Background: The aims of the study were to assess the performance of a clinically available cell-free DNA (cfDNA) assay in a large cohort of pediatric and adult heart transplant recipients and to evaluate performance at specific cut points in detection of rejection.
Methods: Observational, non-interventional, prospective study enrolled pediatric and adult heart transplant recipients from seven centers. Biopsy-associated plasma samples were used for cfDNA measurements. Pre-determined cut points were tested for analytic performance.
Results: A total of 487 samples from 160 subjects were used for the analysis. There were significant differences for df-cfDNA values between rejection [0.21% (IQR 0.12-0.69)] and healthy samples [0.05% (IQR 0.01-0.14), p < .0001]. The pediatric rejection group had a median df-cfDNA value of 0.93% (IQR 0.28-2.84) compared to 0.09% (IQR 0.04-0.23) for healthy samples, p = .005. Overall negative predictive value was 0.94 while it was 0.99 for pediatric patients. Cut points of 0.13% and 0.15% were tested for various types of rejection profiles and were appropriate to rule out rejection.
Conclusion: The study suggests that pediatric patients with rejection show higher levels of circulating df-cfDNA compared to adults and supports the specific cut points for clinical use in pediatric and adult patients with overall acceptable performance.
Background: Urinary CXCL10/creatinine (uCXCL10/Cr) is proposed as an effective biomarker of subclinical rejection in pediatric kidney transplant recipients. This study objective was to model implementation in the clinical setting.
Methods: Banked urine samples at a single center were tested for uCXCL10/Cr to validate published thresholds for rejection diagnosis (>80% specificity). The positive predictive value (PPV) for rejection diagnosis for uCXCL10/Cr-indicated biopsy was modeled with first-positive versus two-test-positive approaches, with accounting for changes associated with urinary tract infection (UTI), BK and CMV viremia, and subsequent recovery.
Results: Seventy patients aged 10.5 ± 5.6 years at transplant (60% male) had n = 726 urine samples with n = 236 associated biopsies (no rejection = 167, borderline = 51, and Banff 1A = 18). A threshold of 12 ng/mmol was validated for Banff 1A versus no-rejection diagnosis (AUC = 0.74, 95% CI = 0.57-0.92). The first-positive test approach (n = 69) did not resolve a clinical diagnosis in 38 cases (55%), whereas the two-test approach resolved a clinical diagnosis in the majority as BK (n = 17/60, 28%), CMV (n = 4/60, 7%), UTI (n = 8/60, 13%), clinical rejection (n = 5/60, 8%), and transient elevation (n = 18, 30%). In those without a resolved clinical diagnosis, PPV from biopsy for subclinical rejection is 24% and 71% (p = .017), for first-test versus two-test models, respectively. After rejection treatment, uCXCL10/Cr level changes were all concordant with change in it-score. Sustained uCXCL10/Cr after CMV and BK viremia resolution was associated with later acute rejection.
Conclusions: Urinary CXCL10/Cr reliably identifies kidney allograft inflammation. These data support a two-test approach to reliably exclude other clinically identifiable sources of inflammation, for kidney biopsy indication to rule out subclinical rejection.