The course and clinical presentation of BD manifest differently in men and women, wherein women exhibit a later onset of BD, more seasonal variations in mood disturbance, and increased susceptibility to relapse- and depression-prone phenotypes (e.g., BD type II, rapid cycling BD, and mixed states) [16-19]. Sutin and Zonderman  reported in the Baltimore Longitudinal Study of Aging that sex moderates the association between depressive symptoms and weight gain. They concluded that women with depressive symptoms present with a greater body mass index (BMI) as well as increased waist and hip circumferences across the adult lifespan when compared to men experiencing similar symptoms. Taken together, these observations indicate that women with BD, as a function of being disproportionately affected by depressive symptoms, may be more susceptible to weight gain [10, 20, 21].
Phenomenological factors can be expanded to include differences in patterns of comorbidity between the sexes among patients with BD. For example, hypothyroidism is reported to be more common in women with BD when compared with men with BD. In a study evaluating the rates of previously diagnosed thyroid disease in 443 individuals with BD compared by sex, race, and manic subtype (i.e., mixed state versus pure manic episodes), hypothyroidism was reported to be significantly higher in female patients, Caucasians, and those of advanced age . A bidirectional association between obesity and peripheral levels of thyroid hormones (i.e., hypothyroidism) has been observed. Thyroid-stimulating hormone (TSH) levels are often increased in individuals with obesity when compared to normal weight individuals . Elevated TSH levels observed in clinical hypothyroidism are associated with weight gain, decreased thermogenesis, and decreased metabolic rate . Moreover, thyroid axis abnormalities are associated with an increased susceptibility to relapse (notably depression) . Hence, it is reasonable to conclude that thyroid abnormalities, which affect women at a higher rate, increase susceptibility to depressive symptoms and episodes as well as obesity.
Replicated evidence has documented that a linear relationship exists between migraine and BMI, particularly in women . In the USA, the prevalence of migraine has been reported to be three times higher in women when compared to men . Women who are obese are at increased risk (i.e., 48%) of experiencing migraine when compared to sex- and age-matched healthy controls . Women who are morbidly obese are at even greater risk (i.e., 275%) of experiencing migraine when compared to women who are obese and normal weight . The bidirectional relationship between migraine and obesity has implications for the BD population. For example, it is reported that individuals with BD, notably BD type II and bipolar spectrum disorders, have elevated rates of migraine relative to the general population [28, 29]. Ortiz-Domínguez et al.  reported that the prevalence of migraine in their mixed cohort of subjects with BD (N = 323) was 24.5%, with a higher rate of migraine in BD type II (34.8%, p < 0.005) when compared to those with BD type I (19.1%, p < 0.005). It has been additionally reported that the rate of bipolar spectrum disorder is significantly elevated in cohorts of migraineurs in a clinical setting .
Several non-mutually exclusive hypotheses have been proposed to explain the co-occurrence of migraine and mood disorders. For example, alterations in central serotonin signaling, implicated in affect regulation, are reported to alter the nociceptive signaling pathways implicated in migraine pain . Moreover, hypothalamic-pituitary-adrenal (HPA) axis abnormalities are the most replicated biological abnormalities in mood disorders and are also identified in migraine populations . Proinflammatory mechanisms may also represent a pathophysiological nexus subserving BD phenomenology, obesity, and migraine . Finally, the reproductive hormonal milieu may also be contributory insofar as women with affective disorders frequently experience migraine during the late luteal phase of the menstrual cycle, which coincides with marked reductions in peripheral estrogen levels, increased sympathetic activity, and reduced central serotonergic and GABAergic activity, potential mechanisms predisposing to body weight gain .
Reproductive life events
The impact of reproductive life events (i.e., menarche, pregnancy, postpartum, and menopausal transition) is a major distinction between men and women as it relates to the differential association between obesity and BD . Reproductive life events may affect propensity toward obesity indirectly by altering illness presentation, course, and outcome, or directly by altering neuroendocrine systems implicated in metabolism and weight homeostasis .
Women with BD are highly susceptible to onset and recurrence of affective episodes during each reproductive life event . Hartlage et al.  demonstrated that premenstrual dysphoric disorder (PMDD), characterized as a cycling of dysphoria, irritability, and mood lability, is a significant risk factor for depressive episodes, with approximately seven out of eight women with PMDD meeting criteria for a mood disorder within two years. Moreover, the risk of recurrence for a depressive or mixed-state episode during pregnancy has been reported to be between 63 and 74%, with 47% of recurrences occurring during the first trimester . The postpartum period is the time of highest vulnerability for risk, with relapse rates exceeding 50% in mixed cohorts [36, 37]. Several factors have been proposed to explain the increased susceptibility to affective episodes associated with pregnancy/postpartum including, but not limited to, medication non-adherence, alterations in drug pharmacokinetics, changes in reproductive hormonal concentrations, and adverse effects on sleep hygiene/architecture [19, 38-40].
Pregnancy is highly associated with an increased propensity to overweight/obesity in the general population. Bastian et al.  reported a dose–response relationship between significantly higher rates of obesity and increasing number of children, with the risk for obesity increasing by 11% following each additional live birth, an observation independent of socioeconomic status (SES) and other covariates. Numerous mechanisms have been proposed to explain this association (e.g., insulin resistance associated with pregnancy, hormonal alterations secondary to fewer ovulatory cycles, increased glucocorticoid activity, and the excess deposit of fat tissue that accumulates in the femoral area during pregnancy). The physiological changes associated with pregnancy often persist years after childbearing .
Motherhood may be associated with changes in diet and physical activity . For example, in a study assessing whether psychological and behavioral changes following pregnancy influenced long-term weight gain, it was reported that mothers who retrospectively recalled eating more following childbirth had significantly greater long-term weight gain [2.78 kg (standard deviation = 1.42)] when compared to those who reported that they had not increased their food intake [−1.15 kg (0.76)] . Mothers who retrospectively reported more access to food postpartum had significantly greater long-term weight gain [1.70 kg (0.87)] when compared to those who did not [−1.37 kg (1.13)] . In addition, mothers who reported decreased exercise following pregnancy relative to their energy expenditure prior to pregnancy were also at greater risk of long-term weight gain . The influence of social networks on propensity to gain weight in women was underscored by an increased propensity to gain weight in postpartum women who reported significantly fewer interpersonal relationships .
Mood symptoms may also worsen during the menopausal transition, with increased irritability, anxiety, depression, and emotional lability, as well as reduced energy, motivation, and concentration [43, 44]. Postmenopausal women with a history of BD represent a subpopulation of women at especially high risk of recurrence, with approximately 55% reporting a worsening of mood episodes . Studies suggest that reproductive hormones including follicle-stimulating hormone, luteinizing hormone, and estrogen may be the chief contributors to mood instability [46, 47]. It has been well documented that the menstrual cycle and the menopausal transition influence susceptibility to relapse and course of BD [43, 44]. Emerging data suggest that the menopausal transition may be a time of increased risk for depression in women with BD .
The effect of reproductive life events on BD presentation/trajectory, course and outcome, and propensity to obesity, is also influenced by iatrogenic factors. For example, preliminary evidence suggests that valproate is associated with menstrual abnormalities (e.g., oligoamenorrhea) as well as a biochemical profile commensurate with polycystic ovarian syndrome (PCOS) [49, 50]. PCOS has been associated with infertility and increased metabolic dysregulation via increases in androgen production and an elevated release of gonadotropin-releasing hormone . Features of PCOS may include insulin resistance, hyperinsulinemia, hyperandrogenism, and menstrual irregularities [52, 53]. PCOS affects approximately 5–10% of reproductive-aged women in the general population; the estimated rates of PCOS are higher in women with BD . Women in the general population with PCOS are differentially affected by all components of the NCEP-ATP-III-defined MeS [52, 55, 56].
Reports of exposure to childhood adversity (e.g., neglect, emotional, physical and/or sexual abuse) in adults with BD are associated with a more complex presentation (e.g., suicidality), clinical course and outcome (e.g., early age at onset, and rapid cycling), as well as increased susceptibility to several comorbid conditions (e.g., substance use disorder) [57, 58]. It is well established that adult women with BD are more likely to report childhood adversity (i.e., sexual abuse) when compared to men with BD . There is growing evidence in support of sex differences in childhood adversity amongst individuals with mood disorders. In a sample of 118 individuals with BD type I treated for a first episode of psychotic mania, the reported prevalence of childhood and adolescent sexual and/or physical abuse was 36.2% for female patients, with 29.8% exposed to sexual abuse and 19.1% to physical abuse, as compared to 16.9% of male patients, with 5.6% exposed to sexual abuse and 14.1% to physical abuse .
Epidemiological studies in the general population have demonstrated that early childhood adversity correlates with overweight/obesity in adulthood . Early adversity (e.g., poverty, social stress, and harsh parenting) has been reported to be associated with elevated adult BMI; more specifically, individuals with childhood major depression have a BMI of 26.1 ± 5.2 compared to a BMI of 24.2 ± 4.1 in healthy comparisons . A retrospective study in adults reported that childhood social stressors may impact the risk for obesity and other adverse medical conditions (e.g., ischemic heart disease, cancer, and liver disease) . Similarly, Alciati et al.  conducted a study investigating the rate of childhood parental loss in obese bariatric surgery candidates and its association with psychiatric diagnoses and weight/eating-related characteristics. Results supported an association between obesity and BD; moreover, early parental loss appeared to be a factor of psychiatric severity as well as weight and BMI values (unrelated to pharmacological effect) in individuals with BD type II .
Taken together, these results suggest that the psychopathology associated with childhood adversity increases the risk of obesity in adulthood. Indeed, several neurobiological mechanisms have been proposed to explain the long-term behavioural and biochemical consequences of early life stressors including, but not limited to, stress axis activation (e.g., HPA) and comorbidity (e.g., binge eating disorder) .
Eating behaviours emerge as a significant lifestyle-related risk factor for overweight/obesity as early as pre-adolescence, which may be a result of dysfunctional reward mechanisms contributing to poor food choices and overeating . Diets high in sugar and saturated fats with low fiber content along with a sedentary lifestyle can further increase an individual's vulnerability to other medical comorbidities (e.g., CVD and diabetes type II mellitus) . Lee et al.  evaluated the eating patterns of school children and reported that those who ate rapidly and engaged in overeating more than twice per week had three times the risk of being overweight. Clinical evidence suggests that children and adolescents with major depressive disorder are at greater risk for developing overweight . It could be conjectured that a subpopulation of children and adolescents with major depressive disorder will later declare themselves as having BD, underscoring the association between BD and obesity early in the illness course .
A cross-sectional study of nutrient intake and physical activity in individuals with BD demonstrated that mean total energy intake was higher in women versus reference subjects, with total daily sucrose intake, percentage of energy from carbohydrate, total fluid intake, and intake of sweetened drinks being slightly higher amongst female individuals when compared to male individuals . Studies evaluating how obese women respond to orosensory pleasure elicited by sweetness when compared to lean women indicate that obesity is associated with slower patterns of habituation, wherein habituation characterizes a response that does not result in either receptor adaptation or effector fatigue following repetitive stimulation [69, 70]. Slower patterns of food habituation are associated with greater energy intake and may contribute to prolonged eating episodes, slower satiation rates, and excessive food consumption . Likewise, differences have been observed in obese women when examining smoking behavior, alcohol consumption, and continuance of physical exercise . A prospective study evaluating diet quality, physical activity, smoking status, and weight fluctuation as predictors of weight change in women and men reported that in women, diet quality interacted with former smoking status . Former smokers with lower diet quality gained an additional 5.2 kg when compared to individuals with higher diet quality . Among women, age and physical activity were stronger predictors of weight change than diet quality .
Obese individuals with binge eating disorder or subthreshold binge eating disorder (i.e., insufficient binge frequency to satisfy DSM criteria) exhibit higher hypomania scores as measured by the Hypomania Checklist (HCL-32) . Moreover, the rate of binge eating disorder is increased in adults with BD, with much higher rates reported in women [73, 74].
Kleinman and colleagues  assessed health-related absence and productivity output of employees with BD compared with those of non-BD and other employee cohorts from a large employer database. Employees with BD had significantly higher absence costs ($1,219) and 11.5 additional lost days per year compared with those without BD. Adjusted annual productivity output was also 20% lower for the BD group. SES has been amply documented to increase women's risk of obesity along a social gradient, wherein women of lower SES who are less educated demonstrate elevated rates of overweight/obesity when compared to their wealthier and more educated compatriots . Moreover, obese individuals with low SES are more likely to report low self-perceived health, quality of life, and intensity of physical activities  when compared to individuals with overweight/obesity who report a higher income. Early social disadvantage has been reported to affect women's weight status and adult BMI more than those of men .
Pertinent to this review is whether women with BD are more likely to be exposed to (weight-gain promoting) psychotropic medication. Women, relative to men, are more likely to be identified as having a mood disorder and be treated with sequential pharmacotherapy [79-81]. Many of the psychotropic medications disproportionately utilized in BD subpopulations are associated with metabolic side effects such as weight gain, insulin resistance, and dyslipidemia.
Lithium is associated with weight gain in up to 60% of treated individuals with BD. Mechanisms of weight gain include fluid retention, increased appetite, or lithium-related subclinical hypothyroidism [22, 82-84]. Valproate is also associated with clinically significant weight gain in approximately 3–20% of treated individuals with BD . Metabolic disturbances related to valproate treatment include insulin resistance, hyperlipidemia, impaired glucose tolerance, and hyperinsulinemia [50, 86].
In a 12-month prospective, naturalistic study of 47 individuals with BD receiving maintenance treatment following their first manic episode, mean weight gain was reported to be 4.76 kg in individuals with BD when compared to healthy age- and sex-matched controls (1.50 kg) . Logistic regression indicated that weight gain in the first six months was significantly associated with the prescription of olanzapine or risperidone, with mean weight gain of 11.38 kg (n = 10) and 4.12 kg (n = 16) by 12 months, respectively; in addition, male patients with BD displayed a trend toward greater weight gain when compared to age- and sex-matched healthy controls (7.20 kg versus 2.03 kg, respectively; F = 3.969, df = 1, p = 0.055) . The investigators also reported that over the 12-month period a comparable number of women from both the BD group and healthy control group experienced clinically significant weight gain [39.1% versus 33.3%, respectively; χ2 = 0.114, df = 1, p = not significant (NS)], with mean weight changes of 2.51 and 1.04 kg, respectively (F =0.506, df = 1, p = NS) .
Covell et al.  reported that individuals treated with clozapine, compared to those treated with first-generation antipsychotic medication, gained more weight after two years, with women exhibiting significantly greater weight gain than men. Likewise, a retrospective study of BMI in individuals treated with antipsychotic agents over extended periods of time demonstrated that the OR for weight gain was significantly higher in women than in men . Female sex was also identified as a predictor for significant weight gain in individuals receiving olanzapine and fluoxetine treatment in combination .
Conventional unimodal antidepressants are highly associated with clinically significant weight gain as well as alterations in glucose and lipid homeostasis. Available evidence indicates that the metabolic hazards of tricyclic antidepressants (TCAs) and monoamine oxidase inhibitors (MAOIs) exceed what is reported for contemporary antidepressants [e.g., selective serotonin reuptake inhibitors (SSRIs)]. Notwithstanding, clinically significant metabolic changes can be seen with SSRIs, notably paroxetine. Taken together, the depression predominance of bipolar expression in women can result in increased exposure to weight-gain promoting antidepressant therapy [91, 92].