Parts of this manuscript were presented as a paper presentation at the Annual Meeting of the Academy of Eating Disorders, Austin, TX, May 2012.
A comparison of women with child-adolescent versus adult onset binge eating: Results from the National Women's Study
Article first published online: 5 JUN 2014
© 2014 The Authors. International Journal of Eating Disorders Published by Wiley Periodicals, Inc.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
International Journal of Eating Disorders
Special Issue: Developmental Risk for Eating Disorders across the Lifespan
Volume 47, Issue 7, pages 836–843, November 2014
How to Cite
Brewerton, T. D., Rance, S. J., Dansky, B. S., O'Neil, P. M. and Kilpatrick, D. G. (2014), A comparison of women with child-adolescent versus adult onset binge eating: Results from the National Women's Study. Int. J. Eat. Disord., 47: 836–843. doi: 10.1002/eat.22309
- Issue published online: 27 OCT 2014
- Article first published online: 5 JUN 2014
- Manuscript Accepted: 18 MAY 2014
- Manuscript Revised: 16 MAY 2014
- Manuscript Received: 5 JAN 2014
- binge eating;
- bulimia nervosa;
- physical assault;
- substance abuse;
Studies of age of first binge have been conducted in clinical samples of patients with bulimia nervosa (BN) and binge eating disorder (BED), but few studies have examined age of first binge using nationally representative samples.
We examined age of first binge and its clinical correlates using data generated from the National Women's Study (n = 3,006). Participants who endorsed ever binge eating (n = 707) were divided into two groups: (1) child-adolescent onset (CO)—age of first binge <18 years, and (2) adult onset (AO)—age of first binge ≥18 years. We hypothesized that CO binge eating would be associated with greater (1) likelihood of developing BN/BED, (2) severity of BN/BED, (3) history of trauma and PTSD, and (4) history of psychiatric comorbidity, such as major depression and substance use.
Of those who ever endorsed binge eating, 212 reported CO (30%) and 495 (70%) reported AO. Although AO binge eating was more common, CO binge eating was associated with higher rates of lifetime BN, greater severity of bulimic symptoms, earlier age of first dieting; earlier age at highest weight, greater likelihood of ED treatment, and higher rates of molestation, physical assault, any direct victimization, lifetime PTSD, and substance abuse.
AO binge eating is more than twice as common as CO binge eating in women, but CO binge eating is associated with higher rates of lifetime BN, greater severity of BN, and higher rates of victimization, PTSD, and substance abuse. © 2014 Wiley Periodicals, Inc. (Int J Eat Disord 2014; 47:836–843)
A binge eating episode is characterized by the consumption of a large amount of food in one sitting accompanied by a subjective loss of control over the eating behavior. Clinically significant binge eating is associated with several DSM-5 defined eating disorders including bulimia nervosa (BN), binge eating disorder (BED), the binge-purge form of anorexia nervosa (AN-BP), and other specified feeding or eating disorder (OSFED). Additionally, binge eating is often associated with the development of obesity and associated medical comorbidity, especially when associated with BED.
Until the 1994 addition of BED in the appendix of the DSM-IV and its acceptance as a full-fledged eating disorder in DSM-5, binge eating has been primarily examined for its role in BN and AN-BP.[1, 3] Although it shares the binge eating component of BN, BED occurs in the absence of compensatory behaviors (i.e., vomiting, laxatives, excessive exercise). Although elevated when compared with normal controls, BED patients have been found to display lower rates of psychopathology, less body dissatisfaction, and less anxiety surrounding their binge eating episodes and subsequent weight gain than BN patients.[4-6] In group comparisons of BED and BN patients, BED patients report less severe eating disorder symptoms as determined by scores on the EDE-Q. These studies suggest that although the two disorders are similar in sharing a core binge eating component, patients diagnosed with BED have a different disorder, a less severe course, and a more positive prognosis than their BN peers.
A limited amount of research has been conducted on the age of onset of binge eating. Most of the existing literature has been done in clinical samples of subjects endorsing a BN or BED diagnosis. Earlier age of onset of BN has been reported to result in greater severity of the disorder as determined by greater binge and purge frequencies, greater body dissatisfaction ratings, and lower Global Assessment of Functioning Scale (GAFS) scores. These authors noted that early onset BN “may be particularly likely to develop in women who come from problematic rearing environments or for whom there is a strong genetic liability to psychiatric illness.” In each of two studies of patients with BED, those that binged first (prior to onset of dieting) were significantly younger at the age of onset of BED.[9, 10] In the study by Spurrell et al. that compared dieting and binge eating behaviors among BED patients, those that began binge eating prior to dieting had a significantly greater number of comorbid Axis I psychiatric disorders and significantly higher rates of substance use disorders and Axis II personality disorders than those who began dieting prior to binge eating. In another group of BED patients, early onset binge eaters were more likely than later-onset binge eaters to have an earlier onset of obesity and dieting, as well as more eating disorder related psychopathology. They were also more likely to report a lifetime history of BN and mood disorder. Reas and Grilo reported the BED patients who binged first were significantly younger at the age of onset of BED than those who dieted first or who became overweight first. Binford et al. showed that early onset binge eating (<13 years old) was associated with greater reductions in post-binge subjective anxiety. In the National Comorbidity Survey Replication Study BED subjects endorsed a lower average age of first binge than BN patients, but this study reported very few details about age of onset in relation to comorbidity. Binge eating behavior has been documented in children,[2, 14, 15] whereas documented cases of prepubertal BN are extremely rare. Although there have been numerous studies conducted with clinical samples of eating disorder patients, varying methodologies make generalizations about early onset binge eating difficult. In addition, there were many nonsignificant differences also noted in these studies. Nevertheless, most studies of early onset binge eating in clinical samples appear to be associated with greater severity of eating pathology and greater psychiatric comorbidity.
PTSD and trauma are well-known mediators of comorbidity in patients with and without eating disorders.[17, 18] The link between victimization and eating disorders was found to be especially strong in participants with BN participating in the National Women's Study. Women with BN were more likely to have reported rape, sexual molestation, and aggravated assault than women who did not have a diagnosis of BN. Lifetime PTSD and subthreshold or partial PTSD (pPTSD) have also been found to occur at an elevated rate in the BN and BED population.[2, 17-20] Subthreshold or pPTSD has been classified in different ways by various authors, but all definitions include the presence of significant PTSD symptoms in one or more of the three main clusters as defined in DSM-IV, i.e., reexperiencing, hyperarousal, and avoidance, but which do not meet the full criteria.[17, 20] PTSD and its symptoms have been consistently reported to contribute to Axis I and Axis II psychiatric comorbidity as well as eating disorder symptomatology.[17, 18, 21-23] Patients with bulimic disorders are well known to have elevated psychiatric comorbidity, and in one study of patients with BED it was found that those with a comorbid axis II disorder exhibited more severe binge eating behavior. Again, there is a limited amount of work published examining age of first binge and its relationship to trauma, PTSD, and psychiatric comorbidity in the general population.
The relationship between traumatic events and age of onset of binge eating has previously been examined in the context of adverse life events occurring prior to symptom onset of eating disorders.[25-28] There are also animal models that demonstrate early life stress can predispose to binge eating behavior. Generally, high rates of adverse life events during the year prior to the onset of an eating disorder have been demonstrated, especially with BN.[27, 28] Other investigators have also demonstrated the link between earlier adversity or maltreatment and later development of bulimic symptoms and/or BED.[29-32] Binge eating and purging behaviors have been hypothesized to be maladaptive affect regulation strategies in response to trauma.[33, 34] Furthermore, one prospective study linked stressful life events to likelihood of relapse in BED patients. The interrelated variables of trauma or adversity, early onset binge eating behavior and the development of BN or BED and their comorbidity constitute an area of research in the field of eating disorders that warrant further examination in nonclinical, representative samples.
This study used a nationally representative sample of women in the United States (US) to further examine the role of age of onset of binge eating and its correlates. Using a representative sample provides more credible results than a clinical sample because it reflects the characteristics of the population from which they are selected. No previous study has specifically examined age of onset of first binge as defined in DSM-IV in relationship to clinical variables using a nationally representative sample of women. We formulated several hypotheses based on existing literature: CO binge eating compared with AO binge eating would be associated with greater (1) likelihood of developing BN and BED, (2) severity of BN and BED, (3) history of trauma, PTSD and its symptoms, and (4) history of psychiatric comorbidity, such as major depression and substance use. In addition, we hypothesized that the age that binge eating first occurred would significantly correlate with the age that traumatic experiences first occurred, the age of first depressive symptoms, and the age of first use of substances of abuse. These statistics would theoretically provide more valuable information in regard to the chronological relationship between these associated behaviors.
Participants in this study were drawn from a national household probability sample of 3,006 women who were participants in the third wave of the National Women's Study (NWS). The weighted sample was predominantly Caucasian (86%) and had a mean (±SD) age of 46.1 ± 17.3 years. A detailed description of the subjects and methodology used in the NWS has been previously published.[19, 26]
Structured telephone interviews were conducted in which participants were screened for current and lifetime experiences with sexual and physical assault, major depression, trauma history, dieting behavior, BN, BED, alcohol abuse/dependence, PTSD, and demographics. All diagnoses were based on DSM-IV criteria. In addition, the presence or absence of binge eating was determined by a series of questions based on DSM-IV criteria for a binge. Those who endorsed ever binge eating were subsequently divided into two groups: (1) CO binge eating, in which age of first binge occurred prior to age 18, and (2) AO binge eating, in which age of first binge occurred at or after the age of 18.
Totally, 4,009 female adults were initially sampled in 1989 and constituted the first wave of the study. In the first stage of the sampling procedure, these participants were gathered using a geographic sampling procedure in which stratified samples of counties in four regions of the country were constructed as primary sampling units. In the second and third stages of the sampling procedure, a random digit dialing technique was used to gather participants from households within each strata of the primary sampling units. Within each household, the female adult with the most recent birthday was selected for interview.
In 1992, Schulma, Ronca, and Bucuvalas, a New York Based survey research firm employed experienced female interviewers to conduct Wave 3 interviews. It was a blind interviewing procedure in which the interviewers were not informed of the purpose of the study. The study had a high content and completion rate (∼75%) for the Wave 1 participants completing the Wave 3 interview. Attrition was mainly due to participant relocation as opposed to refusal to complete the Wave 3 interviewing process. The data were weighted according to estimates of 1992 United States Census figures for age and race to ensure that they were representative of women in the US population. Only Wave 3 respondents were used in this analysis given that Wave 3 is the only wave which contained questions about eating disorders.
The two groups were compared on a number of demographic and clinical variables using an unpaired t-test for parametric data and a chi-square (χ2) test for nonparametric data. The Pearson “r” was calculated for correlational analyses.
A summary of the results is shown in Table 1. Of the 3006 women interviewed, 707 (23.5%) endorsed binge eating at some point in their lives. Of these, 212 reported CO binge eating (30%) and 495 (70%) reported AO binge eating (see Fig. 1). Average age (±SD) of first binge for the two groups was 13.9 ± 2.8 years and 27.7 ± 11.0 years, respectively (p < .001, unpaired t-test). The CO group was significantly younger (29.3 ± 11.0 vs. 35.7 ± 13.3 years, p < .01, unpaired t-test) and less likely to be married (43% vs. 61%, p < .01, χ2) at the time of the interview than the AO group. In addition, the CO group was more likely to have a college or graduate degree (33%) than the AO group (20%, p > .001, χ2). There were no significant differences between the two groups in terms of race, employment or income level.
|CO||AO||t-Value or χ2||Significance|
|Any eating disorder diagnosis (%)||33.5||23.6||2.047||0.041a|
|BED Diagnosis (%)||8.5||7.1||.656||0.512|
|BN Diagnosis (%)||16.5||9.5||2.679||0.008b|
|Ever Purged (%)||61.8||56.6||1.750||0.081|
|Minimum BMI (kg/m2)||21.6 ± 4.2||21.2 ± 3.3||1.324||0.186|
|Maximum BMI (kg/m2)||29.4 ± 7.0||29.7 ± 6.9||−0.499||0.618|
|Current BMI (kg/m2)||26.2 ± 6.3||27.6 ± 6.0||−2.756||0.006b|
|Age of first diet (yrs)||17.0 ± 5.3||25 ± 10.3||−9.961||0.000c|
|Age at highest weight (yrs)||26.4 ± 7.8||33.5 ± 12.1||−7.635||0.000c|
|Age at lowest weight (yrs)||23.1 ± 5.5||23.3 ± 7.1||−0.258||0.797|
|Ever use excessive exercise (%)||34.9||24.4||2.864||0.004b|
|Ever use vomiting (%)||26.4||20.6||1.700||0.090|
|Ever use laxatives (%)||14.2||9.9||1.645||0.1|
|Ever use diuretics (%)||22.2||14.3||2.565||0.01a|
|# of purge types||0.63 ± 0.9||0.45 ± 0.72||2.823||0.005b|
|Worse binge > several weeks (%)||50.5||40.6||2.431||0.006b|
|Ever treated for an ED (%)||11.0||4.0||14.3||0.001c|
|Major lifetime depression (%)||32.8||27.3||1.439||0.151|
|Major depression current (%)||7.6||4.0||1.939||0.053|
|Age at 1st depression (yrs)||22.3 ± 13.4||32.3 ± 18.4||−3.655||0.000c|
|Ever smoked (%)||72.2||59.0||3.346||0.001c|
|Smoke currently (%)||34.4||26.7||2.089||0.037a|
|Alcohol abuse (%)||40.1||25.9||3.082||0.001b|
|Alcohol dependence (%)||20.3||11.5||3.082||0.002b|
|Age of 1st trauma (yrs)||15.4 + 7.8||19.7 + 12.3||−4.824||0.000c|
|Rape victim (%)||23.1||18.8||1.315||0.189|
|Molestation victim (%)||26.4||18.4||2.418||0.016a|
|Physical assault victim (%)||17.5||11.1||2.302||0.022a|
|Direct crime victim (%)||53.3||42.0||2.772||0.006|
|Accident victim (%)||12.1||6.9||2.210||0.000c|
|Disaster victim (%)||3.5||8.8||−2.415||0.000c|
|PTSD and PTSD symptoms|
|PTSD lifetime (%)||31.0||23.0||2.264||0.024a|
|PTSD current (%)||9.0||8.0||0.575||0.566|
|Past re-experiencing symptoms (#)||0.47 ± 0.8||0.4 ± 0.8||0.980||0.327|
|Current re-experiencing symptoms (#)||0.46 ± 0.9||0.34 ± 0.8||1.788||0.074|
|Past avoidance symptoms (#)||1.3 ± 1.6||0.95 ± 1.5||2.497||0.013a|
|Current avoidance symptoms (#)||0.9 ± 0.15||0.8 ± 0.14||0.867||0.386|
|Past arousal symptoms (#)||0.94 ± 0.13||0.66 ± 0.12||2.786||0.005b|
|Current arousal symptoms (#)||0.97 ± 0.14||0.84 ± 0.14||1.121||0.262|
|Amnesia lifetime (%)||21.2||15.0||2.047||0.041a|
|Amnesia current (%)||10.0||6.0||1.496||0.135|
CO binge eating was associated with a higher frequency of an eventual diagnosis of BN (16.5% vs. 9.5% (p < .01, χ2) and any eating disorder diagnosis (33.5% vs. 23.6%, p < .05, χ2) than AO binge eating. CO binge eating was also associated with a lower BMI at the time of the study (26.2 ± 6.3 kg/m2 vs. 27.6 ± 6.0 kg/m2 (p < .01, unpaired t-test), a younger age at first serious diet (17.0 ± 5.3 years vs. 25.0 ± 10.3 years, p < .001, unpaired t-test), and a younger age at which the participant attained their highest adult weight (26.4 ± 7.8 years vs. 33.5 ± 12.1 years, p < .001, unpaired t-test). The CO group was associated with using more types of purging (p < .005, unpaired t-test) and had a higher frequency of using excessive exercise (34.9% vs. 24.4%, p < .005, χ2) and diuretics (22.2% vs. 14.3%, p < .05, χ2) as attempted forms of weight control. CO binge eating was associated with greater severity of binge eating, i.e., at least several times per week or more (50.5% vs. 40.6%, p < .01, χ2). In addition, the CO binge eaters were more likely to receive treatment for an eating disorder (11%) than the AO group (4%, p < .001, χ2).
In regards to comorbid behaviors and conditions, CO binge eating was significantly associated with a younger age at first major depressive episode (22.3 y/o v. 32.3 y/o, p < .001, unpaired t-test). CO binge eating was additionally associated with more alcohol abuse (40.1% vs. 25.9%, p < .001, χ2), alcohol dependence (20.3% vs. 11.5%, p < .02, χ2), a history of smoking (72.2% vs. 59.0%, p = .001, χ2), and present smoking behavior (34.4% vs. 26.7%, p < .05, χ2).
CO binge eating was also associated with a higher frequency of trauma endorsements. This group was more likely to have experienced molestation (26.4% vs. 18.4%, p < .05, χ2), physical assault (17.5% vs. 11.1%, p < .05, χ2), any direct crime victimization experience (53.3% vs. 42.0%, p < .01, χ2), and accidents (12.1% vs. 6.9%, p < .001, χ2). In contrast, AO binge eating was associated with a higher frequency of disaster related trauma (8.8% vs. 3.5%, p < .001, χ2). The age of any first trauma was significantly younger in the CO group compared with the AO group (15.4 + 7.8 years vs. 19.7 + 12.3 years, p < .0001, unpaired t-test).
When examining PTSD symptomology, CO binge eating was associated with a higher frequency of a lifetime history of PTSD (31.0% vs. 23.0%, p < .05, χ2), a greater lifetime history of forgetting all or parts of traumatic events (21.2% vs. 15.0%, p < .05, χ2), and a higher number of past avoidance symptoms (1.3 vs. 0.95, p < .02, χ2) and past arousal symptoms (0.94 vs. 0.66, p < .005, χ2).
Analysis of responses about family history indicated that the CO group endorsed higher rates of having a family member with an “emotional problem” (26.1% vs. 17.9%, p < .02, χ2) and having a family member with a drug problem (16.6% vs. 9.8, p < .02, χ2).
There were no significant difference between the two groups for lifetime diagnosis of BED, the likelihood of having ever engaged in purging behavior, minimum or maximum BMIs, age at lowest weight, and current or lifetime history of major depression, rape trauma, current PTSD, or associated current symptoms of re-exeperiencing, arousal, and avoidance, and history of having a family member with a drinking problem or who had been arrested. There were also no significant differences in rates of natural disasters, other injuries, other life threats, witnessing violence, or other stressful events.
There were significant correlations between the age of first binge and age of first trauma (see Fig. 2) (r = .34, p < .0001), the age of serious accident (r = ., p < .0001), the age of natural disaster (r = .25, p < .001), the age of witnessing violence (r = .3, p < .0001), the age of other stressor (r = .33, p < .0001), the age at homicide survival (r = .54, p < .0001), the age of first serious dieting (r = .69, p < .0001), the age when most overweight (r = .65, p < .0001), the age at lowest weight (r = .27, p < .0001), the age of first depressive symptoms for 2 weeks (r = .5, p < .0001), the age of first alcoholic drink (r = .15, p < .0001), the age of first use of marijuana (r = .15, p < .03), and the age of first use of cocaine (r = .53, p < .0001), but not with the ages of first use of angel dust, LSD, heroin, methadone, or glue.
In looking at the chronological relationship between age of first binge and age of first trauma (2 × 2 table: first trauma before or same year as first binge/first binge before first trauma vs. CO/AO,), we found that the age of first trauma either preceded or occurred during the same age as that of first binge in 48% of CO binge eaters and 78% in AO binge eaters (NS, χ).
Results support our hypotheses that onset of binge eating during childhood or adolescence is associated with a greater likelihood of developing BN and a greater severity of bulimic symptoms as evidenced by greater number of purge types, specific purge behaviors used, and worse binge-purge episodes. In fact, we found that CO binge eating behavior increased the risk for the development of any eating disorder in general. Contrary to the study's hypothesis, no significant difference was found between the CO and AO binge eating groups in regard to the development of subsequent BED. This could be because the sample size of women with BED (n = 30) may have been too small to detect a difference. Perhaps a difference could have been detected using the more relaxed DSM-5 criteria.
Consistent with study hypotheses, CO binge eating behavior was associated with an increased risk for lifetime expression of PTSD symptomatology relative to AO binge eating. Specifically a history of avoidance, hyperarousal, and dissociative-based symptomatology was greater in participants with CO versus AO binge eating behavior. Interestingly, none of the measures of current PTSD or current symptom clusters associated with trauma were significantly different between the groups. It may be that a history of trauma is an etiological risk factor for early onset binge eating behavior as well as the development of PTSD. Results also support that CO binge eating was associated with not only a significant history of any trauma overall, but also with higher rates of specific critical traumatic experiences, including histories of molestation, physical assault, any direct crime victimization, and serious accidents. However, AO binge eating was associated with a greater likelihood of experiencing disaster related trauma. In the National Comorbidity Survey–Replication, women with BN reported significantly more frequent natural disasters than women without an eating disorder. Further investigation of the impact of disaster exposure and disordered eating is warranted in order to determine any specific aspects of disaster that may increase adult risk. However, it may be more likely that disaster-related trauma simply is able to induce PTSD and/or its symptoms, which in turn causes more negative affect and less positive affect that in turn triggers binge eating as a coping strategy. This is in keeping with other literature indicating that any potentially traumatic event that produces PTSD or significant PTSD symptoms can induce binge eating in susceptible individuals.[17, 18]
The hypothesis regarding an increase of major depression and substance use disorders in CO versus AO binge eating behavior was supported by the findings of higher rates of lifetime alcohol abuse, lifetime alcohol dependence, and lifetime and current cigarette smoking, as well as earlier onset of major depressive symptoms. These data are consistent with previous investigations supporting an increased risk for impulsive, risk-taking behavior in individuals with BN. These data also have important implications for prevention and intervention efforts as comorbid substance abuse has been reported to be predictive of poorer treatment outcomes in individuals with BN.[37, 38] These findings in the BN population may generalize to women endorsing binge eating with an earlier age of onset and may warrant further investigation with its relationship to these outcomes. Although the rates of current or lifetime major depression were not significantly different between the two groups, there was a trend for the CO group to have higher rates of current major depression. Perhaps CO binge eating is associated with a more severe and chronic course that includes chronic major depression. However, this finding must be interpreted with caution given its marginal significance and the fact that we did not correct for multiple comparisons (see below).
Previous studies have supported that an early age of BN onset results in greater severity of the disorder as determined by greater binge/purge frequency, greater body dissatisfaction ratings, and lower GAF scores. Data from the present investigation support the contention that binge eating beginning in childhood or adolescence may be associated with a more severe form of BN than binge eating beginning in adulthood. In a study examining purging behavior in AN participants, the early onset participants displayed a significantly lower rate of self-induced vomiting and laxative use as well as no significant difference in amount of excessive exercise used in comparison to late onset participants. These findings suggest that generalized early onset disordered eating symptomology does not, in itself, determine the later severity of eating specific eating disorder analysis. The results obtained from this study suggest that binge eating is a unique variable that should be studied (with and without any associated purging behaviors) in relationship to eating disorder diagnostic outcomes. The lack of significance between age of onset of binge eating and development of BED also contradicts previous findings that BED participants report lower average age of first binge than BN patients. Our findings suggest that earlier age of onset of binge eating is more indicative of the eventual development of BN than BED. Our results indicate that CO onset binge eating may be indicative of a more severe form of ED that has both a greater genetic load and a greater environmental load. The findings that the EO onset group endorsed significantly more emotional problems and drug problems in their families, significantly greater rates of and severity of BN and any eating disorder, significantly greater substance abuse and dependence, significantly higher rates of interpersonal trauma, and significantly higher prevalence rates of lifetime PTSD and its symptoms are compatible with this hypothesis.
This particular study had several limitations. One limitation worth noting was the use of retrospective recall of participants being interviewed. Many of the questions posed to participants required them to reflect on past experiences. These questions leave room for error in misremembering events and experiences and/or inaccurately reporting these events and experiences due to their current understanding of them. Additionally, all data were self-reported which can allow for personal bias to interfere with objective disclosure and interpretation of experiences. It also must be noted that the statistical format of data did not allow for more complex analyses of relationships between certain variables, such as the conversion of diagnoses to DSM-5 or the analysis of age of onset of purging behaviors. Perhaps a difference could have been found between the CO and AO groups on the development of subsequent BED if the more “relaxed” DSM-5 criteria were used. Unfortunately, this was not feasible given the design of the study, which was based on the DSM-IV cutoff of twice weekly binge eating.
Another limitation of our study is that we did not adjust for multiple statistical comparisons. Given that we posed several very specific hypotheses to test, we chose not to apply a correction in order to avoid Type II errors. In addition, some of the analyses included in Table 1 were not hypothesized to be different, e.g., minimum and maximum BMI, but are shown for illustrative purposes. Other analyses were performed to more specifically identify the nature of a difference, e.g., past avoidance, hyperarousal, and re-experiencing symptoms were examined in order to pinpoint the found difference in overall lifetime prevalence of PTSD. Nevertheless, not making adjustments for number of analyses runs the risk of Type I errors, especially for findings of low or marginal significance. We leave it to the reader to put our findings in perspective based on the strength of the t-, χ2-, and p-values.
This study was novel in its examination of a spectrum of correlates related to age of onset of binge eating behavior in a nationally representative sample of adult women, but the exclusion of men is a major limitation that limits the generalizability of our findings to women. Future investigations should include men in the sample in order to explore developmental differences in binge eating across the sexes. This study also encourages prospective, longitudinal examination of variables occurring at several different time points throughout the lifespan to lessen retrospective responding and improve accuracy of self-reports. Several findings also warrant further specified analysis. In particular, the finding that disaster-related trauma was related to AO binge eating contradicts the finding that other trauma variables are related to early onset binge eating. Future studies could also include anorexia nervosa and other specified eating disorders in analyses and expand on the finding that early onset binge eating was related to an increase risk of developing any eating disorder. In order to establish causal relationships, the age at which certain experiences occurred in relation to the age at which binge eating behavior started must be the focus of further examination. These findings also call for a focus on prevention of child maltreatment as part of the prevention effort against the development of binge eating and subsequent bulimic eating disorders.
Furthermore, our results suggest the value of screenings for eating pathology at early ages and the importance of early intervention upon onset of binge eating in the effort to prevent the development of full syndrome eating disorders and related comorbidity, such as substance use disorders, later in life. Our findings also have implications for the clinical assessment of eating disorders. For instance, knowing these correlates of early onset binge eating might help guide clinicians to assess for other known comorbidities. Additionally, clinicians may wish to assess age of onset of binge eating behaviors when assessing patients with BED or BN given that this information holds a great deal of predictive value and may shed light on other difficulties the patient is likely to experience.
- 1American Psychiatric Associations. Diagnostic and statistical manual of mental disorders, 4th ed. Washington, DC: Author, 1994.
- 3American Psychiatric Associations. Diagnostic and Statistical Manual of Mental Disorders, 5th ed. Washington, DC: Author, 2013.
- 5The nature and prevalence of binge eating disorder in a national sample of women. In: Widiger TA, Frances AJ, Pincus HA, First MB, Roth R, David W, editors. DSM-IV Sourcebook, Volume 4, Washington: APA Press, 1998, pp. 515–531., , , , , , et al.
- 18Eating disorders, victimization and comorbidity: Principles of treatment. In: Brewerton TD, editor. Clinical Handbook of Eating Disorders: An Integrated Approach. New York: Marcel Dekker, 2004, pp. 509–545..
- 29Factors associated with binge eating disorder in pregnancy. Int J Eat Disord 2011;44:124–133., , , , , .
- 32Childhood psychological, physical, and sexual maltreatment in outpatients with binge eating disorder: Frequency and associations with gender, obesity, and eating-related psychopathology. Obes Res 2001;9:320–325., .
- 33Bulimia nervosa, PTSD and “forgetting”: Results from the National Women's Study. In: Williams LM, Banyard VL, editors. Trauma and Memory. Durham: Sage Publications, 1999, pp. 127–138., , , .