Donor-Derived Cell-Free DNA Combined with Histology Improves Prediction of eGFR Decline in Kidney Transplant Recipients over Histology Alone
Cedars-Sinai, Los Angeles, CA
Meeting: 2020 American Transplant Congress
Abstract number: A-295
Keywords: Biopsy, Kidney transplantation, Prediction models, Renal function
Session Information
Session Name: Poster Session A: Biomarkers, Immune Assessment and Clinical Outcomes
Session Type: Poster Session
Date: Saturday, May 30, 2020
Session Time: 3:15pm-4:00pm
Presentation Time: 3:30pm-4:00pm
Location: Virtual
*Purpose: Higher Banff inflammation and chronicity scores on kidney transplant (KTx) biopsies (Bx) have been associated with poorer graft survival, but Bx scores alone imperfectly predict outcomes. Incorporating rejection biomarkers may improve prognostic assessment but has yet to be studied. We investigated if integrating donor-derived cell-free DNA (dd-cfDNA, Allosure©; CareDx, Inc.; Brisbane, CA) with Banff Bx scores into a predictive model for eGFR decline can improve model fit vs. Bx scores alone.
*Methods: We identified 185 pts >/= 1 month post-KTx who had dd-cfDNA assessed within one month of Bx (median 13 days, IQR: 6-19 days). We used linear mixed effects models to derive a prediction model of Banff inflammation and chronicity scores and dd-cfDNA on eGFR (CKD-EPI) over time. Nested models were compared using the likelihood ratio test, AIC, and BIC to assess if inclusion of dd-cfDNA into a parsimonious model consisting of Banff Bx scores would improve model fit.
*Results: Bxs were performed a median of 693 days post-KTx (IQR: 99-2132 days). Overall, 25 pts had isolated cell-mediated rejection (CMR), 58 had isolated antibody-mediated rejection (ABMR), 13 had mixed CMR/ABMR, and 89 had no rejection. Univariate models identified significant group x time interactions for cg=3 vs. cg<3 and ci+ct >/= 3 vs. ci+ct<3; statistical interaction with time was non-significant for i, t, v, i-IFTA scores, and log-transformed dd-cfDNA (ln dd-cfDNA) (Table 1). Addition of acute inflammation (i, t, and v-scores) and i-IFTA scores to chronicity scores (cg=3 and ci+ct >/= 3) did not improve model fit. However, a model including ln dd-cfDNA with cg and ci+ct had a better model fit compared to cg and ci+ct alone (Table 2).
*Conclusions: Inclusion of dd-cfDNA, but not inflammation scores, with chronicity scores into a predictive model improved prediction of eGFR decline vs. chronicity scores alone. While there was only a trend toward a significant association of dd-cfDNA and eGFR trajectory, the direction of the dd-cfDNA x time interaction suggests higher dd-cfDNA may be associated with eGFR decline and should be validated in larger datasets.
To cite this abstract in AMA style:
Gillespie M, Haas M, Lim K, Vo A, Peng A, Najjar R, Sethi S, Jordan S, Huang E. Donor-Derived Cell-Free DNA Combined with Histology Improves Prediction of eGFR Decline in Kidney Transplant Recipients over Histology Alone [abstract]. Am J Transplant. 2020; 20 (suppl 3). https://atcmeetingabstracts.com/abstract/donor-derived-cell-free-dna-combined-with-histology-improves-prediction-of-egfr-decline-in-kidney-transplant-recipients-over-histology-alone/. Accessed May 31, 2025.« Back to 2020 American Transplant Congress