Routine In-Situ Hybridization of Kidney Transplant Biopsies for BK Virus (BKV): Is It Justifiable?
Charanjeet Singh, Annie Meares, Juan C Manivel, Jose Jessurun, Stephan E Pambuccian, Behzad Najafian. University of Minnesota, Minneapolis
Background: BKV nephropathy (BKN) is an important cause of renal transplant dysfunction. BKN diagnosis requires observation of characteristic BK nuclear inclusions in tubular or parietal epithelial cells, and confirmatory nuclear staining by BK in-situ hybridization (BKISH) or SV-40 antibody. The triggers for performing BKISH or SV-40, besides suspicious cytopathic changes on biopsy, are +BK blood PCR or prior BKN history. In the absence of evident BK nuclear includions or relevant history, BKN may be missed by pathologists. However, the value of performing BKISH or SV-40 routinely on all renal transplant biopsies (Tx Bx) is not clear. Here we present the data from BKISH performed on renal Tx Bx routinely for one year at our institution.
Design: BKISH was performed on consecutive allograft biopsies over one year. Pathologists, who were not blinded to BKN previous history, were asked to predict presence of BKN on routine light microscopy slides before BKISH results were available. Predictions were compared to BKISH, as the gold standard.
Results: Of the 291 (M/F=166/125) Bx (age 45 ± 16 years) studied, 17 (6%) Bx had + BKISH, the majority of which (82%) were from male patients. Overall, pathologists' prediction was in agreement with BKISH in 91% of biopsies, while %agreement for +BKISH (pathologist call sensitivity) was 59% (Table 1) and for -BKISH (pathologist call specificity) was 93%. The pathologists' positive predictive value for BKN was 36%, while the negative predictive value was 97%. BKN was missed in 41% of +BKISH.
|Positive Prediction||10||18||PPV*= 36%|
|Negative Prediction||7||256||NPV**= 97%|
|Sensitivity= 59%||Specificity= 93%|