Ghrelin does not mediate the somatotroph and corticotroph responses to the stimulatory effect of glucagon or insulin-induced hypoglycaemia in humans

OBJECTIVE: Acylated ghrelin, a gastric peptide, possesses a potent GH- but also significant ACTH/cortisol-releasing activity mediated by the activation of GH secretagogue receptors (GHS-R) at the hypothalamus-pituitary level. The physiological role of ghrelin in the control of somatotroph and corticotroph function is, however, largely unclear. Glucagon is known to induce a clear increase of GH, ACTH and cortisol levels in humans, at least after intramuscular administration. In fact, glucagon is considered to be a classical alternative to insulin-induced hypoglycaemia (ITT) for the combined evaluation of the function of GH and the hypothalamus-pituitary-adrenal (HPA) axis. We aimed to clarify whether ghrelin mediate the GH and corticotroph responses to intramuscular glucagon or ITT, which has recently been reported able to induce a surprising ghrelin decrease. SUBJECTS: To this aim we enrolled six normal young male subjects [age (mean +/- SD): 29.0 +/- 8.0 years, body mass index (BMI) 21.9 +/- 2.5 kg/m(2)]. DESIGN AND MEASUREMENTS: In all the subjects we studied ghrelin, GH, ACTH, cortisol and glucose levels after glucagon (GLU; 0.017 mg/kg intramuscularly), ITT (0.1 IU/kg insulin intravenously) or saline administration. RESULTS: Saline infusion was not followed by any significant variation in ghrelin, GH and glucose levels while ACTH and cortisol showed the expected spontaneous morning trend toward a decrease. GLU administration increased (P < 0.01) circulating GH, ACTH and cortisol as well as insulin and glucose levels. ITT induced an obvious increase (P < 0.01) of GH, ACTH and cortisol levels. The ITT-induced increases in GH and ACTH, but not cortisol, levels were higher (P < 0.01) than those after GLU. Circulating ghrelin levels were not modified by GLU. On the other hand, ghrelin levels underwent a transient reduction (P < 0.01) after insulin-induced hypoglycaemia. CONCLUSIONS: Ghrelin does not mediate the GH and ACTH responses to glucagon or to the ITT. In fact, ghrelin levels are not modified at all by glucagon and transiently decrease during the ITT. These findings support the assumption that ghrelin does not play a major role in the physiological control of somatotroph and corticotroph function.     

Ghrelin secretion is inhibited by glucose load and insulin-induced hypoglycaemia but unaffected by glucagon and arginine in humans

OBJECTIVE: Circulating ghrelin levels are increased by fasting and decreased by feeding, glucose load, insulin and somatostatin. Whether hyperglycaemia and insulin directly inhibit ghrelin secretion still remains matter of debate. The aim of the present study was therefore to investigate further the regulatory effects of glucose and insulin on ghrelin secretion. DESIGN AND SUBJECTS: We studied the effects of glucose [oral glucose tolerance test (OGTT) 100 g orally], insulin-induced hypoglycaemia [ITT, 0.1 IU/kg insulin intravenously (i.v.)], glucagon (1 mg i.v.), arginine (0.5 mg/kg i.v.) and saline on ghrelin, GH, insulin, glucose and glucagon levels in six normal subjects. MEASUREMENTS: In all the sessions, blood samples were collected every 15 min from 0 up to + 120 min. Ghrelin, GH, insulin, glucagon and glucose levels were assayed at each time point. RESULTS: OGTT increased (P < 0.01) glucose and insulin while decreasing (P < 0.01) GH and ghrelin levels. ITT increased (P < 0.01) GH but decreased (P < 0.01) ghrelin levels. Glucagon increased (P < 0.01) glucose and insulin without modifying GH and ghrelin. Arginine increased (P < 0.01) GH, insulin, glucagon and glucose (P < 0.05) but did not affect ghrelin secretion. CONCLUSIONS: Ghrelin secretion in humans is inhibited by OGTT-induced hyperglycaemia and ITT but not by glucagon and arginine, two substances able to increase insulin and glucose levels. These findings question the assumption that glucose and insulin directly regulate ghrelin secretion. On the other hand, ghrelin secretion is not associated with the GH response to ITT or arginine, indicating that the somatotroph response to these stimuli is unlikely to be mediated by ghrelin.     

Post-hypoglycaemic hyperketonaemia does not contribute to brain metabolism during insulin-induced hypoglycaemia in humans

Summary It is controversial as to whether ketone bodies are utilized by the human brain as a fuel alternative to glucose during hypoglycaemia. To clarify the issue, we studied 10 normal volunteers during an experimental hypoglycaemia closely mimicking the clinical hypoglycaemia of patients with Type 1 (insulin-dependent) diabetes mellitus or insulinoma. Hypoglycaemia was induced by a continuous infusion of insulin (0.40 mU·kg^–1·min^–1 for 8 h, plasma insulin 180 pmol/l) which decreased the plasma glucose concentration to approximately 3.1 mmol/l during the last 3 h of the studies. Subjects were studied on two occasions, i. e. spontaneous, counterregulatory-induced post-hypoglycaemic increase in 3--hydroxybutyrate (from 0.2 to 1.1 mmol/l at 8 h), or prevention of post-hypoglycaemic hyperketonaemia (plasma -hydroxybutyrate 0.1 mmol/l throughout the study) after administration of acipimox, a potent inhibitor of lipolysis. In the latter study, glucose was infused to match the hypoglycaemia observed in the former study. The glycaemic thresholds and overall responses of counterregulatory hormones, symptoms (both autonomic and neuroglycopenic), and deterioration of cognitive function (psychomotor tests) were superimposable in the control study in which ketones increased spontaneously after onset of hypoglycaemic counterregulation, as compared to the study in which ketones were suppressed (p=NS). The fact that responses of counterregulatory hormones, symptoms and deterioration in cognitive function were not exaggerated when posthypoglycaemic hyperketonaemia was prevented, indicate that during hypoglycaemia, the counterregulatory-induced endogenous hyperketonaemia does not provide the human brain with an alternative substrate to glucose. Thus, it is concluded that during hypoglycaemia, endogenous hyperketonaemia does not contribute to brain metabolism and function.     

Hypophysiotropic regulation of adrenocorticotropin secretion in response to insulin-induced hypoglycemia

The hypophysiotropic coding of ACTH secretion resulting from insulin-induced hypoglycemia was investigated in urethane-anesthetized fasted rats. The participation of corticotropin-releasing factor (CRF), arginine vasopressin (AVP), and catecholamines in the ACTH response was first investigated by systemic administration of CRF antiserum, an AVP pressor antagonist, or a ganglionic blocking agent. These treatments were without effect on the hypoglycemic response, which was characterized by a 67% fall in systemic glucose levels within 30 min of insulin administration. ACTH secretion in response to insulin-induced hypoglycemia was differentially affected by these pharmacological treatments. Administration of antiserum to CRF abolished the ACTH response, whereas ganglionic blockade was without significant effect. However, administration of a vasopressinergic pressor antagonist significantly attenuated ACTH secretion after insulin treatment. These observations suggested the participation of both CRF and AVP in mediation of the ACTH secretory response to hypoglycemia. Infusion of glucose to counter the hypoglycemia action of insulin injection prevented the ACTH secretory response. Measurement of immunoreactive (ir) CRF, irAVP, and ir-oxytocin in sequential collections of hypophysial portal plasma revealed a significant elevation of irAVP concentration without concomitant elevation of irCRF or ir-oxytocin levels. We propose that CRF functions in a permissive role, maintaining a relatively constant portal concentration and thereby allowing expression of the weaker ACTH-releasing activity of AVP and other secretagogues. Thus, AVP, not CRF, appears to represent the dynamic mediator of ACTH secretion accompanying insulin-induced hypoglycemia. These observations provide additional support for the hypothesis of multifactor stimulus-specific hypophysiotropic coding of ACTH secretion.     

Obestatin inhibits feeding but does not modulate GH and corticosterone secretion in the rat

Obestatin is a recently discovered 23 amino acids peptide derived from the ghrelin gene. As opposed to ghrelin, obestatin was shown to inhibit food intake in mice. The aims of this research were to study the effects of acute obestatin treatment on feeding behavior in the rat and its effects on GH and corticosterone secretion. Our results demonstrate that in young-adult male rats, obestatin effectively blunts the hunger caused by short-term starvation. Obestatin did not modify GH secretion in 10-day-old rats and did not antagonize the GH-releasing effects of hexarelin. Moreover, obestatin administration had no effects on spontaneous corticosterone secretion. In conclusion, these data demonstrate that in young-adult male rats the newly discovered obestatin can inhibit feeding but does not modify GH and corticosterone release in infant rats.     

Pyridostigmine enhances, but does not normalise, the GH response to GH-releasing hormone in obese subjects

Obese patients are characterised by several neuroendocrine abnormalities, including characteristically a decrease in growth hormone responsiveness to GH-releasing hormone. In normal subjects, the GH response to GHRH is enhanced by the acetylcholinesterase inhibitor, pyridostigmine. We have studied the effect of this drug on GH secretion in gross obesity. Twelve obese patients were studied (mean weight 156% of ideal) and compared with a group of 8 normal volunteers. Each subject was initially studied on two occasions, in random order, with GHRH (1-29) NH2 100 micrograms iv alone and following pretreatment with pyridostigmine 120 mg orally one hour prior to GHRH. In obese patients, the GH response to GHRH was significantly blunted when compared to controls (GH peak: 20 +/- 4 vs 44 +/- 16 micrograms/l; mean +/- SEM). After pyridostigmine, the response to GHRH was enhanced in the obese subjects, but remained significantly reduced compared to non-obese subjects treated with GHRH and pyridostigmine (GH peak: 30 +/- 5 vs 77 +/- 20 micrograms/l, respectively). In 6 subjects, higher doses of GHRH or pyridostigmine did not further increase GH responsiveness in obese patients. Our results suggest that obese patients have a disturbed cholinergic control of GH release, probably resulting from increased somatostatinergic tone. This disturbed regulation may be responsible, at least in part, for the blunted GH responses to provocative stimuli.     

Thromboxane A2 could be involved in bronchial hyperresponsiveness to methacholine in asthmatic subjects but not in bronchitic subjects

To determine whether the involvement of thromboxane A2 in bronchial hyperresponsiveness is specific to asthma, we examined the effects of a selective thromboxane synthetase inhibitor (OKY-046) and a cyclooxygenase inhibitor (indomethacin) on bronchial responsiveness to methacholine in patients with bronchial asthma and chronic bronchitis. The provocative concentration of methacholine producing a 20% fall in forced expiratory volume in one second (PC20-FEV1) was measured before and after oral administration of OKY-046 and indomethacin in eight asthmatic and 10 bronchitic subjects. Baseline FEV1 value was not altered by OKY-046 or indomethacin. The geometric mean value of PC20-FEV1 increased significantly (p less than 0.005) from 1.78 to 4.27 mg/ml after OKY-046 in asthmatic subjects, but not in bronchitic subjects. On the other hand, PC20-FEV1 was not altered by indomethacin in all subjects. It was concluded that the involvement of thromboxane A2 in bronchial hyperresponsiveness may be specific to asthma.     

Alprazolam,a benzodiazepine,does not modify the ACTH and cortisol response to hCRH and AVP,but blunts the cortisol response to ACTH in humans

Alprazolam (AL), a benzodiazepine which activates gamma-amino butyrric acid (GABA)-ergic receptors, exerts a clear inhibitory effect on the activity of the hypothalamo-pituitary-adrenal (HPA) axis and is able to markedly reduce the ACTH response to metyrapone-induced inhibition of glucocorticoid feedback. It has been suggested that its inhibitory action could be regulated by CRH or AVP mediated mechanisms. However, the effect of benzodiazepines on the HPA response to CRH or AVP is contradictory. It has been shown that benzodiazepines have specific receptors on the adrenal gland but it is unclear if they mediate biological effects in humans. In order to further clarify the mechanisms underlying the inhibitory effect of benzodiazepine on HPA axis in humans, we studied the effect of AL (0.02 mg/kg po at -90 min) or placebo in 7 healthy young volunteers (7 female, age: 26-34 yr; wt: 50-58 kg, BMI 20-22 kg/m2) on: 1) the ACTH and cortisol responses to hCRH (2.0 microg/kg iv at 0 min) or AVP (0.17 U/kg im at 0 min); 2) the cortisol, aldosterone and DHEA responses to ACTH 1-24 (0.06 and 250 microg iv at 0 and 60 min, respectively). After placebo, the ACTH and cortisol responses to hCRH (peaks, mean+/-SE: 29.8+/-4.4 pg/ml and 199.3+/-19.6 microg/l) were similar to those recorded after AVP (31.7+/-6.5 pg/ml and 164.8+/-18.0 microg/l); the cortisol response to 0.06 microg ACTH (190.4+/-11.8 microg/l) was similar to that recorded after hCRH and AVP but lower (p<0.01) than that after 250 microg ACTH (260.6+/-17.4 microg/l). AL did not modify the ACTH response to both hCRH (42.5+/-7.1 pg/ml) and AVP (33.3+/-2.7 pg/ml), which even showed a trend toward increase. AL also failed to significantly modify the cortisol response to both hCRH (156.3+/-12.7 microg/l) and AVP (119.4+/-14.5 microg/l), which, on the other hand, showed a trend toward decrease. The cortisol peaks after 0.06 microg ACTH were significantly reduced (p<0.02) by AL pre-treatment (115.0+/-7.7 microg/l) which, in turn, did not modify the cortisol response to the subsequent ACTH bolus (214.7+/-16.6 microg/l). The DHEA and aldosterone responses to all the ACTH doses were not significantly modified by AL. In conclusion, these data show that the HPA response to AVP as well as to hCRH is refractory to the inhibitory effect of AL which, in turn, blunts the cortisol response to low ACTH dose. These findings suggest that both CRH- and AVP-mediated mechanisms could underlie the CNS-mediated inhibitory effect of AL on HPA axis; in the meantime, these results suggest that benzodiazepines could also act on adrenal gland by blunting the sensitivity of the fasciculata zone to ACTH.     

Beta-adrenergic agonism does not impair the GH response to acylated ghrelin in humans

Background  Acylated ghrelin (AG) is a physiological GH secretion amplifier, in part stimulating GHRH neurones and antagonizing somatostatin activity. In humans, AG is one of the most potent pharmacological stimuli of GH secretion and, unlike GHRH, is refractory to the inhibitory effect of glucose, free fatty acids (FFA) and somatostatin. Somatotroph secretion is also profoundly modulated by the adrenergic system. Indeed, beta-adrenergic agonists abolish spontaneous and GHRH-stimulated GH secretion. Based on these data, the aim of the present study was to investigate the effects of beta adrenergic agonism on the GH response to AG.
Subjects and measurements  Six young healthy male volunteers underwent: (a) acute AG intravenous (iv) administration (1·0 µg/kg); (b) salbutamol infusion (SLB; 0·06 µg/kg/min iv); (c) AG + SLB; and (d) saline infusion. In all sessions GH levels were assayed every 15 min from time –30 to +210 min.
Results  SLB induced a significant ( P <  0·05) inhibition of spontaneous GH secretion that persisted up to 75 min after SLB withdrawal. AG induced a marked increase ( P <  0·01) in GH that was not modified by SLB.
Conclusions  The GH-releasing effect of AG is refractory to the inhibitory effect of SLB-induced beta-adrenergic receptor activation. Although further studies are needed to confirm these results during the lifespan and particularly during prolonged exposure to beta agonists, the present data clearly suggest that, among GH stimulatory tests, AG administration might be the most suitable in clinical conditions of chronic treatment with beta-2 agonists, such as in asthmatic disease.     

Desensitization of the platelet aggregation response to adrenaline during insulin-induced hypoglycaemia in man

The aim of the present study was to investigate the influence of insulin-induced hypoglycaemia on platelet sensitivity to adrenaline and non-adrenergic agonists in man. Twenty-five healthy male subjects volunteered for the study. To evaluate the effects on platelets of different insulin-adrenaline interrelationships, two experimental models were used. In the first, hypoglycaemia was induced by a 60-min IV infusion of human insulin at the rate of 64 mU m-2 min-1, whereas in the other the same insulin dose was administered as an IV bolus (3.84 U m-2). Throughout the studies, plasma glucose, insulin, and adrenaline were measured together with platelet sensitivity to adrenaline, ADP, platelet activating factor, collagen, and sodium arachidonate. In both studies, hypoglycaemia induced a reduction of platelet sensitivity to adrenaline (p = 0.006 in infusion and p = 0.045 in injection study). In particular, maximal aggregation to adrenaline fell from 47.9 +/- 9.9 (+/- SE) to 31.1 +/- 11.3% at the hypoglycaemic nadir in the infusion study, and from 64.6 +/- 8.2 to 34.6 +/- 10.3% at the hypoglycaemic nadir in the injection study. In the injection study an increase of platelet sensitivity to ADP (p = 0.05), platelet activating factor (p = 0.018), sodium arachidonate (p = 0.035), and collagen (p = 0.027) was also found, in agreement with observations already published using the infusion protocol. Thus, insulin-induced hypoglycaemia increases platelet sensitivity to non-adrenergic agonists and decreases platelet response to adrenaline.     

Plasma cortisol response to ACTH does not accurately indicate the state of hypothalamic-pituitary-adrenal axis

The hypothalamic-pituitary-adrenal function was studied in 55 patients with various pituitary disorders. In particular, the consistency between the responses of plasma cortisol to exogenous ACTH and to insulin hypoglycemia was investigated in 5 patients in whom cortisol response to insulin was absent; four of these patients showed a cortisol response to ACTH of variable degree. These 4 patients had surgical or functional hypothalamus-pituitary disconnection and showed a preserved cortisol response to lysine vasopressin. These data demonstrate the unreliability of ACTH test in assessing hypothalamic-pituitary-adrenal function in hypopituitary patients.     

The impact of fasting and treatment omission on susceptibility to hypoglycaemia in children and adolescents with GH and cortisol insufficiency

Objective Hypoglycaemia may be a frequent occurrence in young GH deficient patients and so we studied the response to fasting in children and adolescents with GH and/or cortisol deficiency. Methods A total of 20 patients (2–18 years) fasted for 14 h (22·00–12·00 h) on two occasions as part of a randomized cross‐over study. Fourteen had pituitary hormone deficiency (PHD) including GH deficiency (GHD). Of the 14 patients, seven were ACTH sufficient (PHDC+) and seven ACTH deficient (PHDC–). Six had primary adrenal failure (PAF). Subjects administered or omitted their normal dose of evening GH and/or morning hydrocortisone. Glucose, insulin, GH, cortisol, ketones and catecholamines were measured at 04·00 h and regularly from 07·00 to 12·00 h. Insulin sensitivity was assessed by HOMA and hypoglycaemia defined as a blood glucose (BG) ≤ 3·3 mmol/l. Results BG was related to age and body mass index on treatment but no subject became hypoglycaemic on or off therapy prior to 07·00 h. Five children (aged 3, 4, 7, 8 and 11 years) were hypoglycaemic between 07·00 and 12·00 h off treatment. There was a positive relationship between GH AUC and minimum BG in patients with PHD on treatment (r² = 0·45, P = 0·012) with increased insulin sensitivity off treatment. Increased cortisol levels were seen in PHDC+ patients off GH (P < 0·001). A negative relationship was observed between minimum BG and adrenaline (r² = 0·37, P = 0·01), ketone bodies (r² = –0·20, P = 0·05) and NEFA (r² = –0·35, P = 0·02). Noradrenaline levels were reduced in patients with PHDC–. Low BMI (on treatment) and young age (off treatment) were determinants of low BG levels in a multiple regression model. Conclusions Unrecognized overnight hypoglycaemia in children and adolescents on pituitary hormone replacement is uncommon but BG levels quickly become abnormal when treatment and meals are omitted. The insulin antagonistic actions of GH are important in preventing hypoglycaemia. Patients with PHD have altered sympathetic nerve activity.     

Antiserum to rat growth hormone (GH)-releasing factor suppresses but does not abolish antisomatostatin-induced GH release in the rat

Injection of SRIF antiserum (oA-SRIF) increases serum GH and TSH levels in urethane-anesthetized rats. These responses require an intact hypothalamus with endogenous GH-releasing factor (GHRF) or TRH, respectively. We examined whether pretreatment of rats with an antiserum against rat GHRF (oA-rGHRF) would abolish the GH response to oA-SRIF, since anti-TRH serum has been shown to abolish TSH response to oA-SRIF. Prior injection of oA-rGHRF reduced oA-SRIF-induced GH response in a dose-related manner in a dose up to 1 ml antiserum, but failed to produce any further suppression at higher doses. The maximum suppression of the GH response was approximately 50%. oA-rGHRF also suppresses basal GH levels significantly. On the other hand, oA-rGHRF completely abolished the GH-releasing activity of 1 microgram synthetic rGHRF when incubated for 30 min at room temperature before injection. The results suggest two conclusions: 1) 43-residue rGHRF is a physiological regulator of both basal and stimulated GH release; 2) failure of oA-rGHRF to completely abolish the GH response to oA-SRIF suggests the presence of other physiological GHRFs in the rat.     

Relationships among the secretion of ACTH, GH, and cortisol during the insulin-induced hypoglycemia test in the normal and obese child.

The behavior of the secretion of ACTH, GH, and plasma cortisol during the insulin-induced hypoglycemia test on normal and obese children was studied. The secretion of the above-mentioned hormones was determined by calculating the integrals of the curves. The mean values of the integrals of the plasma cortisol and ACTH curves do not show any significant differences between the two groups of children. The mean values of the ratios between the integrals of the plasma cortisol and ACTH curves show a significant difference between the two groups (P smaller than 0.01). Since the mean values of the integrals of the plasma cortisol curves are practically the same in both groups, the difference in the above-mentioned ratios refers to the lower values of the integrals of the ACTH curves found in the obese children. This enables us to make the hypothesis that in the obese child the function of the pituitary-adrenal axis, at least during the insulin test, does not differ from the norm thanks to the adaptation of the ACTH secretion to the greater sensitivity of the adrenal glands to this hormone. In both the groups examined there was no correlation between the secretion of ACTH and plasma cortisol, between ACTH and GH, and between plasma cortisol and GH.     

Acute decrease in circulating T3 levels enhances, but does not normalise, the GH response to GHRP-6 plus GHRH in thyrotoxicosis

In thyrotoxicosis there is an impaired GH response to GHRH, normal GH responsiveness to GHRP-6 and lack of synergistic GH response after simultaneous administration of both peptides. We have previously shown that the GHRH-induced GH release in these patients increases after an acute reduction of circulating T3 values with administration of iopanoic acid, a compound that inhibits peripheral conversion of T4 to T3. We have now studied the effect of a decrease in serum T3 levels on the GH response to GHRP-6 (1 microg/kg) plus GHRH (100 microg) in 9 hyperthyroid patients before and after 15 days of treatment with iopanoic acid (3 g every 3 days) and propylthiouracil (600 mg/day). Nine normal subjects were also studied. In all hyperthyroid patients iopanoic acid induced a rapid decrease and normalisation of serum T3 levels. In these subjects peak GH (microg/l; mean +/- SE) and AUC (microg/l x 120 min) values after GHRP-6 plus GHRH were significantly higher on day 15 compared to pretreatment values (peak, 18.3 +/- 3.0 vs 13.4 +/- 1.9; AUC, 1227.9 +/- 212.9 vs 968.5 +/- 160.4; p<0.05). Despite the significant enhancement of the GH responsiveness to GHRP-6 plus GHRH after treatment with iopanoic acid, this response remained significantly blunted when compared to controls both in terms of peak GH (18.3 +/- 3.0 vs 83.7 +/- 15.2; p<0.05) and AUC values (1227.9 +/- 212.9 vs 4956.5 +/- 889.3; p<0.05). In conclusion, our results show that an acute decrease of circulating T3 levels enhances, but does not normalise, the GH response to GHRP-6 plus GHRH in thyrotoxicosis. This could suggest that circulating T3 does not have a major role in the mechanisms involved in the synergistic effect of these peptides.     

Neurohypophyseal secretion in response to cholecystokinin but not meal-induced gastric distention in humans

Exogenous administration of cholecystokinin octapeptide (CCK) is known to decrease food intake and slow gastric emptying in humans and animals. Recent studies have shown that CCK stimulates neurohypophyseal secretion of oxytocin (OT) in rats and arginine vasopressin (AVP) in monkeys, and that gastric distention also stimulates OT release in rats. We therefore studied AVP and OT secretion in 14 normal subjects in response to meal-induced gastric distention and administration of CCK, both separately and in combination, to assess whether these stimuli similarly activated central neurohypophyseal pathways in humans. Neither plasma AVP nor OT concentrations increased after gastric distention produced by ingestion of a large meal. However, a dose-related increase in plasma AVP, but not OT levels, occurred after CCK administration, the threshold CCK dose being 0.05 micrograms/kg body weight. The AVP secretion in response to CCK administration was significantly correlated with subjective aversive symptoms quantified by use of a numeric scale (r = 0.61, P less than 0.001). In 12 of the 14 subjects plasma AVP levels increased in association with symptoms of epigastric pressure and discomfort before the onset of overt nausea or emesis. The combination of CCK and meal-induced gastric distention did not stimulate increases in plasma AVP levels in excess of those produced by CCK administration alone. The results demonstrate that AVP secretion resulting from emetic center activation often is a graded response that can begin in association with milder degrees of visceral discomfort before symptoms of overt nausea or emesis. In addition, the stimulation of AVP secretion by CCK administration, but not by meal-induced gastric distention in association with physiological satiety, suggests that some component of the anorectic effects of exogenous CCK in man likely results from activation of brainstem emetic centers.