Recommend against random cortisol level for AI diagnosis
STEP 1 — Gather Information
Collect clinical signs/symptoms suggestive of adrenal insufficiency (fatigue, hypotension, hyponatremia, hyperkalemia), review timing of last glucocorticoid dose, and note if a random cortisol test is under consideration.
STEP 2 — Rule In / Rule Out
If a random cortisol level is being considered for AI diagnosis, recommend against its use and proceed to obtain an 8–9 AM serum cortisol level; otherwise, proceed with appropriate diagnostic pathway.
STEP 3 — Classify or Stratify
Interpret the 8–9 AM cortisol: <3 µg/dL suggests AI; >15 µg/dL makes AI unlikely; values between 3–15 µg/dL require an ACTH stimulation test.
STEP 4 — Decide
If cortisol <3 µg/dL, diagnose AI and initiate glucocorticoid replacement; if >15 µg/dL, pursue alternative diagnoses; if 3–15 µg/dL, perform ACTH stim test and diagnose AI if peak cortisol ≤18.1 µg/dL (500 nmol/L).
Clinical Guardrails / Mimics / Pitfalls
Do not rely on random cortisol due to circadian variation, estrogen-induced CBG elevation, and glucocorticoid interference; ensure patient is off exogenous glucocorticoids for adequate washout before testing; ACTH stimulation test remains the diagnostic standard when morning cortisol is indeterminate.
Concrete Clinical Example
A 50‑year‑old woman presents with fatigue and hyponatremia; a random cortisol at 14:00 is considered. Instead, an 8 AM cortisol returns 2.5 µg/dL (<3), prompting an ACTH stim test with peak cortisol 10 µg/dL (<18.1), confirming AI; hydrocortisone 15–20 mg daily is started.
Source: Hormonal Replacement in Hypopituitarism in Adults: An Endocrine Society Clinical Practice Guideline, Endocrine Society, 2016, DOI:10.1210/jc.2016-2118