Inadequately Controlled Type 2 Diabetes and Hypercortisolism: Improved Glycemia With Mifepristone Treatment

Abstract

Objective: In many individuals, type 2 diabetes (T2D) remains poorly controlled despite taking multiple glucose-lowering therapies. Several studies have demonstrated that endogenous hypercortisolism is prevalent among these individuals. We tested whether cortisol-directed therapy improves their glycemic control.

Research design and methods: In this prospective, multicenter, double-blind study, 136 individuals with T2D (hemoglobin A1c [HbA1c] 7.5%-11.5% [58-102 mmol/mol] on multiple medications) and hypercortisolism (by dexamethasone suppression test) were randomized 2:1 to the glucocorticoid receptor antagonist mifepristone (300-900 mg once daily; n = 91) or placebo (n = 45) for 24 weeks, with stratification by presence/absence of an adrenal imaging abnormality. The primary end point was the change in HbA1c. Secondary end points included changes in glucose-lowering medications, weight, and waist circumference and safety.

Results: Mean baseline HbA1c in the study cohort was 8.55% (69.9 mmol/mol). At 24 weeks, the least squares mean (LSM) difference from placebo in HbA1c was -1.32% (95% CI -1.81 to -0.83; P < 0.001). Participants receiving mifepristone experienced reductions in body weight and waist circumference (placebo-adjusted LSM differences of -5.12 kg [95% CI -8.20 to -2.03] and -5.1 cm [-8.23 to -1.99], respectively). Of participants on mifepristone, 46% discontinued therapy, compared with 18% on placebo. Adverse events with mifepristone (>10% of participants) included hypokalemia, fatigue, nausea, vomiting, headache, peripheral edema, diarrhea, and dizziness, consistent with mifepristone's known tolerability profile. Increases in blood pressure also occurred.

Conclusions: In individuals with inadequately controlled T2D and hypercortisolism, cortisol-directed medical therapy with mifepristone reduced HbA1c, with a manageable tolerability profile.

Graphical abstract

Figure 1

LSM change from baseline in HbA_1c in all participants (A), participants with adrenal abnormality on computed tomography (CT) scan (B), and participants without adrenal abnormality on computed tomography scan (C). Note that the widths of CIs in B and C have not been adjusted for multiplicity and cannot be used to infer treatment effect.