| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Editorial |
Developmental Endocrinology Branch, National Institutes of Health, Bethesda, Maryland 20892-1103
Address all correspondence and requests for reprints to: Dr. Jeffrey Baron, National Institutes of Health, Building 10-CRC, Room 1-3330, 10 Center Drive, MSC 1103, Bethesda, Maryland 20892-1103. E-mail: jeffrey.baron{at}nih.gov.
GH is used to treat children with GH deficiency and children with specific categories of non-GH-deficient short stature. In both situations, the dose of GH is typically based on the child’s body weight, approximately 25–50 µg/kg·d. The response in terms of growth velocity is quite variable, even within the same diagnostic category, suggesting that individual sensitivity to GH may vary considerably. One possible method to adjust for individual sensitivity might be to titrate the GH dose to serum IGF-I levels. This approach has some theoretical appeal because circulating IGF-I levels are regulated by GH. IGF-I levels tend to be low in GH deficiency and in GH insensitivity and high in states of GH excess. In children treated with GH, IGF-I levels rise in a dose-dependent manner (1). In fact, circulating IGF-I is not just a marker of GH action; it also mediates many of the physiological effects of GH, a concept referred to as the somatomedin hypothesis (2). Based on these concepts, one might predict that we could use serum IGF-I levels to tailor the GH dose to a child’s individual responsiveness.
However, that prediction may not be so straightforward. IGF-I is bound by a family of IGF-binding proteins that affect its bioavailability, yet clinically we generally only measure total IGF-I. Furthermore, circulating IGF-I is thought to reflect primarily hepatic synthesis. However, other tissues produce IGF-I, not always regulated by GH, and this local production may be an important determinant of tissue effects, including linear growth (3). Some consequences of GH are IGF-independent (4, 5). Thus, by titrating to circulating IGF-I, which reflects primarily hepatic output, we might be overtreating some tissues and undertreating others.
In children with non-GH-deficient short stature, the situation might be even more complex. Short stature has many causes, both environmental and genetic, including genetic defects affecting the GH-IGF-I, fibroblast growth factor, C-type natriuretic peptide, PTHrP-Indian hedgehog, and other systems that regulate growth plate function. However, much of short stature is idiopathic. Some children with idiopathic short stature have a low or low-normal serum IGF-I. How often this tendency reflects subtle GH deficiency or GH insensitivity is unclear (6, 7). Some of these children have decreased weight for height, which is not typical of GH deficiency, suggesting that their decreased growth and IGF-I may reflect undernutrition (8, 9, 10). In other children, short stature and low serum IGF-I might simply represent a maturational delay. Still others may have genetic defects outside of the IGF system that affect growth plate function (11). Whether total circulating IGF-I is an appropriate gauge of GH action might depend on the cause of the short stature and decreased IGF-I. For example, undernutrition might have differential effects on the GH responsiveness of the liver compared with growth plate (12).
If theory does not give us much guidance, are there empirical data with which to assess IGF-I-based dosing of GH in children? Unfortunately, there is only indirect evidence. In adult patients, GH dose is typically adjusted based on clinical response, avoidance of side effects, and to achieve an IGF-I level within the normal range for age (13). There is some evidence that such individual adjustment may produce fewer side effects than a fixed-dose regimen (14, 15). However, GH therapy in children differs markedly from therapy in adults in terms of the clinical indications, therapeutic goals, doses, and side effect profiles, and thus observations in adult populations may not apply to children. In children, there is some indirect evidence suggesting that IGF-I levels might be used to gauge GH sensitivity. For example, Kriström et al. (16) reported that, in children starting GH therapy, the increment in IGF-I levels correlated with growth rate on treatment. However, the correlation (R2 = 0.12) was not particularly strong compared with other clinical and biochemical variables analyzed. Furthermore, the correlation might be weaker for the absolute IGF-I value on GH therapy than it was for the change in IGF-I because pretreatment IGF-I correlated negatively with growth rate (16).
In this issue of the Journal of Clinical Endocrinology and Metabolism, Cohen et al. (17) provide important new information about IGF-I-based dosing of GH in children. They performed a 2-yr randomized, controlled trial in prepubertal children with short stature and serum IGF-I SD score (SDS) no greater than –1.0. Some of the children met criteria for GH deficiency, whereas others did not. The subjects were randomized to receive GH in one of three regimens: 1) a conventional regimen in which GH dose was based on body weight (40 µg/kg·d); 2) an IGF(low) regimen in which the GH dose was titrated to achieve an IGF-I SDS = 0; and 3) an IGF(high) regimen in which GH was titrated to achieve an IGF-I SDS = +2. The first lesson from this study was that IGF-I-based dosing of GH is clinically feasible in children; the target IGF-I levels were usually achieved within 6–9 months. In this study, comparison between the conventional group and the IGF(low) group is informative because these two groups received similar doses of GH but differed in the dosing approach; in one group the dose was based on body weight, whereas in the other group it was titrated to a mid-normal IGF-I level. Efficacy, in terms of growth rate, and bone age advancement were similar. One might hope that IGF-based dosing would produce a more uniform growth response, but, based on the presented graph of change in height SDS as a function of time, the IGF(low) group appeared to have at least as much variability in growth response as the conventional group. There were no apparent differences in safety between the groups. However, side effects of GH treatment in children are uncommon, and therefore studies that are powered to assess efficacy differences may not be adequately powered to assess safety differences. In summary, comparison of these two groups does not demonstrate superiority of either the weight-based or the IGF-I-based regimen.
This study also provides a comparison between two different target IGF-I levels—at the upper limit of normal (+ 2 SD) for the IGF(high) group vs. mid-normal (0 SD) for the IGF(low) group. The IGF(high) group showed a greater growth rate than did the IGF(low) group, but this was achieved at the expense of much greater GH doses. A target IGF-I level of 0 SD was generally achieved with a GH dose (mean, 33 µg/kg·d) that is within the range for which we have considerable safety data. In contrast, the GH dose required to maintain an IGF-I at the top of the normal range was more than 3-fold higher, (mean, 110 µg/kg·d), a level for which we have less safety information. The IGF(high) group did not show more rapid bone age advancement. However, a treatment duration of 2 yr may be too short to detect a statistically significant difference in bone age. A study using a GH dosage of 70 µg/kg·d in prepubertal children with idiopathic short stature showed a significant bone age advance and earlier onset of puberty (18). The safety profiles of the two regimens were similar, but again we must remember that large numbers of subjects might be required to show safety differences in this setting. Also, some possible risks, such as development of cancer, might require long-term follow-up. Should we take reassurance from the fact that the IGF-I levels in the IGF(high) group were at the upper limit of normal but not above it? Not necessarily. We should make a distinction between the normal range and the desirable range. Blood pressure, cholesterol, and BMI values at + 2 SD are within the normal range but are still risk factors for cardiovascular disease. Similarly, IGF-I levels in the upper part of the normal range have statistically been associated with increased rates of prostate, breast, and colon cancer in adults (19, 20, 21).
If we were discussing a treatment for a disease with serious morbidity, we might be satisfied that the IGF(high) group showed evidence of increased efficacy without evidence of increased risk. However, this is not the case for short stature. Although short stature may be quite unpleasant for some individuals and carry social disadvantages, it generally does not cause death, serious physical dysfunction, or probably even serious psychological dysfunction (22). Furthermore, the psychological benefits of GH treatment are hard to establish. Because risk must be weighed against benefit, those of us who would treat short stature must adopt highly stringent safety criteria. We are condemned to be constantly looking over our shoulder for even a low risk of a serious adverse outcome.
In summary, we do not have evidence that GH dosing titrated to IGF-I levels is superior to dosing based simply on body weight. What should we do as clinicians until there is clear-cut evidence one way or the other? The principle of Primum non nocere would suggest that we should continue to use weight-based dosing for now and select doses within the range for which we have the most safety data. Titrating the dose downward because of a high serum IGF-I level seems prudent (23, 24). However, outside of a rigorous clinical trial, titrating upward to a GH dose above the well-studied dose range is not justified by existing data. Perhaps in the future, additional well-controlled studies, such as that of Cohen et al. (17) will allow us to individualize the GH dose based on growth rate (23), serum IGF-I concentration, other measures of GH action, etiology of the short stature, and/or genetic determinants of GH responsiveness (25).
Footnotes
Received April 19, 2007.
Accepted May 2, 2007.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
