A comparison between the diagnostic value of gonadotropins, alpha-subunit, and chromogranin-A and their response to thyrotropin-releasing hormone in clinically nonfunctioning, alpha-subunit-secreting, and gonadotroph pituitary adenomas.

We tested the hypothesis of whether chromogranin-A (CGA), an immunohistochemical marker of neuroendocrine tumors, could serve as a serum marker for clinically nonfunctioning pituitary adenomas. Basal and TRH-stimulated concentrations of LH, FSH, alpha-subunit, and CGA were measured in 22 patients with clinically nonfunctioning pituitary adenomas and in 20 control patients with other pituitary tumors. The control group consisted of 9 patients with PRL- and/or GH-secreting adenomas and 11 patients with nonendocrine tumors [5 craniopharyngiomas, 2 (dys)germinomas, 1 astrocytoma, 1 meningioma, 1 neurinoma of the acoustic nerve, and 1 dermoid cyst]. Immunohistochemical staining for CGA was performed on tumor tissue obtained at transsphenoidal surgery in 18 study and 12 control patients. Tissue from 19 of the 22 clinically nonfunctioning adenomas was cultured, and concentrations of LH, FSH, alpha-subunit, and CGA were measured. Immunohistochemical staining for CGA was positive in 15 of 18 clinically nonfunctioning adenomas and negative in all examined control tumors (n = 12). CGA was present in the culture medium of 16 of 18 adenomas in vitro. In 3 adenomas it was present in the absence of detectable amounts of gonadotropins or alpha-subunit. Basal serum levels of gonadotropins and/or alpha-subunit were elevated in 7 of 22 patients with clinically nonfunctioning adenomas and in 4 of 9 control patients with PRL- and/or GH-secreting adenomas. Basal CGA was elevated in 2 study patients and 1 prolactinoma patient. Significant increases in serum gonadotropin and/or alpha-subunit levels in response to TRH occurred in 14 of 21 patients with clinically nonfunctioning adenomas and in 13 of 20 control patients. A significant CGA peak after TRH administration was demonstrated in 6 patients with clinically nonfunctioning pituitary tumors and in none of the controls. We conclude that 1) immunohistochemical staining for CGA is an excellent tool to prove the endocrine origin of clinically nonfunctioning pituitary tumors; 2) in vivo, the gonadotroph origin can be recognized in only a minority of patients who have elevated basal levels of LH, FSH, or alpha-subunit; 3) examination of the effect of TRH on CGA release is a rather insensitive, but specific, diagnostic test, allowing differentiation from nonendocrine pituitary tumors; and 4) the responses of gonadotropins and alpha-subunit to TRH, although more sensitive, are not specific for clinically nonfunctioning pituitary adenomas and are probably only reliable in cases of total hypopituitarism.