Figure 4.

Biopsy-core parameters increased the likelihood of BKVAN diagnosis. The likelihood of BKVAN detection was higher with multiple biopsy cores (A) and a longer maximal biopsy core (B; P<0.001). The lesser detection rate with four biopsy cores was explained by multiple shallow cores from difficult cases undergoing additional passes (A). **P<0.01 versus one core. The relationship between glomerular number (cortical adequacy) and BK infection was weaker and nonlinear, where infection could be detected in medullary samples that contained very few glomeruli (C; P<0.001). Glomerular number was associated with longer cores (F). Biopsy polyomavirus (pvl; pvl score category) assessed by SV40T and cytopathologic abnormalities (D; left and right bars, respectively) was also greater when biopsy medulla was present. ***P<0.001 versus cortex. Longer biopsy cores associated with higher BKVAN detection rates; however, samples containing medulla enhanced the diagnostic likelihood compared with cortical samples (E; upper line with black squares and lower line with open circles, respectively; P<0.001 for differences). 95% CI, 95% confidence interval; SS, sum of squares.