consider these analyses to be exploratory; thus, explicit predictions are not made.

Methods

Sample

In 1975, a stratified random sampling plan was undertaken to obtain a representative sample of families residing in two upstate New York counties for a study of familial and social indicators of child development and behavior. Primary sampling units stratified by urban/rural residence, ethnicity and median income as per census data were created; a systematic sample of primary sampling unit was drawn in each county, with probabilities proportional to the number of households and equal across strata. That sampling plan yielded 1141 eligible households (that is, households with at least one child between the ages of 1 and 10); 976 (85.5%) agreed to participate. One parent (usually the mother) was interviewed about the target study child and provided family background information. The current study sample of 544 women was drawn from the 948 biological mothers interviewed in 1975 when they were between the ages of 19 and 44 (average age 32); follow-up interviews were conducted in 1983, 1985–1986, 1991–1994 and 2002–2005. Additional information regarding methodology of the CIC Study may be found in previous reports41, 42 and on the study website (http://nyspi.org/childcom).

Mothers in the 1975 sample are predominantly Caucasian (>93; 6% African American; <1% other) and span a wide range of socioeconomic levels. Because at that time the study was not intended to be longitudinal no surnames were recorded; consequently, 722 (76.2%) of the 948 mothers interviewed in 1975 were located and re-interviewed in 1983. Attrition in 1983 was non-random, with those lost disproportionately less educated (24.1 vs 17.1% with less than a 12th grade education) (χ=5.61, d.f.=1, P<0.05), unmarried (25.0 vs 11.2% separated, divorced or never married) (χ=26.69, d.f.=1, P<0.001), non-Caucasian (9.8 vs 5.0%) (χ=7.02, d.f.=1, P<0.01), and younger when they became pregnant with study offspring (mean age 25.6 years (s.d. 6.9) vs mean age 27.3 years (s.d. 6.8)) (two-tailed t=−3.92, d.f.=1, P<0.001). However, weight did not vary significantly between mothers lost to first follow-up in 1983 and those re-interviewed (mean weight 126.7 lbs (s.d.=21.2) vs mean weight 126.8 lbs (s.d.=19.7), respectively: two-tailed t=−0.03, d.f.=1, ns).

Of the 722 women interviewed in both 1975 and 1983, 558 were re-interviewed in 2002–2005; information obtained from families and confirmed by official records indicated that 71 had died since 1983, thus accounting for 629 or 87.1% of those eligible. The remaining 93 women were not re-interviewed owing to refusal to participate (29), illness (25) and failure to locate (21), or time constraints of the study (18). Data used in the current study are based on the 1975 interview and follow-up interviews in 1983 and 2002–2005. The current study sample is comprised of the 544 women with complete data relevant to this research (1975, 1983, 2002–2005). Compared to those 544 women in the current study sample, the 93 women lost to the most recent wave in 2002–2005 were disproportionately less educated (24.6 vs 14.5% with less than a 12th grade education) (χ=7.28, d.f.=1, P<0.01), whereas the 71 deceased women were disproportionately non-Caucasian (11.1 vs 2.9%) (χ=6.61, d.f.=1, P<0.01), less educated (29.2 vs 14.5% with less than a 12th grade education) (χ=10.03, d.f.=1, P<0.01), and older when they became pregnant with study offspring (mean age 30.4 years (s.d. 5.7) vs mean age 25.6 years (s.d. 5.3)) (two-tailed t=7.58, d.f.=1, P<0.001). No other significant differences between those two groups were noted in study variables obtained in the first two waves.

Procedure

Face-to-face interviews were conducted in the women's homes by trained interviewers with a minimum of 15 years experience. Written informed consent was obtained in adherence to institutional guidelines and after interview procedures has been fully explained.

Measures

Body mass index. In 1975, the women were asked how much they weighed just prior to pregnancy with the study child, about 5 years earlier on average, when they were average age 27 (s.d. =5.3). Current height and weight was obtained in 2002–2005 at average age 59 (s.d.=6.3). BMI was calculated by dividing weight in kilograms by the square of height in meters (kg m−2).

Other potential risk factors for psychopathology. Among the covariates examined was age at outcome, race (set at 0=Caucasian, 1=other) and education level (set at 0=12th grade education, 1=<12th grade education). Prior depressive symptoms were measured with a nine-item index of symptom severity based on the HSCL-90 Depression Scale43 (difficulty concentrating, sleep/appetite problems, low energy, sadness, loss of interest, self-blame, hopelessness, loneliness), each symptom rated on a five-point scale (1=‘not at all bothered by in the last year,’ 5=‘extremely bothered by…’). Used to assess sub-clinical depression in 1983 in CIC Study mothers, this scale has a high internal consistency (Cronbach's alpha [α] 0.85) and has been related significantly to age, cohort and social role differences in an overlapping sample of women.44, 45 Moreover, it is strongly associated with subsequent GAD and MDD in the current study sample (as shown in Tables 3 and 4, respectively); thus, it is used to adjust for prior psychopathology in the analyses.

Table 3 BMI at average age 27 (baseline BMI) and odds of subsequent GAD at average age 59 in a community sample of 544 women
Table 4 BMI at average age 27 (baseline BMI) and odds of subsequent Major Depressive Disorder (MDD) at average age 59 in a community sample of 544 women

Other covariates were assessed concurrently with GAD and MDD at average age 59. Marital status was dichotomized (set at 0=married, 1=unmarried (divorced, separated or widowed)). Information about chronic medical conditions (categorized as autoimmune, cancer, cardiovascular, diabetes, gastrointestinal, musculoskeletal, neurological or respiratory disease) that interfered with role function or for which medication was taken in the past 12 months was obtained by interview and dichotomized (set at 0=absent, 1=present). Presence of a chronic medical condition has been related to self-reported health and depressive syndrome in CIC Study mothers.46 Perceived social support was assessed by four items (for example, How much can you rely on your friends or family for help with a serious problem? How much can you open up to them if you need to talk about your worries?), each rated on a 4-point Likert scale (1=‘not at all,’ 4=‘a lot’), and then dichotomized at 1 standard deviation (s.d.) below the sample mean (set at 0=average/high social support, 1=low social support). Financial strain was assessed by two items (for example, Do you feel threatened financially?), each rated on a 4-point Likert scale (1=‘not at all,’ 4=‘very’) and then dichotomized at 1 s.d. below the sample mean (set at 0=no financial strain, 1=financial strain). Associations between social support and depression47 and between financial strain and self-reported health46 also have been reported in CIC mothers.

Psychopathology at outcome. Women responded to 10 and 16 interview items covering DSM-IV diagnostic criteria for GAD and MDD, respectively, at average age 59. Criteria for GAD were excessive anxiety and worry most days in the past 6 months that could not be explained by specific fears or medication or substance use or confined to features of another Axis I disorder, and an inability to control the anxiety/worry. Also required was a minimum of three of six symptoms (for example, symptoms of tension such as a pounding heart, clammy hands, an upset stomach, trouble sleeping, or muscle tension; feeling irritable, keyed up or on edge; difficulty concentrating) and impaired functioning in occupational, social, or interpersonal domains. Criteria for MDD were depressed mood, or being hopeless or suicidal, or loss of interest/lack of enjoyment in all things most of the day nearly every day for at least 2 weeks in the past year that could not be explained by medication, substance use or stressful events. Also required was a minimum of five of nine symptoms that included depressed mood (for example, tiring easily, feeling fatigued or having less energy than usual; feeling guilty or to blame for things that happened; feeling worthless and without value; continuing thoughts about death or suicide); and impaired functioning in occupational, social or interpersonal domains during the depressed state. Prospective associations between this measure of MDD and social support and neuroticism in an overlapping sample of CIC mothers support its validity.47

Statistical analysis

Differences in proportions were tested with χ2 tests, and differences in means by two-sample t-tests. Binary logistic regression was used to estimate associations between meeting DSM-IV criteria for GAD and MDD at average age 59 and a baseline BMI 30 and 25 (that is, at average age 27), controlling for other potential risk factors that included age, race, education level, prior depressive symptoms (that is, at average age 38) and marital status, chronic disease, low social support, financial strain and attained BMI assessed at outcome (that is, at average age 59). Regression analyses were conducted with the SPSS (15.0) for Windows software (binary logistic regression module) (SPSS Inc.).48

Generalized anxiety disorder and MDD each were examined in a separate set of parallel models. First, bivariate associations between each covariate and the disorders were examined to determine crude or unadjusted odds of GAD and MDD (model 1). If a bivariate association was not significant, that covariate was not considered in any subsequent model for the disorder; subsequent models were contingent upon a significant bivariate association between baseline BMI and disorder. With the exception of attained BMI, all significant covariates were examined simultaneously with baseline BMI 30 (model 2) and 25 (model 3), thus yielding adjusted odds for GAD and MDD. Finally, to determine whether more proximal elevations in BMI would mediate predictive associations between baseline BMI and GAD or MDD, attained BMI was added to models 2 and 3 where effects of baseline BMI remained significant. Attained BMI values 30, 25 and continuous were examined.

Scales assessing prior depressive symptoms and attained continuous BMI were standardized in the regression analyses to facilitate interpretation. Alpha was set at 0.05 for all analyses.

Results

Sample frequencies (%) or means (s.d.) and ranges for all study variables are shown in Table 1; continuous BMI at baseline, albeit not considered in the main analyses, is included to examine normative change in relative body weight. The correlation matrix in Table 2 shows the degree of association between all study covariates employed in the regression analyses, with associations within and between baseline BMI values and attained BMI values being of greater magnitude than associations between other potential risk factors or between other potential risk factors and baseline or attained BMI.

Table 1 Frequency (%) or mean (s.d.) and range of study variables in a community sample of 544 women
Table 2 Correlation coefficients among BMI and other potential risk factors for psychopathology examined

Normative change in relative body weight from ages 27 to 59

Mean (s.d.) BMI increased by six units, from 21.9 (3.6) at baseline to 28.0 (6.3) about three decades later, and was highly correlated over that lengthy interval (r=0.65, P<0.0001). At baseline, 17 women (3.1%) had a BMI 30, whereas 173 (31.8%) had a BMI 30 at average age 59; corresponding proportions for a BMI 25 were 84 (15.4%) and 343 (63.1%). The correlations between a BMI 30 or a BMI 25 at average age 27 and attained BMI values at average age 59 shown in Table 2 also indicate that elevations in baseline BMI were highly predictive of elevations in BMI three decades later. Additional analyses (not tabulated) demonstrated that, compared to women with a baseline BMI <30, women with a baseline BMI 30 were disproportionately among those with an attained BMI 30 (94.1 vs 29.8%) (χ=31.4, d.f.=1, P<0.0001) or 25 (100 vs 61.9%) (χ=10.3, d.f.=1, P<0.0001), and had a higher attained continuous BMI (mean=37.9 (s.d.=7.0) vs mean=27.7 (s.d.=6.0)) (two-tailed t=6.9, d.f.=1, 542, P<0.0001) three decades later. Similarly, compared to women with a baseline BMI <25, women with a BMI 25 were disproportionately among those with an attained BMI 30 (75.0 vs 23.9%, P<0.0001) or 25 (94.0 vs 57.4%, P<0.0001), and had a higher attained continuous BMI (mean=35.4 (s.d.=7.5) vs mean=26.7 (s.d.=5.0), P<0.0001) three decades later.

Bivariate associations between covariates and GAD and MDD

At average age 59, 27 (5.0%) and 42 (7.7%), women were diagnosed with GAD or MDD, respectively (Table 1). Associations between covariates and disorder are shown in Table 3 (for GAD) and Table 4 (for MDD). At the bivariate level, odds of both GAD and MDD rose significantly among women with a BMI 30 at baseline and those who were younger, had less than a 12th grade education, or had prior elevated depressive symptoms, and in women who were unmarried, had a chronic disease, or experienced low social support or financial strain at average age 59. Unadjusted odds of GAD but not MDD also rose significantly among women with a BMI 25; thus, a BMI 25 was not considered in any subsequent MDD model. Because there was no significant effect of race for either GAD or MDD, this variable was not considered any further.

Adjusted associations of BMI with GAD and MDD three decades later

General anxiety disorder. Odds of GAD rose significantly with a baseline BMI 30 (odds ratio (OR)=6.27, 95% confidence interval (CI)=1.39–28.16) but not a baseline BMI 25 (OR=2.15, 95% CI=0.78–5.93) independent of other risk factors, including age, education, prior depressive symptoms, marital status, chronic disease, low social support and financial strain (Table 3). Adjusted odds of GAD also remained significantly elevated among younger women, women with elevated prior depressive symptoms and women with a chronic disease in both model 2 (BMI 30) and model 3 (BMI 25). The Hosmer–Lemeshow test,49 which determines whether predicted probabilities for covariates match observed probabilities in binary logistic regression models, was used to assess goodness-of-fit; the large P-values obtained here indicate a good match in GAD model 2 (χ=5.22, P=0.78) and in GAD model 3 (χ=4.56, P=0.80).

To determine whether attained BMI mediated the association between a baseline BMI 30 and GAD, bivariate relations between attained BMI and GAD were examined first. As shown in Table 5, no attained BMI value among 30, 25 or continuous was significantly related to GAD; thus, mediational effects could not be tested.

Table 5 Bivariate associations between body mass index (BMI) at average age 59 (attained BMI) and concurrent psychopathology in a community sample of 544 women

Major depressive disorder

Odds of MDD rose significantly with a baseline BMI 30 (OR=5.25, 95% CI=1.41–19.58) independent of other risk factors, including age, education, prior depressive symptoms, marital status, chronic disease, low social support and financial strain (Table 4). Odds of MDD also remained significantly elevated among younger women, women with less than a 12th grade education, women with elevated prior depressive symptoms and women under concurrent financial strain. As with the GAD models, the large P-value obtained here with the Hosmer–Lemeshow test49 supports the adequacy of the MDD adjusted model for describing the data (χ=4.05, P=0.85).

To determine whether attained BMI mediated the association between a baseline BMI 30 and MDD, bivariate relations between attained BMI and MDD were examined first: As shown in Table 5, only attained continuous BMI significantly increased the odds of MDD, with the likelihood of meeting diagnostic criteria increasing by 47% with each standard deviation increase in attained continuous BMI (that is, an increase equivalent to 6.3 BMI units). With the addition of attained continuous BMI in adjusted model 2 shown in Table 4, odds of MDD remained significant for a baseline BMI 30 (OR=3.95, 95% CI=1.04–15.52), whereas the bivariate association between attained continuous BMI and MDD was no longer significant after adjusting for other risk factors, whether or not a baseline BMI 30 was included among the covariates (data not tabulated).

Discussion

The primary purpose of this study was to investigate prospective associations between obesity and psychopathologic features assessed at the clinical level in a representative sample of women. The key finding of a long-term impact of obesity on both GAD and MDD independent of other more substantiated risk factors implicated in psychopathology is consistent with our main hypothesis and expands the available database on the temporal link between obesity and mental health consequences in women. In particular, the predictive effect of obesity on MDD is compatible with reports of an increased risk of incident major depression and other indicators of mental distress over a 5-year interval in obese adults aged 50 and older,17, 39 and provides added support for obesity as a risk factor for subsequent major depression among adult women. Moreover, our findings also are among the first to establish a temporal connection between obesity and anxiety disorder, and as such have heuristic value with respect to furthering research in this area.

Perhaps what was most striking about these findings was the three-decade period over which obesity predicted psychopathology in this sample, a considerably lengthier interval than has been examined previously. Obesity appears to be a persistent condition that often precedes depression among women,22, 50 and has been associated with increased exposure to stressful life experiences in adulthood, including poor interpersonal and occupational outcomes and reduced economic circumstances.51 Obese women report lowered quality of life6 and less engagement in and enjoyment of pleasurable experiences and activities,52 and often are the targets of recurrent public ridicule and discrimination.53 The most common and frequently reported source of interpersonal stigma is close family members;54 thus, wives and mothers may experience more daily hassles over excess weight than unmarried or childless women. Moreover, unlike other stigmatized groups, obese women appear to internalize the negative perceptions toward obesity held by society and to devaluate themselves.55 It may be that, at least for some obese women, increasing cumulative stress burden and poor self-concept eventually lower the threshold at which stress exposure may precipitate depression.

The significant predictive effect of the less severe cutpoint of a BMI 25 on GAD was attenuated after adjusting for other putative risk factors; moreover, the heightened odds of MDD given a baseline BMI 25 were not significant even prior to adjustment. This finding is compatible with reports that associations between less severe levels of elevated relative body weight and psychopathologic features are either weaker13 or no longer hold once adjusted for other relevant risks.38 It also suggests that, for women, a more severely elevated BMI may precipitate adverse life experiences that are significantly greater in number or severity.

Examination of change in BMI across the three decades indicated that the proportion of women with a BMI 30 rose from 3.1 to 31.8%, thus increasing by about 10 times; corresponding proportions for women with a BMI 25 were 15.4 and 63.1%, quadrupling the rate of those either overweight or obese. The rise in relative body weight with age found here is compatible with cross-sectional reports of spiraling overweight and obesity rates.23

We also found that, albeit odds of comorbid MDD increased with elevations in attained continuous BMI at average age 59, the association was not independent of other significant covariates. This finding is compatible with reports based on both cross-sectional12 and short-term prospective38 data, suggesting that increased risk of depression may be limited to more extreme elevations in relative body weight. Nonetheless, at average age 59, neither a BMI 30, a criterion met by 173 (31.9%) women, nor a BMI 25, a criterion met by 343 (63.1%) women, significantly increased the odds of comorbid MDD at the bivariate level (Table 5). Thus, elevations in continuous BMI (but not a BMI 30 or 25) at average age 59 significantly increased the odds of comorbid MDD, at least prior to adjusting for other risk factors, whereas a baseline BMI 30 significantly increased both unadjusted and adjusted odds of MDD three decades later. Such findings bolster the argument that risk factors implicated in the etiology of psychopathologic features may differ from those that exacerbate or prolong them.56 Given the more than three-decade interval between assessment of baseline BMI and attained BMI in this sample, they also indicate that age may be a factor in the relation between obesity and psychological distress in women.

These findings should be interpreted with the following cautions in mind. Self-reported BMI is measured with some bias; however, several validation studies suggest that this bias is unlikely to affect conclusions about relations between BMI and psychopathology, particularly in longitudinal studies.57, 58, 59 Although this study is prospective with respect to the association of BMI with psychopathology, the use of pre-pregnancy weight to calculate baseline BMI also may have contributed to misclassification in the direction of underestimation owing to memory-related bias. Nonetheless, research on the validity of long-term recall of body weight supports the relative accuracy and usefulness of such data as a proxy for measured weight in epidemiologic studies, and corroborates its predictive utility.59, 60, 61, 62 For example, Casey et al.60 found that, among 50-year-old participants, the accuracy of recalled height and weight measured in childhood, adolescence and at ages 30 and 40 did not differ significantly from the accuracy of current self-reported height and weight that also was measured at the same time. Similarly, Norgan and Cameron62 reported that recall of previous height and weight over a 27- to 37-year interval (measured at 18 to 24 years old) was not significantly influenced by the passage of time. Study data were based on a primarily Caucasian sample of mothers; consequently, confirmation requires replication in more diverse samples of women with other racial backgrounds and women without offspring. Analyses were based on a one-time assessment of mental disorders; thus, the impact of BMI on incident disorder could not be ascertained. We were, however, able to adjust for prior depressive symptoms which were highly related to both GAD and MDD and remained independent of all other covariates in the adjusted models. Moreover, elevations in BMI and disorder rates may wax and wane over the lengthy interval examined here as a consequence of changing risk and protective factors; thus, data based on more than two assessment points are needed to capture potential change and reciprocity of effects. Another limitation was the small cell sizes owing to small proportions of women in the obese category at baseline and who met diagnostic criteria for GAD or MDD at outcome, which limited statistical power and precluded investigating whether associations between obesity and psychopathology were conditional on other risk factors.

Nonetheless, the current study is based on long-term prospective data obtained from a representative sample of women. The study design also incorporates two BMI cutpoints of differing severity, DSM-IV diagnostic criteria for GAD and MDD, controls for established risk factors of those disorders, and considers the impact of baseline and attained BMI simultaneously. In addition, these findings are among the first to support and extend existing findings based on prospective data regarding the causal direction of the link between obesity and depressive and anxiety disorders among women, and as such have critical clinical and public health implications related to the individual and societal costs of obesity to the mental health status of women.

References

  1. 1

    Crisp AH, McGuiness B . Jolly fat: relationship between obesity and psychoneurosis in the general population. BMJ 1976; 1: 7–9.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  2. 2

    Crisp AH, Queenan M, Sittampain Y, Harris G . ‘Jolly fat’ revisited. J Psychosom Res 1980; 24: 233–241.

    CAS  Article  Google Scholar 

  3. 3

    Hällström T, Noppa H . Obesity in women in relation to mental illness, social factors and personality traits. J Psychosom Res 1981; 25: 75–82.

    Article  PubMed  Google Scholar 

  4. 4

    John U, Meyer C, Rumpf H-J, Hapke U . Relationships of psychiatric disorders with overweight and obesity in an adult general population. Obes Res 2005; 13: 101–109.

    Article  Google Scholar 

  5. 5

    Palinkas LA, Wingard DL, Barrett-Connor E . Depressive symptoms in overweight and obese older adults: a test of the ‘jolly fat’ hypothesis. J Psychosom Res 1996; 40: 59–66.

    CAS  Article  Google Scholar 

  6. 6

    Han TS, Tijhuis MAR, Lean MEJ, Seidell JC . Quality of life in relation to overweight and body fat distribution. Am J Public Health 1998; 88: 1814–1820.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  7. 7

    Heo M, Pietrobelli A, Fontaine KR, Sirey JA, Faith MS . Depressive mood and obesity in US adults: comparison and moderation by sex, age, and race. Int J Obes 2006; 30: 513–519.

    CAS  Article  Google Scholar 

  8. 8

    Istvan J, Zavela K, Weidner G . Body weight and psychological distress in NHANES I. Int J Obes 1992; 16: 999–1003.

    CAS  Google Scholar 

  9. 9

    Johnston E, Johnson S, McLeod P, Johnston M . The relation of body mass index to depressive symptoms. Can J Public Health 2004; 95: 175–183.

    Google Scholar 

  10. 10

    Jorm AF, Korten AE, Christensen H, Jacomb PA, Ridgers P, Parslow RA . Association of obesity with anxiety, depression, and emotional well-being: a community survey. Aus N Z J Public Health 2003; 27: 434–440.

    Article  Google Scholar 

  11. 11

    Carpenter KM, Hasin DS, Allison DB, Faith MS . Relationships between obesity and DSM-IV major depressive disorder, suicide ideation, and suicide attempts: results from a general population study. Am J Public Health 2000; 90: 251–257.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  12. 12

    Simon GE, Von Korff M, Saunders K, Miglioretti DL, Crane PK, van Belle G et al. Association between obesity and psychiatric disorders in the US adult population. Arch Gen Psychiatry 2006; 63: 824–830.

    Article  PubMed  PubMed Central  Google Scholar 

  13. 13

    Onyike CU, Crum RA, Lee HB, Lyketsos CG, Eaton WW . Is obesity associated with major depression? Results from the Third National Health and Nutrition Examination Survey. Am J Epidemiol 2003; 158: 1139–1147.

    Article  Google Scholar 

  14. 14

    Herva A, Laitinen J, Miettunen J, Veijola J, Karvonen JT, Läksy K et al. Obesity and depression: results from the longitudinal Northern Finland 1966 Birth Cohort Study. Int J Obes 2006; 30: 520–527.

    CAS  Article  Google Scholar 

  15. 15

    Anderson SE, Cohen P, Naumova E, Jacques PF, Must A . Adolescent overweight and risk for subsequent major depressive disorder and anxiety disorder: prospective evidence. Psychosomatic Med 2007; 69 (in press).

  16. 16

    Mustillo S, Worthman C, Erkanli A, Keeler G, Angold A, Costello EJ . Obesity and psychiatric disorder: developmental trajectories. Pediatrics 2003; 111: 851–859.

    Article  Google Scholar 

  17. 17

    Roberts RE, Deleger S, Strawbridge WJ, Kaplan GA . Prospective association between obesity and depression: evidence from the Alameda County Study. Int J Obes 2003; 27: 514–521.

    CAS  Article  Google Scholar 

  18. 18

    Bryan J, Tiggemann M . The effect of weight-loss dieting on cognitive performance and psychological well-being in overweight women. Appetite 2001; 36: 147–156.

    CAS  Article  PubMed  Google Scholar 

  19. 19

    Klem ML, Wing RR, McGuire MT, Seagle HM, Hill JO . Psychological symptoms in individuals successful at long-term maintenance of weight loss. Health Psychol 1998; 17: 336–345.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  20. 20

    Dixon JB, Dixon ME, O’Brien P . Depression in association with severe obesity. Arch Intern Med 2003; 163: 2058–2065.

    Article  Google Scholar 

  21. 21

    van Gemert WG, Severeijins RM, Greve JW, Groenman N, Soeters PB . Psychological functioning of morbidly obese patients after surgical treatment. Int J Obes 1998; 22: 393–398.

    CAS  Article  Google Scholar 

  22. 22

    Friedman MA, Brownell KD . Psychological correlates of obesity: moving to the next research generation. Psychol Bull 1995; 117: 3–20.

    CAS  Article  Google Scholar 

  23. 23

    Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM . Prevalence of overweight and obesity in the United States, 1999–2004. JAMA 2006; 295: 1549–1555.

    CAS  Article  Google Scholar 

  24. 24

    Foreyt JP, Goodrick GK . The ultimate triumph of obesity. Lancet 1995; 346: 134–135.

    CAS  Article  PubMed  Google Scholar 

  25. 25

    Kotke TE, Wu LA, Hoffman RS . Economic and psychological implications of the obesity epidemic. Mayo Clin Proc 2003; 78: 92.

    Article  Google Scholar 

  26. 26

    Wyatt SB, Winters KP, Dubbert PM . Overweight and obesity: prevalence, consequences, and causes of a growing public health problem. Am J Med Sci 2006; 331: 166–174.

    Article  Google Scholar 

  27. 27

    Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath Jr CW . Body mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999; 391: 1097–1105.

    Article  Google Scholar 

  28. 28

    Fontaine KR, Redden DT, Wang C, Westfall AO, Allison DB . Years of life lost due to obesity. JAMA 2003; 289: 187–193.

    Article  Google Scholar 

  29. 29

    Mokdad AH, Bowman BA, Ford ES, Vinicor F, Marks JS, Koplan JP . The continuing epidemics of obesity and diabetes in the United States. JAMA 2001; 286: 1195–1200.

    CAS  Article  Google Scholar 

  30. 30

    Must A, Spadano J, Coakley EH, Field AE, Colditz G, Dietz WH . The disease burden associated with overweight and obesity. JAMA 1999; 282: 1523–1529.

    CAS  Article  Google Scholar 

  31. 31

    Pi-Sunyer FX . Medical hazards of obesity. Ann Intern Med 1993; 119: 655–660.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  32. 32

    Horwath E, Cohen RS, Weissman MM . Epidemiology of depressive and anxiety disorders. In: Tsuang MT, Tohen M (eds). Textbook in Psychiatric Epidemiology, (2nd edn). Wiley-Liss: New York, 2002, pp 389–426.

    Chapter  Google Scholar 

  33. 33

    National Heart, Blood, and Lung Institute. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults: The Evidence Report 1998. National Institutes of Health Publication 98–4083. National Institutes of Health: Bethesda, MD.

  34. 34

    Blazer DG, Kessler RC, McGonagle KA, Swartz MS . The prevalence and distribution of major depression in a national community sample: the National Comorbidity Survey. Am J Psychiatry 1994; 151: 979–986.

    CAS  Article  PubMed  Google Scholar 

  35. 35

    Kessler RC . The effects of stressful life events on depression. Annu Rev Psychol 1997; 48: 191–214.

    CAS  Article  PubMed  Google Scholar 

  36. 36

    Kessler RC . Stress, social status, and psychological distress. J Health Soc Behav 1979; 20: 259–272.

    CAS  Article  PubMed  Google Scholar 

  37. 37

    Roberts RE, Kaplan GA, Shema SJ, Strawbridge WJ . Prevalence and correlates of depression in an aging cohort: the Alameda County Study. J Gerontol B Psychol Sci Soc Sci 1997; 52: S252–S258.

    CAS  Article  PubMed  Google Scholar 

  38. 38

    Roberts RE, Kaplan GA, Shema SJ, Strawbridge WJ . Are the obese at greater risk for depression? Am J Epidemiol 2000; 152: 163–170.

    CAS  Article  Google Scholar 

  39. 39

    Roberts RE, Strawbridge WJ, Deleger S, Kaplan GA . Are the fat more jolly? Ann Behav Med 2002; 24: 169–180.

    Article  Google Scholar 

  40. 40

    Bromberger JT, Harlow S, Avis N, Kravitz HM, Cordal A . Racial/ethnic differences in the prevalence of depressive symptoms among middle-aged women: the study of women's health across the nation (SWAN). Am J Public Health 2004; 94: 1378–1385.

    Article  PubMed  PubMed Central  Google Scholar 

  41. 41

    Cohen P, Cohen J . Life Values and Adolescent Mental Health. Lawrence Erlbaum Associates: Mahwah, NJ, 1996.

    Google Scholar 

  42. 42

    Kogan LS, Smith J, Jenkins S . Ecological validity of indicator data as predictors of survey findings. J Soc Serv Res 1977; 1: 117–132.

    Article  Google Scholar 

  43. 43

    Derogatis LR, Lipman RS, Rickels K, Uhlenhuth EH, Covi L . The Hopkins Symptom Checklist (HSCL): a self-report inventory. Behav Sci 1974; 19: 1–15.

    CAS  Article  PubMed  Google Scholar 

  44. 44

    Kasen S, Cohen P, Chen H, Castille D . Depression in adult women: age changes and cohort effects. Am J Public Health 2003; 93: 2061–2066.

    Article  PubMed  PubMed Central  Google Scholar 

  45. 45

    Kasen S, Cohen P, Berenson K, Chen H, Dufur R . Dual work and family roles and depressive symptoms in two birth cohorts of women. Soc Psychiatry Psychiatr Epidemiol 2005; 40: 300–307.

    Article  PubMed  Google Scholar 

  46. 46

    Chen H, Cohen P, Kasen S . Cohort differences in self-rated health: evidence from a three-decade community-based, longitudinal study of women. Am J Epidemiol 2007; 166: 439–446.

    Article  PubMed  Google Scholar 

  47. 47

    Sneed JR, Kasen S, Cohen P . Early life risk factors of late-onset depression. Int J Geriatr Psychiatry 2007; 22: 663–667.

    Article  PubMed  Google Scholar 

  48. 48

    SPSS, Inc. SPSS 15.0 Base User's Guide. Prentice-Hall: Chicago, IL, 2006.

  49. 49

    Hosmer DW, Lemeshow S . Applied Logistic Regression. Wiley: New York, 1989.

    Google Scholar 

  50. 50

    Fabricatore AN, Wadden T . Psychological aspects of obesity. Clin Dermatol 2004; 22: 332–337.

    Article  Google Scholar 

  51. 51

    Viner RM, Cole TJ . Adult socioeconomic, educational, social, and psychological outcomes of childhood obesity: a national birth cohort study. BMJ doi:10.1136/bmj.38453.422049.EO (published online 17 May 2005).

  52. 52

    Pagoto SL, Spring B, Cook JW, McChargue D, Schneider K . High BMI and reduced engagement and enjoyment of pleasant events. Pers Individ Dif 2006; 40: 1421–1431.

    Article  Google Scholar 

  53. 53

    Puhl R, Brownell KD . Bias, discrimination, and obesity. Obes Res 2001; 9: 788–805.

    CAS  Article  Google Scholar 

  54. 54

    Puhl RM, Brownell KD . Confronting and coping with weight stigma: an investigation of overweight and obese adults. Obesity 2006; 14: 1802–1815.

    Article  Google Scholar 

  55. 55

    Wang SS, Brownell KD, Wadden TA . The influence of stigma of obesity on overweight individuals. Int J Obes 2004; 28: 1333–1337.

    CAS  Article  Google Scholar 

  56. 56

    Eaton WW . Studying the natural history of psychopathology. In: Tsuang MT, Tohen M, Zanner GEP (eds). Textbook in Psychiatric Epidemiology. Wiley-Liss: New York, 1995. pp 157–177.

    Google Scholar 

  57. 57

    Stevens J, Keil JE, Waid LR, Gazes PC . Accuracy of current, 4-year, and 28-year body weight in an elderly population. Am J Epidemiol 1990; 132: 1156–1163.

    CAS  Article  Google Scholar 

  58. 58

    Rowland ML . Reporting bias in height and weight data. Statist Bull 1989; 70: 2–12.

    CAS  Google Scholar 

  59. 59

    Stunkard AJ, Albaum JM . The accuracy of self-reported weights. Am J Clin Nutr 1981; 34: 1593–1599.

    CAS  Article  PubMed  Google Scholar 

  60. 60

    Casey VA, Dwyer JT, Coleman JT, Krall EA, Gardner J, Valadian I . Accuracy of recall by middle-aged participants in a longitudinal study of their body size and indices of maturation earlier in life. Ann Hum Biol 1991; 18: 155–166.

    CAS  Article  PubMed  Google Scholar 

  61. 61

    Casey VA, Dwyer JT, Berkey CS, Coleman JT, Gardner J, Valadian I . Long-term memory of body weight and past weight satisfaction: a longitudinal follow-up study. Am J Clin Nutr 991; 53: 1492–1498.

    Google Scholar 

  62. 62

    Norgan NG, Cameron G . The accuracy of body weight and height recall in middle-aged men. Int J Obes 2000; 24: 1695–1698.

    CAS  Article  Google Scholar 

Download references

Acknowledgements

This study was supported by grant funds from the National Institute of Child Health and Human Development HD-40685. The research was approved by the New York State Psychiatric Institute Institutional Review Board. Members of the sample participated under conditions of informed consent in adherence to institutional guidelines.

Author information

Affiliations

Authors

Corresponding author

Correspondence to S Kasen.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Kasen, S., Cohen, P., Chen, H. et al. Obesity and psychopathology in women: a three decade prospective study. Int J Obes 32, 558–566 (2008). https://doi.org/10.1038/sj.ijo.0803736

Download citation

Keywords

  • body mass index
  • anxiety
  • depression
  • longitudinal study
  • women

Further reading

Search

Quick links