INTRODUCTION

Binge-eating disorder (BED) is defined by recurrent episodes of binge eating, characterised by eating an amount of food that is definitely larger than what most people would eat, in a discrete period of time (usually less than 2 hours), under the same circumstances, accompanied with a sense of lack of control (^{American Psychiatric Association 2013}; ^{Hilbert 2019}). Life-time prevalence of BED varies from 1 to 4%, with higher rates seen on women, though this gender bias is not as pronounced as in other eating disorders. Additionally, prevalence seems to be higher in adults rather than in adolescents (^{Davis 2015}; ^{Keski-Rahkonen & Mustelin, 2016}; Lindvall ^{Dahlgren & Wisting, 2017}). Hereditability accounts for 41% to 57% of BED etiology (^{Mayhew et al., 2018}), and other risk factors include childhood overweight, dieting, body dissatisfaction, negative affect, family disfunction and personality traits such as perfectionism (^{Bakalar et al., 2015}).

BED is associated with multiple medical comorbidities beyond obesity, such as metabolic syndrome, sleep disorders, functional gastrointestinal symptoms, asthma, among others (^{Mitchell 2016}; ^{Olguin et al., 2017}). BED is also associated with important psychiatric disorders and some of their characteristic psychopathology, of which the most prevalent are anxiety, depression, substance abuse and borderline personality disorders (Lindvall ^{Dahlgren & Wisting, 2017}). Health-related quality of life of BED patients is significantly worse and is linked to greater healthcare costs (^{Ágh et al., 2016}).

Negative emotions and maladaptive emotion regulation, along with body dissatisfaction, overconcern with weight and shape, grater levels of food craving, eating as a coping strategy, and eating for purposes of reward enhancement play a nuclear role on BED psychopathology (^{Dingemans & Danner, 2017}; ^{Lewer et al., 2017}; ^{Leslie et al., 2018}). In fact, negative emotions like depressive an anxiety symptoms, sadness, anger and frustration (specially in an interpersonal context (^{Truglia et al., 2006}) seem to be linked to BED, and their severity can be related with frequency and severity of binge behaviour (^{Arnow et al., 1992}; Dingemans & Danner, 2017; ^{Rosenbaum & White, 2013}). The importance of negative emotions as a trigger for binge episodes, and binge behaviour as a down-regulator of these negative emotions, serving as a tool to regulate affect, are components common to many BED models and theories (Dingemans & Danner, 2017; ^{Kittel et al., 2015}).

Anxiety, and particularly anxiety sensitivity (a cognitive style in which physiological anxiety symptoms are perceived as dangerous), seem to have an important role in negative emotions and their regulation (^{Taylor, 1995}; ^{Morillo Baro et al., 2016}; ^{Reigal Garrido et al., 2018}). Some have even proposed that binge eating behaviours may serve as maladaptive strategies to regulate anxiety, much like nail-biting (^{Rosenbaum & White, 2013}; ^{Tanaka et al., 2008}). Approximately 50-60% of patients with binge eating disorder do not respond to the conventional treatments available (Linardon & Wade, 2018; Wilson, Grilo, & Vitousek, 2007). Thus, adaptive strategies to cope with stress and anxiety, like physical activity (PA), may be important in reducing binge eating episodes, BED symptoms and comorbidities ^{Deboer et al., 2012}; ^{Blanchet et al., 2018}; ^{Vancampfort et al., 2014}).

In a recent systematic review (^{Blanchet et al., 2018}), it was observed that PA has a role in reducing BED episodes, improving anxiety and physical health. However, there is still evidence about this relationship in the literature. Therefore, further research is needed to clarify this issue. In this sense, our aim is to examine the relationship between PA and anxiety in patients with BED. Through a regression analysis, we expect to show that PA may improve BED symptoms through 3 mechanisms: 1) improvement in the food reward system (through biochemical pathways comparable to those used in substance abuse), reducing cravings; 2) down-regulation of negative affect (through PA effects on neurotransmitters such as serotonin and endorphin); 3) and improvement in appetite control, an well-known characteristic of PA. Correlations between anxiety-BED and anxiety-PA are established, but there is a need of more data on the role of PA in anxiety in BED patients.

MATERIAL AND METHODS

The present study is characterized by a cross-sectional study, in which exposure to the factor or cause is present to the effect at the same time or time interval analyzed. They apply to investigations of effects for causes that are permanent, or for factors dependent on permanent characteristics of individuals, such as the effect of PA on a particular disease (^{Campana et al., 2001}). These studies describe a situation or phenomenon at an undefined moment, only represented by the presence of a disease or disorder, such as, for example, changes in symptoms in people with diabetes (^{Haddad, 2004}). Thus, there is no need to know the time of exposure of a cause to generate the effect, the cross-sectional model is used when exposure is relatively constant over time and the effect (or disease) is chronic. Therefore, this model presents itself as a photograph or snapshot taken in a population by means of sampling, examining the presence or absence of exposure and the presence or absence of the effect (or disease) in the members of the sample (^{Fletcher, 2003}).

RESULTS

Participated of this study 32 patients (27 F and 5 M), age between 18-65 years old, with BMI varying between 27.05 and 56.69. Thirteen of them revealed being active while 19 revealed being sedentary. In addition, taking into account anxiety, 5 from 32 patients did not present anxiety and 27 presented some level of anxiety (for further details, please see table 1).

Table 1.  Descriptive data of sample. 

Table 2 shows the descriptive statistics of the values obtained by the subjects on the HADS, EDEQ and IPAQ. We also indicate the values of internal consistency, evaluated with the Cronbach's alpha coefficient. The values are all higher than the minimum of 0.60.

Table 2.  Average values of HADS, EDE-Q and IPAQ (N=32). 

Note: METs: metabolic equivalent task.

The correlation coefficients between the dimensions of EDE-Q and anxiety are almost all significant (except restriction), positive and moderate. Thus, the higher the eating disorder values, the higher the anxiety levels (see table 3).

Table 3.  Correlation between Anxiety and EDE-Q dimensions (N=32). 

^*p<0.05

^**p<0.01

^***p<0.001

To check the factors that correlate with the anxiety levels, and to control the variables, we used a linear regression where the dimensions of EDE-Q, IPAQ were used as independent variables. The association is positive between anxiety (i.e., scale indicates that the higher the score, the higher the present anxiety) and the lack of PA (nominal variable 1-exercise, 2-no exercise). The same is true with concern about weight. Thus, the direction of variation in anxiety and lack of exercise is the same, so a positive association between the two variables means that the less exercise and the more worry about the body, the more anxiety. It was concluded that 29.7% of the anxiety variance is correlated with a model composed by PA and concern with weight. When R² was adjusted, we found 43.1% of the anxiety variance correlated with the model PA and concern weight (see table 4).

Table 4.  Linear regression for Anxiety, IPAQ and EDE-Q (N=32). 

^*p<0.05

^**p<0.01

R² adj: R² adjusted; B: unstandardized coefficient; Beta: standardized coefficient; S: standard error; VIF: variance inflation factor; DW: Durbin-Watson

DISCUSSION

The aim of our study is to verify if there is relationship between PA and anxiety in BED patients, contributing to the existing evidence of PA as a therapeutic tool for this disorder. It is hypothesized that PA may improve BED symptoms through 3 mechanisms: 1) improvement in the food reward system (through biochemical pathways comparable to those used in substance abuse), reducing cravings; 2) down-regulation of negative affect (through PA effects on neurotransmitters such as serotonin and endorphin); 3) and improvement in appetite control, an well-known characteristic of PA. Correlations between anxiety-BED and anxiety-PA are established, but there is a need of more data on the role of PA in anxiety in BED patients.

Data were assessed between levels of PA, anxiety, and BED, corroborating previous studies (^{Hrabosky et al., 2007}). Additionally, the term supported the anxiolytic effect of PA in patients with disorders, as well as different populations (^{Bartley et al., 2013}; ^{Broocks et al., 1998}; ^{Blumenthal et al., 1999}; ^{Karssemeijer et al., 2017}; ^{Wedekind et al., 2010}; ^{Wegner et al., 2014}; ^{Tajik et al., 2017}; ^{Stubbs et al., 2017}; ^{Hiles et al., 2017}).

From this perspective, previous data from ^{Levine et al (1996}) showed that PA may be an important aspect in the treatment of BED. In accordance with this study, the systematic review by de ^{Blanchet et al (2018}) found benefits of PA in treating patients with BED, showing decrease in episodes and other comorbidities correlated with BED, such as weight concern, which was one of the symptoms that related to the levels of PA in our findings.

In contrast, the study by ^{Vancampfort et al. (2014}) found that muscle fatigue and pain after 6 minutes walking test (MWT) were significantly higher in obese people with Binge compared with the obese and nonobese control group. Thus, the authors mentioned that people with BED face barriers that make it difficult to adhere to the practice of PA, which may explain the low levels of PA found in this population.

Above all, qualitative studies also highlight the combined therapy between exercise and diet as a way to generate a new perception of Binge patients on diet and PA (^{Bakland et al., 2019}; ^{Pettersen et al., 2017}), suggesting that the positive experience with PA can positively influence the compensatory behavior of these individuals. In physiological terms, the review by ^{King et al. (2011)} points out that PA causes mechanisms that may contribute to appetite control, such as gastric emptying, increase of appetite-related peptides, regulation of substrate oxidation resulting from hormonal regulation and neural.

In addition, a more recent review by ^{Blundell et al. (2015}) found that exercise promotes regulation of energy demand, control of eating, increased resting metabolic rate and muscle mass, which is associated with decreased leptin and other adipocytes and, consequently, body fat. In contrast, the meta-analysis by ^{Lavagnino et al. (2016}) found that the increase in BMI is related to a decrease in activity of inferior frontal gyrus, ventromedial prefrontal cortex and dorsolateral prefrontal cortex, which are associated with inhibitory control. These data also showed that comparative studies found reduced activation in these areas in obese participants with BED compared to obese participants without BED.

PRACTICAL APPLICATIONS

Our findings support that PA correlates with anxiety in patients with BED, suggesting that PA may decrease symptoms, such as the appetite, weight and body shape concerns. From the practical point of view, the results emphasize the importance of the trainer or gymnasium staff promoting the necessary support for the prescription of training, so that it favors the training adequacy and also of making decisions about the best intensity and type of exercises to be prescribed. In line with this, we advise the minimum recommendation proposed by the American College of Sports Medicine (ACSM), 150 min.week^-1of moderate aerobic activity or 60 min.week^-1of vigorous activity to promote health (^{ACSM, 2013}).

This work have limitations concerning the characteristics of the sample, so that the results may be influenced by the low number of participants, the heterogeneity of their demographic characteristics, and the wide age range of the participants, which are conceivable to limit our conclusions, and do not allow its extension to other groups of patients with eating disorders. Therefore, it is suggested to replicate the study with more homogeneous samples. Another limitation is its transversal naturalization, which only allows the adjustment of the model in a certain time cut, while it is presumed that longitudinal designs allow to draw solid conclusions about the causal relationships and are opportunities for our better understanding of the causal process of phenomena of interest.