Division of Pediatric Endocrinology and Diabetes Duke University Medical Center Durham, North Carolina 27710
Address all correspondence and requests for reprints to: Michael Freemark, M.D., Box 3080, Department of Pediatrics, Duke University Medical Center, Durham, North Carolina 27710. E-mail: freem001{at}mc.duke.edu.
Our society prizes stature like a jeweler values precious stones. Tall men and women embody "power," "beauty," and "authority." In contrast, short adults are "cute" and inconsequential; if aroused, they may become "Napoleonic." Tall children are "mature" and "strong" and garner admiration and respect, whereas short children are too often infantilized, ignored, teased, or bullied. It is hardly surprising, then, that short children and their parents and physicians would seek ways to increase height to an "acceptable" range to meet societal demands and reap its rewards.
Previous studies demonstrated that GH administration can enhance short-term height velocity in otherwise normal short children. That long-term GH therapy can also increase adult height was suggested by a meta-analysis of 11 studies involving a total of 413 young people. However, the mean gain in final height relative to pretreatment predicted height was only 0.4 SD or 2.7 cm, approximately 1 in. (summarized in Ref. 1). In many cases, these studies suffered from a lack of randomization, high dropout rates, and the use of historical, rather than placebo-treated, controls.
In this issue of JCEM, Leschek et al. (2) present a randomized, double-blind, placebo-controlled trial examining the effects of GH in children and adolescents with "idiopathic short stature." They demonstrate that 4.4 yr of GH treatment administered thrice weekly increased ultimate height in boys and girls by a mean of 0.51 SD, equivalent to 3.7 cm (1.5 in.). There were no serious complications of therapy. However, fasting glucose concentrations were significantly higher (95.3 vs. 88.2 mg%) on mornings that followed the previous days GH injection.
The findings of this study no doubt contributed to the recent decision of the U.S. Food and Drug Administration to approve GH for treatment of children with idiopathic short stature. Current guidelines require that GH be administered only to boys whose height predictions fall at or below 5 ft. 3 in. or to girls with height predictions at or below 4 ft. 11 in. A registry of patients, maintained by pharmaceutical companies, will monitor the effects and potential complications of therapy.
Although the use of GH as a physiological replacement for GH deficiency is unquestioned, the administration of pharmacological doses of GH for treatment of idiopathic short stature is problematic. The term "idiopathic short stature" carries the implication of disease; compare, for example, the term "idiopathic epilepsy." But most children with idiopathic short stature have no apparent defects in GH production, secretion, or action, and short stature per se confers no specific dysfunction or disability; indeed, otherwise normal short men may have reduced risks of prostate cancer and, if thin, may live longer than tall, heavy men (3, 4). Moreover, previous reports of serious psychological difficulties and cognitive deficits among short children referred to pediatric endocrine clinics have not always been confirmed in randomly selected populations (5). The major liability of short stature is susceptibility to discrimination. But discrimination is a "disease" of society, not of the short individual. In theory, societal intolerance should be addressed by enacting and enforcing antidiscrimination laws, reeducating the public, and counseling the family rather than by medicating the child.
In the absence of disease, there is no rationale for defining a cutoff for treatment. For example, how does one justify treating a boy whose height prediction is 5 ft. 3 in. but not one whose height prediction is 5 ft. 3 and 1/32 in.? More important, the use of GH in very short children may create an unending cycle of catch-up; increases in the ultimate heights of very short children would necessitate reclassification of some previously normal children (short but with height predictions exceeding current cutoffs) as "idiopathically short." This may be the only circumstance in which treatment of one group of children creates illness in another previously healthy group.
The long-term benefits of GH in otherwise normal short children appear quite limited; the mean increase in ultimate height in the study by Leschek et al. (2) (3.7 cm) was comparable to the mean increase in height in previous studies (1) of GH treatment of normal short children (2.7 cm). Moreover, there were no significant differences between the GH-treated and placebo groups in final adult height SD values relative to gender-adjusted midparental (target) height SD values. However, one half of the children had a significant response to GH; the remaining half had little or no increase in ultimate height. Together with the finding that the growth response correlated with pretreatment IGF-I levels and bone age delay, the results suggest that the term "idiopathic short stature" comprises a heterogeneous group containing normal and GH-resistant subjects (less responsive) as well as patients with subtle, poorly defined defects in GH production and/or secretion (more responsive).
It will be argued that the dose of GH used by Leschek et al. (2) (0.2 mg/kg·wk), the age of onset of treatment (916 yr), and the duration of therapy (4.4 yr) do not provide for optimal increases in ultimate height. Quite likely true; some investigators recommend doses of 0.3750.48 mg/kg·wk administered on a daily basis. But higher doses of GH in younger children raise the risk of potential complications. For example, many of those treated with higher doses of GH will have been born small for gestational age and may therefore be at higher risk for insulin resistance, glucose intolerance, and the metabolic syndrome in later life (6). The long-term risks of prolonged treatment of children with higher doses of GH are unknown, but the increases in fasting glucose concentrations in subjects treated with low doses of GH raise concerns. In their enthusiasm to stimulate linear growth, some physicians will no doubt increase GH doses above those recommended and thereby raise serum IGF-I levels. Chronic exposure to supraphysiological GH and IGF-I concentrations may induce arthralgias, carpal tunnel syndrome, and other features of acromegaly. In theory, there may also be increased risks of hypertension and certain malignancies (711). Given our historical experience with thalidomide, diethylstilbestrol, fenfluramine, troglitazone, hormone replacement therapy, and many other medications initially considered safe, we must ensure that the benefits of GH clearly exceed its short- and long-term risks.
One potential consequence of expanding the base for GH therapy is the economic burden inflicted upon society at large. At a cost approximating $40/mg of hormone, the yearly cost of treating a short child with GH ranges from $5,000 to $40,000 (approximate mean, $18,00020,000). The cost of treating even 0.1% of the childhood population runs into the hundreds of millions of dollars. Such expenditures cannot be justified for treatment of normalchildren, particularly when risks are incurred.
Yet some children with idiopathic short stature likely have true, albeit poorly defined, abnormalitiesin GH production, secretion, or turnover. How can these subjects be identified, and how should they be managed? The initial clinical assessment may be enlightening: short children with the lowest pretreatment growth velocities and IGF-I levels and greatest delays in bone maturation are most likely to benefit from GH therapy (2). Radiological studies may also be informative, particularly when the GH response to pharmacological stimulation is equivocal (710 ng/ml); those with small pituitary glands and midline defects are most likely to have reduced plasma concentrations of IGF-I and IGF binding protein-3 (12, 13) and to respond to GH (14). Finally, the initial response to GH treatment may be instructive. Those that show striking increases in growth velocity during the first 612 months of therapy are most likely to profit from long-term treatment; in contrast, treatment should be discontinued if there is no significant increase in growth rate during the first year. In either case, careful monitoring for potential complications and full disclosure of costs and possible risks are mandatory.
Received April 7, 2004.
Accepted April 7, 2004.
References