The function of aromatase inhibitors (AIs) is to block the aromatase enzyme (CYP 19) which promotes the conversion of androgens to estrogens. Due to the resulting decrease in production of estrogen, AIs have been used for many years in the treatment of estrogen-responsive breast cancer. Estrogen also plays an important role in regulating longitudinal growth during puberty by promoting the closure of epiphyses. If the production of estrogen is blocked, the fusion of the epiphyses may be delayed, potentially leading to a prolonged period of longitudinal growth and increased final adult height. This physiological effect has been described in cases with inactivating mutations of the aromatase gene in which young men became very tall with unfused epiphyses even upon reaching adulthood (1).
The AIs are currently being used as an off-label treatment for short stature in adolescent males. According to some studies, the use of AIs can lead to a significant increase in predicted adult height (PAH) as well as an increased height standard deviation score (HtSDS) for bone age (BA) (2–4). However, data available regarding the actual attainment of final height following AI therapy are limited. A recent follow-up study showed that the initial difference in PAH had disappeared 4.2 years post-treatment (5).
Of additional concern is the limited safety profile of AI usage in the pediatric population. Untoward effects previously reported include possible alterations in vertebral bone growth and vertebral anomalies (5), decrease in high-density lipoprotein (4, 6, 7), and increase in hematocrit (8). The limited human data thus far available suggest no adverse effect of AI usage on insulin sensitivity (6, 7). However, a recent animal model study found that female Wistar rats treated with continuous 90-day administration of higher dose letrozole prior to puberty did show signs of insulin resistance including decreased insulin sensitivity and enlarged adipocytes (9). Potential effects on cognition and memory have not been extensively evaluated, though a recent trial suggested that AI therapy did not impact the cognitive performance (10).
In the current study, we performed a chart review of adolescent males who received AI therapy for short stature and/or rapid pubertal progress, to learn more about the efficacy of AIs as well as to document any untoward effects of this treatment.
Study design and subject selection
This retrospective chart review was approved by the institutional review board. All adolescent males followed in the outpatient clinics of the Department of Pediatric Endocrinology, Diabetes, and Metabolism at New York Presbyterian-Morgan Stanley Children’s Hospital who were treated with AI therapy (anastrozole or letrozole) were considered candidates for the chart review, and were required to meet the following inclusion criteria:
short stature [defined as the height at least 2 standard deviation (SD) below average, or at least 2 SD below mid-parental target height (MPTH)] and/or
rapid pubertal progression with a BA ≥13 years.
Due to the retrospective nature of this chart review, we were unable to apply stricter inclusion criteria (such as treatment offered only to those boys with predicted height less than the MPTH). Subjects were excluded if they had a treatment duration of less than 12 months (n=8), if they were receiving adjunctive therapy with growth hormone (n=15), if they had clinically significant scoliosis (due to the difficulty in comparing heights after surgical treatment) (n=2), or if they were receiving or had recently received hormonal medications such as testosterone or Lupron (n=2). One additional subject with the diagnosis of gonadal dysgenesis was also excluded (n=1). Some 55 charts were initially reviewed and 27 among them were selected for analysis based on the aforementioned inclusion and exclusion criteria. All 27 subjects selected had received the same dose of anastrozole, i.e., 1 mg daily (three subjects who had received letrozole were excluded based on the above-mentioned criteria).
Anthropometric, hormonal, and metabolic data from years 2006 to 2012 were collected from charted visits up to 10 months prior to starting AI therapy, during AI therapy and up to 10 months following cessation of AI therapy, when available. All data collected were entered into a secure study database and subsequently de-identifed by the co-authors. Outcome measures were changes in PAH, HtSDS, ratios of BA/chronological age (CA) post-therapy, near-final height, and effects on growth, pubertal development, hormonal and metabolic markers. Bone ages had been read by up to two endocrinologists and an average was taken if there was a discrepancy between the readings. The PAH was calculated from the BAs by the method of Bayley and Pinneau, using the Atlas of Greulich and Pyle (11).
PrePAH was based on the most recent BA prior to the start of therapy (preBA) while postPAH was based on the last available BA following the start of therapy (postBA). Subjects were divided into two groups based on the length of therapy (LOT): Rx group 1 for <18 months (n=7) and Rx group 2 for 18–30 months (n=20). For purposes of analysis, LOT was taken at the time of the postBA, in which most subjects were at or immediately following the actual cessation of therapy. However, in eight subjects LOT was taken 2–5 months prior to cessation of therapy, and in one subject (Rx group 1), it was taken 9 months prior to the actual cessation of therapy.
Continuous variables were summarized with means and standard errors. Treatment group differences were assessed with independent t-tests (Rx group 1 or 2) while within-group pre- to post-changes and peak vs. pre-changes were assessed with two-tailed t-test against the null hypothesis of zero change. All subjects which contributed data were included in all analyses and no adjustment was made for multiple comparisons. A p-value <0.05 was considered statistically significant.
PAH, HtSDS, and bone age progression
The average preBA and postBA values were 13.6±0.1 and 15.3±0.2 years, respectively. The average CA values at the time of preBA and postBA were 13.7±0.1 and 15.4±0.1 years, respectively. The average MPTH was 172.52±0.81 cm. The average LOT was 21 months, but ranged from 14 to 30 months.
The primary outcome measures were assessed for all subjects together, and for Rx group 1 and Rx group 2 separately (Table 1). Overall, there was no significant change in PAH, HtSDS, or the ratio of BA/CA for all subjects and for those in Rx group 1 post-therapy. In Rx group 2, there was a small increase in PAH that approached significance (2.03±1.02 cm, p=0.11), no change in HtSDS, and a small but significant decrease in the ratio of BA/CA post-therapy (0.03±0.01, p=0.02). The ratio of postBA-preBA/postCA-preCA (ΔBA/ΔCA) was 0.86±0.08 for all subjects, 1.14±0.21 for Rx group 1, and 0.76±0.08 for Rx group 2. The post-therapy PAH remained significantly different from MPTH in all subjects (169.42±1.04, p=0.005), in Rx group 1 (166.55±1.57, p=0.008), and in Rx group 2 (170.40±1.22, p=0.013).
We considered near-final height to have been reached in 8 out of the 27 subjects. Among these, 4 subjects had a BA ≥16 years and growth velocity ≤1 cm/year when compared to the height measured at a visit at least 3 months prior. Two subjects had growth velocities >1 cm/year but were included as their BAs were 17 and 18 years. Two subjects had growth velocities ≤1 cm/year with BAs of 15 and 15.5 years obtained 2 and 1.5 years before, respectively. One of the latter two patients was initially classified as part of the <18 months group (Rx group 1) because their post-therapy BA was obtained at 15 months post-treatment. However, we included him as one of the eight patients who had reached near-final height because at 26 months his growth velocity had declined to less than 1 cm/year.
Six out of eight subjects had discontinued AI therapy prior to the near-final height measurement, and the other two subjects completed the therapy within 1 month of the near-final height measurement.
The average LOT for this subgroup (the eight subjects above) was 20.6 months. The average near-final height was 163.96±1.75 cm, which was 6.73±1.40 cm (range: 1.52–13.54 cm) lower than the MPTH of the sample (169.88±2.13, p=0.002). The near-final height was less than or equal to the prePAH in all eight subjects individually, and overall the average near-final height was not statistically different from the prePAH (163.96±1.75 vs. 165.86±1.42 cm, p=0.07).
Growth and pubertal development
The average height pre- and post-therapy values were 152.40±1.02 and 163.14±0.86 cm, respectively (p<0.0001). There was no change in the body mass index standard deviation score pre- and post-therapy (0.028±0.177 vs. 0.020±0.152, p=0.58). There was a significant decrease in the growth velocity post-therapy (7.9±0.6 vs. 3.3±0.3 cm/year, p<0.0001). There was a statistically significant but clinically unremarkable difference between Tanner stage pre- and post-therapy (3.2±0.1 vs. 4.2±0.1, p<0.0001). Testicular volume increased significantly pre- vs. post-therapy (11.8±0.6 vs. 17.7±0.5 mL, p<0.0001).
Hormonal and metabolic parameters
On therapy, testosterone increased to a peak of 802.0±51.6 ng/dL, which was elevated from pre-therapy (349.4±42.0 ng/dL, p<0.0001). Post-therapy testosterone was 457.2±54.5 ng/dL, which was statistically different from pre-therapy (p=0.03).
On therapy, estradiol decreased to a nadir of 0.30±0.06 ng/dL, which was different from the pre-therapy value of 1.14±0.27 ng/dL (p=0.02). Post-therapy estradiol was 1.20±0.18 ng/dL, which was not different from pre-therapy (p=0.78).
Insulin-like growth factor-1
Overall, there was no statistically significant change in insulin-like growth factor-1 (IGF-1) pre- vs. post-therapy (448.4±22.7 vs. 429.8±45.4, p=0.40).
Low-density lipoprotein (LDL) was 85.0±4.9 mg/dL pre-therapy, which was different from peak values while on therapy, 95.1±4.5 (p=0.02), but was not decreased overall post-therapy, 82.6±5.2 mg/dL (p=0.62). High-density lipoprotein (HDL) was 54.6±2.7 mg/dL pre-therapy and decreased to a nadir of 43.0±2.2 mg/dL on therapy (p<0.0001). Post-therapy HDL was 50.5±3.7 mg/dL and was lower than pre-therapy HDL (p=0.02). There was neither a statistical decrease in cholesterol (p=0.71), nor an increase in triglycerides (TG) post-compared to pre-therapy (p=0.32); however these values were not routinely assessed in fasting state so are difficult to interpret.
Pre-therapy, hematocrit was 41.3±0.6% and was different from the peak value on therapy 45.9±0.6% (p<0.0001), and there was a small increase in post-therapy compared to pre-therapy hematocrit (43.4±1.5% vs 41.3±0.6%, p<0.03). To be noted, post-therapy measurements were taken a minimum of 3 months after cessation of therapy.
No clinically notable abnormalities were seen for hemoglobin A1C, glucose, and liver function tests. Two subjects stopped anastrozole due to coincident anxiety issues, which persisted 4–6 months after cessation of therapy. One subject stopped anastrozole due to heart palpitations and shortness of breath; however, a cardiology evaluation suggested exercise-induced bronchospasm as the likely diagnosis. One subject noted worsening in acne on anastrozole and considered stopping therapy, however, the acne subsequently improved while still on therapy. Apart from one subject that missed a week of anastrozole therapy due to insurance coverage issues, no other issues with compliance were noted upon chart review (Table 2).
In this chart review, we have shown that AI therapy did not change the PAH or HtSDS in our subject sample, however, there may be a delay in BA progression that is more pronounced with longer duration of therapy. The fact that we only observed both a significant decrease in the delta BA/CA post-therapy along with a ΔBA/ ΔCA of 0.77 in Rx group 2 lends support to this suggestion. Prior studies have also shown a more prominent change in BA with longer duration of therapy (4), as well as ΔBA/ΔCA value approaching 0.62 after 2 years of therapy (2). We did observe BA advancement in Rx group 1 (where ΔBA/ΔCA was >1), however, this initial advancement of BA in the shorter duration of treatment likely reflects an inability to alter the ongoing prior effects of estrogen at the bone receptor level.
The near-final heights achieved by a smaller subject sample were substantially lower than their MPTHs, and for some, were even less than the pre-therapy prediction. We hesitate to speculate on an etiology for these findings as they are based on a limited number of subjects.
Hormonal alterations associated with treatment were as anticipated. The large increase in testosterone during therapy does not seem to be associated with any complications, and levels return to just slightly above baseline after the therapy is completed. The small decrease in HDL during therapy has been noted in other studies of AIs (4, 6, 7, 12). Furthermore, increased serum testosterone has been associated with lower HDL in adolescent males, perhaps via a mechanism involving apolipoproteins A-I and A-II (13). The small decrease in LDL post-therapy was not clinically significant. The alterations observed in cholesterol and TGs are difficult to interpret given that the samples were not routinely fasting, but prior studies of AIs have shown no change in these parameters (6, 7). Erythrocytosis has been previously reported with AI therapy and is likely to be considered as an androgen-mediated process (8, 14). The increase in hematocrit values reported here, however, was not clinically significant. The decrease in growth velocity, as expected, has also been noted in a previous study (5). To be noted, we observed no significant changes in IGF-1 with therapy. Previous studies have reported a range of inconsistent findings related to circulating IGF-1 levels; including increased IGF-1 levels in the placebo group compared to those on AIs post-therapy (2, 15), increased levels in both AI and placebo groups (4), and unchanged IGF-1 levels with AI therapy (6).
Finally, an inherent pitfall was that the study design was a retrospective chart review rather than a randomized placebo-controlled trial. The majority of our practitioners had a personal preference to prescribe a lower-potency AI, arimidex (3, 4) so our conclusions, as such, are limited and differ from other studies which mainly used letrozole (3, 4, 9). This study design also resulted in a heterogeneous patient population. Finally, as this was not a prospective protocol, the results should be interpreted with a modicum of caution.
In summary, we cannot from this pilot data suggest that AI therapy results in any changes in PAH in the attainment of MPTH, or that it improves final height when compared to the pre-therapy prediction. Given the lack of long-term safety data on AIs, one must continue to weigh the risks and benefits of such height augmentation therapy in the pediatric/adolescent population and pursue more prospective long-term outcome analyses.
Acknowledgments: This article was presented in part at the PAS/SPR Annual Meeting, April 2012.
Conflict of interest statement
Competing interests: The authors declare that they have no competing interests.
Authors’ contributions: KJ participated in the design and coordination of the study, carried out the chart review, performed part of the statistical analysis, and drafted the manuscript. SO conceived of the study, participated in its design and coordination, and helped to draft the manuscript. DM performed majority of the statistical analysis and assisted in the interpretation of the data. TC assisted in data collection, initial and final revision, and analyses. CY assisted in the chart review. IF, GA, SL, AH, and AS participated in the study design. All authors read and approved the final manuscript.
Grants/Fellowships: Dr. Sharon Oberfield currently holds a T32 Training Grant in Pediatric Endocrinology: T32 DK06552-08.
Disclosure Statement: The authors have nothing to disclose.
Morishima A, Grumbach MM, Simpson ER, Fisher C. Aromatase deficiency in male and female siblings caused by novel mutation and the physiological role of estrogens. J Clin Endocrinol Metab 1995;12:3689–98.Google Scholar
Hero M, Norjavaara E, Dunkel L. Inhibition of estrogen biosynthesis with a potent aromatase inhibitor increases predicted adult height in boys with idiopathic short stature: a randomized controlled trial. J Clin Endocrinol Metab 2005;90:6396–402.Google Scholar
Hero M, Wickman S, Dunkel L. Treatment with the aromatase inhibitor letrozole during adolescence increases near-final height in boys with constitutional delay of puberty. Clin Endocrinol 2006;64:510–3.Google Scholar
Mauras N, Gonzalez de Pijem L, Hsiang HY, Desrosiers P, Rapaport R, et al. Anastrozole increases predicted adult height of short adolescent males treated with growth hormone: a randomized, placebo-controlled, multicenter trial for one to three years. J Clin Endocrinol Metab 2008;93:823–31.Web of ScienceGoogle Scholar
Hero M, Toiviainen-Salo S, Wickman S, Mäkitie O, Dunkel L. Vertebral morphology in aromatase inhibitor-treated males with idiopathic short stature or constitutional delay of puberty. J Bone Miner Res 2010;25:1536–43.Web of ScienceGoogle Scholar
Wickman S, Saukkonen T, Dunkel L. The role of sex steroids in the regulation of insulin sensitivity and serum lipid concentrations during male puberty: a prospective study with a P450-aromatase inhibitor. Eur J Endocrinol 2002;146:339–46.Google Scholar
Hero M, Ankarberg-Lindgren C, Taskinen M, Dunkel L. Blockade of oestrogen biosynthesis in peripubertal boys: effects on lipid metabolism, insulin sensitivity, and body composition. Eur J Endocrinol 2006;155:453–60.Google Scholar
Hero M, Wickman S, Hanhijärvi R, Siimes MA, Dunkel L. Pubertal upregulation of erythropoiesis in boys is determined primarily by androgen. J Pediatr 2005;146:245–52.Google Scholar
Maliqueo M, Sun M, Johansson J, Benrick A, Labrie F, et al. Continuous administration of a P450 aromatase inhibitor induces polycystic ovary syndrome with a metabolic and endocrine phenotype in female rats at adult age. Endocrinology 2013;154:434–45.Web of ScienceGoogle Scholar
Greulich W, Pyle S. Radiographic atlas of skeletal development of the hand and wrist, 2nd ed. Stanford, CA: Stanford University Press, 1959.Google Scholar
Morrison JA, Sprecher DL, Biro FM, Apperson-Hansen C, Dipaola LM. Serum testosterone associates with lower high-density lipoprotein cholesterol in black and white males, 10 to 15 years of age, through lowered apolipoprotein AI and AII concentrations. Metabolism 2002;51:432–7.Google Scholar
Rochira V, Zirilli L, Madeo B, Maffei L, Carani C. Testosterone action on erythropoiesis does not require its aromatization to estrogen: insights from the testosterone and estrogen treatment of two aromatase-deficient men. J Steroid Biochem Mol Biol 2009;113:189–94.Google Scholar
Mauras N, Welch S, Rini A, Klein KO. An open label 12-month pilot trial on the effects of the aromatase inhibitor anastrozole in growth hormone (GH)-treated GH deficient adolescent boys. J Pediatr Endocrinol Metab 2004;17:1597–606.Google Scholar
About the article
Published Online: 2014-04-16
Published in Print: 2014-07-01