The Interplay of Gender, Mood, and Stress Hormones in the Association between Emotional Eating and Dietary Behavior^,

See corresponding article on page 1264

Human and animal studies indicate that stress may affect eating behavior in various ways, under different circumstances. Generally, intense emotions trigger inhibited food intake through fight-or-flight responses, which result in glucose release into the blood stream, thus suppressing the feelings of hunger (1). Conversely, impaired cognitive eating control, emotion-congruent modulation of eating, and eating to regulate emotions have been suggested mechanisms underlying abnormal eating behavior observed in groups of individuals (2). In such instances, stress enhances intake of palatable and “nutrient poor” foods, high in sugars and saturated fat content (3–5), thereby reducing intakes of fruits and vegetables and overall dietary quality (6, 7). Alternatively denoted as “stress-related eating,” “eating comfort foods,” and “emotional eating,” this behavior is a direct consequence of negative mood states, such as sadness, loneliness, and concern (5, 8–10). In the long run, such eating behavior contributes substantially to weight gain at the individual level (11), thereby increasing the burden of obesity at the population level (12–14). Generally speaking, abnormal eating behaviors leading to increased risk of obesity include a combination of emotional eating (i.e., eating as a result of negative mood states), uncontrolled eating (i.e., not being able to control the amount of food consumed as a result of external triggers), and low cognitive restraint (i.e., inability to restrict food intake in order to control body weight) (10).

In this issue of The Journal of Nutrition, Camilleri et al. (15) examined the relation between emotional eating and consumption of energy-dense snacks by using data on French adults collected from a large cross-sectional survey. The analysis involved multiple linear and logistic regression models in which gender and depressive symptoms were included as potential effect modifiers of the main association. The key finding was that among depressed women, emotional eating was associated with greater consumption of energy-dense snacks. This synergism between emotional eating and depressive status was not found among men.

Several instruments to assess emotional eating were used in the current study and in the past literature that either incorporated a limited range of emotions or mood states or inquired further about binge eating episodes (16–21). The measure of emotional eating used in the study by Camilleri et al. (15) is a subscale of the French version of the Three-Factor Eating Questionnaire (TFEQ-R21) (20); the 2 other subscales were “uncontrolled eating” and “cognitive restraint” subscales. Similar to the Emotional Eating Scale (21), this subscale tackles the direct response to a wide range of emotions in terms of urge to eat. Those affective states include anger/frustration, anxiety, and depression. A study in generally overweight men and women indicated that the Emotional Eating Scale (21) has a 4-factor structure (Anger, Anxiety, Depression, and Somatic), accounting for 60% of the total variance, with ∼40% of this total variance being explained by the Depression factor (22). Consequently, depression is a major contributor to emotional eating. It ensues that the association between depressive symptoms and intake of non–health-promoting foods is potentially mediated by emotional eating. In fact, using structural equations modeling techniques and data from Finnish men (n = 2312) and women (n = 2674) aged 25–74 y, a recent study found that both emotional eating and physical activity self-efficacy are important mediators between depressive symptoms and adiposity (5). Another study from the same group of researchers in 25- to 64-y-old Finnish men (n = 1679) and women (n = 2035) from the FINRISK 2007 Study (Dietary, Lifestyle and Genetic Determinants of Obesity and Metabolic Syndrome substudy) came to the conclusion that emotional eating was also a mediating factor between depressive symptoms and increased intake of sweet foods but not reduced intake of fruits and vegetables (23). Recent methodologic advancements allow to test mediation while relaxing the exact additivity of effects assumption between exposure (e.g., emotional eating) and mediator (e.g., depression) variables (24, 25), thus disentangling those 2 potential effects.

Moreover, in the study by Camilleri et al. (15), the observed synergism between depression and emotional eating in relation to consumption of high-energy-dense snacks was restricted to women. Given the large sample size in both genders, inadequate statistical power to detect the association among men is an unlikely explanation. Moreover, another large cross-sectional study also observed this relation to be stronger among women (11). An alternative explanation is that men respond to stress by decreased appetite (26, 27), whereas women are more susceptible to engage in emotional eating behavior as a function of their menstrual cycle and hormonal phase (28) as well as a possible gene-depression interaction present only in women (29).

Over many years, emotions and mood have been shown to influence food choice. This observation was primarily explained via sensory, physiologic, and psychological pathways (30). For instance, eating a meal may reduce arousal and irritability and increase calmness and positive affect, whereas a small meal portion size may have opposing effects on mood (30). Sweetness and fatty structure can also improve mood by activating brain opioidergic, dopaminergic, and benzodiazepine/γ-amino butyric acid neurotransmission (30). In fact, dopamine neurotransmission reflects the “wanting” motivational aspect of eating, whereas opioid and benzodiazepine/γ-aminobutyric acid are linked to the “liking” hedonic aspect of food-related stimuli (31).

A distinction is generally made between the acute effects of food on mood and its chronic influence on human neuropsychology. For instance, after fasting overnight, individuals who consumed a high-carbohydrate diet (i.e., less fat and protein) experienced a surge in the stress hormone cortisol (32). Conversely, if those individuals consume a carbohydrate-rich diet over a period of 10 d, a lower average plasma cortisol is often observed compared with that with a high-protein diet (33). This phenomenon was explained by an adaptive response in the hypothalamic-pituitary-adrenal axis function and was replicated in animal studies (34). Indeed, when rodents were given a choice between food types, they tended to reduce their intake of low-fat, low-sugar feed pellets after exposure to stressors, given that sweetened milk, or high-fat feed pellets, thought to be more palatable, was available (35–42). The underlying mechanism for this eating behavior is that glucocorticoids in rodents, including cortisol, can trigger motivation to consume high-fat and high-sugar comfort foods (43–45), an effect mediated in part by corticotrophin-releasing factor in the central nervous system (46, 47). This effect may also result from the interplay between cortisol and insulin’s relative concentrations (9, 44, 45, 48, 49).

Evidence from human studies is accumulating slowly for a relation between cortisol concentrations and dietary intakes. For instance, key findings of a recent large cross-sectional study included associations between cortisol and perceived stress in nondiabetics only and a significant association of cortisol with saturated fat intake overall and with sweet foods in diabetic participants (8). Another recent retrospective study found that urinary free 24-h cortisol concentrations were positively correlated with daily carbohydrate and fat intake as well as with weekly consumption of starchy foods among obese/overweight women, even after adjustment for BMI (50).

The evidence discussed so far argues for a directionality of the association between mood and food intake in which depressive symptoms, perceived stress, cortisol, and emotional eating affect dietary behavior. However, reverse causality is also commonly examined, whereby certain nutrients (e.g., folate and other B-vitamins, n–3 FAs, vitamin D, antioxidants) are thought to affect mood, with specific underlying biologic mechanisms (51–56). Thus, future studies should examine those relations longitudinally by assessing change over time in either mood or dietary intakes in relation to baseline exposure. Moreover, mediation and moderation should be tested simultaneously between mood states and emotional eating in relation to dietary intakes. Finally, the role of cortisol in mediating the effects of emotional eating and/or depressive symptoms on dietary intake should also be examined in longitudinal studies.