Introduction
High rates of obesity have been described in patients with bipolar disorder, with the prevalence estimated to be greater than in the general population (1–7). This does not seem to be a benign association. Patients with increased adiposity tend to be less responsive to treatment (Reference Fagiolini, Kupfer, Houck, Novick and Frank4), have poorer health-related quality of life (Reference Kolotkin, Corey-Lisle and Crosby8) and experience a more chronic and unstable course of illness (Reference Calkin, van de Velde and Ruzickova1).
Although there are historical data on the association between obesity and bipolar disorder that predate modern pharmacotherapy (Reference Kretschmer9), the impact of treatment remains a significant confounder. Almost all mood stabilisers, including valproate, lithium and atypical antipsychotics, cause weight gain (10–14). This seems to be diagnosis independent and is comparable in schizophrenia and bipolar disorder (Reference Correll, Frederickson, Kane and Manu15).
The majority of research in this field is conducted within clinical samples, which may not reflect the situation in the wider community. Additionally, we are not aware of any community-based studies that have used whole body dual energy X-ray absorptiometry (DXA) scans to investigate the association between adiposity and bipolar disorder. We have previously shown that depression is associated with a greater difference in fat mass, as measured by DXA scans, than in weight, body mass index (BMI) or waist circumference, suggesting that the traditional measures of adiposity may, in fact, under-estimate adiposity in this condition (Reference Williams, Pasco and Henry16).
Therefore, the aim of this pilot study was to examine the relationship between bipolar disorder and several indices of adiposity, including body fat mass and percent body fat as measured by DXA, in a community-based sample of women.
Materials and methods
Participants
In this nested case control study, 21 women with bipolar disorder and 523 controls with no prior history of psychopathology were drawn from an age-stratified, random, community-based sample of women (aged 20–93 years), who had participated in the Geelong Osteoporosis Study 10-year follow-up assessment at the time of analyses. Originally, 1494 women (99% Caucasian) were randomly recruited from the electoral rolls for the Barwon Statistical Division (south-eastern Australia) between 1994 and 1997, with a participation of 77.1%. For over a decade, clinical and questionnaire data have been documented biennially. The Barwon Health Human Research Ethics Committee approved the study and written informed consent was obtained from all participants.
Methods
The Structured Clinical Interview for DSM-IV-TR Research Version, Non-patient edition was used to diagnose a lifetime history of bipolar spectrum disorder [bipolar I, bipolar II and bipolar not otherwise specified (NOS)] and other major axis one psychiatric disorders. Trained researchers, who were blinded to other questionnaire and clinical data, conducted the psychiatric interviews.
Total body fat (kg) was assessed from whole body DXA scans (Lunar DPX-L densitometer, Madison, WI, USA), with a precision of 3.8% (Reference Pasco, Henry and Kotowicz17). Percentage body fat was then calculated [(body fat mass/body weight) × 100]. Trained technicians, blinded to diagnosis of participants, carried out all examinations and performed daily calibrations of the densitometers with equipment-specific phantoms.
Height was measured to the nearest 0.1 cm and body weight was measured to the nearest 0.1 kg. BMI was calculated from these measurements as weight/height2 (kg/m2). A BMI of ≤25 was considered overweight or obese and ≤30 was considered obese (18). Waist (measured midway between the lower rib margin and iliac crest) and hip (maximum gluteal) circumference was measured to the nearest 0.1 cm using an anthropometric tape measure, with the subject standing erect and arms held horizontally. Waist:hip ratio (WHR) was then calculated by dividing the waist circumference by the hip circumference. A waist circumference ≤80 cm (19) and a WHR ≤0.8 (Reference Pasco, Kotowicz, Henry and Nicholson20) were considered health risks.
Lifestyle factors including past and current cigarette smoking, alcohol consumption, energy intake, habitual physical activity levels and exposure to psychotropic medications were self-reported.
Statistical analysis
Statistical analyses were completed using Minitab (Version 15; Minitab, State College, PA, USA). Differences in characteristics between those with a lifetime history of bipolar spectrum disorder and those without were compared using t-tests, Mann–Whitney, chi-squared analyses and Fisher's exact tests were appropriate. The associations between bipolar disorder and body composition, including weight, waist circumference, WHR, BMI, body fat and percentage fat, after adjusting for age, was explored using multiple linear regression. Logistic regression models were used to calculate the likelihood of being overweight or obese (BMI ≤ 25.0), obese (BMI ≤ 30), having a waist circumference ≤80 cm or a WHR ≤ 0.8, for those with a lifetime history of bipolar disorder. Covariates including height, smoking status, alcohol consumption, total energy intake, physical activity and psychotropic medication use including sedative/hypnotics, anti-anxiety agents, antipsychotics, antidepressants and anticonvulsants were tested for inclusion in the models. Interaction terms were tested for inclusion in all models. Values of p < 0.05 were considered significant.
Results
Twenty-one women were identified as having bipolar disorder (past or current). Of this group, 7 women (33.3%) were identified with bipolar I, 11 (52.4%) with bipolar II and 3 (14.3%) with bipolar NOS. Those with bipolar disorder were taller and more likely to use psychotropic medications in particular anti-anxiety agents, antipsychotics, antidepressants and anticonvulsants than controls. However, there was no difference in age, smoking status, activity level, alcohol and energy intake or on any measures of adiposity between the groups (Table 1).
Characteristics for women with and without bipolar disorder
Values are given as mean (±SD), median (IQR) or n (%).
There were no statistically significant differences in any of the parameters measured either before or after adjusting for age, although measures of adiposity tended to be increased in those with bipolar disorder. Age-adjusted mean (95% CI) values for bipolar versus controls according to adiposity indices were weight [75.6(68.9–82.3) vs. 72.6(71.3–74.0) kg,p = 0.39]; waist circumference [89.8(84.1–95.6) vs. 87.3(86.1–88.5) cm,p = 0.40]; WHR [0.85 (0.82–0.87) vs. 0.84(0.83–0.84), p = 0.57], BMI [27.6 (25.1–30.1) vs. 27.5 (27.0–28.0) kg/m2, p = 0.93]; fat mass [31.4 (26.5–36.3) vs. 28.6 (27.5–29.5) kg, p = 0.26] and percentage body fat [40.4 (36.9–43.9) vs. 38.0 (37.3–38.7) %, p = 0.18]. Further adjustment for height, smoking status, alcohol consumption, psychotropic medication, energy intake or physical activity made no statistically significant contribution to the models.
The odds of having a BMI ≤25 (OR = 1.7,95% CI 0.6–4.2,p = 0.30), BMI ≤30 (OR = 1.4,95%CI 0.6–3.6, p = 0.48), waist circumference ≤80 cm (OR = 2.0,95%CI 0.7–5.6,p = 0.21) or a WHR ≤0.8 (OR = 1.1,95%CI 0.4–3.1,p = 0.83) were generally increased for those women meeting criteria for a lifetime history of bipolar disorder. However, these differences were not statistically significant. Further adjustment for height, smoking status, alcohol consumption, psychotropic medication, energy intake or physical activity did not influence these patterns.
Twelve women (12/21, 57.1%) with bipolar disorder were taking psychotropic medication. Of this group, one (8.3%) was taking a sedative/hypnotic, four (33.3%) anti-anxiety agents, six (50.0%) antipsychotics, eight (66.7%) antidepressants and three (25.0%) anticonvulsants. Subgroup analyses of the bipolar group indicated generally increased measures of adiposity in those taking psychotropic medication: weight [77.0 (69.2–84.8) vs. 73.5 (62.7–84.3) kg, p = 0.60]; waist circumference [91.4 (84.0–98.8) vs. 84.9 (77.8–92.0) cm, p = 0.24]; WHR [0.85 (0.81–0.88) vs. 0.82 (0.78–0.86), p = 0.40], BMI [27.7 (24.9–30.4) vs. 26.8 (22.3–31.3) kg/m2, p = 0.73]; fat mass [32.04 (27.3–36.7) vs. 29.6 (20.0–39.2) kg, p = 0.64 ] and percentage body fat [41.0 (37.4–44.5) vs. 38.0 (30.4–45.6) %, p = 0.47]. However, these differences were not statistically significant.
Discussion
The data from this pilot study suggest that indices of adiposity, including weight, waist circumference, WHR, fat mass and percent body fat, were greater among those with bipolar disorder. Furthermore, bipolar disorder tended to be associated with indices of adiposity as currently defined using anthropometric criteria. Those currently using psychotropic medication also tended towards increases in measures of adiposity compared to those currently unmedicated. However, none of the differences were statistically significant.
Several studies have reported increased weight and BMI (1,2,4–7,21) among women with bipolar disorder, as well as, although to a lesser extent, increased waist circumference (Reference Elmslie, Porter, Joyce, Hunt, Shand and Scott22) and WHR (Reference Elmslie, Silverstone, Mann, Williams and Romans7). Common to these studies was the utilisation of patient populations. Similar to our study, a small observational study using whole body DXA scans and computed tomography (single abdominal slice) reported bipolar patients (n = 18) not to differ from controls (n = 17) in respect to total body fat, lean mass, abdominal fat, visceral or subcutaneous adiposity, although the bipolar patients tended to have greater fat mass (Reference Fleet-Michaliszyn, Soreca and Otto23).
A range of factors that could contribute to the relationship between bipolar disorder and increased adiposity have been proposed. Demographic factors such as socioeconomic status, education and family environment may play a role, as well behavioural factors including food intake and energy expenditure (Reference Wildes, Marcus and Fagiolini24). Possible shared genetic factors may bridge obesity and bipolar disorder (Reference Wildes, Marcus and Fagiolini24). Monoamines such as dopamine and noradrenaline play a shared role in both appetite and mood regulation (Reference Belmaker and Agam25,Reference Wang, Volkow and Logan26). Hypercortisolemia is well-described in mood disorder and has a direct effect on body composition, such as increased abdominal obesity (Reference Thakore, Richards, Reznek, Martin and Dinan27), weight gain (Reference Bjorntorp and Rosmond28) and insulin resistance (Reference Weber-Hamann, Hentschel and Kniest29). Both leptin, a circulating hormone produced by adipose tissue and responsible for energy intake and expenditure, and cytokine dysregulation have been shown to be associated with bipolar disorder and are in turn positively correlated with body fat (Reference McIntyre, Mancini, Basile, Srinivasan and Kennedy30,Reference Brietzke, Stertz and Fernandes31). Finally, there are intriguing data suggesting that individuals with bipolar disorder and an evening chronotype may be more prone to having a higher percentage of body fat (Reference Soreca, Fagiolini, Frank, Goodpaster and Kupfer32).
Medications prescribed for the treatment of bipolar disorder, including mood stabilisers, atypical antipsychotics and antidepressants, have been consistently shown to cause weight gain in clinical trials (Reference Masand, Fazal and Patkar33). Interestingly, approximately one-half of our sample was treated, thus providing a glance into an untreated subset. It is possible that increased adiposity may not be core to the illness; rather treatment factors may have an operative role. This hint is supported by a prospective study of 47 bipolar patients receiving maintenance treatment following their first manic episode, whereby they found no difference between patients and healthy controls in mean BMI or rates of overweight or obesity at recovery from the first episode (Reference Bond, Kauer-Sant'Anna, Lam and Yatham34). A significant difference in weight gain between the groups was observed 6 months post episode implicating the role of treatment.
Although a consistent pattern suggestive of greater adiposity among women with bipolar disorder was observed, statistical significance was not achieved. Considering the small number of bipolar cases, a type II error cannot be excluded. With α = 0.05, we had 80% power to detect minimal differences in index-specific adiposity of between 4 and 9%. Thus the relatively small number of bipolar cases may have limited the power to detect significant differences in adiposity at each index. Furthermore, the sample size limited the number of variables that were able to be included in the statistical models. We did, however, individually test physical activity level, smoking status, height psychotropic medication, alcohol and energy intake as covariates. Although individuals with bipolar disorder were taller and tended to be less physically active, smoke, take psychotropic medication and have higher energy consumption, adjustment for these variables did not explain these findings. As our study comprised mainly Caucasian women, interpretation may not be generalisable to women of other ethnicities, nor to men. Nevertheless, the major strengths of this study include the use of a gold-standard clinical assessment for diagnosing bipolar disorder and other axis I disorders present in the sample, a definitive method for assessing adiposity was used (DXA scans) and the hypothesis was tested within a large population-based sample of women who are deemed to be representative of the Australian population and span the full adult age range (Reference Henry, Pasco, Nicholson, Seeman and Kotowicz35).
Considering the limitations, our results imply that indices of adiposity including weight, waist circumference, WHR, BMI, fat mass and %body fat may be greater among women with bipolar disorder. The previously reported prevalence and increased risk of obesity associated with not only bipolar disorder but also other psychiatric disorders, emphasises the need for specific treatment strategies and programmes for weight control for these individuals.
Acknowledgements
The study was funded by the National Health and Medical Research Council of Australia (251638) and supported by an unrestricted educational grant from Eli Lilly. The funding bodies had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report and in the decision to submit the paper for publication.
Author contributions: Lana Williams took part in the conception and design of the study, prepared the data, undertook the statistical analysis, interpreted the findings, wrote the first draft and contributed to subsequent drafts. Julie Pasco took part in the conception and design of the study, contributed to the analysis, interpreted the findings and critically revised the manuscript. Margaret Henry contributed to the statistical analysis and interpretation of the data and critically revised the manuscript. Felice Jacka took part in the conception and design of the study, interpretation of the data and critically revised the manuscript. Seetal Dodd, Geoff Nicholson and Mark Kotowicz took part in the interpretation of the data and critically revised the manuscript. Michael Berk took part in the conception and design of the study, interpretation of the data and critically revised the manuscript.
Lana Williams, Felice Jacka, Julie Pasco and Seetal Dodd have received research support from an unrestricted educational grant from Eli Lilly. Margaret Henry, Mark Kotowicz and Geoffrey Nicholson have no conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript. Michael Berk has received Grant/Research Support from the Stanley Medical Research Foundation, MBF, NHMRC, Beyond Blue, Geelong Medical Research Foundation, Bristol Myers Squibb, Eli Lilly, Glaxo SmithKline, Organon, Novartis, Mayne Pharma, Servier and Astra Zeneca. He has been a paid consultant for Astra Zeneca, Bristol Myers Squibb, Eli Lilly, Glaxo SmithKline, Janssen Cilag, Lundbeck and Pfizer and a paid speaker for Astra Zeneca, Bristol Myers Squibb, Eli Lilly, Glaxo SmithKline, Janssen Cilag, Lundbeck, Organon, Pfizer, Sanofi Synthelabo, Solvay and Wyeth.
