Predictors of graft failure after first detection of de novo donor-specific HLA antibodies in kidney transplant recipients.

BACKGROUND: De novo donor-specific antibodies (dnDSA) may cause antibody-mediated rejection and graft dysfunction. Little is known about the clinical course after first detection of dnDSA during screening in asymptomatic patients. We aimed to assess the value of estimated glomerular filtration rate-eGFR and proteinuria to predict graft failure in patients with dnDSA and their potential utility as surrogate endpoints.

METHODS: All 400 kidney transplant recipients with dnDSA at our center (01/03/2000-31/05/2021) were included in this retrospective study. The dates of graft loss, rejection, doubling creatinine, ≥30% eGFR decline, proteinuria ≥ 500 mg/g and ≥ 1000 mg/g were registered from first dnDSA appearance.

RESULTS: During 8.3 years of follow-up, graft failure occurred in 33.3% of patients. Baseline eGFR and proteinuria correlated with 5-year-graft loss (AUC-ROC 0.75 and 0.80, p < 0.001). Creatinine doubled after a median of 2.8 years (1.5-5.0) from dnDSA, and the time from doubling creatinine to graft failure was 1.0 years (0.4-2.9). Analyzing eGFR reduction ≥ 30% as surrogate endpoint (148/400), the time from dnDSA to this event was 2.0 years (0.6-4.2), with a positive predictive value (PPV) of 45.9% to predict graft loss, which occurred after 2.0 years (0.8-3.2). The median time from proteinuria ≥ 500 mg/g and ≥ 1000 mg/g to graft failure was identical-1.8 years, with PPV of 43.8% and 49.0%. Composite endpoints did not improve PPV. Multivariable analysis showed that rejection was the most important independent risk factor for all renal endpoints and graft loss.

CONCLUSIONS: Renal function, proteinuria and rejection are strongly associated with graft failure in patients with dnDSA and may serve as surrogate endpoints.