Longitudinal Reliability of Focus Glycoprotein G-Based Type-Specific Enzyme Immunoassays for Detection of Herpes Simplex Virus Types 1 and 2 in Women

Abstract

Serologic assays that utilize herpes simplex virus (HSV) type-specific glycoproteins G-1 (HSV-1) and G-2 (HSV-2) to discriminate between antibodies against HSV-1 and HSV-2 are sensitive and specific. However, the high rates of seroreversion, defined as the change in an individual's antibody status from positive to negative over time, previously reported in longitudinal evaluations of glycoprotein G type-specific tests suggests that their use in HSV acquisitional studies would be problematic. To further explore the reliability of the glycoprotein G-based serologic tests, we evaluated HSV-1 and HSV-2 enzyme immunoassays from Focus Technologies in a longitudinal cohort of 1,207 young women from Pittsburgh, Pa. On enrollment of the women in the study, HSV-1 and HSV-2 antibodies were detected in 46.6 and 24.9% of the women, respectively. Among the women with at least three visits, 3.4% (15 of 447) of those who were HSV-1 antibody positive had a subsequent negative result while fewer than 1% (2 of 227) of those who were HSV-2 antibody positive seroreverted. The median of mean positive index values for women who seroreverted to HSV-1 antibody was lower than that for women who remained seropositive (1.25 versus 7.06; P < 0.001). Similarly, the median of mean positive index values for women whose HSV-2 antibody status reverted from positive to negative was lower than that for those women who did not serorevert (1.83 versus 7.46; P = 0.02). Comparative Western blot analysis demonstrated that the lower positive index values, seen more often among the HSV seroreverters, often signified false-positive immunoassay results. Overall, the seroreversion rates were low; the use of glycoprotein G-based serologic tests for the measurement of HSV-1 and HSV-2 antibodies in incidence studies therefore appears warranted.

It is estimated that over 50 million people in the United States have genital herpes (7). A 30% increase in the seroprevalence of herpes simplex virus type 2 (HSV-2) infection in this country over the past two decades demonstrates that the epidemic continues unabated (9). Furthermore, an increasing proportion of primary genital herpes infections are caused by herpes simplex virus type 1 (HSV-1) (11). The ability to diagnose these HSV infections has been enhanced by the advent of type-specific serologic tests. HSV-1 and HSV-2 exhibit a high degree of cross-reactivity, but an envelope glycoprotein, gG, is antigenically distinct between the two viruses, and the type-specific antibody response to gG provides evidence of infection with HSV-1 or HSV-2 or both (5). Type-specific gG serologic assays include Western blotting (2, 6), immunoblot strips (21), and enzyme-linked immunosorbent assays (ELISAs) (14, 16, 17). Although the Western blot assay is considered to be the “gold standard,” it is expensive and labor-intensive and its interpretation is operator dependent (5). In comparison, ELISAs are quicker, less expensive, and better suited for high-volume screening.

The objective of the present study was to determine the longitudinal reliability of type-specific HSV-1 (gG-1) and HSV-2 (gG-2) ELISAs in a cohort of nonpregnant women. In cross-sectional studies, HSV-1 and HSV-2 gG enzyme immunoassays have demonstrated both high sensitivity and specificity (4, 8, 14, 16, 17). However, prior evaluations of type-specific assays have reportedly detected frequent changes in an individual's gG antibody status from positive to negative over time. This phenomenon has been termed seroreversion (5). Women tested in each trimester of pregnancy demonstrated seroreversion rates of 3% for HSV-1 and 25% for HSV-2, while Thai military recruits had seroreversion rates of 6.6% for HSV-1 and 14.9% for HSV-2 when tested by a recombinant gG Western blot assay (1, 19). The authors of the latter investigation concluded that until the reliability of gG serologic tests in longitudinal studies is better understood, it would be prudent to confine their use to cross-sectional analyses. Although provocative, examination of a subset of 33 specimens from their cohort revealed that four of the specimen sets may have contained sera from more than one individual; therefore, misidentification of specimens may, in part, account for the high rates of detected seroreversion.

To study the risk factors for acquisition of HSV in women, we tested a cohort over time for HSV-1 and HSV-2 antibodies by gG type-specific ELISAs (Focus Technologies). These assays use baculovirus recombinant gG constructs, are Food and Drug Administration approved for testing sexually active adults and pregnant women, and are commercially available. In comparison to Western blot assays, the sensitivity and specificity of the Focus HSV-2 ELISA are 98 and 95%, respectively, while those of the Focus HSV-1 ELISA are 98 and 94%, respectively (16, 17).

MATERIALS AND METHODS

Study population and design.

The analysis was conducted on 3,829 serum samples collected between 1998 and 2000 for The Streptococcal Initiative, an investigation of the risk factors associated with vaginal acquisition of group B Streptococcus. The evaluation of the gG type-specific HSV ELISAs results was supported by a contract and a grant from the National Institute of Allergy and Infectious Diseases and approved by the Institutional Review Board of the Magee-Womens Hospital of Pittsburgh, Pa. In this longitudinal cohort study, 1,207 nonpregnant women 18 to 30 years of age were recruited from three Pittsburgh, Pa., area sites: the University of Pittsburgh Student Health Clinic, the Allegheny County Health Department Clinic, and the Family Health Council Clinic of Aliquippa. Blood was obtained from the women at the time of enrollment and at the next three 4-month intervals. At least three serum samples were obtained from 958 of the 1,207 women enrolled.

Laboratory Tests.

Focus (Cypress, Calif.) HSV-1 and HSV-2 IgG gG ELISAs were performed as specified by the manufacturer. Serum samples and controls were incubated in either HSV-1 or HSV-2 recombinant gG antigen-coated polystyrene microwells. This allowed specific antibody present in the samples to bind with the adhered antigen. Peroxidase-conjugated anti-human immunoglobulin G (IgG) was added to react with specific IgG, and the subsequent addition of a substrate reagent resulted in a color change in wells that contained positive samples. The color change was quantified by a spectrophotometric reading of optical density (OD), and sample OD readings were compared with reference cutoff OD readings to determine the results. Index values greater than 1.10 were considered positive, while values less than 0.90 were negative. Sera with equivocal results (an index value between 0.90 and 1.10 inclusive) were retested. The final result for a particular specimen was considered equivocal if the repeat index value again fell between the inclusive values of 0.90 and 1.10. For women who had made four visits and whose HSV-1 or HSV-2 serostatus had shifted from positive to negative, all serum samples were submitted to the University of Washington Virology Laboratory for comparative HSV Western blot (WB) testing (2). The mean of the positive index values was calculated for each woman, and then the medians of these values were compared between women consistently positive for HSV-1 or HSV-2 and women who seroreverted in either assay, using the Mann-Whitney U test.

RESULTS

At the time of their enrollment, 46.6 and 24.9% of the women harbored HSV-1 and HSV-2 antibodies, respectively. Among the women who had made at least two follow-up visits, 3.4% of the women who were HSV-1 antibody positive (15 of 447) had a subsequent negative result (Table 1). Another 6 women had an equivocal HSV-1 ELISA result after a positive index value had been recorded. In comparison, 0.9% of the women who had made at least three visits (2 of 227) reverted from a positive to negative HSV-2 index value, while only one woman had an equivocal value after a positive result.

TABLE 1.

Pattern frequencies for HSV-1 and HSV-2 gG type-specific ELISA results among women who made at least two follow-up visits

Result pattern	No. (%) with result for:
	HSV-1 (n = 956)	HSV-2 (n = 958)
Consistently positive	432 (45.2)	225 (23.5)
Consistently negative	473 (49.5)	701 (73.2)
Seroconversion	30 (3.1)	29 (3.0)
Seroreversion	15 (1.6)	2 (0.2)
Seropositive with subsequent equivocal value	6 (0.6)	1 (0.1)

Quality control measures supplied by the manufacturer of these ELISAs included the measurement of both high and low positive control index values each time the assays were performed. If the high positive control index value was less than 3.5, the test results were considered invalid and it was necessary to repeat the assay. As the next step in data analysis, we calculated the percentage of the positive index values at any visit which were greater than the high positive control index value of 3.5. The two immunoassays performed similarly; 80% of the HSV-1-positive values were greater than 3.5, as were 83% of the HSV-2-positive index values (Table 2). Using a positive index value cutoff of 3.0 rather than 3.5, we found that only 16 and 13% of the HSV-1- and HSV-2-seropositive results, respectively, were below this lower cutoff value.

TABLE 2.

Distribution of HSV-1 and HSV-2 gG type-specific ELISA index values for study participants with a positive index value for either assay at any visit

Index value	No. (%) with index value for:
	HSV-1 (n = 1,825)	HSV-2 (n = 979)
1.5	98 (5.4)	37 (3.8)
>1.5 to 3.5	266 (14.6)	133 (13.6)
>3.5	1,461 (80.0)	809 (82.6)

We then analyzed the data to learn if the index values of samples obtained preceding seroreversion were lower than the index values obtained in women who remained seropositive for either assay. Table 3 lists the individual index values for all women with four serum samples whose HSV-1 or HSV-2 ELISA results reverted from positive to negative at any visit during the investigation. As predicted, the mean positive index values prior to seroreversion were lower than the mean positive index values calculated in women who did not serorevert. The median of mean positive index values among the women whose HSV-1 antibody status reverted from positive to negative was lower than for women who retained their HSV-1 seropositivity (1.25 versus 7.06; P < 0.001) (Table 4). Similarly, the median of mean positive index values for women whose HSV-2 antibody status reverted from positive to negative was 1.83, significantly lower than the median of 7.46 found among women who did not serorevert from a positive HSV-2 result during the investigation (P < 0.02).

TABLE 3.

HSV-1 and HSV-2 gG type-specific ELISA index values and Western blot analysis results by visit for women whose antibody status reverted from positive to negative by ELISA^a

Patient ID and virus	Index value/Western blot resultd at visit:
	Baseline	4 mo	8 mo	12 mo
HSV-1
1	1.23/0	0.96/0	0.59/0	0.46/0
2	1.23/0	0.87/0	0.87/0	2.08/0
3b	1.74/1	0.54/0	0.80/0	0.84/1
4	1.53/0	1.43/0	1.24/0	0.64/0
5c	3.30/1	3.79/1	3.66/1	0.13/0
6	0.16/0	0.60/0	1.18/0	0.42/0
7	0.91/0	0.93/1	1.26/1	0.83/1
8	1.55/0	0.86/0	0.68/0	0.49/0
9	0.86/0	1.19/0	0.78/0	0.70/0
10	1.43/0	0.78/0	0.78/0	0.85/0
11	1.26/1	0.72/1	1.21/1	0.97/1
HSV-2
1	0.22/1	2.52/1	0.60/1	0.47/1
2	1.16/0	1.07/0	0.75/0	1.11/0

^aData for women who made four visits and hence had four serum samples.

^bInterpretation of the Western blot results indicated that samples from the baseline and 12-month visits were not from the same subject as those from the 4- and 8-month visits.

^cInterpretation of the western blot results indicated that the sample from the 12-month visit was not from the same subject as those from the baseline and the 4- and 8-month visits.

^dWestern blot results were as follows: 0, antibody to neither HSV-1 or HSV-2; 1, antibody to HSV-1; 2, antibody to HSV-2.

TABLE 4.

HSV-1 and HSV-2 gG type-specific ELISA median mean positive index values for women who remained antibody positive for HSV-1 or HSV-2 compared to the mean positive index values for women whose antibody status reverted from positive to negative is either assay

Status of patients	Median positive index value (range) fora:
	HSV-1	HSV-2
Remained positive	7.06 (1.21-10.07)	7.46 (1.34-13.58)
Serorevertered	1.25 (1.14-2.24)	1.83 (1.13-2.52)

^aP < 0.001 for HSV-1 and P = 0.02 for HSV-2.

Along with its display of individual index values, Table 3 details the comparison between ELISA and Western blot results for women whose HSV-1 or HSV-2 antibody status reverted from positive to negative by ELISA. For the 11 women whose HSV-1 ELISA seroreverted, 7 of the serum sets showed no evidence for HSV-1 infection by Western blot analysis, indicating that the ELISA results probably represent false-positive index values (identification numbers 1, 2, 4, 6, 8, 9, and 10). The Western blot results from another two of the HSV-1 serum sets suggested that they contained serum samples from different women, and these results were excluded from all seroreversion analyses (identification numbers 3 and 5). Finally, 2 of the 11 HSV-1 serum sets contained a negative HSV-1 ELISA index value while Western blot analysis of the same sample was consistent with HSV-1 infection; therefore, these index values were probably falsely negative (identification numbers 7 and 11). A similar comparison of the HSV-2 ELISA and Western blot results revealed the presence of false-positive index values in both of the serum sets.

DISCUSSION

Type-specific gG serologic testing for the diagnosis of HSV-1 and HSV-2 infections has a variety of applications. For the two women who were determined by ELISA to have seroconverted to HSV-2-positive status, Western blot analysis failed to detect gG-2 antibodies during any visit. Many infected individuals who are not cognizant of their infection continue to shed HSV and may be an important reservoir for transmission (13). Type-specific antibody screening is the most practical way to identify these silent carriers of infection. The mothers of infants who develop neonatal herpes frequently deny a history of genital herpes; type-specific serologic testing could be employed to identify mothers at highest risk of infecting their infants (15). Type-specific HSV diagnostic screening tests may be offered to high-risk populations as a component of HIV prevention programs, since HSV-2 infections increase the risk of human immunodeficiency virus type 1 HIV-1 acquisition (10). Finally, evidence that antiviral treatment of HSV can suppress genital shedding suggests another utilization of type-specific serologic testing for the diagnosis of herpes infections (20).

It is important to achieve an improved understanding of the accuracy and limitations of the type-specific HSV serologic assays. False-negative results may occur in patients recently infected with HSV. Other groups in which false-negative results have been reported include individuals with AIDS and recipients of solid-organ transplants (8, 18). As our investigation demonstrates, false-positive results are more likely to occur when index values are in the low positive range. Therefore, it is possible that low-level, nonspecific IgG binding may be responsible for these inaccuracies. Because of the psychological distress that often occurs when individuals are diagnosed with genital herpes, repeat HSV ELISA or confirmatory Western blot analysis may be recommended when a clinician is counseling a patient with a low positive index value. Our investigation also highlights the importance of accurate handling and identification of specimens in longitudinal studies, since Western blot analysis identified two sets among the putative HSV-1 ELISA seroreverters that probably contained serum from more than one individual. Errors in specimen labeling, either in the clinic or in the laboratory, would lead to tests being interpreted erroneously as seroreversions or acquisitions.

We found the seroreversion rates of the Focus HSV-1 and HSV-2 IgG gG ELISAs to be lower in our cohort of young women than was previously reported for other G-based assays (19). Seroreversions did occur with both ELISAs, but they were infrequent and occurred exclusively following tests which had index values of less than 2.24 for HSV-1 and 2.52 for HSV-2. Our examination of the data revealed that the majority of positive index values were greater than 3.0 (84% for HSV-1 and 87% for HSV-2) and that women who seroreverted by ELISA were more likely to have positive index values. It would have been necessary to perform comparative Western blot analyses on all seropositive samples to determine if positive index values greater than 3.0 always represented true-positive infections. However, because of the higher risk for false-positive results with lower positive index values, we recommend that the Focus HSV immunoassays be repeated when a positive index value from 1.1 to 3.0 is obtained. As with all diagnostic tests, the utilization of HSV ELISAs and the interpretation of their results depend on clinical acumen as well as assay reliability. In individuals in whom no lesions are detectable but recent HSV acquisition is suspected, a delay of several weeks in repeating the test would be more likely to provide a definitive diagnosis. Overall, we feel that these HSV assays performed reliably in our longitudinal investigation and that further use in incidence studies is justified.