Introduction

When we consider reasons for overweight and obesity, the early environment is particularly important to address [1]. Eating patterns are often potentially established in childhood and contribute to epigenetic changes and subsequent weight difficulty development [2, 3]. Research has reliably demonstrated associations between childhood maltreatment and body mass index (BMI) [4,5,6]. Stressful and traumatic childhood abuse experiences highly increase the risk of adulthood overweight [7,8,9,10]. The association between childhood traumatic experiences and adulthood obesity has been confirmed [11]. While childhood sexual and physical abuse were also hypothesized to be risk factors in multifactorial models of obesity, the role of emotional abuse gradually gained more attention. Childhood emotional abuse means a sustained, repetitive, inappropriate emotional response to the child’s experience of emotion. Among all kinds of childhood trauma, emotional abuse is highly prevalent and easily occurs since inappropriate emotional responses are instantaneous, less effort is spent by abusers, and the consequence of emotional hurt is insidious and difficult to detect [12]. Childhood emotional abuse is central to understanding the latent effects of child maltreatment, and its potential importance in the etiology of obesity needs further investigation [13].

One possible mediating factor in the relationship between childhood emotional abuse and adult obesity is anxiety. Childhood emotional abuse has a long-term effect on psychiatric performance [14,15,16,17]. Specifically, childhood emotional abuse is particularly relevant to the development of anxiety and depression [5, 18,19,20]. Furthermore, the study showed that obese people are worse in indicators of happiness, perceived mental health, life satisfaction, positive affect, negative affect, optimism, feeling loved and cared for, and depression [21]. Obesity may have long-term implications for mental distress at a clinical level over the adult years [22]. Some studies have shown that obesity is associated with an increase in lifetime anxiety disorders [23, 24]. A meta-analysis review of cross-sectional studies confirmed the association between obesity and anxiety and concluded that obesity was also associated with past-year and lifetime anxiety prevalence [25].

Another possible mediating factor between childhood emotional abuse and obesity is bulimia, which is an eating disorder behavior indicated by a tendency to eat a large amount of food in a short time [26]. Bulimia has a similar meaning to bingeing describing a tendency of excessive or uncontrolled indulgence, especially in food or drink. It is also considered a key symptom of the eating disorder. The literature consistently suggests a close association between bulimia and obesity [2, 27,28,29]. Bulimia is the most shared direct risk factor for obesity. Distressing psychological states such as anxiety and depression are also likely to increase indulgent food intake, frequent emotional overeating, and bulimia, which are unhealthy eating behaviors that contribute to high rates of obesity [30]. For example, a study has shown that greater attachment anxiety is predictive of a heavier body mass index [31]. A positive correlation has been examined between social anxiety disorder and binge eating frequency [32, 33]. In weight-loss surgery candidates, higher attachment anxiety is associated with a greater incidence of bulimia [34]. In addition, a systematic literature review indicated that some studies demonstrate an association between depression and binge eating disorder, but carefully designed studies are required [35]. While many studies have suggested a negative emotional effect on bulimia, the role of anxiety may be more important for future research [36]. Therefore, we hypothesize that anxiety is a more important mediating factor in the present study.

The prevailing view is that the relationships between anxiety, overeating, and body mass index can be explained in terms of emotion regulation [37]. The emotion regulation system connects to eating behavior by balancing different mental dimensions. Due to early adverse emotional experiences, individuals tend to be hyperactivated to potentially upsetting/stressful negative social cues. They are relatively poor at managing their emotions and thus more likely to be anxious [38, 39]. Therefore, to ‘soothe’ themselves, some anxious individuals rely on external sources of affect regulation such as food, while others may choose to rely on smoking, substance misuse, etc. [40]. Studies have proven that anxiety is specifically related to emotional eating among weight-loss surgery candidates [41]. Obesity seems to involve higher emotional dysregulation than normal weight conditions [42]. Emotion regulation is essential in the relationship between anxiety and bulimia, as it could represent a risk factor for the worsening of problems related to overeating and excessive body weight [34]. Thus, the importance of the emotion regulation process has been considered in the success of weight-loss treatment and could provide significant clinical information and therefore be part of the obesity diagnostic criteria and therapeutic program [43].

The etiology and maintenance of obesity have been substantially advanced. Based on the preproved potential connections between childhood trauma, adverse mood state, and bulimia, we hypothesize that there are explicit multiple mediation models linking all the possible variables of childhood emotional abuse, anxiety/depression, and bulimia and obesity (body mass index). Here, we sampled some weight-loss surgery candidates at hospitals and well-matched healthy controls to establish multiple mediation models to (1) add an approval of the association between childhood emotional abuse and obesity; (2) reveal the potential pathway between childhood emotional abuse and obesity, in which anxiety/depression could serve as intermediate factors; and (3) compare the model fit to test whether anxiety/depression play an equal main mediating effect in this relationship. The multiple mediation models will help clinicians continue to disentangle interactions of these factors to further facilitate our understanding of eating psychopathology.

Method

Participants

From September 2020 to January 2021, obese patients who were going to have weight-loss surgery at the Department of Bariatric & Metabolic Surgery, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, were informed about the present study. Inclusion criteria for the obesity are (1) meet the indication of bariatric surgery: BMI ≥ 32.5 or BMI between 32.5 to 27.5 with comorbidity of metabolic syndromes; (2) aged above 18 years old; (3) able to read and understand the description of each item of the questionnaire; and (4) voluntarily participated in the survey and signed the informed consent form were invited to participate in the present study to answer a set of clinical scales. Department of Bariatric & Metabolic Surgery, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital handed out healthy control recruitment advertisements in the nearby community. Citizens who were interested in participating in the research would contact the experimenters directly. Inclusion criteria for the healthy controls are (1) with normal figures and do not meet the indication of bariatric surgery; (2) without any eating disorders; (3) demographically matched with obese patients’ characteristics (with similar means of age, education years and similar gender ratio); (4) voluntarily participated in the survey and signed the informed consent. Both participants in obesity and healthy control were first interviewed by a professional psychiatrist. Participants examined with psychiatric disorders would be excluded from the research. The present study was approved by the Ethics Committee of Shanghai Sixth People’s Hospital, NO 2020–219-(1). All procedures followed the Declaration of Helsinki. Eventually, we analyzed the clinical data from 37 obese patients and 37 healthy people.

Clinical measurement

Basic demographic information (age, years of education, height, weight) and a series of clinical scales (Childhood Trauma Questionnaire, CTQ; Beck Anxiety Inventory, BAI; Beck Depression Inventory, BDI; and Eating Disorders Inventory, EDI) were collected. Body mass index (BMI, equal to weight (in kilograms) divided by height (in meters) squared) was calculated to describe the severity of obesity. All demographic information and clinical scales were presented to participants in the form of an online questionnaire designed by a professional psychologist. Participants answered all online questionnaires using their cell phones in the examination room with the supervision of experimenters. Demographic information on height and weight was self-reported by participants, experimenters would invite them to use the height and weight gauge in the examination room to measure these indexes when they are not sure.

The Childhood Trauma Questionnaire (CTQ) [44] is designed for adolescents and adults to obtain a brief, reliable, and valid assessment of traumatic experiences in childhood [45, 46]. It assesses the incidents of abuse and neglect in childhood, including physical abuse, emotional abuse, sexual abuse, emotional neglect, and physical neglect [47]. The total Cronbach’s α of the Chinese version of the CTQ is 0.73 [44]. The CTQ has 28 items, including a minimization–denial subscale of 3 items, and each item adopts a 5-point Likert score from 1 “never” to 5 “always” according to the frequency of the experiences that occurred [46]. Scores for each of the categories include 5 items, ranging from 5 to 25. A higher CTQ score indicates more severe childhood trauma. The total CTQ Cronbach’s α of the present sample is 0.623, and the Cronbach’s α values of the physical abuse, emotional abuse, sexual abuse, emotional neglect, and physical neglect subscales are 0.648, 0.793, 0.641, 0.746, and 0.462, respectively.

The Beck Anxiety Inventory [48, 49] assesses the severity of generalized anxiety symptoms [50]. It has good reliability, validity, internal consistency, and convergence [51, 52]. The total Cronbach’s α of the Chinese version of the BAI is 0.95 [49]. The BAI has 21 items, with each response based on a 4-point Likert scale ranging from 0 “not at all” to 3 “severely”. A higher score indicates greater anxiety severity. The total BAI Cronbach’s α of the present sample is 0.920.

The Beck Depression Inventory (BDI) [53] version 2 is a widely used clinical instrument to evaluate depression severity in normal populations [54,55,56]. It has good reliability and validity [55]. The total Cronbach’s α of the Chinese version of the BDI is 0.94 [53]. The BDI has 21 items, and each item consists of four self-evaluative statements scored from 0 to 3, with an increasing score indicating greater depression severity. The total BDI Cronbach’s α of the present sample is 0.901.

The Eating Disorder Inventory version 2 (EDI-2) measures eating disorder symptoms and the cognitive and behavioral characteristics of anorexia nervosa and bulimia [65,66,67].

Anxiety is thought to be an important mediator of emotional abuse in childhood and obesity in adulthood. Childhood emotional abuse, involving a repeated pattern of caregiver behavior or a serious incident, transmits negative information to the child that he or she is worthless, flawed, unloved, unwanted, endangered, or only of value in meeting another’s needs [68]. Spurning, intimidating and terrorizing, confining and isolating, exploiting and corrupting, denigrating emotional needs, and neglecting health needs manifest negative impacts on a child’s emotions and daily functionality and seriously undermine a child’s future adaptation [69,70,71]. Mechanisms linking emotional abuse with anxiety include maladaptive self-experience, such as with resilience and self-esteem. Psychological maltreatment reduces children’s psychological resilience, which is a positive resource adolescents can utilize to manage stressful challenges [72,73,74]. Emotional abuse causes low self-esteem, including negative evaluations of oneself [75, 76]. Undeveloped self-experience highly increases susceptibility to develo** anxiety and depression by causing a series of difficulties identifying emotions and emotional awareness and is more likely to induce anxiety in adulthood [77]. These findings may explain the evidence of a higher prevalence of lifetime diagnosed anxiety in obesity [78].

Bulimia symptoms are considered emotionally induced psychosomatic symptoms, and the tendency to bulimia in obese people is closely related to obesity. To further prove the connection sequence of anxiety and bulimia, we built and examined a model in which bulimia was the first multiple mediating variable and anxiety was the second multiple mediating variable (SFig. 1). The model fitting result was less satisfactory than the original model (anxiety was the first multiple mediator variable; see Supplementary material: the regression coefficient of emotional abuse → bulimia → anxiety/depression → BMI was not significant, STables 1 and 2). According to previous studies, anxiety disorders commonly have an onset in childhood and frequently exist before eating disorders [79, 80]. The model fitting and clinical evidence indicated that childhood emotional abuse might primarily lead to anxious traits. The recurring anxious emotional state triggers more bulimic behavior and further leads to obesity.

Depression is thought to be a mediating factor outside of anxiety between childhood emotional abuse and adult obesity. Our study described one indirect pathway of childhood emotional abuse contributing to obesity and demonstrated that anxiety plays an important mediating role in this relationship. This result provides a new perspective for treating obese patients with adverse early life events. Beyond bariatric surgery, psychological intervention is also helpful in reducing the influence of predisposing pathogenic factors. In future treatment, it would be beneficial to offer obese patients psychological therapy to reduce their anxiety and bulimic behavior. Anxiety and obesity are the two most common related health problems [81]. It is highly possible that anxiety disorders would lead to weight gain. For stressed individuals, the dysregulation of the hypothalamic–pituitary–adrenal axis contributes to subsequent getting weight [82, 83]. Symptoms of anxiety stimulate a craving for high-sugar and high-fat foods [84,85,86]. Anxiety-related chronic conditions might also have an influence on functional health, which may cause physical inactivity leading to excess weight. Previous studies also reported that anxiety is strongly associated with binge eating and emotional eating [33]. Obese patients eat more when they feel anxious, and the aroused effect is significantly reduced after gluttonous eating [87]. Therefore, anxiety is a critical factor in the childhood emotional abuse–obesity relationship, since the high likelihood of an anxious emotional state triggers bulimic behavior. In contrast, the connection between depression and bulimia is ambiguous. A depressive state does not always increase eating. In a sample of depressed patients, only 14% indicated an increase in appetite, while in 66%, appetite decreased, and in 20%, it showed no change [88]. Bradley M. Appelhans et al. reported that more severe depression is associated with more inferior diet quality [89]. For these reasons, we believe that anxiety plays a vital role in leading these obese patients to perform more bulimic behavior, which could release their anxious impulses but cause excessive fat accumulation.

Many studies have suggested that unhealthy eating habits, including overeating and bulimia, could be the result of a failure to attempt to regulate negative emotions [90,91,92]. Lack of emotion regulation could lead to a breakdown in the autoregulation of other personal areas, including those linked to the control of eating behavior [93, 94]. Emotion regulation is the ability to regulate one’s own positive or negative emotions to diminish, attenuate, maintain, or amplify their content [95]. For obese patients, when their emotions are dysregulated, maladaptive behaviors can be adopted to encourage them to overeat in response to anxiety [94, 96]. The emotionally driven eating model explains inappropriate eating behaviors by suggesting dysfunctional emotional and cognitive processing as causes of overeating [97]. The deficit in the regulation of eating behavior can be attributed to a failure in emotional regulation that can lead to overeating behavior to compensate for an inability to employ proper cognitive strategies to avoid a negative emotional state [92, 98]. Therefore, the ability to regulate anxiety or dysphoric mood is associated with binge eating and emotional eating in overweight individuals and has been considered a critical target to reduce excess body weight [33, 99, 100]. The clinical implications of the proposed multiple mediation models in this research strengthen the obesity treatment idea that improving obese patients’ emotion regulation ability is an effective target to rectify unhealthy bulimia behavior. Some pilot randomized controlled trial studies have achieved some therapeutic effects. For example, Berking and Whitley developed emotion regulation training (EuREKA), which is an innovative intervention program for children and adolescents that aims to examine the effectiveness of emotion regulation training when combined with a multidisciplinary obesity treatment in inpatient-treated 10- to 14-year-old youngsters [101, 102]. Obese youngsters of the EuREKA program exhibited less emotional eating behavior and improved weight loss and weight-loss maintenance, causally proving that emotion regulation intervention can be applied in clinical practice [103].

Some limitations of the present study should be noted. First, the obese participants recruited in the sample are patients seeking bariatric surgery in the hospital, which only presents a subpopulation of the obese. Compared to the obese, candidates for bariatric surgery display advantageous personality features and lower rates of psychopathology [104]. Second, its sample size is relatively small. Future studies need to collect a larger sample to make a firmer conclusion. Third, the present study is a cross-sectional investigation. All participants estimated their childhood experiences based on their retrospective memory. Longitudinal designs and interventional experiments should be adopted in future studies to reveal sequential causality. Finally, the data collection was based on the self-report questionnaire, which inevitably led to reported biases even though we strictly controlled the response quality. More objective indicators of neuroimaging are necessary. Childhood maltreatment reduces left-side hippocampal volumes [105,106,107] and the functional integrity of white matter tracts [108, 109]. Increased insula activation is involved in the neurological processing of food-related stimuli [110, 111]. Diminished frontostriatal activity contributes broadly to emotion, motivation, and movement processes and, importantly, is thought to underlie self-regulatory control [112]. More functional connections between related brain areas need to be confirmed to further substantiate the multiple mediation models of the present study. More evidence from random clinical trial research about emotional regulation training in obesity treatment will also make the current conclusion more stable.

In conclusion, obese patients experienced more childhood emotional abuse and were more anxious, depressive, and bulimic than healthy people. Childhood emotional abuse may contribute to adulthood obesity, potentially mediated by anxiety and bulimia. In obesity treatment, psychological interventions such as emotion regulation training would be helpful to reduce anxious emotions and thus decrease bulimic behavior.