Detection of Kidney Allograft Rejection Using Urinary Chemokines.

Abstract

Background: Urinary chemokines CXCL9 and CXCL10 have shown promise for detecting kidney allograft rejection, but the demonstration of their added value beyond standard of care patient monitoring requires further study.

Methods: We prospectively enrolled adult patients who underwent kidney transplantation in 7 transplant referral centers between July 2018 and December 2019 (ClinicalTrials.gov, NCT03582436). We quantified urinary CXCL9 and CXCL10 protein levels at the time of kidney allograft biopsies in the first year post-transplantation using an automated immunoassay platform. The primary outcome was allograft rejection defined according to the international Banff 2019 classification.

Results: Overall, 733 kidney transplant patients (64% male, 36% female) were included in the main analysis, with 1,549 biopsies paired with a urine sample. The cumulative incidence of rejection was 10%. For detecting allograft rejection, urinary CXCL9 and CXCL10 demonstrated areas under the receiver operating characteristic curve (AUROC) of 0.70 (95% confidence interval [CI], 0.64-0.75) and 0.64 (95% CI, 0.58-0.71), respectively. Adding urinary CXCL9 to a standard of care model improved discrimination for allograft rejection (AUROC 0.75 [percentile bootstrap CI 0.70-0.79] to 0.78 [percentile bootstrap CI 0.73-0.83]), while urinary CXCL10 did not. There was no improvement of overall fit with the addition of urinary CXCL9 (Brier score changed from 0.056 [95% CI, 0.046-0.067] to 0.054 [95% CI, 0.045-0.064]), as this tended to overestimate the risk for allograft rejection. In sensitivity analyses restricting to only acute/active forms of rejection or to a single randomly selected biopsy per patient, urinary chemokines did not show additional value beyond the standard of care. In addition, existing chemokine-based models showed low to moderate performance for the detection of allograft rejection.

Conclusions: Urinary CXCL9 demonstrated limited clinical utility, while urinary CXCL10 provided no additional value beyond standard of care monitoring for detecting allograft rejection within the first year after kidney transplantation.