Background.
Donor-derived cell-free DNA (dd-cfDNA%) is a biomarker of early acute lung allograft dysfunction (ALAD), with a value of ≥1.0% indicating injury. Whether dd-cfDNA% is a useful biomarker in patients >2 y posttransplant is unknown. Our group previously demonstrated that median dd-cfDNA% in lung recipients ≥2 y posttransplant without ALAD was 0.45%. In that cohort, biologic variability of dd-cfDNA% was estimated by a reference change value (RCV) of 73%, suggesting that change exceeding 73% may be pathologic. In this study, we aimed to determine whether dd-cfDNA% variability or absolute thresholds are optimal for detecting ALAD.
Methods.
We prospectively measured plasma dd-cfDNA% every 3 to 4 mo in patients ≥2 y post–lung transplant. ALAD was defined as infection, acute cellular rejection, possible antibody-mediated rejection, or change in forced expiratory volume in 1 s >10%, and was adjudicated retrospectively. We analyzed area under the curve for RCV and absolute dd-cfDNA% and reported performance of RCV ≥73% versus absolute value >1% for discriminating ALAD.
Results.
Seventy-one patients had ≥2 baseline measurements of dd-cfDNA%; 30 developed ALAD. RCV of dd-cfDNA% at ALAD had a greater area under the receiver operator characteristic curve than absolute dd-cfDNA% values (0.87 versus 0.69, P = 0.018). Test characteristics of RCV >73% for ALAD diagnosis were sensitivity 87%, specificity 78%, positive predictive value 74%, and negative predictive value 89%. In contrast, dd-cfDNA% ≥1% had sensitivity 50%, specificity 78%, positive predictive value 63%, and negative predictive value 68%.
Conclusions.
Relative change in dd-cfDNA% has improved test characteristics for diagnosing ALAD compared with absolute values.