Changes in growth hormone, insulin, insulinlike growth factors (IGFs), and IGF-binding protein-1 in chronic fatigue syndrome

Changes in Growth Hormone, Insulin, Insulinlike Growth Factors (IGFs), and IGF-Binding Protein-1 in Chronic Fatigue Syndrome Teresa J. Allain, Jenny A. Beam, Patsy Coskeran, Jennifer Jones, Anna Checkley, Joan Butler, Simon Wessely, and John P. Miell Chronic fatigue syndrome (CFS) is characterized by severe physical and mental fatigue of central origin. Similar clinical features may occur in disorders of the hypothalamopituitary axis. The aim of the study was to determine whether patients with CFS have abnormalities of the growth hormone/insulinlike growth factor (GH-IGF) axis basally or following hypothalamic stimulation with insulin-induced hypoglycemia. We compared levels of GH, IGF-I, IGF-1I, IGF-binding protein-1 (IGFBP-1), insulin, and C-peptide in nondepressed CFS patients and normal controls. We found attenuated basal levels of lGF-I (214 +- 17 vs. 263.4 +- 13.4 ~g/L, p = .036) and lGF-lI (420 +_ 19.8 vs. 536 +- 24.3 Ixg/L, p = .02) in CFS patients and a reduced GH response to hypoglycemia (peak GH; 41.9 +- 11.5 vs. 106.0 +- 25.6 mUlL, p = .017). Insulin levels were higher (7.6 +__1.0 vs. 4.3 +_ 0.8 mU/L, p = .02) and IGFBP-1 levels were lower (19.7 +- 4.6 vs. 43.2 -L-_2.7 mg/L, p = .004) in CFS patients compared with controls. This study provides preliminary data of abnormalities of the GH-IGF axis in CFS. It is not apparent whether these changes are components of a primary pathological process or are acquired secondary to behavioral aspects of CFS such as reduced physical activity.

© 1997 Society of Biological Psychiatry Key Words: Chronic fatigue syndrome, hypoglycemia, stress, growth hormone, insulinlike growth factor BIOL PSYCHIATRY1997;41:567--573

Introduction Chronic fatigue syndrome (CFS) is a debilitating disorder characterized by persistent fatigue and fatigability of more From the Department of Medicine, Kings College School of Medicine and Dentistry, London, United Kingdom (TJA, PC, AC, JPM); Department of Clinical Biochemistry, Kings College School of Medicine and Dentistry', London, United Kingdom (JJ); Department of Psychological Medicine, King's College School of Medicine and Dentistry, London, United Kingdom (JAB, SW); and Institute of Child Health, London, United Kingdom (JB). Address reprint requests to Dr. Simon Wessely, Department of Psychological Medicine, Kings College School of Medicine and Dentistry, Bessemer Road, London SE5 9RS, United Kingdom. Received January 23, 1995; revised February 1, 1996.

© 1997 Society of Biological Psychiatry

than 6 months duration in the absence of alternative clinical explanations (Schluederberg et al 1992). Associated features include subjective cognitive disturbance, sleep disturbance, myalgia, and depression. No single cause has yet been identified, and there is also uncertainty about the pathophysiological mechanism underlying the fatigability. To understand the pathophysiology of CFS is, therefore, a priority, both to facilitate diagnosis and to suggest effective treatment strategies. Although fatigability and muscle pain are prominent features of CFS, there is no evidence for a specific muscle disorder in these patients (Edwards et al 1993). Further0006-3223/97/$17.00 PII S0006-3223(96)00074-I

568

BIOLPSYCHIATRY

T.J. Allain et al

1997;41:567-573

more the pattern of fatigue is of central, rather than peripheral, origin (Wessely and Powell 1989). In consequence a number of recent studies have addressed the issue of whether CFS, and the closely related primary fibromyalgia syndrome (FMS) (Goldenberg et al 1990), are due to abnormalities of the hypothalamopituitary axis. These studies suggest that there may be both pituitary and adrenal cortical impairment in CFS. (Ferracioli et al 1990; Demitrack et al 1991; Bakheit et al 1992; Griep et al 1993; Beam et al 1995). Growth hormone (GH), either acting directly or through its major mediator, insulinlike growth factor-I (IGF-I), is an important anabolic agent and is a prerequisite for normal muscle homeostasis. GH-deficient adults experience a number of symptoms in common with CFS patients including fatigue and depression (Lamberts et al 1992). Bennett and colleagues (Bennett et al 1992, 1994) have shown significantly reduced levels of IGF-I in patients suffering from the FMS and attenuated GH responses to stimulation with clonidine and dopamine in a small group of FMS patients with low IGF-I levels. Insulin-induced hypoglycemia is a standardized model of stress, which causes hypothalamic stimulation and leads to increased levels of growth hormone, prolactin, and cortisol in man (Fish et al 1986). The majority of anabolic actions of GH are mediated through systemic and local production of IGF-I, and the biological activity of IGF-I is in turn modulated by levels of IGF-binding proteins (IGFBPs), particularly IGFBP-1 (reviewed in Langford and Miell 1993). In view of our observations of hypothalamopituitary dysfunction in CFS (Beam et al 1995), and those of Bennett and colleagues of abnormalities of GH/IGF in FMS (Bennett et al 1992, 1994), it was decided to establish whether abnormalities of GH responsiveness to hypothalamic stimuli are present in CFS. It was hypothesized that CFS patients may mount defective stress responses reflected by reduced GH/IGF levels either basally or in response to hypothalamic stimulation by hypoglycemic stress in comparison with normal controls. Psychiatric disorders, most commonly depression, are

present in up to 75% of CFS patients (Wessely and Powell 1989). Depression is associated with well-defined abnormalities of the hypothalamopituitary axis (Gold et al 1988). Therefore, in order not to confound or mask any specific changes attributable to CFS, we only investigated nondepressed CFS patients in this study.

Methods

Subjects Patients with CFS were recruited from the specialist CFS clinic at King's College Hospital run by one of the authors (SW). All patients fulfilled recent consensus criteria for the diagnosis of CFS (Sharpe et al 1991). The Oxford criteria probably identify a similar group of patients to the 1994 Centers for Disease Control (CDC) criteria (Fukuda et al 1994), which became available after the conclusion of this study. Chart review confirmed that all of the sample would have fulfilled these criteria, although not all would have fulfilled the now obsolete 1988 CDC criteria (Holmes et al 1988). All had been characterized as typical CFS as part of a recent multicenter diagnostic exercise (Wilson A, Lloyd A, Hickie I, personal communication). The patients completed the General Health Questionnaire, a cutoff of 3 - 4 being used for probable psychological disorder (Goldberg 1972). They also completed a questionnaire specifically designed for the assessment of fatigue (Chalder et al 1993). A cutoff of 3 - 4 is used to indicate clinically significant fatigue; the maximum score is 11. Patients were interviewed on two occasions 1 month apart by experienced psychiatrists (SW and JB), using the relevant sections of the Schedule for Affective Disorders (Spitzer and Endicott 1978) and the Hamilton Depression Rating Scale (Hamilton 1967). Individual scores for the General Health Questionnaire, Chalder fatigue score, and the Hamilton Depression Rating Scale are given in Table 1. No patients were depressed according to operational criteria for major depression defined in DSM-1II-R (American Psychiatric Association 1987) or in the opinion of the

Table 1. Clinical Characteristics of Patients with Chronic Fatigue Syndrome Patient LB PB GW MD HR SM JM JM

Sex F M M F M M M F

Length of Past antidepressant Currentpsychiatric Hamilton Age (years) illness(years) Viralonset treatment diagnosis score GHQ 44 32 39 25 27 41 49 32

GHQ, General Health Questionnaire.

4 1 2 l 8 1 15 4

No Yes Yes Yes No Yes Yes Yes

No No Yes No No No No No

No No Yes No No No No No

6 4 13 8 6 3 5 3

3 7 9 7 11 7 3 2

Chalder fatigue score 10 10 8 11 11 11 11 l0

GH/IGF Changes in CFS

BIOLPSYCHIATRY

569

1997;41:567-573

two consultants; however, 3 patients (GW, MD, and HR) had a past history of depression. These 3 patients were no longer depressed and in other neuroendocrine tests their response patterns were in keeping with the diagnosis of CFS in contrast to depression (Cleare et al 1995), so their inclusion in the current study was felt to be appropriate. Normal control subjects were recruited from students and members of staff at King's College Hospital and the Institute of Psychiatry. Eight patients and 9 controls underwent insulin-induced hypoglycemia. Prerequisites for entry to the study for both patients and controls were: age 18-60 years, a normal electrocardiogram, normal electrolytes, renal function, liver function, thyroid function, and 09:00-hours cortisol, no history of epilepsy, cardiovascular disease, or any other intercurrent illness, and no medications other than vitamin supplements. Use of the oral contraceptive pill was also excluded. All women were tested during the follicular phase of their menstrual cycle (days 3-5). One patient had previously been prescribed antidepressant drugs for symptomatic purposes, even though he did not fulfill the criteria for depression. The patient had discontinued this treatment more than 12 weeks prior to the study. All subjects gave informed consent. The study was approved by both the King's College Hospital and the Bethlem Royal and Maudsley Hospital ethical committees.

Insulin Hypoglycemia Test Subjects attended the outpatient endocrine investigation unit at 09:00 hours having fasted from midnight. On arrival at the unit an intravenous cannula was sited in the forearm. Subjects remained supine throughout the tests. Following a half hour-rest period, at 09:30 hours (time 0), soluble insulin (Velosulin, Leo Laboratories, Aylesbury, U.K.) 0.15 U/kg IV was given as a single bolus injection. All subjects became hypoglycemic with a blood glucose concentration < 2 mmol/L within 30 min. Blood glucose concentration was measured at times 0, 20, 30, 45, 60, 90, and 120 min. All other measurements were made at times 0, 30, 60, 90, and 120 min. All patients were offered oral glucose (1 g) after the 30-min sample (once they had become hypoglycemic) to reverse the symptoms of hypoglycemia. Equal numbers accepted in the patient and control groups (patients 6/8, 75%; controls 7/9, 78%). All samples were allowed to clot at 4°C, separated by centrifugation, and serum stored at -20°C prior to analysis.

Assays Serum IGF-I was measured, after acid-ethanol extraction of its binding proteins, by radioimmunoassay (RIA) using

a polyclonal rabbit antiserum (R557A) raised against purified human IGF-I. The level of detection of this assay was 5 p~g/L; the interassay coefficients of variation (CVs) were 9.0, 4.5, and 6.2% at analyte levels of 654, 231, and 78.4 ~g/L, respectively, with an intraassay coefficient of variation of 4% at 231 p~g/L. IGF-II was measured by radioimmunoassay after extraction of binding proteins and displacement of IGF-II from binding proteins by addition of excess "cold" IGF-I (Blum et al 1988). The intra- and interassay CVs at an analyte level of 520 Ixg/L were 3.9 and 10.2% respectively, and the assay has 0.05% cross-reactivity with IGF-I and a sensitivity of 0.018 ng/tube. Serum levels of IGFBP-1 were measured by a specific RIA. Purified antigen was obtained from Dr. S. Drop (Rotterdam, Holland). Tracer was prepared by iodination of antigen using the chloramine-T method followed by separation on a short Sephadex G75 column. Antiserum was used at a final dilution of 1:10000, which bound approximately 60% of iodinated tracer. Bound and free antigen were separated using a solid phase second antibody, donkey antirabbit coated cellulose (Sac-Cel, Wellcome, U.K.). There was no cross-reactivity with glycosylated or nonglycosylated IGFBP-3, IGF-I, or IGF-II. Minimum detection limit of the assay was 6 ~g/L. The interassay CV at 55 i~g/L was 6.2% and the intraassay at 35 p~g/L 4%. Immunoreactive insulin was determined by a doubleantibody radioimmunoassay (Diagnostics Products Corporation plc, Oxford, U.K.); the within-assay CVs were 10.0 and 4.9% at 16 and 95 mU/L, and between-assay CVs 12.0 and 5.0% at 5 and 15 mU/L. C-peptide was measured by RIA using reagents purchased from Diagnostics Products Corporation. This assay has a sensitivity of 0.06 ng/mL. The intraassay CV at 4.28 ng/mL was 2.1% with an interassay CV of 3.6% at 3.92 ng/mL (1 ng/mL is equivalent to 331 pmol/L). Serum GH was measured by immunoradiometric assay (IRMA) (NETRIA, London) using the I.S. 80/505. The limit of detection of the assay was 1.0 mU/L; the withinassay coefficient of variation was < 5% at > 2 mU/L. The between-assay CVs were 11.3, 5.6, and 7.7% at 1.5, 28, and 59 mU/L, respectively.

Statistical Analysis Results are given as means _+ SEM. Comparisons of hormone estimations at each time point between groups was made by analysis of variance (ANOVA) for repeated measures, and where p < .05 the computation was completed using Fisher's least significant difference test. Growth hormone data were found to be nonparametric, and therefore the analysis was carried out on logarithmi-

570

BIOLPSYCHIATRY

T.J. Allain et al

1997;41:567-573

300 ¸

7" 6"

IGF-I (Fg/I) 250"

5" Glucose (retool/I)

4-

200'

3 2

150

.3o

~

~

~

~

~

1 .~o

time (rn|n)

Figure 1. Changes in serum insulinlike growth factor-I (IGF-I) in chronic fatigue syndrome patients (©, n = 8) and controls (0, n = 9) following soluble insulin 0.15 U/kg IV. Points = means, bars = SEMs.

cally transformed data. For ease of interpretation non-logtransformed data are presented in the text and figures. Time course changes within each group were assessed by paired Student's t test.

Results The clinical characteristics of the patients are shown in Table 1. The patients comprised 5 men and 3 women, mean age 36.1 years (range 2 5 - 4 9 years), and the controls comprised 3 men and 6 women, mean age 26.8 years (range 23-36 years; 3 subjects aged 23, 2 aged 24, 1 each aged 25, 31, 32, and 36 years). The controls were significantly younger than the patient group (p < .01). There was no significant difference in body mass index between the two groups (patients 25.3 _ 1.2 vs. controls 23.7 +. 1.2). Thyroid function tests were equivalent in the two groups (thyroxine: patients 13.4 ___ 0.7 vs. controls 12.6 +, 0.9 pmol/L; thyroid-stimulating hormone patients 2.0 _ 0.3 vs. controls 1.5 _ 0.2 mU/L). Basal IGF-I levels were significantly lower in the patient group (214 +__ 17 Ixg/L) than the control group (263.4 _+ 13.4 i~g/L, p = .036). There were no significant alterations in IGF-I levels in either group during the period of the insulin stress test (Figure 1). Basal IGF-II levels were also lower in the patient group (420 + 19.8 ixg/L) than the control subjects (536 +. 24.3 Ixg/L, p = .02)--no further measurement of immunoreactive IGF-II levels was made during the period of the test. Initial glucose levels were similar in the two groups (control: 3.7 +_ 0.1 mmol/L; CFS: 3.9 - 0.1 mmol/L; p = .51), and the glucose nadirs attained at 30 rain in both groups were not different (control: 1.3 _ 0.2 mmol/L; CFS: 1.3 _ 0.1 mmol/L; p = .93; Figure 2). Basal insulin levels were higher in the CFS group (7.6 +- 1.0 mU/L) than the control group (4.3 -s- 0.8 mU/L, p =

;

~

~,

~

,~

tlme (mln)

Figure 2. Changes in blood glucose in chronic fatigue syndrome patients (©, n = 8) and controls (0, n = 9) following soluble insulin 0.15 U/kg IV. Points = means, bars = SEMs. .02) with a concomitant suppression of basal IGFBP-1 levels (19.7 +- 4.6 vs. 43.2 +- 2.7 I~g/L, p = .004) (Figures 3 and 4). C-peptide estimations reflected the differences in basal insulin (CFS: 596 _ 79; control: 352 +- 43 pmol/L; p --- .014). Insulin levels in the two groups were similar at 30 min (CFS: 98.5 --- 18.5; control: 110.7 _ 13.2 mU/L), but C-peptide levels were significantly higher in the CFS group (118 _ 14 pmol/L) than the control group (69 _ 8 pmol/L, p = .007), suggesting less effective suppression of endogenous insulin in the former group. Thereafter, both glucose and insulin levels were higher in the CFS patients at 60, 90, and 120 min compared with controls (glucose: 60 min 3.7 _ 0.4 vs. 2.5 +- 0.2 mmol/L, p = .02, 90 min 5.6 _ 0.8 vs. 3.5 _ 0.5 mmol/L, p = .04, 120 min 5.4 _ 0.7 vs. 4.5 +- 0.5 mmol/L, p = .3; insulin: 60 min 57.4 _ 14.1 vs. 20.5 _ 4.5 mU/L, p = .02, 90 min 57.8 _+ 13.7 vs. 26.9 +. 3.8 mU/L, p = .03, 120 min 73.7 _ 23.5 vs. 26.5 +- 4.1 mU/L, p = .05). Growth hormone responses to induced hypoglycemia are shown in Figure 5. Although there was no difference in basal GH estimations, the GH peak (attained at 60 min) was significantly higher in the normal controls (106.0 _ 25.6 mU/L) than in the 140 120 100 Insulln (mUll)

80'

60 40 ,,T.

X

20 0 tln~l (rain)

Figure 3. Changes in serum insulin in chronic fatigue syndrome patients (O, n = 8) and controls (0, n = 9) following soluble insulin 0.15 U/kg IV. Points = means, bars = SEMs.

GH/IGF Changes in CFS

B]OLPSYCHIATRY

571

1997;41:567-573

50'

40' IGFBP-1

(~gn)

30.

20"

10

-.,o

o

~,

6o

~

,;o

time (rain)

Figure 4. Changes in serum insulinlike growth factor-binding protein-1 (IGFBP-1) in chronic fatigue syndrome patients ((3, n = 8) and controls (O, n = 9) following soluble insulin 0.15 U/kg IV. Points = means, bars = SEMs.

CFS patients (41.9 ___ 11.5 mU/L, p = .017). The mean GH over the duration of the test was also significantly lower in patients (22.0 ___ 6.5 vs. 53.8 ___ 11.3 mU/L, p = .032).

Discussion This study has demonstrated interesting differences in the G H - I G F axis between nondepressed patients with CFS and controls. Patients with CFS had lower levels of GH in response to stress in that, during insulin-induced hypoglycemia, their peak GH response and their mean GH output were significantly lower. GH levels decline following puberty, although the spontaneous GH secretion rate does not differ significantly over the age range 2 0 - 4 0 years (Finkelstein et al 1972), and age does not affect the GH response to insulin-induced hypoglycemia (Brunswick et al 1988). Although the mean age of our patients was higher than our

150

-

Growth lO0Hormone (mUlL)

time

(min)

Figure 5. Serum growth hormone levels following soluble insulin 0.15 U/kg IV in chronic fatigue syndrome patients ((3) and controls (O). Points = means, bars = SEMs.

controls, there was considerable overlap in the age ranges of the two groups. It seems unlikely, therefore, that the reduction in GH response to hypoglycemia in the CFS patients is simply accounted for by the difference in ages between our patients and controls. Attenuated GH responses have also been demonstrated in depressed patients (Sacher et al 1971), although this is not a consistent observation (Brunswick et al 1988); however, in these studies the glucose nadir following exogenous insulin administration was higher in the depressed subjects, which may account for the observed differences. In our study the glucose nadir was equivalent in patients and controls, so this mechanism does not account for the attenuated GH response we observed. An alternative explanation is that there is central dysregulation of the normal hypothalamic stress response manifested by impaired GH secretion. Results of the cortisol and prolactin responses to insulininduced hypoglycemia in an overlapping group of nondepressed CFS patients has already been published (Beam et al 1995). It was found that the prolactin response to hypoglycemia was also attenuated in these patients, consistent with an impaired hypothalamic stress response. Consistent with the trend toward lower baseline levels of GH, CFS patients had lower baseline IGF-I levels. These levels remained lower than controls throughout the insulin stress test. This finding is in agreement with that of Bennett and colleagues (1992, 1994), who described lower levels of somatomedin-C (IGF-I) taken at random in 285 patients with FMS compared with controls. IGF-I levels also vary with age, showing a steady decline after puberty (Blum 1993); however, given the considerable overlap in ages between our patient and control groups, age differences are again unlikely to account for the differences we observed. Furthermore, in Bennett and colleagues' studies (1992, 1994) the patient and control groups were wellmatched for age, and differences in somatomedin-C were still observed. Baseline IGF-II levels were also significantly lower in the CFS patients. IGF-II levels do not decline significantly in adulthood, so this difference cannot be attributed to the difference in age between the two groups (Blum 1993). IGF-II is a less potent somatomedin that shares common carrier proteins with IGF-I. Although little is known about its physiological regulation and in vivo actions, it would appear to be an important growth factor for neural tissue, since it is present in many human brain areas, whereas IGF-I has not been detected in adult human brain (Haselbacher et al 1985). Recently low levels of IGF-II that rise on treatment have been demonstrated in patients with hypothyroidism (Miell et al 1993). Since fatigue is a prominent feature of both hypothyroidism and CFS, it may be that central IGF-II is involved in neural pathways concerned with the experience of fatigue.

572

BIOLPSYCHIATRY

T.J. Allain et al

1997;41:567-573

IGFBP-1 is a ubiquitous IGF-binding protein whose levels are normally inversely related to those of insulin. Patients with CFS had higher baseline insulin levels and correspondingly lower IGFBP-1 levels. The basal glucose levels were the same in the two groups, suggesting the possibility of insulin resistance in CFS relative to the controls. This was further borne out during the insulin hypoglycemia test, where both insulin and blood glucose levels were higher in CFS patients. C-peptide levels were also higher in the CFS patients, confirming that the observed elevation in insulin was due to increased production of endogenous insulin, as opposed to reduced clearance of the exogenously administered insulin. The mechanism of possible insulin resistance in CFS is not apparent. The patients in our study had none of the known causes of secondary insulin resistance (Moiler and Flier 1991). In particular, in comparison with controls there was no difference in body mass index, no evidence of hypercortisolemia, and, as previously stated, levels of GH were in fact lower. It is possible that CFS patients had higher levels of catecholamines, as we did not make any measurements of catecholamines or their metabolites; however, a previous study of monoamine metabolism in CFS

found lower levels of plasma noradrenaline metabolites in CFS patients compared with controls (Demitrack et al 1992), and in our own study we found no elevation of other components of the stress response, making this an unlikely explanation. In conclusion, we have demonstrated significantly reduced levels of IGF-I and IGF-II, a reduced GH response to hypoglycemia, elevated glucose levels despite higher insulin levels, and lower IGFBP-1 levels in nondepressed patients with CFS compared with controls. This study provides further evidence for hypothalamic dysfunction in chronic fatigue syndrome. It is not apparent whether these changes are components of a primary pathological process or are acquired secondary to some of the behavioral aspects of CFS, such as reduced physical activity. It is clearly important to extend this work to ascertain whether it is a primary or secondary phenomenon since, if primary abnormalities of the G H - I G F axis are confirmed, it may suggest treatment strategies for this complex disorder.

J.P. Miell is supported by the Wellcome Trust. This project was funded by the Linbury Trust.

References American Psychiatric Association (1987): Diagnostic and Statistical Manual of Mental Disorders, 3rd ed rev. Washington, DC: American Psychiatric Press. Bakheit AMO, Behan PO, Dinan TG, Gray CE, O'Keane VO (1992): Possible upregulation of hypothalamic 5-hydroxytryptamine receptors in patients with postviral fatigue syndrome. Br Med J 304:1010-1012. Beam JA, Allain T, Coskeran P, et al (1995): Neuroendocrine responses to D-fenfluramine and insulin-induced hypoglycaemia in chronic fatigue syndrome. Biol Psychiatry 37:245252. Bennett RM, Clark SR, Campbell SM, Burckhardt CS (1992): Low levels of somatomedin C in patients with the fibromyalgia syndrome. Arthritis Rheum 35:1113-1116. Bennett RM, Cook DM, Clark SR, Campbell SM, Burckhardt CS (1994): Low somatomedin C in fibromyalgia: An analysis of clinical specificity and pituitary/hepatic responses. 57th Meeting of the American College of Rheumatology, San Antonio, Texas. Blum WF (1993): Insulin-like growth factors and their binding proteins. In Ranke MB (ed), Functional Endocrinologic Diagnostics in Children and Adolescents. Mannheim: J&J Verlag, pp 101-117. Blum WF, Ranke MB, Beirich JR (1988): A specific radioimmunoassay for insulin-like growth factor. II: The interference of IGF binding proteins can be blocked by excess IGF-I. Acta Endocrinol 118:374-380. Brunswick DJ, Frazer A, Koslow SH, et al (1988): Insulin-

induced hypoglycaemic response and release of growth hormone in depressed patients and healthy controls. Psychol Med 18:79-91. Chalder T, Berelowitz G, Pawlikowska T, et al (1993): Development of a fatigue scale. J Psychosom Res 37:147-153. Cleare AJ, Beam J, Allain T, et al (1995): Contrasting neuroendocrine responses in depression and chronic fatigue syndrome. J Affect Disord 35:283-289. Demitrack AM, Dale JK, Straus SE, et al (1991): Evidence for impaired activation of the hypothalamic-pituitary-adrenal axis in patients with chronic fatigue syndrome. J Clin Endocrinol Metab 73:1224-1234. Demitrack MA, Gold PW, Dale JK, Krahn DD, Kling MA, Straus SA (1992): Plasma and cerebrospinal fluid monoamine metabolism in patients with chronic fatigue syndrome: Preliminary findings. Biol Psychiatry 32:1065-1077. Edwards RHT, Gibson H, Clague JE, Helliwell T (1993): Muscle histopathology and physiology in chronic fatigue syndrome. In Bock GR, Whelan J (eds), Chronic Fatigue Syndrome (Ciba Foundation symposium 173). Chichester, U.K.: Wiley, pp 102-131. Fish H, Chemow B, O'Brien J (1986): Endocrine and neurophysiologic responses of the pituitary to insulin-induced hypoglycaemia: A review. Metabolism 35:763-780. Ferracioli G, Cavalieri F, Salaffi F, et al (1990): Neuroendocrinologic findings in primary fibromyalgia (soft tissue chronic pain syndrome) and in other chronic rheumatic conditions (rheumatoid arthritis, low back pain). J Rheumatol 17:869873.

GH/IGF Changes in CFS

Finkelstein J, Roffwarg H, Boyar R, Kream J, Hellman L (1972): Age-related change in the twenty-four-hour spontaneous secretion of growth hormone. J Clin Endocrinol Metab 35:665-671. Fukuda K, Straus S, Hickie I, Sharpe M, Dobbins J, Komaroff AI (1994): The chronic fatigue syndrome: a comprehensive approach to its definition and study. Ann Intern Med 121: 953-959. Gold PW, Goodwin FK, Chrousos GP (1988): Clinical and biochemical manifestations of depression. Relation to the neurobiology of stress. N Engl J Med 319:413-420. Goldberg D (1972): Detecting Psychiatric Illness by Questionnaire (Maudsley Monograph 22). Oxford: Oxford University Press. Goldenberg DL, Simms RW, Geiger A, Komaroff AL (1990): High frequency of fibromyalgia seen in patients with chronic fatigue seen in primary care practice. Arthritis Rheum 33: 381-389. Griep EN, Boersma JW, de Kloet ER (1993): Altered reactivity of the hypothalamic-pituitary-adrenal axis in the primary fibomyalgia syndrome. J Rheumatol 20:469-474. Hamilton N (1967): Development of a rating scale for primary depressive illness. Br J Soc Exp Psychol 6:278-296. Haselbacher GK, Schwab ME, Pasi A, Humbel RE (1985): Insulin-like growth factor-II (IGF-II) in human brain: Regional distribution of IGF-II and of higher molecular mass forms. Proc Natl Acad Sci USA 82:2153-2159. Holmes GP, Kaplan JE, Gantz NM (1988): Chronic fatigue syndrome: A working case definition. Ann Intern Med 108: 387-389.

BIOL PSYCHIATRY 1997;41:567-573

573

Lamberts SWJ, Valk NK, Binnerts A (1992): The use of growth hormone in adults: A changing scene. Clin Endocrinol (Oxf) 37:111-115. Langford KS, Miell JP (1993): The insulin-like growth factor-I/ binding protein axis: Physiology, pathophysiology and therapeutic manipulation. Eur J Clin Invest 23:503-516. Miell JP, Taylor AM, Zini M, Maheshwari HG, Ross RJM, Valcavi R (1993): Effects of hypothyroidism and hyperthyroidism on insulin-like growth factors (IGFs), growth hormone- and IGF:binding proteins. J Clin Endocrinol Metab 76:950-955. Moiler DE, Flier JS (1991): Insulin resistance-mechanisms, syndromes and implications. N Engl J Med 325:938-948. Sacher EJ, Finkelstein J, Hellman L (1971): Growth hormone responses in depressive illness: I. Response to insulin tolerance test. Arch Gen Psychiatry 25:263-269. Schluederberg A, Straus S, Peterson P, et al (1992): Chronic fatigue syndrome research: Definition and medical outcome assessment. Ann Intern Med 117:325-331. Sharpe M, Archard L, Banatvala J, et al (1991): Chronic fatigue syndrome: Guidelines for research. J Royal Soc Med 84:118121. Spitzer R, Endicott J (1978): Schedule for Affective Disorders and Schizophrenia. New York: New York State Psychiatric Institute, Wessely S, Poweli R (1989): Fatigue syndromes: A comparison of chronic "postviral" fatigue with neuromuscular and affective disorders. J Neurol Neurosurg Psychiatry 52:940-948,