PSLBI-30 SARS-CoV-2 Screening and Detection: Assessment of surrogate virus neutralization test (svNT) diagnostic appropriateness in multispecies SARS-CoV-2 virus screening and detection

Abstract

Abstract The National Veterinary Services Laboratories (NVSL) have been testing animal samples since the 2020 SARS-CoV-2 outbreak. The current “gold standard” for serodetection of SARS-CoV-2 is the virus neutralization (VN) test. This method requires specialized training, more time to complete, and specialized facilities (BSL 3). Therefore, the SARS-CoV-2 VN test is not widely run outside of NVSL. To facilitate increased testing capacity during the outbreak, the surrogate virus neutralization test (sVNT) was selected mid-2020. This commercially available kit uses an ELISA format, giving a percent inhibition (%I) cut-off value to indicate “negative” or “positive,” allowing for faster testing turnaround. Despite its widespread use in many species, an in-depth, multispecies analysis of the data comparing sVNT to VN results has not been reported. Such an analysis is useful to assess the diagnostic appropriateness of the sVNT test for SARS-CoV-2 virus screening and detection across multiple species. Results were considered from all serological samples submitted to NVSL from March 2020 to May 2023 for routine SARS-CoV-2 diagnostic testing. Of these, samples with results for both VN and sVNT testing were considered for further analysis (n = 942). Samples were sorted by the species and family levels, and were assessed for rates of “true positive,” “false positive,” “true negative,” and “false negative” by sVNT, using paired VN results as the indicator of the true status of the sample. Additionally, the increased sVNT cut-off threshold for %I (20%I to 30%I for “positive”) was assessed for appropriateness by species using mean %I values for VN-positive samples. Increasing sVNT %I cut-off threshold for “positive” results from 20%I to 30%I resulted in 58 samples being correctly re-classified as “negative” by sVNT, rather than “positive” at the 20%I cut-off threshold. The mean %I values for most species fell between 40 to 80%I, except for Felidae, which had a mean %I of 93%I, and a minimum %I of 40.6%I for VN-positive samples. The true negative (sVNT/VN-NEG) rates for most species fell between 97to 100%, with 0 to 3% being false negatives (svNT-NEG, VN-POS), with very few exceptions. However, the least false positive (sVNT-POS;VN-NEG) rate was 8.8%, ranging up to 100% false positives for Equidae and one species of Cervidae. While the mean %I for most species indicates the cut-off value is appropriate (for those species where sVNT is an appropriate diagnostic), overall, the results highlight a lack of precision and specificity of the sVNT for many animal species. This indicates a lack of diagnostic appropriateness of the sVNT as a SARS-CoV-2 screening tool for certain species and a need for refinement in cut-off values for other species. The sVNT as a screening tool varies considerably by species for serodetection of SARS-CoV-2, results should be interpreted with caution, and use of other confirmatory tests is recommended.