Growth hormone deficiency in anorexia nervosa

Predictors of late menarche and adult height in children with anorexia nervosa

OBJECTIVES

The onset of anorexia nervosa (AN) during childhood can affect the timing of puberty and adult height. The aim of the study was to evaluate the determinants of late menarche and adult height in children with AN.

PATIENTS AND METHODS

We carried out a retrospective, longitudinal, university hospital-based study. All prepubertal or early pubertal girls diagnosed with AN between 1998 and 2002 were selected for the study. Participants (n = 33) were studied at a median age of 21 (19.8-24.3) years. AN was diagnosed at 11.8 (10.7-12.3) years.

RESULTS

Patients with AN reached menarche at significantly greater ages than their mothers [15.4 (13.5-16.8) vs. 13.2 (12.0-14.5) years, P < 0.01]. Chronological age at onset of AN and lowest body mass index (BMI) were important independent predictive factors for delayed menarche (P < 0.01). Adult height was 165.0 (163.0-172.0) cm, 2.5 (-1.5 to 5.0) cm above target height. Twelve patients (36%) did not reach their target height and had a median height deficit of -3.9 cm with respect to their target height. The duration of hospitalization, a marker of disease severity and chronicity, was an independent predictor of the difference between adult height and target height for a given individual (beta coefficient = -0.07; P = 0.01). The other factors studied (i.e. age at onset of AN, pubertal stage at diagnosis of AN, lowest BMI reached, associated comorbidity if any, type of AN, age at menarche) had no significant effect on adult height.

CONCLUSION

The intensity of the disease affects the timing of menarche but not adult height in most patients. Hospitalization, despite often being an effective means of managing AN, does not reduce the impact of AN on growth.

Disturbances in growth hormone secretion and action in adolescents with anorexia nervosa

Women in whom anorexia nervosa develops during adolescence have failure of linear growth associated with low levels of insulin-like growth factor I (IGF-1). To investigate the pathophysiology of growth retardation in adolescents with anorexia nervosa, we measured basal growth hormone (GH), growth hormone-binding protein (GHBP), IGF-1, and insulin-like growth factor binding protein-3 (IGFBP-3) in three groups of patients: (1) 28 recently hospitalized female adolescents with anorexia nervosa, (2) 23 of the same patients after partial weight restoration, and (3) 28 healthy control subjects matched for age, sex, and pubertal stage. Fasting GH levels in group 1 did not differ significantly from those in group 3. In contrast, serum GHBP (p < 0.001), IGF-1 (p < 0.001), and IGFBP-3 (p < 0.01) were significantly lower in group 1 than in group 3. Serum GHBP and IGFBP-3 levels were positively correlated with body mass index. Serum GHBP levels were low in patients in all five pubertal stages and even in those shown to have adequate GH secretion. In group 2 (after refeeding) the serum IGF-1 concentration increased significantly and GHBP and IGFBP-3 returned to normal. We conclude that patients with anorexia nervosa have diminished GH action resulting in decreased secretion of IGF-1. The positive correlation with body mass index and the reversibility with refeeding suggest that these changes are secondary to malnutrition. Altered GH function that occurs during the years of active growth can explain the growth retardation seen in anorexia nervosa.

Blunted growth hormone responses to clonidine in adolescent girls with early anorexia nervosa. Evidence for an early hypothalamic defect

We studied growth hormone (GH) levels in 14 adolescents with anorexia nervosa (AN), 12-20 years of age. All had amenorrhea and the duration of symptoms ranged from 2-24 months. There was no difference in unstimulated GH between the girls with AN and endocrinologically normal girls. However, the GH response to the alpha-adrenergic agent clonidine was blunted (less than 10 ng/ml) in 11 of 14 girls with AN. Our results of a normal basal GH level and blunted response to stimulation is in contrast to some reports of high basal GH levels and an exaggerated response to pituitary stimulation although it agrees with other studies. Our patient population was younger and at an earlier stage of the disease (12 with symptoms for less than 1 year) than the AN patients previously reported to have elevated GH secretion. In view of studies that report an elevated GH response to growth hormone releasing factor, our results suggest that the AN patient's abnormalities in GH secretion may result from an abnormal hypothalamic response to adrenergic stimuli. This hypothalamic abnormality might also contribute to the growth failure seen in adolescent girls with AN.

Laboratory tests and nutritional assessment. Protein-energy status

1. There is no one definitive laboratory test, group of tests, or indices that are satisfactory for the assessment of protein-calorie status. 2. Clinical evaluation remains the simplest, most widely available, most reproducible, and wisest method. It is satisfactory for the majority of clinical situations. 3. The laboratory should be used selectively and to complement clinical evaluation. Routine testing must be relegated to research protocols. 4. Always include some assessment of inflammation (for example, C-reactive protein, erythrosedimentation rate [ESR]), because its presence affects the interpretation of all the other tests. 5. The pathophysiologic effects of the underlying disease, especially in hospitalized patients, will affect the interpretation of every one of the laboratory tests. 6. Nutritional status often impacts more on the interpretation of commonly performed laboratory tests than laboratory tests impact on the assessment of nutrition.

Which children should have growth hormone therapy?

Existing criteria for the diagnosis of growth hormone (GH) deficiency do not identify all children who can be made to grow faster with GH. There is a spectrum of GH secretion in short slowly growing normal children and in some cases the secretion may be inadequate to promote optimum growth in height. A diagnostic/therapeutic trial of GH therapy, with auxological monitoring, may be the only means of identifying some patients who will benefit. Assessment of the place for GH therapy in the treatment of short stature requires special knowledge of childhood diseases, growth, and endocrinology.

Short stature in anorexia nervosa patients

In the treatment of many anorexia nervosa (AN) patients, we observed clinically that these girls were short in stature. In order to test this hypothesis, we reviewed the height records of all patients seen with a previous diagnosis of AN in an eight-year period. Of 104 patients, 85 were suitable for study and were compared to a control group of 85 age-matched subjects. Seventy-six percent of the AN patients (mean age 15.8 years) were below the 50th height percentile for age; 14.1% were below the 5th percentile; and 25.8% were between the 25th and 49th percentile. Parental height was available for 35 subjects. The AN patients were significantly shorter than their parents, and comparison suggests an impairment of growth rather than a familial etiology. As 80% of subjects developed AN after menarche, malnutrition does not appear to be responsible for the observed height deficits.

Anorexia nervosa presenting as growth retardation in adolescents

Two adolescent males and one adolescent female were referred for evaluation of short stature. Extensive laboratory studies, often performed over several years, were unremarkable. After several years of observation, the diagnosis of anorexia nervosa was established. We emphasize the necessity to consider anorexia nervosa in the differential diagnosis of growth retardation in young adolescents.

Somatomedin activity and growth hormone secretion. I. Changes related to body weight in anorexia nervosa

Anorexia nervosa in childhood and adolescence, associated with impaired growth can be considered as a form of chronic malnutrition. Twelve patients aged 11 to 17 years were investigated. In spite of increased GH secretion, plasma somatomedin activities were diminished in 8 cases. Plasma T4, T3 and arginine stimulated insulin secretion were also decreased. In six patients who gained weight a significant negative correlation was found between weight deficit and plasma somatomedin activity. Prolonged administration of growth hormone in one case did not stimulate the generation of somatomedin activity. Nutrition and increased insulin secretion could plan a role in changes observed during recovery.