OBJECTIVE: The aim was to evaluate the validity of current remission criteria in acromegaly, a random GH level of <2.5 microg/l, a glucose-suppressed GH level of <1 microg/l and a normal IGF-I level. DESIGN: In forty-one patients treated for acromegaly (23 males and 18 females, 20-69 years) and 94 healthy subjects (50 males and 44 females, 20-78 years), basal GH and IGF-I levels and nadir GH levels after 75 g oral glucose were evaluated in decade blocks; these were assayed by sensitive immunoradiometric assays. RESULTS: Basal GH levels varied widely from 0.022 to 10.4 in healthy subjects and were >2.5 microg/l in 19%. The mean post-glucose GH nadir was 0.067+/-0.009 microg/l (range 0.003-0.4 microg/l) and the upper limit of the GH nadir was 0.26 microg/l (means+2 S.D.) in healthy subjects. Thirty-five patients with acromegaly had high-for-age IGF-I levels in relation to our healthy subjects. In this group, 15 (42.9%) patients had basal GH levels of <2.5 microg/l, 14 (40%) patients had nadir GH levels of <1 microg/l, and three (8.6%) patients had GH suppression to <0.26 microg/l which was defined as normal GH suppression in our healthy subjects. Only six patients with acromegaly had normal-for-age IGF-I levels and all of these patients had basal GH levels of <2.5 microg/l and all but one had nadir GH levels of <0.26 microg/l. CONCLUSIONS: A basal or random GH level of <2.5 microg/l is not a reliable criterion for remission in acromegaly and the currently accepted normal upper limit of 1 microg/l for post-glucose GH suppression is too high. Post-glucose nadir GH levels, measured with sensitive assays, can be <1.0 microg/l in 40% and basal GH levels can be <2.5 microg/l in 43% of the active acromegalic patients. IGF-I levels appeared to correlate better with a nadir GH cut-off of 0.26 microg/l rather than 1 microg/l in the determination of disease activity.