INTRODUCTION

Binge eating disorder (BED) and bulimia nervosa (BN) are complex eating disorders (EDs) that share the criteria of recurrent binge eating episodes, which are characterized by the excessive consumption of food within a distinct, usually short, period of time, accompanied by a sense of loss of control. Inappropriate compensatory behaviors to prevent weight gain are present in BN but do not occur in BED¹.

BED and binge eating symptoms have been associated with overweight, obesity and various metabolic disorders such as diabetes, hypertension, hypercholesterolemia, and hypertriglyceridemia^2-6. Moreover, BN has been associated with multiple ill-health conditions such as renal and electrolyte disturbances⁷, cardiovascular diseases⁸, polycystic ovarium syndrome⁹and, like other EDs, higher risk of premature mortality¹⁰. Despite of the subject relevance, to our knowledge, only one Brazilian population-based study, conducted in Rio de Janeiro, assessed the frequency and correlates of EDs; the estimated lifetime prevalence of DSM-5 BN, BED and recurrent binge eating episodes were, respectively, 0,7%, 1,4%, and 6,2%¹¹.

EDs are frequently associated with other psychopathology and role impairment, yet often remain undiagnosed and under-treated¹². Reports from the World Health Organization World Mental Health Surveys show lifetime prevalence estimates of 1.0% for BN and 1.9% for BED across surveys in 14 countries in 4 continents¹³. Associations of BN and BED with equivalently high rates of current overweight and obesity, and high comorbidity with other DSM-IV/CIDI disorders, contrasting with low proportions of people with EDs that actually receive treatment, were also reported¹³.

Given the negative consequences of EDs and related comorbidities, there is a clear need to promote their prevention, early identification and prompt treatment. In this context, research on the epidemiology of eating disorders can provide a foundation for planning and strengthening public health policies. Furthermore, research addressing population-based samples of urban areas, known to be more affected by metabolic conditions, can offer greater insight into the processes of illness associated with binge eating conditions.

In the light of these considerations, the present study aims to investigate the lifetime, 12-months and 30-day prevalence estimates of BN, BED and any binge eating disorders or behaviors (ABEDB), as well as their distribution according to sociodemographic characteristics, and the comorbidity of these EDs with chronic conditions in a representative sample of the adult population (aged 18 and older) residing in the São Paulo Metropolitan Area, analyzing data from the São Paulo Megacity Mental Health Survey¹⁴.

METHODS

RESULTS

Prevalence estimates of eating disorders and behaviors according to sociodemographic variables are reported in Table 1. Overall lifetime, 12-month and 30-day prevalence estimates were, respectively, 2.0% (SE 0,3), 0.9% (SE 0.2) and 0.4% (SE 0.1) for BN, 4,7% (SE 0,3),1.8% (SE 0.3) and 1.2% (SE 0.2) for BED and 9.0% (SE 0.5), 3.4% (SE 0.4) and 2.0% (0.3) for ABEDB. Women presented significantly higher rates of all eating disorders compared to men, with the exception of 12-month BN. Although prevalences were systematically higher among younger cohorts for all eating conditions, differences in age categories were statistically significant only for lifetime ABEDB. No significant differences in prevalences were observed for marital status, education and income categories (Table 1).

Prevalence estimates of eating disorders and behaviors according to chronic diseases and conditions are reported in Table 2. Higher prevalence estimates of BEDB were seen among those with most chronic conditions, although not all differential distributions showed statistical significance. Significantly higher prevalences of lifetime BN were seen in those with gastrointestinal ulcer (8.2%%; SE 2.8), headaches (3.5% SE 0.6), neck or back problems (3.3%; SE 0.6) and hypertension (3.1%, SE 0.7), compared to those without these conditions. For lifetime BED, was higher among individuals that had chronic pulmonary disease (13.2%; SE 6.0), arthritis or rheumatism (10.0%; SE 2.1), gastrointestinal ulcer (9.2%; SE 2.6), neck or back problems (8.9%, SE 1.3), chronic pain (7.7%, SE 1.2) and headaches (6.6%, SE 0.9). Finally, lifetime prevalence of ABEDB was higher among those who reported gastrointestinal ulcer (18.1%; SE 3.1), neck or back problems (14.0%; SE 1.9), headaches (12.8%; SE 1.2) and chronic pain in other sites (11.7%, SE 1.5).

Significantly higher prevalence estimates of binge eating disorders and behaviors were observed in individuals of higher BMI categories, with the exception of 30-day BN (Table 2).

Poisson regression models of BEDB adjusted for sex and age are described in Table 3. Significant associations of BEDB and several chronic conditions were observed, as well as with overweight and obesity. PRs of all BEDB were systematically higher among individuals with gastrointestinal ulcer, with Prs ranging from 2.2 to 8.4. The PRs of lifetime, 12-month and 30-day BED and ABEDB were significantly higher among individuals with obesity, and in individuals that reported 12-month arthritis or rheumatism, neck or back problems, other chronic pain. Lifetime and 12-month BN were also associated with obesity and neck or back problems. The PRs of lifetime, 12-month and 30-day BN and ABEDB, as well as of lifetime BED, were significantly higher among overweight individuals. Lifetime, 12-month and 30-day ABEDB were also associated with a medical diagnosis of diabetes or glucose intolerance, although BN and BED individually were not (except for 12-month and 30-day BED in the crude model). Lifetime and 12-month BN and ABEDB were associated with the diagnosis of hypertension, while ABEDB and BED appear significantly associated with heart disease. Only 30-day BN was associated with lifetime heart attack. Lifetime BED was also associated with the diagnosis of chronic pulmonary disease. Lifetime BN and ABEDB, as well as 30-day ABEDB was associated with frequent or intense headaches.

DISCUSSION

This is the first epidemiological study estimating the prevalence of binge eating disorders and behaviors in the general population of the largest Metropolitan Area in South America (9.0%; SE 0.52). It, further, examined their associations with chronic conditions and nutrition status. The estimated lifetime prevalences of BN and BED in this Brazilian sample (2.0% and 4.7%, respectively) were considerably higher than global averages from the WHO World Mental Health Surveys, which reported 1.0% for BN and 1.9% for BED across surveys in multiple countries¹³. Disorganized urbanization in addition to significant socio-economic inequalities may help explain the high frequency of EDs reported in Latin America, and especially in metropolitan areas^20,21 A more recent systematic review of international studies reported lifetime prevalence estimates of 0.57% for DSM-IV BN and 1.72% for DSM-IV BED, and 1.41% for DSM-5 BN and 1.04% for DSM-5 BED²², also lower than the São Paulo estimates reported here.

Higher prevalence estimates of EDs among women found in this study have been widely documented^23-25.The 12-month prevalence estimates of 0.9% and 1.8% for DSM-IV BN and BED, respectively, were similar to those reported in a more recent study in another metropolitan area in the southeast of Brazil, that used DSM-5 criteria and estimated point prevalences of 0.7% for BN and 1.4% for BED, taking in consideration the differences between DSM-IV and DSM-5 criteria for BN and BED in the frequency of recurrent binge eating symptoms¹¹. Few prevalence studies on eating disorders have been conducted in Brazil, indicating the need for further researches, especially including rural samples²⁶.

Consistent with our findings, while using different methods, the study in Rio de Janeiro also reported associations of BED with several medical conditions, such as arthritis/rheumatism, spine problems and chronic muscle pain, whilst BN was also associated with hypertension and headaches¹¹. In our study, however, ABEDB were associated with diabetes or glucose intolerance in the adjusted model, but not BN or BED individually, although the association of BED and diabetes has been reported by other authors using both different and similar methods^11,13,27.

Results were mostly convergent with global estimates in population-based samples, as BN and BED have been associated with musculoskeletal conditions, other chronic pain conditions, diabetes, hypertension and ulcers in multiple countries¹³.

The associations of most BEDB with overweight and all but one with obesity were significant and corroborate findings from other studies²⁸. Binge eating behavior has been prospectively associated with subsequent overweight/obesity in cohort studies, with OR 1.73 (95%CI 1.11-2.69)²⁹ and OR 1.9 (95% CI 1.0-3.5)³⁰, respectively, while lifetime BED was associated with severe obesity in cross-sectional analysis¹². The Saudi Arabian National Mental Health Survey reported that respondents with overweight or obesity were more likely to be diagnosed with BED, compared to underweight individuals or those with normal weight, and overweight individuals were more likely to be diagnosed with BN, compared to those in the underweight, normal, or obesity BMI categories³¹. Their findings are in accordance with ours, although we also reported a higher prevalence of BN among those with obesity compared to normal weight individuals. Besides, the Saudi Arabia study revealed that people with overweight or obesity were significantly more likely to be diagnosed with BED in the preceding 12 months, reinforcing the bidirectional character of the associations^13,31.

The strong association of 30-day BN with reported history of heart attacks found in this study corroborates findings from the WHO World Mental Health Surveys, where BN was associated with subsequent heart attack¹³. The association of BN with ischemic heart disease has been described in a Canadian longitudinal cohort study of women, as well as the association with other cardiovascular conditions⁸. Damage done to the heart from BN may be attributed to the tendency to purge and the electrolyte imbalance caused by excessive vomiting and inappropriate use of medications³⁵.

Even though this study showed significantly higher prevalences of BN and ABEDB, but not BED, among individuals with a diagnosis of hypertension, BED has been associated with hypertension in other investigations^11,13,36. However, in a sensitive analysis adjusting by BMI, the associations with hypertension found in our study lost statistical significance. On the other hand, the association of heart disease with both ABEDB and BED reported herein remained significant when adjusted by BMI, suggesting that binge eating behavior is associated to heart disease regardless of the presence of obesity or overweight. This may be linked to the frequent consumption of ultra-processed and hyperpalatable food during binge episodes, reported in Brazilian and other population samples around the globe^37-39.

Although the diagnosis of diabetes or glucose intolerance was associated with ABEDB in all time frames, that association only remained significant for 30-day ABEDB when adjusted by BMI. This implies high BMI as a mediator factor in the association between binge eating behavior and diabetes, explained by the insulin resistance caused by chronic inflammation, especially in visceral obesity^{40, 41}.

Results showed comorbidity of chronic pain with BED and ABEDB, and while significantly higher prevalence estimates of lifetime BN and ABEDB, but not BED, were found among those who reported frequent or intense headaches, other studies have reported associations of BED with headaches, as well as with chronic pain^11,27,36.

In addition, the significantly higher prevalence of lifetime BED among those with chronic pulmonary disease may be related to overlapping neurobiological and behavioral patterns displayed by individuals with food addiction or BEDB and smoking/vaping habits^42,43.

It is worth acknowledging that associations of BEDB with arthritis or rheumatism, neck or back problems, other chronic pain and gastrointestinal ulcer found here remained significant when adjusted for BMI (data not shown).

Results from the latest Vigitel³², a Brazilian serial health survey carried out by the Ministry of Health annually by telephone contact in state capitals and the Federal District, showed high percentages of adults with excess weight (BMI ≥25: 61.4%), including overweight (37.1%) and obesity (24.3%), with similar findings in the city of São Paulo (63.0%, 38.7% and 24.3%, respectively)³³. However, these indicators are much larger than those observed in 2006, where the overall Brazilian rates of excess weight was 42.6% and obesity was 11,8%, and, for São Paulo state capital, rates were 44.3% and 11,0%, respectively³⁴, which are slightly lower than those found in this study from the São Paulo Metropolitan area (47,1% and 13,8%, respectively). This expressive national increase (1.03% and 0.69% per year, respectively)³⁴, reinforces the urgent need for public healthcare policies regarding the prevention and control of this steadily growing epidemic and its related conditions.

The results reported in this paper should be interpreted taking into account several limitations. As a cross-sectional study, the direction of associations was not ascertained; although the age of onset of all eating disorders was enquired, we did not know when the chronic conditions assessed started, nor the history of weight gain. Further, weight and height measures used to calculate BMI were self-reported. Individuals with eating disorders may be more likely to misestimate their weight, as this may be a sensitive issue for them, and frequently have a distorted body image^44,45. Indeed, an overestimation of self-reported height and underestimation of weight has been found in this sample, in general, resulting in a systematic reduction of BMI⁴⁶. Nevertheless, if this was the case, the association of BEDB with obesity and overweight would be diminished in our study, increasing type I error. Regarding clinical diagnoses, the concordance of the WMH-CIDI with the Structured Clinical Interview for DSM-IV (SCID)⁴⁷ was found to be unbiased for 12-month WMH-CIDI prevalence estimates, but generally conservative for lifetime disorders⁴⁸; however, no assessment of the consistency of WMH-CIDI eating disorders diagnoses was conducted. Furthermore, embarrassing behaviors or emotional contents, such as those related to binge eating and purging, and psychological suffering, are likely to be underreported, especially in cross-sectional assessments, and even more so in non-clinical interviews conducted within the respondent’s household⁴⁹. Recall bias may also impair the accuracy of retrospective information, although the questionnaire was rebuilt taking into consideration a series of strategies to minimize this type of error¹⁵. Moreover, people with mental problems are known to be more reluctant to participate in epidemiological surveys of this sort⁵⁰. It is not possible, however, to evaluate this hypothesis and determine its impact, as information on non-responders (28.7%) was limited to the household listing (age, gender, and family relation to household informant). All these plausible biases were, therefore, likely to have yielded conservative rates of psychiatric morbidity, according to the adopted system of classification. Finally, sample sizes for some crosstabulations were too small, reducing the statistical power to identify significant associations on disorders with low prevalence, such as BN, 30-day diagnoses and less common chronic conditions.

Despite of these limitations, the findings reported herein represent an important contribution to the understanding of the prevalence and distribution of binge eating disorders and behaviors and their frequent comorbidity with chronic conditions in the Brazilian population, offering insights on the public health and clinical implications which may be helpful for guiding public policies and clinical interventions. We hope this sheds a light on the importance of a multidisciplinary approach in the prevention, screening and treatment of these conditions, as to minimize the burden and functional impairment in the general population.

CONCLUSION

The prevalence of BEDB was systematically higher among women, and comorbidity with multiple chronic conditions was commonly observed. Significant associations with high BMI and multiple chronic conditions, including metabolic and cardiovascular diseases, and other conditions not directly related with obesity, such as chronic pain and gastric ulcers. Other studies may unravel more aspects of those associations, and investigate the comorbidity of EDs with other conditions, such as mental disorders, especially in Brazil, where few population-based studies on this topic have been conducted.