Drs. Avgerinos and Cutler [1] fail to address a central issue [2]: The diagnosis of the Cushing syndrome and the differential diagnosis of its causes do not depend on a single diagnostic test result but rather on the consistency of multiple test results in indicating a single cause. Matrix analysis addresses the following type of question: What is the probability that an individual patient has ectopic as opposed to pituitary ACTH secretion if his or her basal urine free cortisol excretion is greater than w µgrams/24 h and decreases by less than x% after high-dose dexamethasone administration, his basal midnight plasma ACTH level is greater than y pg/mL, and his plasma 11-deoxycortisol level increases by less than z% after metyrapone administration? An individual patient might fail to meet the authors' criteria for 100% specificity for any one test, but the combination of less stringent criteria for multiple tests may have equal or greater diagnostic power.

The authors report that evaluating only results of 24-hour urine collections with creatinine excretion within 10% yields cut-off criteria identical to those for the whole series. The condition of a patient whose actual excretion falls near the test cut-off point will be misdiagnosed if the patient's urine collection is incomplete or excessive [2].

Inferior petrosal sinus sampling does not differentiate normal persons, patients with pseudo-Cushing syndrome, or patients with primary adrenal Cushing syndrome and incomplete hypothalamic-pituitary-adrenal suppression from those with the pituitary Cushing syndrome, but merely shows (in patients with proven ACTH-dependent Cushing syndrome) whether the source of the ACTH is pituitary or nonpituitary. Surprisingly, the authors did not mention the high-dose dexamethasone suppression test result. The unilateral iodocholesterol uptake suggests that the adenoma had caused functional atrophy of the normal adrenal cortices; one might anticipate that the patient's cortisol secretion would not suppress to the degree the authors demand of patients with the pituitary Cushing syndrome. The iodocholesterol scan, moreover, does not reliably show all functioning adrenal adenomas.