P140Proper setting of cut-off value for the lymphocyte crossmatch by flow cytometry in kidney transplantation

Abstract

Aim The flow cytometric lymphocyte crossmatch (FCXM) has become a routine test performed prior to kidney transplantation. Many studies have reported that a proportion of transplant recipients experienced early graft loss despite a negative complement-dependent cytotoxicity (CDC) crossmatch test. The FCXM is able to detect Human Leukocyte Antigen (HLA) antibodies of low titer or of a non-complement binding nature that may be responsible for graft loss. It is important to report positive and negative results using appropriate cut-off values. We reviewed the results of FCXM with CDC and panel reactive antibody (PRA) and attempted to set clinically useful cut-off values for FCXM. Methods From January 2015 to October 2017, 518 kidney recipients’ results of FCXM, CDC, and PRA (Screening test, Panel kit, and single antigen bead assay, LABScreen Luminex kits, One Lambda, Canoga Park, CA, USA) were analyzed. The PRA test was performed simultaneously from 403 patients, and 159 patients who were negative for HLA antibodies in the all PRA tests. Fifty-seven patients were positive for CDC test results and 37 patients were both positive in T cells and B cells. In 20 patients, only B cells CDC results were positive, and 13 patients were negative for FCXM test results. Results The final interpretation of FCXM result was decided through CDC, PRA result and existence of donor specific antibody (DSA). T cell FCXM result was interpreted as negative in 451 patients and positive in 67 patients according to DSA. The appropriate cut-off value of the Mean fluorescent intensity (MFI) ratio using the ROC curve for 518 patients was 1.86 (sensitivity 100%, specificity 98.2%, Youden’s index 0.982). However, patients who underwent desensitization treatment due to DSA or ABO blood type incompatibility were included. The cut-off value obtained using the ROC curve for 446 patients who have not experienced desensitization treatment was 2.45 (sensitivity 100%, specificity 99.3%, Youden’s index 0.993). Generally used, Median + 2 SD value of MFI ratio of patients judged as negative FCXM was 1.60. Conclusions Based on the statistical method used, various cut-off values could be obtained. It is important to determine the clinically significant cut-off value based on the data accumulated in each laboratory, not the simple median value from negative patients.