Donor–derived cell-free DNA associated with increased risk of chronic lung allograft dysfunction and mortality: Are absolute levels better than percentage?
Kellie Phipps, Hyesik Kong, Han Su, Alan Sanders, Michael Keller, Muhtadi Alnababteh, Pali Shah, Shambhu Aryal, Moon Kyoo Jang, Woojin Park, T. Andargie, J. B. Orens, Steven D. Nathan, Xin Tian, Sean Agbor-Enoh
Abstract
BACKGROUND: There is a need for reliable methods to risk-stratify patients for chronic lung allograft dysfunction (CLAD) and death. A proof-of-concept study reported the predictive utility of early post-transplant levels of percent donor-derived cell-free DNA (dd-cfDNA), a measure affected by host cell-free DNA changes. This study assesses the predictive utility of dd-cfDNA in and beyond the early post-transplant period using absolute dd-cfDNA, a measure not affected by host cell-free DNA changes. METHODS: This multicenter study (Genome Research Alliance for Transplantation, NCT02423070) of lung transplant patients measured serial absolute dd-cfDNA in copies/ml and percent dd-cfDNA. We performed multivariable Cox regression analyses to determine the association between dd-cfDNA and the composite outcome of moderate-to-severe CLAD and/or death; the early post-transplant model used mean dd-cfDNA levels within days 30 to 91. A time-dependent model used dd-cfDNA values beyond the early post-transplant period. RESULTS: The cohort included 242 patients and 2,838 serial dd-cfDNA assessments. Absolute dd-cfDNA had similar or better performance than percent dd-cfDNA to risk-stratify patients for moderate-to-severe CLAD/death. High absolute dd-cfDNA levels in the early post-transplant period (hazard ratio 1.29, 95% confidence interval 1.04-1.60, p = 0.018) or at any time post-transplant (hazard ratio 1.51, 95% confidence interval 1.32-1.73, p < 0.001) were associated with an increased risk of moderate-to-severe CLAD/death, and absolute dd-cfDNA levels >250 copies/ml were associated with an elevated risk of moderate-to-severe CLAD/death. CONCLUSIONS: The study reports dd-cfDNA risk assessment models for moderate-to-severe CLAD/death; absolute values showed similar or better predictive performance than percent.