|
An ELISA assay
using a glycoprotein G peptide for the detection of HSV-2 antibodies
performed better than two assays using the intact protein.
Performance
characteristics of a glycoprotein G based oligopeptide (peptide 55)
and two different methods using the complete glycoprotein as assays
for detection of anti-HSV-2 antibodies in human sera.
Nilsen A, Ulvestad E, Marsden H,
Langeland N, Myrmel H, Matre R, Haarr L.
J Virol Meth. 2003;107:21-27.
Summary:
Question
What are the performances of three different HSV-2 glycoprotein G ELISA
assays for the detection of anti-HSV-2 IgG antibodies in human serum?
Design
This study describes a direct comparison of the performance of an ELISA
assay that uses a peptide portion of the HSV-2 G glycoprotein (gG-2) as
the test antigen with those of two ELISA assays that use the complete
intact protein for the detection of antibodies to HSV-2 in a panel of
previously characterized human sera.
Participants
Thirty-six individuals who had culture-documented genital infection by
HSV-2 more than 6 months earlier, 30 patients who had antibodies to HSV-1
but not HSV-2, and 32 who were anti-HSV negative were tested. The HSV-1
positive and HSV negative patients had no history of genital herpes and
their serostatus was determined using Enzygost Anti HSV/IgG and IgM (Behringer
Diagnostics, Marburg, Germany) and an HSV-1 and 2 differentiation
immunoblot test (Focus Technologies, Cypress, CA).
Description of Tests and Diagnostic
Standard
Three ELISA assays were used for specific detection of anti-HSV-2 IgG
antibodies. 1) The Ho method, employing native gG-2 selected by lectin
affinity, was performed as described by the author except that 0.5% Tween
in PBS was used for washing. The antigen was diluted 1:6000 and the serum
1:250. 2) The Gull method, using native gG-2 selected by monoclonal
antibody affinity, was performed according to the manufacturer's
instructions (Gull Laboratories, Salt Lake City, UT). This kit is no
longer commercially available. 3) The peptide 55 method used a branched
oligopeptide corresponding to residues 561-578 of gG-2 as the antigen.
Each well was coated with 100 ng of peptide 55, the serum was diluted 1:5,
incubated with biotinylated anti-human IgG, and detected with streptavidin-horseradish
peroxidase.
Performance of the three ELISA tests was
investigated at all combinations of sensitivity-specificity pairs using
receiver-operating characteristics (ROC) methodology. Cut-off levels were
established following the manufacturers' instructions and at a level where
the tests showed the highest proportion of correct results. The interassay
and intra-assay variability were measured by analysis of six replicates of
3 serum samples (HSV negative, HSV-2 low positive, and HSV-2 high
positive) on three different days.
Main Outcome Measures
The coefficient of variation (CV), the measure of agreement (?), and the
sensitivity and specificity for the detection of anti-HSV-2 antibodies
were calculated for the three ELISAs.
Main Results
The interassay CVs for the three tests were 8.2, 9.1, and 8.9 for the Ho,
peptide 55, and Gull methods, respectively. The intra-assay CVs for the
same tests were 12.6, 15.7, and 11.0, respectively. The measure of
agreement (?) value was 0.849 for the Gull and Ho, 0.914 for the peptide
55 and Ho, and 0.891 for the peptide 55 and Gull test comparisons. The
sensitivity and specificity of each assay calculated using the
manufacturers' cutoff and the optimum cut-off determined from the ROC
analyses are shown in the table.
Authors' Conclusions
Establishing new cut-off values for the
ELISAs based on diagnostic efficiency of the tests increased the
performance characteristics for all three tests. Although the performance
of all three assays was satisfactory, the ELISA assay using HSV-2
glycoprotein G peptide 55 gave superior efficiency and accuracy compared
to two different methods based upon antigenicity of the complete gG-2
protein.
Source of funding:
Norwegian Council of Universities' Programme for Development Research and
Education, GlaxoSmithKline, and the Norwegian Cancer Society
For correspondence: Arvid Nilsen,
Department of Dermatology, Haukeland University Hospital, N-5021 Bergen,
Norway. E-mail address: arvid.nilsen@haukeland.no.
|
.gif){kind=small}