Is measuring serum cortisol important in the work‑up of a patient with suspected Cushing’s syndrome?

Measuring Serum Cortisol in Cushing's Syndrome Work-Up

Measuring blood serum cortisol is NOT a first-line screening test for Cushing's syndrome, but it plays a critical role at specific points in the diagnostic algorithm—particularly for measuring ACTH levels to determine etiology and for midnight serum cortisol when late-night salivary cortisol is not feasible. 1, 2

Initial Screening: Serum Cortisol is NOT Recommended

The diagnostic work-up for Cushing's syndrome should begin with one or more of three validated first-line screening tests, none of which involve random serum cortisol measurement 2, 3:

• Late-night salivary cortisol (LNSC): Sensitivity 92-100%, specificity 93-100%, collected at bedtime on 2-3 consecutive nights 2, 4
• 24-hour urinary free cortisol (UFC): Sensitivity >90%, requires 2-3 complete collections due to 50% random variability 2, 5
• Overnight 1-mg dexamethasone suppression test (DST): Sensitivity >90%, with abnormal threshold ≥1.8 μg/dL at 8 AM 2, 5

Random morning serum cortisol has no role in screening because it overlaps significantly between normal subjects and Cushing's patients, even though Cushing's patients have elevated levels. 4 The circadian rhythm makes timing critical, and random measurements lack diagnostic utility.

When Serum Cortisol IS Essential: Determining Etiology

Once hypercortisolism is confirmed by abnormal screening tests, measuring morning plasma ACTH is absolutely essential to differentiate ACTH-dependent from ACTH-independent causes 1, 2:

• Low/undetectable ACTH (<5 ng/L or <1.1 pmol/L): Indicates ACTH-independent Cushing's syndrome (adrenal source) → proceed to adrenal CT or MRI 1, 2
• Normal or elevated ACTH: Indicates ACTH-dependent Cushing's syndrome (pituitary or ectopic source) → proceed to pituitary MRI 1, 2

This single ACTH measurement fundamentally directs the entire subsequent work-up and cannot be omitted.

Midnight Serum Cortisol: Alternative When LNSC Not Feasible

Midnight serum cortisol can be used as an alternative screening test when late-night salivary cortisol is not feasible, particularly in hospitalized patients 1, 2. Normal midnight serum cortisol should be <1.8 μg/dL (<50 nmol/L) in sleeping individuals 5. However, this requires inpatient admission for proper collection, making it less practical than LNSC for outpatient screening 2.

Post-Dexamethasone Serum Cortisol: Part of DST

Serum cortisol measurement the morning after dexamethasone administration is integral to the DST, one of the three first-line screening tests 2, 5. This is not a "random" serum cortisol but a specific measurement at 8 AM following 1 mg dexamethasone at midnight, with failure to suppress below 1.8 μg/dL indicating possible Cushing's syndrome 2, 5.

Measuring dexamethasone levels concomitantly with cortisol improves test interpretability by confirming adequate drug absorption and excluding false-positive results from abnormal metabolism. 1, 2

Critical Pitfalls to Avoid

• Never use random serum cortisol for screening: The overlap between normal and pathologic values renders it useless for initial diagnosis 4
• Oral contraceptives falsely elevate total serum cortisol: Estrogen increases cortisol-binding globulin (CBG), raising total cortisol while free cortisol remains normal—this is the most important clinical pitfall 5
• Night-shift workers should not undergo midnight cortisol testing: Disrupted circadian rhythm invalidates the test 2, 5
• Pseudo-Cushing's states (severe obesity, depression, alcoholism, uncontrolled diabetes) can cause mildly elevated cortisol and false-positive results 1, 2

The Diagnostic Algorithm

1. Screen with 2-3 first-line tests (LNSC, UFC, or DST)—not random serum cortisol 2, 3
2. If screening tests abnormal, repeat 1-2 tests to confirm 1, 2
3. Once confirmed, measure morning ACTH to determine etiology 1, 2
4. Proceed based on ACTH results: Low ACTH → adrenal imaging; Normal/high ACTH → pituitary MRI 1, 2

Why Multiple Tests Are Necessary

Episodic Cushing's syndrome is highly prevalent in mild disease, with 65 of 66 confirmed Cushing's patients having at least one normal test result. 6 The probability of having Cushing's syndrome despite one negative test was 92% for LNSC, 88% for UFC, and 54% for nighttime plasma cortisol 6. This underscores why multiple tests on separate occasions are required, especially when clinical suspicion remains high despite initial normal results 6, 7.