RESUMEN
Many clinical laboratories have replaced virus isolation in cell-culture (VIC) for cytomegalovirus (CMV) by quantitative-nucleic-acid testing (QNAT), rendering clinically relevant CMV-replication difficult to distinguish from CMV-shedding or latent infection. We compared direct VIC in 1109 consecutive bronchoalveolar lavage fluids (BALFs) and a well-validated CMV-QNAT (Basel-CMV-UL111a-77bp). In the retrospective Group 1 (N = 694) and Group 2 (N = 303), CMV-QNAT was performed within 48 h from 2-fold and 10-fold concentrated total nucleic acid (TNA) eluates, respectively. In Group 3 (N = 112), 2-fold and 10-fold concentrated TNA eluates were prospectively analyzed in parallel to VIC. CMV was detected by VIC in 79 of 694 (11%) and 26 of 303 (9%) of Groups 1 and 2, but in 114 of 694 (16%) and 57 of 303 (17%) by CMV-QNAT, respectively. Median CMV loads were significantly higher in VIC-positive than in VIC-negative BALF. The likelihood for CMV detection by VIC was 85% for BALF CMV- loads >4 log10 copies/ml. In the prospective Group 3, CMV was detected by VIC in 10 of 112 (9%), and in 14 of 112 (13%) and 20 of 112 (18%) by CMV-QNAT, when using 2-fold and 10-fold concentrated TNA eluates, respectively. Notably, CMV was undetectable by CMV-QNAT in 10 VIC-positive cases of Groups 1 and 2, but in none of Group 3. We conclude that CMV-QNAT can be adopted to BALF diagnostics but requires several careful steps in validation. CMV-QNAT loads >10 000 copies/ml in BALF may indicate significant CMV replication as defined by VIC, if short shipment and processing procedures can be guaranteed. Discordance of detecting CMV in time-matched plasma samples emphasises the role of local pulmonary CMV replication, for which histopathology remains the gold standard of proven CMV pneumonia.