Effects of Aromatase Inhibition in Hypogonadal Older Men: A Randomized, Double-Blind, Placebo-controlled Trial

Sherri-Ann M. Burnett-Bowie; Kristen C. Roupenian; Melissa E. Dere; Hang Lee; Benjamin Z. Leder


Clin Endocrinol. 2009;70(1):116-123. 

In This Article


There were no significant differences at baseline in the subjects in the two groups, except for small differences in E2 levels and total abdominal fat area by CT (Table 1). Additionally, there were no significant differences between the 69 subjects who completed the study and the 19 who were withdrawn.

Gonadal Steroids

Anastrozole therapy increased mean serum testosterone levels at all time points (P < 0·0001 vs. placebo) (Fig. 2a). Specifically, serum testosterone increased from 11·2 ± 3·3 nmol/l at baseline to 18·2 ± 4·8 nmol/l at month 3 (P < 0·0001). From months 3-12, testosterone remained significantly higher than baseline and placebo but decreased to 17·6 ± 5·0 nmol/l at 6 months and 16·5 ± 5·1 nmol/l at 12 months. The month 12 testosterone was significantly lower than the month 3 level (P = 0·03). None of the men in the anastrozole group had testosterone levels at any time point that exceeded the normal range for young healthy controls.

Figure 2.

Mean (± SE) (a) testosterone (b) bioavailable testosterone (c) dihydrotestosterone (d) E2 (e) SHBG, and (f) LH with anastrozole (solid line) vs. placebo (dashed line). *P < 0·05 as compared to placebo.

Aromatase inhibition also increased both bioavailable testosterone and DHT significantly (P < 0·0001 vs. placebo) (Fig. 2b,c). Bioavailable testosterone increased from 2·7 ± 0·8 nmol/l at baseline to 5·4 ± 1·7 nmol/l at month 3 (P < 0·0001). Bioavailable testosterone then decreased to 4·9 ± 2·2 nmol/l at month 12 (P = 0·09 vs. month 3). DHT increased from 1·9 ± 0·7 nmol/l at baseline to 3·4 ± 1·1 nmol/l at month 3 (P < 0·0001). DHT then decreased to 3·0 ± 1·2 nmol/l at month 12 (P = 0·025 vs. month 3). The 12-month bioavailable testosterone and DHT levels remained significantly different from baseline and placebo (P < 0·0001).

Anastrozole therapy modestly but significantly lowered E2 levels (P = 0·0004 vs. placebo) (Fig. 2d). Mean E2 decreased from 55·8 ± 15·4 pmol/l at baseline to 42·2 ± 13·6 pmol/l at 3 months (P < 0·0001). Unlike the pattern seen with the androgens, however, E2 levels did not revert towards baseline between months 3 and 12 (45·5 ± 15·4 pmol/l at month 12). In the anastrozole group, the on-treatment E2 ranged from 14·7 to 95·4 pmol/l. Anastrozole therapy significantly lowered mean SHBG levels (P = 0·0013 vs. placebo for overall effect, P = 0·093 to P = 0·862 for individual time points) (Fig. 2e). Aromatase inhibition increased mean LH levels from 4 ± 3 IU/l at baseline to 6 ± 4 IU/l at month 3 and remained stable thereafter (P = 0·0009 vs. placebo) (Fig. 2f).

Mean testosterone, bioavailable testosterone, DHT, E2, SHBG and LH were unchanged in the placebo group at all time points.

Other End-points

There were no differences in fat or lean mass between the anastrozole and placebo groups (Fig. 3a,b). Similarly, anastrozole therapy did not change total abdominal fat and thigh muscle area (Fig. 3c,d) or strength (2 ± 12%vs. 1 ± 11% change in 1-RM strength in the anastrozole and placebo groups, respectively (P = 0·6)). PSA levels were stable in both groups (Fig. 4). There was no difference in AUA-I scores between the groups (P = 0·66), although, as compared to baseline, mean AUA-I scores increased slightly in the placebo group (P = 0·05). Similarly, there was no difference between the groups in the change in HCT or lipid levels. As detailed below, PSA increases necessitating subject withdrawal were balanced between groups. At baseline all 69 subjects (34 in the treatment group and 35 in the placebo group) had a positive response to the ADAM questionnaire. At month 12, however, 7 subjects in the treatment group and 11 subjects in the placebo group had a negative response to the ADAM questionnaire.

Figure 3.

Mean (± SE) change in (a) total fat mass by DXA (b) total lean mass by DXA (c) total abdominal fat area by CT, and (d) thigh muscle area by CT with anastrozole or placebo for 12 months. No significant differences found between groups. DXA, dual X-ray absorptiometry; CT, computerized tomography.

Figure 4.

Mean (± SE) change in prostate specific antigen with anastrozole (solid line) or placebo (dashed line). No significant difference found between groups.

Subject Withdrawal and Serious Adverse Events

Nineteen subjects withdrew during the study: 11 in the anastrozole group and 8 in the placebo group. Six subjects withdrew consent for nonmedical reasons. Of the remaining 13, seven were withdrawn because of prespecified safety criteria related to the prostate (three in the anastrozole group, four in the placebo group). PSA increased in six subjects and BPH symptoms increased in one. One subject (treatment group) was withdrawn at month 6 after then disclosing that he had been diagnosed with obstructive sleep apnea (an exclusion criteria) prior to starting the study. Five serious adverse events resulted in subject withdrawal. One subject (placebo group) was diagnosed with prostate cancer after his PSA increased between months 0 and 3; one subject (anastrozole group) was diagnosed with pancreatic cancer at month 5; one subject (anastrozole group) was diagnosed with hepatitis A at month 5; one subject (anastrozole group) was diagnosed with a pulmonary embolism at month 6; and one subject (anastrozole group) had an embolic stroke at month 8.

Compliance with the drug was excellent as measured by pill counts and medication diaries. With one exception, subjects missed fewer than four doses of study medicine over the 4-month period prior to a study visit. One subject stopped taking his medication for 2 months prior to his final study visit.