AN ESTIMATED 22% OF PERSONS in the United States are infected with herpes simplex virus type 2 (HSV-2).1 Recently, some authorities have advocated a national HSV-2 prevention campaign to diminish this burden of disease.2 Because most persons with antibody to HSV-2 are unaware of their infection,3–7 public health programs to control genital herpes will likely require efforts to identify persons with unrecognized infections. Newer serologic tests for HSV-2 that detect antibodies to glycoprotein G2 (gG2) have reported sensitivities of 94% to 100% and specificities of 94% to 98%.8–13 A study performed in our sexually transmitted disease (STD) clinic reported that the positive predictive value (PPV) of the Gull/Meridian HSV-2 enzyme-linked immunosorbent assay (ELISA), a type-specific glycoprotein G-2 assay that is no longer available, was 96% among patients who agreed to undergo serologic testing for HSV-2.11 However, PPV is a function of both test performance and the prevalence of infection in the population tested. The seroprevalence of HSV-2 varies widely, with a higher prevalence observed in women than in men, in blacks relative to whites, and in older persons. As a result, even within an STD clinic, prevalence is heterogeneous, and widespread HSV-2 testing is likely to involve testing many persons at relatively low risk for infection in whom the ELISA PPV may be low.14
In 2001, we began routinely offering all patients in our STD clinic HSV-2 testing using the Focus HSV-2 ELISA. Because of uncertainties about how the test would perform in lower prevalence subpopulations, we froze sera from all men testing positive for HSV-2 in our clinic over an 11-month period, retested them in a second laboratory using the Focus HSV-2 ELISA, and performed confirmatory tests using the University of Washington Western blot.15 We then developed a new algorithm for interpreting the HSV-2 ELISA and instituted that algorithm in our clinic. In this article, we present data on test result agreement between the HSV-2 ELISA and Western blot in a relatively low prevalence population of male STD clinic patients, propose a new algorithm for interpreting HSV-2 serologies that includes a newly defined group of indeterminate results, and present HSV-2 Western blot results in persons with indeterminate HSV-2 ELISAs prospectively identified in our clinic over a 24-month period.
As a matter of clinic policy, persons evaluated at the Public Health–Seattle & King County (PHSKC) STD clinic since March 30, 2001, have been routinely offered HSV-2 serologic testing as part of evaluations for STD. Patients are charged a $25 fee for herpes simplex virus type 1 (HSV-1) and HSV-2 testing that is separate from the cost of the rest of their visit fee.
This article presents data from 2 studies: an initial analysis defining the PPV of the Focus HSV-2 ELISA (PPV study) and a follow-up study assessing an algorithm for interpreting HSV-2 ELISA results based on findings from the PPV study (follow-up study). The study population for the PPV study included men who tested for HSV between March 30, 2001, and February 28, 2002. Men were chosen as a study population because they could easily be identified in the laboratory based on information on laboratory slips and were known to have a lower prevalence of infection than women.1,11 If patients were tested more than once during the study period and one test was positive, the first positive test result was used for the calculation of prevalence.
Based on initial analyses of these data, we began routinely performing HSV-2 Western blots on sera from patients in our clinic testing positive by Focus ELISA for HSV-1 and indeterminate for HSV-2 (defined as an index value of 1-1-2.9). These sera define the follow-up study population.
All sera were initially tested using the Focus ELISA for HSV-1 and HSV-2 in the PHSKC laboratory and then stored at 2° to 8°C. Duplicate HSV-1 and HSV-2 Focus ELISAs as well as Western blots for HSV-2 were then performed at the University of Washington Diagnostic Virology Laboratory (UWDVL). ELISA tests for HSV-1 and HSV-2 were performed according to the manufacturer's instructions (Focus Technologies, Cypress, CA). Western blot for HSV-2 was performed on all sera as previously described.16 Specimens with a HSV-2 ELISA index value of less than 3.0 were tested using the Focus Immunoblot according to the manufacturer's instructions. Specimens with discordant results on HSV-2 ELISA and Western blot were tested using HSV-2 avidity and inhibition assays as previously described.17,18
Western blot was considered the gold standard for diagnosing HSV-2 for calculations of PPV. Chi-square was used to test for an association between test performance and the presence of antibody to HSV-1. The correlation between HSV-2 ELISA test results from the 2 laboratories was calculated using the Pearson correlation coefficient. Median HSV-2 index values in specimens with positive and negative HSV-2 Western blot were compared using the Wilcoxon rank sum test. All statistical procedures were performed using the SAS system (release 8.02).
The University of Washington Human Subjects Review Committee approved all study procedures.
Positive Predictive Value Study
Between March 30, 2001, and February 28, 2002, 5263 men were seen in the PHSKC STD clinic for a total of 8467 visits. Of these men, 1047 (20%) had a Focus ELISA performed for HSV-1 and HSV-2, of whom 140 (13%) tested HSV-2-positive. Tested men were predominately white (77%) and without symptoms of STD (65%). Serologic tests were positive in 20 (45%) of 44 men in whom clinicians made a clinical diagnosis of genital herpes and in 120 (12%) of 969 men without such a diagnosis (P <0.0001). In the absence of a clinical diagnosis of genital herpes, serologic tests were more often positive in men with a self-reported history of genital herpes than in those without such a history (54% vs 11.4%, P <0.0001). However, symptoms or signs of genital lesions in the absence of a clinical diagnosis of genital herpes were not associated with an elevated risk of a positive HSV-2 seropositivity (12% vs 12%, P = 0.89). Sera from 108 (77%) of 140 men with positive serologic tests for HSV-2 were available for Western blot testing and are included in the PPV study. (Sera from other men were not stored by the laboratory.) Sociodemographic and clinical characteristics of men included in the study are shown in Table 1.
Among the 108 specimens found to be HSV-2 ELISA-positive by the PHSKC laboratory, 106 (98%) tested ELISA-positive in the UWDVL using the manufacturer's recommended index value cutoff of >1.1 to define positivity, and 2 had index values in the equivocal range (index value 0.9–1.1). HSV-2 EIA results from the 2 laboratories were strongly correlated (R = .80, P <0.0001), although values at the UWDVL tended to be slightly higher. Using Western blot, 86 (80%) specimens tested HSV-2-positive, 17 (16%) tested HSV-2-negative, and 5 (5%) gave atypical Western blot patterns.
Excluding sera with atypical Western blot results and defining sera testing positive by Western blot as true positives, the PPV for the HSV-2 ELISA was 84% and 85% in the PHSKC and UWDVL, respectively. HSV-2 ELISA-positive specimens with positive HSV-2 Western blots had higher median HSV-2 ELISA index values than for those with negative HSV-2 Western blots in both the PHSKC laboratory (5.6 vs 1.6, P <0.0001) and the UWDVL (8.0 vs 2.0, P <0.0001). Western blot positivity also was more common in persons without antibody to HSV-1 detected by ELISA than in persons with HSV-1 antibody (93% vs 76%, P = 0.02) and in men with a self-reported history of genital herpes or symptoms or signs of genital lesions than in men without such a history or clinical presentation, although this latter difference was not significant (88% vs 80%, P = 0.30).
Raising the HSV-2 ELISA index value used to define positivity increased the PPV of the test but decreased the test's sensitivity (Table 2). For example, if sera with ELISA index values ≥3.0 are defined as positive and those with index values of 1.2 to 2.9 as indeterminate, the PPV of the ELISA in the PHSKC laboratory was 98% (72 of 73 Western blot-positive); 30 sera (29%) had indeterminate ELISA results. Using these criteria, 84% of Western blot-positive sera would be identified as positive by the ELISA.
Algorithms for altering the interpretation of the HSV-2 ELISA based on HSV-1 serostatus or history and clinical presentation are presented in Figures 1 and 2. Algorithm 1 interprets the HSV-2 ELISA based on HSV-1 serostatus (Fig. 1). Sera were interpreted as being HSV-2-positive if the ELISA index value was ≥3.0 in the presence of antibody to HSV-1 or >1.1 in the absence of antibody to HSV-1; sera were otherwise interpreted as having an indeterminate result. An alternative algorithm based on patient history and clinical presentation is presented in Figure 2. Table 3 compares the 2 algorithms with the standard test interpretation using an index value >1.1 to define positivity, as well as a revised test interpretation using an index value cutoff of >3.0 to define positivity for all sera, and an algorithm that incorporates both clinical information and HSV-1 serostatus. As shown, simply raising the index value for all sera improves the ELISA's PPV, but results in 29% of specimens being classified as having an indeterminate result. Compared with defining positivity based on an index value of >3.0, integrating either clinical information or HSV-1 serostatus results in fewer sera being classified as having indeterminate results. Incorporation of both clinical information and HSV-1 serostatus further raises the PPV but decreases sensitivity slightly and results in more sera being classified as indeterminate.
The Focus Immunoblot for HSV-2 was performed on 28 samples with ELISA index values of <3.0 as measured in the PHSKC laboratory, of which 20 tested positive (Table 4). (Two specimens were not available for Focus Immunoblot testing.) Among 22 specimens with positive ELISAs for HSV-1 in the PHSKC laboratory, the Focus Immunoblot was positive for HSV-2 in 9 (100%) of 9 HSV-2 Western blot-positive specimens and 7 (64%) of 11 HSV-2 Western blot-negative specimens. Using Western blot as a gold standard to define HSV-2 positivity, the PPV of the Focus Immunoblot for HSV-2 in persons with antibody to HSV-1 and an HSV-2 ELISA index value 1.1 to 2.9 was 56% (9 of 16); the negative predictive value (NPV) of Focus Immunoblot in this group was 100% (4 of 4). Among 21 specimens from persons without a history of genital herpes or signs or symptoms of genital lesions and HSV-2 ELISA index values 1.1 to 2.9, the Focus Immunoblot was positive in 7 (64%) of 11 Western blot-positive specimens and 5 (50%) of 10 Western blot-negative specimens; the PPV of Focus Immunoblot for HSV-2 in persons without clinical evidence suggestive of HSV-2 was 58% (7 of 12); the NPV in this group was 56% (5 of 9).
Table 4 presents results of supplementary HSV-2 testing using the Focus Immunoblot, an avidity assay, and an inhibition assay on 15 specimens with discordant HSV-2 ELISA and Western blot results. (Insufficient specimen was available to test the other 2 specimens with discordant Western blot and ELISA results.) Nine specimens were interpreted as having supplementary testing consistent with an initial false-positive HSV-2 ELISA, 3 had tests suggestive of possible early HSV infection (i.e., seroconversion), results were indeterminate for 2, and one person had culture-confirmed HSV-2.
Based on findings from the PPV study, beginning in October 2002, our clinic began defining HSV-2 ELISA results as indeterminate if the HSV-2 index value was >1.1 to 2.9 in a person with antibody to HSV-1 by ELISA. HSV-2 Western blots were routinely performed on all persons with these indeterminate HSV-2 ELISA results.
A total of 14,966 persons were seen for 20,047 new problem visits at the PHSKC STD clinic between October 2002 and September 2004. Of these persons, 589 were men who had sex with men (MSM) tested for HSV-2 but not HSV-1 as part of a study and were excluded from analysis. Of the remaining 14,338 persons, 3394 (24%) tested for HSV-1 and HSV-2 at least once using the Focus ELISA. Among 1012 women tested, 229 (23%) tested HSV-2-positive, 23 (2.3%) had indeterminate HSV-2 results, and 4 had equivocal HSV-2 results (i.e., index value 0.9–1.1). Among 2382 men tested, 336 (14%) tested HSV-2-positive and 57 (2.4%) had indeterminate HSV-2 results, and 10 had equivocal HSV-2 results. HSV-2 Western blots were performed on 70 sera with indeterminate HSV-2 results, of which 4 had atypical HSV-2 Western blots, 55 had negative HSV-2 Western blots, and 11 tested HSV-2 Western blot-positive (PPV = 17% among those with nonequivocal results). (Ten sera were not sent for HSV-2 Western blot.) Excluding persons with atypical Western blots, 4 (22%) of 18 indeterminate sera from women and 7 (15%) of 48 indeterminate sera from men tested HSV-2 Western blot-positive. Seventeen (30%) of 56 persons with indeterminate ELISAs and negative Western blots returned to the clinic for repeat HSV-2 testing at a median of 189 days (range, 10–556 days); 11 had negative repeat ELISAs, 2 had a persistently indeterminate ELISAs and negative Western blots, 2 had equivocal ELISAs with negative Western blots, and one had an indeterminate ELISA and Western blot.
The Focus HSV-2 ELISA has previously been shown to be sensitive and specific for the diagnosis of HSV-2 in high-prevalence populations. However, the test's performance has not been extensively studied in low-prevalence populations. Among men testing positive for HSV-2 as part of routine clinical care in our STD clinic, a group with a 13% prevalence of HSV-2 by ELISA, only 84% had a positive confirmatory HSV-2 Western blot. This estimate of the test's PPV is similar to a PPV of 81% for the clinical diagnosis of first-episode genital herpes reported in a previous study19 and is consistent with estimates of the test's performance presented in the manufacturer's package insert, but may not be adequate in the judgment of some clinicians. Discordant ELISA and Western blot results were more common in persons with antibody to HSV-1, in those with low ELISA index values, and in persons without a history or symptoms suggestive of genital herpes. These findings led us to adopt an HSV-2 ELISA testing algorithm that included classifying low index value (i.e., <3.0) HSV-2-positive ELISA results as indeterminate in persons with positive ELISA tests for HSV-1. Subsequent evaluation of that algorithm in both men and women confirmed that few persons with indeterminate HSV-2 ELISA results had positive HSV-2 Western blots.
How to use gG-2 based tests in lower HSV-2-prevalence populations is controversial. Eing and colleagues evaluated the performance of different HSV-2 ELISAs and estimated how these tests might perform in a population with a 10% prevalence of HSV-2.12 Comparing different confirmatory testing strategies, they concluded that only initial testing by a gG2 ELISA followed by Western blot testing of all positives would yield a PPV over 80%. However, the HSV-2 Western blot is not widely available, making this approach impractical in most settings. Prince observed that among 11 specimens with HSV-2 ELISA index values of 1.0 to 3.0, 5 had negative Western blots.20 Similarly, a study of sera taken from persons in 7 different nations found marked heterogeneity in the performance of the Focus ELISA, and PPV was improved when the index value cutoff defining positivity was increased.21 Our findings agree with these previous reports documenting poor agreement between HSV-2 ELISA results in the low-positive range and the HSV-2 Western blot, and extend these observations by suggesting possible algorithms of how test interpretation might be altered based on clinical findings or HSV-1 serologic status.
The presented algorithms increased the PPV of the HSV-2 ELISA in our clinic to over 90%, but resulted in approximately 20% of persons previously classified as HSV-2-infected being classified as having indeterminate test results. Which of these algorithms is preferable is likely to vary depending on clinical settings. Basing ELISA interpretation on clinical circumstance would likely require that laboratories report index values and that clinicians either know or would be provided information on how to interpret the ELISA based on the presence or absence of a history of genital herpes or signs or symptoms of genital lesions. Basing interpretation on the presence or absence of antibody to HSV-1, the approach we currently use in our clinic, allows the laboratory to report a more definitive result without additional clinical information, but involves the expense of additional testing. (To save money, HSV-1 testing could be restricted to persons with low index values for HSV-2.) Of note, although the current Centers for Disease Control and Prevention STD treatment guidelines suggest that the Focus Immunoblot may be used as a confirmatory test for ELISA,22 our findings do not support that recommendation.
Our study has several limitations. We initially studied a selected population of low-prevalence STD clinic patients: men who were willing to pay for an HSV-2 test. As a result, our findings reflect the test's performance as we currently use it in our STD clinic and should not be interpreted to define how the test would perform if used to test all STD clinic patients, a group in which the HSV-2 seroprevalence is higher. Also, our PPV study did not include women. The prevalence of HSV-2 is higher in women, and PPV is typically higher in higher-prevalence populations. Consistent with this, in the follow-up study evaluating our current STD clinic algorithm, the ratio of positive to indeterminate HSV-2 ELISA results was higher in women than in men. However, among persons we defined as having indeterminate HSV-2 ELISA results, the prevalence of Western blot positivity was low and comparable in men and women, suggesting that poor agreement between Western blot and the HSV-2 ELISA among HSV-1-positive persons with low HSV-2 index values affects women and men similarly. Because we only performed Western blots on specimens with HSV-2-positive ELISA results, we cannot calculate the sensitivity or specificity of the test using the new index value cutoff we suggest. Also, we cannot say with certainty whether persons who tested positive for HSV-2 by Focus ELISA and negative by Western blot were free of HSV-2 infection. The time from acquisition of infection to seroconversion is shorter for Focus ELISA than Western blot, and it is possible that some of our patients were seroconverting.23 Indeed, at least one patient with discordant HSV-2 ELISA and Western blot results had a positive culture for HSV-2, whereas 3 others had avidity and inhibition assay results that previous studies suggest could be consistent with seroconversion.17,18 As a result, the algorithms we present should be regarded as a conservative approach to interpreting the Focus ELISA, one that seeks to minimize the potential for false-positive results.
In conclusion, using Western blot as a gold standard, we observed that the Focus ELISA had a low PPV among predominantly white, male STD clinic patients, a population with a relatively low prevalence of HSV-2. The PPV was likely higher in women tested in our clinic, in whom the prevalence of HSV-2 was higher, although the absolute risk of a false-positive result was similar in men and women at approximately 2.5%. Based on these findings, clinicians using the test, particularly in low-prevalence populations, should consider selectively using a higher index value to define positivity based either on HSV-1 serostatus or on the presence or absence of clinical findings suggestive of genital herpes. Although these algorithms merit testing in other populations, in our clinic, this approach defines patients whose ELISA results are >90% likely to be confirmed by a positive HSV-2 Western blot. This approach also minimizes the number of persons with indeterminate results. Some patients with indeterminate ELISAs can be definitively diagnosed with HSV-2 using the HSV-2 Western blot, and a newly commercially available inhibition assay also may be useful in defining people who are likely to have false-positive ELISA results.17 However, even when these supplemental tests are available, the true HSV-2 status of some persons with indeterminate HSV-2 ELISAs is uncertain. We advise patients with indeterminate serologic tests for HSV-2 to return for clinical evaluation and virologic testing should they develop genital lesions, and for follow-up serologic testing in the absence of symptoms. Uncertainty in how to interpret discordant HSV-2 ELISA and Western blot results emphasizes the need to develop new HSV-2 diagnostic tests.
1. Fleming DT, McQuillan GM, Johnson RE, et al. Herpes simplex virus type 2 in the United States, 1976 to 1994. N Engl J Med 1997; 337:1105–1111.
2. Corey L. Raising the consciousness for identifying and controlling viral STDs: Fears and frustrations—Thomas Parran Award Lecture. Sex Transm Dis 1998; 25:58–69.
3. Koutsky LA, Ashley RL, Holmes KK, et al. The frequency of unrecognized type 2 herpes simplex virus infection among women. Implications for the control of genital herpes. Sex Transm Dis 1990; 17:90–94.
4. Koutsky LA, Stevens CE, Holmes KK, et al. Underdiagnosis of genital herpes by current clinical and viral-isolation procedures. N Engl J Med 1992; 326:1533–1539.
5. Oliver L, Wald A, Kim M, et al. Seroprevalence of herpes simplex virus infections in a family medicine clinic. Arch Fam Med 1995; 4:228–232.
6. Wald A, Koutsky L, Ashley RL, Corey L. Genital herpes in a primary care clinic. Demographic and sexual correlates of herpes simplex type 2 infections. Sex Transm Dis 1997; 24:149–155.
7. Cowan FM, Johnson AM, Ashley R, Corey L, Mindel A. Relationship between antibodies to herpes simplex virus (HSV) and symptoms of HSV infection. J Infect Dis 1996; 174:470–475.
8. Ashley RL, Wu L, Pickering JW, Tu MC, Schnorenberg L. Premarket evaluation of a commercial glycoprotein G-based enzyme immunoassay for herpes simplex virus type-specific antibodies. J Clin Microbiol 1998; 36:294–295.
9. Groen J, Van Dijk G, Niesters HG, Van Der Meijden WI, Osterhaus AD. Comparison of two enzyme-linked immunosorbent assays and one rapid immunoblot assay for detection of herpes simplex virus type 2-specific antibodies in serum. J Clin Microbiol 1998; 36:845–847.
10. Eis-Hubinger AM, Daumer M, Matz B, Schneweis KE. Evaluation of three glycoprotein G2-based enzyme immunoassays for detection of antibodies to herpes simplex virus type 2 in human sera. J Clin Microbiol 1999; 37:1242–1246.
11. Whittington WL, Celum CL, Cent A, Ashley RL. Use of a glycoprotein G-based type-specific assay to detect antibodies to herpes simplex virus type 2 among persons attending sexually transmitted disease clinics. Sex Transm Dis 2001; 28:99–104.
12. Eing BR, Lippelt L, Lorentzen EU, et al. Evaluation of confirmatory strategies for detection of type-specific antibodies against herpes simplex virus type 2. J Clin Microbiol 2002; 40:407–413.
13. Turner KR, Wong EH, Kent CK, Klausner JD. Serologic herpes testing in the real world: Validation of new type-specific serologic herpes simplex virus tests in a public health laboratory. Sex Transm Dis 2002; 29:422–425.
14. Grimes DA, Schulz KF. Uses and abuses of screening tests. Lancet 2002; 359:881–884.
15. Ho DW, Field PR, Irving WL, Packham DR, Cunningham AL. Detection of immunoglobulin M antibodies to glycoprotein G-2 by western blot (immunoblot) for diagnosis of initial herpes simplex virus type 2 genital infections. J Clin Microbiol 1993; 31:3157–3164.
16. Ashley RL, Militoni J, Lee F, Nahmias A, Corey L. Comparison of Western blot (immunoblot) and glycoprotein G-specific immunodot enzyme assay for detecting antibodies to herpes simplex virus types 1 and 2 in human sera. J Clin Microbiol 1988; 26: 662–667.
17. Hogrefe W, Su X, Song J, Ashley R, Kong L. Detection of herpes simplex virus type 2-specific immunoglobulin G antibodies in African sera by using recombinant gG2, Western blotting, and gG2 inhibition. J Clin Microbiol 2002; 40:3635–3640.
18. Ashley-Morrow R, Friedrich D, Krantz E, Wald A. Development and utility of a type specific antibody avidity test based on herpes simplex virus type 2 glycoprotein G
19. Langenberg AG, Corey L, Ashley RL, Leong WP, Straus SE. A prospective study of new infections with herpes simplex virus type 1 and type 2. Chiron HSV Vaccine Study Group. N Engl J Med 1999; 341:1432–1438.
20. Prince HE, Ernst CE, Hogrefe WR. Evaluation of an enzyme immunoassay system for measuring herpes simplex virus (HSV) type 1-specific and HSV type 2-specific IgG antibodies. J Clin Lab Anal 2000; 14:13–16.
21. Ashley-Morrow R, Nollkamper J, Robinson NJ, Bishop N, Smith J. Performance of Focus ELISA tests for herpes simplex virus type 1 (HSV-1) and HSV-2 antibodies among women in ten diverse geographical locations. Clin Microbiol Infect 2004; 10:530–536.
22. Centers for Disease Control and Prevention. Sexually transmitted treatment guidelines 2002. MMWR Morb Mortal Wkly Rep 2002; 51.
23. Ashley RL. Performance and use of HSV type-specific serology test kits. Herpes 2002; 9:38–45.