Evaluation of automated nucleic acid extraction methods for virus detection in a multicenter comparative trial

Five European veterinary laboratories participated in an exercise to compare the performance of nucleic acid extraction robots. Identical sets of coded samples were prepared using serial dilutions of bovine viral diarrhoea virus (BVDV) from serum and cell culture propagated material. Each laboratory extracted nucleic acid from this panel using available robotic equipment (12 separate instruments, comprising 8 different models), after which the processed samples were frozen and sent to a single laboratory for subsequent testing by real-time RT-PCR. In general, there was good concordance between the results obtained for the different automated extraction platforms. In particular, the limit of detection was identical for 9/12 and 8/12 best performing robots (using dilutions of BVDV infected-serum and cell culture material, respectively), which was similar to a manual extraction method used for comparison. The remaining equipment and protocols used were less sensitive, in an extreme case for serum, by a factor of 1000. There was no evidence for cross-contamination of RNA template in any of the negative samples included in these panels. These results are not intended to replace local optimisation and validation, but provide reassurance to laboratories to indicate that the best performing optimised nucleic acid extraction systems can have similar performance.