Intervals for the new Brahms Chromogranin A (CGA) assay

neuroendocrine tumours. To improve analytical performance and decrease test turnaround time, we introduced the new fully automated Brahms CGA assay performed on Kryptor platform. As CGA results from different assays are not comparable, we performed a study for establishing reference intervals for the new assay. Methods: Fresh serum samples were obtained from 200 healthy blood donors and immediately measured for CGA. Shapiro-Wilk and Wilcoxon rank sum tests were employed to assess distribution of CGA values and compare groups, respectively. Multiple regression models were used to evaluate the influence of age and sex on CGA concentrations, including the interaction between the two factors. Results: 4 elevated CGA values were statistical outliers and in 3 of those individuals an interfering condition possibly increasing CGA was identified (hyperthyroidism, vitamin D supplementation, use of hormonal contraceptives). After their exclusion the remaining values from 196 subjects [99 males and 97 females; median age (range) 44 years (19-67)] were analyzed. Median (range) CGA concentration was 41.6 μg/L (15.9-146.2), with no gender-related difference. Although deviating from normal frequency distribution, the visual examination of data did not suggest log transformation. Regression analysis confirmed the lack of gender influence, showing however that CGA concentrations increased with age (P <0.001). The lack of biological interaction between age and sex excluded the hypothesis that menopausal status may influence CGA release. Aiming to decide if reference values should be partitioned by age, we compared CGA concentrations in subjects <45 (n=99) and ≥45 years (n=97). Higher CGA concentrations were found in older people (mean±SD: 49.1±18.6 μg/L vs. 41.8±19.4 μg/L, P=0.0006). Accounting for manufacturer’s declared imprecision at CGA range of 80-120 μg/L (CV <7%) and the estimated upper reference limit (97.5th percentile - URL) for subjects ≥45 (98.5 μg/L) and <45 (87.0 μg/L), we however decided to adopt a single URL for overall population (93.7 μg/L; 90%CI: 79.0- 114.1). Conclusions: In healthy subjects age but not gender may affect CGA release. However this does not appear to require agerelated reference limits.