Question
How well does the commercially available ELISA assay, HerpeSelect, perform compared to the Western blot assay for the detection of HSV 2 type-specific antibodies in samples from HIV-positive and HIV-negative men and women from the Rakai District in Uganda?

Design
The results of the HerpeSelect HSV-2 ELISA were compared to the Western blot assay for the detection of HSV 2 type-specific antibodies in samples from HIV-1 positive and HIV-negative men and women in the Rakai District of Uganda.  The results of the Western blot assay were used to adjust the ELISA index cutoff values to obtain optimal performance of the ELISA.

Participants
Sera from 248 HIV-1 positive and 496 HIV-negative adults aged 15 to 19 years in the Rakai District, Uganda, who were enrolled in a population-based randomized controlled trial of presumptive sexually transmitted disease treatment, were tested.  

Description of Tests and Diagnostic Standard
The HerpeSelect HSV-2 ELISA (Focus Technologies, Cypress,CA) was performed according to the manufacturer’s protocol.  Positive and negative results were determined using index cutoff values ranging from 1 to 15.  The manufacturer’s index cutoff value was 1.1.  The Western blot assay was performed as previously described.  Samples with atypical Western blot results were considered negative for HSV-2 for the statistical analyses.  

Main Outcome Measures
The sensitivity and specificity of the ELISA were assessed and a receiver operating curve (ROC) was determined by using Western blot as the gold standard.  

Main Results
The HSV-2 seroprevalence was 62% by Western blot.  The sensitivity and specificity values of the ELISA compared to the Western blot for 744 samples were calculated using index cutoff values of 1, 1.1, 1.5, 2 to 4 by increments of 0.1, and 5 to 15 by whole number (table) and the results were plotted on a ROC curve.  The ROC curve indicated that the index cutoff value for optimal ELISA performance was 3.4.  The HSV-2 seroprevalence by Western blot was 71% and 59% for HIV-1 positive and HIV-negative subjects, respectively (P < 0.001).  The ELISA median index values for HIV-positive subjects did not differ significantly from those for HIV-negative subjects among either HSV-2 positive or negative subjects as determined by Western blot.  

Performance of HSV-2 ELISA compared to Western blot at different ELISA index cutoff values for 744 HIV positive and HIV negative serum samples from men and women in Uganda

Authors’ Conclusions
Depending on the setting where the HSV-2 ELISA is used, a different index cutoff value may need to be selected.  Where Western blot testing is available to confirm a positive result, a value of 1.1 will maximize sensitivity.  In the absence of a confirmatory Western blot test, a cutoff value of 2.2 is more appropriate.    

Source of funding:  Grants from the National Institutes of Health and the HIV Prevention Trials Network.  

For correspondence:  Thomas C. Quinn, Ross Building 1159, 720 Rutland Ave., Baltimore, MD 21205.  E-mail address:   HYPERLINK "mailto:tquinn@jhmi.edu" tquinn@jhmi.edu.