hypercortisolism common in hard-to-treat diabetes

 

 

In patients with difficult-to-control type II diabetes, an impressive study found that about one quarter of them have hypercortisolism, and mifepristone improves their hyperglycemia (see dm hypercortisolism DiabCar2025 in dropbox or https://doi.org/10.2337/dc24-2841).

thanks to Dr Elvira Isganaitis for bringing this to my attention.

 

Details:

-- the CATALYST study enrolled 1063 participants with type II diabetes and a hemoglobin A1c ranging from 7.5%-11.5% and who were on two or more glucose lowering medications with or without micro/macro vascular complications or taking multiple blood pressure lowering medications

-- these patients were screened with a 1 mg dexamethasone suppression test (DST)

    -- common causes of false positives in DST evaluations were excluded, including taking oral contraceptives, excessive alcohol consumption, severe untreated sleep apnea, severe psych/medical/surgical illness, night shift work or hemodialysis/end-stage renal disease 

-- mean age 61, 45% female, 71% white/19% Black/24% Latino

-- BMI 34, waist circumference 113 cm

-- hemoglobin A1c 8.8%, blood pressure 128/75

-- glucose-lowering medications: at least three in 71%, insulin plus other meds in 70%, at least two glucose lowering meds plus at least 1 micro/macro-vascular complication 50%, at least 2 glucose-lowering meds and at least 2 blood pressure lowering meds 46%

-- glucose-lowering meds taken: metformin 74%, insulin 73%, SGLT2s 52%, GLP-1 receptor agonists (any dose) 48%, GLP-1 maximum dose 19%, GLP-1 (any dose) or tirzepatide, 57%, sulfonylureas 29%, pioglitazone 11%, tirzepatide alone 10%, DPP-4 inhibitors 10%

 

-- primary endpoint: prevalence of hypercortisolism, defined as a post-DST cortisol >1.8 μg/dL (50 nmol/L) [this means that endogenous cortisol production was not suppressed by the dexamethasone]

    -- in addition, they assessed the characteristics associated with hypercortisolism as well as a percentage of participants with hypercortisolism and adrenal imaging abnormalities

 

Results:

-- 1057 participants completed the overnight 1 mg DST with dexamethasone levels at least 140 ng/mL [this is to ensure that they got sufficient dexamethasone to suppress the morning cortisol level]

    -- dexamethasone threshold was not met in one person

    -- repeat DST was needed but not repeated in 5 people

-- repeated DST utilizing 4mg dexamethasone was used to achieve the 140 ng/dL threshold in 25 people

    -- of these, 17 people had post-DST cortisol at least 1.8 μg/dL (unsuppressed, hypercortisolemic) and 8 had suppressed cortisol (ie, did not have hypercortisolism)

-- CT scans were completed in 219 of the 252 people with unsuppressed cortisol (ie, >1.8 μg/dL)

 

-- post-DST cortisol was unsuppressed (abnormal, reflecting hypercortisolism) in:

    -- overall population: 252 of the 1057 participants: prevalence 23.8% (21.3-26.5%)

    -- patients with cardiac disorders: 33.3%

        -- on univariate analysis, there was increased association of hypercortisolism with coronary artery disease (16.7% vs 9.8% if no hypercortisolism), atrial fibrillation (8.3% vs 3.0%), heart failure (6.7% vs 1.4%), hypertension (89.3% vs 78.8%); all of these differences were highly statistically significant

    -- patients taking three or more blood pressure lowering medications: 36.6% [ie, those with more resistant-to-treat hypertension had even a higher association with hypercortisolism]

 

-- CT scans in individuals with hypercortisolism:

    -- adrenal imaging abnormalities: 34.7% of those with hypercortisolism

    -- on multivariate analysis, increased odds of hypercortisolism with abnormal adrenal CT exams:

        --  greater waist circumference: 2.3% increase, OR 1.023 (1.004-1.042), p=0.02

        -- use of b-blockers: 3-fold risk, OR 2.928 (1.560-5.497), p=0.0008 [but were more b-blockers given as treatment for palpitations, which are more common in those with hypercortisolism??]

    -- and decreased odds of hypercortisolism with abnormal adrenal CT exams:

        -- each 5-mmHg increase in diastolic BP: 0.5% decrease, 0.953 (0.921-0.986), p=0.006

        -- use of insulin and SGLT-2 inhibitors: 63% decrease, 0.365 (0.186-0.718), p=0.004

        -- each 5 ng/dL decrease in ACTH level: 13% decrease, 0.867 (0.771-0.974), p=0.02

    -- no differences found by eGFR, liver function tests or lipid concentrations

 

-- odds ratios of hypercortisolism in different patients, all having statisticllly significant p<0.03:

    -- those on SGLT2 inhibitors: OR 1.558

    -- those on maximum dose GLP-1 receptor agonists: OR 1.544

    -- those on tirzepatide: OR 1.981 [?reverse causation: those on tirzepatide, the most powerful of the GLP-1 type meds, may well have had the most difficult-to-control diabetes??, same with those on max doses of other GLP-1 meds]

    -- those on a higher number of blood pressure lowering medications: OR 1.390

    -- older age: OR 1.316

    -- BMI <30: OR 1.639

    -- non-Latino ethnicity: OR 3.718

    -- those on fibrates: OR 2.676

    -- those on analgesics: OR 1.457

 

Commentary

-- even with the availability of phenomenal newer diabetes medications (GLP-1 receptor agonists, SGLT2 inhibitors) which have led to shockingly great diabetes control as well as providing cardiovascular and renal protection, in addition to the use of multiple medications for diabetes when needed, still about one quarter of US individuals with type-2 diabetes have hemoglobin A1c >8%

-- people with hypercortisolism have several associated problems that can lead to poorer glycemic control: insulin resistance, impaired pancreatic b-cell function, inhibiting the pancreatic insulin release by GLP-1 (as stimulated by the GLP-1 receptor agonists) or GIP (the additional glucose-dependent insulinotropic polypeptide stimulated by tirzepatide), and increasing hepatic glucose output

    -- and, hypercortisolism itself is associated with increased cardiovascular disease and overall mortality

-- this all suggests the hypercortisolism led to more severe and hard-to-treat diabetes and hypertension

    -- as an anecdote, i had a 50yo man who had severe obesity (weight up to 600 pounds), hard-to-treat hypertension (blood pressure in the 190+ mmHg range despite 4 medications), hard-to-treat diabetes (A1c in the 10-12% range despite 3 diabetes meds), whom i diagnosed with Cushing's disease by this clinical constellation as well as his prominent purple abdominal striae and "buffalo hump". He had a pituitary adenoma causing this problem, had surgery, and literally within months lost about 100 pounds and had excellent diabetes and hypertension control on few or no meds. this is certainly a more extreme example of the hypercortisolism than in the current study, but it highlights the rather profound effects of high systemic cortisol levels

    -- though Cushing's disease has been the poster-child for hypercortisolism, the presence of hypercortisolism without the classic more-extreme clinical findings found in Cushing's has been recognized for the past few decades, often in the setting of comorbidities of diabetes and hypertension. This hypercortisolism found in CATALYST is more typically associated with autonomous cortisol production by the adrenals

-- the findings in this CATALYST trial have been found in smaller studies in the past, with hypercortisolism in 5-33% of people with difficult-to-control diabetes and predated the use of the newer diabetes meds (GLP-1's and SGLT2's)

 

-- the results of the CATALYST trial were quite impressive: in this large study, patients with hard-to-control diabetes had a 23.8% chance of having hypercortisolism per an abnormal DST, those individuals had a 34.7% change of abnormal adrenal CT scans, with 65.8% having unilateral adrenal nodules

 

-- There was another impressive article from the same group of researchers who then assessed the effect of taking mifepristone, a glucocorticoid receptor antagonist, for the treatment of the patients identified above with diabetes and hypercortisolism, finding a pretty dramatic improvement in their glycemia (see dm hypercortisolism mifepristone rx DiabCare2025 in dropbox, or https://doi.org/10.2337/dc25-10550.)  It turns out that mifepristone actually does have multiple clinical uses as a synthetic steroid beyond pregnancy termination, including for Cushing syndrome and uterine leiomyomas: https://www.ncbi.nlm.nih.gov/books/NBK557612/.   :

    -- in this prospective, multicenter, double-blind study, 136 individuals with type 2 diabetes who had hypercortisolism by DST identified in the first CATALYST study were included

    -- as above, the patients had hemoglobin A1c of 7.5%-11.5%, with an mean of 8.8%

    -- patients were randomized to mifepristone 300-900 mg once a day (n=91) or placebo (n=45) for 24 weeks, with stratification for presence/absence of an adrenal imaging abnormality

    -- primary endpoint was change in A1c, secondary endpoints including changes in glucose lowering medications, weight, and waist circumference and safety

    -- findings at 24 weeks:

        -- difference in hemoglobin A1c was -1.32 percentage points in those on the mifepristone versus placebo

        -- reduction in body weight: -5.12 kilograms (-8.20 to -2.03)

        -- reduction in weight circumference: -5.1 cm (-8.23 to -1.99)

        -- relationship of glycemic control by adrenal CT scan results:

            -- patients with adrenal abnormalities on CT: difference in baseline A1c from placebo -1.41% (-2.34% to -0.48%)

            -- patients without adrenal abnormalities on CT: difference in baseline A1c from placebo -1.29% (-1.88% to -0.70%)

                -- Both of these were highly statistically significant improvements over placebo, and they had overlapping confidence intervals (ie no statistically significant difference between those with or without CT abnormalities)

       -- there were also decreases in the group on mifepristone in the use of fast-acting insulin, long-acting insulin and sulfonylureas (ie, the improvements in Aic with the mifepristone was actually even better, since these individuals often needed fewer glucose lowering meds). In patients on insulin and sulfonylureas, adding the mifepristone can lead to rapid glucose reductions, so it is important to follow the glycemic effects of mifepristone closely and modify these diabetes meds

        -- although the initial CATALYST study started with low dose mifepristone (300mg) with increase to 600mg and then to 900mg if needed, the glycemic benefits tended to begin early in treatment, prior to loss of significant weight and with low dose mifepristone

        -- discontinuation of therapy: 46% on mifepristone and 18% on placebo

        -- adverse effects in at least 10% of participants (though most adverse effects were grade 1 or 2 and manageable):

            --mifepristone: hypokalemia, fatigue, nausea, vomiting, headache, peripheral edema, diarrhea, dizziness (all expected, per other studies).  Also increases in blood pressure (increase in systolic blood pressure of 8.0 mmHg, a not insignificant issue, though this was counteracted in the study primarily with spironolactone); this increased BP was less evident in those with baseline systolic BP <130 mmHg.

            -- 3 people had euglycemic diabetic ketoacidosis on SGLT-2 meds, associated with decreased food intake and the resulting reduction in insulin doses, but no prolonged hospitalization; the ketoacidosis resolved, and patients stayed in the study

-- So, the study completes the circle: A significant number of patients with uncontrolled diabetes have hypercortisolism which is associated with significantly decreased response to medications for glycemia or blood pressure control; these patients, some of whom have adrenal abnormalities on CT scan, do respond clinically to a medication that reverses the hypercortisolism

-- these results from mifepristone in terms of changes in A1c, weight, and waist circumference were also found in participants in the SEISMIC study of people with overt features of hypercortisolism (Cushing’s syndrome): see cushings mifepristone helps SEISMIC JClinEndMetab2012 in dropbox, or https://pubmed.ncbi.nlm.nih.gov/22466348/ or doi.org/10.2337/dc25-1055

-- also, other studies have found that the weight loss associated with mifepristone was associated with more loss of visceral fat vs subcutaneous fat and muscle (visceral fat is the bad actor, associated with systemic inflammation and worse cardiovascular outcomes)

-- one complicating issue is that uncontrolled diabetes, hypertension, and obesity can be associated with abnormal DST results, with some studies finding elevated cortisol levels in patients with kidney disease, heart failure, even depression and stress. and these conditions are remarkably common and could undercut the direct association between hypercortisolism and hard-to-treat diabetes and hypertension. But this study of mifepristone suggests that just finding abnormal DST results and treating with mifepristone does reinforce the benefit of DST testing and treating with mifepristone

 

Limitations:

-- as a reference, it would have been useful to have a control group in the study with patients having controlled diabetes, to be certain that abnormal DST testing is basically confined to those with uncontrolled diabetes.

-- this study mostly involved non-Hispanic white patients, limiting the generalizability of the results. Were their differences reflect differing diets/exercise/stress levels/other diseases causing high cortisol levels?

-- lots of early terminations in the study may have distorted the results. unclear why so many people terminated their involvement. ?adverse effects ?which ones if that was the cause?

 

So, pretty striking trials because of the large percent of people with inadequately controlled diabetes/hypertension who had hypercortisolism

    -- and that mifepristone was so effective in improving glycemic control

    -- unfortunately, no information about improved hypertension control. Or ability to decrease meds for hypertension

        -- it would be interesting/useful to know if people with hard-to-control hypertension alone would benefit from assessing for hypercortisolism, and treating with mifepristone if positive

    -- what about patients who have minimal response to GLP-1 receptor agonists in terms of diabetes, and perhaps weight loss (an uncommon finding in my practice, but i do have several such patients)? should  these individuals be worked up for hypercortisolism??

    -- it was impressive that 2/3 of those with abnormal CT scans had adrenal nodules:

        -- were these the cause of the hypercortisolism, or were they benign “incidentalomas”?

        -- if adrenal assessment did show these were functioning adenomas producing cortisol, it may be useful to perform surgery in order to treat the hypercortisolism and its adverse sequelae (heart disease, mortality, consequences of inadequate diabetes or hypertension control), instead of perhaps life-long mifepristone? Though, the majority of patients did not have unilateral nodules, so perhaps more investigation (eg pituitary imaging) might reveal a surgically reversable cause??

    -- but,  some people with hypercortisolism will likely benefit from long-term mifepristone, which raises the issue of its long-term availability; there are strong efforts in our current political situation to limit its availability since it is used in medically-treated abortions (politics dominating medicine, yet again...)

        --and this is an issue for people living in many states at this time:

 

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu 

to get access to all of the blogs:

 

 go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

 

if you would like to see the articles in this blog, please email me. 

 

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list