Growth Hormone Neurosecretory Disfunction in Mayor Depressive

Rosana Fiasche, Hugo L. Fideleff, Julio Moizeszowicz, Patricia Frieder, Stella Maris Pagano, Matilde Hollland
Endocrinology Unit T. Alvarez Hospital, Buenos Aires, Argentina
Psychopharmacology Foundation (FundoPsi), Buenos Aires Argentina

Psychoneuroendocrinology 1995, 20: 727-733



Neurotransmitter impairments in MDI can also affect hormonal neuroregulation. Therefore, we decided to study the integrated concentration of growth hormone (IC-GH) and its 24-h secretory profile in this pathology. Ten women with major depressive illness (MDI) (three premenopausal and seven postmenopausal) were evaluated. Samples were obtained every 30 min using a constant withdrawal pump. Growth hormone (GH) pulses were analyzed by Cluster System. Twenty-four IC-GH was evaluated as area under the curve (AUC) and the following results were found:
Depressed (D) = 429.15 + 367.9 vs. controls (C) = 1281.07 + 379.77 (p < .008); nocturnal IC-GH: D = 220 + 274.0 vs. C = 739.52 + 378.15 (p < .02). No statistically significant differences were found between D and C in diurnal IC-GH or in the number of nocturnal or diurnal pulses. Adrenal (cortisol at 0800h, 2300h and post-suppression 1 mg of dexamethasone) and thyroid (T3, T4 0800h and 1700h TSH) evaluations did not show statistically significant differences between D and C women. In conclusion, patients with MDI present a decrease in total GH secretion at the expense of the nocturnal period, probably due to changes in the neurotransmitters that would be involved in depression.

Keywords-Depression; Endogenous depression; Growth hormone; GH neurosecretory dysfunction; GH secretory profiles.


It is known that the changes in neurotransmission occurring in major depressive illness can also affect the neuroregulation of the various hormonal axes. The most consistent neuroendocrine abnormalities present in endogenous depression have been hypercortisolism, lack of cortisol suppression by dexamethasone, and blunted TSH response to TRH. The results obtained from studies of other axes, such as the prolactin and dynamic tests have been performed to study the hypothalamic-pituitary-somatotropic axis, and the GH response to insulin-induced hypoglycemia, l-dopa, clonidine and other hypothalamic neuropeptides such as TRH, GHRH and LHRH have been described. Even though basal GH levels are normal (Brambilla et al., 1988, some studies have described different alterations in the 24-h GH secretory profile. Mendlewicz et al. (1985) reported diurnal hypersecretion with no changes in the amount of GH secreted at night, whereas Rubin et al. (1990) did not find significant differences in the 24-h secretory profiles. Other authors studied only the nocturnal period (Jarret et al., 1990; Schildkraut et al., 1975; Steigert et al., 1989) agreeing that GH secretion appeared to be lower during this period. Because of these contradictory results, we decided to study diurnal and nocturnal GH secretion in woman with major depressive illness in order to gain better knowledge of the somatotropic impairments that may occur in this illness.


Ten women with major depressive illness (three premenopausal and seven postmenopausal), normal weight and no systemic illness were evaluated. None of the patients had any history of alcoholism, drug addiction or endocrinopathies. Four volunteer women (one premenopausal and three postmenopausal) were studied as controls. They did not have a personal or family history of psychiatric illness.
The psychiatric diagnosis was made in accordance with the criteria established by the Diagnostic and Statistically Manual III Edition Revised (DSMIII R) for major depressive illness (1987). The severity of depression was assessed using the Hamilton Scale (Hamilton, 1987) the same day hormonal study was performed. Patients 1, 2, 5, 7, 8, 9 and 10 were diagnosed as unipolar and nos 3, 4 and 6 as bipolar. Five of them had never received antidepressant medication and the remaining five were studied after a 30-day drug-free period. These patients received different associations of a tricyclic antidepressive (desipramine, imipramine or clomipramine), neuroleptics (levomepromazine or clozapine) and benzodiazepines (clobazam, flunitrazepam or clonazepam). Patients 4, 5 and 6 received eight series of electroconvulsive therapy (ECT). In the case of patient no. 6 it was associated to carbamazepine.
The mean age for patients was 49.8 years (range: 40-59 years) and for controls 49.5 years (range: 45-54 years). The patients reported no major alterations in their alimentary behavior and their body mass index (BMI) was 24.72 (range: 20.5-28.2). Control subjects had normal weight and BMI of 24.8 (range: 20.8-27).
For GH evaluation, patients entered the laboratory at 0800h the day test was performed. A sterile non-thrombogenic catheter was inserted into a forearm superficial vein and connected to a constant withdrawal pump to determine integrated concentrations (CORMED ML-6) (Kowarsky et al., 1971). The first sample was discarded to avoid possible effects related to the venipuncture stress. Samples were collected every 30 min and frozen at –20°C until assayed. Patients were ambulatory during daytime and their food intake was not restricted.
The thyroid and adrenal axes were also studied. Morning T3 and T4 (0800h) and TSH (0800h and 1700h) were measured to evaluated the thyroid axis. In patient no.5 a TRH test was performed because she presented elevated 0800h TSH levels. We administered 200µg of TRH (Elea lab., Buenos Aires, Argentina) intravenously obtaining blood samples cortisol with samples obtained at 0800h and 2300h. Dexamethasone (1.0 mg) was administered at 2300h, measuring plasma cortisol the next day at 1600h. GH was measured by a double antibody radioinmunoassay (RIA) with reagents from Diagnostic Products Corporation (DPC, California, USA). The interassay quality control showed the following coefficients of variation (CV): CV % GH: 10.2% (mean =2.6 ng/ml) and 6.3% (mean = 18 ng/ml) with a sensitivity of: 0.76 ng/ml.
T3 and T4 were measured by RIA with reagents from DPC and TSH was measured by IRMA with two monoclonal antibodies and a magnetic solid-phase separation (reagents from SERONO Diagnostic S.A., Coinsins, Switzerland).
The interassay quality control showed the following 3 coefficients of variation:
CV % T3: 6.1%(mean = 115 ng/dl) and 4.5 % (mean = 207 ng/dl).
CV % T4: 4.1 % (mean = 3.95 µg/dl); 4.2 % (mean = 9.8 µg/dl and 5.5 % mean = 16 µg/dl.
CV % TSH: 5.8 % (mean = 1.02 µIU/ml); 4.4 % (mean = 2.8 µIU/ml) and 4.6 % (mean 15.5 µIU/ml).
Cortisol was measured with DPC reagents and the interassay coefficients of variation were as follows: CV % cortisol: 5.2 % (mean = 16.5 µg/dl and 9 % (mean =1 µg/dl).
Pulses were evaluated using the Cluster Analysis program described by Veldhius & Johnson (1986). The total number of pulses, diurnal (1000-2200h) and nocturnal (2200-1000h) area under the curve (AUC) were evaluated in all patients. The statistical calculations were made using the Mann-Whitney and Wilcoxon tests.


Depressed patients secreted less GH over 24 h than controls. The statistical analysis of the results obtained showed that the 24-h IC-GH, evaluated as AUC, was 429.15 + 367.9 in depressives vs. 1281.07 + 379.77 in controls (p < .008). This difference persisted for the nocturnal period, being 220 + 274.01 in depressed patients vs. 739.52 + 378.15 in controls (p < .02).
All the sleep periods occurred during this time.
Conversely, no significant differences were found between both groups in the amount of GH secreted during the diurnal period (Fig. 1). There were no significant differences between depressives and controls in the number of diurnal, nocturnal and 24-h pulses. No significant differences were found in diurnal vs. nocturnal IC-GH. The decrease in the 24-h AUC found in patients vs. normal controls is shown individually in Fig. 2, following a sequential order based on the amount of GH secreted, and in the secretory profiles of each subject. We found that the severity of depressive symptoms evaluated by the Hamilton Rating Scale did not correlate with the IC-GH in 24 h ( r_ = .3590) nor with the nocturnal IC-GH (r_ = .2529). Adrenal evaluation was normal in nine of the 10 patients studied, but one of them did not achieve good post-dexamethasone suppression at 1600h. No significant differences were found between depressives and controls for cortisol levels at 0830h, 2300h or post-dexamethasone. The TSH morning measurement (0830h) was elevated in one of the 10 patients studied, with no changes in the evening concentration (1700h). This patient showed a normal TSH response to the TRH test. No abnormalities were observed in any of the parameters evaluated in controls. No significant differences were found between depressed patients and controls in T3, T4 morning TSH and evening TSH levels. Evening TSH in depressed patients was significantly decreased with respect to morning levels (p < 0.2).

Figure 1: Growth Hormone (GH) relationship between depressive patients (n=10) and controls (n=4)


GH daily production (measured as AUC) shows that depressed patients present a statistically significant decreased in hormonal production over 24 h. This decreased in total GH occurred mainly at the expense of nocturnal GH. Diurnal secretion also showed a trend towards decreased but did not attain statistical significance when compared with the control group. These findings are in partial agreement with those of other authors (Galard et al., 1988; Jarret et al., 1990; Schildkraut et al., 1975). Jarret et al. (1990) reported a decrease in GH during the nocturnal period without evaluating the 24-h secretory profile. However, Mendlewicz et al. (1985) found a diurnal hypersecretion which did not decreased during the nocturnal period. These discrepancies could be attributed both to the different methods used for the research and to the heterogeneity of the patients evaluated. Mendlewicz divided depressed patients into unipolar and bipolar, finding changes in unipolar patients more marked than in bipolar ones. Other authors make no distinction between both groups (Rocatagliata et al., 1982; Zohar et al., 1985), which, together with the diversity of criteria used in research methods, makes it impossible to obtain comparable results. The bipolar patients included in our study were not found to show secretory profile that differentiated them from the rest of the group. What is worth highlighting is the case of patient no. 5, whose secretory profile is similar to those of the control group. All possibilities of diagnostic error are excluded in this patient, since she presented the most severe depression picture, melancholia and history of suicide attempts.
The marked difference between this secretory profile and the rest of the depressed patients suggest that there might be other factors ----hitherto unidentified---- that could be related to various inherent neurochemical abnormalities. On the basis of this hypothesis, different subpopulation could be detected by studying a significantly larger number of cases. Besides, due to individual characteristics of the patient, the wash-out period for the previously received drug (carbamazepine) may have been insufficient or the electroconvulsive therapy may have been produced persistent changes.
Another variable to be considered in depressed patients is the frequency of pulses or secretory peaks. Mendlewicz found a larger number of peaks associated with diurnal hypersecretion, while we did not find any differences in the number of peaks as compared with the control group. It should be noted that peak detection is directly related to the sampling method, since the sampling frequency can bring about differences in peak detection. On this basis, implementing what Albertsson-Wikland & Rosberg (1988) had described and according to our own experience in short stature (Fideleff et al., 1992), we considered a 30 min frequency to be the most appropriate to identify the detectable GH performed every 20 min, Holl et al. (1991) observed a 64 % decrease in the pulses detected by the Cluster System. Hence, our system might not have been sensitive enough to detect small variations in secretory profiles.
Other factors, such as hormonal status, nutritional intake and sleep patterns influence GH secretion.
As to hormonal status, some investigators (Mendlewicz et al.,1985; Rocatagliata et al., 1982) performed their studies in male patients in order to prevent estrogens from influencing the GH secretory profile. Others (Rubin et al., 1990) included patients of both sexes but analysed both groups separately. In spite of the known stimulatory effects of estrogens on GH (Faria et al., 1992) we did not find that premenopausal women secreted more GH than postmenopausal women. Ho et al. (1987) report that elderly women secrete less GH than young women and that difference is unrelated to estrogen effects. On the other hand, according to Iranmanesh et al. (1991), the age and the body mass index (BMI) account for 60 % of the variability observed in the GH production rate. For this reason, our study included only female patients with a small age range, since it was limited to women between 40 and 59 years.
Regarding the GH-sleep relationship, it is known that GH is secreted with the EEG slow waves. Even thought we did not evaluated this phenomenon, we observed that all controls and eight of the 10 depressed patients presented a secretory period in the first half if the night. The remaining two depressed patients (case nos 9 10) showed a dissociation of the sleep/GH release relationship, which is not striking, since it is usually observed from 50 years of age (Carlson et al., 1972) which affects their quality of life, leading to social isolation, emotional instability and disturbances in their sexual life, all these being symptoms also observed in depression. Symptoms improved in these patients after treatment with recombinant GH (Almqvist et al., 1986; Bengtsson et al., 1993; McGauley, 1989).
With regard to thyroid evaluation, only one patient showed increased TSH values. However, she had no clinical manifestations of hypothyroidism, normal T3 and T4 values, negative antibodies, normal TSH response to TRH, and a decrease in evening TSH. The depressed patients studied by our group also presented this decrease in evening values. Which would confirm the normality of the thyroid axis.
We conclude that under the conditions of this research, patients with major depressive illness show a decrease in total GH, especially at the expense of the nocturnal period. This decrease in GH could have contributed to the depressive symptomatology of the patients studied. Such impairments in the secretory pattern could be due to changes in the activity of one or more neurotransmitter system which would be involved in the pathophysiology of this disease.

Figure 2: Integrate concentration of Growth Hormone (IC-GH) in depressive patients (n=10) and controls (n=4)