Histopathological Discordance with Sequentially Rising Donor-Derived Cell-Free DNA

Introduction: Plasma donor-derived cell-free DNA (dd-cfDNA) fraction is a promising, non-invasive predictive tool for allograft kidney rejection. An elevated dd-cfDNA fraction when combined with HLA de novo donor-specific antibody (dnDSA) can further be used to predict ABMR. We discuss a case where the dd-cfDNA and dnDSA inform decisions to seek a biopsy that is ultimately inconsistent with the predictive models for ABMR. Case Description: A 73-year-old male with ESKD due to Type 2 diabetes had a living, unrelated kidney transplant (KT). There were no pre-formed anti-HLA antibodies, and HLA A-B-DR mismatch was 2-2-1. Induction was with basiliximab. He started tacrolimus (TAC), mycophenolate (MMF) 750 mg BID and prednisone. Two months post KT, MMF was reduced to 500 mg BID due to leukopenia. Three months post KT, MMF was paused for neutropenic fever and then resumed at 250 mg BID. He developed low titer BK viremia. TAC trough goal was reduced to 4-6 mcg/L. The prednisone dose was reduced from 10 mg to 5 mg daily. The patient began serial checks of dd-cfDNA, with a dd-cfDNA fraction of 0.19%. 13 months post KT, the repeated dd-cfDNA fraction increased to 0.28% with a new finding of a single dnDSA, DQ2 (8000MFI). Although he had stable creatinine without proteinuria, there was concern for subclinical acute rejection. An allograft biopsy showed mild glomerulitis without evidence of acute rejection but with findings of mild glomerulitis. There was concern for microvascular injury of immunologic causes, so serial measurement of dd-cfDNA and dnDSA continued. 16 months post-KT, the dd-cfDNA rose to 2.1%. Although creatinine remained stable without proteinuria, a second kidney biopsy was done. There was focal, minimal peritubular capillary staining for C4d without morphological evidence of rejection. 28 months post-KT, creatinine has remained stable without proteinuria. Discussion: Although novel tools exist in predicting allograft kidney rejection, they are imperfect and require careful interpretation and application. There was marked discordance between the probability of active rejection, as indicated by the combination of elevated and sequentially rising fractions of dd-cfDNA and dnDSA, and the absence of histological, immunofluorescence, and electron microscopic evidence of rejection on two kidney allograft biopsies 4 months apart.