Inhibiting estrogen to slow the closing on bone plates in adolescent boys

posted by Loco on WCBB

Inhibiting estrogen to slow the closing on bone plates in adolescent boys
The Lancet, June 2, 2001 v357 i9270 p1743
A specific aromatase inhibitor and potential increase in adult height in boys with delayed puberty: a randomised controlled trial. (Brief Article) Sanna Wickman; Ilkka Sipila; Carina Ankarberg-Lindgren; Ensio Norjavaara; Leo Dunkel.
Abstract: The authors discuss inhibiting estrogen to slow the closing on bone plates in adolescent boys. Using testosterone and letrozole compared against testosterone and an inert substance, they determined that the inhibition of estrogen did slow bone maturation, allowing time for correction of growth disorders.

Full Text: COPYRIGHT 2001 The Lancet Publishing Group, a division of Elsevier Science Ltd.

Summary

Background The role of oestrogens in the closure of growth plates in both sexes is unequivocal. We postulated that inhibition of oestrogen synthesis in boys with delayed puberty would delay maturation of the growth plates and ultimately result in increased adult height.

Methods We did a randomised, double-blind, placebo-controlled study in which we treated boys with constitutional delay of puberty with testosterone and placebo, or testosterone and letrozole. Boys who decided to wait for the spontaneous progression of puberty without medical intervention composed the untreated group.

Findings letrozole effectively inhibited oestrogen synthesis and delayed bone maturation. Progression of bone maturation was slower in the letrozole group than in the placebo group. In 18 months, bone age had advanced 1.1 (SD 0.8) years in the untreated group and 1.7 (0.9) years in the group treated with testosterone and placebo, but only 0.9 (0.6) years in the letrozole group (p=0.03 between the treatment groups). Predicted adult height did not change significantly in the untreated group and in the placebo group, whereas in the group treated with letrozole the increase was 5.1 (3.7) cm (p=0.004).

Interpretations Our findings suggest that if oestrogen action is inhibited in growing adolescents, adult height will increase. This finding provides a rationale for studies that aim to delay bone maturation in several growth disorders.

Lancet 2001; 357: 1743-48

Introduction

Raised concentrations of sex steroids are generally assumed to be the primary cause of accelerated pubertal growth. However, the role of androgens and oestrogens in the regulation of pubertal growth in boys is obscure. The growth-enhancing effect of androgens was established by studies in which a non-aromatisable androgen, oxandrolone, was shown to accelerate growth in boys with constitutional delay of puberty,(1,2) and by an investigation in which infusion of another non-aromatisable androgen, dihydrotestosterone proved to induce ulnar growth in prepubertal or early pubertal boys.(3) Evidence for an effect of oestrogens on growth in boys was obtained from a study in which a low-dose infusion of oestradiol accelerated ulnar growth in prepubertal or early pubertal boys.(4)

The role of oestrogens in the growth of boys was further clarified in three different reports of cases in which oestrogen action was suppressed by mutations in the genes for the oestrogen-receptor a5 or for the enzyme P-450 aromatase.(6,7) At the age of 24 years or older, all these men were taller than 190 cm, had unfused epiphyses of the long bones, and were still growing. In the men with P450-aromatase deficiency, treatment with oestrogens induced epiphyseal closure in 6 months(8) and 9 months.(7) Furthermore, a man with a mutation in the gene for oestrogen-receptor a had no recollection of pubertal growth acceleration despite normal testosterone concentrations.(5) These case reports confirm that oestrogens are essential hormones for epiphyseal closure in men. They also suggest that, although oestrogens are important hormones in the pubertal growth spurt, they do not participate in linear growth.

Several study results have shown that some boys with constitutional delay of puberty do not exploit their genetic growth potential.(9-12) In these boys, androgen treatment does not increase adult height.(11,13,14)

As evidence for the role of oestrogens in epiphyseal closure seems unequivocal, we postulated that boys with constitutional delay of puberty would attain greater adult height if oestrogen action were suppressed. Thus, we can expect that such boys would achieve an adult height closer to their genetic growth potential.

Methods

Patients

Between January and December, 1998, 38 boys were referred to the outpatient clinic at the Hospital for Children and Adolescents, University of Helsinki, Finland, for assessment of delayed puberty, short stature, or both. Three boys refused to participate in the study, and two were excluded because they had already reached mid puberty (figure 1). The last outpatient visit was done in June, 2000. Diagnosis of constitutional delay of puberty was defined as a Tanner genital or pubic hair stage observed at an older age than the mean +2 SD for healthy Finnish boys,(15) or a testis volume of less than 4 mL after 13.5 years of age. At entry, none of the boys had had any pubertal increase in growth velocity. Medical history, clinical examination, or routine laboratory tests all failed to reveal any signs of chronic illnesses accounting for the delayed puberty. 25 (76%) boys had a family history of delayed puberty. None of the boys had received any previous sex-hormone treatment. Two boys were receiving inhaled corticosteroid treatment for asthma.

Ten boys with a mean age of 15.0 (SD 0.7) years decided to wait for spontaneous progression of puberty without medical intervention, and thus composed the untreated group. 23 boys with a mean age of 15.1 (0.8) years wanted medical intervention, and we randomly assigned them to receive one or other of the two treatments. One group (12 boys) received Testosterone Enanthate (Testoviron-Depot-250, Schering, Berlin, Germany) six times at a dose of 1 mg/kg intramuscularly every 4 weeks, and placebo orally once daily for 12 months. The other group (11 boys) received testosterone enanthate (as above) and, in addition, a specific and potent, fourth-generation aromatase inhibitor, letrozole (Femar, Novartis AG, Stein, Switzerland) 2.5 mg orally once daily for 12 months. The project was done as a randomised, double-blind, placebo-controlled study between the treated groups. Written informed consent was obtained from the patient and from his guardian. The protocol was approved by the ethical committee of the Hospital for Children and Adolescents and the National Agency for Medicines.

Study protocol

We examined patients at the start of the study, at 2 months (about 7 days after the third testosterone injection), at 5 months (about 7 days after the sixth testosterone injection), at 12 months, and at 18 months. We measured height on a Harpenden stadiometer with 0.1 cm precision. Three of us (SW, IS, LD) were unaware of the groups and we estimated bone ages by Greulich and Pyles' method.(16) We first ranked bone age X-ray films of every time point in successive order according to maturation, and then ascertained bone age for each film. We calculated adult height predictions by the Bayley-Pinneau method;(17) the table for boys with average skeletal maturity was used, since the bone ages in most of the boys exceeded the range of bone ages reported for boys with retarded skeletal maturity. We assessed pubertal stages according to Tanner,(18) and calculated testis volumes from the formula length3width230.52.(19) Results are presented as means of the measurements for the two testes. The bone mineral densities of the first through fourth lumbar vertebrae and femoral neck were measured by dual-energy X-ray absorptiometry with a Hologic QDR-1000 instrument (Hologic, Waltham, MA, USA).

Nine boys in the untreated group, ten in the placebo group, and 11 in the letrozole group completed the 2-month follow-up; eight, 11, and 11 boys, respectively, the 5-month follow-up; eight, 11, and ten, the 12-month follow-up; and seven, ten, and ten the 18-month follow-up. One boy in the letrozole group was judged non-compliant and his results were excluded from analyses (figure 1).

We took venous blood samples between 07.30 h and 10.15 h. Serum 17b-oestradiol concentrations were measured by a modified radioimmunoassay by use of coated-tube technology (Spectria oestradiol, Orion Diagnostica, Espoo, Finland) after diethyl ether extraction (700 mL serum and 5 mL diethyl ether).(20) The detection limit of the assay was 6 pmol/L. We measured serum luteinising hormone (LH) and follicle-stimulating hormone (FSH) concentrations by ultrasensitive immunofluorometric assay (Wallac, Turku, Finland),(21) and serum testosterone and 5a-dihydrotestosterone concentrations by radioimmunoassay after separation of steroid fractions on a Lipidex-5000 microcolumn (Packard-Becker, BV Chemical Operations, Groningen, Netherlands).(22) We measured serum inhibin-B concentrations by ELISA (Serotec, Oxford, UK), and serum insulin-like growth factor I and insulin-like growth factor-binding protein 3 concentrations by radioimmunoassay (DiaSorin, Stillwater, MN, USA, and Nichols Institute Diagnostics, San Juan Capistrano, CA, USA, respectively).

Statistical analysis

Values are shown as mean (SD) apart from the values for Tanner stages, which are presented as median (range). Analyses were done with the SPSS statistical software for Windows, Release 8.0.2 (SPSS, Chicago, IL, USA). For analysis of serial measurements, summary measures--ie, differences from the start--were calculated for all boys, and these values were treated as raw data for an appropriate statistical analysis. The differences in the summary measures were only compared between treatment groups. One-way analysis of variance, Student's unpaired t test, Kruskal-Wallis non-parametric analysis of variance, or Mann-Whitney U tests were used as appropriate. Student's paired t test or the Wilcoxon matched pairs signed-rank sum test were used for the analyses of changes within groups during follow-up. Pearson correlation coefficient was used to investigate the relation between growth velocity and hormone concentration. All statistical tests were two-sided. Primary outcome measures were height velocity, bone age, and predicted adult height. A p value of less than 0.05 was judged significant.

Results

There were no differences between the groups in chronological age, height, bone age, predicted adult height, pubertal stage, testis volume, bone mineral density, and mean concentrations of serum 17b-oestradiol, LH, FSH, testosterone, 5a-dihydrotestosterone, inhibin B, insulin-like growth factor I, and insulin-like growth factor-binding protein 3 at the start of the study (table 1).

Characteristic No treatment Testosterone and
(n=10) placebo (n=12)

Chronological age (years) 15.0 (0.7) 15.0 (0.8)
Height (cm) 154.3 (4.4) 151.9 (8.3)
Bone age (years) 12.7 (0.8) 12.6 (1.4)
Predicted adult height (cm) 178.3 (4.3) 174.9 (8.3)
Pubertal stage (G;P)[*] 2 (2-3); 2 (1-2) 2 (2-3); 1 (1-2)
Testis volume (mL) 5.9 (2.7) 6.9 (4.3)
Serum 17b-oestradiol (pmol/L) 15.3 (9.7) 16.4 (10.2)
Serum LH (IU/L) 2.0 (1.3) 1.9 (1.2)
Serum FSH (IU/L) 2.7 (1.0) 2.7 (1.8)
Serum testosterone (nmol/L) 10.3 (10.6) 11.9 (10.0)
Serum 5a-dihydrotestosterone 1.2 (0.5) 1.1 (0.9)
(nmol/L)
Serum inhibin B (ng/L) 153.7 (38.3) 176.1 (43.3)
Serum IGF-I (nmol/L) 27.4 (12.1) 28.3 (9.4)
Serum IGFBP3 (mg/L) 3.7 (0.8) 3.8 (0.5)

Characteristic Testosterone and
letrozole (n=10)

Chronological age (years) 15.2 (0.8)
Height (cm) 155.3 (6.7)
Bone age (years) 13.1 (0.6)
Predicted adult height (cm) 176.5 (5.5)
Pubertal stage (G;P)[*] 2 (2-3); 1 (1-2)
Testis volume (mL) 5.5 (1.9)
Serum 17b-oestradiol (pmol/L) 12.8 (9.6)
Serum LH (IU/L) 2.5 (1.1)
Serum FSH (IU/L) 3.5 (2.2)
Serum testosterone (nmol/L) 9.5 (11.1)
Serum 5a-dihydrotestosterone 1.0 (0.6)
(nmol/L)
Serum inhibin B (ng/L) 161.2 (51.1)
Serum IGF-I (nmol/L) 30.3 (10.8)
Serum IGFBP3 (mg/L) 3.7 (0.5)

Values are mean (SD), apart from pubertal stage which is median (range).
IGF-I=insulin-like growth factor I; IGFBP3=insulin-like growth
factor-binding protein 3.[*] According to Tanner; G=genital stage;
P=pubic hair stage.(18)

Table 1: Baseline characteristics
In the untreated group, serum 17b-oestradiol concentrations rose by 90%, from 15.3 to 29.0 pmol/L (p=0.008) in 12 months, whereas in the testosterone with placebo group they rose by 130% from 16.4 to 37.9 pmol/L (p=0.02) in 5 months (figure 2). By contrast, in the testosterone with letrozole group, no such increase was seen during treatment for 12 months. 17b-oestradiol concentrations were higher in the placebo group than in the letrozole group from 2 to 12 months. After discontinuation of letrozole, 17b-oestradiol concentration increased in the letrozole group, and at 18 months the concentrations in all three groups were very similar.

[GRAPH OMITTED]

In the untreated group, bone age increased 1.1 (SD 0.8) years in 18 months (figure 3). In the placebo group, the respective change was 1.7 (0.9) years, and in the letrozole group it was only 0.9 (0.6) years. In the treatment groups, boys on testosterone and placebo grew slightly faster during the first 5 months than those given testosterone and letrozole (table 2). After 5 months, boys in both treatment groups grew with similar velocity. No correlation was seen between height velocity during the first 5 months and serum insulin-like growth factor I, insulin-like growth factor-binding protein 3, 17b-oestradiol, testosterone, or 5a-dihydrotestosterone concentration at 5 months. The upper part/lower part body ratio, measured by sitting height as a proportion of total height, remained unchanged in all groups.

[GRAPH OMITTED]

No treatment Testosterone
and placebo

Growth velocity (cm/year)
0-5 months 7.3 (2.4) (8) 9.9 (1.7) (11)
5-12 months 6.6 (2.8) (8) 7.9 (2.5) (11)
12-18 months 6.5 (3.4) (7) 6.5 (1.6) (10)

Predicted adult height (cm)
0 months 178.3 (4.3) (10) 174.9 (8.3) (12)
12 months 176.8 (3.3) (8) 174.4 (8.0) (11)
18 months 180.3 (4.9) (7) 175.2 (6.6) (10)

Testis volume (mL)
0 months 5.9 (2.7) (10) 6.9 (4.3) (12)
5 months 8.1 (3.9) (8) 8.9 (4.9) (11)
12 months 12.0 (5.2) (8) 13.4 (5.6) (11)
18 months 14.7 (6.1) (7) 18.6 (6.4) (10)

Pubertal stage--G;P[*]
0 months 2 (2-3); 2 (1-2) 2 (2-3); 1 (1-2)
(10) (12)
5 months 2 (2-4); 2 (2-3) 3 (2-4); 2 (1-4)
(8) (11)
12 months 3 (2-5); 2 (2-3) 4 (2-5); 3 (2-4)
(8) (11)
18 months 4 (2-5); 3 (2-5) 4 (2-5); 4 (2-5)
(7) (10)

Testosterone p
and letrozole

Growth velocity (cm/year)
0-5 months 7.3 (2.8) (10) 0.02
5-12 months 7.1 (2.3) (9) 0.50
12-18 months 8.3 (2.3) (9) 0.06

Predicted adult height (cm)
0 months 176.5 (5.5) (10) . .
12 months 178.0 (6.2) (9) 0.40
18 months 182.1 (6.4) (9) 0.04

Testis volume (mL)
0 months 5.5 (1.9) (10) . .
5 months 11.5 (5.1) (10) 0.01
12 months 16.8 (3.7) (9) 0.0005
18 months 19.1 (3.3) (9) 0.20

Pubertal stage--G;P[*]
0 months 2 (2-3); 1 (1-2) . .
(10)
5 months 3 (2-4); 2 (1-4) . .
(10)
12 months 4 (3-5); 3 (2-4) . .
(9)
18 months 4 (3-5); 4 (2-5) . .
(9)

Values are mean (SD), except for pubertal stage which is median (range).
Numbers in brackets show the number of boys at each time-point.[*] According to Tanner; G=genital stage; P=pubic hair stage.(18)
Growth velocity: p value refers to difference between treatment groups.
Predicted adult height and testis volume: p value refers to difference
between treatment groups regarding changes in value from the start to
the time-point indicated by p value.

Table 2: Growth velocity, predicted adult height, testis volume,
and pubertal stage during follow-up
The predicted adult heights in the untreated group (p=0.4) and in the placebo group (p=0.8) did not change significantly within 18 months (table 2; figure 4), whereas, in the letrozole group, an increase of 5.1 cm (p=0.004) in predicted adult height was seen (in one patient, the predicted adult height fell by 3.5 cm, the increases in the other boys ranged from 2.5 cm to 8.8 cm). Change in predicted adult height between treatment groups differed significantly (p=0.04).

[GRAPH OMITTED]

Puberty advanced in all three groups during follow-up (table 2). The increase in testis volume was greater in the letrozole group than in the placebo group in the first 5 months and at 12 months follow-up. The clinical stages of puberty progressed in a similar way in the two treatment groups. Gynaecomastia was seen in none of the boys in the untreated group, in two in the placebo group, and in two in the letrozole group.

Bone density of lumbar vertebrae and of femoral neck did not change significantly in the untreated group during follow-up. Mineral density of the lumbar vertebrae had increased in both treatment groups at 12 and 18 months; the magnitude of the increases was similar. Bone density of the femoral neck increased within 12 and 18 months in the placebo group, but did not change significantly in the letrozole group. Changes in femoral-neck bone density did not differ significantly between treatment groups.

The gonadotropin concentrations in the untreated group rose during follow-up (figure 5). The pattern of change in gonadotropin concentration in the two treatment groups was divergent; in the placebo group, mean LH concentration fell from 1.9 (1.2) to 0.31 (0.4) IU/L at 2 months, whereas, in the letrozole group, mean concentration rose from 2.5 (1.1) to 7.8 (7.7) IU/L. At 12 months, LH concentration had increased to the pretreatment value in the placebo group, whereas it remained raised in the letrozole group. After discontinuation of all treatments, LH concentrations were similar in all three groups. FSH concentrations changed in parallel with those of LH (figure 5).

[GRAPH OMITTED]

Testosterone concentration in the untreated group increased during follow-up (figure 5). In the two treatment groups, values were already increased in the first 2 months, but the magnitude of the changes differed significantly (p=0.003); in the placebo group there was a rise of 33% from 11.9 (10.0) to 15.8 (9.7) nmol/L (p=0.02), whereas in the letrozole group the respective increase was 420%, from 9.5 (11.1) to 49.1 (34.9) nmol/L (p=0.001). In the letrozole group, the high concentration was sustained until discontinuation of letrozole, after which it fell to a value closely similar to that of the other groups. During follow-up, 5a-dihydrotestosterone concentrations had similar patterns of changes to testosterone concentrations, although the magnitude of the changes was smaller.

Inhibin-B concentration in the untreated group increased during follow-up (figure 5). In the placebo group, the concentration did not change significantly. By contrast, in the letrozole group, the concentrations increased in 12 months, but after discontinuation of the letrozole treatment, did not differ from pretreatment concentration. Consistent with the difference in increases in testis volume and the divergent pattern of change in gonadotropin concentration, the changes in inhibin-B concentration in the two treatment groups were different in the first 5 months (p=0.01).

In the untreated group, serum insulin-like growth factor I concentration did not change and insulin-like growth factor-binding protein 3 concentration increased during follow-up (table 3). In the placebo group, increases in both insulin-like growth factor I and insulin-like growth factor-binding protein 3 concentrations were seen. In the letrozole group, insulin-like growth factor I concentration did not change during follow-up, and insulin-like growth factor-binding protein 3 concentration increased only after discontinuation of letrozole.

No treatment Testosterone
and placebo

Serum IGF-I (nmol/L)
0 months 27.4 (12.1) (10) 28.3 (9.4) (12)
2 months 28.7 (7.8) (7) 34.0 (7.7) (10)[*]
5 months 25.9 (5.8) (8) 34.5 (7.7) (11)[*]
12 months 29.3 (9.3) (8) 34.3 (9.5) (11)[*]
18 months 27.9 (6.8) (7) 31.9 (8.4) (10)

Serum IGFBP3 (mg/L)
0 months 3.7 (0.8) (10) 3.8 (0.5) (12)
2 months 3.7 (0.7) (8) 4.1 (0.5) (9)[*]
5 months 3.8 (0.6) (8) 4.3 (0.6) (11)[]
12 months 3.9 (0.7) (8) 4.3 (0.5) (11)[c]
18 months 4.5 (0.5) (7)[*] 4.7 (0.6) (10)[]

Testosterone p
and letrozole

Serum IGF-I (nmol/L)
0 months 30.3 (10.8) (10) ..
2 months 25.6 (4.3) (8) 0.01
5 months 25.2 (5.2) (10) 0.01
12 months 27.4 (2.8) (8) 0.06
18 months 34.1 (3.7) (9) 0.90

Serum IGFBP3 (mg/L)
0 months 3.7 (0.5) (10) ..
2 months 3.6 (0.7) (10) 0.02
5 months 3.4 (0.6) (10) 0.0004
12 months 3.5 (0.6) (9) 0.008
18 months 4.4 (0.5) (9)[] 0.80

Values are mean (SD). Numbers in brackets show the number of boys at each
time-point. IGF-I=insulin-like growth factor I; IGF BP3=insulin-like
growth factor-binding protein 3. Change within group from the start
to indicated time-point:[*] p [less than] 0.05; [] p [less than] 0.001;
[c] p [less than] 0.01. p value refers to difference between treatment
groups regarding changes in concentration from start to time-point
indicated by p value.

Table 3: Serum IGF-I and IGFBP3 concentrations during follow-up
Discussion

letrozole is an effective inhibitor of oestrogen synthesis in boys of pubertal age, as previously seen in men.(23) Moreover, treatment is well tolerated. Potentially deleterious effects on bone density have to be taken into account when manipulating human growth by inhibition of oestrogen action. We saw only minor differences in bone density between treatment groups, and in the letrozole group bone density increased more than in the untreated group.

Consistent with our original hypothesis, we saw that inhibition of oestrogen synthesis delayed bone maturation. It is noteworthy that, in the boys treated with testosterone and letrozole, bone maturation was slower, despite the considerably higher androgen concentrations, than in the boys treated with testosterone alone. This finding confirms the view that oestrogens are more important than androgens in bone maturation in pubertal boys, and agrees with findings in men who do not have oestrogen action.(5-7) Furthermore, even after discontinuation of all treatments, progression of bone maturation was slower in boys treated with testosterone and letrozole than in the boys treated with testosterone alone, indicating that the effect of treatment outlasts the period of treatment.(24)

Our results lend support to the notion that testosterone accelerates growth via an oestrogen-dependent mechanism. The growth-enhancing effect of oestrogens possibly results from stimulation of growth hormone secretion by oestrogens.(25) Consistently, in our study, insulin-like growth factor I and insulin-like growth factor-binding protein 3 concentrations increased in the placebo group during treatment, but did not change in the letrozole group. Our findings further suggest that other factors, in addition to oestrogens, are implicated in pubertal growth acceleration in boys, because during letrozole treatment most grew with normal pubertal growth velocity despite low 17b-oestradiol concentrations. However, the normal pubertal growth velocity during letrozole treatment could also have resulted from activation of oestrogen receptors. If aromatisation of androgens is inhibited, steroid biosynthesis is directed to produce 5a-dihydrotestosterone and subsequently 3b-androstanediol, which is a weak oestrogen that binds to oestrogen receptors(26) and could, therefore, have oestrogenic effects.

The findings that predicted adult height did not change either in the boys who received no treatment or in those who were treated with testosterone alone, and are consistent with those of previous studies, which have shown that androgen treatment does not increase adult height.(11,13,14) By contrast, the rise in predicted adult height in boys treated with testosterone and letrozole lends support to our primary hypothesis that inhibition of oestrogen synthesis in growing adolescents increases adult height. We suggest that boys with constitutional delay of puberty can achieve an adult height closer to their genetic growth potential if oestrogen actions are inhibited. We need to establish whether treatment with aromatase inhibitors can be used in boys with delayed puberty and genetic short stature, or in patients with various growth disorders resulting in short stature--eg, boys with precocious puberty or congenital adrenal hyperplasia with greatly advanced bone age. For some boys with constitutional delay of puberty, the aromatase inhibitors alone, without exogenous testosterone, might be sufficient treatment.

Our findings accord with the view that oestrogens regulate LH secretion negatively in early and mid-pubertal boys.(27) Androgens regulate LH secretion negatively in late pubertal boys(28) and in men.(29) Our results suggest that androgens have a minor role compared with oestrogens in regulating LH and FSH secretion in early and mid-pubertal boys. Our results further suggest that the negative feedback regulation between FSH and oestrogens, seen in men,(23,30) is already operative in early-pubertal and mid-pubertal boys. The increase in FSH concentrations during letrozole treatment is probably not due to a diminished negative feedback signal from inhibin B, for inhibin-B concentrations increased concomitantly with FSH concentrations during letrozole treatment. Our finding, that inhibin-B concentration was unchanged concomitantly with the profoundly suppressed gonadotropin concentrations, confirms the notion that inhibin-B secretion in men is in part independent of gonadotropin secretion.(31)

Neither treatment had an adverse effect on testis size or inhibin-B concentration, suggesting that they did not adversely affect maturing spermatogenesis. Furthermore, the two treatments similarly advanced the appearance of secondary sexual characteristics, despite the considerably higher androgen concentrations in the boys treated with testosterone and letrozole.

Mean testosterone concentrations at the start of follow-up were about 10 nmol/L, which indicates that some boys were already at early or mid puberty. Blood samples were drawn between 7.30 h and 10.15 h, thus these concentrations do not reflect the mean diurnal testosterone concentrations, which are much lower.(32,33) Testosterone secretion in pubertal boys has a clear diurnal rhythm with a peak early in the morning, the concentrations being much lower in the afternoon and evening.(32,33) In some boys, puberty might have advanced spontaneously in a short time without any treatment. However, we believe that it was justified to treat these boys since none had had pubertal increase in growth velocity before treatment and all treated boys wanted medical intervention.

Our findings suggest that an increase in adult height can be attained in growing adolescents by inhibition of oestrogen action. This result provides a rationale for studies aimed at delaying maturation of growth plates and increasing adult height in several growth disorders.

Contributors

Leo Dunkel created the study design. Sanna Wickman and Leo Dunkel designed and ran the study. Sanna Wickman, Ilkka Sipila, and Leo Dunkel identified bone ages. Carina Ankarberg-Lindgren and Ensio Norjavaara were responsible for 17b-oestradiol measurements. Statistical analyses were done by Sanna Wickman, supervised by Leo Dunkel. All investigators contributed to writing of the report.

Acknowledgments

This study was supported by the Foundation for Paediatric Research, Helsinki, Finland.

Hospital for Children and Adolescents, University of Helsinki, Ph 281, FIN-00029 Hus, Finland (S Wickman MD, I Sipila MD, L Dunkel MD); and Goteborg Paediatric Growth Research Centre, Institute for the Health of Women and Children, Goteborg University, Goteborg, Sweden (C Ankarberg-Lindgren BSc, E Norjavaara MD)