Insulin-like growth factor-binding protein-3 (IGFBP-3) is a GH-dependent protein that binds insulin-like growth factor-I (IGF-I) in the circulation and modulates its action at the tissue level. Because IGFBP-3 is a stable and specific marker of somatotroph function, we hypothesized that it would be a useful biochemical marker in the diagnosis of patients with acromegaly and the assessment of surgical cure. We, therefore, investigated the sensitivity of serum IGFBP-3 levels to detect GH excess in 44 patients with clinical acromegaly and pathologically confirmed somatotroph adenomas, including a cohort of 18 patients with untreated disease evaluated before transsphenoidal surgery and medical therapy. IGFBP-3 levels were compared to IGF-I and random GH levels before and after transsphenoidal surgery. Concordance among IGFBP-3, IGF-I, and GH suppressibility by glucose was also determined. In addition, the response of IGFBP-3 to glucose suppression was investigated. All 18 patients with untreated acromegaly had elevated serum IGFBP-3 levels, ranging from 4,186-10,026 micrograms/L (mean +/- SD, 6566 [plusm] 1800 micrograms/L). There was no overlap with the age-adjusted normative ranges (P = 0.0001 in patients 18-55 yr old and P = 0.0176 in patients > 55 yr old) or with the levels obtained in age-comparable controls (P = 0.0001). In 11% of untreated patients with clinical findings of acromegaly and a pathologically confirmed adenoma, IG-FBP-3 levels were elevated, although GH was suppressed to less than 2 micrograms/L with glucose. In these patients, IGF-I levels were either normal or minimally elevated and considered nondiagnostic. IGFBP-3 and IGF-I levels were correlated in patients with untreated acromegaly (r = 0.650; P = 0.0162) and after transsphenoidal surgery (r = 0.644; P = 0.0001). Neither IGF-I nor IGFBP-3 correlated with random GH levels before surgery. However, both IGF-I (r = 0.471; P = 0.0001) and IGFBP-3 (r = 0.259; P = 0.041) correlated with random GH levels in patients studied more than 1 month after transsphenoidal surgery. IGFBP-3 and IGF-I levels were concordant with GH suppressibility by glucose (P = 0.0039) and IGFBP-3 decreased with glucose suppression in 7 of 10 patients. These data indicate that IGFBP-3 is a sensitive physiological marker of somatotroph function and is concordant with glucose suppression and IGF-I levels before and after transsphenoidal surgery.(ABSTRACT TRUNCATED AT 250 WORDS)