If a child is significantly below average stature for his or her age group, should hormone therapy be routinely offered?

Yes. The definition of deficiency is arbitrary and outdated. Established in the age of growth hormone's scarcity, it was partly a device to ration it.

Not all short children should be treated. We will treat children who do not spontaneously release sufficient growth hormone. Current tests do not identify all these children, but their rate of growth is consistent with the rate found in growth-hormone deficient children rather than normal children.

Under the classic definition, a child shouldn't be treated unless he has abnormally low serum growth-hormone levels in response to two pharmacologic stimuli. But the test is imprecise. And even if done accurately, standard tests may miss children with a real deficiency. Some with poor growth rates have normal growth-hormone responses but 24-hour secretion profiles indistinguishable from those with the classic deficiency.

When we treat such kids, their response to the hormone is indistinguishable from that of children with classic deficiency. I think that proves there are kids with a non-classic deficiency for whom normal growth depends on receiving this treatment.

There are risks, but experience over the past 25 years shows it's relatively safe. There may be a risk of hypothyroidism. But that's not difficult to treat, and it probably occurs only in children who already had borderline thyroid function. Diabetes isn't a major risk with conventional doses, which aren't nearly as high as the growth-hormone levels of patients who have growth-hormone-producing tumors, who are at high risk of diabetes.

Epidemiologically, there may be a link between growth-hormone treatment and increased leukemia risk, but the link is weak and may reflect factors that cause both growth-hormone deficiency and leukemia, such as brain tumors treated with radiation or chemotherapy. -- Gilbert P. August, MD Attending pediatric endocrinologist, Children's National Medical Center, Washington; professor of pediatrics, George Washington University

No. Neither risks nor benefits are known well enough to warrant giving these children growth-hormone therapy outside a controlled clinical trial.

It's clear that growth-hormone therapy can add to children's medical problems. It can lower glucose tolerance and accelerate the onset of diabetes. It can also lower thyroid function. We don't know enough about its long-term effects. A worldwide survey suggested it may be associated with a slightly higher leukemia risk.

If the child has a clearly identified deficiency, we know he's not going to grow normally without treatment, so these risks are worth taking. But if there's no clear deficiency, the risks are not worth taking, especially since it's not certain that the hormone will enhance final height.

Studies show that about a third of children without the classic deficiency who are given growth hormone do not experience any significant changes in growth velocity over the short term. Why expose them to the unknown risk therapy poses when there's no apparent benefit?

Some people advocate expanding the definition to include children whose secretion profiles are abnormally low. Our study, however, showed that the overlap between normal and hormone-deficient children is so great that it's not reasonable to rely on this test to tell them apart. -- Susan R. Rose, MD Research assistant professor of pediatrics, University of New Mexico; adjunct scientist in developmental endocrinology, National Institute of Child Health and Human Development