Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`O R I G I N A L
`
`A R T I C L E
`
`E n d o c r i n e C a r e
`
`Mifepristone, a Glucocorticoid Receptor Antagonist,
`Produces Clinical and Metabolic Benefits in Patients
`with Cushing’s Syndrome
`
`Maria Fleseriu, Beverly M. K. Biller, James W. Findling, Mark E. Molitch,
`David E. Schteingart, and Coleman Gross, on behalf of the SEISMIC Study
`Investigators
`
`Oregon Health & Science University (M.F.), Portland, Oregon 97239; Massachusetts General Hospital
`(B.M.K.B.), Boston, Massachusetts 02114; Medical College of Wisconsin (J.W.F.), Milwaukee, Wisconsin
`53226; Northwestern University Feinberg Medical School (M.E.M.), Chicago, Illinois 60611; University of
`Michigan (D.E.S.), Ann Arbor, Michigan 48109; and Corcept Therapeutics (C.G.), Menlo Park, California 94025
`
`Context: Cushing’s syndrome (CS) is a disorder associated with significant morbidity and mortality
`due to prolonged exposure to high cortisol concentrations.
`
`Objective: Our objective was to evaluate the safety and efficacy of mifepristone, a glucocorticoid
`receptor antagonist, in endogenous CS.
`
`Design and Setting: We conducted a 24-wk multicenter, open-label trial after failed multimodality
`therapy at 14 U.S. academic medical centers and three private research centers.
`
`Participants: Participants included 50 adults with endogenous CS associated with type 2 diabetes
`mellitus/impaired glucose tolerance (C-DM) or a diagnosis of hypertension alone (C-HT).
`
`Intervention: Mifepristone was administered at doses of 300-1200 mg daily.
`
`Main Outcome Measures: We evaluated change in area under the curve for glucose on 2-h oral
`glucose test for C-DM and change in diastolic blood pressure from baseline to wk 24 for C-HT.
`
`Results: In the C-DM cohort, an area under the curve for glucose (AUCglucose) response was seen in 60%
`of patients (P ⬍ 0.0001). Mean ⫾ SD glycated hemoglobin (HbA1c) decreased from 7.43 ⫾ 1.52% to
`6.29 ⫾ 0.99% (P ⬍ 0.001); fasting plasma glucose decreased from 149.0 ⫾ 75.7 mg/dl (8.3 ⫾ 4.1 mmol/
`liter) to 104.7 ⫾ 37.5 mg/dl (5.8 ⫾ 2.1 mmol/liter, P ⬍ 0.03). In C-HT cohort, a diastolic blood pressure
`response was seen in 38% of patients (P ⬍ 0.05). Mean weight change was ⫺5.7 ⫾ 7.4% (P ⬍ 0.001) with
`waist circumference decrease of ⫺6.78 ⫾ 5.8 cm (P ⬍ 0.001) in women and ⫺8.44 ⫾ 5.9 cm (P ⬍ 0.001)
`in men. Overall, 87% (P ⬍ 0.0001) had significant improvement in clinical status. Insulin resistance,
`depression, cognition, and quality of life also improved. Common adverse events were fatigue, nausea,
`headache, low potassium, arthralgia, vomiting, edema, and endometrial thickening in women.
`
`Conclusions: Mifepristone produced significant clinical and metabolic improvement in patients
`with CS with an acceptable risk-benefit profile during 6 months of treatment. (J Clin Endocrinol
`Metab 97: 2039 –2049, 2012)
`
`ISSN Print 0021-972X ISSN Online 1945-7197
`Printed in U.S.A.
`Copyright © 2012 by The Endocrine Society
`doi: 10.1210/jc.2011-3350 Received December 12, 2011. Accepted March 8, 2012.
`First Published Online March 30, 2012
`
`Abbreviations: AE, Adverse event; AI, adrenal insufficiency; AUCglucose, area under the
`curve for glucose; BDI, Beck Depression Inventory; CD, Cushing’s disease; C-DM, patients
`with CS and T2DM/IGT; C-HT, patients with CS and a diagnosis of HTN; CI, confidence
`interval; CS, Cushing’s syndrome; DBP, diastolic blood pressure; DRB, data review board;
`ET, early termination; FPG, fasting plasma glucose; HbA1c, glycated hemoglobin; HDL-C,
`high-density lipoprotein cholesterol; HOMA-IR, homeostatic model assessment of insulin
`resistance; HTN, hypertension; IGT, impaired glucose tolerance; mITT, modified intent-to-
`treat; MRI, magnetic resonance imaging; oGTT, oral glucose tolerance test; T2DM, type 2
`diabetes mellitus.
`
`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`jcem.endojournals.org
`
`2039
`
`1
`
`TEVA1029
`
`

`

`Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`2040
`
`Fleseriu et al.
`
`Mifepristone and Cushing’s Syndrome
`
`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`Cushing’s syndrome (CS), is a serious endocrine disor-
`
`der that may be caused by a pituitary [Cushing’s
`disease (CD)] or nonpituitary (ectopic) ACTH-secreting
`tumor or by an adrenal neoplasm. If inadequately treated,
`CS is associated with a 3.8- to 5.0-fold higher mortality
`than the general population (1–3). Regardless of cause,
`surgery is usually the treatment of choice; however,
`complete removal of the neoplasm may not be possible
`(4, 5). Adjunctive radiotherapy for CD may take years
`to control excess cortisol (6). Laparoscopic bilateral
`adrenalectomy represents another treatment option.
`No medical treatments were approved by the U.S. Food
`and Drug Administration for CS when the study was
`conducted, but off-label use of several medications is
`common, including dopamine agonists, somatostatin
`analogs, and the adrenal steroidogenesis inhibitors (ke-
`toconazole, metyrapone, mitotane, and etomidate) (4,
`7). Ketoconazole and mitotane are effective in many
`patients, but in CD, doses may need progressive in-
`creases due to escape from cortisol blockade. The tol-
`erability of these drugs, especially at higher doses, limits
`their use in some patients (8, 9).
`(11␤-[P-(dimethylamino)phenyl]-17␤-
`Mifepristone
`hydroxy-17-(1-propynyl)estra-4,9-dien-3-one) is a pro-
`gesterone receptor antagonist that has glucocorticoid re-
`ceptor antagonist activity at higher concentrations, with
`more than three times the binding affinity for the gluco-
`corticoid receptor than dexamethasone (10, 11). It does
`not bind to the mineralocorticoid receptor (9). Case re-
`ports and small retrospective studies of mifepristone treat-
`ment in CS document improvements in abnormal glucose
`metabolism, psychiatric symptoms, and the somatic
`changes associated with CS; hypokalemia was the most
`commonly reported side effect (9, 12–25). Based on these
`preliminary findings, an open-label, prospective, multi-
`center, 6-month study of the safety and efficacy of mife-
`
`pristone was conducted in patients with endogenous CS
`refractory to other therapies.
`
`Patients and Methods
`
`Patients
`Adults with confirmed endogenous CS who had associated
`type 2 diabetes mellitus (T2DM), impaired glucose tolerance
`(IGT), or a diagnosis of hypertension (HTN) were enrolled (Fig.
`1). Endogenous hypercortisolism was defined as elevated urinary
`free cortisol on at least two 24-h collections and elevated late-
`night salivary cortisol and/or lack of suppression with dexameth-
`asone. T2DM was defined as a fasting plasma glucose (FPG) of
`at least 126 mg/dl (ⱖ7.0 mmol/liter) on two measurements or a
`2-h plasma glucose of at least 200 mg/dl (ⱖ11.1 mmol/liter) after
`a 75-g oral glucose tolerance test (oGTT), and IGT was defined
`as 2-h oGTT glucose value of 140 –199 mg/dl (7.8 –11.0 mmol/
`liter). HTN was defined as systolic blood pressure over 140 mm
`Hg and/or diastolic blood pressure (DBP) over 90 mm Hg or
`pharmacologically treated HTN.
`At least two of the following signs or symptoms of Cushing’s
`were also necessary for inclusion: Cushingoid appearance (moon
`facies, dorsocervical fat pad, and plethora), increased body
`weight or central obesity, proximal muscle weakness, low bone
`mineral density (T score ⬍ ⫺1.0), psychiatric symptoms, and
`skin changes (hirsutism, violaceous striae, or acne).
`Patients were excluded for poorly controlled diabetes mellitus
`[glycated hemoglobin (HbA1c) ⱖ 11%], poorly controlled HTN
`(⬎170/110 mm Hg), use of drugs to treat hypercortisolism within
`1 month of baseline (mitotane for adrenal carcinoma was allowed
`ifonstabledose ⱖ1monthbeforeentry),uncorrectedhypokalemia,
`or uncontrolled hypothyroidism or hyperthyroidism; also excluded
`were women with a uterus who required anticoagulants or had
`hemorrhagic disorders, endometrial hyperplasia, carcinoma, or
`polyps. Increases or additions of antihyperglycemic medications
`during the study were not permitted for patients with T2DM/IGT.
`For patients with HTN, increases or additions of antihypertensive
`medications were not permitted with the exception of mineralo-
`corticoid receptor antagonists, which were allowed for treating hy-
`pokalemia, a known side effect of mifepristone (9). Changes in or
`initiation of antidepressant or lipid-lowering medications were not
`allowed.
`
`The study was approved by the insti-
`tutional review board at each center and
`was registered with www.clinicaltrials.
`gov (NCT00569582). All patients pro-
`vided written informed consent.
`
`Design
`This was a 24-wk, open-label, multicenter
`study of mifepristone administered as a single
`daily oral dose. Treatment began at 300 mg/d;
`if no significant clinical improvement was
`noted by the investigator, doses could be in-
`creased to 600 mg/d on d 14, 900 mg/d at wk
`6, and 1200 mg/d at wk 10. Dose interruption
`and reduction were specified in the protocol
`for the following adverse events (AEs): adre-
`nal insufficiency (AI), severe hypokalemia,
`
`FIG. 1. Enrollment: ITT/safety population.
`
`2
`
`

`

`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`jcem.endojournals.org
`
`2041
`
`Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`and vaginal bleeding. Temporary glucocorticoid rescue for sus-
`pected AI was also allowed.
`
`Assessments
`The primary endpoint for patients with CS and T2DM/IGT
`(C-DM cohort) was the change in area under the curve for glucose
`(AUCglucose) on oGTT from baseline to wk 24. Response was de-
`fined as at least a 25% decrease in AUCglucose, an amount consid-
`ered clinically meaningful improvement in glucose control (26).
`AUCglucose was used because both patients with T2DM and pa-
`tients with IGT were enrolled, and HbA1c and FPG would not be
`uniformlyapplicable.Inpatientsreceivingmedicationsfordiabetes,
`administration occurred before the oGTT (other than short-acting
`insulin and glucagon-like peptide-1 analogs). The primary endpoint
`for patients with CS and a diagnosis of HTN (C-HT cohort) was the
`change in DBP from baseline to wk 24; response was defined as DBP
`decrease of at least 5 mm Hg (mean of two sequential readings).
`Patients with both T2DM/IGT and HTN were included only in the
`C-DM cohort.
`Key secondary endpoints included clinical response graded by
`an independent data review board (DRB) at wk 6, 10, 16, and 24
`compared with baseline. The DRB consisted of three CS experts
`who evaluated the following assessments: glucose homeostasis,
`blood pressure, lipids, weight and body composition change, clin-
`ical appearance (acne, hirsutism, striae, and Cushingoid appear-
`ance) (27, 28) as rated by the investigators, strength, and neuro-
`psychological [Beck Depression Inventory (BDI)-II and Trail
`Making Test] (29–31) and quality of life [Short-Form 36 Health
`Survey version 2 (SF-36)] (32) parameters. The DRB also reviewed
`standardized photographs of 34 consenting patients. Visit number
`after baseline and mifepristone dose were blinded. Each DRB mem-
`ber categorized patient overall status at follow-up visits as worse
`(⫺1), unchanged (0), or having clinically significant improvement
`(⫹1) from baseline. If the reviewers’ median score was ⫹1, the
`patient was considered to have clinical improvement.
`Blood, urine, and saliva samples were analyzed by a central
`laboratory (Quest Diagnostics, Collegeville, PA). AUCglucose and
`AUCinsulin were determined using the linear trapezoidal rule;
`homeostatic model assessment of insulin resistance (HOMA-IR)
`was calculated (33). Urinary and salivary cortisol levels were
`assayed with liquid chromatography tandem mass spectrometry
`[normal ranges, respectively, are 2– 42.4 ␮g/24 h (5.5–117
`nmol/24 h) and ⱕ0.09 ␮g/dl (2.5 nmol/liter)]; serum cortisol
`[normal range is 4 –22 ␮g/dl (110 – 607 nmol/24 h)], and ACTH
`(normal range is 5–27 pg/ml (1.1–5.9 pmol/liter) for females and
`7–50 pg/ml (1.5–11 pmol/liter) for males] were measured with
`immunochemiluminometric assay.
`AEs were reviewed every visit, and patients were monitored
`with vital signs, physical exams, and blood tests; transvaginal
`ultrasounds were conducted at baseline, wk 24 [or early termi-
`nation (ET)], and 6 wks after last dose. Pituitary magnetic res-
`onance imaging (MRI) was performed at screening and at wk 10
`and 24 (or ET) in patients with CD. Body composition was mea-
`sured using dual-energy x-ray absorptiometry at baseline and wk
`24 or ET using Hologic (Bedford, MA) or GE Lunar (Madison,
`WI) instruments; results were submitted to a central reading site
`for quality control and analysis.
`
`Statistics
`Patients who took at least one dose of study medication com-
`prised the safety population (n ⫽ 50). A modified intent-to-treat
`
`(mITT) population (patients who received ⱖ30 d of study med-
`ication) was used for analyses of efficacy (n ⫽ 46). The completer
`population included participants who completed through wk 24
`and were at least 80% compliant with study medication (n ⫽ 33).
`Because there was no placebo group in this study, a responder
`analysis was conducted. Responder rates were tested against an
`a priori threshold of 20%, which was chosen based on very low
`spontaneous response rates in this patient population (⬍5%)
`(34). The null hypothesis was to be rejected if the lower bound
`of the one-sided binomial 95% confidence interval (CI) of re-
`sponder rates was over 20%. Because mifepristone blocks rather
`than lowers cortisol, alternative quantitative endpoints (other
`than cortisol) were assigned at study entry based on inclusion in
`either C-DM or C-HT cohorts. Two abnormal oGTTs were re-
`quired for inclusion in the C-DM group; patients with a diagnosis
`of HTN and without T2DM/IGT were included in the C-HT
`group. For statistical analysis, response was defined as at least
`25% reduction in AUCglucose for C-DM patients or at least 5 mm
`Hg reduction in DBP in C-HT patients comparing baseline with
`wk 24/ET. For patients who did not complete the study or have
`an ET visit, the last available data were used. ANOVA and t tests
`were used for analyses of other endpoints. Nonparametric sta-
`tistical testing was employed for nonnormally distributed data.
`Change in oGTT curves over the course of the study was modeled
`by a hierarchical linear mixed model that took into consideration
`the correlation within subjects. SAS statistical software versions
`9.1.3 and 9.2 (Cary, NC) were used. Data are shown as mean ⫾
`SD unless otherwise stated.
`
`Results
`
`Patients
`From January 2008 to January 2011, 50 patients with CS
`were enrolled at 17 U.S. centers; 34 completed the study.
`Forty-three patients had a pituitary source of CS (42 with
`unsuccessful pituitary surgery, 18 with pituitary radiation,
`and one without previous surgery), four had ectopic ACTH
`secretion, and three had adrenal cortical carcinoma. Baseline
`characteristics are detailed in Tables 1 and 2. The mean
`dose ⫾ SD at the final study visit was 732 ⫾ 366 mg/d. Twen-
`ty-two subjects received the maximum dose of 1200 mg/d.
`Dose interruptions occurred in 42% of patients with median
`duration of 2 d (range 1–39 d). There were 18 dose reduc-
`tions in 12 patients; reductions occurred most commonly in
`300-mg decrements (317 ⫾ 114 mg).
`
`Primary efficacy analyses
`
`Patients with T2DM/IGT
`In the C-DM mITT population, AUCglucose decreased
`by at least 25% on oGTT in 15 of 25 (60%) patients from
`baseline to wk 24/ET (95% CI lower bound 42%, P ⬍
`0.0001) with a median decrease of 36% [30330.0 mg/
`dl䡠120 min (1683.3 mmol/liter䡠120 min) to 20655.0 mg/
`dl䡠120 min (1146.4 mmol/liter䡠120 min)] as well as com-
`parable reductions in plasma glucose levels (Fig. 2 and
`
`3
`
`

`

`Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`2042
`
`Fleseriu et al.
`
`Mifepristone and Cushing’s Syndrome
`
`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`TABLE 1. Demographics and body measurements at baseline (ITT/safety population)
`
`Characteristic
`Sex 关n (%)兴
`Male
`Female
`Race 关n (%)兴
`Black or African-American
`White
`Ethnicity 关n (%)兴
`Hispanic or Latino
`Not Hispanic or Latino
`Age (yr)
`Mean ⫾ SD
`Median
`Range
`Height (cm)
`Mean ⫾ SD
`Median
`Range
`Weight (kg)
`Mean ⫾ SD
`Median
`Range
`BMI (kg/m2)
`Mean ⫾ SD
`Median
`Range
`Waist circumference, cm
`Mean ⫾ SD
`Median
`Range
`Etiology of CS
`CD 关n (%)兴
`Ectopic ACTH 关n (%)兴
`Adrenal cancer 关n (%)兴
`
`C-DM (n ⴝ 29)
`
`C-HT (n ⴝ 21)
`
`Overall (n ⴝ 50)
`
`7 (24.1)
`22 (75.9)
`
`6 (20.7)
`23 (79.3)
`
`2 (6.9)
`27 (93.1)
`
`44.4 ⫾ 13.71
`41.0
`26 –71
`
`168 ⫾ 12.11
`168
`143.5–190.5
`
`105 (33.54)
`102
`61.3–198.7
`
`37.4 (11.18)
`35.1
`24.1– 66.4
`
`124 (21.73)
`120
`97.9 –178.4
`
`24 (82.8)
`3 (10.3)
`2 (6.9)
`
`8 (38.1)
`13 (61.9)
`
`2 (9.5)
`19 (90.5)
`
`2 (9.5)
`19 (90.5)
`
`46.7 ⫾ 8.83
`46.0
`26 – 67
`
`166 ⫾ 8.84
`163
`154.0 –185.4
`
`91.4 (21.10)
`88.2
`62.7–150.5
`
`33.4 (7.44)
`31.8
`24.5–53.6
`
`111 (15.77)
`104
`88.5–153.5
`
`19 (90.5)
`1 (4.8)
`1 (4.8)
`
`15 (30.0)
`35 (70.0)
`
`8 (16.0)
`42 (84.0)
`
`4 (8.0)
`46 (92.0)
`
`45.4 ⫾ 11.85
`45.0
`26 –71
`
`167 ⫾ 10.81
`166
`143.5–190.5
`
`99.5 (29.55)
`92.4
`61.3–198.7
`
`35.7 (9.90)
`33.5
`24.1– 66.4
`
`119 (20.31)
`115
`88.5–178.4
`
`43 (86.0)
`4 (8.0)
`3 (6.0)
`
`The C-DM group included subjects with T2DM and/or IGT at screening and d 1 asdetermined by two or more abnormal oGTT. The C-HT group
`included subjects with a diagnosis of HTN at screening but without T2DM and/or IGT.
`in DBP (95% CI lower bound 21%, P ⬍ 0.05; Table 3).
`Table 3). Similar reductions in AUCglucose were observed
`in the C-DM ITT and completer populations. The most
`Four patients (two responders) received spironolactone
`common doses among responders at wk 24/ET were 600
`during the study; one nonresponder was on spironolac-
`mg (40%) and 1200 mg (40%), followed by 300 mg
`tone at entry and remained on a stable dose throughout the
`(13.3%) and 900 mg (6.7%). In exploratory analyses we
`study.
`found no relationship between the incremental change in
`dose from baseline and AUCglucose (see Supplemental Fig.
`1, published on The Endocrine Society’s Journals Online
`web site at http://jcem.endojournals.org).
`
`Secondary endpoints
`
`Patients with HTN
`In the C-HT mITT cohort, eight of 21) patients (38.1%
`achieved the primary endpoint of at least 5 mm Hg decline
`
`TABLE 2. Biochemistry at baseline (ITT/safety population)
`
`Clinical improvement
`Theoverallclinicalimprovementresponserateasassessed
`by the DRB in the mITT population was 87% (95% CI lower
`bound 76%, P ⬍ 0.0001); response rates were similar in the
`C-DM and C-HT cohorts (Table 3). Thirty-three patients
`
`Biochemistry
`ACTH (pg/ml)
`24 h UFC (␮g/24 h)
`Serum cortisol (␮g/dl)
`Late-night salivary cortisol (␮g/dl)
`
`CD
`
`Ectopic ACTH
`
`Adrenal cancer
`
`Overall
`
`63 (51)
`139 (137)
`21.2 (6.0)
`0.29 (0.29)
`
`153 (140.3)
`2471 (3266)
`42.6 (14.3)
`1.90 (2.26)
`
`812 (559)
`37.4 (15.4)
`1.02 (0.58)
`
`66 (66)
`366 (1049)
`23.9 (10.0)
`0.47 (0.83)
`
`To convert values of ACTH to picomoles per liter, multiply by 0.22; urinary free cortisol (UFC) to nanomoles per 24 h, multiply by 2.759; cortisol to
`nanomoles per liter, multiply by 27.59.
`
`4
`
`

`

`Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`jcem.endojournals.org
`
`2043
`
`FIG. 2. Changes in glycemic parameters. A, Significant decreases in AUCglucose were observed in the C-DM cohort from baseline to each
`subsequent visit including wk 24/ET (P ⬍ 0.001). Data are shown as mean ⫾ SD. B, Significant decreases were also seen in plasma and fasting
`plasma glucose (P ⫽ 0.03), as measured by oGTT from baseline to wk 24. The oGTT response curves at each visit were statistically different
`compared with baseline. Mean data are shown. To convert glucose values to millimoles per liter, multiply by 0.0555.
`
`(72%) had a median score of ⫹1 at wk 24 or ET. Eleven
`patients by wk 6 and another six patients by wk 10 had a
`median score of ⫹1 with responses maintained throughout
`the remainder of the study (Initial clinical improvement re-
`sponse by dose and visit are shown in Supplemental Fig. 2).
`Three patients had a nonsustained improvement (median
`score of ⫹1 decreased to 0 at wk 24 or ET). One patient was
`rated as being worse at the final visit (early termination at wk
`10) than at baseline.
`
`Other glucose-related endpoints
`FPG decreased from 149.0 ⫾ 74.7 mg/dl (8.3 ⫾ 4.1
`mmol/liter) at baseline to 104.7 ⫾ 37.5 mg/dl (5.8 ⫾ 2.1
`mmol/liter) at wk 24 (P ⬍ 0.03). Antidiabetic medica-
`tions were reduced in seven of 15 patients. Of 12 pa-
`tients taking insulin, five reduced their daily dose by at
`least half. Eighteen of 25 C-DM patients (72%) had at
`least a 25% reduction from baseline in AUCglucose or a
`reduction in antidiabetic medication (95% CI ⫽ 50.6 –
`
`TABLE 3. Summary of responder analyses (mITT population)
`
`Responder
`关n (%)兴
`
`Nonresponder
`关n (%)兴
`
`Lower bound one-sided
`95% exact binomial CI (%)
`
`Statistics (mITT population)
`C-DM (n ⫽ 25)
`Participants with or without a 25% reduction
`from baseline in AUCglucose at wk 24/ET
`C-HT (n ⫽ 21)
`Participants who had ⱖ5 mm Hg reduction
`from baseline in DBP at wk 24/ET
`C-HT and C-DM with HTN at screening (n ⫽ 40)
`Participants who had ⱖ5 mm Hg reduction
`from baseline in DBP at wk 24/ET
`Participants who had a reduction in
`antihypertensive medications at wk 24/ET
`Participants who had either ⱖ5 mm Hg reduction
`from baseline in DBP or had a reduction in
`antihypertensive medications at wk 24/ET
`Median clinical improvement score of ⫹1
`at any reviewed visitb
`Combined cohorts (n ⫽ 46)
`C-DM (n ⫽ 25)
`C-HT (n ⫽ 21)
`a 95% CI ⫽ 36.1– 68.5.
`b For overall clinical improvement (median DRB score ⫹1) at any reviewed visit, the null hypothesis was to be rejected in favor of the alternative if
`the lower limit of the 95% exact one-sided binomial CI for the responder rate was at least 30%.
`
`41.7
`
`20.6
`
`P value
`
`⬍0.0001
`
`⬍0.05
`
`15 (60)
`
`10 (40)
`
`8 (38.1)
`
`13 (61.9)
`
`17 (42.5)
`
`23 (57.5)
`
`11 (27.5)
`
`29 (72.5)
`
`21 (52.5)a
`
`19 (47.5)
`
`40 (87.0)
`23 (92.0)
`17 (81.0)
`
`6 (13.0)
`2 (8.0)
`4 (19.0)
`
`75.9
`76.9
`61.6
`
`⬍0.0001
`
`5
`
`

`

`Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`2044
`
`Fleseriu et al.
`
`Mifepristone and Cushing’s Syndrome
`
`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`87.9%). The mean baseline HbA1c of 7.43 ⫾ 1.52% in
`the C-DM group decreased to 6.29 ⫾ 0.99% at wk
`24/ET (P ⬍ 0.001) (Fig. 3A). Twelve C-DM patients had
`an HbA1c over 7% at baseline (mean 8.53 ⫾ 1.11%);
`of these, nine achieved an HbA1c below 7%, including
`six reaching an HbA1c of 6% or below. C-DM and
`C-HT patients were insulin resistant and demonstrated
`rapid and significant
`improvements in AUCinsulin,
`which continued throughout
`the study (Fig. 3B);
`HOMA-IR demonstrated improvements in insulin sen-
`sitivity (Fig. 3C).
`
`Weight and body composition
`In the mITT population (n ⫽ 46), mean ⫾ SD body
`weight change from baseline (99.5 kg) to wk 24/ET was
`⫺5.7 ⫾ 7.4% (P ⬍ 0.001) (Fig. 4A). Twenty-four patients
`lost at least 5% of their baseline weight, 12 of whom lost
`
`at least 10%; 10 patients gained an average of 3.6 ⫾ 3.9%.
`Waist circumference decreased by ⫺6.8 ⫾ 5.8 cm (P ⬍
`0.001) in women and ⫺8.4 ⫾ 5.9 cm in men (P ⬍ 0.001)
`(Fig. 4B). Mean percent total body fat declined by 3.6% by
`wk 24 (P ⬍ 0.001). Absolute fat mass declined by 13.9%
`(P ⬍ 0.001) for the total body, 15.6% (P ⬍ 0.001) for the
`trunk, and 17.1% (P ⬍ 0.001) for the abdominal region
`(Fig. 4C).
`
`DBP and antihypertensive medications (C-HT and
`C-DM with HTN)
`In addition to the 21 C-HT patients, 19 C-DM patients
`had a diagnosis of HTN at study entry; 42.5% (17 of 40)
`of these had a reduction in DBP of at least 5 mm Hg at wk
`24/ET compared with baseline, and 27.5% had reductions
`in antihypertensive medications (50% of patients with a
`diagnosis of HTN were taking at least two antihyperten-
`
`FIG. 3. Changes in glucose-related outcomes. A, HbA1c significantly decreased from baseline to wk 24/ET (P ⬍ 0.001); B and C, a significant
`reduction in AUCinsulin (B) and significant improvements in HOMA-IR (C) were also observed. Data are shown as mean ⫾ SD. To convert insulin
`values to picomoles per liter, multiply by 6.945.
`
`6
`
`

`

`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`jcem.endojournals.org
`
`2045
`
`Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`FIG. 4. Changes in weight and body composition. Results demonstrated a significant reduction in body weight from baseline to wk 24/ET (P ⬍
`0.001) (A), significant decreases in overall waist circumference for females and males (P ⬍ 0.001) (B), and improvements in body composition (C).
`Data are shown as mean ⫾ SE.
`
`sive medications at baseline). Overall, 52.5% (95% CI ⫽
`36.13– 68.49%) had either a response in DBP or a reduc-
`tion in antihypertensive medications (Table 3). However,
`there were no significant differences in mean systolic
`blood pressure and DBP from baseline to wk 24/ET among
`C-HT patients (129.5 ⫾ 16.3/82.9 ⫾ 11.4 vs. 129.9 ⫾
`19.0/82.8 ⫾ 13.2 mm Hg) or in C-DM patients with a
`diagnosis of HTN (137.7 ⫾ 24.0/86.4 ⫾ 15.3 vs. 132.2 ⫾
`16.7/82.4 ⫾ 13.2 mm Hg). Eight of 12 patients with DBP
`of at least 90 mm Hg at study entry had a reduction of at
`least 5 mm Hg (median decline ⫺14 mm Hg, range ⫺26.5
`to ⫺5.5 mm Hg); only one (C-DM patient) of the eight
`received additional antihypertensive therapy. AEs of in-
`creased blood pressure were reported in 12 patients, nine
`(75%) of whom had evidence of mineralocorticoid recep-
`
`tor activation (edema, hypokalemia, and/or need for spi-
`ronolactone to control hypokalemia).
`
`Mood, cognition, and quality of life
`Median BDI-II depression scores improved in the mITT
`population (baseline 14.5, range 0 – 49; wk 24/ET 9.5,
`range 0 –36; P ⬍ 0.001). For patients with at least mild
`depression at baseline (BDI-II ⱖ 14, n ⫽ 24), median
`BDI-II scores improved from 23 (range 14 – 49) to 12
`(range 0 –34) (P ⬍ 0.001). Cognition scores were mea-
`sured by the Trail Making Test at wk 24/ET; there were
`improvements in both Trail A (median decrease of 4.0 sec,
`P ⬍ 0.01) and Trail B (median decrease of 12 sec, P ⬍
`0.01). Quality of life improved at wk 24/ET as measured
`by SF-36 mental composite scores (mean 40.0 ⫾ 14.5 vs.
`
`7
`
`

`

`Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`2046
`
`Fleseriu et al.
`
`Mifepristone and Cushing’s Syndrome
`
`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`45.4 ⫾ 12.5, P ⫽ 0.01) and physical composite scores
`(mean 34.9 ⫾ 11.0 vs. 39.1 ⫾ 10.8, P ⫽ 0.02).
`
`Hormone and pituitary MRI scan changes
`During mifepristone treatment, 72% of the 43 patients
`with CD had at least a 2-fold increase in ACTH, cortisol,
`or both; 28% had smaller increases. These changes were
`observed early (by d 14), plateaued from wk 10 –24, and
`declined to baseline levels at the follow-up visit 6 wk after
`discontinuation of mifepristone. Increases in ACTH of at
`least 2-fold were observed in 62.8% of patients; 33.6%
`had lesser increases, and 4.7% had no change. Late-night
`salivary cortisol increased 7.92-fold (1.43) at wk 16, and
`urinary free cortisol increased 7.70-fold (15.29) at wk 24/
`ET. At the 6-wk follow-up visit, ACTH and cortisol (se-
`rum and urine) declined to near baseline levels. Patients
`with ectopic ACTH secretion did not demonstrate in-
`creases in ACTH and cortisol in response to mifepristone.
`Pituitary MRIs were obtained in 41 CD patients; 17 had
`visible tumors, 10 of which were macroadenomas, and the
`remaining 24 did not have demonstrable tumors after sur-
`gery. MRIs were stable at wk 10 and 24 in all cases except
`one. This patient had an aggressive pituitary tumor at
`baseline that was increased in size at wk 10, leading to
`treatment discontinuation.
`
`Safety
`Overall, AEs were reported in 88% of patients during
`mifepristone treatment, most commonly nausea (48%),
`fatigue (48%), headache (44%), decreased blood potas-
`sium (34%), arthralgia (30%), vomiting (26%), periph-
`eral edema (26%), HTN (24%), dizziness (22%), de-
`creased appetite (20%), and endometrial
`thickening
`(20%). The majority of AEs were considered mild or mod-
`erate. Seven patients discontinued mifepristone because of
`an AE; fatigue was the only cause of discontinuation for
`more than one patient (n ⫽ 2). Interruptions or reductions
`in mifepristone due to AEs, most commonly nausea (n ⫽
`6), occurred in 40% of patients; there were interruptions
`or reductions for protocol-specified events in four subjects
`(two for AI, one for severe hypokalemia, and one for vag-
`inal bleeding). After dose interruption or reduction before
`wk 10, there were increases in dose in one of four and two
`of five patients, respectively; after wk 10, dose escalation
`did not occur after an interruption for an AE except in a
`single patient. Four patients experienced progression of
`preexisting metastatic malignancy that resulted in death.
`AI was reported in two patients. One occurred during
`an infection and responded to withdrawal of mifepristone;
`the other resolved with mifepristone withdrawal and
`dexamethasone administration (6 –9 mg by mouth daily
`for 6 d). Neither episode was associated with hypoglyce-
`
`mia or hypotension, and mifepristone was restarted at a
`lower dose. Analysis of AEs and concomitant medications
`identified five other instances of two or more symptoms
`possibly consistent with AI during which glucocorticoids
`were administered. Dexamethasone doses for these epi-
`sodes ranged from 2– 8 mg daily in tapering amounts for
`1–12 d. Vaginal bleeding was observed during the study in
`five premenopausal women. Prolonged metrorrhagia was
`observed in two of them after discontinuing mifepristone.
`Endometrial thickening was reported as an AE in 10
`women. Three women underwent dilatation and curettage
`for unresolved endometrial thickening.
`Twenty-two patients had a serum potassium level less
`than 3.5 mEq/liter (⬍3.5 mmol/liter), but only three ex-
`perienced severe hypokalemia [ⱕ2.5 mEq/liter (ⱕ2.5
`mmol/liter)] during mifepristone treatment, including one
`serious AE [potassium 2.1 mEq/liter (2.1 mmol/liter)]. Hy-
`pokalemia occurred in patients with both ACTH-depen-
`dent and independent CS. Four (one adrenal cancer and
`three ectopic ACTH) of seven patients with nonpituitary
`CS experienced hypokalemia during treatment. Hypoka-
`lemia was often associated with alkalosis and edema and
`generally responded to potassium replacement (10 – 420
`mEq daily); all nonpituitary CS patients received potas-
`sium supplementation. Overall, spironolactone (50 – 400
`mg daily) was used by 14 patients; it was started or in-
`creased in 11 patients for hypokalemia while taking mife-
`pristone, including one patient with adrenal cancer and
`two patients with ectopic ACTH secretion. Reversible de-
`creases in high-density lipoprotein cholesterol (HDL-C)
`and increases in TSH were observed. The mean change in
`HDL-C from baseline [62.3 ⫾ 27.8 mg/dl (1.61 ⫾ 0.72
`mmol/liter)] to wk 24/ET was ⫺14.2 ⫾ 11.9 mg/dl (0.37 ⫾
`0.31 mmol/liter) (P ⬍ 0.001); there were small declines in
`low-density lipoprotein cholesterol and triglycerides that
`were not statistically significant. Eight patients had unde-
`tectable TSH at baseline; of the remaining 42 patients,
`eight had increases in TSH above normal (three with
`TSH ⬎ 10 ␮U/liter, one with TSH of 32 ␮U/liter). Six
`weeks after mifepristone discontinuation, both HDL-C
`and thyroid function tests reverted to baseline levels.
`
`Discussion
`
`Cushing’s syndrome is a complex endocrine condition
`with serious sequelae, including cardiovascular mortality,
`fractures, proximal myopathy, insulin-resistant hypergly-
`cemia, and neuropsychiatric and neurocognitive disorders
`(35, 36). Transsphenoidal pituitary surgery with adenoma
`resection is initially successful in 65–90% of patients with
`ACTH-secreting microadenomas when performed by ex-
`
`8
`
`

`

`J Clin Endocrinol Metab, June 2012, 97(6):2039 –2049
`
`jcem.endojournals.org
`
`2047
`
`Downloaded from https://academic.oup.com/jcem/article-abstract/97/6/2039/2536684 by National Library of Medicine user on 13 March 2019
`
`pert surgeons, but approximately 20 –25% have persistent
`hypercortisolism or recurrence postoperatively; cure rates
`are lower and recurrence rates are higher for macroad-
`enomas (4). Morbidity and mortality in patients with CD
`are related to cortisol excess and rarely to the ACTH-
`secreting pituitary tumor mass. When surgery fails to re-
`verse hypercortisolemia, medical treatment can suppress
`cortisol overproduction and improve clinical manifesta-
`tions. Bilateral adrenalectomy promptly resolves hypercor-
`tiso