Bayesian Evaluation of Solana HSV 1+2/VZV Assay Compared to Viral Culture and Commercial PCR Assay for Cutaneous or Mucocutaneous Specimens.

Results from the Solana HSV 1+2/VZV assay for the detection of herpes simplex virus 1 (HSV-1), HSV-2, and varicella-zoster virus (VZV) in cutaneous or mucocutaneous specimens were compared with that of viral culture and a commercial PCR assay (RealStar alpha herpesvirus PCR kit). Three hundred two mucocutaneous specimens, for which HSV-1, HSV-2, or VZV viral culture or PCR detection have been requested, were randomly selected and prospectively processed on the Solana assay and viral culture or the RealStar assay. Discordant results between culture and the Solana assay were further analyzed using the RealStar assay. A Bayesian latent class model was developed to estimate the performance of each method. Viral culture detected 123 positive specimens (85 HSV-1, 36 HSV-2, and 2 VZV), while the Solana assay detected 27 additional positive specimens (4 HSV-1, 11 HSV-2, and 12 VZV), in agreement with the RealStar PCR assay. The estimated sensitivity of the Solana assay according to our model was 92.7% to 98.7%, 87.1% to 97.8%, and 94.9% to 98.8% (95% confidence interval [CI]) for HSV-1 HSV-2, and VZV, respectively, while the estimated sensitivity of viral culture was 85.2% to 95.0%, 73.6% to 89.6%, and 30.9% to 45.8% (95% CI), respectively. A nonsignificant tendency toward increased sensitivity was noted for the Solana assay compared with culture for HSV-1 and HSV-2, and the Solana assay was significantly more sensitive than culture for the detection of VZV. The Solana assay performed comparably to the RealStar assay. Processing time was reduced with the Solana assay compared with viral culture.