Optimizing blood sampling protocols in patients with acromegaly for the estimation of growth hormone secretion

Context: Optimal blood sampling schemas of GH for the estimation of the 24-hour secretion rate have not been established in acromegalic patients. Objective: By censoring available 24-hour GH serum profiles, we investigated the reliability of such simplified schemas. Design, Subjects, and Methods: We used 24-hour serum GH concentration profiles obtained with 10-minute sampling in a large cohort of healthy subjects (n = 130; mean age, 42; range, 18-77 years) and acromegalic patients (n = 87; mean age, 48; range 18-72 years). Patient categories were active disease, surgically cured, and somatostatin analog-treated individuals. The regression coefficients of determination (R2) and the linear slopes (=) between 24-hour secretion rates or mean concentrations (144 samples) on the one hand and mean values with less frequent sampling on the other, decreased in controls and in patients a short (1-2 weeks) or long (2-5 years) time after successful surgery. By contrast, the regression parameters remained essentially unchanged in patients with active acromegaly and those under GH suppressive treatment. Excellent correlations (R2-0.90) without GH underestimation existed between mean GH of daytime profiles and mean GH of 24-hour profiles and GH secretion rates estimated by deconvolution in patients with active acromegaly and patients treated with somatostatin analogs. Conclusion: Simplified sampling schemes, particularly a day profile, can be used for the estimation of GH secretion in patients with active acromegaly and under medical treatment. However, in healthy controls and patients after successful surgery, prolonged and frequent sampling schemes, at least at 2-hour intervals, reliably reflect 24-hour secretion.