Impact of analytical variability in growth hormone (GH) and insulin-like growth-factor I (IGF-I) assays on clinical decisions in patients with acromegaly

Abstract

Serum concentrations of GH and IGF-I are used in diagnosis of acromegaly and monitoring of treatment efficacy. Problems with comparability of immunoassay results are acknowledged in published guidelines, but information on clinical impact of such problems is limited. In a nationwide study, we compared local GH- and IGF-I results to those from a central laboratory. Furthermore, we independently asked the participating physicians and the local labs about the assays the IGF-I reference ranges used. We also investigated the agreement between local classification of disease activity and IGF-I concentrations measured by the central lab. 23 centers participated in this study, sending 313 samples (125 male/188 female, mean 12 samples/center; range 1–69). Nichols Advantage GH and IGF-I assays were used by the central lab. Classification of disease activity by local physicians was „controlled“ (119 cases), „active“ (107), „unsure“ (35) or „not provided“ (52). Local labs reported 7 different assays for GH and 6 for IGF-I, Nichols Advantage and DPC Immulite 2000 being most frequently used. The information on the local assay method provided by the local physician was incorrect in most cases. The name of the assay manufacturer was correct in 207/313 samples for GH and 241/313 samples for IGF-I. Mean deviation between local and central assessment was -2.6% (range -187.2 to 188.0%) for GH and -3.4% (range -56.5 to 92.7%) for IGF-I. Regression analysis revealed overall correlation between methods (R2 0.77–0.97), but difference plots demonstrated a large scatter, especially at lower concentrations. The local judgement of disease activity was not supported by central lab IGF-I in 20.2% (locally controlled) and 14.0% (locally active) of cases. In conclusion, we observed limited correctness in assay methodology reported through local physicians. We confirm the dramatic discrepancies for GH and IGF-I results between labs. The clinical decision seems to be affected by the lack of assay standardization and poor quality of normative data in more than 30% of cases.