makes me wish i was still a teen.
Inhibition of Estrogen Biosynthesis with a Potent Aromatase Inhibitor Increases Predicted Adult Height in Boys with Idiopathic Short Stature: A Randomized Controlled Trial
Matti Hero, Ensio Norjavaara and Leo Dunkel
Hospital for Children and Adolescents, University of Helsinki and Helsinki University Central Hospital (M.H.), Helsinki 00029 HUS, Finland; Goteborg Pediatric Growth Research Center, Institute for the Health of Women and Children, Goteborg University (E.N.), Goteborg S-41685, Sweden; and Department of Pediatrics, Kuopio University Hospital (L.D.), Kuopio 70211, Finland
Address all correspondence and requests for reprints to: Dr. Leo Dunkel, Department of Pediatrics, Kuopio University Hospital, P.O. Box 1777, 70211 Kuopio, Finland. E-mail:
leo.dunkel@kuh.fi.
Context: In males as well as in females, estrogen is an essential regulator of bone maturation, growth plate fusion, and cessation of longitudinal growth. Therefore, an increase in predicted adult height (PAH) may be achieved in short boys by blocking estrogen biosynthesis.
Objective: We tested the hypothesis that a decrease in the rate of bone maturation and an increase in PAH can be achieved in boys with idiopathic short stature (ISS) by the method of blocking estrogen biosynthesis with an aromatase inhibitor. Secondarily, we investigated the effects of aromatase inhibition on bone mineralization.
Design: This was a prospective, double-blind, randomized, placebo (Pl)-controlled clinical study.
Setting: The study was performed at a university hospital out-patient clinic.
Patients: Thirty-one boys, aged 9.0–14.5 yr, with ISS were studied.
Intervention: The boys were treated with the aromatase inhibitor letrozole (Lz; 2.5 mg/d) or Pl for 2 yr.
Main Outcome Measure: The main outcome measure was the change in PAH after 24 months of treatment.
Results: PAH increased by 5.9 cm (P < 0.0001), and height SD score for bone age increased by 0.7 SD score (P < 0.0001) in the Lz-treated boys, whereas no changes occurred in the respective measures in Pl-treated boys. Areal bone mineral density of the lumbar spine and femoral neck, assessed by dual-energy x-ray absorptiometry, increased in a similar fashion in both groups during the treatment, whereas bone mineral apparent density increased only in those taking Lz (median increase, 4.3%; P = 0.009).
Conclusions: Treatment with the aromatase inhibitor Lz delays bone maturation and improves PAH in boys with ISS. No adverse effects on bone mineralization were evident after 2 yr of treatment.
THE ROLE OF estrogen in the regulation of growth has been clarified considerably during the past decade. Case reports of patients with estrogen insensitivity (1) or estrogen deficiency (2, 3, 4) have substantiated the fact that in the absence of an estrogen effect on the growth plate, epiphyses remain open, and longitudinal growth continues for an exceptionally long period of time. On the basis of these observations, it has become possible to postulate that longi- tudinal growth can be modulated by blocking estrogen biosynthesis with aromatase inhibitors, which inhibit aromatization of C19 androgens [mainly androstenedione and testosterone (T)] to C18 estrogens.
The results of studies on the relatively weak aromatase inhibitor testolactone in treatment of congenital adrenal hyperplasia (5), familial male-limited precocious puberty (6), and McCune-Albright syndrome (7) have indeed suggested that aromatase inhibitors may be effective in delaying bone maturation and increasing predicted adult height (PAH), at least in children with peripheral precocious puberty. However, these studies were not controlled (6), or data regarding the changes in PAH were inconclusive (5, 7). Third generation aromatase inhibitors now available offer significant advantages over testolactone in terms of potency, safety, and tolerability. To date, only one placebo (Pl)-controlled study has evaluated the efficacy of a modern aromatase inhibitor in delaying bone maturation and increasing PAH. In that study, 1-yr treatment with letrozole (Lz) in combination with T in boys with delayed puberty significantly delayed bone maturation and increased PAH by 5.1 cm (.
Because the influence of inhibition of estrogen biosynthesis on bone maturation and PAH in boys with growth failure has remained obscure, we performed a randomized, double-blind, Pl-controlled trial in which boys with idiopathic short stature (ISS) were treated with Lz or Pl for 2 yr. We tested the hypothesis that blocking estrogen biosynthesis with an aromatase inhibitor would delay bone maturation and consequently improve PAH, and this could eventually result in greater adult height. Secondarily, we investigated the effects of aromatase inhibition on gonadotropin secretion and bone mineralization in boys during prepuberty and during the transition from peripuberty to puberty.
Subjects
The study population comprised short boys examined and followed up at the out-patient clinic for pediatric endocrinology at the Hospital for Children and Adolescents, University of Helsinki (Helsinki, Finland); the boys had no signs of underlying disease accounting for the short stature. They were identified through a systematic review of growth charts and medical records. When reassessed by the researchers, boys with no signs of chronic or endocrine illness in medical history, clinical examination, and routine laboratory tests were considered to have ISS and were eligible for recruitment. Inclusion criteria were calendar age of 9.0–14.5 yr and height at least 2 SD below the mean for age or at least 2 SD below midparental target height. Those with a bone age of more than 14 yr were excluded. GH deficiency was ruled out by a GH stimulation test if this was suspected on the basis of slow growth velocity or subnormal serum IGF-I or IGF-binding protein-3 concentrations. Being small for gestational age at birth was not considered a criterion of exclusion. Four boys taking Lz and three taking Pl were small for gestational age at birth.
Between May 2001 and May 2002, 40 boys were assessed for eligibility. After initial assessment and provision of information, 31 boys with ISS were enrolled. Before initiation of treatment, written informed consent was obtained from the boy and his guardian(s). One boy in the Lz group and one in the Pl group were receiving continuous low-dose inhaled corticosteroid treatment, and three Lz-treated boys and one Pl-treated boy were receiving seasonal inhaled corticosteroid treatment for asthma. Apart from these treatments, none of the boys had any treatment known to affect growth or bone maturation. After 6 months of treatment, one boy taking Pl was diagnosed with diabetes mellitus and excluded from the study.
Study protocol
The boys were randomized in a double-blind manner to receive either Lz (Femar, Novartis AG, Stein, Switzerland; 2.5 mg, orally, once daily) or Pl (orally, once daily) for 2 yr. Randomization was carried out in blocks of 10 at the hospital pharmacy by a computer-generated randomization list. Researchers and subjects were blind to treatment assignment until the end of follow-up, when the randomization code was revealed to the researchers after the data had been entered into the computer.
The boys were examined at entry and every 6 months thereafter for 2 yr. Follow-up visits included physical examinations, all performed by the same physician (M.H.) throughout the follow-up period. Stage of puberty was evaluated according to Tanner (9), and testis volumes were calculated by the formula: length x width2x 0.52 (10). Height was measured on a Harpenden stadiometer with 0.1-cm precision.
The study protocol was approved by the ethics committee of the Hospital for Children and Adolescents and by the National Agency for Medicines.
Radiological measurements
The bone mineral densities (BMD) of the lumbar spine and femoral neck were assessed every 6 months by dual-energy x-ray absorptiometry with a Hologic QDR 4500W device (Hologic, Waltham, MA). Areal BMD (grams per meter squared) measured by dual-energy x-ray absorptiometry normalizes bone mineral content values for the projected area, but does not account for differences in bone depth in the region measured. Areal BMD values are thus confounded by differences in bone size. To correct for this, bone mineral apparent densities (BMAD; grams per meter cubed) of the lumbar spine were also calculated by the formula: bone mineral content/(area projected)1.5 (11).
Bone ages were determined by Greulich and Pyle’s method (12). Adult height predictions were calculated by the Bayley-Pinneau method (13), a method producing a percentage of height attained relative to final height for bone ages of 6.0–18.5 yr. Two boys had a bone age less than 6 yr at the start of the study, and their adult height predictions were calculated by extrapolating data from Bailey and Pinneau’s tables.
Biochemical measurements
Venous blood samples were drawn between 0730 and 1000 h. Concentrations of serum FSH, LH, IGF-I, and inhibin B were determined directly after follow-up visits. Serum T and estradiol concentrations were measured in sera stored at –20 C until required. Serum T concentrations were determined by a modified RIA as previously described (14). Serum estradiol concentrations were quantified with a modified RIA (Spectria E2, Orion Diagnostica, Espoo, Finland) with a detection limit of 1.2 pg/ml (4.5 pmol/liter) as previously described (15). Interassay coefficients of variation (CVs) were 40% and 12% at 1.1 and 6.0 pg/ml (4.2 and 22 pmol/liter), respectively, and less than 16% above 6.0 pg/ml (22 pmol/liter). The intraassay CVs were less than 14% at 1.3–35.4 pg/ml (4.6–130 pmol/liter). Serum FSH and LH levels were measured by ultrasensitive immunofluorometric assays (Wallac, Turku, Finland) as previously described (16). Serum IGF-I concentrations were determined by RIA (DiaSorin, Stillwater, MN). Interassay CVs were less than 16% at 69–252 µg/liter (9–33 nmol/liter). Serum inhibin B levels were measured using a commercially available immunoenzymometric assay (Serotec, Oxford, UK) with a detection limit of 15.6 ng/liter. Interassay CVs at concentrations of 229, 85, and 42 ng/liter were 6.9%, 10.1%, and 12.0%, respectively. The intraassay CV was less than 5%.
Statistical analysis
The primary end point of the study was the efficacy of Lz in improving PAH, as measured by change in PAH after 24 months of treatment. Second, we explored the role of estrogen in regulation of the hypothalamopituitary-testicular axis and of bone mineralization in males during prepuberty and during the peripuberty to puberty transition.
For calculation of required sample size, the smallest clinically significant treatment effect on PAH was set at 5 cm. The SD of the change in PAH was set at 4.7 cm, based on the results of our previous study (. With a significance level of 5% and a power of 80%, the smallest appropriate sample size was 28 subjects. Assuming a maximum drop-out rate of 10%, we chose to recruit at least 31 boys.
In a previous study we showed that in pubertal boys, Lz treatment enhances gonadotropin secretion by decreasing the level of estrogen-mediated negative feedback. Consequently, Lz expedites testicular growth and enhances T secretion (16). Based on these effects of Lz after the start of puberty, for analysis of pituitary-testicular function the boys were subdivided into two groups: those who remained prepubertal (testis volume, 2 ml) throughout the follow-up period and those who entered puberty (testis volume, >2 ml) within 18 months after the onset of treatment. According to these criteria, seven of 16 and six of 14 boys remained prepubertal, and nine of 16 and eight of 14 progressed in puberty in the Lz and Pl groups, respectively.
Two-sided tests of hypotheses were used, and P < 0.05 was considered significant. Values, unless otherwise stated, are expressed as the mean ± SD. Analyses were conducted with SPSS statistical software for Windows (release 10.0, SPSS, Inc., Chicago, IL). Differences in baseline characteristics were evaluated by unpaired t test. Within-group changes and between-group differences in bone age progression, height SD score (SDS) for bone age, and PAH were analyzed by paired and unpaired t tests, respectively. Within-group changes in serum FSH, LH, T, estradiol, inhibin B, and IGF-I concentrations; testis volume; and BMD were assessed by paired t tests. Between-group differences in T, estradiol, inhibin B, and IGF-I were analyzed by repeated measures ANOVA, using treatment (Lz/Pl) as a between-subjects factor. Logarithmic transformation was applied when appropriate. Between-group differences in testis volume and between-group differences in changes in FSH and LH in pubertal boys were analyzed with the unpaired t test. Within-group changes in Tanner genital (G) and pubic hair (P) stages were evaluated by the Friedman test, followed by multiple comparison tests. Between-group differences in Tanner G and P stages were evaluated by the Mann-Whitney U test.
Baseline characteristics
At baseline, no significant differences appeared in age, height, weight, bone age, height for bone age, pubertal stage, testis volume, or midparental target height between the groups (Table 1). No boy had had a pubertal increase in growth velocity before the start of treatment.
View this table:
[in this window]
[in a new window]
TABLE 1. Baseline characteristics of study subjects
Auxology
Lz- and Pl-treated boys grew at a similar velocity during the study (Fig. 1A). During the second year of treatment, those Lz-treated boys who had entered puberty by the 12 month point (n = grew faster than the Lz-treated boys still prepubertal at that point (n = 8; 6.7 vs. 4.5 cm/yr; P = 0.04). Correspondingly, those Pl-treated boys who were pubertal at the 12 month point (n = 5) grew faster than the prepubertal Pl-treated boys (n = 9) during the second year of follow-up (7.4 vs. 4.6 cm/yr; P = 0.004). Growth velocities did not differ between pubertal Lz- and Pl-treated boys (6.7 vs. 7.4 cm/yr, respectively; P = 0.5 or between the prepubertal Lz- and Pl-treated boys (4.5 vs. 4.6 cm/yr) during the second year of treatment.
View larger version (21K):
[in this window]
[in a new window]
FIG. 1. Growth velocity (A), rate of bone age progression (BA/CA; B), height SDS for bone age (C), and PAH (D) during the study period. Values are the mean ± SEM. *, P < 0.05; , P < 0.01 (differences between the treatment groups). CA, Calendar age.
In Lz-treated boys, bone age progressed by only 1.24 yr during the 2-yr treatment, whereas in the Pl-treated boys, bone age progressed by 2.05 yr during the same period (change in bone age/change in calendar age, 0.62 vs. 1.02; P = 0.04; Fig 1B). Consequently, height SDS for bone age increased by 0.7 SDS in the Lz group, with no change in the Pl group (Fig. 1C). In a similar fashion, PAH increased by 5.9 cm (P < 0.0001) in Lz-treated boys, but did not change in those treated with Pl during the study. Although PAHs did not differ between groups at the start of the study, at the end of the treatment period, the PAH of the Lz-treated group was higher than the respective measure for the Pl-treated boys (172.8 vs. 166.9 cm; P = 0.03; Fig. 1D).
Lz treatment appeared effective regardless of bone age, because bone age at the start of the study did not correlate in Lz-treated boys with the change in height SDS for bone age (r = 0.05; P = 0.85) or with the change in PAH (r = –0.06; P = 0.83). Moreover, the increases in PAH during the 2-yr treatment were similar in prepubertal and pubertal Lz-treated boys (7.2 vs. 4.8 cm, respectively; P = 0.17).
Boys taking Pl and those taking Lz gained weight in a similar fashion during treatment (3.7 vs. 4.2 kg/yr; P = 0.37), with no differences in body mass index between the groups (data not shown).
Pubertal maturation and hormonal changes
During the study, similar proportions of boys in the Lz and Pl groups entered puberty. At the start, 13 of 16 (81%) Lz-treated and 13 of 14 (93%) Pl-treated boys were prepubertal. Respectively, after 2 yr of treatment, seven of 16 (44%) and six of 14 (43%) remained prepubertal. In those who entered puberty within 18 months after the study started, no significant differences in Tanner G stage appeared between groups (Table 2). Lz-treated pubertal boys, however, had a higher Tanner P stage 12 and 18 months after the start of treatment (Table 2). Their testis volumes increased more rapidly, which was already evident 6 months after the start of treatment (Table 2). In fact, all except one of the pubertal Lz-treated boys had a testis volume of 10 ml or more after 2 yr of treatment, and all Pl-treated boys had testis volumes under 10 ml.
View this table:
[in this window]
[in a new window]
TABLE 2. Progression of puberty in the letrozole- and placebo-treated boys entering puberty within 18 months after start of treatment
In boys who remained prepubertal throughout the study, no differences in concentrations of FSH, LH, T, estradiol, IGF-I, or inhibin B appeared between treatment groups (data not shown). Simultaneously, the respective values in the pubertal boys changed differentially in the Lz and Pl groups, except for those of inhibin B (Fig. 2). In the Pl-treated boys entering puberty within the first 18 months of treatment, the serum LH concentration gradually increased from 0.6 to 2.4 IU/liter (P = 0.0002). In contrast, in the respective group of boys taking Lz, the concentration of LH increased more rapidly, from a baseline value of 1.0 to 4.5 IU/liter (P = 0.002) as early as 6 months after the start of treatment; the LH level remained higher than that in Pl-treated pubertal boys during the treatment period (Fig. 2). In a similar fashion, concentrations of FSH were higher in Lz-treated than in Pl-treated pubertal boys (Fig. 2). Serum T concentrations increased rapidly in the former and exceeded the concentrations in the latter after the start of treatment (Fig. 2). After 24 months of treatment, the mean T concentration in the Lz-treated pubertal boys was 892 ng/dl (range, 17.3–1385 ng/dl), 30.9 nmol/liter (range, 0.6–48.0 nmol/liter). In contrast, serum estradiol concentrations remained at pretreatment level in the Lz-treated pubertal boys throughout the treatment period, whereas the respective concentrations in the Pl-treated boys showed a trend toward an increase 18 months after the start of treatment (Fig. 2).
View larger version (23K):
[in this window]
[in a new window]
FIG. 2. Serum (S-) concentrations of FSH, LH, T, estradiol, inhibin B, and IGF-I in Lz-treated (n = 9) and Pl-treated (n = boys entering puberty within 18 months after the onset of treatment. Values are the mean ± SEM. *, P < 0.05; , P < 0.01 (differences between the treatment groups). To convert to Systeme International units: T (ng/dl) x 0.0347 = nmol/liter; estradiol (pg/ml) x 3.67 = pmol/liter; IGF-I (µg/liter) x 0.131 = nmol/liter.
Serum IGF-I concentrations changed differentially in the Lz- and Pl-treated pubertal boys during the treatment (Fig. 2). This resulted in higher IGF-I concentrations in the pubertal Pl-treated boys at 18 (P < 0.001) and 24 (P < 0.05) months after the start of treatment.
Bone mineralization
Areal BMD in the lumbar spine and femoral neck increased similarly during the study in both groups (Fig. 3, A and B). Simultaneously, BMAD of the lumbar spine increased in those taking Lz (median increase, 4.3%; P = 0.009), but not in those taking Pl (median increase, 0.5%; P = 0.21; Fig. 3C). In both treatment groups, pubertal boys gained more BMD in the lumbar spine and femoral neck than prepubertal boys (P < 0.01–0.05). Areal BMD in the lumbar spine increased by 0.07 and 0.06 g/cm2, and areal BMD in the femoral neck increased by 0.07 and 0.05 g/cm2 in pubertal Lz- and Pl-treated boys, respectively, whereas the former increased by only 0.03 g/cm2 (in both groups) and the latter by 0.01 and 0.03 g/cm2 in prepubertal Lz- and Pl-treated boys.
View larger version (14K):
[in this window]
[in a new window]
FIG. 3. Changes in BMD in lumbar spine (A) and femoral neck (B) and in BMAD of the lumbar spine (C) in short boys treated with Lz or Pl. Boxes represent interquartile ranges (50% of values) with medians; whiskers represent the total ranges of values, excluding outliers (circles). *, P < 0.05; , P < 0.01; , P < 0.001 (changes from the baseline value).
In a test of the hypothesis that suppression of estrogen biosynthesis delays bone maturation and consequently increases PAH in short boys, we compared the effect on PAH of the aromatase inhibitor Lz with that of Pl in adolescent boys with ISS. We found that 2-yr treatment with Lz effectively delayed bone maturation and, consistent with our hypothesis, improved PAH as much as 5.9 cm. Our findings on the crucial role of estrogens in bone maturation in males are in agreement with findings in men with estrogen insensitivity (1) and estrogen deficiency (2, 3, 4, 17) and with our previous observations of Lz treatment in boys with delayed puberty (. In contrast with our findings, a recent study of 12-month treatment with the aromatase inhibitor anastrozole in adolescent boys with GH deficiency failed to find an effect on PAH (1. This is potentially explained by the shorter duration of treatment in the study by Mauras et al. (1 and the somewhat lower potency of anastrozole compared with Lz (19).
It is of note that treatment with Lz delayed bone maturation and improved PAH similarly in prepubertal and pubertal boys regardless of bone age, implying that estrogen plays a role in the maturation process of the growth plate in childhood. This is an interesting finding, because before puberty in boys, aromatase activity is low in the testis (20) and probably also in extraglandular tissues (21); hence, the effects of circulating, intracrine, and paracrine estrogen on growth and bone maturation have been assumed to become important only after the start of puberty in males (22). The circulating level of estradiol, however, was not significantly lower in prepubertal Lz-treated boys than in prepubertal Pl-treated boys, possibly the result of the limited accuracy of the assay at prepubertal low concentrations of estradiol. Using this sensitive RIA for estradiol in prepubertal boys, we observed somewhat higher serum estradiol concentrations than previously reported with the same method (23). The reason for this difference is unknown, and it may have masked possible between-group (and longitudinal) differences in true concentrations. However, because several tissues are capable of converting androstenedione and T to estrogens, the circulating estradiol concentration may poorly reflect the paracrine or intracrine effects of estrogen in the tissues.
In accord with our previous findings in early and midpubertal boys (16), we observed that suppression of estrogen biosynthesis by Lz decreases the negative feedback control of gonadotropin secretion and raises concentrations of serum FSH and LH after the onset of puberty. This increase in gonadotropin secretion results in a supraphysiological rise in T concentrations and rapid testicular growth. The increase in T level may accelerate the progression of puberty, as suggested by the higher Tanner P stage in Lz-treated boys 12 and 18 months after start of treatment. This treatment did not, however, affect the onset of puberty, because similar proportions of boys in the Lz and Pl groups entered puberty during the 2-yr treatment. This is in agreement with the concept that the control of GnRH secretion before the onset of puberty is mediated via the central nervous system rather than by sex steroids.
In pubertal boys, puberty-associated stimulation of the GH-IGF-I axis, normally mediated by estrogen (24), was inhibited in those treated with Lz, as evidenced by the opposite changes in circulating IGF-I concentrations in the treatment groups. Because an increase in estrogen level and subsequent augmentation of the GH-IGF-I axis are considered to be the principal regulators of the pubertal growth spurt in males (25, 26), it is interesting that the pubertal Pl- and Lz-treated boys grew at similar velocities. The observed growth velocity in the latter strongly suggests that in the presence of low (prepubertal) IGF-I and estradiol concentrations, high androgen concentrations are able to enhance growth velocity. In support of the view that androgens have direct growth-stimulating effects in the growth plate during puberty, androgen receptors have indeed been located in human growth plates (27).
The areal BMD of the lumbar spine and femoral neck increased in a similar fashion in both groups of boys, whereas the BMAD of the lumbar spine increased significantly only in those treated with Lz. These findings suggest that 2-yr treatment with an aromatase inhibitor does not adversely affect bone mineralization in adolescent boys. In adult males, however, estrogen plays a significant role in regulating bone mineralization and, in particular, bone resorption, as indicated by the significantly reduced BMD in men with estrogen insensitivity (1) or estrogen resistance (2, 3, 4) and by the results of direct interventional studies (28, 29). Because T contributes to the maintenance of bone formation in men (2, it can be speculated that the high androgen concentrations in the Lz-treated boys may have compensated for any increase in their bone resorption. Bone turnover, however, was not assessed in our study.
ISS is a heterogeneous entity that includes short children with familial short stature, constitutional delay of growth and puberty, and normal GH secretion with normal or decreased concentrations of serum IGF-I. Therefore, the condition of ISS includes short children with delayed bone age, and so did our study population. Because bone age was similarly delayed in the treatment groups at the start of the study, this most likely did not contribute to the observed differences in bone age progression and PAH between the treatment groups.
In conclusion, treatment with the aromatase inhibitor Lz effectively delayed bone maturation and improved PAH in adolescent boys with ISS. Aromatase inhibitors may thus provide a new and effective treatment modality for growth disorders of various etiologies. Lz was well tolerated, and no adverse effects on bone mineralization were observed. However, until longer-term follow-up data on boys treated with an aromatase inhibitor are available, we consider that the use of aromatase inhibitors in children and adolescents should be limited to clinical trials. Moreover, adult height data for patients treated with an aromatase inhibitor during adolescence are needed to determine whether the gain in PAH also leads to greater adult height.
