In normal subjects the main form of circulating insulin-like growth factor (IGF) is the 150-kDa complex. This complex is formed by the IGF peptide, the acid-stable IGF-binding protein-3 (IGFBP-3), and the acid-labile subunit (ALS). Experimental and clinical data have demonstrated that ALS is primarily under the control of GH and plays a critical role in maintaining constant levels of circulating IGF-I. In this study we evaluated ALS, IGF-I, and IGFBP-1, -2, and -3 in 45 acromegalic patients in basal conditions and, in 37 of these, twice after surgical therapy compared with 100 age- and sex-matched control subjects to estimate their value as parameter of GH secretory state. The results demonstrated that in acromegaly before treatment all parameters (ALS, 523 +/- 26; IGF-I, 129 +/- 6; IGFBP-1, 0.7 +/- 0.1; IGFBP-3, 234 +/- 21; nmol/L; mean +/- SEM) but IGFBP-2 were significantly different (P<0.0001) from those in healthy subjects (ALS, 281 +/- 4; IGF-I, 22 +/- 1; IGFBP-1, 1.6 +/- 0.1; IGFBP-3, 91 +/- 3). IGF-I was more sensitive (100%) than ALS (89%), and both were more predictive of disease status than IGFBP-3, in that 27% of the patients had IGFBP-3 levels within the normal range. Considering the ALS/IGFBP-3 molar ratio, almost 55% of ALS circulated in a free form in active acromegaly. Before treatment, the IGF-I/IGFBPs (-1 + -2 + -3) molar ratio, which can be regarded as free, biologically active, IGF-I, was greatly increased (0.77 +/- 0.06; P<0.0001) compared with that in control subjects (0.23 +/- 0.01). After surgery, all 10 patients with controlled disease showed normalization of ALS (100% sensitivity), whereas 9 of them had normal IGFBP-3; reevaluation after varying lengths of time showed all these parameters within the normal range. In the 27 patients with active disease, IGF-I and ALS were more predictive of disease status (91% and 83% negative predictive values, respectively) than IGFBP-3 (53%). The basal ALS concentration correlated only with IGFBP-3 (r = 0.70; P<0.001). In postsurgery samples (first control) a statistically significant (P<0.001) correlation was found between mean GH values as well as minimum GH after oral glucose tolerance test and ALS (r = 0.72 and 0.83, respectively), IGF-I (r = 0.69 and 0.77), IGFBP-3 (r = 0.50 and 0.72), and IGFBP-2 (r = -0.36 and -0.63). Similarly, IGF-I, IGFBP-3, and ALS were positively correlated among themselves and negatively correlated with IGFBP-2 (P<0.001). In conclusion, in the diagnosis of acromegaly, the measurement of total IGF-I appears to be the most sensitive parameter among the subunits of the 150K complex, and IGFBP-3 the least sensitive. For ALS, this subunit is quite sensitive and appears to be a useful parameter in reassessment after surgical treatment.