Effects of metformin and leuprolide acetate on insulin resistance and testosterone levels in nondiabetic postmenopausal women: a randomized, placebo-controlled trial

Abstract

Objective: To determine whether insulin sensitizers lower androgen levels and whether androgen suppression improves insulin resistance in nondiabetic postmenopausal women.

Design: Randomized, double-blind, placebo-controlled study.

Setting: Clinical and Translational Research Center of a university hospital.

Patient(s): Thirty-five postmenopausal women aged 50-79 years with insulin resistance and higher T levels.

Intervention(s): Subjects were randomized to metformin plus leuprolide acetate (LA) placebo, LA plus metformin placebo, or LA placebo plus metformin placebo in a 1:1:1 fashion during a 12-week period.

Main outcome measure(s): Insulin sensitivity (M) assessed by euglycemic-hyperinsulinemic clamp and free T by equilibrium dialysis.

Result(s): In those randomized to metformin, free T decreased by 19% compared with placebo, along with an expected improvement in M. Total T also decreased significantly, whereas sex hormone-binding globulin (SHBG) did not change. In those randomized to LA, the percent change in M was not different from placebo, despite a 48% relative decrease in free T levels.

Conclusion(s): These data are the first to establish a causal link between insulin resistance and T in postmenopausal women. They confirm that treatment of insulin resistance decreases T production in this population and demonstrate that pharmacologic lowering of T does not affect insulin resistance.

Figure 1

Longitudinal data for (A) free testosterone, (B) calculated free testosterone, (C) total testosterone, (D) SHBG, (E) HOMA-IR and (F) insulin by group. Mean values with SE bars are presented. P-values are for longitudinal comparisons to placebo of log-transformed values. Abbreviations: SHBG, sex hormone binding globulin; HOMA-IR, homeostasis model assessment of insulin resistance

Figure 2

Longitudinal data for (A) LH, (B) FSH, (C) estradiol, and (D) DHEAS by group. Mean values with SE bars are presented. P-values are for longitudinal comparisons to placebo. DHEAS values were log-transformed for analysis.