Functional hyperandrogenism may be associated with a mild increase in body iron stores. Iron depletion exerts a beneficial effect on metabolic endpoints in other iron overload states. (i) To determine the effect of iron depletion on the insulin sensitivity and frequency of abnormal glucose tolerance in patients with functional hyperandrogenism submitted to standard therapy with combined oral contraceptives (COC). ii) To assess the overall safety of this intervention. Randomized, parallel, open-label, clinical trial. Academic hospital. Adult women with polycystic ovary syndrome or idiopathic hyperandrogenism. After a 3-month run-in period of treatment with 35 μg ethinylestradiol plus 2 mg cyproterone acetate, participants were randomized (1:1) to 3 scheduled bloodlettings or observation for another 9 months. Changes in insulin sensitivity index and frequency of prediabetes/diabetes, and percentage of women in whom bloodletting resulted in plasma hemoglobin <120 g/L and/or hematocrit <0.36. From 2015 to 2019, 33 women were included by intention-to-treat. During the follow-up, insulin sensitivity did not change in the whole group of women or between study arms [mean of the differences (MD): 0.0 (95%CI: -1.6 to 1.6)]. Women in the experimental arm showed a similar odds of having prediabetes/diabetes than women submitted to observation [odds ratio: 0.981 (95%CI: 0.712 to 1.351)]. After bloodletting, 4 (21.1%) and 2 women (10.5%) in the experimental arm had hemoglobin (Hb) levels <120 g/L and hematocrit (Hct) values <0.36, respectively, but none showed Hb <110 g/L or Hct <0.34. Scheduled bloodletting does not improve insulin sensitivity in women with functional hyperandrogenism on COC.