Abstract
Objective
Bulimia nervosa (BN) and binge eating disorder (BED) are associated with significant health impairment. Premenstrual syndrome (PMS) and premenstrual dysphoric disorder (PMDD) comprise both psychological (disturbances in mood and affect) and physiological (bloating and changes in appetite) symptoms that may trigger binge eating and/or purging.
Method
Female participants were drawn from the Collaborative Psychiatric Epidemiological Surveys, conducted from 2001–2003. Weighted multivariable logistic regression modeled the association between lifetime PMS and PMDD and lifetime odds of BN or BED.
Results
Among 8,694 participants, 133 (1.0%) had BN and 185 (1.8%) BED. Additionally, 366 (4.2%) had PMDD and 3,489 (42.4%) had PMS. Prevalence of PMDD and PMS were 17.4% and 55.4% among those with BN, 10.7% and 48.9% among those with BED and 3.4% and 59.1% among those with subthreshold BED. After adjustment for age, race/ethnicity, income, education, body mass index, age at menarche, birth control use and comorbid mental health conditions, PMDD was associated with 7-times the odds of BN (OR 7.2, 95% CI 2.3, 22.4) and PMS with 2-times the odds of BN (OR 2.5, 95% CI 1.1, 5.7). Neither PMDD nor PMS were significantly associated with BED.
Discussion
Women with PMS and PMDD have a higher odds of BN, independent of comorbid mental health conditions. PMS and PMDD may be important comorbidities to BN to consider in clinical settings, and future research should investigate whether PMS and PMDD affect the onset and duration of bulimic symptoms as well as the potential for shared risk factors across disorders.
Young women are disproportionately affected by bulimia nervosa (BN) and binge-eating disorder (BED), disorders which increase the risk of chronic health problems (1, 2), reduce quality of life (3), and increase the risk of suicide attempts (4) and mortality (5). BN affects approximately 1.5% of women (6) and BED approximately 3.5% of women (6) in the United States. Median duration of BN and BED have been estimated in the WHO World Mental Health Surveys as 6.5 (IQR 2.2, 15.4) years and 4.3 (IQR 1.0, 11.7) years, respectively (2). Additionally, both have high comorbidity with mood disorders, anxiety disorders, and other mental health conditions (2, 7), which may affect the severity of both BN and BED, lengthen time to recovery (8), and hinder treatment outcome (9). Premenstrual syndrome (PMS) and the more severe premenstrual dysphoric disorder (PMDD) are potential comorbidities to BN and BED that may affect the onset, severity and duration of binge-type eating disorders among women.
PMDD and PMS affect approximately 3–8% and 20–32% of women, respectively (10). The menstrual cycle has two phases: the follicular phase, which is marked by a surge in estrogen and begins at menstruation and ends at ovulation, and the luteal phase, which is marked by a surge in progesterone and a secondary surge in estrogen and follows ovulation (11). PMDD and PMS are marked by changes in both psychological and physiological symptoms in the late luteal phase of the menstrual cycle (12). Psychological symptoms, including alterations in mood and affect, and physiological symptoms, including abdominal bloating, edema and changes in appetite and cravings (13, 14), may exacerbate symptoms of binge-type eating disorders. PMDD, newly moved from the appendix to an officially recognized disorder in the Diagnostic and Statistical Manual, Fifth Edition (DSM-5), is a more severe form of PMS, in which symptoms lead to significant interference in usual activities or in functioning at work, school, or in interpersonal relationships (15). To be characterized as having PMS or PMDD, symptoms must remit within a few days after the onset of menses and remain normal during the follicular phase (approximately the 2 weeks after menstruation) (15, 16). If the symptoms experienced during the late-luteal phase represent worsening of an underlying mental health condition, women may be diagnosed with premenstrual exacerbation of the underlying condition (e.g., major depressive disorder) (17), rather than with PMS or PMDD (16). Although the etiology of premenstrual conditions remains largely unknown, research has suggested that, rather than experiencing differences in menstrual function and ovarian hormone levels (18), women with premenstrual conditions may have an increased sensitivity to the normal fluctuation of hormone levels during the menstrual cycle, particularly as relates to γ-aminobutyric acid A (GABA-A) receptor sensitivity and the serotonergic system (19). This hypothesis is at least partly supported by the demonstrated efficacy of selective serotonin reuptake inhibitors (SSRIs) in promoting remission of PMS/PMDD symptoms in randomized clinical trials (20).
There is reason to believe that PMS or PMDD symptoms could serve as maintenance factors for binge-type eating disorders. Similar to other eating disorder comorbidities (2, 7), the physiological and psychological symptoms associated with premenstrual conditions may serve as a trigger for binge eating and/or purging (21), potentially affecting the onset of binge-type eating disorders as well as serving as a maintenance factor impeding recovery. Potential triggers include physiological changes associated with PMS/PMDD, such as bloating and changes in body weight (which could be perceived as “fatness” by individuals with eating disorders) and increases in appetite and cravings (which could lead directly to binge eating) (13, 14). Studies have found that women with PMDD report significantly higher calorie intake and uncontrolled eating in the late-luteal vs. follicular phases of the menstrual cycle (22), and greater food cravings and positive emotional response to sweet and high-fat foods in the late-luteal phase (23). Additionally, psychological changes associated with PMS and PMDD, such as alterations in mood and affect (13, 14), may serve as triggers. Among healthy reproductive-aged women, weight preoccupation was found to be significantly higher in the pre-menstrual and menstrual phases of the menstrual cycle, independent of inter-individual variation in estradiol and progesterone levels (24).
Little past research has investigated the potential role of premenstrual conditions in the onset, severity and duration of binge-type eating disorders. A few studies have documented an exacerbation of bulimic symptoms during the premenstrual phase (25–27), during which increases in the ovarian hormones progesterone and estrogen may be associated with depressed mood, irritability, and anger (13). For example, a prospective study following eight women with BN for one menstrual cycle found that bulimic symptoms were significantly worsened in the luteal phase (25). However, past research into this topic has been limited by small sample size, and has not explored the potential role of premenstrual conditions in the onset and duration of binge-type eating disorders.
There is also reason to believe that PMS or PMDD may have shared risk factors for binge-type eating disorders. Specifically, eating disorders typically onset during or after puberty and are more common in females, which suggests that the epigenetic changes that underlie the activation of ovarian hormones at puberty may play an etiological role (28, 29). It is also plausible that women who have increased sensitivity to fluctuations in ovarian hormones through development of PMS and PMDD are more likely to experience the symptoms of disordered eating that accompany the activation of ovarian hormones.
Although PMS and PMDD have the potential to exacerbate symptoms of binge-type eating disorders, few studies have looked at their association, and none have quantified the degree of comorbidity. Therefore, the present study addresses these gaps in the literature by investigating the cross-sectional association between a lifetime history of PMS and PMDD, and a lifetime history of BN, BED and subthreshold BED in an ethnically diverse, nationally representative sample from the Collaborative Psychiatric Epidemiological Surveys. To address the potential for PMS and PMDD to inhibit recovery from BN and BED (e.g., by serving as a maintenance factor for disordered eating), we additionally investigated the association of PMS and PMDD with duration of lifetime binge-type eating disorder symptoms.
METHODS
Study Population
The Collaborative Psychiatric Epidemiological Surveys (CPES) consists of three independent surveys conducted between 2001–2003: The National Comorbidity Survey Replication (NCS-R); the National Survey of American Life (NSAL)-focused on African American and Afro-Caribbean populations; and the National Latino and Asian American Survey (NLAAS)—focused on Latino and Asian American populations (30). Each survey employed a similar design, methodology, and set of assessment tools. The main aims of these surveys were to document the prevalence and correlates of mental health conditions in the United States and associated utilization of mental health services.
All three surveys used a four-stage national area probability sample, consisting of nested samples conducted in the following order: 1) US Metropolitan Statistical Areas (MSAs) and counties, 2) area segments within each selected MSA or county, 3) housing units within each selected area segment, and 4) eligible respondents in each selected housing unit. Surveys were administered in the participant’s home when possible, utilizing a computer-assisted personal interview (CAPI) format. Surveys included comparable content for the assessment of prevalence of mental health conditions following criteria laid out in DSM-IV (31). All DSM-IV diagnoses were assessed using the World Health Organization’s World Mental Health Survey update to the Composite International Diagnostic Interview (WMH-CIDI) (30). Interviews were conducted by lay interviewers in participants’ homes. All participants in the Collaborative Psychiatric Epidemiological Surveys provided informed consent, and this secondary analysis was reviewed and approved by Partners Healthcare Institutional Review Board.
Binge-Type Eating Disorders
The outcomes of interest in this study were BN, BED and subthreshold BED (which is associated with similar levels of eating pathology and general psychopathology compared to full threshold BED) (32). Importantly, as data were collected before the 2013 publication of DSM-5, BN and BED were defined according to DSM-IV criteria. We excluded all participants who had a lifetime history of anorexia nervosa (AN) from analysis due to low prevalence in this dataset in addition to the low body weight associated with a diagnosis of AN, which may interfere with normal menstrual function, including inducing amenorrhea, which could affect the presence of PMDD and PMS. Therefore, participants with lifetime BN, BED and subthreshold BED in addition to lifetime AN were excluded from analysis (among the 25 participants with AN who had complete data, 8 also had a lifetime history of bulimia and 2 had a lifetime history of BED).
Prior to completing the diagnostic module for BN and BED, participants were assessed with a screening question asking whether they had ever in their lifetime engaged in a period of eating binges which occurred at least twice per week for several months. Among those that answered that they had engaged in a period of binge eating, participants met DSM-IV criteria for lifetime BN if, in the absence of AN, they reported a period of binge eating (i.e., consuming a large amount of food while feeling a loss of control) at least twice per week for 3 months or longer that was accompanied by inappropriate compensatory behavior. Participants met criteria for lifetime BED if, in the absence of AN and BN, they reported a period of binge eating at least twice per week for six months or longer that was associated with at least three of the five DSM-IV binge features. Participants were classified as having lifetime subthreshold BED if, in the absence of AN, BN, or BED, they reported a period of binge eating at least twice per week for three months or longer (33). Of note, a subset of individuals defined in our study as having DSM-IV sub-threshold BED may also have met criteria for full-threshold BED under DSM-5 if they reported duration of 3 months or longer (but not 6 months or longer) and met all other criteria for binge eating disorder, including at least 3 of 5 binge features. For the purpose of analysis, participants were classified according to their most severe lifetime eating disorder (exclusive of AN), in the order of BN, BED, subthreshold BED and no lifetime history of a DSM-IV eating disorder. For example, women with a lifetime history of both BN and BED would be classified as having a lifetime history of BN. For BN, the CIDI has moderate to good reliability (test-retest kappa=0.64 and interrater kappa=0.78) (34) and validity (35) as compared to clinician diagnoses.
Premenstrual syndrome and premenstrual dysphoric disorder
To screen women for the premenstrual syndrome module, participants were first asked whether they ever had a time in their life when their mood became worse in the week before their menstrual period and returned to normal within a few days of the start of their menstrual period. Because standard diagnostic criteria for PMS requires remission of PMS-related symptoms during the follicular phase of the menstrual cycles cycle (16), the measure employed in the Collaborative Psychiatric Epidemiological Surveys included mood-related symptoms of both PMS and premenstrual exacerbation of existing mental health conditions, and therefore our analyses included women with either PMS or premenstrual exacerbation not associated with significant interference in daily functioning.
PMDD was defined in this study according to DSM-IV criteria (31) (also consistent with DSM-5 criteria), and endorsed when participants met all of the following criteria: 1) presence of PMS as defined above, 2) worsening of mood occurring in at least 7 months/year, 3) worsening of mood occurring most or all of the time in the week prior to menstruation, 4) mood symptoms accompanied by physiological symptoms, including difficulty concentrating, tiredness or appetite or sleep changes, and 5) worsening of mood leading to interference in work, social life or personal relationships some, most or all of the time. Similar to the classification of PMS using the CIDI, PMDD as measured by the CIDI included both PMDD as defined by DSM-IV and premenstrual exacerbation of existing mental health conditions leading to significant interference in daily functioning, and therefore our analyses included women with either PMDD or premenstrual exacerbation associated with significant interference in daily functioning.
Covariates
Demographic factors assessed include age (years), race/ethnicity (Latino, non-Latino black, non-Latino white, another race/ethnicity), educational attainment (less than high school, high school, at least some college), annual household income (<$15,000, $15,000–<$35,000, $35,000–<$75,000, ≥ $75,000), partner status (married/cohabitating, divorced/separated/widowed, never married) and occupational status (employed, unemployed, not in the workforce). Reproductive factors assessed include age at menarche (years), having ever used hormonal birth control and having ever given birth. Weight status was assessed using body mass index (BMI) (<18.5, 18.5–<25, 25–<30, ≥ 30 kg/m2). Comorbid mental health conditions were defined according to DSM-IV criteria and include lifetime history of major depressive disorder (MDD), dysthymia, bipolar disorder, agoraphobia, panic disorder, generalized anxiety disorder, social phobia, substance use disorders (defined as having at least one of the following: alcohol abuse, alcohol dependence, substance abuse or substance dependence) and posttraumatic stress disorder.
Statistical Analysis
All analyses were weighted to account for sampling probabilities inherent in the survey design. Numbers and frequencies or means and standard deviations were calculated for descriptive statistics, with distributional differences between groups assessed using Rao-Scott chi-square tests or weighted least-squares linear regression. Lifetime history of an eating disorder was modeled as a series of dummy variables, including 1) lifetime BN, 2) lifetime BED in the absence of BN, 3) subthreshold BED (lifetime history of a period of binge eating with episodes at least twice per week for 3 months or longer, in the absence of BN and BED), and 4) no history of a DSM-IV eating disorder. Bivariate multinomial logistic regression was used to model the association between lifetime history of PMS and PMDD and BN, BED and subthreshold BED. Two adjusted models were run: an a priori model and a full model. For the a priori model, we identified potential confounders from the literature (2, 36) including age, race/ethnicity, income, education, age at menarche, lifetime use of hormonal birth control and lifetime diagnosis of major depressive disorder. For the full model, we retained all a priori covariates. Additional potential confounders, including demographic factors, menstrual cycle characteristics and comorbid mental health conditions other than major depressive disorder, were assessed by adding each independently into the a priori model, and retaining them if their addition to the a priori resulted in a 10% or greater change in one of the main effect estimates. Confounders identified using this method include a lifetime diagnosis of dysthymia, bipolar disorder, social phobia and substance use disorder. In separate subsets of participants with a lifetime history of BN, BED, and subthreshold BED, respectively, we used linear regression to model the association of PMS and PMDD with duration of each condition, similarly by first using bivariate linear regression to model the duration of each condition in relation to lifetime history of PMS and PMDD, next using an a priori model, and finally using a full model, both adjusting for the covariates identified above. All analyses were conducted in SAS 9.3 (SAS Institute Inc., Cary, NC).
RESULTS
Of 20,013 total Collaborative Psychiatric Epidemiological Surveys participants, 11,463 (51.9%) were female and eligible for inclusion in this study. Of these, 9,111 (94.8%) were asked questions about both eating disorders and PMS/PMDD (519 non-Latina white participants in National Survey of American Life were not assessed for eating disorders or PMS/PMDD (37) and 1833 of participants in National Comorbidity Survey Replication did not complete the long-form survey which included questions on PMS/PMDD). Furthermore, 392 (3.6%) were missing data on PMS/PMDD and 146 (0.9%) were missing data on eating disorders, resulting in a final study population of 8,694 participants. Women who were excluded from analysis were more likely to be older (mean 49.6 vs. 43.4 years, p<0.001), be non-Latina white (86.8% vs. 62.4%), have a lower household income (13.3% vs. 23.8% with income ≥ $75,000), have completed high school or at least some college (84.7% vs. 81.4%), be divorced, separated or widowed (28.3% vs. 24.5%) and not have children in the household (74.3% vs. 65.3%). Among women included in the analysis, mean age at menarche was 12.7 (SEM 0.03) years. A total of 5,607 (65.3%) had ever used hormonal birth control, and 6,669 (74.7%) had given birth. A total of 1,749 (19.9%) had a lifetime history of major depressive disorder.
Table 1 presents the results of bivariate analyses. In total, 133 (1.0%) participants had a lifetime history of BN, 185 (1.8%) had a lifetime history of BED, 106 (0.6%) had a lifetime history of subthreshold BED, and 8270 (96.6%) had no history of an eating disorder. Average duration of each condition was 13.0 (SEM 1.5, range 0–49) years for BN, 13.7 (SEM 1.8, range 0–56) years for BED and 6.4 (SEM 1.6, range 0–63) years for subthreshold BED. Women with a lifetime history of BN or subthreshold BED were younger, less likely to be non-Latina white, and had lower educational attainment compared to women without a history of an eating disorder. Women with BED and subthreshold BED were less likely to be married, and women with BN and BED were more likely to smoke, than women without a history of an eating disorder. Women with BN and BED additionally had a significantly lower age at menarche, and were significantly more likely than women without a history of an eating disorder to have ever used hormonal birth control. All comorbid mental health conditions assessed were more common among women with an eating disorder than among those without, with particularly high prevalence among those with BN. The most common comorbidities of BN were major depressive disorder (38.2%), posttraumatic stress disorder (34.9%), social phobia (33.5%) and dysthymia (27.4%). Finally, 366 (4.2%) met criteria for PMDD, 3,489 (42.4%) met criteria for PMS in the absence of PMDD, and 4,839 (53.4%) did not meet criteria for PMS or PMDD.
Table 1.
Study population characteristics by eating disorder diagnosis (n=8694).
| BN (n=133) | BED (n=185) | Subthreshold BED (n=106) | No binge eating (n=8270) | ||||||
|---|---|---|---|---|---|---|---|---|---|
|
|
|||||||||
| n | % | n | % | n | % | n | % | P | |
| Age (weighted mean, SEM) | 35.6 | 1.3 | 41.6 | 1.9 | 38.0 | 2.3 | 45.5 | 0.4 | <0.001 |
| Race/ethnicity | |||||||||
| Asian | 12 | 5.9 | 15 | 3.8 | 17 | 12.8 | 1067 | 4.7 | <0.001 |
| Latina | 32 | 18.7 | 53 | 15.1 | 26 | 22.8 | 1684 | 11.6 | |
| Non-Latina black | 62 | 21.5 | 67 | 13.4 | 51 | 31.0 | 3220 | 12.2 | |
| Non-Latina white | 24 | 45.3 | 49 | 65.0 | 10 | 29.3 | 2197 | 69.3 | |
| Other race/ethnicity | 3 | 8.6 | 1 | 2.8 | 2 | 4.2 | 102 | 2.3 | |
| Income | |||||||||
| <$15,000 | 47 | 32.2 | 56 | 19.2 | 39 | 34.0 | 1991 | 19.7 | 0.16 |
| $15,000–<$35,000 | 42 | 26.8 | 55 | 28.3 | 26 | 24.7 | 2322 | 24.0 | |
| $35,000–<$75,000 | 27 | 23.3 | 40 | 27.7 | 25 | 24.1 | 2341 | 32.2 | |
| $75,000+ | 17 | 17.8 | 34 | 24.8 | 16 | 17.1 | 1616 | 24.1 | |
| Education | |||||||||
| 0–11 years | 32 | 30.6 | 61 | 26.8 | 33 | 33.6 | 1756 | 17.1 | <0.001 |
| 12 years | 48 | 26.2 | 47 | 20.5 | 25 | 26.8 | 2426 | 31.3 | |
| 13+ years | 53 | 43.3 | 77 | 52.7 | 48 | 39.6 | 4088 | 51.6 | |
| Marital Status | |||||||||
| Married/cohabitating | 55 | 52.8 | 80 | 49.1 | 45 | 43.9 | 4040 | 54.8 | 0.008 |
| Divorced/separated/widowed | 37 | 27.7 | 48 | 17.1 | 29 | 25.5 | 2205 | 25.4 | |
| Never married | 41 | 19.4 | 57 | 33.8 | 32 | 30.6 | 2025 | 19.8 | |
| Work status | |||||||||
| Employed | 85 | 65.3 | 107 | 65.2 | 54 | 54.3 | 5114 | 59.2 | 0.14 |
| Unemployed | 15 | 8.6 | 18 | 8.1 | 19 | 17.5 | 691 | 7.2 | |
| Not in labor force | 33 | 26.1 | 60 | 26.7 | 33 | 28.1 | 2456 | 33.5 | |
| Ever smoke (yes vs. no) | 59 | 52.0 | 89 | 54.6 | 44 | 37.3 | 2885 | 42.2 | 0.032 |
| Body mass index (kg/m2) | |||||||||
| <18.5 | 3 | 4.1 | 8 | 3.7 | 3 | 2.4 | 318 | 5.7 | <0.001 |
| 18.5–<25 | 33 | 26.1 | 34 | 18.4 | 41 | 42.8 | 3208 | 42.5 | |
| 25–<30 | 28 | 22.7 | 50 | 31.2 | 27 | 25.0 | 2296 | 27.6 | |
| ≥30 | 64 | 47.2 | 88 | 46.7 | 33 | 29.7 | 2243 | 24.2 | |
| Reproductive factors | |||||||||
| Age at menarche (weighted mean, SEM) | 12.4 | 0.2 | 12.2 | 0.2 | 12.7 | 0.2 | 12.7 | 0.04 | 0.004 |
| Ever use birth control (yes vs. no) | 100 | 77.4 | 122 | 77.2 | 64 | 60.9 | 5321 | 67.3 | 0.029 |
| Given birth (yes vs. no) | 98 | 70.3 | 138 | 73.1 | 87 | 75.3 | 6346 | 74.7 | 0.88 |
| Mental health comorbidities | |||||||||
| Lifetime major depression (yes vs. no) | 50 | 38.2 | 65 | 30.2 | 23 | 33.9 | 1659 | 19.3 | <0.001 |
| Lifetime dysthymia | 27 | 27.4 | 29 | 13.4 | 10 | 12.1 | 395 | 4.7 | <0.001 |
| Lifetime bipolar disorder | 17 | 11.8 | 11 | 7.4 | 2 | 1.6 | 149 | 1.9 | <0.001 |
| Agoraphobia | 24 | 13.8 | 21 | 7.9 | 6 | 2.9 | 279 | 2.6 | <0.001 |
| Generalized anxiety disorder | 26 | 17.4 | 28 | 12.6 | 8 | 5.8 | 751 | 9.3 | 0.042 |
| Panic disorder | 26 | 13.6 | 29 | 13.4 | 9 | 11.6 | 470 | 5.6 | <0.001 |
| Social phobia | 41 | 33.5 | 50 | 27.2 | 20 | 25.4 | 967 | 11.3 | <0.001 |
| Substance abuse | 23 | 21.1 | 38 | 20.9 | 10 | 10.9 | 613 | 7.9 | <0.001 |
| Posttraumatic stress disorder | 41 | 34.9 | 36 | 21.3 | 14 | 12.0 | 822 | 9.0 | <0.001 |
| Premenstrual conditions | |||||||||
| PMS/PMDD | |||||||||
| PMDD | 23 | 17.4 | 25 | 10.7 | 6 | 3.4 | 312 | 3.9 | <0.001 |
| PMS, no PMDD | 72 | 55.4 | 93 | 48.9 | 53 | 59.1 | 3271 | 42.0 | |
| No PMDD or PMS | 38 | 27.3 | 67 | 40.4 | 47 | 37.4 | 4687 | 54.0 | |
PMS, PMDD and prevalence of eating disorders
In unadjusted analysis, the prevalence of both PMDD and PMS were greater among those with BN, BED or subthreshold BED as compared to those without BN, BED or subthreshold BED. Among those without BN, BED, or subthreshold BED, prevalence of PMDD was 3.9% and PMS was 42.0%. Comparatively, among those with BN, 17.4% had PMDD and 55.4% had PMS; among those with BED, 10.7% had PMDD and 48.9% had PMS; and, among those with subthreshold BED, 3.4% had PMDD and 59.1% had PMS. Table 2 presents the results of the multinomial logistic regression models for lifetime prevalence of PMS and PMDD and lifetime prevalence of binge-type eating disorders. In bivariate models, PMDD was associated with almost 9-times the odds of BN (OR 8.7, 95% CI 4.0, 19.1) and PMS was associated with 2.6-times the odds of BN (OR 2.6, 95% CI 1.5, 4.6). Additionally, PMDD was significantly associated with greater than 3-times the odds of BED (OR 3.6, 95% CI 1.8, 7.4) and PMS was associated with 2-times the odds of subthreshold BED (OR 2.0, 95% CI 1.2, 3.5). Results were largely unchanged in the a priori adjusted model, with PMDD associated with 9-times the odds of BN (9.4, 95% CI 3.9, 22.5) and PMS associated with 2.7-times the odds of BN (OR 2.7, 95% CI 1.4, 5.2). In the a priori model, PMDD was significantly associated with more than 3-times the odds of BED (OR 3.6, 95% CI 1.7, 7.6) and PMS was associated with 2-times the odds of subthreshold BED (OR 2.1, 05% CI 1.2, 3.7). After combining participants with full BED and subthreshold BED into one group, both PMDD and PMS were associated with a greater odds of combined full and subthreshold BED in both the unadjusted and a priori adjusted models.
Table 2.
Association of premenstrual syndrome and premenstrual dysphoric disorder with bulimia nervosa and binge-eating disorders; multinomial logistic regression analyses
| BN (n=133)
|
BED (n=185)
|
Subthreshold BED (n=106)
|
|||||||
|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | |
| Bivariate | |||||||||
| PMS/PMDD | |||||||||
| PMDD | 8.72 | (3.98, 19.14) | <0.001 | 3.64 | (1.79, 7.41) | <0.001 | 1.26 | (0.50, 3.15) | 0.63 |
| PMS, no PMDD | 2.61 | (1.47, 4.62) | 0.001 | 1.55 | (0.94, 2.57) | 0.087 | 2.03 | (1.17, 3.53) | 0.012 |
| No PMDD or PMS | ref | ref | ref | ||||||
| A priori multivariable model | |||||||||
| PMS/PMDD | |||||||||
| PMDD | 9.35 | (3.89, 22.49) | <0.001 | 3.55 | (1.67, 7.55) | 0.001 | 1.50 | (0.56, 4.03) | 0.42 |
| PMS, no PMDD | 2.72 | (1.44, 5.15) | 0.002 | 1.63 | (0.95, 2.79) | 0.08 | 2.10 | (1.18, 3.73) | 0.011 |
| No PMDD or PMS | ref | ref | ref | ||||||
| MDD | 2.00 | (1.13, 3.56) | 0.018 | 1.49 | (0.91, 2.44) | 0.12 | 2.77 | (1.48, 5.17) | 0.001 |
| Full multivariable model | |||||||||
| PMS/PMDD | |||||||||
| PMDD | 7.17 | (2.30, 22.42) | 0.001 | 2.29 | (0.90, 5.80) | 0.08 | 1.65 | (0.48, 5.63) | 0.43 |
| PMS, no PMDD | 2.51 | (1.11, 5.67) | 0.027 | 1.27 | (0.68, 2.39) | 0.45 | 1.82 | (0.80, 4.18) | 0.16 |
| No PMDD or PMS | ref | ref | ref | ||||||
| MDD | 1.28 | (0.52, 3.16) | 0.59 | 1.23 | (0.66, 2.26) | 0.52 | 3.14 | (1.28, 7.71) | 0.012 |
| Dysthymia | 4.88 | (1.87, 12.72) | 0.001 | 1.44 | (0.65, 3.19) | 0.37 | 2.21 | (0.70, 7.03) | 0.18 |
| Bipolar | 2.27 | (0.76, 6.76) | 0.14 | 1.73 | (0.61, 4.90) | 0.30 | 0.80 | (0.15, 4.14) | 0.79 |
| Social phobia | 2.31 | (0.98, 5.42) | 0.055 | 1.77 | (0.93, 3.38) | 0.09 | 1.88 | (0.74, 4.80) | 0.19 |
| Substance abuse | 0.93 | (0.43, 2.00) | 0.85 | 1.83 | (0.98, 3.41) | 0.059 | 1.03 | (0.30, 3.59) | 0.96 |
A priori model adjusted for age, race/ethnicity, income, education, body mass index, age at menarche, lifetime use of birth control and lifetime history of major depressive disorder (MDD).
Full model adjusted for a priori covariates and additionally for dysthymia, bipolar disorder, social phobia and substance use disorder.
In the full adjusted model, the association between PMDD and BN was attenuated but remained significant (OR 7.2, 95% CI 2.3, 22.4), as was the association between PMS and BN (OR 2.5, 95% CI 1.1, 5.7). The association between PMDD and BED and PMS and subthreshold BED were no longer significant. After combining participants with full and subthreshold BED into one group, both PMDD and PMS were no longer significantly associated with combined full and subthreshold BED. Among the five comorbid mental health conditions adjusted for in the full model, all of which were significantly associated with eating disorders in bivariate analyses (Table 1), only dysthymia (OR 4.9, 95% CI 1.9, 12.7) remained significantly associated with BN.
PMS, PMDD and duration of eating disorders
Table 3 presents the results of the linear regression models for duration of each eating disorder in the presence of PMS and PMDD. In bivariate models, lifetime history of PMDD versus no PMS or PMDD was associated with a significant 11-year greater duration of BN (mean duration of BN 20.5, 95% CI 12.6, 28.4 years among those with PMDD; mean duration of BN 9.4, 95% CI 4.1, 14.7 years among those without PMDD, p<0.01). In contrast, PMS was not significantly associated with duration of BN, and neither PMDD nor PMS were significantly associated with duration of BED or subthreshold BED. In multivariable models, the duration of BN for women with PMDD as compared to those without PMS remained significant in the a priori model (MD 6.6, 95% CI 0.4, 12.8) but was no longer significant in the full model (MD 6.6, 95% CI −0.2, 13.3).
Table 3.
Association of premenstrual syndrome and premenstrual dysphoric disorder with duration of bulimia nervosa and binge-eating disorder symptoms; linear regression analyses
| BN (n=133)
|
BED (n=185)
|
Subthreshold BED (n=106)
|
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Mean difference | 95% CI | P | Mean difference | 95% CI | P | Mean difference | 95% CI | P | |
| Bivariate | |||||||||
| PMS/PMDD | |||||||||
| PMDD | 11.1 | (1.5, 20.8) | 0.024 | −5.3 | (−15.4, 4.7) | 0.29 | −0.6 | (−11.4, 10.2) | 0.91 |
| PMS, no PMDD | 2.7 | (−3.5, 8.8) | 0.39 | −5.2 | (−13.3, 2.9) | 0.21 | −3.9 | (−11.6, 3.8) | 0.32 |
| No PMDD or PMS | ref | ref | ref | ||||||
| A priori multivariable model | |||||||||
| PMS/PMDD | |||||||||
| PMDD | 6.6 | (0.4, 12.8) | 0.038 | 0.2 | (−9.2, 9.6) | 0.96 | 0.2 | (−8.0, 8.4) | 0.96 |
| PMS, no PMDD | −2.5 | (−6.9, 2.0) | 0.28 | −2.7 | (−8.2, 2.8) | 0.33 | −3.9 | (−9.1, 1.4) | 0.15 |
| No PMDD or PMS | ref | ref | ref | ||||||
| MDD | −1.5 | (−4.9, 1.8) | 0.37 | −2.4 | (−7.6, 2.8) | 0.36 | −4.5 | (−10.7, 1.8) | 0.16 |
| Full multivariable model | |||||||||
| PMS/PMDD | |||||||||
| PMDD | 6.6 | (−0.2, 13.3) | 0.057 | 2.4 | (−8.0, 12.8) | 0.65 | −6.9 | (−15.7, 2.0) | 0.13 |
| PMS, no PMDD | −2.6 | (−8.0, 2.9) | 0.36 | −2.5 | (−9.2, 4.3) | 0.47 | −8.4 | (−15.5, −1.3) | 0.022 |
| No PMDD or PMS | ref | ref | ref | ||||||
| MDD | −4.0 | (−8.9, 0.9) | 0.11 | −1.5 | (−8.1, 5.1) | 0.65 | −2.5 | (−14.1, 9.0) | 0.66 |
| Dysthymia | 5.9 | (1.2, 10.6) | 0.015 | 5.0 | (−3.8, 13.7) | 0.26 | −0.9 | (−12.6, 10.8) | 0.88 |
| Bipolar | −3.6 | (−11.7, 4.5) | 0.39 | 3.6 | (−7.7, 14.9) | 0.53 | 3.0 | (−21.3, 27.3) | 0.80 |
| Social phobia | 0.6 | (−3.2, 4.4) | 0.75 | 1.5 | (−6.5, 9.5) | 0.71 | 3.7 | (−5.2, 12.6) | 0.41 |
| Substance abuse | −1.3 | (−5.8, 3.2) | 0.57 | 4.5 | (−1.3, 10.2) | 0.13 | 14.6 | (4.0, 25.2) | 0.008 |
A priori model adjusted for age, race/ethnicity, income, education, body mass index, age at menarche, lifetime use of birth control and lifetime history of major depressive disorder.
Full model adjusted for a priori covariates and additionally for dysthymia, bipolar disorder, social phobia and substance use disorder.
DISCUSSION
In this cross-sectional data set drawn from a large, nationally-representative, ethnically diverse epidemiological survey, we found that both PMS and PMDD were significantly associated with a higher odds of BN in multivariable models adjusting for age, race/ethnicity, household income, education, age at menarche, lifetime use of birth control and major depressive disorder. These associations remained significant after adjusting for lifetime history of other comorbid mental health conditions (dysthymia, bipolar disorder, social phobia and substance use disorder). Additionally, PMDD was significantly positively associated with BED and PMS with subthreshold BED in models adjusting for sociodemographic factors, reproductive health factors and major depressive disorder, although results for these associations were attenuated and no longer significant when adjusting for other comorbid mental health conditions. Finally, PMDD was significantly associated with a longer reported duration of BN in bivariate models and models adjusting for sociodemographic factors, reproductive health factors and major depressive disorder, but not in models adjusting for other comorbid mental health conditions.
This is the first study to evaluate the association between PMS, PMDD and eating disorders in a nationally representative epidemiological sample. Our findings of an association between lifetime PMDD and lifetime BN adds to previous research that has reported an exacerbation of bulimic symptoms during the late luteal phase among women with and without premenstrual symptoms (25–27, 38, 39) and evaluating eating- and body image-related symptoms among women with PMDD but who do not have diagnosed eating disorders (22–24). As these studies suggest, it is possible that women with PMDD are at greater risk for developing binge eating because they experience greater pre-menstrual negative affect and appetite for high-calorie foods, which, in turn, leads to binge eating (22, 39, 40). Alternatively, it is possible that some women are particularly vulnerable to the impact of ovarian hormones, and that variations in ovarian hormones may increase risk for both binge eating and negative affect, thereby increasing risk of both disorders simultaneously. The high degree of comorbidity found in this study for both PMDD and PMS with BN suggests that PMS and PMDD may be important conditions to screen for and take into consideration in treatment of BN. While we found a significantly longer duration of bulimic symptoms among women with PMDD in the a priori adjusted models (mean difference 6.6, 95% CI 0.4, 12.8), prospective studies are needed to evaluate whether women with BN and concurrent comorbid PMDD have a longer time to recovery than women with BN who do not have comorbid PMDD.
Although our data suggest comorbidity between premenstrual conditions and BN, future research is needed to indicate whether onset of premenstrual conditions precede onset of BN, and therefore may plausibly contribute to risk of BN. Experience of premenstrual symptoms after puberty may directly contribute to risk of BN through regular exacerbation of subclinical disordered eating, increasing the likelihood of development of full-blown BN. Early identification and treatment of PMDD and PMS among young women with subclinical binge eating could therefore be an important step in preventing development of full-blown BN. It is also possible that the sensitivity to fluctuations in ovarian hormones indicated by PMS and PMDD may indicate a shared genetic risk with BN, and that women who experience one or both may share a similar genetic vulnerability to the presence of ovarian hormones (28). Future research identifying the mechanisms underlying this vulnerability may suggest therapeutic approaches that are beneficial to women with both bulimia nervosa and premenstrual conditions.
Given the greater prevalence of BED versus BN in our sample, it seems unlikely that the lack of association between BED and premenstrual dysphoria was a result of lower statistical power in the BED group. Indeed, when we combined participants with full and subthreshold BED into one group, the results remained non-significant in the fully adjusted model. One possible reason for this differential finding is a higher level of dietary restraint in the BN group. Specifically, a previous study identifying hormonal predictors of eating disorder symptoms across the menstrual cycle among community women highlighted a trend-level interaction between estradiol and dietary restraint, in which emotional eating was elevated in the premenstrual phase among individuals with high restraint scores only (39). It is therefore possible that the individuals with BN (who typically exhibit high levels of dietary restraint) are more vulnerable than those with BED (who may or may not exhibit high levels of dietary restraint) to the impact of ovarian hormones, putting them at greater risk for both emotional eating in the premenstrual phase and for premenstrual dysphoria.
Although our study is unique in its description of the comorbidity of PMS and PMDD with eating disorders, our data must be interpreted in the context of certain limitations. Potential confounding of the relationship between premenstrual conditions and BN by other mental health comorbidities is a potential concern, although the Collaborative Psychiatric Epidemiological Surveys are uniquely suited to evaluate this confounding, as they include validated DSM-IV diagnoses of a broad range of mood, anxiety and other mental health conditions. In our analyses, after assessing a range of mental health conditions as confounders, we found that major depressive disorder, dysthymia, bipolar disorder, social phobia and substance abuse disorder appeared to partially explain the association between premenstrual conditions and bulimia by attenuating the effect estimate, although the association between premenstrual conditions and bulimia remained statistically significant. A second potential limitation is the use of the WMH-CIDI for assessing eating disorders. Specifically, our assessment of BN, BED, and subthreshold BED was specific to DSM-IV and would therefore have missed clinically significant presentations that now appear in DSM-5, including DSM-5 BED with binge eating less than twice per week, or BN/BED of limited frequency or duration (now examples of Other Specified Feeding or Eating Disorder). Additionally, the CIDI has been shown to underdiagnose AN and BN (35), both of which may be reflected in the low prevalence of eating disorders in our sample.
A third limitation is the exclusion of 2,744 potentially eligible women, who differed significantly from the included participants on several demographic characteristics, due to missing data. Although 1,833 of these participants completed the short-form for the National Comorbidity Survey Replication, and weights were incorporated to account for missing data, the additional 911 missing participants could have introduced a selection bias. A fourth limitation of our study was the cross-sectional design, restricting the ability to demonstrate a temporal relationship between premenstrual conditions and BN and BED. Finally, accuracy of recall is a concern in this study due to the survey design, as participants may have difficulty remembering past symptoms and their age at onset. For the analysis of BN and BED duration, misclassification of age at onset could have led to attenuation of the effect estimate. Recall bias is possible as participants with premenstrual conditions, BN or BED may be more likely to recall other psychological symptoms, which could exaggerate the effect estimates. This potential bias is partially addressed by adjusting for additional psychological comorbidities, such as major depressive disorder, which may serve as a partial indicator of this form of recall bias. Our findings are likely generalizable to all women at risk for premenstrual conditions and without a lifetime history of AN, due to the use of a nationally-representative survey.
Our finding of a significant association between lifetime PMDD and PMS and lifetime BN, independent of comorbid mental health conditions, in a nationally-representative survey suggests that women with BN are more likely to experience the cyclical physiological and psychological effects of PMDD than women who have no history of a DSM-IV eating disorder. Given past research highlighting a predictable exacerbation of bulimic symptoms in the late-luteal phase of the menstrual cycle, comorbid premenstrual conditions may be an important consideration in the course of treatment for binge-type eating disorders. Future research incorporating a longitudinal design would help to better elucidate the potential for premenstrual conditions to affect the onset and duration of BN and other eating disorders. Additionally, the high degree of comorbidity between PMDD and BN found in this study suggests that evaluating the presence of and potential exacerbation of binge-type eating disorder symptoms by PMDD may be important in clinical practice.
Acknowledgments
Time for this manuscript was funded in part by a grant from the National Institute of Mental Health (NIMH K23 MH096029-01A1) awarded to Dr. Luana Marques
Footnotes
Disclosure of Conflicts: None of the authors have a conflict of interest to disclose.
References
- 1.Westmoreland P, Krantz MJ, Mehler PS. Medical Complications of Anorexia Nervosa and Bulimia. The American journal of medicine. 2015 doi: 10.1016/j.amjmed.2015.06.031. [DOI] [PubMed] [Google Scholar]
- 2.Kessler RC, Berglund PA, Chiu WT, Deitz AC, Hudson JI, Shahly V, et al. The prevalence and correlates of binge eating disorder in the World Health Organization World Mental Health Surveys. Biological psychiatry. 2013;73(9):904–14. doi: 10.1016/j.biopsych.2012.11.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Agh T, Kovacs G, Pawaskar M, Supina D, Inotai A, Voko Z. Epidemiology, health-related quality of life and economic burden of binge eating disorder: a systematic literature review. Eating and weight disorders : EWD. 2015;20(1):1–12. doi: 10.1007/s40519-014-0173-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Pisetsky EM, Thornton LM, Lichtenstein P, Pedersen NL, Bulik CM. Suicide attempts in women with eating disorders. Journal of abnormal psychology. 2013;122(4):1042–56. doi: 10.1037/a0034902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Arcelus J, Mitchell AJ, Wales J, Nielsen S. Mortality rates in patients with anorexia nervosa and other eating disorders. A meta-analysis of 36 studies. Archives of general psychiatry. 2011;68(7):724–31. doi: 10.1001/archgenpsychiatry.2011.74. [DOI] [PubMed] [Google Scholar]
- 6.Smink FR, van Hoeken D, Hoek HW. Epidemiology of eating disorders: incidence, prevalence and mortality rates. Current psychiatry reports. 2012;14(4):406–14. doi: 10.1007/s11920-012-0282-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ulfvebrand S, Birgegard A, Norring C, Hogdahl L, von Hausswolff-Juhlin Y. Psychiatric comorbidity in women and men with eating disorders results from a large clinical database. Psychiatry research. 2015 doi: 10.1016/j.psychres.2015.09.008. [DOI] [PubMed] [Google Scholar]
- 8.Fichter MM, Quadflieg N, Hedlund S. Long-term course of binge eating disorder and bulimia nervosa: relevance for nosology and diagnostic criteria. The International journal of eating disorders. 2008;41(7):577–86. doi: 10.1002/eat.20539. [DOI] [PubMed] [Google Scholar]
- 9.Vall E, Wade TD. Predictors of treatment outcome in individuals with eating disorders: A systematic review and meta-analysis. The International journal of eating disorders. 2015;48(7):946–71. doi: 10.1002/eat.22411. [DOI] [PubMed] [Google Scholar]
- 10.Biggs WS, Demuth RH. Premenstrual syndrome and premenstrual dysphoric disorder. American family physician. 2011;84(8):918–24. [PubMed] [Google Scholar]
- 11.Franz WB., 3rd Basic review: endocrinology of the normal menstrual cycle. Primary care. 1988;15(3):607–16. [PubMed] [Google Scholar]
- 12.O’Brien S, Rapkin A, Dennerstein L, Nevatte T. Diagnosis and management of premenstrual disorders. BMJ (Clinical research ed) 2011;342:d2994. doi: 10.1136/bmj.d2994. [DOI] [PubMed] [Google Scholar]
- 13.Dennerstein L, Lehert P, Backstrom TC, Heinemann K. Premenstrual symptoms -- severity, duration and typology: an international cross-sectional study. Menopause international. 2009;15(3):120–6. doi: 10.1258/mi.2009.009030. [DOI] [PubMed] [Google Scholar]
- 14.ACOG committee opinion. Premenstrual syndrome. Number 155--April 1995 (replaces no. 66, January 1989) Committee on Gynecologic Practice. American College of Obstetricians and Gynecologists. International journal of gynaecology and obstetrics: the official organ of the International Federation of Gynaecology and Obstetrics. 1995;50(1):80–4. [PubMed] [Google Scholar]
- 15.Association AP. Diagnostic and Statistical Manual of Mental Disorders:: DSM-5: ManMag. 2003. [Google Scholar]
- 16.Yonkers KA, O’Brien PS, Eriksson E. Premenstrual syndrome. The Lancet. 2008;371(9619):1200–10. doi: 10.1016/S0140-6736(08)60527-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hartlage SA, Brandenburg DL, Kravitz HM. Premenstrual exacerbation of depressive disorders in a community-based sample in the United States. Psychosomatic medicine. 2004;66(5):698–706. doi: 10.1097/01.psy.0000138131.92408.b9. [DOI] [PubMed] [Google Scholar]
- 18.Rubinow DR, Schmidt PJ. Gonadal steroid regulation of mood: the lessons of premenstrual syndrome. Frontiers in neuroendocrinology. 2006;27(2):210–6. doi: 10.1016/j.yfrne.2006.02.003. [DOI] [PubMed] [Google Scholar]
- 19.Rapkin AJ, Akopians AL. Pathophysiology of premenstrual syndrome and premenstrual dysphoric disorder. Menopause international. 2012;18(2):52–9. doi: 10.1258/mi.2012.012014. [DOI] [PubMed] [Google Scholar]
- 20.Marjoribanks J, Brown J, O’Brien PM, Wyatt K. Selective serotonin reuptake inhibitors for premenstrual syndrome. The Cochrane database of systematic reviews. 2013;6:Cd001396. doi: 10.1002/14651858.CD001396.pub3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Pinkerton JV, Guico-Pabia CJ, Taylor HS. Menstrual cycle-related exacerbation of disease. American journal of obstetrics and gynecology. 2010;202(3):221–31. doi: 10.1016/j.ajog.2009.07.061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ko CH, Yen CF, Long CY, Kuo YT, Chen CS, Yen JY. The late-luteal leptin level, caloric intake and eating behaviors among women with premenstrual dysphoric disorder. Psychoneuroendocrinology. 2015;56:52–61. doi: 10.1016/j.psyneuen.2015.03.002. [DOI] [PubMed] [Google Scholar]
- 23.Yen JY, Chang SJ, Ko CH, Yen CF, Chen CS, Yeh YC, Chen CC. The high-sweet-fat food craving among women with premenstrual dysphoric disorder: emotional response, implicit attitude and rewards sensitivity. Psychoneuroendocrinology. 2010;35(8):1203–12. doi: 10.1016/j.psyneuen.2010.02.006. [DOI] [PubMed] [Google Scholar]
- 24.Hildebrandt BA, Racine SE, Keel PK, Burt SA, Neale M, Boker S, et al. The effects of ovarian hormones and emotional eating on changes in weight preoccupation across the menstrual cycle. The International journal of eating disorders. 2015;48(5):477–86. doi: 10.1002/eat.22326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lester NA, Keel PK, Lipson SF. Symptom fluctuation in bulimia nervosa: relation to menstrual-cycle phase and cortisol levels. Psychol Med. 2003;33(1):51–60. doi: 10.1017/s0033291702006815. [DOI] [PubMed] [Google Scholar]
- 26.Gladis MM, Walsh BT. Premenstrual exacerbation of binge eating in bulimia. The American journal of psychiatry. 1987;144(12):1592–5. doi: 10.1176/ajp.144.12.1592. [DOI] [PubMed] [Google Scholar]
- 27.Price WA, Torem MS, DiMarzio LR. Premenstrual exacerbation of bulimia. Psychosomatics. 1987;28(7):378–9. doi: 10.1016/s0033-3182(87)72511-0. [DOI] [PubMed] [Google Scholar]
- 28.Klump KL, Burt SA, McGue M, Iacono WG. Changes in genetic and environmental influences on disordered eating across adolescence: a longitudinal twin study. Archives of general psychiatry. 2007;64(12):1409–15. doi: 10.1001/archpsyc.64.12.1409. [DOI] [PubMed] [Google Scholar]
- 29.Klump KL. Puberty as a critical risk period for eating disorders: A review of human and animal studies. Hormones and behavior. 2013;64(2):399–410. doi: 10.1016/j.yhbeh.2013.02.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Pennell BE, Bowers A, Carr D, Chardoul S, Cheung GQ, Dinkelmann K, et al. The development and implementation of the National Comorbidity Survey Replication, the National Survey of American Life, and the National Latino and Asian American Survey. Int J Methods Psychiatr Res. 2004;13(4):241–69. doi: 10.1002/mpr.180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Association AP. Diagnostic and Statistical Manual of Mental Disorders. 4. Washington DC: American Psychiatric Association; 1994. [Google Scholar]
- 32.Thomas JJ, Vartanian LR, Brownell KD. The relationship between eating disorder not otherwise specified (EDNOS) and officially recognized eating disorders: meta-analysis and implications for DSM. Psychological bulletin. 2009;135(3):407. doi: 10.1037/a0015326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Hudson JI, Hiripi E, Pope HG, Jr, Kessler RC. The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biological psychiatry. 2007;61(3):348–58. doi: 10.1016/j.biopsych.2006.03.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Wittchen HU. Reliability and validity studies of the WHO--Composite International Diagnostic Interview (CIDI): a critical review. J Psychiatr Res. 1994;28(1):57–84. doi: 10.1016/0022-3956(94)90036-1. [DOI] [PubMed] [Google Scholar]
- 35.Thornton C, Russell J, Hudson J. Does the Composite International Diagnostic Interview underdiagnose the eating disorders? The International journal of eating disorders. 1998;23(3):341–5. doi: 10.1002/(sici)1098-108x(199804)23:3<341::aid-eat11>3.0.co;2-0. [DOI] [PubMed] [Google Scholar]
- 36.Cohen LS, Soares CN, Otto MW, Sweeney BH, Liberman RF, Harlow BL. Prevalence and predictors of premenstrual dysphoric disorder (PMDD) in older premenopausal women. The Harvard Study of Moods and Cycles. Journal of affective disorders. 2002;70(2):125–32. doi: 10.1016/s0165-0327(01)00458-x. [DOI] [PubMed] [Google Scholar]
- 37.Jackson JS, Torres M, Caldwell CH, Neighbors HW, Nesse RM, Taylor RJ, et al. The National Survey of American Life: a study of racial, ethnic and cultural influences on mental disorders and mental health. International journal of methods in psychiatric research. 2004;13(4):196–207. doi: 10.1002/mpr.177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Klump KL, Keel PK, Racine SE, Burt SA, Neale M, Sisk CL, et al. The interactive effects of estrogen and progesterone on changes in emotional eating across the menstrual cycle. Journal of abnormal psychology. 2013;122(1):131. doi: 10.1037/a0029524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Klump KL, Keel PK, Burt SA, Racine SE, Neale MC, Sisk CL, Boker S. Ovarian hormones and emotional eating associations across the menstrual cycle: an examination of the potential moderating effects of body mass index and dietary restraint. International Journal of Eating Disorders. 2013;46(3):256–63. doi: 10.1002/eat.22084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Reed SC, Levin FR, Evans SM. Changes in mood, cognitive performance and appetite in the late luteal and follicular phases of the menstrual cycle in women with and without PMDD (premenstrual dysphoric disorder) Horm Behav. 2008;54(1):185–93. doi: 10.1016/j.yhbeh.2008.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]