Statistics from Altmetric.com
Editor,—The recent interesting and valuable article by Johnston and colleagues1 describing the outcome of a trial of recombinant growth hormone (GH) and low dose oestrogen in girls with Turner syndrome (TS) concluded that low dose oestrogen before planned induction of puberty was not beneficial for adult height. However, they extend their conclusions by the cautious word that although the majority of girls might benefit from GH treatment, a “realistic appraisal” suggests “modest” benefit. Although evidence to the contrary is fully discussed in their paper, this generalisation might lead the reader to doubt the efficacy of GH in TS.
The best known of the trials of GH in TS is that of Rosenfeld and colleagues2 who followed their patients until the age of 17–18 years (near final height). Although they started this trial with a randomised untreated control arm who grew at a rate of 3.8 cm per year in contrast to girls in the treatment arms who grew more rapidly, the former were placed in a treatment arm of the study. Therefore historical controls were needed for comparison of near final height. The historical controls achieved an adult height of 142.2 (6.0) cm, comparable with their original projected adult of 142.2 (6.1) cm. The group treated with GH alone gained 8.4 (4.5) cm height and the group treated with GH and oxandrolone gained 10.3 (4.7) cm over their projected heights. The benefit from GH treatment seemed to be more than modest, so why the discrepancy between the US results and those of Johnston et al? There could be a number of reasons but a striking contrast is in the use of oestrogen; Rosenfeld and colleagues did not induce puberty until a minimum age of 14 years and at least three years of GH treatment. Johnstonet al induced puberty at 12 years and many of the girls had already had low dose oestrogen for some years, the very purpose and design of the study.
Chernausek and colleagues3 have thrown light on the timing of the use of oestrogen in girls who received GH treatment. They found that the number of years on GH treatment prior to introduction of oestrogen was a strong predictor of height gained (the equation was given simply: height gain in cm = 2.1 × years on GH before oestrogen; p < 0.0001; r 2 41%).
There is no doubt that the lack of a prospective randomised control study with an untreated arm until adult height has raised important doubts about the efficacy of GH for improved adult height. These doubts have been increased because of clinicians' experience of treating individual girls subsequent to the licensing of GH for TS. The availability of GH treatment for TS girls led to the treatment of a much older population compared to the US trial, and oestrogens were often introduced close to the onset of GH treatment. The results were “modest” or of no benefit.
To overcome the problem of being unable now to run a study with an untreated arm, Sas and colleagues4 cleverly devised a randomised dose response study. The lowest dose of GH was 4 IU/m2/day in a group of girls who started GH at 7.9 (0.9) years, oestrogen at 12.7 (0.6) years, and completed 93.3 (8.5) months of GH treatment. For this standard dose group their projected heights were 146.2 (7.5) cm and their achieved last heights were 158.8 (7.1) cm. The group receiving 8 IU/m2/day had significantly greater gains over projected heights and greater latest heights. This seems to be good evidence that there is a GH effect and that the gains are clinically useful.
What then should be our “best” practice in 2001? Based on the evidence of the thorough trials discussed above, we feel that it is justified to make efforts to diagnose girls with TS early so that they can receive at least four years of oestrogen free GH treatment with a standard dose. The issues involved in the timing of pubertal induction are complex and not just related to height as an outcome, but one should be aware of Chernausek's analysis of the relationship of oestrogen free years and height gained.
However, although cohorts of TS girls may incur significant benefit in adult height, there remains considerable variability in response, both in the short and long term, between individuals. A reasonable approach would be for the child and the parents to be given an estimate of the expected response in the first and subsequent years, and should there be a serious shortfall in achieved response, then issues of treatment adherence, tissue resistance, and other incidental diseases need investigation. Ranke and colleagues5 have shown that a major predictor of growth response in the second, third, and fourth years of GH is the first year response, and therefore the end of the first year of GH treatment is an appropriate time for reassessment of likely long term benefit. If the factors inhibiting first year response cannot be satisfactorily addressed, it is unlikely that there will be more than a modest effect on adult height, and then the patient, parents, and doctor may agree on cessation of treatment.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.