The Importance of Kidney Medullary Tissue for the Accurate Diagnosis of BK Virus Allograft Nephropathy.

BACKGROUND AND OBJECTIVES: The published tissue adequacy requirement of kidney medulla for BK virus allograft nephropathy diagnosis lacks systematic verification and competes against potential increased procedural risks from deeper sampling. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: We evaluated whether the presence of kidney medulla improved the diagnostic rate of BK nephropathy in 2244 consecutive biopsy samples from 856 kidney transplants with detailed histologic and virologic results. RESULTS: Medulla was present in 821 samples (37%) and correlated with maximal core length (r=0.35; P<0.001). BK virus allograft nephropathy occurred in 74 (3% overall) but increased to 5% (42 of 821) with medulla compared with 2% (32 of 1423) for cortical samples (P<0.001). Biopsy medulla was associated with infection after comprehensive multivariable adjustment of confounders, including core length, glomerular number, and number of cores (adjusted odds ratio, 1.81; 95% confidence interval, 1.02 to 3.21; P=0.04). In viremic cases (n=275), medulla was associated with BK virus nephropathy diagnosis (39% versus 19% for cortex; P<0.001) and tissue polyomavirus load (Banff polyomavirus score 0.64±0.96 versus 0.33±1.00; P=0.006). Biopsy medulla was associated with BK virus allograft nephropathy using generalized estimating equation (odds ratio, 2.04; 95% confidence interval, 1.05 to 3.96; n=275) and propensity matched score comparison (odds ratio, 2.24; 95% confidence interval, 1.11 to 4.54; P=0.03 for 156 balanced pairs). Morphometric evaluation of Simian virus 40 large T immunohistochemistry found maximal infected tubules within the inner cortex and medullary regions (P<0.001 versus outer cortex). CONCLUSIONS: Active BK virus replication concentrated around the corticomedullary junction can explain the higher detection rates for BK virus allograft nephropathy with deep sampling. The current adequacy requirement specifying targeting medulla can be justified to minimize a missed diagnosis from undersampling.