menstrual abnormalities and subclinical eating disorders

8/7/2019 menstrual abnormalities and subclinical eating disorders

1/6

Menstrual Cycle Abnormalities and Subclinical Eating Disorders: APreliminary Report

RICHARD E. KREIPE, MD, JAINE STRAUSS, PHD, CHRISTOPHER H. HODG-MAN, MD, AND RICHARD M. RYAN, PH D

Menstrual dysfunction is a common concomitant of anorexia nervosa and bulimia. Initialinvestigations emphasized the role of weight loss and lean/fat ratio in amenorrhea. Subsequentstudies suggest a more complex interaction between eating disorders and menstrual status.However, in past investigations, menstrual abnormalities have been confounded with lowweight. We conducted two studies to ascertain the prevalence of menstrual abnormalities in agroup of women with subclinical eating pathology versus an age-, education-, and weight-matched group of normal controls. In Study I, 93.4% of the subclinical subjects reported ahistory of menstrual abnormality as compared to 11.7% of the normal controls. In Study II,100% of the subclinical subjects, versus 15.0% of the controls, reported an abnormal menstrualhistory. These data suggest that menstrual dysfunction often occurs in women with abnormaleating attitudes but without weight loss or diagnosable eating pathology. Several hypothesesfor this finding are proposed.

INTRODUCTION

This paper concerns a serendipitousfinding. In the course of collecting controlsubjects for a study on anorexia nervosaand bulimia, we encountered a subsetwho scored high on the Eating AttitudesTest yet did not evidence diagnosable eat-ing pathology. The psychological featuresof this group, which we labeled "subclin-ical eating disorders,"were quite interest-ing. However, we were most struck by aphysiological finding: these women re-ported a history of significant menstrualdisruption. Understanding the relation-ship between menstrual cycle abnormal-

From the University of Rochester Medical Center,Rochester, New York.

Address reprint requests to: Jaine Strauss, Ph.D.,Department of Psychology, Williams College, Wil-liamstown, MA 01267.

Received July 14,1988; revision received Septem-ber 20, 1988.

ities and subclinical eating pathology isthe focus of this paper.

Amenorrhea, or the cessation or inhi-bition of menstrual periods, has long beenassociated with anorexia nervosa (1). Infact, the DSM-III-R (2) now includesamenorrhea as a necessary condition forthe diagnosis of anorexia nervosa in fe-males. Researchers have also noted theprevalence of menstrual abnormalities inother types of eating disorders. In a studyof normal-weight bulimics, Johnson et al.(3) detected amenorrhea in 20% and men-strual irregularity in 51% of their subjects.Similar findings have been reported inother bulimic samples (4).

Menstrual abnormalities have alsoemerged in high-risk, nonclinical sam-ples. Garner and Garfinkel (5) reportedelevated rates of amenorrhea, menstrualirregularity, and pathological eating atti-tudes among low-weight ballet dancers.Gadpaille et al. (6) noted increased eatingpathology and major affective disorder inamenorrheic runners compared withnonamenorrheic runners. However, in

Psychosomatic Medicine 51:81-86 (1989) 81

0033-3l74/89/5lOl-0081$02.00/0Copyright by the American Psycho


2/6

R. E. KREIPEetal.

both of these studies, menstrual functionwas confounded with low weight; amen-orrheic subjects weighed significantly lessthan their menstruating counterparts.

Weight is clearly important in regulat-ing menstrual status, yet recent reports(e.g., Refs. 7-10) indicate that eating dis-orders are associated with menstrual dys-

function in ways that cannot be ac-counted for on the basis of weight orweight change alone. Past research hasfocused exclusively on clinical or high-risk samples, and menstrual dysfunctionhas been confounded by low weight. Thisreport presents two preliminary studieswhich examine the extent to which men-strual abnormalities are associated withsubclinical eating pathology in normal-weight, nonclinical populations.

STUDY I

MethodSubjects for Study I were white females, aged 16-

31, who weighed no more than 110% of ideal bodyweight (IBW). Eighty participants comprised threegroups: 48 with diagnosed eating disorders (includ-ing 17 with restrictive anorexia nervosa, 14 withbulimic anorexia nervosa, and 17 with normal-weight bulimia nervosa), 15 with subclinical eatingdisorders (SEDs), and 17 normal controls (C).

The SED and C groups were recruited from intro-ductory psychology classes and from signs posted atthe University of Rochester Medical Center adver-tising a study entitled "Thoughts and Feelings aboutBody Image." The Eating Attitudes Test (EAT-26)(11) was used to screen for undiagnosed eating pa-thology. The 17 subjects who scored below the rec-ommended cut-off score (19 points) were consideredCs. The 16 subjects scoring above this level wereinterviewed to determine whether they had evermet DSM-III criteria for either anorexia nervosa orbulimia nervosa.1 One subject was diagnosed with

restrictive anorexia nervosa. The remaining 15 sub-jects had never fulfilled the criteria for either eatingdisorder and were considered SEDs. All subclinicalsubjects reported frequent dieting and preoccupa-tion with weight and shape. In addition, two patternsof pathological eating were discernible. The firstpattern was characterized by excessive starvation,exercise, vomiting, or laxative abuse, without bing-ing or significant weight loss. The second patterninvolved occasional binging and purging.

All three groups were indistinguishable on age (x= 21.2, SD = 3.7), education (x = 13.3, SD = 2.0),marital status (84.5% single), and age of menarche(x = 13.1, SD = 1.4). In addition, the SED and Cgroups were matched for current (x = 91.2,SD = 3.7)and lowest (x = 83.6, SD = 6.8) percentage of IBW(12). The groups did differ, however, in terms oftheir highest percentage of IBW (x = 102.1% and96.33%, SD = 4.6 and 8.8, for subclinicals and con-trols, respectively; t = 2.36, p < 0.03) and t he differ-ence between their highest and lowest percentagesof IBW (x = 20.6% and 10.6%. SD = 8.1 and 5.0,respectively; ( = 4.12, p < 0.001).

After signing a consent form, subjects completeda battery of self-report measures including the EAT-

26 and provided basic clinical data, including weighthistory and menstrual functioning.2 Amenorrheawas defined as the cessation of menses for at leastfour consecutive months. Menstrual disruption dueto physical conditions, such as diagnosed polycysticovary syndrome or pregnancy, was excluded fromconsideration.

Results

As shown in the upper portion of Table1, the three groups differed dramatically

on menstrual history (x2

= 94.5, p


3/6

MENSTRUAL ABNORMALITIES AND EATING DISORDERS

TABLE 1. Menstrual Abnormalities in Diagnosed Eating Disorders, Subclinical Eating Disorders, andNormal Controls

Study 1Amenorrhea

CurrentPast

OligomenorrheaTotal

Study IIAmenorrhea

CurrentPast

OligomenorrheaTotal

Diagnosedeating disorders

n = 48

37 (77.0)a

5(10.5)3 (6.2)

45 (93.7)

Subclinicaleating disorders

n = 75

0

10 (66.7)4 (26.7)

14 (93.4)n = 7

0

2 (28.6)5 (71.4)7 (100 0)

Normalcontrols

n = 17

0

02(11.7)2(11.7)n = 20

0

1 (5.0)2(10.0)3(15.0)

1 Numbers in parentheses are percentages.

additional 26.7% of the SEDs had experi-enced oligomenorrhea, as compared to11.8% of the Cs. One SED subject reporteda pregnancy resulting in a stillbirth andone C subject had two uncomplicated,full-term deliveries; no other subjects hadbeen pregnant. Overall, abnormal men-strual function characterized93.4% of theSEDs but only 11.8% of the Cs.

To clarify the relationship betweenweight change and menstrual abnormali-ties, we tabulated the frequency of men-strual disruption by weight change sepa-rately from the SED andC groups. All 15SED subjects had experienced weightfluctuations of at least 10% IBW, includ-ing the only subject who reported no his-tory of menstrual irregularities. Sevensubjects (46.7%) had weight fluctuationsof greater than 20% IBW, and all had ahistory of secondary amenorrhea. Thus,there appears to be a strong relationshipbetween weight shifts and menstrual dis-ruption among subclinical subjects. Thepattern was much less clear among con-

trol subjects. The two C subjects (11.8%)with a history of menstrual disruption(oligomenorrhea) had not experienced ap-preciable weight fluctuations. Con-versely, subjects with a history of weightshifts did note menstrual disturbances;nine subjects (52.4%) with weight changesof 10-20% IBW and one subject (5.9%)with a fluctuation of more than 20% IBWreported no menstrual disruptions.

Discussion

Clearly, subjects with dysfunctionaleating attitudes reported significant men-strual disruption. However, it is likelythat these subjectswhile indistinguish-able from the controls in terms of theircurrent or lowest percentage of IBWhadindeed experienced weight pathology.Their higher weights and weight fluctua-tions suggest that the subclinical subjectsmay have been below their physiological



4/6

R. E. KREIPEetal.

"normal" weight; even though they wereat "normal" weight based on standardizedtables, these women may have been un-derweight relative to their natural set-points.

Alternatively, the subclinical subjectsmay have gone on more "crash diets,"resulting in menstrual disruption due to

sudden and dramatic weight fluctuations(13).3 While it is noteworthy that altera-tions in weight do not appear to be asso-ciated with menstrual changes amongcontrol subjects, the confounding of priorweight fluctuations and group member-ship attenuates the clarity of these find-ings.

An additional difficulty in interpretingthese data arises from the title of thestudy. It is likely tha t "Thoughts and Feel-ings about Body Image" inadvertently at-tracted a biased sample. We undertookStudy II to examine whether the relation-ship between menstrual disruption andsubclinical eating pathology would persistin the absence of any weight pathology ina less skewed sample of women.

STUDY II

MethodSubjects for Study II were students in a large,

upper-level psychology course who elected to par-

ticipate in a project entitled "Construct validation"in exchange for extra credit. Ninety-nine femaleparticipants completed a large battery of psycholog-ical instruments including the Eating Disorders In-ventory (EDI) (13) and a m enstrua l h istory form.4 As

in Study I, amenorrhea was defined as cessation ofmenstruation for at least four consecutive monthsand did not include menstrual disruption due tophysical conditions.

Our analyses focus on 27 subjects, all of whomweighed between 90 and110% of IBW. We classified7 subjects as SEDs and 20 as Cs based on their EDIprofiles. The EDI consists of eight subscales withempirically derived norms; three of the subscalesassess eating disturbances.SED subjects had to score

above the 99% confidence interval for controls onall three eating disorder subscales. In an attempt toensure that these subjects did not, in fact, sufferfrom diagnosable eating pathology, we required thatthey also score below the 99% confidence intervalfor anorexia nervosa on at least one eating disordersubscale.5 To be classified as Cs, subjects had toscore within or below the 99% confidence intervalfor controls on all three eating disorder subscales.8

The resulting groups did not differ on their current(x = 93.3, SD = 7.6), lowest (x = 87.3, SD = 6.7), orhighest (x = 98.0, SD = 9.1) percentage ofIBW, nordid they differ in the magnitude of their weightfluctuations (x = 10.7%, SD = 5.5). The groups alsodid not differ on age of menarche (x = 12.5, SD =

1.6), and none of the subjects reported a history ofpregnancy.

Results

As shown in the lower portion of Table1, the SED and C subjects did indeedmanifest differences in menstrual func-tion (x2 = 16.09, p < 0.0001). None of thesubjects was currently amenorrheic.However, 28.6% of the SEDs vs. 5.0% ofthe Cs had a history of secondary amen-orrhea. The mean duration of secondary

3 We are indebted to an anonymous reviewer forthis explanation and reference.

4 Information about the additional measures inthe battery can be obtained from Richard Ryan.

5 Because of confidentiality concerns, we wereunable to interview subjects to determine more ac-curately the nature of their eating pathology.

"Seventy-two subjects were excluded from anal-yses because: a) they could not be clearly classifiedas SED or C, b) they would not provide weightinformation, or c) they weighed more than 110% ofIBW.

84 Psychosomatic Medicine 51:81-86 (1989)


5/6

MENSTRUAL ABNORMALITIES AND EATING DISORDERS

amenorrhea was 8.0 months in the SEDsand 4.0 months in the Cs. An additional71.4% of the SEDs vs. 10.0% of the Cshad experienced menstrual irregularity.Thus, 100% of the SEDs and 15.0% of theCs reported abnormal menstrual histories.

Of the 99 original participants, 11(11.1%) reported having stopped men-

struating for three consecutive months ormore. Of these 11,10(90.9%) scored abovethe normal range on at least one of theEDI eating disorder subscales.

Discussion

The subjects in study II, unlike those inStudy I, were not explicitly recruited toparticipate in an investigation of eatingattitudes. Thus, it is unlikely that the

subject population was skewed towardwomen with particular weight-relatedconcerns. In addition, the subclinicalgroup in Study II did not report anyweight pathology relative to the controls.Yet the results of StudyII largely replicatethe findings obtained in Study I. Womenwith disturbed eating attitudes had ex-perienced considerably more menstrualabnormalities than their weight-matchedpeers.

GENERAL DISCUSSION

There are obvious shortcomings to bothof the studies reported here. Both reliedon small subject pools and involved ret-rospective recall of menstrual history.Neither measured lean/fat ratio, exerciselevel, gonadotropic hormone levels, or di-etary intake, and Study II did not includesubject interviews. Neither study exam-ined the temporal relationship between

the eating and menstrual disturbances.However, despite their limitations, theseare among the first investigations inwhich differences in menstrual statuswere not confounded with differences inweight. Moreover, both studies demon-strated that disturbances of the hypothal-amic-pituitary-gonadal axis were in the

histories of women with subclinical eat-ing disorders.There are several ways to account for

this co-occurrence. First, the menstrualirregularities noted here, like the "psy-chogenic" amenorrhea of earlier years(14), may result from generalized emo-tional stress. According to this view, themenstrual dysfunction and the eating dis-turbance would be parallel symptoms ofthe same psychological process. Second,the menstrual irregularity might be theresult of a specific behavioral pattern,such as erratic dietary intake or excessiveexercise. Third, the endocrine mecha-nism underlying the menstrual disturb-ance might also interfere with appetiteregulation. Finally, the hypothalamic-pi-tuitary-gonadal axis might be morecausally linked to the etiology of eatingpathology than has been previously sus-pected. A poorly regulated hypothalamic-pituitary-gonadal axis, reflected in men-strual irregularity, might provide thepathophysiologic substrate necessary forthe genesis and maintenance of eatingpathology. The viability of these specula-tions awaits further research.

SUMMARY

Weight loss has often been implicatedin the etiology of menstrual abnormalitiesin eating disordered women. Recent re-search suggests that weight or weight



6/6

R. E. KREIPEetal.

change may not be sufficient to explainthe complex relationship between patho-logical eating attitudes and menstrualdysfunction, yet, in past investigations,menstrual abnormalities have beencon-founded with low weight. The two studiesreported here examined the prevalenceof amenorrhea and oligomenorrhea in

women with abnormal eating attitudesbut without diagnosable eating pathologyor weight pathology (subclinical eatingdisorders). We compared these womenwith age-, education-, and weight-

matched control subjects (controls).InStudy I, 93.4% of the subclinical subjectsreported a history of menstrual abnor-mality, as compared to only 11.7% of thecontrols. Similarly, in Study II, 100% ofthe subclinical subjects and 15.0% of thecontrols had experienced menstrual dys-function. Despite methodological prob-

lems, these data suggest that this elevatedprevalence of oligomenorrhea and amen-orrhea in subclinical eating disorders can-not be explained solely on the basis ofweight or weight loss.

REFERENCES

1. Morton R: Phthisiologia: or, a Treatise of Consump tions. London, Smith and Walford, 1694, pp 8 -92. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders. Third Edition-

Revised. Washington, DC, APA, 19873. Johnson C, Stuckey M, Lewis L, Schwartz D: Bulimia: A descriptive survey of 316 cases. Int J Eating Dis

2:3-16, 1983

4. Loro A: Binge-eating: A cognitive-behavioral treatment approach. In Hawkins W, Fremouw W, ClementP (eds), The Binge-Purge Syndrome: Diagnosis, Treatment, and Research. New York, 1984, pp 183-210

5. Garner D, Garfinkel P: Sociocultural factors in the development of anorexia nervosa. Psychol Med10:647-656

6. Gadpaille W, Sanborn C, Wagner W: Athletic amenorrhea, major affective disorders, and eatingdisorders. Am J Psychiatry 144:939-942,1987

7. Frisch R, McArthur J: Menstrual cycles: Fatness as a determinant in minimum weight for heightnecessary for their maintenance or onset. Science 185:949, 1974

8. Garfinkel P, Garner D: Anorexia Nervosa: A Multidim ension al Perspe ctive. New York, Brun ner/M azel,1982

9. Abrah am S, Beumont P, Fraser 1, Llewellyn-Jones D: Body weight, exercise and men strual status amongballet dancers in training. Br J Obstet Gynecol 89:507-510,1982

10. Falk J, Halmi K: Amenorrhea in anorexia nervosa: Examination of the critical body weight hypothesis.Biol Psychiatry 17:799-806, 1982

11 . Garner D, Garfinkel P: The Eating Attitud es Test. An index of the sympto ms of anorex ia nervosa.

Psychol Med 9:273-279, 197912. Metropolitan Life Insurance Society of Actuaries: Body build and blood pressure study. New York,

Metropolitan Life Insurance Society of Actuaries, 195913. Fichter M, Pirke K: Hypo thalamic p ituita ry function in starving hu man subjects. In Pirke K, Ploog D

(eds), The Psychobiology of Anorexia Nervosa. New York, Springer-Verlag, 1984, pp 124-13514. Garner D, Olmsted M, Polivy J: Development and validation of a multidimensional eating disorder

inven tory for anorexia nervo sa. Int J Eating Dis 2:11-20, 198315. Reifenstein E: Psychogenic or "hypothalamic" amenorrhea. Med Clin North Am 30:234, 1946

86 Psychosomatic Medicine 51:81-86 (1989)

Documents

menstrual abnormalities and subclinical eating disorders