OBJECTIVE To assess the outpatient investigations, overnight urinary growth hormone (uGH) excretion, random serum GH and insulin like growth factor 1 (IGF-1), and GH indices from the oral glucose tolerance test (OGTT) (fasting, nadir and mean GH), as measures of mean GH secretion in treated acromegaly, in comparison with a GH day series, which served as a gold standard. METHODS Prospective cross-sectional study, with patients admitted to a metabolic ward for the following investigations: random GH, IGF-1, 6 point GH day series (day 1), 9 h timed overnight uGH excretion, OGTT with GH response (day 2). Agreements between the mean GH during the day series and the other outcome measures, and the diagnostic performance of the latter, for the presence or absence of active acromegaly (mean GH during day series > or = 5 or < 5 mU/l, respectively) were determined. METHODS 26 patients with treated acromegaly (11 with inactive acromegaly off drug therapy). METHODS Serum GH and uGH were measured by immunoradiometric assays and IGF-1 by radioimmunoassay. RESULTS Agreements with the mean GH during the day series were perfect for the nadir GH during the OGTT with a 2 mU/l cutoff (Cohen's kappa (kappa) = 1, P < 0.00001), almost perfect for the fasting and mean GH throughout the OGTT (both kappa = 0.92, P < 0.0001) and random GH (kappa = 0.85, P < 0.0001), and substantial for the nadir GH with a 5 mU/l cutoff (kappa = 0.77, P < 0.0001), IGF-1 (kappa = 0.62, P < 0.001) and overnight uGH excretion (kappa = 0.61, P = 0.002). Nadir GH with a 2 mU/l cutoff was completely accurate for diagnosing the presence or absence of active acromegaly (positive and negative predictive values (% +/- standard error percentage) 100 +/- 8% and 100 +/- 10%). None of the outpatient tests used alone was an adequate diagnostic test (positive and negative predictive values: overnight uGH excretion -86 +/- 10% and 75 +/- 13%; random GH -100 +/- 11% and 85 +/- 11%; IGF-1 -92 +/- 10% and 71 +/- 13%) and so combinations of tests were assessed. The best was overnight uGH excretion plus random GH (positive and negative predictive values 88 +/- 9% and 100 +/- 12%). Using all three outpatient investigations, the positive predictive value of three raised results was 100 +/- 13%. CONCLUSIONS In treated acromegaly, residual GH secretion can be reliably assessed with the OGTT, using standard diagnostic criteria. It can also be assessed on an outpatient basis with overnight uGH excretion and random GH, as direct measures, and IGF-1. If these are all normal, active acromegaly is excluded. Three raised results denote active acromegaly, and one or two raised results would need further investigation with a GH day series.