The accuracy of blood nucleic acid testing to predict BK polyomavirus nephropathy varies by clinical context

Abstract

BACKGROUND: BK polyomavirus (BKPyV) reactivation causes a destructive allograft nephropathy (BKVAN) with limited treatment options. Quantitative nucleic acid testing (QNAT) of BKPyV-DNA in blood within screening programs monitors for viral reactivation but has imprecise real-world authentication. METHODS: This prospective, single-gate, single-center, population-based cohort study evaluated the test performances of QNAT (the index test) against biopsy-proven BKVAN, verified by universal Simian virus 40 T antigen (SV40T) immunohistochemistry (the optimal reference test), in 5339 consecutive kidney transplant samples. RESULTS: BKVAN occurred in 205 tissue samples (3.8%) obtained from 1542 recipients (6.9%); classified as polyomavirus nephropathy class 1 in 26.3%, class 2 in 67.8%, and late class 3 in 5.9%. Contemporaneous QNAT+ was detected in 12.7% of cases and was associated with viremia persistence, SV40T+, and cytopathic effect using multivariable analysis. The sensitivity of 97.7% and specificity of 90.7% for QNAT+ to predict BKVAN were unaffected by assay methodology. SV40T+ BKVAN was QNAT+ in all patients except for 6 false-negative cases. A negative qualitative NAT during protocol surveillance sampling (n = 3667) with a specificity of 90.7% demonstrated a 99.9% negative predictive value for excluding BKVAN at a 2.6% prevalence. In contrast, QNAT exceeding 1.0E+04 BKPyV-DNA copies/mL in viremic patients (n = 679) showed a 56.3% sensitivity and a 54.6% positive predictive value. Mild BKVAN was often misclassified, with a positive likelihood ratio of 2.9. CONCLUSIONS: Test performance of QNAT in transplant recipients varies markedly according to clinical context. A negative QNAT in surveillance programs reliably rules out BKVAN. However, although positive results confirm BKV reactivation, QNAT levels could not accurately distinguish tissue-invasive infection at any threshold value.