Editorials

Growth hormone in growth hormone deficiency

Start treatment early and give it for long enough

Papers p 70

With the advent of an unlimited supply of recombinant DNA growth hormone some 15 years ago endocrinologists and paediatricians hoped that the major goal in treating children with growth hormone deficiencythat is, a near normal adult heightwould finally be achievable.

Carel and colleagues report in this issue the adult height after "long term" recombinant growth hormone treatment for idiopathic isolated growth hormone deficiency.¹ The investigators were able to analyse entry data on all French children with growth hormone deficiency whose treatment started between 1987 and 1992 and stopped in 1996 under the auspices of the French national programme, Association France Hypophyse. They then were able to record adult heights for 76% of these patients. Gain in height was on average a disappointing 1.1 (0.9 SD). Overall, the treatment of a child for about three years (certainly not "long term") was associated with an estimated gain in height of only 4.2 cm. Faced with these unimpressive results, the authors conclude with understandably muted enthusiasm that (a) the effect of growth hormone is unclear in many patients treated for idiopathic isolated growth hormone deficiency and that many patients may have simple constitutional delay of growth and development, and (b) only patients with "severely and permanently" altered growth hormone secretion should be treated with growth hormone.

We have been treating growth hormone deficiency with growth hormone in children for over four decades.² The paper by Carel et al would suggest that we still do not know whom and how to treat. What then are the possible factors contributing to these disappointing results? The roughly 2800 children in this study underwent routine growth hormone stimulation tests and were treated only if they had a maximum growth hormone peak of less than 10 µg/l. Unfortunately, 25% of the children included in the study had either neurosecretory dysfunction or simply inadequate criteria for growth hormone deficiency. Neurosecretory dysfunction is still an ill defined term. In this context it meant that they had normal stimulated growth hormone values (above 10 µg/l) yet their spontaneous overnight growth hormone secretion was low. The authors recommend cogently that the diagnostic criteria for growth hormone deficiency should be redefined. They propose that peak levels be pegged at 2-4 µg/l, that oestrogen priming is used for growth hormone testing, and that more attention is paid by physicians to the causes of hypopituitarism. These are all well intended suggestions. However, had these criteria been applied to the group the French researchers studied less than a paltry 3% would have been eligible for treatment with growth hormone.

The Growth Hormone Research Society has published consensus guidelines for diagnosing and treating growth hormone deficiency in childhood and adolescence.³ These guidelines do not recommend oestrogen priming because it is really an unphysiological manoeuvre in the prepubertal child. Recognising the well known problems of basing the diagnosis of growth hormone deficiency on measurements of growth hormone alone, the diagnosis of this condition in childhood requires not only endocrine but also auxological assessment and growth hormone determinations. Growth hormone measurements should be combined with insulin like growth factor-1 (IGF-1) measurements.⁴ No IGF-1 data are presented in the French study. A growth hormone assay of below 10 µg/l has been used to support the diagnosis. However, it is well known that this threshold needs to be lowered when new monoclonal assays are used. Since we do not know what assays were used in the French study, this argument too becomes mute.

We have known for a long time that there is a continuum of growth hormone secretion that ranges from a moderate deficiency to a severe one, as seen specifically in congenital growth hormone deficiency presenting in infancy and congenital or acquired multiple pituitary hormone deficiency. There will always be overlap between normal children and those with growth hormone deficiency, and therefore an approach taking into account auxology, IGF-1 as well as IGF binding proteins measurements, growth hormone levels, age, and bone age will afford better diagnostic criteria. In the absence of a gold standard therefore the recommendations of the Growth Hormone Research Society conclude that it is important that the clinician integrates all of the available dataclinical, auxological, radiological, and biochemicalwhen making a diagnosis.

The difficulty of diagnosing growth hormone deficiency especially in the immediate, peripubertal period, has been well recognised. Falsely low growth hormone levels in provocation tests may frequently occur, particularly in overweight children. The factors that have been found to influence or predict the response of treatment with growth hormone include the severity of the deficiency, genetic target height (that is the sex adjusted average of parents height), age at start of treatment, duration of treatment, and the dose of growth hormone.⁵ All of these predicted methods have limited accuracy and have a substantial error in individual cases, limiting the usefulness for the individual patient. The more profoundly deficient patients, however, do grow better.⁶

Short term prediction models that more precisely predict the long term growth related effects of treatment with growth hormone are desirable. A new model incorporating not just the conventional factors, but also serum IGF-1 and IGF binding protein-3, as well as urinary markers of bone metabolism markersthus correlating the growth hormone stimulated bone turnover together with the auxological responsemay improve the accuracy of prediction models as they affect the sensitivity of individual patients to growth promoting effects of growth hormone.⁷

The overly pessimistic conclusions that growth hormone therapy is inappropriate in most children so treated do not take into account that the patients were older and that they were treated for too short a time. In growth hormone deficiency, as in Turner's syndrome, there are now studies clearly indicating that the two major factors guaranteeing a more successful treatment outcome are early onset of treatment allowing for longer duration of treatment and a higher dose of growth hormone. ^{8 9} In growth hormone deficiency, adult height in 121 subjects for males and females was 0.7 SDS compared to mid-parental target height scores 0.6 and 0.4, respectively. Both numbers indicate a much more successful therapeutic outcome, and the children reached adult heights in males of 171.6 +/ 8.2 cm and in females 158.5 +/7.1 cm. Total gain in height was 2.4 and 2.7 SDS respectively. The mean duration of treatment was 6.2 yearsthe duration of treatment was thus twice as long as the French study and the dose of treatment was also twice as much, that is, 0.3 mg/kg/week (0.9 IU/kg/week compared to 0.14 mg/kg/week). Similar conclusions can be drawn from a long term study in Turner's syndrome published by Dutch investigators.⁹ These much more robust responses indicate that we should not conclude that growth hormone is ineffective when treatment offered is too late and too little.

We clearly have to hone our diagnostic criteria (evaluate IGF-1 levels) and should avail ourselves of recent advances in molecular endocrinology allowing more refined diagnosis of particular gene defects as causes of short stature.^10-12 In real estate dealings, it is "location, location, location," that countsin growth hormone therapy it's "duration, duration, duration" that counts. That approach in conjunction with an appropriate growth hormone dose should net more encouraging results while the search for further refinement in diagnostic and therapeutic criteria continues.

Paul Saenger, professor of paediatrics. 

Department of Pediatrics, Division of Pediatric Endocrinology, Children's Hospital at Montefiore/Albert Einstein College of Medicine, Bronx, NY 10467, USA (phsaenger{at}aol.com)