Discontinuation of estrogen replacement therapy in GH-treated hypopituitary women alters androgen status and IGF-I

Abstract

Objective and design: Compared with their male counterparts, healthy females secrete more growth hormone (GH) and those with GH-deficiency have lower insulin-like growth factor I (IGF-I) levels and are less responsive to GH substitution. To test whether this gender difference is related to sex hormones we measured androgen status and IGF-I related parameters in 38 hypopituitary women (mean (range) age 41.5 (20–58) years) during continued GH substitution as compared with a control group of 38 healthy women matched for age and menopausal status. Twenty six patients were studied twice: with estrogen replacement and after 28 days of estrogen discontinuation in a randomised design. Results: The patients were androgen deficient compared with controls (median, range), dehydroepiandrosterone sulphate (DHEAS): 185 (99–7800) nmol/l vs 4400 (820–13 000) nmol/l, P = < 0.001; androstenedione: 0.5 (0.1–7.1) nmol/l vs 4.3 (1.6–8.8) nmol/l, P = < 0.001; dihydrotestosterone (DHT): 0.13 (0.09–0.54) nmol/l vs 0.55 (0.09–0.89) nmol/l, P = < 0.001; testosterone: 0.28 (0.09–1.56) nmol/l vs 1.1 (0.71–2.24) nmol/l, (P = < 0.001); free testosterone: 0.004 (0.001–0.030) nmol/l vs 0.016 (0.001–0.030) nmol/l, P = < 0.001. The circulating levels of IGF-I, IGF-II, IGF-binding protein 1 (IGFBP-1), and IGFBP-3 did not differ between patients and controls. The subgroup of patients receiving hydrocortisone (HC) replacement (n = 24) had significantly lower levels of androgens (suppressed by 80–100%) as well as IGF-I and IGFBP-3 as compared with the patients not receiving HC. IGF-I was correlated to free testosterone in patients (r = 0.57, P = 0.0005) as well as controls (r = 0.43, P = 0.008), and free testosterone was a significant positive predictor of IGF-I. Estrogen discontinuation induced an increase in IGF-I (167 ± 15 vs 206 ± 14 μg/l, P = 0.005 and IGFBP-3 (3887 ± 139 vs 4309 ± 138 μg/l, P = 0.0005). Estrogen discontinuation was associated with a significant increase in median (range) free testosterone (0.004 (0–0.02) vs 0.0065 (0–0.03) nmol/l, P = 0.001) and a significant decrease in median (range) sex-hormone binding globulin (SHBG; 93 (11–278) vs 55.5 (20–142) nmol/l, P = 0.001). ΔIGF-I correlated with ΔSHBG (r = −0.45 P = 0.033) and ΔIGFBP-3 (r = 0.67 P = < 0.001). In a regression model ΔE2, Δtestosterone, ΔSHBG and ΔIGFBP-3 explained 93% of the variation in ΔIGF-I. Conclusions: Androgen levels are low in hypopituitary women and free testosterone correlates with IGF-I. Discontinuation of estrogen replacement in these patients induces elevations in IGF-I as well as free testosterone, and ΔIGF-I correlated positively with Δfree testosterone. These effects may contribute to the gender differences observed in the GH–IGF axis in healthy adults as well as in the responsiveness of hypopituitary patients to GH substitution.