Higher Banff inflammation and chronicity scores on kidney transplant biopsies are associated with poorer graft survival, although histology alone has limitations in predicting outcomes. We investigated if integrating donor-derived cell-free DNA (dd-cfDNA, Allosure; CareDx, Inc.) with Banff biopsy scores into a predictive model for estimated glomerular filtration rate over time can improve prognostic assessment versus histology alone.
Methods: We identified 180 kidney transplant patients with dd-cfDNA assessed within 1 mo of biopsy. Using linear mixed-effects models, a prediction model of Banff histology scores and dd-cfDNA on estimated glomerular filtration rate over time was derived. Nested models were compared using the likelihood-ratio test, Akaike Information Criterion, and Bayesian Information Criterion to assess if inclusion of dd-cfDNA into a model consisting of Banff biopsy scores would improve model fit.
Results: Univariate models identified significant covariate-by-time interactions for cg = 3 versus <3 (coefficient: -1.3 mL/min/1.73 m2/mo; 95% confidence interval [CI], -2.4 to -0.2; P = 0.02) and ci + ct ≥ 3 versus <3 (coefficient: -0.7 mL/min/1.73 m2/mo; 95% CI, -1.3 to -0.1; P = 0.03) and a trend toward significant covariate-by-time interaction for dd-cfDNA (coefficient: -0.5 mL/min/1.73 m2/mo; 95% CI, -1.0 to 0.1; P = 0.08). Addition of acute inflammation (i, t, and v), microvascular inflammation (g and ptc), and inflammation in area of interstitial fibrosis and tubular atrophy scores to chronicity scores (cg ≥ 3 and ci + ct ≥ 3) did not improve model fit. However, a model including dd-cfDNA with cg and ci + ct with covariate-by-time interactions had a better model fit compared with cg and ci + ct alone (likelihood-ratio test statistic = 21.1; df = 2; P < 0.001).
Conclusions: Addition of dd-cfDNA to Banff biopsy scores provided better prognostic assessment over biopsy characteristics alone.
Copyright © 2020 The Author(s). Transplantation Direct. Published by Wolters Kluwer Health, Inc.