Mean GH profile is more accurate than single fasting GH in the evaluation of acromegaly disease control during somatostatin receptor ligands therapy

Purpose This study aims to compare the accuracy of mean GH profile (GHP) < 2.5 ng/ml and single fasting GH (SGH) < 1 ng/ml in the evaluation of disease control in acromegaly patients during somatostatin receptor ligands (SRLs) therapy. Methods We retrospectively enrolled 100 acromegaly patients, 68 responder, and 32 partial responder to SRLs. Controlled disease has been defined as IGF-I levels within age-related normal limits, while partial response as pathological IGF-I values despite a reduction >= 50%. In all patients, GHP, SGH, IGF-I, and IGFBP-3 were evaluated. Results Median GHP levels (1.2 ng/ml, IQR 0.5-2.3 ng/ml) were lower (p = 0.001) than SGH (1.9 ng/ml, IQR 1.0-3.6 ng/ml). Accuracy of GHP was 81%, whereas that of SGH was 55%, with a Kappa index of 0.520 and 0.237, respectively. In multivariable analysis GHP (p = 0.002) and IGFBP-3 (p = 0.004), but not SGH, were independently associated with normal IGF-I levels. At receiver-operator characteristic curve (ROC) analysis GHP cut-off sensitivity and specificity were 94.1% and 50.0%, respectively, while SGH sensitivity and specificity were 35.3% and 93.7%, respectively. Finally, in obese patients the GH cut-off level (both as SGH and GHP) associated to good disease control was significantly different with respect to not obese ones. Conclusions GHP associates with IGF-I (and therefore with appropriate control of disease) with higher accuracy than SGH. When GH evaluation is needed, the measurement of mean GHP should be preferred and use of BMI-related cut-offs is suggested.