B-168 Establishment of reference intervals of serum immunoglobulin and complement by Hoffmann and refineR algorithms and evaluation of their feasibility

Abstract Background The Chinese health industry standard, WS/T 645, provide reference intervals of serum immunoglobulin G (IgG), IgA, IgM, complement 3 (C3) and C4 in adults. These RIs were established using a direct sampling method that involved recruiting apparently healthy individuals, which is time-consuming and costly. Contrarily, indirect methods utilize statistical techniques to delineate the data distribution of the "healthy population" from pooled patient data to establish the RI. This study has calculated the adult reference intervals of Ig G, IgA, IgM, C3 and C4 using Hoffmann and refineR methods, and compared them with the reference interval reported by WS/T 645 to explore the clinical application value of indirect methods. Methods The physical examination data were collected at Peking Union Medical College Hospital from 2013 to 2023. The Box-Cox was used to normalize the data distribution, the Tukey method was used to remove outliers. Mann-Whitney test and Kruskal-Wallis test were used to compare the differences between gender and age, respectively. Quantile regression analysis was used to examine the effects of age and gender. Hoffmann and refineR methods were employed to calculate RIs of the five parameters. Results There were 82 251, 82 483, 49 236, 20 027 and 19 942 test results for IgG, IgA, IgM, C3 and C4, respectively. Gender was significantly correlated with IgG, IgA, IgM, C3, and C4 levels, and age was significantly correlated with IgA, IgM, C3, and C4 levels. According to Hoffmann and refineR methods, the total reference interval of IgG was 7.58∼16.85 g/L and 7.54∼16.88 g/L; IgA 0.97∼4.26 g/L and 0.97∼4.34 g/L; IgM for women 0.44∼2.57 g/L and 0.44∼2.62 g/L, for men 0.33∼1.91 g/L and 0.31∼1.87 g/L; C3 0.73∼1.43 g/L and 0.73∼1.44 g/L; C4 0.11∼0.32 g/L. Gender – and/or age-specific reference intervals were further calculated according to the bias ratio values. The reference intervals calculated by Hoffmann and refineR are in good agreement. The upper and lower limits of reference intervals for IgA and C3, the lower limits of IgG, the upper limits of IgM and C3 calculated by indirect method have good comparability with those reported by standards. Conclusion The total and gender - and/or age-specific reference intervals of IgG, IgA, IgM, C3 and C4 in our hospital population were established by Hoffmann and refineR methods, and the reference intervals established by the two indirect methods were comparable. The reference intervals calculated using indirect methods were also comparable with those established using the direct sampling method reported in current health standards.