AM Cortisol: Drugs That Distort This Test

What AM Cortisol Measures and Why the Draw Window Matters

AM cortisol captures the daily peak of the hypothalamic-pituitary-adrenal (HPA) axis. Cortisol follows a circadian rhythm that peaks between 6:00 and 8:00 AM and drops to its nadir near midnight, a pattern first characterized in detail by Weitzman and colleagues [1]. The standard blood draw occurs between 7:00 and 9:00 AM for this reason.

A result below 3 mcg/dL at 8:00 AM strongly suggests adrenal insufficiency, while a value above 18 mcg/dL generally excludes it. Values between 3 and 18 mcg/dL land in an indeterminate zone that typically triggers a cosyntropin stimulation test [2]. The 2016 Endocrine Society Clinical Practice Guideline for adrenal insufficiency states: "A serum cortisol level measured between 0800 and 0900 h that is clearly below the reference range is supportive of adrenal insufficiency, but a value within the reference range does not exclude partial insufficiency" [2].

Most hospital immunoassays measure total cortisol, which is roughly 90% bound to CBG and 10% free. Any drug that shifts CBG concentration, alters adrenal enzyme activity, or suppresses ACTH will move the number without changing true adrenal function. The clinical trap is straightforward: the test looks abnormal, the patient gets worked up (or, worse, started on replacement steroids), and the real culprit was a medication list nobody reviewed at the time of the order.

Exogenous Glucocorticoids: The Single Biggest Distorter

Prednisone, dexamethasone, hydrocortisone, methylprednisolone, and inhaled or topical steroids all suppress the HPA axis by negative feedback on CRH and ACTH. The result is a falsely low AM cortisol.

Duration matters. A 5-day burst of prednisone 40 mg may suppress the axis for only a few days. Chronic use at doses above 7.5 mg prednisone-equivalent for longer than 3 weeks, however, can suppress the axis for months [3]. A 2015 meta-analysis by Broersen et al. (37 studies, N = 3,753) found that adrenal suppression occurred in 48.7% of patients on any form of exogenous glucocorticoid therapy, including 7.8% of those on inhaled corticosteroids alone [4].

The practical problem: some steroid formulations cross-react with cortisol immunoassays. Hydrocortisone is chemically identical to cortisol and will register directly on the assay. Prednisolone shows variable cross-reactivity depending on the platform (Roche Elecsys, Abbott Architect, Siemens Immulite each behave differently) [5]. If a patient cannot stop steroids before testing, a liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay eliminates cross-reactivity and should be specifically ordered.

Hold guidance for exogenous steroids before AM cortisol testing:

• Short course (under 2 weeks): hold 48 to 72 hours before draw.
• Moderate course (2 to 4 weeks): hold at least 1 week; consider cosyntropin stim test instead.
• Chronic use (over 4 weeks): taper under physician supervision, then test at least 24 hours after last physiologic-dose hydrocortisone or 48 hours after last prednisone dose.
• Inhaled / high-potency topical: hold 24 hours if possible; if not, note on the lab requisition.

Estrogen and Oral Contraceptives: The Overlooked CBG Effect

Estrogen-containing oral contraceptives (OCPs), hormone replacement therapy (HRT), and pregnancy all raise CBG by stimulating hepatic production. Because most assays measure total cortisol (bound plus free), a higher CBG means a higher total cortisol reading without any true increase in biologically active free cortisol [6].

The magnitude is not subtle. A 1999 study by Meulenberg and Hofman found that women on combined OCPs had mean morning total cortisol levels approximately 75% higher than matched controls not on OCPs, while salivary free cortisol (which reflects unbound cortisol) was not significantly different [7]. This means an OCP user with a "normal" pre-treatment cortisol of 12 mcg/dL could return a value of 21 mcg/dL, prompting an unnecessary Cushing syndrome workup.

The Endocrine Society's 2008 guideline on Cushing syndrome screening explicitly recommends against using total serum cortisol as a first-line screening test in women on estrogen-containing medications [8]. Dr. Lynnette Nieman, Senior Investigator at the National Institutes of Health, has stated: "Oral estrogen is probably the most common cause of a falsely elevated cortisol measurement in premenopausal women. If you can't stop it, measure salivary cortisol or urinary free cortisol instead" [8].

Bottom line: hold OCPs or oral estrogen for 6 weeks before drawing AM cortisol if the question is Cushing syndrome. If the question is adrenal insufficiency (low cortisol concern), the estrogen effect artificially raises the value, so a low reading on OCPs is actually more concerning, not less.

Opioids and the Suppressed Axis

Chronic opioid use suppresses the entire HPA axis at the hypothalamic level, reducing both CRH and ACTH release. The downstream effect is a genuinely low cortisol. This is not an assay artifact. It is a real pharmacologic suppression of adrenal output.

A 2013 study by Li et al. (N = 109 chronic pain patients on long-term opioids) found that 29% had an 8 AM cortisol below 5 mcg/dL [9]. Morphine, methadone, and fentanyl appear to carry the highest risk. Buprenorphine, a partial agonist, may be less suppressive, though data remain limited [10].

The clinical challenge: these patients often have fatigue, nausea, and orthostatic symptoms that overlap completely with adrenal insufficiency. A low AM cortisol in a chronic opioid user could represent true opioid-induced adrenal insufficiency (OIAI) or could simply reflect the pharmacologic suppression that will reverse if the opioid is tapered. AACE recommends cosyntropin stimulation testing to distinguish the two [11]. A normal stimulated cortisol response (peak above 18 mcg/dL at 30 or 60 minutes) suggests the adrenals can still respond when pushed, even if basal output is suppressed.

Antiepileptics and CYP3A4 Inducers

Phenytoin, carbamazepine, phenobarbital, and rifampin induce CYP3A4 and CYP3A5, the primary enzymes responsible for cortisol metabolism in the liver. Faster metabolism means lower circulating cortisol levels despite normal adrenal production [12].

Phenytoin is the best-studied offender. A study by Putignano et al. found that patients on chronic phenytoin had mean AM cortisol levels 35% lower than age-matched controls [12]. This effect also extends to the dexamethasone suppression test (DST): phenytoin accelerates dexamethasone clearance, leading to false-positive DST results (cortisol fails to suppress because the dexamethasone was metabolized too quickly, not because the patient has Cushing syndrome) [13].

Rifampin presents the same problem. Patients being evaluated for adrenal disorders while on rifampin for tuberculosis treatment will have artificially low basal cortisol and unreliable DST results. The Endocrine Society recommends using a 48-hour low-dose DST with measurement of dexamethasone levels, or switching to CRH stimulation testing, when CYP3A4 inducers cannot be discontinued [8].

Drugs That Directly Inhibit Adrenal Steroidogenesis

Three medications block cortisol production at the enzymatic level within the adrenal gland itself.

Ketoconazole inhibits multiple cytochrome P450 enzymes in the steroidogenic pathway, including CYP11A1, CYP17A1, and CYP11B1 [14]. This effect is so reliable that ketoconazole is used therapeutically to lower cortisol in Cushing syndrome, with FDA approval under the brand name Isturisa (osilodrostat is a separate agent; ketoconazole is used off-label in the U.S. but approved in Europe as Ketoconazole HRA). A dose of 400 to 1 to 200 mg/day can reduce cortisol by 50% or more [14]. Even antifungal-dose ketoconazole (200 mg/day) may lower AM cortisol enough to confuse adrenal testing.

Etomidate, the anesthetic induction agent, inhibits 11β-hydroxylase. A single induction dose suppresses cortisol for 12 to 24 hours. This was demonstrated dramatically in a 1983 study by Ledingham and Watt, which linked continuous etomidate sedation in ICU patients to increased mortality from adrenal suppression [15]. Any AM cortisol drawn within 24 hours of etomidate administration is unreliable.

Mitotane, used for adrenocortical carcinoma, is directly adrenolytic. It destroys adrenal cortical cells and also raises CBG, creating a double distortion: true cortisol production falls while measured total cortisol may appear deceptively preserved due to elevated CBG [16]. Free cortisol or salivary cortisol is mandatory for monitoring patients on mitotane.

Megestrol Acetate and Synthetic Progestins

Megestrol acetate, commonly prescribed as an appetite stimulant in cancer and HIV-related cachexia, has intrinsic glucocorticoid receptor activity. At doses of 400 to 800 mg/day, it suppresses the HPA axis in a manner similar to exogenous prednisone [17]. A study by Mann et al. found that 30 of 46 patients (65%) on megestrol 800 mg/day developed biochemical adrenal suppression, defined as an AM cortisol below 5 mcg/dL or a subnormal cosyntropin stimulation response [17].

This is frequently missed. Megestrol is categorized as a progestin, not a steroid, and clinicians may not think to list it alongside prednisone or dexamethasone when reviewing medications before an AM cortisol draw. The effect is dose-dependent and reversible after discontinuation, but recovery may take 2 to 4 weeks.

Psychiatric Medications: SSRIs, Antipsychotics, and Mifepristone

Selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs) modestly activate the HPA axis, particularly in the first 4 to 8 weeks of treatment. A meta-analysis by Lenze et al. found a transient cortisol increase of approximately 10 to 15% after SSRI initiation, which typically normalizes by 12 weeks [18]. This is rarely enough to trigger a false Cushing diagnosis, but it can nudge borderline values upward.

Atypical antipsychotics (olanzapine, quetiapine, risperidone) have variable effects. Some increase cortisol acutely; others suppress it with chronic use. The clinical relevance for AM cortisol interpretation is modest compared to the drug classes above.

Mifepristone (Korlym), a glucocorticoid receptor antagonist approved for hyperglycemia in Cushing syndrome, creates a paradox: it blocks cortisol action at the receptor, so the body compensates by massively increasing ACTH and cortisol production. AM cortisol levels can rise 2 to 3 fold during mifepristone treatment [19]. A high AM cortisol on mifepristone does not mean the patient has uncontrolled Cushing syndrome. It means the drug is working. Clinical response (blood glucose, weight, blood pressure) must be used to monitor efficacy, not cortisol levels [19].

How to Get an Accurate AM Cortisol Despite Polypharmacy

When drugs cannot be stopped, the testing strategy needs to change. Total serum cortisol is the most drug-vulnerable cortisol measurement. Several alternatives resist specific interference patterns.

Salivary cortisol reflects free (unbound) cortisol and is unaffected by CBG changes from estrogen or mitotane [20]. Late-night salivary cortisol is the standard screening test for Cushing syndrome precisely because it bypasses CBG interference. However, salivary cortisol can still be affected by drugs that alter cortisol production itself (ketoconazole, opioids, exogenous steroids).

24-hour urinary free cortisol (UFC) measures the fraction of cortisol that escapes protein binding and is filtered by the kidneys. Like salivary cortisol, it is CBG-independent but still affected by drugs that change true cortisol output [8].

Cosyntropin stimulation test assesses adrenal reserve by administering synthetic ACTH (250 mcg IV) and measuring cortisol at 0, 30, and 60 minutes. A peak cortisol above 18 mcg/dL indicates intact adrenal function regardless of what the basal AM cortisol shows [2]. This test is the gold standard when basal cortisol is uninterpretable due to medication effects.

LC-MS/MS cortisol assay eliminates cross-reactivity from exogenous steroids. If the concern is that prednisone or hydrocortisone is interfering with the immunoassay, specify LC-MS/MS on the order.

Decision path for the ordering clinician: If the patient is on estrogen or mitotane, order salivary or urinary free cortisol. If the patient is on exogenous glucocorticoids, use LC-MS/MS or cosyntropin stimulation. If the patient is on opioids, draw the AM cortisol but follow with cosyntropin stimulation if the result is low. If the patient is on a CYP3A4 inducer and a DST is planned, check dexamethasone levels concurrently or switch to CRH stimulation.

Quick-Reference: Drug Classes and Direction of Distortion

Drugs that falsely lower AM cortisol: exogenous glucocorticoids, opioids, megestrol acetate, ketoconazole, etomidate, mitotane, metyrapone.

Drugs that falsely raise total AM cortisol: oral estrogen / OCPs (via CBG), mifepristone (via HPA feedback), SSRIs (modest, transient).

Drugs that accelerate cortisol clearance (lower levels, confound DST): phenytoin, carbamazepine, phenobarbital, rifampin, pioglitazone.

The minimum medication hold before an interpretable AM cortisol: 24 hours for short-acting opioids, 48 to 72 hours for short-course steroids, 6 weeks for estrogen-containing OCPs, and "discuss with endocrinology" for chronic steroids, mitotane, or megestrol. A morning cortisol drawn at 10:30 AM from a non-fasting patient on prednisone and oral contraceptives is, for diagnostic purposes, noise.