In patients with treated acromegaly, improved survival is associated with serum GH concentrations below 2 microgram/L (5 mU/L). A principal aim of therapy in acromegaly is to achieve a GH level less than 2 microgram/L, as such levels are thought to be "safe." However, such GH levels do not always equate with normalization of plasma insulin-like growth factor I (IGF-I), although epidemiological data linking survival or morbidity to IGF-I levels are at present lacking. The aims of this study were 1) to further define the nature of GH release in those acromegalic patients who achieve mean GH concentrations below 2 microgram/L post therapy, 2) to examine the effect of different therapeutic interventions on the 24-h GH profile (surgery alone or radiotherapy), and 3) to determine the relationship between the various characteristics of the 24-h GH profile and IGF-I production in acromegalic subjects who have achieved GH below 2 microgram/L. Spontaneous 24-h GH secretion was measured using both a conventional immunoradiometric assay (limit of detection, 0.4 microgram/L) and an ultrasensitive assay (limit of detection, 0.002 microgram/L). The GH data have been analyzed by several methods: 1) the pulse detection algorithm Cluster, 2) a distribution method for detection of peak [the observed concentration 95%, i.e. the threshold at or below which GH concentrations are assessed to be 95% of the time, as calculated by probability analysis (OC 95%)] and trough (OC, 5%) GH activity, 3) deconvolution analysis, and 4) approximate entropy analysis. GH was sampled every 20 min for 24 h, along with basal IGF-I and IGF-binding protein-3, in 21 treated acromegalic patients with a mean GH below 2 microgram/L [ACR; 9 women and 12 men; median age (range), 49 (31-76) yr] and 16 healthy controls [C; 6 women and 10 men; age, 50 (30-75) yr]. Mean 24-h serum GH concentrations were [median (range)]: ACR, 1.1 (0.04-1.5) microgram/L; C, 0.4 (0.02-3.3) microgram/L (P = 0.28). GH pulse frequency was: ACR, 11 (1-14)/24 h; C, 10 (8-18)/24 h (P = 0.41). In the GH profiles the mean heights of the GH peaks were: ACR, 1.2 (0.05-2.8) microgram/L; C, 0.8 (0.02-5.1) microgram/L (P = 0.91), and the mean GH valley nadirs were: ACR, 0.65 (0.03-1.1) microgram/L; C, 0.09 (0.01-1.8) microgram/L (P < 0.02). The OC 95% was: ACR, 1.0 (0.04-3.8) microgram/L; C, 1.0 (0.02-10) microgram/L (P = 0.65), and the OC 5% was: ACR, 0.09 (0.01-0.6) microgram/L; C, 0.01 (0.001-0.4) microgram/L (P < 0.001). The median IGF-I was: ACR, 227 (100-853) microgram/L; C, 156 (89-342) microgram/L (P < 0.005). Approximate entrophy values were: ACR, 1.06 (0.35-1.45); and C, 0.57 (0.27-1.19); P < 0.05. In the acromegaly group a significant positive correlation was found between IGF-I and the calculated GH secretory burst amplitude in the radiotherapy subset (r = 0.85; P < 0.0005) as well as between IGF-I and both the mean GH valley nadir (r = 0.60; P < 0.004) and the trough (OC 5%) GH activity for the acromegalic patients as a whole (r = 0.55; P < 0.02). We conclude that in treated acromegaly (GH, <2 microgram/L), 1) IGF-I (by approximately 50%) and basal GH secretion (by 5-fold) remain significantly elevated compared with control values despite similar mean 24-h GH concentrations; 2) the calculated GH secretory pulse amplitude, mean GH valley nadir, and OC 5% correlate positively with IGF-I; 3) the greater mean GH valley nadir and OC 5% in acromegalic patients compared with controls may account for the raised IGF-I; and 4) radiotherapy is unlikely to normalize the GH secretory pattern, which underlies the persisting elevated IGF-I levels.