#1390 Differences in the detection of anti-HLA antibodies between two single antigen bead methods in kidney transplantation patients

Abstract

Abstract Background and Aims Two semiquantitative single-antigen bead (SAB) methods are currently available for the detection of anti-HLA antibodies and the estimation of their activity, depending on complement binding. However, data on accordance of the two methods on specificities detection and their reactivity evaluation are scarce. The purpose of this study is to compare the two methods in terms of anti-HLA antibodies specificities detection and of their complement binding capacity Method Sera from 97 patients with a positive Panel Reactive Antibody Test (>5%) was tested by two SAB methods, Immucor (IC) and One-lambda (OL), for class I and II anti-HLA antibody specificities and their ability to bind complement. The patients were either on a waiting list for kidney transplantation or transplanted patients tested routinely or after a clinical event. Statistical analysis of the results and comparison of the two methods followed. Results OL detected more class I positive specificities per patient [25 (9-34) vs. 18 (7-28) with IC, p < 0.001], while the two methods showed no difference in class II anti-HLA antibodies. IC detected 1608 out of 8148 and 1136 out of 7275 totally examined, positive anti-HLA class I and II antibody specificities respectively, with a median mean fluorescence intensity (MFI) of 4195 (1995-11272) and 6706 (2647-13184) respectively. Similarly, OL detected 1942/8148 and 1247/7275 positive specificities of anti-HLA class I and II antibodies respectively, with MFI 6185 (2855-12099) and 9498 (3630-17702)] respectively. In the IC method, 428/1608 [MFI 13900 (9540-17999)] and 409/1136 [MFI 11832 (7128-16531)] positive class I and II specificities, respectively, bound the C3d fraction of complement. Similarly, OL detected 485/1942 [MFI 15452 (9369-23095)] and 298/1247 [MFI 18852 (14415-24707)] class I and II specificities respectively that bound the C1q fraction of complement. In the analysis of the specificities that were positive by both methods, MFI showed a strong positive linear correlation between the two methods (r = 0.61, p < 0.001 and r = 0.54, p = 0.001 for classes I and II respectively). Accordingly, data were analyzed with Principal Components Analysis, in which negative results were also included. In the projection of the plane defined by the first two principal components the vector corresponding to the IC method for class I antibodies was congruent with the vectors of the complement binding detection methods, C3d and C1q. In contrast, the OL vector was nearly perpendicular to the other vectors, indicating the existence of non-complement-binding specificities that are not detected by IC. A similar picture was found for antibodies against HLA class II (Figure). Conclusion OL was more sensitive for detecting class I and II anti-HLA antibodies despite the fact that there was no difference in the number of class II specificities per patient. However, the additional specificities detected by OL, potentially do not bind complement. The MFI of the complement binding specificities were higher in both methods. The two methods were equivalent in detecting complement-binding class I specificities, however the C3d IC method was more sensitive in detecting complement-binding class II anti-HLA antibodies.