The menopausal transition—A possible window of vulnerability for eating pathology




No published studies, to our knowledge, have examined the association of menopausal status with eating disorders and body image in women. We assessed these associations in a large sample of middle-aged women.


We administered an anonymous questionnaire to a randomly selected nonclinical sample of women aged 40–60 in Innsbruck, Austria. The questionnaire covered demographic items, menopausal status, weight history, measures of body image, and current eating disorders as diagnosed by DSM-IV criteria. Using modified WHO criteria, we classified the respondents' current stage of menopausal transition as premenopausal (N = 192), perimenopausal (N = 110), or naturally postmenopausal (N = 134). In a separate analysis, we also examined the small group of women with surgically induced menopause (N = 12).


The three groups were similar in all demographic features except age, and did not differ significantly on current body mass index (BMI), weight-control behaviors, or dieting history after age adjustment. However, perimenopausal women reported a significantly greater prevalence of eating disorders as compared to premenopausal women. Perimenopausal women also reported significantly higher self-ratings of “feeling fat” and higher Body Shape Questionnaire scores than premenopausal women. Women with surgically induced menopause also showed an elevated prevalence of eating and body image pathology.


Our data suggest that the menopausal transition is associated with an increased prevalence of eating disorders and negative body image. Menopause, like puberty, may perhaps represent a window of vulnerability to these conditions, likely because of changes in hormonal function, body composition, and conceptions of womanhood. © 2013 Wiley Periodicals, Inc. (Int J Eat Disord 2013; 46:609–616)


The menopausal transition is a phase of complex neuroendocrine, physical, mental, and social changes that may render women particularly vulnerable to psychological distress.[1] The onset of powerful and sometimes chaotic hormonal fluctuations, together with novel psychosocial stressors—the loss of reproductive ability and the need to adapt to new meanings of womanhood—all appear likely to increase the risk for the onset of psychiatric disorders. The best studied of these disorders is depression.[2-6] For example, one large prospective study of 3,302 multiethnic women showed significantly elevated depression scores in the perimenopausal period as compared to the premenopausal period, even after adjustment for possible confounding variables.[5] Another recent study found that women with a lifetime history of depression were more likely to develop perimenopausal depression than women without such a history.[6]

However, the overall psychiatric effects of the menopausal transition are still incompletely understood. Certainly, there is some general agreement that women experience a heightened risk for psychiatric conditions during periods of marked hormonal change, such as puberty, the postpartum period, and the menopausal transition.[7, 8] However, aside from some studies rating general mental health, very few studies have examined the association of menopause with other specific psychiatric disorders, and none, to our knowledge, has assessed the association of menopause with eating disorders. This issue deserves to be investigated, given that several recent epidemiological studies have now shown that eating disorders and disordered eating are surprisingly common in middle-aged and older women.[9-16] These studies have revealed that the prevalence of eating disorders in middle-aged women may be comparable to that in younger women, except that the balance of eating-disorder diagnoses among middle-aged women is shifted more towards diagnostic categories associated with normal weight or overweight (bulimia nervosa, binge-eating disorder, eating disorder not otherwise specified). In light of these findings, we sought to explore the association between menopausal status and eating disorders, including associated pathology, in a large population of women between 40 and 60 years of age.



Using census bureau data, we recruited a random population sample of 1,500 women aged 40–60 years in Innsbruck, Austria, for an anonymous questionnaire survey regarding eating behavior and body image. We sent all women an initial letter describing the study, explaining that we wanted to study middle-aged women with regard to eating behavior, body image, menopause, and physical and mental health. This letter included a “non-participation-card” for those not wishing to join. This card was returned by 224 (14.9%) of the women, leaving 1,276 women who were mailed questionnaires. Of these, 715 (56%) returned their questionnaires within 3 months. All respondents signed the associated informed consent form, which was approved by the Ethics Commission of the University Innsbruck.

Using modified WHO criteria[17] and material from related epidemiological publications,[18, 19] we classified the respondents' current stage of menopausal transition as (1) “premenopausal” if they reported that they were continuing to menstruate regularly throughout the past 12 months, with no change in their normal pattern (N = 192); (2) “perimenopausal” if they reported the recent onset of amenorrhea or of menstrual irregularities (changes in the length of the menstrual cycle, or changes in menstrual flow, duration, or intensity) lasting for at least 3 months but less than 12 months (N = 110); and (3) “naturally postmenopausal” if they reported amenorrhea of natural onset (i.e., not attributable to medical interventions) for more than 12 months (N = 134). These classifications were based on participants' responses to a general screening question: “Have you noticed a change in your menstrual cycle in the last 2 years? Please check the appropriate response; you may check more than one item.” The questionnaire then offered nine specific response options, such as “YES—for at least the last 3 months but no more than 12 months, my menstrual periods have been much shorter or much longer than in the past” or “YES—for at least the last 3 months but no more than 12 months, my menstrual flow has been significantly weaker or stronger than in the past.”

In a separate analysis, we examined women reporting surgically induced menopause, defined as amenorrhea because of surgical removal of both ovaries with or without hysterectomy (N = 12). We excluded women with amenorrhea because of other gynecological interventions such as hysterectomy or other uterine surgery in the absence of bilateral oophorectomy (N = 65); and women using birth-control pills, other hormonal treatments (e.g., hormone replacement therapy), or intrauterine devices for birth control (N = 91). We also excluded 111 women whose questionnaire responses were equivocal or insufficient to permit classification into the above categories.

Study Instrument

The questionnaire (available from the authors on request) was kept as brief as possible in order to maximize the participation rate. Thus, we used simple questions and simplified or abbreviated versions of various rating scales, as described below. We first assessed five demographic variables (age, marital status, number of children, country of origin, and level of education). These demographic questions were deliberately limited and broadly defined to ensure anonymity of the respondents. Weight history was obtained from self-reports of height, current weight, highest weight (except pregnancies), and desired weight. From these data, we calculated current, maximum, and desired Body Mass Index (BMI).

In order to generate diagnoses of eating disorders that conformed to threshold DSM-IV criteria,[20] we assessed current “eating disorders” using questions adapted from the German version of the Structured Clinical Interview for DSM-IV (SCID)[21] for anorexia nervosa, bulimia nervosa, and eating disorder not otherwise specified (EDNOS). Examples of EDNOS, as provided by DSM-IV, include (1) meeting full criteria for anorexia nervosa but with regular menses, (2) meeting full criteria for anorexia nervosa, with significant weight loss, but still remaining in the normal weight range, (3) meeting full criteria for bulimia nervosa, but with eating binges less frequent than two times a week or lasting less than 3 months, (4) inappropriate compensatory (purging) behaviors for weight loss in the absence of full eating binges and despite normal weight, (5) chewing and spitting out food, and (6) binge-eating disorder. The last of these examples, binge-eating disorder, is defined by specific proposed criteria in the Appendix to DSM-IV.[20] Therefore we diagnosed and classified binge-eating disorder separately from the other forms of EDNOS. We did not ask for age of onset of eating disorders because this could not be determined accurately without asking respondents to specify the age of onset of each individual symptom, thus making the questionnaire too time-consuming.

In women not meeting full DSM-IV criteria for an eating-disorder diagnosis, but nevertheless reporting some eating-disorder symptoms, we broadly diagnosed current “subthreshold eating disorders” if they reported (1) recurrent eating binges, associated with loss of control over eating (both criteria as defined by DSM-IV), while not fully meeting the remaining DSM-IV criteria for bulimia nervosa or binge-eating disorder or (2) purging behaviors such as use of laxatives, diuretics, appetite suppressants, and/or vomiting, while overweight (note that EDNOS may be diagnosed only if such behaviors occur in the presence of normal weight). Women reporting neither an eating disorder nor a subthreshold eating disorder were classified as “normal eaters.”

We also administered the first three subscales of the Eating Disorder Inventory (EDI): (1) Drive for Thinness, (2) Bulimia, and (3) Body Dissatisfaction.[22] We assessed, current weight-control behavior and perception of body image using questions modified from the Diagnostic Survey for Eating Disorders (DSED).[23] Weight-control behavior included any regular activity to control body weight, including physical activity specifically intended for weight loss and or maintenance, consumption of diet foods, use of artificial sweeteners, or regular counting of calories.

To assess body shape preoccupations typically associated with eating disorders, we administered the Body Shape Questionnaire (BSQ),[24] a 34-item self-report measure. A score of less than 80 on this instrument suggests no significant concern with body shape; 80–110 indicates mild concern; 111–140, moderate concern; and greater than 140, marked concern.

We rated perceived health on the basis of three open questions asking respondents to list, in their own words, (1) current medical conditions (e.g., hypertension, diabetes, arthritis, etc.), (2) current psychological disorders (e.g., depression, anxiety, substance abuse, etc.), and (3) all lifetime surgical interventions. We assessed cigarette consumption by asking, “How many cigarettes do you smoke per day?” Response options were: none, sometimes (less than daily), or four categories of daily use (1–5, 6–10, 11–20, or >20 cigarettes/day). For alcohol consumption, we asked, “How much alcohol do you currently drink?” Response options were: never, rare, occasional, sometimes, often, or three categories of daily (1–2, 3–5, or >5 glasses per day).

To assess current depressive symptomatology, we administered the Center for Epidemiologic Studies Depression Scale (CES-D),[25] a widely used self-report scale covering depressive symptoms experienced in the last week. A score of 16 or above on the CES-D indicates clinically significant depressive symptomatology.[25]

We assessed menopausal symptoms using a simplified version of the Menopausal Rating Scale (MRS).[26] This scale covers 11 items in 3 domains: (1) psychological symptoms (depressed, irritable, anxious, exhausted), (2) somato-vegetative symptoms (sweating/flushing, cardiac complaints, sleeping disorders, joint and muscle complaints), and (3) urogenital symptoms (sexual problems, urinary complaints, vaginal dryness). Normally this scale offers five response options for each item (from “not present” to “severe”), but in accordance with our strategy to keep the questionnaire as brief and simple as possible, we restricted each response to a dichotomous “yes” or “no.” Thus the total score on the MRS could range from 0 to 11 “yes” responses.

Statistical Analysis

Statistical analysis was performed using SPSS (version 18).[27] To compare women of the three menopausal groups (premenopause, perimenopause, and postmenopause) with respect to demographic features, we used the Kruskal–Wallis test for continuous and ordinal variables, and the chi-square test for nominal variables. Post-hoc pairwise group comparisons were performed using the Mann–Whitney U test and Fisher's exact-test, respectively.

Table 1. Demographic, medical, psychological, and menopausal variables
 Group 1Pre-MPbGroup 2Peri-MPcGroup 3Post-MPdSignificance of Differencesa
 N = 192 (44%)N = 110 (25%)N = 134 (31%)Overall 1 vs. 2 1 vs. 32 vs.3
  1. a

    Comparisons for age are unadjusted; all remaining comparisons are age-adjusted.

  2. b


  3. c


  4. d


  5. e

    Modified scoring method as described in the text.

Demographic variables       
Age categories (years), N (%)   0.000   
40–45126 (66)35 (32)2 (2)    
46–5054 (28)43 (40)15 (11)    
51–5510 (5)26 (24)66 (50)    
56–601 (1)4 (4)50 (38)    
Married or in partnership, N (%)153 (80)83 (76)94 (71)ns   
One or more children, N (%)144 (75)83 (76)100 (75)ns   
Education >12 years, N (%)87 (46)41 (38)54 (41)ns   
European country of origin, N (%)180 (96)107 (99)129 (98)ns   
Medical variables       
Perceived health condition, N (%)       
Reporting >1 current medical illness90 (47)39 (36)50 (37)ns   
Reporting >1 current psychiatric condition39 (20)20 (18)37 (28)ns   
Current cigarette smoking, N (%)   ns   
Daily (1 to >20 cigarettes)47 (25)33 (31)41 (32)    
Sometimes (less than daily)10 (5)7 (7)5 (4)    
None129 (70)66 (62)84 (64)    
Current alcohol consumption, N (%)   ns   
Often or daily (>1–2 glasses/day)27 (14)17 (16)31 (23)    
Sometimes (less than daily) or rarely134 (71)77 (72)82 (62)    
None28 (15)13 (12)20 (15)    
Psychological variables       
Depression, CES-D, M (SD)10.9 (9)12.1 (9)12.2 (10)ns   
Depression, CES-D, cutoff >16, N (%)44 (23)31 (28)35 (28)ns   
Menopausal variables       
Menopause Rating Scale,e M (SD)       
Psychological symptoms (score: 0–4)0.1 (0.6)0.3 (0.8)1.0 (1.3)0.0000.0020.0000.000
Somato-vegetative symptoms (score: 0–4)0.2 (0.6)0.7 (1.0)1.6 (1.0)0.0000.0000.0000.000
Urogenital symptoms (score: 0–3)0.04 (0.3)0.2 (0.5)0.6 (0.8)0.0000.0010.0000.035
All symptoms (score: 0–11)0.4 (1.3)1.2 (1.8)3.2 (2.5)0.0000.0000.0000.000

As the three groups differed significantly with regard to age, the potentially confounding effect of this variable was taken into account in all subsequent analyses. We used analysis of covariance for group comparisons involving continuous variables, logistic regression for binary-dependent variables, and ordinal regression for ordinal-dependent variables, always with adjustment for age. For comparisons regarding body image, we additionally adjusted for the potential confounder BMI. Post-hoc pairwise group comparisons were performed only if the overall comparison of the three groups had yielded statistical significance (P < 0.05). This sequential testing procedure grants, in the case of three groups, that the family-wise alpha-level of 0.05 is retained without correction for multiple testing.[28]


Demographic, Medical, Psychological, and Menopausal Variables

The three groups of women were similar on all demographic characteristics except age, which differed as expected because of the definition of the groups (Table 1). Self-ratings of medical and psychiatric disorders, alcohol and cigarette consumption, and even CES-D scores also did not differ significantly among groups. On each of the three domains of the MRS, as well as MRS total score, the number of symptoms endorsed showed a continuous, statistically significant increase from premenopause to perimenopause to postmenopause.

Body Weight, Weight Control, Eating Behavior, and Body Image

The three groups showed no significant differences on measures of current or past body weight or on dieting behavior after adjustment for age (Table 2), but differed markedly in their prevalence of eating disorders, with 9% of perimenopausal women reporting a DSM-IV eating disorder as compared to only 2% the premenopausal women (P = 0.007). Perimenopausal women also showed a slightly but not significantly higher prevalence of eating disorders when compared to the postmenopausal women (9% vs. 5%; P = 0.32). Indeed, when including cases of subthreshold eating disorders, fully 15% of the perimenopausal women reported at least some abnormalities of eating behavior. In addition, perimenopausal women differed significantly from premenopausal women in reported feelings of fatness and in BSQ scores, even after adjustment for current BMI.

Table 2. Body weight, weight control, eating behavior, and body image
 Group 1Pre-MPbGroup 2Peri-MPcGroup 3Post-MPdSignificance of Differencesa
 N = 192 (44%)N = 110 (25%)N = 134 (31%)Overall 1 vs. 2 1 vs. 32 vs.3
  1. a

    All comparisons are adjusted for age; comparisons on body–image measures are further adjusted for current BMI.

  2. b


  3. c


  4. d


  5. e

    See text for definition of terms.

  6. f

    Eating Disorder Inventory.

  7. g

    Body Shape Questionnaire.

Body weight       
Current body mass index (BMI), M (SD)22.9 (3.8)24.2 (5.1)24.4 (4.4)ns   
Lifetime maximum BMI, M (SD)24.7 (4.6)26.2 (5.9)26.1 (4.6)ns   
Desired BMI, M (SD)21.5 (2.2)22.2 (2.7)22.5 (2.5)ns   
Weight control       
Current weight control behavior, N (%)172 (90)93 (85)111 (84)ns   
Lifetime restrictive dieting, N (%)   ns   
Often (>20 times)6 (3)4 (4)6 (5)    
Never, rarely, or sometimes (0–20 times)177 (97)100 (96)128 (96)    
Eating behavior   0.0190.006nsns
Normal eating, N (%)181 (94)93 (85)120 (90)    
Subthreshold eating disorders,e N (%)7 (4)7 (6)7 (5)    
-Binge eating with loss of control553    
-Purging when overweight/obese224    
Eating disorders (DSM-IV), N (%)4 (2)10 (9)7 (5)0.0320.007nsns
-Anorexia nervosa000    
-Bulimia nervosa223    
-Binge eating disorder142    
-Eating disorder not otherwise specified142    
-Anorexia nervosa without amenorrhea010    
-Purging when normal weight132    
EDIf—total of 3 subscales, M (SD)14.0 (8.9)15.8 (10.0)14.3 (9.7)ns   
EDI—drive for thinness2.5 (3.9)3.4 (4.7)2.5 (3.6)ns   
EDI—bulimia0.3 (0.8)0.5 (1.3)0.5 (1.9)ns   
EDI—body dissatisfaction11.1 (5.6)12.2 (5.7)11.6 (6.1)ns   
Body image measures       
Satisfaction with weight? N (%)   ns   
Satisfied95 (50)49 (45)57 (43)    
Moderately satisfied60 (31)30 (27)41 (31)    
Dissatisfied37 (19)31 (28)36 (27)    
Feeling of fatness, N (%)   0.0490.0300.027 ns
Very fat11 (6)12 (11)14 (11)    
Moderately fat61 (32)46 (42)49 (37)    
Not fat at all119 (62)52 (47)70 (53)    
BSQgscore, M (SD)67 (29)78 (35)69 (30)0.0130.010nsns
“I really like my body,” N (%)   0.075   
Agreement136 (71)63 (58)86 (65)    
Do not know or disagreement56 (29)46 (42)47 (35)    

However, we found no statistically significant differences between groups on the EDI.

Looking across all three menopausal status groups, and using linear regression with adjustment for age, we found that women with eating disorders and subthreshold eating disorders differed significantly from women with normal eating in number of positively endorsed MRS items [eating disorders mean (SD): 3.0 (2.9) vs. normal eating: 1.5 (2.2); P < 0.001; subthreshold eating disorders: 2.5 (2.9) vs. normal eating: 1.5 (2.2); P = 0.002] and in depression scores on the CES-D [eating disorders: 22.4 (10.9) vs. normal eating: 10.8 (8.8); P < 0.001; subthreshold eating disorders: 18.8 (11.3) vs. normal eating: 10.8 (8.8); P < 0.001]. The eating disorder and subthreshold eating-disorder groups did not differ significantly from one another on either these measures (P > 0.25).

Finally, we performed a separate analysis of the small group of women (N = 12) with surgically induced menopause. This group showed no significant differences from the other three groups on any of the demographic variables in Table 1, save for the fact that only four (25%) had children. These women showed a high prevalence of eating disorders [eating disorders: N = 2: one with bulimia nervosa, one with EDNOS (anorexia nervosa with BMI of 20); subthreshold eating disorders: N = 1 use of laxatives with obesity]. When compared to the premenopausal women, the surgically menopausal women also reported significantly greater feelings of fatness [18% vs. 6% feeling “very fat,” 73% vs. 30% “moderately fat,” and 9% vs. 65% “not fat at all”; P = 0.002 after adjustment for age and BMI by ordinal regression] and significantly higher BSQ scores [101 (44) vs. 67 (29); P < 0.001 by linear regression after adjustment for age and BMI].


We assessed the association of menopausal status with eating disorders and associated pathology in a population-based sample of 436 community women, aged 40–60 years, in Innsbruck, Austria. To our knowledge, this is the first study to examine the association of eating disorders, body image, and menopausal status in women at midlife. We found no significant differences among premenopausal, perimenopausal, and naturally postmenopausal women on a range of demographic and health measures, aside from the expected difference among groups in age (which was used as an adjustment variable for all further comparisons). However, the perimenopausal women displayed a markedly higher prevalence of eating disorders than the premenopausal women, and also scored significantly higher than premenopausal women on two measures of body image. These latter differences were particularly striking when it is considered that they were significant even after adjustment for both age and BMI—suggesting that the greater body-image distress of the perimenopausal women was attributable to factors over and above the effects of weight gain associated with perimenopause. Somewhat surprisingly, however, we failed to find statistical significances on the EDI. On inspection of the participants' responses, it appeared that this finding might have been because of the fact that the three chosen subscales of the EDI were more sensitive to symptoms of anorexia and bulimia nervosa (which were uncommon in the sample), as opposed to binge-eating disorder and EDNOS (which accounted for the majority of eating-disorder cases in the sample).

Finally, we also noted a high prevalence of eating disorders and body image pathology in the small group of 12 women reporting surgically induced menopause.

Several limitations of our study should be considered. First, 14.9% of the initial sample of women declined to receive a questionnaire, and of the women who did receive the questionnaire, 56.0% responded. Although the guaranteed anonymity of the questionnaire presumably encouraged candid responses, women with eating disorders, depression, or other psychiatric disorders may have been more reluctant to respond, thus causing us to underestimate the true prevalence of psychopathology. However, barring the unlikely possibility that women with psychopathology in one of the three menopause-status groups were selectively less likely to respond to the questionnaire than women with comparable psychopathology in the other groups, failure to respond would not bias our comparisons among the three groups. Similar considerations would apply to the 111 cases that were excluded because they provided insufficient information to classify menopausal status. The exclusion of these cases might be expected to reduce the overall statistical power of the study, but would not bias comparisons among the groups, barring the unlikely possibility of a systematic association between eating-disorder status, menopausal status, and the provision of insufficient menopausal information on the questionnaire.

Second, we did not obtain hormonal validation of menopausal status in our respondents because of the anonymous study design. Therefore at least some respondents were probably misclassified. However, such misclassification likely occurred in both directions (i.e., some perimenopausal women misclassified as nonperimenopausal and some nonperimenopausal women classified as perimenopausal). Such nondifferential misclassification error would typically bias the results toward the null, and thus would be unlikely to have created a false-positive finding in this study.

Third, we defined “subthreshold eating disorders” very broadly, using criteria of our own design, rather than adopting previously published criteria for “subthreshold” disorders, which are typically more narrow than our subthreshold criteria here.[29-31] However, as noted above in our analyses of MRS and CES-D scores, the group with subthreshold eating disorders differed significantly from the group of normal eaters on both of these measures, but showed no significant differences from the group meeting the full threshold DSM-IV criteria for eating disorders. This finding suggests that women with even broadly defined subthreshold eating disorders may share important features with women meeting the full threshold criteria for an eating disorder.

Fourth, our analysis was based on responses to an anonymous questionnaire, and thus we could not perform follow-up interviews to confirm the validity of responses. Although the questionnaire utilized some scales with established psychometric properties,[22-26] other parts of the questionnaire had not been similarly tested. In particular, we used questions adapted from the SCID to diagnose eating disorders, even though these questions are not normally administered in a self-report format, because we were not aware of any alternative self-report instrument with established psychometric properties that would generate DSM-IV diagnoses of eating disorders. Although SCID-based questions have been used in at least one other previous large self-report study,[15] the psychometric properties of this approach have not been established. Subsequent studies using a two-stage design (questionnaires with follow-up interviews) will be required to confirm and expand upon the findings of our preliminary investigation here.

While subject to these limitations, our data offer preliminary evidence that the menopausal transition may be associated with an increased prevalence of eating disorders and associated pathology. The causes of this phenomenon, however, remain speculative. It seems likely that both biological factors (e.g., hormonal changes; feeling that one's body is “out-of-control”) and psychosocial factors (the transition to a new stage of womanhood with different meanings) contribute to an increased prevalence of eating pathology. Indeed, menopause might be conceptualized to resemble puberty, in that both stages of life are times of major hormonal and psychological changes. Thus puberty and menopause—the “entrance” and the “exit” from reproductive life—may represent complementary windows of vulnerability for the development of eating disorders and associated pathology. Of course, given the cross-sectional nature of the present study, it is difficult to determine the specific causal factors that contribute to this apparent vulnerability during menopause; subsequent longitudinal observations will be required to clarify the nature of these factors.

Body image concerns are frequently associated with the development of eating disorders, and thus it is not surprising that we found feelings of fatness and body image concerns (assessed by the BSQ) to be elevated in perimenopausal women as compared to premenopausal women, even after adjustment for age and BMI. At present, published data in this area remain limited and somewhat conflicting.[32] Deeks and McCabe[33] found postmenopausal women to be less positive regarding their appearance compared to premenopausal women, whereas McLaren[34] showed that postmenopausal women were more satisfied with their weight than were premenopausal women.

Interestingly, we found no significant differences among groups in self-ratings of depressive symptoms, despite the extensive literature showing an increased risk of depression during the menopausal transition (e.g., see Refs. [4] and [5]). Although unexpected, our finding in this regard appears somewhat consistent with a recent report by Gibbs et al.,[35] who showed that hormonal vulnerability, in and of itself, does not account for the phenomenon of perimenopausal depression. Instead, these authors posited that depression arises from a complex interplay between hormonal vulnerability, psychosocial resources, overall well-being, and demands on coping resources. Notably, we did find a strong association between self-rated depressive symptoms and the presence of eating disorders.

Our findings have implications for both research and clinical practice. In the past, eating disorders have often been considered to be uncommon in middle-aged to older women, and an association between eating disorders and menopause, to our knowledge, has not previously been demonstrated in the literature. Our study suggests that clinicians should be alert to the apparently heightened risk of eating disorders in perimenopausal women, and should take care to explore this possibility when assessing such patients. Given that many patients suffer covertly from eating disorders, and do not spontaneously disclose their symptoms to clinicians, cases of perimenopausal eating disorders may well be missed and left untreated if not specifically sought. It would also seem important to devote further research to the apparent association of eating disorders with menopause because such research may offer new insight into the complex interplay of biological and psychosocial factors that give rise both to body image disorders in general and eating disorders in particular.


The authors thank Dr. Kerstin Lackner-Seifert, PhD, Department of Psychology, Innsbruck University, Innsbruck, Austria, Dr. Nadja Frey, MD, Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Innsbruck, Austria, and Prof. Ludwig Wildt, MD, Department of Gynecological Endocrinology and Reproductive Medicine, Innsbruck Medical University, Innsbruck, Austria.

Declaration of Interest

The study was supported by the public health services of the City of Innsbruck (Austria). None of the authors had any commercial interest related to study planning or conduct.

Earn CE credit for this article!

Visit: for additional information. There may be a delay in the posting of the article, so continue to check back and look for the section on Eating Disorders. Additional information about the programis available at