Growth hormone response to GRF 1-44 in children following cranial irradiation for central nervous system tumors

The growth hormone (GH) responses to (A) GRF 1-44, 1 microgram/kg i.v., (B) L-dopa and either arginine, insulin, or glucagon, and (C) exercise were evaluated in 10 children (3 girls, 7 boys; ages 10 years to 15 years, 8 months), 2-10.75 years following cranial irradiation for medulloblastoma (8 patients), pineoblastoma (1 patient), and a fourth ventricular ependymoma (1 patient). Nine of the 10 children had abnormal growth rates. All children were euthyroid at the time of the study. The mean 0-60-min peak GH response to GRF (10.06 +/- 2.6 ng/ml) in the patients was less than the mean peak GH response (29 +/- 2.3 ng/ml) in the control children (n = 7). In 6 patients (5 with poor growth rates), a decreased GH response was noted to GRF and all other tests. Of the remaining patients, all with poor growth rates, two patients demonstrated an adequate response to GRF and pharmacologic testing; one patient had a normal GH response to GRF with a low GH response to pharmacologic testing; and one patient had a low response to GRF, despite a normal response to both exercise and pharmacologic testing. The decrease in mean peak GH response to GRF in the patient population confirms that radiation to the hypothalamic-pituitary region produces abnormalities in growth hormone release. Furthermore, in these patients, discordant GH responses to GRF and pharmacologic or physiologic tests can be observed. The abnormality in growth hormone release may result from a hypothalamic dysfunction in GRF release and/or damage to GH secretory pituicytes.     

Levels of growth hormone and growth-hormone-releasing factor in cord blood

In order to evaluate the mechanism of high growth hormone (GH) secretion in perinatal life, the levels of GH and growth-hormone-releasing factor (GRF) in cord blood were determined. Plasma-immunoreactive GRF was measured by a double antibody RIA method. The levels (mean +/- SD) of GH, GRF and somatostatin (SRIF) were 23.4 +/- 10.2 ng/ml, 49.5 +/- 11.7 and 41.5 +/- 10.4 pg/ml, respectively; they were remarkably higher than those of healthy adults. In statistical analysis, there were no significant relationships among the levels of GH, GRF or SRIF. We speculate that a high GRF release from the hypothalamus might increase the secretion of GH in the perinatal period.     

Growth hormone secretion following administration of growth hormone releasing hormone in constitutional short stature and idiopathic growth hormone deficiency

A stimulation test using 1 microgram growth-hormone-releasing factor (GRF 1-29 X NH2)/kg bodyweight was performed in children with familial short stature and in children with constitutional delay of growth and development. The GH secretion induced by this means was not different in these groups, but there was a difference in the response between normal children and children with idiopathic growth hormone deficiency (GHD). GH secretion after GRF administration was significantly lower in the GHD group than in the other groups. However, 6 of 24 patients with GHD responded to the test with a normal increase in GH (greater than 10 ng/ml), and 11 with an intermediate response (2-10 ng/ml). Thus, the test does not differentiate individual patients with defective growth hormone secretion from normal short children.     

Growth hormone secretion in patients with constitutional delay of growth and pubertal development

Growth hormone levels were measured every 30 minutes during sleep over 9 hours in 20 prepubertal patients with constitutional delay of growth and puberty (CGD) and in 10 age-matched controls, all of whom had had normal GH responses to an orally administered dose of clonidine. We found no significant difference in the mean 9-hour overnight GH concentration between groups (4.5 +/- 1.8 ng/ml (mean +/- SD) in the CGD group, 4.4 +/- 2.8 ng/ml in the control group). Total GH output (258 +/- 99 U vs 222 +/- 135 U), total number of nocturnal GH pulses (3.6 +/- 0.8 vs 3.3 +/- 1.3), mean peak GH response during nocturnal sampling (13 +/- 1.2 ng/ml vs 13.2 +/- 1.3 ng/ml), and basal somatomedin C concentrations were not different in the children with growth delay and controls. We conclude that prepubertal patients with constitutional delay of growth and puberty secrete GH normally and do not seem to have any abnormality in GH regulation.     

Somatocrinin and children in 1984. Application to the etiological diagnosis of somatotropin deficiencies

A single acute IV injection (1 microgram/kg) of the synthetic replicate of Somatocrinin (GRF) in 40 children with growth hormone (GH) deficiency induces a marked plasma GH increase, although heterogeneous. Clinical tolerance is excellent. Compared to Propranolol + Glucagon (P + G), GRF induces a better GH response. It also discriminates better idiopathic GH deficiency (n = 13), where mean GH peak = 6.5 ng/ml (3.3 after P + G) from GH deficiency secondary to a brain tumor (n = 24) where mean GH peak = 15.5 ng/ml (5.0 after P + G) GRF induces a slight Prolactin (Prl) increase, more obvious when basal Prl is elevated. However there is no correlation between GH and Prl responses to GRF even with basal hyperprolactinemia. GH response to GRF seems to slowly decrease after radiation therapy. GRF is a new potent, well tolerated secretagogue of GH and improves the diagnostic quality of the etiology of GH deficiency.     

Growth hormone testing in short children and their response to growth hormone therapy

Because current concepts of growth hormone (GH) testing and GH treatment have become controversial, we investigated the GH secretory patterns in children with normal and short stature. Twenty-four-hour serum GH levels were evaluated in three groups of children. Group 1 was composed of children with normal height (mean height = 0.02 SD, n = 33); group 2 was composed of short children (less than 5th percentile, n = 63) with normal results on provocative GH testing; and group 3 was composed of short children (less than 5th percentile, n = 7) with subnormal results on provocative GH testing. Mean +/- SD (range) GH levels during 24-hour studies of GH secretion were 1.6 +/- 1.1 (0.5 to 5.6), 1.8 +/- 1.2 (0.6 to 6.3), and 0.9 +/- 0.4 (0.5 to 1.7) ng/ml in groups 1, 2, and 3, respectively. No statistical difference existed in mean GH levels between groups 1 and 2 or between groups 1 and 3. The mean GH concentration from 24-hour studies in group 2 children did not correlate with chronologic age, height standard deviation, growth rates, or insulin-like growth factor 1 levels. The linear growth rate of 26 of 28 children in group 2 who received GH therapy for 6 months improved by 2 cm/yr or more; the mean +/- SD growth rate was 4.0 +/- 1.3 and 8.8 +/- 2.0 cm/yr during control and treatment periods, respectively, for these 28 children. Mean GH levels from testing did not predict response to GH during 6 months of therapy. Children with slower growth rates responded better to GH therapy (p less than 0.05). We conclude that (1) in 24-hour studies, GH levels in normal children overlapped with those of short children, including those with classic GH deficiency, (2) in 24-hour studies, GH levels did not predict responses of linear growth to short-term GH treatment, nor did they correlate with children's heights or growth velocities, and (3) the majority of short children in group 2 treated with GH for 6 months had an increase in linear growth velocity, the mean +/- SD change being 4.8 +/- 2.0 cm/yr.     

Growth hormone response to clonidine in obese children.

Basal and stimulated serum growth hormone (GH) levels after exercise, insulin induced hypoglycemia (IIH) and oral clonidine were evaluated in 20 (16 M, 4 F) normal statured obese (body mass index greater than or equal to 25 kg/M2) children. Basal serum GH levels (mean +/- SEM, 2.0 +/- 0.38 ng/ml) were not different from basal levels in non-obese children. The mean peak levels were 3.16 +/- 1.17 ng/ml, 2.15 +/- 0.36 ng/ml and 3.15 +/- 1.12 ng/ml (+/- SEM) after exercise, IIH and oral clonidine, respectively. The positive responses (peak level of serum GH greater than 7 ng/ml) were seen in 10% with exercise, in 10% with clonidine and in none with IIH test. These observations suggest that GH response to oral clonidine is subnormal in obese children.     

Growth hormone response to growth hormone-releasing hormone (hp GHRH1-44) as an index of growth hormone secretory dysfunction after prophylactic cranial irradiation for acute lymphoblastic leukemia (24 grays)

The growth hormone response to growth hormone releasing hormone hp GHRH1-44 (2 micrograms/kg i.v.) was studied in 19 prepubertal children who had been irradiated with 24 Gy for acute lymphoblastic leukemia (ALL) or lymphosarcoma (LS) at a mean chronological age of 4 10/12 years (limits 10/12 to 9 years). They were evaluated after a mean time interval of 4 8/12 +/- 3/12 years and compared to 14 prepubertal children with constitutional short stature (CSS). The individual responses to GHRH were decreased in all but three of the irradiated children. The mean GH response was 16.7 +/- 2.5 ng/ml as compared to 52.6 +/- 8.5 ng/ml in the control group (p less than 0.001). The GH response to GHRH was not correlated with the GH response to arginine-insulin tolerance test (AITT). A decreased response to GHRH with values between 12.5 and 19.4 ng/ml was observed in four cases with normal growth rates and normal GH responses to AITT. These results suggest that an impaired GH response to GHRH is a frequent finding after cranial irradiation for ALL or LS and may be the only sign of GH secretory dysfunction. It is probably indicative of early hypothalamic impairment of GH secretion.     

Effect of growth hormone-releasing factor on growth hormone release in children with radiation-induced growth hormone deficiency

Five male children who received cranial irradiation for extrahypothalamic intracranial neoplasms or leukemia and subsequently developed severe growth hormone (GH) deficiency were challenged with synthetic growth hormone-releasing factor (GRF-44), in an attempt to distinguish hypothalamic from pituitary dysfunction as a cause of their GH deficiency, and to assess the readily releasable GH reserve in the pituitary. In response to a pulse of GRF-44 (5 micrograms/kg intravenously), mean peak GH levels rose to values higher than those evoked by the pharmacologic agents L-dopa or arginine (6.4 +/- 1.3 ng/mL v 1.5 +/- 0.4 ng/mL, P less than .05). The peak GH value occurred at a mean of 26.0 minutes after administration of GRF-44. These responses were similar to those obtained in children with severe GH deficiency due to other etiologies (peak GH 6.3 +/- 1.7 ng/mL, mean 28.0 minutes). In addition, there was a trend toward an inverse relationship between peak GH response to GRF-44 and the postirradiation interval. Prolactin and somatomedin-C levels did not change significantly after the administration of a single dose of GRF-44. The results of this study support the hypothesis that cranial irradiation in children can lead to hypothalamic GRF deficiency secondary to radiation injury of hypothalamic GRF-secreting neurons. This study also lends support to the potential therapeutic usefulness of GRF-44 or an analog for GH deficiency secondary to cranial irradiation.     

Spontaneous and provoked growth hormone (GH) secretion and insulin-like growth factor I (IGF-I) concentration in patients with beta thalassaemia and delayed growth

Growth retardation in children with thalassaemia major is multifactorial. We studied the growth hormone (GH) response to provocation by clonidine and glucagon, measured the circulating concentrations of insulin, insulin-like growth factor-I (IGF-I), IGF-binding protein-3 (IGFBP3), and ferritin, and evaluated the spontaneous nocturnal (12 h) GH secretion in prepubertal patients with thalassaemia and age-matched children with constitutional short stature (CSS) (height SDS < -2, but normal GH response to provocation). The anatomy of the hypothalamic pituitary area was studied in patients with abnormal GH secretion using MRI scanning. Children with thalassaemia had significantly lower peak GH response to provocation by clonidine and glucagon (8.8 +/- 2.3 micrograms/l and 8.2 +/- 3.1 micrograms/l respectively) than did controls (17.6 +/- 2.7 micrograms/l and 15.7 +/- 3.7 micrograms/l respectively). They had significantly decreased circulating concentrations of IGF-I and IGFBP3 (68.5 +/- 19 ng/ml and 1.22 +/- 0.27 mg/l respectively) compared to controls (153 +/- 42 ng/ml and 2.16 +/- 0.37 mg/l respectively). Seven of the thalassaemic children had a GH peak response of < 7 micrograms/l after provocation. Those with a normal GH response after provocation also had significantly lower IGF-I and IGFBP3 concentrations than controls. Analysis of their spontaneous nocturnal GH secretion revealed lower mean (2.9 +/- 1.77 micrograms/l) and integrated (2.53 +/- 1.6 micrograms/l) concentrations compared to controls (4.9 +/- 0.29 micrograms/l and 5.6 +/- 0.52 micrograms/l respectively). Five of them had mean nocturnal GH concentration < 2 micrograms/l and four had maximum nocturnal peak below 10 micrograms/l. These data denoted defective spontaneous GH secretion in some of these patients. MRI studies revealed complete empty sella (n = 2), marked diminution of the pituitary size (n = 4), thinning of the pituitary stalk (n = 3) with its posterior displacement (n = 2), and evidence of iron deposition in the pituitary gland and midbrain (n = 7) in those patients with defective GH secretion (n = 9). Serum ferritin concentration was correlated significantly with the circulating IGF-I (r = -0.47, p < 0.01) and IGFBP3 (r = -0.43, p < 0.01) concentrations. These data prove a high prevalence of defective GH secretion in thalassaemic children associated with structural abnormality of their pituitary gland.     

Circulating somatomedin-C levels and the effect of growth hormone-releasing factor on plasma levels of growth hormone and somatostatin-like immunoreactivity in obese children

The effect of growth hormone-releasing factor (GRF) 1-44 on growth hormone and somatostatin release in plasma has been studied in 20 obese children. Twenty age and sex-matched children with normal weight served as controls. Mean peak growth hormone response in obese children after 1 g/kg body wt. GRF 1-44 was significantly lower than in controls (23.7±3.6 ng/ml vs. 41.1±3.0 ng/ml;P<0.01), as were mean integrated growth hormone response areas (1544±272 ng×ml^-1×2 h vs. 2476±283 ng×ml^-1×2 h;P<0.01). Mean plasma levels of somatostatin-like immunoreactivity did not change after GRF in both goups. Mean somatomedin-C levels in obese children were significantly higher compared to controls (1.6±0.4 U/ml vs. 0.86±0.4 U/ml;P<0.01). Somatomedin-C levels were not related to the integrated growth hormone responses. In conclusion there is no relation between somatomedin-C levels and the reduced growth hormone-releasing effect of GRF in obese children. GRF does not alter peripheral somatostatin-like immunoreactivity levels either in normal or obese children.Abbreviations Sm-C somatomedin C - GRF growth hormone-releasing factor - GH growth hormone - SLI somatostatin-like immunoreactivity     

Growth hormone secretion in response to the administration of thyrotropin releasing hormone (TRH) in children with insulin-dependent diabetes mellitus]

In this study the Authors examined the response in growth hormone (GH) to thyrotrophin releasing hormone (TRH) administration in a group composed of 29 children (17 males, 12 females) suffering from insulin-dependent diabetes mellitus (IDDM) (group 1). All subjects were prepubertal, had a chronological age of 8.82 +/- 1.76 years (m +/- SD), a bone age of 8.60 +/- 1.65 years; the time elapsed since the diagnosis was 2.45 +/- 1.51 years, glycosylated hemoglobin (HbA1c) was 7.33 +/- 1.80%. Some of the same subjects (all those with a response in GH to TRH higher than 4 ng/ml; no. 11; group 2) were examined again 12-18 months later; as controls, 13 short children were also examined (group 3). All the subjects of the three groups showed a TSH peak ranging from 10-25 microU/ml, whereas GH peak resulted higher than 4 ng/ml ("paradoxical" response) in 6 subject of the group 1 and in an only subjects of the group 2. All the responders of the 3 groups showed a value in HbA1c higher than 8%. A significant difference was not present between males and females in GH and TSH values. Cortisol levels and glycaemia remained almost constant during the performance of the tests. By considering all the groups, TSH and GH values during TRH-test were not correlated with glycaemia, chronological age, bone age, the time elapsed since the diagnosis, height, height velocity, HbA1c values. In conclusion, our data demonstrated that "paradoxical" response in GH to TRH administration was present only in some subjects and particularly in those with a poor metabolic control of the disease.     

Oral clonidine: an effective growth hormone provocative test

Twenty normal statured healthy children (8 M; 12 F) aged 9-16 years were subjected to growth hormone (GH) provocative tests. The mean basal GH level was 2.0 +/- 0.42 ng/ml (+/- SEM). The mean peak levels of GH were 11.9 +/- 2.19 ng/ml (+/- SEM) after exercise, 9.82 +/- 2.81 ng/ml (+/- SEM) after insulin and 15.2 +/- 2.54 ng/ml (+/- SEM) after oral clonidine. A significant rise (peak level greater than 7 ng/ml) of serum GH was found in 70, 80 and 85% of children after exercise, insulin and oral clonidine tests, respectively. The observation in the present study indicates that oral clonidine test, a safer, easier and more economical test than insulin hypoglycemia, is equally potent and can be done in out patients.     

Administration of low-dose estrogen rapidly and directly stimulates growth hormone production

We tested the concept that estrogen directly stimulates growth hormone (GH) production by determining whether low-dose treatment with ethinyl estradiol increases the GH reserve, as assessed by levodopa administration, without inhibiting somatomedin-C (Sm-C) levels. Twenty-three prepubertal short normal children underwent levodopa tests before and after being treated with ethinyl estradiol. One bedtime dose of ethinyl estradiol (20 to 40 micrograms/sq m, n = 8) resulted in a significant increase in GH levels during levodopa testing, with no significant change in Sm-C levels (0.27 +/- 0.03 vs 0.36 +/- 0.1 units/mL). Two days of a comparable ethinyl estradiol dose (n = 12) raised the mean basal GH level (2.4 +/- 0.4 vs 9 +/- 3 ng/mL) and had a similar effect on peak GH response, without affecting the mean Sm-C level. Eighteen of the 23 patients responded (maximum GH level, greater than or equal to 7 ng/mL) to levodopa before estrogen; all 20 children who received ethinyl estradiol priming in a dose of 20 micrograms/sq m or more also responded. We conclude that low-dose estrogen therapy rapidly stimulates GH production without decreasing Sm-C plasma levels. These results support the concept that the estrogen effect is direct. This action may be important for the stimulation of growth by estrogen. This effect can be conveniently employed to enhance the specificity of the levodopa test for profound GH deficiency.     

Linear growth, growth-hormone secretion and IGF-I generation in children with neglected hypothyroidism before and after thyroxine replacement

We studied growth hormone (GH) stimulation and insulin-like growth factor -I (IGF-I) generation tests in 15 children with neglected congenital hypothyroidism (CH) (age = 6.4 +/- 4.2 years) and measured their growth parameters for >1 years after starting thyroxine (T4) replacement. One year after treatment, height SDS (HtSDS) increased from -4.3 +/- 2.5 to -2.7 +/- 2.3. Peak GH response to clonidine increased from 3.2 +/- 1.2 ng ml(-1) to 7.62 +/- 1.38 ng ml(-1) after treatments. Basal and peak IGF-I response to GH increased from (34.66 +/- 17.3 ng ml(-1) and 58.4 +/- 36.99 ng ml(-1), respectively) before treatment to (130.6 +/- 97.8 ng ml(-1) and 193.75 +/- 122.5 ng ml(-1), respectively). HtSDS increments were correlated significantly with basal free T4 concentrations (r = 0.622, P < 0.01). In summary, after long period of hypothyroidism, T4 replacement produced significant, although incomplete, catch-up growth through a partial recovery of GH- IGF-I axis.     

Administration of arginine plus growth hormone releasing hormone to evaluate growth hormone (GH) secretory status in children with GH deficiency

BACKGROUND: Diagnosis of growth hormone deficiency (GHD) in childhood is usually based on growth hormone (GH) response to at least two provocative stimuli. The aim of this study was to determine whether sequential administration of arginine (Arg) plus GH releasing hormone (GHRH) could be a useful tool in evaluating GHD in children. METHODS: Thirty patients with short stature (mean age 9.0 years) with decreased growth rate were tested for GHD with Arg and the insulin tolerance test (ITT). Patients with confirmed GHD (peak GH <8 ng/ml) were subsequently tested with Arg + GHRH. RESULTS: Maximum GH stimulation for Arg and ITT was 6.3 (1.0-7.8) and 6.7 (0.5-7.7) ng/ml, respectively. Peak GH for the Arg + GHRH test was 36.3 (4.3-84.5) ng/ml and significantly different from the other provocative tests. Peak GH values for the three tests were not significantly correlated between tests or with clinical parameters. There were no significant differences in Arg + GHRH results between children with or without abnormal hypothalamic-pituitary MRI scans. CONCLUSION: Arg + GHRH gave higher GH levels than insulin or Arg alone. Because of the different causes of childhood GHD (hypothalamic and/or pituitary dysfunction), the Arg + GHRH test is unsuitable .for evaluating GHD and deciding whether GH replacement therapy is indicated.     

Effect of a high-fat meal on the growth hormone response to exercise in children

Exercise-induced growth hormone (GH) secretion may significantly modulate growth and development in children. Altered physiological GH responses, therefore, may reduce the beneficial effects of exercise. High-fat food ingestion before exercise blunts the GH response in adults, but it is unknown whether this occurs in children. We therefore performed standard exercise tests, following a high-fat meal or placebo, in 12 children, age 11-15 (6 M, 6 F). GH, insulin-like growth factor-I, glucose, insulin, glucagon, cortisol, epinephrine and interleukin-6 samples were drawn at baseline, end-exercise, and 30 and 60 min post-exercise. While GH was similar at baseline in all experiments, the exercise-induced GH peak was lower after the high-fat meal (6.7 +/- 1.6 ng/l vs 11.8 +/- 2.4 ng/l, p <0.02). Other exercise responses were not affected by prior fat ingestion. A high-fat meal before exercise, therefore (a common event in Western societies), may reduce the growth factor response to exercise in children, with potential implications for growth and development.     

Low dose of insulin for assessment of growth hormone and cortisol release in short children

OBJECTIVE: In 55 prepubertal children with growth failure, aged 8.62 +/- 2.89 years, we evaluated the efficacy of a test using only half the usual dose of insulin by comparing the results with those obtained during a classical arginine tolerance test, performed separately. PATIENTS AND METHODS: The patients were randomly divided into two groups: group A consisting of 37 children received 0.05 U/kg insulin, while group B consisting of 18 patients received 0.1 U/kg insulin. Each child received the same dose of arginine per kg during the second test. RESULTS: Serum growth hormone (GH) peak levels were significantly (p < 0.01) lower in children of group A (6.59 +/- 4.10 ng/ml) than in those of group B (10.12 +/- 5.80 ng/ml). No differences of GH peak levels were found in patients of the two groups after arginine infusion. The injection of 0.05 U/kg insulin induced a significantly (p < 0.0001) lower percent decrease of serum glucose than 0.1 U/kg. No difference of the percent increase of serum cortisol induced by insulin at 0.05 U/kg and 0.1 U/kg was observed. CONCLUSION: The diagnosis of GH deficiency in children can be supported by a blunted GH response after two or more pharmacological stimuli including hypoglycaemia induced by only half the usual dose of insulin.     

The catecholamine response to hypoglycemia in children with isolated growth hormone deficiency syndromes and multiple pituitary hormone defects

We examined the catecholamine response to insulin-induced hypoglycemia in 46 short children evaluated for growth hormone (GH) deficiency by both pharmacologic stimulation and integrated concentration of GH. Twelve patients had quantitatively normal GH secretion by both pharmacologic stimulation and integrated concentration of GH (GHNORM). Twenty-two patients had normal GH to pharmacologic stimulation but subnormal integrated concentration of GH (GHND). Twelve patients had GH deficiency by both tests (GHD): six had isolated GH deficiency (GHD type 1) and six had multiple hormone deficiencies (GHD type 2). There was no significant difference between the peak epinephrine, norepinephrine, and cortisol responses of GH-NORM, GHND, and GHD type 1 patients. The mean peak epinephrine response of GHD type 2 patients was significantly lower (564 +/- 561 pg/ml, p less than 0.03) compared to the other patient groups. There was no significant difference between the peak norepinephrine levels between GHD type 2 patients and the remaining groups. There was no correlation between decrease in blood glucose and either increase in growth hormone, catecholamine, or cortisol concentrations. There was a significant correlation between log peak epinephrine and peak cortisol response (r = 0.53, p less than 0.0002) of the 46 subjects. Neither the basal nor stimulated catecholamine levels correlated with the integrated concentration of cortisol. We conclude that isolated GH deficiency is not associated with impairment of the catecholamine response to hypoglycemia; impairment of the epinephrine response to hypoglycemia is only associated with multiple pituitary hormone deficiencies; in children, the degree of glucose lowering is not correlated with the magnitude of peak GH, catecholamine, or cortisol responses.     

Magnetic resonance imaging in the diagnosis of growth hormone deficiency.

Forty-six patients with idiopathic growth hormone deficiency were examined by magnetic resonance imaging at a mean (+/- SEM) age of 9 +/- 1 years (range 15 days to 20 years). They were classified into two groups according to MRI images: group 1 (n = 29) had pituitary stalk interruption syndrome and group 2 (n = 17) had normal pituitary anatomy. All patients with pituitary stalk interruption had a pituitary height at less than -2 SD for age; three had no visible anterior pituitary lobe. By contrast, the pituitary height was less than normal in only 10 patients (60%) with normal pituitary anatomy. Growth hormone deficiency was transient in one of the seven patients with normal pituitary anatomy and height. The group with pituitary stalk interruption had the first symptom of growth hormone deficiency at an earlier age (2.8 +/- 0.6 vs 5.5 +/- 1.2 years; p less than 0.001), were of smaller stature (-4 +/- 0.2 vs -3 +/- 0.2 SD; p less than 0.01) and had lower GH peak response to provocative testing (3 +/- 0.4 vs 5 +/- 0.5 ng/ml; p less than 0.001) than did the group with normal pituitary anatomy. Their pituitary gland was also shorter (2.5 +/- 0.2 vs 3.5 +/- 0.2 mm; p less than 0.01). All the patients with multiple pituitary deficiencies except one (n = 19) belonged to this group. One girl with pituitary stalk interruption and deficiencies in growth hormone and thyroid-stimulating hormone had advanced central precocious puberty. We conclude that the evaluation of the shape and height of the pituitary gland by MRI is an additional tool for the diagnosis of growth hormone deficiency. The presence of pituitary stalk interruption confirms this diagnosis and is predictive of multiple anterior pituitary deficiencies. The lack of a significant increase in perinatal abnormalities in this group and the association of pituitary stalk interruption with microphallus and with facial or sella abnormalities suggest that this appearance may have an early antenatal origin. The finding of a familial case of pituitary stalk interruption suggests a genetic origin.