I thought this article might be of interest.
Volume 76, Issue 2 , August 2001, Pages 310-316
Marsha D. Marcus Ph.D., Tammy L. Loucks, M.P.H.a and Sarah L. Berga M.D.
Objective: To determine whether mood, attitudes, or symptoms of disordered eating discriminated women with functional hypothalamic amenorrhea (FHA) from those with organic causes of amenorrhea and eumenorrhea.
Design: Cross-sectional comparison of women with FHA, women with organic amenorrhea, and eumenorrheic control women.
Setting: Clinical research center in an academic medical institution.
Patient(s): Seventy-seven women ≥18 years old with time since menarche ≥5 and ≤25 years were recruited by advertisement.
Intervention(s): Ovulation was confirmed in eumenorrheic control women. Causes of anovulation were carefully documented in amenorrheic participants and LH pulse profiles were obtained to document the diagnosis of FHA. All participants were interviewed and completed questionnaires.
Main Outcome Measure(s): Self-report measures of dysfunctional attitudes, coping styles, and symptoms of depression and eating disorders.
Result(s): Women with FHA reported more depressive symptoms and dysfunctional attitudes than did eumenorrheic women, but not significantly more than women with organic amenorrhea. However, women with FHA reported significantly more symptoms of disordered eating than did either anovulatory or ovulatory women.
Conclusion(s): The findings are consistent with the hypothesis that FHA is precipitated by a combination of psychosocial stressors and metabolic challenge.
Functional hypothalamic amenorrhea or anovulation (FHA) is the cessation of menses and ovulation in women with no identifiable organic cause for the amenorrhea. This common and, in theory, reversible condition accounts for an estimated 15%–35% of cases of secondary amenorrhea [1 and 2].
Recent work has clarified the pathophysiology of FHA and confirmed the relationship between stress and FHA. The proximate cause of FHA is insufficient hypothalamic pulsatile GnRH release, which leads to reduced pituitary secretion of gonadotropins, LH, and FSH, with resulting anovulation. The best endocrine evidence that stress desynchronizes the GnRH neuronal network has been provided by studies showing that the activity of the hypothalamic-pituitary-adrenal axis is increased in women with FHA [3, 4, 5 and 6]. Moreover, we have documented that cortisol levels are increased in women with FHA but not those with eumenorrhea or organic forms of anovulation .
Since Reifenstein  defined hypothalamic amenorrhea as a syndrome in which "overt or latent psychological disturbances" disrupt menstrual functioning, numerous factors, including environmental stressors , personality traits [10 and 11], psychological disorders [12 and 13], exercise [14 and 15], low body weight, and weight loss [16, 17 and 18] have been implicated in the etiology of FHA. In an effort to integrate available data, Berga et al. [4 and 5] proposed a model of FHA in which psychogenic stress and metabolic challenge interact to disrupt or suppress GnRH drive. Specifically, we hypothesized that psychobiological characteristics predispose women to sustained activation of central neural processes in response to commonplace problems. We posited that the combination of mild energy deprivation induced by nutritional or calorie restriction or exercise and psychosocial distress, such as performance pressure, unrealistic goals, and negative attributions, acts synergistically to provoke a constellation of hypothalamic alterations, including disruption of GnRH drive.
In a preliminary investigation, Giles and Berga  reported that dysfunctional attitudes (e.g., perfectionistic performance standards and concerns about the judgments of others) discriminated women with FHA (n = 9) from those with organic amenorrhea (n = 6) and eumenorrheic controls (n = 8). Women with functional or organic causes of amenorrhea tended to report more depressive symptoms and difficulty coping with daily stress compared with eumenorrheic women.
Amenorrheic women were more likely than control women to report a history of mood disorder (relative risk, 2.7), but the difference in risk was not significant. In addition, women with FHA reported more regular exercise and mild weight loss in conjunction with amenorrhea, raising the question of whether women with FHA are more likely to have subthreshold levels of disordered eating.
The sample size in this initial investigation was small, symptoms of disordered eating were not assessed systematically, and women with FHA and those with eumenorrhea were not matched on levels of self-reported exercise. We therefore sought to replicate and extend the findings of our previous study in a larger group of women with FHA. Given our hypothesis that metabolic and psychosocial stressors act synergistically to produce FHA, we expected to find greater degrees of psychological dysfunction and disordered eating in the FHA group.
Women with FHA and eumenorrheic control women were matched on levels of self-reported exercise, and we evaluated whether women with FHA were more likely than control women or women with organic causes of anovulation to report behaviors and attitudes that characterize disordered eating. As in the previous study, we used a comparison group of women with organic causes of anovulation to control for the distress of reproductive impairment.
Materials and methods
Seventy-seven women—28 with FHA, 24 with amenorrhea of documented organic origin, and 25 eumenorrheic women—were recruited by advertisement. All participants signed consent forms approved by the Magee-Women’s Hospital Institutional Review Board. Entry criteria were age since menarche ≥5 and ≤25 years, chronologic age ≥18 years, nonsmoker status, no use of concurrent medications or illicit drugs, no major medical problems other than amenorrhea, time since last delivery ≥6 months, no lactation for ≥6 months, day-awake/night-asleep schedule, and no current Axis I psychiatric disorder. In addition, to exclude women with recent weight changes that could explain functional amenorrhea, participants were required to be 90%–110% of ideal body weight , without a weight loss or gain of more than 10 pounds in the past year or since the onset of amenorrhea, and to report current exercise totaling ≤10 h/wk and running ≤10 miles/wk.
The diagnosis of functional hypothalamic amenorrhea was established by ruling out organic causes, such as hyperprolactinemia or hyperandrogenic anovulation. The LH pulse frequency was significantly lower in women with FHA than in eumenorrheic women [mean [±SE], 5.96 ± 0.71 LH pulses/day vs. 14.08 ± 0.57 LH pulses/day; t(51) = 8.8, P≤001]. Participants with organic amenorrhea included women with the polycystic ovary syndrome (n = 19), premature ovarian failure (n = 1), hyperprolactinemia (n = 1), the Kallmann Syndrome (n = 2), or hydroencephaly (n = 1).
To avoid bias, we enrolled all women with organic causes of anovulation who met other inclusion and exclusion criteria. Eumenorrheic women were studied during days 2–7 of their menstrual cycle and were required to have a history of regular menstrual cycle intervals (28–32 days) and a midluteal progesterone level ≥30 nmol/L during the preceding and current menstrual cycle. Ovulatory and FHA women were similar in age, body mass index, and exercise level. Other biochemical and neuroendocrinologic features of these groups have been reported elsewhere [7 and 39].
Participant characteristics are summarized in Table 1. Participants were predominantly white (81.9%) and well educated (mean, 15.5 ± 2.0 years of education). Groups did not differ in age (F(2,74) = .06; P=.95) or years of education (F(2,71)=.40; P=.67) but did differ significantly in body mass index (BMI) (F(2,74)=3.9; P=.024). The Scheffe tests indicated that women with FHA and ovulatory women had similar BMIs; women with organic causes of anovulation had significantly higher BMIs than did women with FHA, but all women were 90%–110% of ideal body weight.
To rule out current psychiatric disorders, participants were interviewed by using the Structured Interview for the Diagnostic and Statistical Manual of Mental Disorders-IV , a semi-structured interview with established reliability and validity. We used four measures to assess mood, cognitive factors, and coping.
The Beck Depression Inventory  is a widely used 16-item questionnaire designed to assess the severity of self-reported depressive symptoms. Items measure vegetative, cognitive, and mood symptoms. Scores ≥ 10 indicate clinically significant depressive symptoms.
The Hamilton Rating Scale for Depression  is a clinical rating scale for assessing severity of depression. Items relate to mood, self-reproach, sleep disturbance, psychomotor activity, anxiety, loss of interest, change of appetite, decreased libido, loss of insight, and somatic concerns. Scores ≤ 7 indicate absence of depression, whereas scores ≥ 14 indicate clinically significant depression.
The Dysfunctional Attitude Scale  is a 40-item measure that assesses cognitions commonly found in persons predisposed to emotional disorders, particularly depression. These include perfectionistic performance standards, belief in the likelihood of desired outcomes, rigid ideas about events that should occur, and concern about the judgment of others. Scores on the Dysfunctional Attitude Scale range from 40 to 280; higher scores indicate greater dysfunction. Two robust primary factors have been derived from the Dysfunctional Attitude Scale: need for approval and perfectionism .
The Self-Control Scale  is a 36-item measure of learned resourcefulness or ability to cope. Four dimensions are assessed: cognitive strategies, problem solving, delay of gratification, and self-efficacy expectations. Higher scores on the Self-Control Scale reflect better coping ability.
We also sought to determine whether symptoms of disordered eating might discriminate women with FHA from those with organic amenorrhea. A subset of participants (n = 50) completed the Eating Disorder Inventory and the Bulimia Test—Revised to assess maladaptive eating behaviors and attitudes and beliefs that characterize persons with disordered eating.
The Eating Disorder Inventory [27 and 28] is a well-validated and reliable 64-item self-report measure designed to assess the range of psychological and behavioral traits common in persons with disordered eating. The Eating Disorder Inventory has eight subscales: drive for thinness (excessive concern with dieting, preoccupation with weight, and fear of weight gain), bulimia (tendencies to think about and engage in binge eating), body dissatisfaction (dissatisfaction with shape and size of stomach, hips, thighs, and buttocks), ineffectiveness (feelings of general inadequacy, insecurity, worthlessness, and lack of control), perfectionism (belief that one’s personal achievements should be superior), interpersonal distrust (general feelings of alienation and reluctance to form close relationships), interoceptive awareness (confusion in recognizing and responding to emotional states and visceral sensations related to hunger and satiety), and maturity fears (desire to retreat to the security of childhood).
The Bulimia Test—Revised (BULIT-R) [29 and 30] is a psychometrically sound 36-item self-report questionnaire that assesses behaviors and attitudes specific to bulimia nervosa. Scores ≥104 on BULIT-R are suggestive of diagnosable bulimia nervosa; however, the investigators have suggested that scores ≥85 warrant further evaluation.
Participant characteristics were compared by using analysis of variance for continuous variables and nonparametric techniques for categorical variables. Analysis of variance also was used to evaluate differences among participant groups; the Scheffe post hoc test was used to test for significance of pairwise comparisons when the overall F ratio was significant. Finally, we tabulated the frequency of participant group responses to individual BULIT-R items to illustrate examples of specific behaviors endorsed by women with FHA.
Measures of mood, attitudes, and coping
Table 2 shows a summary of scores on measures of depression, dysfunctional attitudes, and coping. Groups differed significantly in scores on the Beck Depression Inventory (F(2,69)=4.5; P=.02) and Hamilton Rating Scale for Depression (F(2,71)=4.3; P=.02). The Post hoc Scheffe tests revealed that women with FHA reported significantly higher levels of depressive symptoms on the Beck Depression Inventory and Hamilton Rating Scale for Depression than did eumenorrheic women. Scores on the Beck Depression Inventory and Hamilton Rating Scale for Depression in women with organic causes of anovulation fell mid-way between those of women with FHA and those of eumenorrheic women but did not differ significantly from the scores of either group.
The pattern of results was similar for the Dysfunctional Attitude Scale (DAS) and Self-Control Scale. There were significant between group differences overall on the DAS, F(2,69)=5.0, P=.01. Scheffe’s tests indicated that FHA women reported significantly more dysfunctional attitudes on the DAS in comparison to eumenorrheic control subjects. Although DAS scores of FHA women were higher than scores of OA women, the difference was not significant. Women with FHA also had significantly higher scores on the need for approval subscale of the Dysfunctional Attitude Scale (F(2,69)=4.9; P=.01) but not the perfectionism subscale (F(2,69)=2.8; P=.07) than did ovulatory control women. Women with organic causes of anovulation women had scores on the two Dysfunctional Attitude Scale subscales that were intermediate between those of women with FHA and eumenorrheic women.
Finally, the overall group difference in Self-Control Scale scores was not significant (F(2,70)=2.8, P=.07), but Self-Control Scale scores were lower in women with FHA than in eumenorrheic women, and scores of women with FHA and those with organic causes of anovulation were similar.
Measures of eating disorder symptoms
A subset of 50 participants completed the Eating Disorder Inventory and BULIT-R. As in the overall sample, women with FHA, those with organic causes of anovulation, and eumenorrheic controls did not differ in age (F(2,48)=.57; P=.57) or years of education (F(2,47)=.60, P=.55). However, in contrast to the sample as a whole, the BMI in the subgroup did not differ in women with FHA, those with organic causes of anovulation, and eumenorrheic women (21.1, 22.7, and 21.9 kg/m2, respectively; F(2,48)=2.3; P=.11).
Table 3 shows group comparison data on the Eating Disorder Inventory and BULIT-R. Women with FHA, those with organic causes of anovulation, and eumenorrheic women differed significantly on four of the eight subscales of the Eating Disorder Inventory: drive for thinness (F(2,47)=7.55,; P=.001), bulimia (F(2,47)=7.63; P=.001), ineffectiveness (F(2,47)=4.65; P=.014); and interoceptive awareness (F(2,47)=5.30; P=.008). Scheffe’s tests revealed that women with FHA had significantly higher scores on each of these subscales than did women with organic causes of anovulation or eumenorrheic women; scores of women with organic causes of anovulation and eumenorrheic women did not significantly differ.
Similar findings were observed on the BULIT-R. Compared with women with organic causes of anovulation or eumenorrheic women, women with FHA reported significantly more bulimic symptoms on the BULIT-R (F(2,47)=13.01; P=.000). Half of the women with FHA indicated that they had binge-eating episodes at least 2–3 times per month (Table 4). Similarly, women with FHA had greater preoccupation with size, shape, weight, and food. No women self-reported laxative or diuretic misuse, but 18.8% (n = 3) of women with FHA repeated self-induced vomiting.
To our knowledge, our study is the most complete current assessment of behavioral and psychological correlates in women with FHA. Our major finding is that symptoms of disordered eating discriminate women with FHA from those with organic amenorrhea or eumenorrhea. Specifically, women with FHA report more concerns about dieting and weight, fear of weight gain, and tendencies to engage in binge eating.
Women with FHA also have higher levels of psychological characteristics associated with disordered eating, such as feelings of general inadequacy, insecurity, and lack of control over life and confusion in identifying and responding to food-related body sensations and emotional states. Moreover, they display dysfunctional attitudes compared with eumenorrheic women, but the presence of dysfunctional attitudes in women with organic causes of anovulation indicates that some of this increase may be attributable to reproductive impairment rather than a specific feature of FHA.
Other investigations have implicated weight change and exercise in the pathogenesis of FHA [12, 13, 14 and 15]. Furthermore, previous studies have found high rates of eating disorders among amenorrheic [31 and 32] and infertile women [33 and 34]. However, in contrast to previous studies, we excluded women with frank eating disorders and those who reported significant weight change or high rates of exercise.
As expected, the overall symptoms reported by participants with FHA did not reflect clinical psychopathology. For example, the average BULIT-R score of women with FHA was 67.1, which is well below the cut-off score of 85 recommended by the investigators as indicating need for evaluation for a diagnosable eating disorder . Nevertheless, even though we excluded women with diagnosable eating disorder from our FHA group, several participants reported significant symptoms of disordered eating. Thus, there were subtle but detectable differences in eating disorder symptoms between women with FHA and those with regular cycles or other forms of amenorrhea. This finding is consistent with the hypothesis that mild metabolic challenge sensitizes the reproductive axis to the effects of subsequent or ongoing psychosocial stressors .
Our findings partially replicate those of our earlier study, which documented elevated rates of depressive symptoms and dysfunctional attitudes in amenorrheic women. Reported depressive symptoms were similar in the current study and the previous study . For example, Beck Depression Inventory scores of women with FHA averaged 7.9 in the current study and 7.4 in the previous study. In our earlier study, however, depression scores did not differ among women with FHA, those with organic amenorrhea, and those with eumenorrhea, perhaps reflecting the small sample.
Our current findings indicate that women with FHA or organic amenorrhea have significantly more symptoms of depression than do ovulatory controls. However, levels of depressive symptoms in FHA and organic causes of anovulation women in both studies were similar, suggesting that distress may be associated with amenorrhea of functional or organic origin.
As was the case with the symptoms of eating disorders, the overall level of depressive symptoms among amenorrheic women was lower than the clinical threshold suggestive of diagnosable disorder (e.g., scores on the Beck Depression Inventory ≥ 10 are associated with clinically relevant depression). This finding indicates that amenorrheic women have more symptoms than do eumenorrheic women but that differences are relatively subtle.
Likewise, women with FHA reported more dysfunctional attitudes, particularly those associated with need for approval. Specifically, women with FHA were more likely than eumenorrheic women to endorse attitudes that are prevalent among persons vulnerable to depression, such as perfectionistic standards and concern about the judgment of others. Women with organic forms of amenorrhea reported dysfunctional attitude scores that fell between those of FHA and eumenorrheic women, and their scores did not differ significantly from those in either anovulatory group. Finally, we observed a nonsignificant trend for anovulatory women to report less learned resourcefulness or ability to cope compared with ovulatory women, replicating the pattern identified in measures of depressive symptoms and dysfunctional attitudes.
Taken together, our findings about mood, attitude, and coping measures corroborate that reproductive impairment in general is associated with mild psychological dysfunction. As noted, these findings suggest that although the degree of psychosocial distress is marginally greater in women with FHA than in those with organic causes of anovulation, psychological dysfunction is not specific to FHA.
In summary, we found that women with FHA report elevated rates of eating disorder symptoms, potentially maladaptive cognitive traits, and depressive symptoms. These findings are consistent with the hypothesis that metabolic challenge and the stress associated with daily living are involved in the pathogenesis of FHA and that metabolic challenge may amplify the endocrine effects of psychological challenge.
Since both the disordered eating behavior and psychological symptoms reported by women with FHA are in theory reversible and amenable to psychological treatment, these women may benefit from psychological interventions that target disordered eating and the other psychological factors. However, because women with FHA do not usually meet criteria for the diagnosis of a psychiatric disorder, physicians may not recognize these patients as needing psychological intervention. Moreover, since the difficulties experienced by patients with FHA may be subtle, mental health professionals may not feel that treatment is warranted or that treatment would be supported by third-party payers.
Accordingly, we developed and are testing a cognitive-behavioral intervention adapted from an empirically supported treatment for disordered eating  in women with FHA. The 16-week program focuses on adopting healthy eating behavior; modifying dysfunctional attitudes related to eating, shape and weight; and learning more effective strategies for coping with everyday problems. If this cognitive-behavioral intervention is effective in reducing cortisol levels and restoring hypothalamic GnRH drive, it may provide an alternative to expensive and risky infertility treatments sought by women with FHA that may lead to multiple gestation, preterm labor, and intrauterine growth retardation . Moreover, women with FHA have hypothalamic hypothyroidism that corrects only when activation of the hypothalamic-pituitary-adrenal axis is ameliorated . The ongoing presence of functional hypothyroidism  may impair fetal cognitive development or cause other subtle impairments in fetal development .
Even when pregnancy is not an issue, amenorrheic women are vulnerable to osteopenia and osteoporosis . Sex-steroid replacement is less or not efficacious in promoting bone accretion in the presence of the metabolic derangements that accompany stress, such as those that have been documented in women with FHA .
Finally, the long-term consequences of hypercortisolism seen in women with FHA are not benign, as cortisol is a potent neurotoxic hormone [42 and 43]. Thus, psychological interventions may favorably affect bone health and other health factors. Finally, the cognitive-behavioral intervention may promote general well-being by enhancing problem-solving and coping strategies in women with FHA.
Our findings also have important implications for obstetricians and gynecologists who care for amenorrheic women. The symptoms of disordered eating and associated psychological characteristics reported by many women with FHA are subtle, and information about height and weight alone is unlikely to provide sufficient information to detect these disturbances. Therefore, screening measures such as the BULIT-R or Eating Disorder Inventory, which are readily obtainable [28 and 30], may provide important information in the assessment of patients with amenorrhea. Certainly, routine questions about dietary restriction; preoccupation with eating, weight, and shape; and binge eating are indicated in the evaluation of all amenorrheic patients. Referrals for nutrition consultation may be useful in the treatment of this common health problem.
Similarly, psychological treatment may be helpful; however, in the absence of a diagnosable psychiatric disorder, insurance reimbursement for treatment may be problematic. Our data documenting the morbidity associated with subthreshold eating disorders are therefore important, since they strongly suggest that subclinical symptoms are appropriate targets for psychological intervention.