LHRH Responsiveness in Anorexia Nervosa: Intactness Despite Prepubertal Circadian LH Pattern
JACK L. KATZ, MD, ROBERT M. BOYAR, MD, HOWARD ROFFWARG, MD, LEON HELLMAN, MD, AND HERBERT WEINER, MD
Consistent with previous findings, six women with primary anorexia nervosa who underwent 24-hour endocrine studies were all observed to have immature (prepubertal or pubertal) patterns of circadian luteinizing hormone (LH) secretion. This abnormality was present despite negligible weight deficit (but active symptomatology) in three of the women. Regardless of the extent of immaturity of circadian pattern, each woman showed a completely normal release of LH in response to the single intravenous administration of luteinizing hormone releasing hormone (LHRH). These findings provide further evidence for the intactness of pituitary function and for the possibility of a functional hypothalamic disturbance in anorexia nervosa. The discrepant finding in certain other studies of inadequate LH response to LHRH in women with anorexia nervosa is reviewed, and the potential influences of differing variables in these studies and the possibility of a heterogeneous illness are emphasized.
INTRODUCTION
The plasma and urinary levels of two pituitary polypeptides, luteinizing hormone (LH) and follicle stimulating hormone (FSH), are low during the active phase of primary anorexia nervosa (1^1). Two explanations can account for these findings: (1) the synthesis of LH and FSH by the pituitary may be normal, but these hormones are not being released in normal fashion because the hypothalamic mechanism that controls their release by
From the Institute for Steroid Research and the Department of Oncology (Drs. Boyar and Hellman) and the Department of Psychiatry (Drs. Katz, Roffwarg, and Weiner), Montefiore Hospital and Medical Center, 111 East 210th Street, Bronx, New York 10467. This work was supported in part by grant RR-53 from the General Clinical Research Centers Branch of the National Institutes of Health and by research grants CA 07304, ROH 00331, DH 06209, and HL 14734 from the United States Public Health Service. Address reprint requests to: Dr. J. Katz. Received for publication 917176; revision received 1/20/77.
the pituitary is defective or (2) the hypothalamic mechanism may be intact, but the pituitary is defective in its synthesis or release of LH and FSH. Indirect evidence would suggest that the former hypothesis is more likely to be correct. In women with primary anorexia nervosa, the circadian pattern of LH secretion is immature (5,6), and such patterning is presumed to be under hypothalamic control. Moreover, these patients fail to show the normal initial rise in plasma LH in response to clomiphene citrate (7,8), an agent that acts on the median hypothalamus by the competitive blocking of estrogens. The usual delayed rise in LH is also absent during the active phase and continues to be so even with satisfactory weight restoration (9), a finding that suggests a functional failure of the anterior hypothalamus to respond to the direct stimulating effect of clomiphene (10). Luteinizing hormone releasing hormone (LHRH), the hypothalamic decapeptide hormone that induces pituitary release of the gonadotropins, has now been
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977) Published by Elsevier North-Hoi land. Inc.
241
JACK L. KATZetal.
purified, identified, and synthesized (11). It can be administered in humans to determine whether the pituitary is normally responsive to LHRH. Thus, a normal release of LH (and, although less reliably, FSH) in women with anorexia nervosa given LHRH would provide evidence that the second explanation for low plasma and urinary gonadotropins is not likely to be valid. To date, however, such studies of LHRH responsiveness in these women have provided controversial findings. Some investigators have found normal responses (12), or at least normal responsiveness after 3-5 days of "priming" with intravenous (i.v.) LHRH (13), others have described variable responses (14-16), and still others have reported a clearly impaired response during the active phase, particularly when patient weight falls below 25% of ideal weight (17-19). There is no ready explanation for these discrepant findings, but possible considerations include (1) these studies are not all comparable because differing doses of LHRH were administered (25-100 /Ag); (2) in some studies, administration was intravenously, in others subcutaneously; and (3) in some studies, a single injection was employed, in others rapid intravenous drip was the method, and in still others prolonged intravenous "priming" for a number of days was given prior to rapid i.v. administration. Also, it is not certain when during the day LHRH was given in many of the studies. Since we have shown that there can be marked fluctuations in LH levels throughout the 24-hour period (5), the lack of standardization in this regard may also have contributed to the inconsistent findings. The possibility also exists that patients with anorexia nervosa do not constitute a homogeneous group. These patients can 242
vary considerably in their psychological and physiological characteristics; indeed a subgroup appears to exist that has gonadal dysgenesis (Turner's syndrome) (20), a condition that can have concomitant aberrations in gonadotropin output (21). Our investigative group also entertained another explanation for the discrepant findings. In normal prepubertal children, in whom plasma LH levels remain low throughout the 24-hour sleep-wake cycle, the administration of LHRH does not produce a significant increment inLH (22,23). Since we have demonstrated that adult women with active anorexia nervosa are characterized by immature (either prepubertal or pubertal) circadian LH secretory patterns (5), we postulated that the variable responses to LHRH reported in such women might be a function of their respective 24-hour LH secretory patterns, one variable that has never been considered in any of the previous studies. More specifically, we sought to test the hypothesis that adult woman with anorexia nervosa whose LH patterns were prepubertal would have deficient LH responses to LHRH administration, whereas those with maturer, i.e., pubertal, patterns would manifest normal LH releases. SUBJECTS Six white, educated women, 17V2. to 26 years old, diagnosed as having primary anorexia nervosa according to Bruch's criteria (24), were studied as inpatients on our Clinical Research Center. They ranged from 0 to 46% below their ideal weights, and age of onset ranged from 14 to 19 years of age. The patients had taken no medication of any kind for at least 2 months prior to admission or received any during the period of study. None was known to have Turner's syndrome (karyotyping done in three patients), although one (DD) was eventually diagnosed as having Cushing's syndrome and ultimately required bilateral adrenalectomy. Tables 1 and 2 summarize their physical, clinical, and endocrine features.
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA TABLE 1.
Age, Height, Weight, and Endocrine Data for Six Women with Active Primary Anorexia Nervosa Patient AA
BB
CC
DD
17
19
19
61.4
63.6
25
26
47.7 152.4
27.3 160.4
15 56.8 17% 42.7 152.4
EE
FF
14
17
Age at illness onset (years) Weight at illness onset (kg) Age at study (years) Weight at study (kg) Height (cm) % below ideal weight at study Lowest weight (ke) (age)
12
0
46
10
7
36.4 (22)
38.6 (22)
22.7 (22)
39.5 (15Vi)
45.5 (21)
50.0 160.0 2 44.5 (17'/2)
Circadian LH pattern
Late pubertal
Early pubertal
Prepubertal
Prepubertal
Midpubertal
Midpubertal
LH response to LHRH
Normal
Normal
Normal
Normal
Normal
Normal
59.1 23'A 48.2 165.1
METHODS Shortly before 10:00 a.m. on the first morning of study, a heparinized catheter was inserted into an antecubital vein and connected to a longer tube, which extended into an adjoining room. The precise details of obtaining and handling blood specimens via this technique in our laboratory have been elaborated elsewhere (25), but its most cogent feature is that it permits frequent blood sampling without disturbance to (actually awareness by) the subject. Beginning at about 10:00 a.m., a blood sample was obtained every 20 min for the next 24 hours. At night, thepatienthad polygraphic monitoring of sleep onset and stages, which were scored according to standardized criteria (26). After completing the 24-hour study at about 10:00 a.m. the next morning, the patient received 100 ^.g of LHRH administered rapidly through the indwelling intravenous catheter. Blood samples were then obtained every 15 min for the first hour and every 30 min for the second hour following LHRH administration. Plasma LH concentration was determined by the double-antibody radioimmunoassay method of Midgley (27), as modified in our laboratory (28), and is expressed in milli-International Units per milliliter (mlU/ml). The coefficient of variation for duplicate samples is approximately 10%, and all plasma samples from each 24-hour study were assayed simultaneously.
Criteria for Maturity of Circadian LH Secretory Pattern and LH Response to LHRH Based on the previous work by our group in normal children, adolescents, and adults (29), the following
63.6 26 50.0 163.8
56.8 23
are the features that characterize the circadian LH secretory pattern in each developmental stage (FSH patterning being less reliable). (1) Prepubertal Pattern. LH pulses are small throughout sleep and awake stages and there is no significant sleep-awake variation in mean LH concentration, both values usually falling between 2 and 6mIU/ml. (2) Early to Midpubertal Pattern. LH pulses begin to increase moderately in amplitude during sleep (particularly slow wave sleep) so that mean sleep LH concentration increases to about 7-12 mlU/ml, significantly greater than the mean awake concentration of about 3-7 mlU /ml. (3 J Mid- to Late Pubertal Pattern. LH pulses show a marked increase in amplitude now throughout both sleep and awake stages, but mean sleep LH concentration still remains significantly higher than mean awake concentration, typical values being 16-24 mlU/ml vs. 7-12 mIU/ml, respectively. (4J AduJt Pattern. Pulses moderate in amplitude throughout both sleep and awake stages and the difference in mean sleep-awake concentrations is no longer present, the 24-hour mean value now ranging typically between 7 and 24 mlU /ml (the phase of the menstrual cycle influencing where in this broad range the mean will be). It should be noted that the basis for biological staging here is primarily the maturity of breast development (30), not age or presence or absence of menarche (although obviously a rough correlation will exist with these variables as well). Because of the physical (i.e., breast) maturity of our patients, all of whom
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
243
TABLE 2. Clinical Data for Six Women with Active Primary Anorexia Nervosa Patient AA Menstrual status
< znr
o o 3 n
Clinical course
viedi Q n>
CC
EE
FF
Menarche: 12; periods regular; amenorrhea at 22Vb through present
Menarche: IPA; periods regular; amenorrhea at age 20 through present
Menarche: 12; periods regular; persistent amenorrhea at age 15V2 after 6 month use of oral contraceptive
Menarche: 12; periods irregular; amenorrhea at age 20 x 3 years after 1 year use of oral contraceptive; currently irregular
Menarche: 13Vi; periods regular; amenorrhea at age 15% x 4V4 years; currently irregular
Began dieting in context of tensions at home; currently: obsessive dieting alternating with binging and vomiting, distorted body image, hyperactivity, insomnia, rapid weight changes (fluid retention and diuresis)
Began dieting in context of new job and urging by friend; currently: obsessive dieting alternating with binging (no vomiting), distorted body image, insomnia, rapid weight changes (fluid retention and diuresis)
Began dieting as college sophomore when father commented she was "getting fat"; currently: obsessive dieting alternating with binging and vomiting, food stealing and hoarding, distorted body image, hyperactivity, insomnia, cold intolerance
Began dieting in context of first sexual relationship; currently: obsessive dieting alternating with binging (no vomiting), distorted body image, hyperactivity, cold intolerance, rapid weight changes (fluid retention and diuresis)
Began dieting in context of first sexual relationship; could not diet successfully so forced self to vomit regularly; currently: vomiting compulsively after meals, binging, distorted body image, hyperactivity, insomnia, cold intolerance, rapid weight changes (fluid retention and diuresis)
Began dieting to lose weight for ballet lessons; currently: obsessive dieting alternating with binging and vomiting, food stealing and hoarding distorted body image, insomnia, hyperactivity
Depressive, anxious, hysterical, impulsive ("borderline") features
Compulsive features
Schizoid, depressive, paranoid, impulsive ("borderline") features; karyotyping normal
Hysterical features; karyotyping normal; scalp hair loss but hirsute elsewhere; found to have Cushing's syndrome eventually requiring bilateral ad renal ectomy
Depressive, impulsive, dependent ("borderline") features; karyotyping normal
Hysterical, depressive, impulsive ("borderline") features; father and paternal grandfather have manicdepressive illness
o
Comments
DD
Menarche: 15V2; periods regular; amenorrhea at age 20 x years; currently irregular
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA were also past age 17 and had reached menarche a number of years earlier, they should normally have manifested fully adult circadian LH secretory patterns. The findings of Dickerman and co-workers (23) indicate that normal prepubertal girls show a negligible increase in plasma LH level in response to a single intravenous bolus of LHRH, whereas normal pubertal girls will respond with a significant LH increment to a plasma level that will be at least severalfold that of the base-line value. The response becomes progressively greater with each maturer stage of puberty and is observed to reach its peak about 30-45 min after infusion. FSH response to LHRH follows the same pattern as LH but tends to be less reliable and of smaller magnitude.
1000
Figs. la-f.
1200
MOO
16O0
1800
2000 2200 2400 CLOCK TIME
RESULTS Even though three of the six patients were at weights approximating ideal values, by virtue of binge eating counterbalancing their excessive dieting, all six were found to have abnormal circadian patterns of LH secretion. By the criteria for stages of LH pattern maturity, two of the women (CC, DD) had a prepubertal, one (BB) an early pubertal, two a midpubertal (EE, FF), and one (AA) a late pubertal circadian pattern. Nevertheless, all six, including the woman (CC) who had been
0200
0400
O6O0 0800
1000
1200
1400
1600
Twenty-four-hour sleep-wake pattern of plasma LH and FSH levels and gonadotropin responsiveness to 100 /xg of i.v. LHRH in each of six women with active primary anorexia nervosa. Shown here, patient AA.
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
245
JACK L. K A T Z e t a l .
1000 1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 CLOCK TIME
1000 1200
Fig. lb. Patient BB.
response to intravenous LHRH is fully normal in at least some women with anorexia nervosa. Not only does this normal response appear to be independent of weight, it is independent of maturity of circadian LH secretory pattern, thereby refuting our original hypothesis. Thus, an explanation for the low plasma and urinary levels of gonadotropins must be sought in a source other than the pituitary—presumably at the DISCUSSION hypothalamic level—assuming one can make inferences about physiologic funcOur findings are consistent with those tion from responsiveness to a pharof Wiegelman and Solbach (12) that LH macologic dose.
chronically and severely ill for 7 years and was 46% below her ideal weight, showed brisk, ample, fully normal releases of LH in response to i.v. administration of LHRH (see Table 1 and Figs. 1 a-f). No significant difference in responsiveness existed between those with prepubertal and those with pubertal patterns. Typically, FSH responses were similar to those of LH, but of smaller magnitude.
246
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA
1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 1000 1200 1400 CLOCK TIME Fig. lc. Patient CC.
This study still leaves unanswered the question of why some investigators have found deficient responses in their patients with anorexia nervosa. For instance, a very careful recent study by Beaumont and co-workers (31), which also utilized a 100-ju.g intravenous injection of LHRH, showed a definite correlation between percent of ideal weight and LH response in their women with anorexia nervosa: those below 70% had grossly deficient responses, those at 70%-79% had low normal responses, and those above 80% had accentuated responses. Five of our patients were indeed within 12% of their ideal weights, but one was 46%
below her ideal weight, had not gained any weight prior to the study, and had a very satisfactory response. Halmi and Sherman (16) showed that level of circulating estrogen does not appear to be a relevant variable. In addition to the possible explanations presented in the Introduction, we would emphasize that differences in chronicity and in nature of eating behavior (i.e., dietary content and pattern of eating, dieting, binging, vomiting, etc., as contrasted with merely relative or absolute weight) might be important variables. For instance, AN patients usually subsist on high protein, low carbohydrate diets, but those who are given to binging generally
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
247
JACK L. KATZ et al.
1200 1400
1600
1800 2000 2200 2400 0200 0400 0600 0800 CLOCK TIME
1000 1200 1400
Fig. Id. Patient DD.
do so on carbohydrates; Crisp and Stonehill (32) have observed empirically that adequate carbohydrate refeeding in particular is vital for restoration of menses in these women, but nature of carbohydrate intake has never been considered in any studies of LHRH responsiveness. It should be noted that all our patients were subject to binging. Also tenable certainly is the argument that anorexia nervosa is not a single illness and that biological, as
248
well as psychological, differences exist between subgroups of such patients. SUMMARY
The availability now of synthetic LHRH for human administration permits discrimination between pituitary and hypothalamic sources for diminished levels of plasma gonadotropins. Because anorexia nervosa has been consistently
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA
LH,mIU/ml
1000 1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 1000 CLOCK TIME Fig. l e . Patient EE.
characterized by low plasma (and urinary) LH and FSH levels, we have administered LHRH to six women with this illness; the women had varying degrees of weight loss and chronicity. Other investigators had reported variable responses to LHRH and we therefore postulated that intactness of response might correlate with degree of maturity of circadian LH secretory response, a finding characteristic of normal children and adolescents. However, while two women had prepubertal patterns and
1200 1400 1600
one only an early pubertal pattern, all six showed normal responses to LHRH. These data would appear to provide further confirmation of a hypothalamic disturbance in at least some patients with anorexia nervosa. We believe that patients described as unresponsive to LHRH in other studies may reflect the existence of biologically different subgroups with this illness, different courses or diets among patients, or differences in method or amount of LHRH administration.
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
249
JACK L. KATZetal.
\J 1000
1200
1400
1600
1800
2000 2200 2400 0200 0400 0600 0800 CLOCK TIME
1000 1200
Fig. If. Patient FF.
REFERENCES 1. Bliss EL, Migeon CJ: Endocrinology of anorexia nervosa. J Clin Endocrinol Metab 17:766-776, 1957 2. Russell GFM, Loraine JA, Bell ET, Harkness RA: Gonadotropin and oestrogen excretion in patients with anorexia nervosa. J Psychosom Res 9:79-85, 1965 3. Warren MP, VandeWiele RL: Clinical and metabolic features of anorexia nervosa. Am J Obstet Gynecol 117:435-449, 1973 4. Kanis JA, Brown P, Fitzpatrick K, et al.: Anorexia nervosa: a clinical, psychiatric, and laboratory study. Q J Med 53:321-328, 1974 5. Katz JL, Boyar RM, Weiner H, et al.: Toward an elucidation of the psychoendocrinology of anorexia nervosa, in EJ Sachar (ed.), Hormones, Behavior, and Psychopathology. New York, Raven Press, 1974, pp. 263-283 6. Kalucy RS: Paper read at First International Congress of Obesity, London, 1974 7. Marshall JC, Fraser TR: Amenorrhea in anorexia nervosa: assessment and treatment with clomiphine citrate. Br. Med J 4:590-592, 1971
250
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA 8. Beaumont PJV, Carr PJ, Gelder MG: Plasma levels of luteinizing hormone and of immunoreactive oestrogens in anorexia nervosa: response to clomiphene citrate. Psychol Med 3:495-501, 1973 9. Russell GFM, Wakeling A: The endocrine and menstrual response to clomiphene citrate in patients with anorexia nervosa. Paper read at Tenth European Conference on Psychosomatic Research, Edinburgh, 1974 10. Katz JL, Weinter H: A functional, anterior hypothalamic defect in primary anorexia nervosa? Psychosom Med 37:103-105, 1975 11. Schally AV, Arimura A, Kastin AJ: Hypothalamic regulatory hormones. Science 179:341-350, 1973 12. Wiegelmann W, Solbach HG: Effects of LHRH on plasma levels of LH and FSH in anorexia nervosa. Horm Metab Res 4:404, 1972 13. Yoshimoto Y, Moridera K, Imura H: Restoration of normal pituitary gonadotropin reserve by administration of luteinizing-hormone-releasing hormone in patients with hypogonadotropic hypogonadism. N Engl J Med 292:242-245, 1975 14. Mecklenburg RS, Loraiux L, Thompson RH, et al.: Hypothalamic dysfunction in patients with anorexia nervosa. Medicine 53:147-159, 1974 15. Frankel RJ, Jenkens JS: Hypothalamic-pituitary function in anorexia nervosa. Acta Endocrinol 78:209-221, 1975 16. Halmi KA, Sherman BM: Gonadotropin response to LH-RH in anorexia nervosa. Arch Gen Psychiatry 30:875-878, 1975 17. Jequier AM, Jacobs HS, Kurtz AJ, et al.: Gonadotropin secretion in endocrine disorders: response to luteinizing hormone-follicle-stimulating hormone releasing hormone. Proc Soc Endocrinol 244:233, 1975 18. Warren MP, Jewelewicz R, Dyrenfurth I, et al. The significance of weight loss in the evaluation of pituitary response to LH-RH in women with secondary amenorrhea. J Clin Endocrinol Metab 40:601-611, 1975 19. Palmer RL, Crisp AH, Mackinnon PCB, et al.: Pituitary sensitivity to 50 /xg LS /FSH-RH in subjects with anorexia nervosa in acute and recovery stages. Br Med J 1:179-182, 1975 20. Kron L, Katz JL, Gorzynski G, Weiner H: Gonadal dysgenesis and anorexia nervosa: further evidence of a relationship. Arch Gen Psychiatry 34:332-335, 1977 21. Faggiano M, Minozzi M, Lombardi G, etal.: Two cases of chromatin positive variety of ovarian dysgensis (XO /XX mosaicism) associated with hGH deficiency and marginal impairment of other hypothalamic pituitary functions. Clin Genet 8:324-329, 1975 22. Winter JSD: Analysis of clinical studies with LH-RH in children and adolescents. Am J Dis Child 130:590-592, 1976 23. Dickerman Z, Prager-Lewin R, Laron Z: Response of LH and FSH to synthetic LH-RH in children at various pubertal stages. Am J Dis Child 130:634-638, 1976 24. Bruch H: Anorexia nervosa and its differential diagnosis. J Ment Nerv Dis 141:555-566, 1965 25. Hellman L, Nakada F, Curti J, et al.: Cortisol is secreted episodically by normal man. J Clin Endocrinol Metab 30:411-422, 1970 26. Rechtschaffen A, Kales A (eds): A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects (NIH Publication 204). Washington, D.C., Government Printing Office, 1968 27. Midgley AR Jr: Radioimmunoassay: a method for human chorionic gonadotropin and human luteinizing hormone. Endocrinology 79:10-18, 1966 28. Boyar R, Perlow M, Hellman L, et al.: Twenty-four patterns of luteinizing hormone secretion in normal men with sleep stage monitoring. J Clin Endocrinol Metab 35:73-81, 1972 29. Boyar R, Finkelstein J, Roffwarg, H, et al.: Synchronization of augmented luteinizing hormone secretion with sleep during puberty. N Engl J Med 287:582-586, 1972 30. Tanner JM: Growth at Adolescence (ed 2). Oxford, Blackwell Scientific Publication, p. 32 31. Beaumont PJV, George GCW, Pimstone BL, Vinik Al: Body weight and the pituitary response to hypothalamic releasing hormones in patients with anorexia nervosa. J Clin Endocrinol Metab 43:487^196, 1976 32. Crisp AH, Stonehill E: Relation between aspects of nutritional disturbance and menstrual activity in primary anorexia nervosa. Br Med J 3:149-151, 1971
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
251
JACK L. KATZ, MD, ROBERT M. BOYAR, MD, HOWARD ROFFWARG, MD, LEON HELLMAN, MD, AND HERBERT WEINER, MD
Consistent with previous findings, six women with primary anorexia nervosa who underwent 24-hour endocrine studies were all observed to have immature (prepubertal or pubertal) patterns of circadian luteinizing hormone (LH) secretion. This abnormality was present despite negligible weight deficit (but active symptomatology) in three of the women. Regardless of the extent of immaturity of circadian pattern, each woman showed a completely normal release of LH in response to the single intravenous administration of luteinizing hormone releasing hormone (LHRH). These findings provide further evidence for the intactness of pituitary function and for the possibility of a functional hypothalamic disturbance in anorexia nervosa. The discrepant finding in certain other studies of inadequate LH response to LHRH in women with anorexia nervosa is reviewed, and the potential influences of differing variables in these studies and the possibility of a heterogeneous illness are emphasized.
INTRODUCTION
The plasma and urinary levels of two pituitary polypeptides, luteinizing hormone (LH) and follicle stimulating hormone (FSH), are low during the active phase of primary anorexia nervosa (1^1). Two explanations can account for these findings: (1) the synthesis of LH and FSH by the pituitary may be normal, but these hormones are not being released in normal fashion because the hypothalamic mechanism that controls their release by
From the Institute for Steroid Research and the Department of Oncology (Drs. Boyar and Hellman) and the Department of Psychiatry (Drs. Katz, Roffwarg, and Weiner), Montefiore Hospital and Medical Center, 111 East 210th Street, Bronx, New York 10467. This work was supported in part by grant RR-53 from the General Clinical Research Centers Branch of the National Institutes of Health and by research grants CA 07304, ROH 00331, DH 06209, and HL 14734 from the United States Public Health Service. Address reprint requests to: Dr. J. Katz. Received for publication 917176; revision received 1/20/77.
the pituitary is defective or (2) the hypothalamic mechanism may be intact, but the pituitary is defective in its synthesis or release of LH and FSH. Indirect evidence would suggest that the former hypothesis is more likely to be correct. In women with primary anorexia nervosa, the circadian pattern of LH secretion is immature (5,6), and such patterning is presumed to be under hypothalamic control. Moreover, these patients fail to show the normal initial rise in plasma LH in response to clomiphene citrate (7,8), an agent that acts on the median hypothalamus by the competitive blocking of estrogens. The usual delayed rise in LH is also absent during the active phase and continues to be so even with satisfactory weight restoration (9), a finding that suggests a functional failure of the anterior hypothalamus to respond to the direct stimulating effect of clomiphene (10). Luteinizing hormone releasing hormone (LHRH), the hypothalamic decapeptide hormone that induces pituitary release of the gonadotropins, has now been
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977) Published by Elsevier North-Hoi land. Inc.
241
JACK L. KATZetal.
purified, identified, and synthesized (11). It can be administered in humans to determine whether the pituitary is normally responsive to LHRH. Thus, a normal release of LH (and, although less reliably, FSH) in women with anorexia nervosa given LHRH would provide evidence that the second explanation for low plasma and urinary gonadotropins is not likely to be valid. To date, however, such studies of LHRH responsiveness in these women have provided controversial findings. Some investigators have found normal responses (12), or at least normal responsiveness after 3-5 days of "priming" with intravenous (i.v.) LHRH (13), others have described variable responses (14-16), and still others have reported a clearly impaired response during the active phase, particularly when patient weight falls below 25% of ideal weight (17-19). There is no ready explanation for these discrepant findings, but possible considerations include (1) these studies are not all comparable because differing doses of LHRH were administered (25-100 /Ag); (2) in some studies, administration was intravenously, in others subcutaneously; and (3) in some studies, a single injection was employed, in others rapid intravenous drip was the method, and in still others prolonged intravenous "priming" for a number of days was given prior to rapid i.v. administration. Also, it is not certain when during the day LHRH was given in many of the studies. Since we have shown that there can be marked fluctuations in LH levels throughout the 24-hour period (5), the lack of standardization in this regard may also have contributed to the inconsistent findings. The possibility also exists that patients with anorexia nervosa do not constitute a homogeneous group. These patients can 242
vary considerably in their psychological and physiological characteristics; indeed a subgroup appears to exist that has gonadal dysgenesis (Turner's syndrome) (20), a condition that can have concomitant aberrations in gonadotropin output (21). Our investigative group also entertained another explanation for the discrepant findings. In normal prepubertal children, in whom plasma LH levels remain low throughout the 24-hour sleep-wake cycle, the administration of LHRH does not produce a significant increment inLH (22,23). Since we have demonstrated that adult women with active anorexia nervosa are characterized by immature (either prepubertal or pubertal) circadian LH secretory patterns (5), we postulated that the variable responses to LHRH reported in such women might be a function of their respective 24-hour LH secretory patterns, one variable that has never been considered in any of the previous studies. More specifically, we sought to test the hypothesis that adult woman with anorexia nervosa whose LH patterns were prepubertal would have deficient LH responses to LHRH administration, whereas those with maturer, i.e., pubertal, patterns would manifest normal LH releases. SUBJECTS Six white, educated women, 17V2. to 26 years old, diagnosed as having primary anorexia nervosa according to Bruch's criteria (24), were studied as inpatients on our Clinical Research Center. They ranged from 0 to 46% below their ideal weights, and age of onset ranged from 14 to 19 years of age. The patients had taken no medication of any kind for at least 2 months prior to admission or received any during the period of study. None was known to have Turner's syndrome (karyotyping done in three patients), although one (DD) was eventually diagnosed as having Cushing's syndrome and ultimately required bilateral adrenalectomy. Tables 1 and 2 summarize their physical, clinical, and endocrine features.
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA TABLE 1.
Age, Height, Weight, and Endocrine Data for Six Women with Active Primary Anorexia Nervosa Patient AA
BB
CC
DD
17
19
19
61.4
63.6
25
26
47.7 152.4
27.3 160.4
15 56.8 17% 42.7 152.4
EE
FF
14
17
Age at illness onset (years) Weight at illness onset (kg) Age at study (years) Weight at study (kg) Height (cm) % below ideal weight at study Lowest weight (ke) (age)
12
0
46
10
7
36.4 (22)
38.6 (22)
22.7 (22)
39.5 (15Vi)
45.5 (21)
50.0 160.0 2 44.5 (17'/2)
Circadian LH pattern
Late pubertal
Early pubertal
Prepubertal
Prepubertal
Midpubertal
Midpubertal
LH response to LHRH
Normal
Normal
Normal
Normal
Normal
Normal
59.1 23'A 48.2 165.1
METHODS Shortly before 10:00 a.m. on the first morning of study, a heparinized catheter was inserted into an antecubital vein and connected to a longer tube, which extended into an adjoining room. The precise details of obtaining and handling blood specimens via this technique in our laboratory have been elaborated elsewhere (25), but its most cogent feature is that it permits frequent blood sampling without disturbance to (actually awareness by) the subject. Beginning at about 10:00 a.m., a blood sample was obtained every 20 min for the next 24 hours. At night, thepatienthad polygraphic monitoring of sleep onset and stages, which were scored according to standardized criteria (26). After completing the 24-hour study at about 10:00 a.m. the next morning, the patient received 100 ^.g of LHRH administered rapidly through the indwelling intravenous catheter. Blood samples were then obtained every 15 min for the first hour and every 30 min for the second hour following LHRH administration. Plasma LH concentration was determined by the double-antibody radioimmunoassay method of Midgley (27), as modified in our laboratory (28), and is expressed in milli-International Units per milliliter (mlU/ml). The coefficient of variation for duplicate samples is approximately 10%, and all plasma samples from each 24-hour study were assayed simultaneously.
Criteria for Maturity of Circadian LH Secretory Pattern and LH Response to LHRH Based on the previous work by our group in normal children, adolescents, and adults (29), the following
63.6 26 50.0 163.8
56.8 23
are the features that characterize the circadian LH secretory pattern in each developmental stage (FSH patterning being less reliable). (1) Prepubertal Pattern. LH pulses are small throughout sleep and awake stages and there is no significant sleep-awake variation in mean LH concentration, both values usually falling between 2 and 6mIU/ml. (2) Early to Midpubertal Pattern. LH pulses begin to increase moderately in amplitude during sleep (particularly slow wave sleep) so that mean sleep LH concentration increases to about 7-12 mlU/ml, significantly greater than the mean awake concentration of about 3-7 mlU /ml. (3 J Mid- to Late Pubertal Pattern. LH pulses show a marked increase in amplitude now throughout both sleep and awake stages, but mean sleep LH concentration still remains significantly higher than mean awake concentration, typical values being 16-24 mlU/ml vs. 7-12 mIU/ml, respectively. (4J AduJt Pattern. Pulses moderate in amplitude throughout both sleep and awake stages and the difference in mean sleep-awake concentrations is no longer present, the 24-hour mean value now ranging typically between 7 and 24 mlU /ml (the phase of the menstrual cycle influencing where in this broad range the mean will be). It should be noted that the basis for biological staging here is primarily the maturity of breast development (30), not age or presence or absence of menarche (although obviously a rough correlation will exist with these variables as well). Because of the physical (i.e., breast) maturity of our patients, all of whom
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
243
TABLE 2. Clinical Data for Six Women with Active Primary Anorexia Nervosa Patient AA Menstrual status
< znr
o o 3 n
Clinical course
viedi Q n>
CC
EE
FF
Menarche: 12; periods regular; amenorrhea at 22Vb through present
Menarche: IPA; periods regular; amenorrhea at age 20 through present
Menarche: 12; periods regular; persistent amenorrhea at age 15V2 after 6 month use of oral contraceptive
Menarche: 12; periods irregular; amenorrhea at age 20 x 3 years after 1 year use of oral contraceptive; currently irregular
Menarche: 13Vi; periods regular; amenorrhea at age 15% x 4V4 years; currently irregular
Began dieting in context of tensions at home; currently: obsessive dieting alternating with binging and vomiting, distorted body image, hyperactivity, insomnia, rapid weight changes (fluid retention and diuresis)
Began dieting in context of new job and urging by friend; currently: obsessive dieting alternating with binging (no vomiting), distorted body image, insomnia, rapid weight changes (fluid retention and diuresis)
Began dieting as college sophomore when father commented she was "getting fat"; currently: obsessive dieting alternating with binging and vomiting, food stealing and hoarding, distorted body image, hyperactivity, insomnia, cold intolerance
Began dieting in context of first sexual relationship; currently: obsessive dieting alternating with binging (no vomiting), distorted body image, hyperactivity, cold intolerance, rapid weight changes (fluid retention and diuresis)
Began dieting in context of first sexual relationship; could not diet successfully so forced self to vomit regularly; currently: vomiting compulsively after meals, binging, distorted body image, hyperactivity, insomnia, cold intolerance, rapid weight changes (fluid retention and diuresis)
Began dieting to lose weight for ballet lessons; currently: obsessive dieting alternating with binging and vomiting, food stealing and hoarding distorted body image, insomnia, hyperactivity
Depressive, anxious, hysterical, impulsive ("borderline") features
Compulsive features
Schizoid, depressive, paranoid, impulsive ("borderline") features; karyotyping normal
Hysterical features; karyotyping normal; scalp hair loss but hirsute elsewhere; found to have Cushing's syndrome eventually requiring bilateral ad renal ectomy
Depressive, impulsive, dependent ("borderline") features; karyotyping normal
Hysterical, depressive, impulsive ("borderline") features; father and paternal grandfather have manicdepressive illness
o
Comments
DD
Menarche: 15V2; periods regular; amenorrhea at age 20 x years; currently irregular
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA were also past age 17 and had reached menarche a number of years earlier, they should normally have manifested fully adult circadian LH secretory patterns. The findings of Dickerman and co-workers (23) indicate that normal prepubertal girls show a negligible increase in plasma LH level in response to a single intravenous bolus of LHRH, whereas normal pubertal girls will respond with a significant LH increment to a plasma level that will be at least severalfold that of the base-line value. The response becomes progressively greater with each maturer stage of puberty and is observed to reach its peak about 30-45 min after infusion. FSH response to LHRH follows the same pattern as LH but tends to be less reliable and of smaller magnitude.
1000
Figs. la-f.
1200
MOO
16O0
1800
2000 2200 2400 CLOCK TIME
RESULTS Even though three of the six patients were at weights approximating ideal values, by virtue of binge eating counterbalancing their excessive dieting, all six were found to have abnormal circadian patterns of LH secretion. By the criteria for stages of LH pattern maturity, two of the women (CC, DD) had a prepubertal, one (BB) an early pubertal, two a midpubertal (EE, FF), and one (AA) a late pubertal circadian pattern. Nevertheless, all six, including the woman (CC) who had been
0200
0400
O6O0 0800
1000
1200
1400
1600
Twenty-four-hour sleep-wake pattern of plasma LH and FSH levels and gonadotropin responsiveness to 100 /xg of i.v. LHRH in each of six women with active primary anorexia nervosa. Shown here, patient AA.
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
245
JACK L. K A T Z e t a l .
1000 1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 CLOCK TIME
1000 1200
Fig. lb. Patient BB.
response to intravenous LHRH is fully normal in at least some women with anorexia nervosa. Not only does this normal response appear to be independent of weight, it is independent of maturity of circadian LH secretory pattern, thereby refuting our original hypothesis. Thus, an explanation for the low plasma and urinary levels of gonadotropins must be sought in a source other than the pituitary—presumably at the DISCUSSION hypothalamic level—assuming one can make inferences about physiologic funcOur findings are consistent with those tion from responsiveness to a pharof Wiegelman and Solbach (12) that LH macologic dose.
chronically and severely ill for 7 years and was 46% below her ideal weight, showed brisk, ample, fully normal releases of LH in response to i.v. administration of LHRH (see Table 1 and Figs. 1 a-f). No significant difference in responsiveness existed between those with prepubertal and those with pubertal patterns. Typically, FSH responses were similar to those of LH, but of smaller magnitude.
246
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA
1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 1000 1200 1400 CLOCK TIME Fig. lc. Patient CC.
This study still leaves unanswered the question of why some investigators have found deficient responses in their patients with anorexia nervosa. For instance, a very careful recent study by Beaumont and co-workers (31), which also utilized a 100-ju.g intravenous injection of LHRH, showed a definite correlation between percent of ideal weight and LH response in their women with anorexia nervosa: those below 70% had grossly deficient responses, those at 70%-79% had low normal responses, and those above 80% had accentuated responses. Five of our patients were indeed within 12% of their ideal weights, but one was 46%
below her ideal weight, had not gained any weight prior to the study, and had a very satisfactory response. Halmi and Sherman (16) showed that level of circulating estrogen does not appear to be a relevant variable. In addition to the possible explanations presented in the Introduction, we would emphasize that differences in chronicity and in nature of eating behavior (i.e., dietary content and pattern of eating, dieting, binging, vomiting, etc., as contrasted with merely relative or absolute weight) might be important variables. For instance, AN patients usually subsist on high protein, low carbohydrate diets, but those who are given to binging generally
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
247
JACK L. KATZ et al.
1200 1400
1600
1800 2000 2200 2400 0200 0400 0600 0800 CLOCK TIME
1000 1200 1400
Fig. Id. Patient DD.
do so on carbohydrates; Crisp and Stonehill (32) have observed empirically that adequate carbohydrate refeeding in particular is vital for restoration of menses in these women, but nature of carbohydrate intake has never been considered in any studies of LHRH responsiveness. It should be noted that all our patients were subject to binging. Also tenable certainly is the argument that anorexia nervosa is not a single illness and that biological, as
248
well as psychological, differences exist between subgroups of such patients. SUMMARY
The availability now of synthetic LHRH for human administration permits discrimination between pituitary and hypothalamic sources for diminished levels of plasma gonadotropins. Because anorexia nervosa has been consistently
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA
LH,mIU/ml
1000 1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 1000 CLOCK TIME Fig. l e . Patient EE.
characterized by low plasma (and urinary) LH and FSH levels, we have administered LHRH to six women with this illness; the women had varying degrees of weight loss and chronicity. Other investigators had reported variable responses to LHRH and we therefore postulated that intactness of response might correlate with degree of maturity of circadian LH secretory response, a finding characteristic of normal children and adolescents. However, while two women had prepubertal patterns and
1200 1400 1600
one only an early pubertal pattern, all six showed normal responses to LHRH. These data would appear to provide further confirmation of a hypothalamic disturbance in at least some patients with anorexia nervosa. We believe that patients described as unresponsive to LHRH in other studies may reflect the existence of biologically different subgroups with this illness, different courses or diets among patients, or differences in method or amount of LHRH administration.
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
249
JACK L. KATZetal.
\J 1000
1200
1400
1600
1800
2000 2200 2400 0200 0400 0600 0800 CLOCK TIME
1000 1200
Fig. If. Patient FF.
REFERENCES 1. Bliss EL, Migeon CJ: Endocrinology of anorexia nervosa. J Clin Endocrinol Metab 17:766-776, 1957 2. Russell GFM, Loraine JA, Bell ET, Harkness RA: Gonadotropin and oestrogen excretion in patients with anorexia nervosa. J Psychosom Res 9:79-85, 1965 3. Warren MP, VandeWiele RL: Clinical and metabolic features of anorexia nervosa. Am J Obstet Gynecol 117:435-449, 1973 4. Kanis JA, Brown P, Fitzpatrick K, et al.: Anorexia nervosa: a clinical, psychiatric, and laboratory study. Q J Med 53:321-328, 1974 5. Katz JL, Boyar RM, Weiner H, et al.: Toward an elucidation of the psychoendocrinology of anorexia nervosa, in EJ Sachar (ed.), Hormones, Behavior, and Psychopathology. New York, Raven Press, 1974, pp. 263-283 6. Kalucy RS: Paper read at First International Congress of Obesity, London, 1974 7. Marshall JC, Fraser TR: Amenorrhea in anorexia nervosa: assessment and treatment with clomiphine citrate. Br. Med J 4:590-592, 1971
250
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
LHRH RESPONSIVENESS IN ANOREXIA NERVOSA 8. Beaumont PJV, Carr PJ, Gelder MG: Plasma levels of luteinizing hormone and of immunoreactive oestrogens in anorexia nervosa: response to clomiphene citrate. Psychol Med 3:495-501, 1973 9. Russell GFM, Wakeling A: The endocrine and menstrual response to clomiphene citrate in patients with anorexia nervosa. Paper read at Tenth European Conference on Psychosomatic Research, Edinburgh, 1974 10. Katz JL, Weinter H: A functional, anterior hypothalamic defect in primary anorexia nervosa? Psychosom Med 37:103-105, 1975 11. Schally AV, Arimura A, Kastin AJ: Hypothalamic regulatory hormones. Science 179:341-350, 1973 12. Wiegelmann W, Solbach HG: Effects of LHRH on plasma levels of LH and FSH in anorexia nervosa. Horm Metab Res 4:404, 1972 13. Yoshimoto Y, Moridera K, Imura H: Restoration of normal pituitary gonadotropin reserve by administration of luteinizing-hormone-releasing hormone in patients with hypogonadotropic hypogonadism. N Engl J Med 292:242-245, 1975 14. Mecklenburg RS, Loraiux L, Thompson RH, et al.: Hypothalamic dysfunction in patients with anorexia nervosa. Medicine 53:147-159, 1974 15. Frankel RJ, Jenkens JS: Hypothalamic-pituitary function in anorexia nervosa. Acta Endocrinol 78:209-221, 1975 16. Halmi KA, Sherman BM: Gonadotropin response to LH-RH in anorexia nervosa. Arch Gen Psychiatry 30:875-878, 1975 17. Jequier AM, Jacobs HS, Kurtz AJ, et al.: Gonadotropin secretion in endocrine disorders: response to luteinizing hormone-follicle-stimulating hormone releasing hormone. Proc Soc Endocrinol 244:233, 1975 18. Warren MP, Jewelewicz R, Dyrenfurth I, et al. The significance of weight loss in the evaluation of pituitary response to LH-RH in women with secondary amenorrhea. J Clin Endocrinol Metab 40:601-611, 1975 19. Palmer RL, Crisp AH, Mackinnon PCB, et al.: Pituitary sensitivity to 50 /xg LS /FSH-RH in subjects with anorexia nervosa in acute and recovery stages. Br Med J 1:179-182, 1975 20. Kron L, Katz JL, Gorzynski G, Weiner H: Gonadal dysgenesis and anorexia nervosa: further evidence of a relationship. Arch Gen Psychiatry 34:332-335, 1977 21. Faggiano M, Minozzi M, Lombardi G, etal.: Two cases of chromatin positive variety of ovarian dysgensis (XO /XX mosaicism) associated with hGH deficiency and marginal impairment of other hypothalamic pituitary functions. Clin Genet 8:324-329, 1975 22. Winter JSD: Analysis of clinical studies with LH-RH in children and adolescents. Am J Dis Child 130:590-592, 1976 23. Dickerman Z, Prager-Lewin R, Laron Z: Response of LH and FSH to synthetic LH-RH in children at various pubertal stages. Am J Dis Child 130:634-638, 1976 24. Bruch H: Anorexia nervosa and its differential diagnosis. J Ment Nerv Dis 141:555-566, 1965 25. Hellman L, Nakada F, Curti J, et al.: Cortisol is secreted episodically by normal man. J Clin Endocrinol Metab 30:411-422, 1970 26. Rechtschaffen A, Kales A (eds): A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects (NIH Publication 204). Washington, D.C., Government Printing Office, 1968 27. Midgley AR Jr: Radioimmunoassay: a method for human chorionic gonadotropin and human luteinizing hormone. Endocrinology 79:10-18, 1966 28. Boyar R, Perlow M, Hellman L, et al.: Twenty-four patterns of luteinizing hormone secretion in normal men with sleep stage monitoring. J Clin Endocrinol Metab 35:73-81, 1972 29. Boyar R, Finkelstein J, Roffwarg, H, et al.: Synchronization of augmented luteinizing hormone secretion with sleep during puberty. N Engl J Med 287:582-586, 1972 30. Tanner JM: Growth at Adolescence (ed 2). Oxford, Blackwell Scientific Publication, p. 32 31. Beaumont PJV, George GCW, Pimstone BL, Vinik Al: Body weight and the pituitary response to hypothalamic releasing hormones in patients with anorexia nervosa. J Clin Endocrinol Metab 43:487^196, 1976 32. Crisp AH, Stonehill E: Relation between aspects of nutritional disturbance and menstrual activity in primary anorexia nervosa. Br Med J 3:149-151, 1971
Psychosomatic Medicine Vol. 39, No. 4 (July-August 1977)
251
