The impact of different calibration matrices on the determination of insulin-like growth factor 1 by high-resolution-LC-MS in acromegalic and growth hormone deficient patients

Abstract

ObjectivesCalibration is an important source of variability in liquid chromatography mass spectrometry (LC-MS) methods for insulin-like growth factor 1 (IGF-1). This study investigated the impact of different calibrator matrices on IGF-1 measurements by LC-MS. Moreover, the comparability of immunoassays and LC-MS was assessed. Design & MethodsCalibrators from 12.5 to 2009 ng/ml were prepared by spiking WHO international Standard (ID 02/254 NIBSC, UK) into the following matrices: native human plasma, fresh charcoal-treated human plasma (FCTHP), old charcoal-treated human plasma, deionized water, bovine serum albumin (BSA), and rat plasma (RP). A validated in-house LC-MS method was calibrated repeatedly with these calibrators. Then, serum samples from 197 growth hormone excess and deficiency patients were analysed with each calibration. ResultsThe seven calibration curves had different slopes leading to markedly different patient results. The largest differences in IGF-1 concentration from the median (interquartile range) was observed with the calibrator in water and the calibrator in RP (336.4 [279.6–417.0] vs. 112.5 [71.2–171.2], p < 0.001). The smallest difference was observed with calibrators in FCTHP and BSA (141.8 [102.0–198.5] vs. 127.9 [86.9–186.0], p < 0.049). Compared to LC-MS with calibrators in FCTHP, immunoassays showed relevant proportional bias (range: −43% to −68%), constant bias (range: 22.84 to 57.29 ng/ml) and pronounced scatter. Comparing the immunoassays with each other revealed proportional bias of up to 24%. ConclusionsThe calibrator matrix is critical for the measurement of IGF-1 by LC-MS. Regardless of the calibrator matrix, LC-MS shows poor agreement with immunoassays. Also, the agreement between different immunoassays is variable.