Prevalence and Correlates of Herpes Simplex Virus Type 2 Infection in a Population-Based Survey of Young Women in Low-Income Neighborhoods of Northern California

HERPES SIMPLEX VIRUS type 2 (HSV-2) is one of the most common causes of genital ulcer disease worldwide,1 though infection is often asymptomatic.2 HSV-2 is transmitted primarily through sexual contact. After acute primary infection, HSV-2 can persist for life and can periodically reactivate to produce genital lesions and increased viral shedding. A severe consequence of HSV-2 infection, particularly of acute primary infection, in pregnant women is potential transmission to the newborn. Neonatal infection can manifest as a generalized systemic infection, encephalitis or localized central nervous system infection, or infection of the skin, eyes, and mouth. In the United States, the presence of HSV-2-specific antibodies is associated with the female sex, black race, older age, and lower socioeconomic status.2-5 Behavioral risk factors for HSV-2 infection are sexual orientation, number of sexual partners, years of sexual activity, age of sexual onset, and history of other sexually transmitted diseases (STDs).4,6,7

Herpes simplex virus type 2 infection has gained particular attention as a hypothesized causal factor in the transmission of HIV.8-12 The disruption of genital mucosa caused by HSV-2 infection may provide a mechanism for easier acquisition of HIV.13,14 Coinfection with HSV-2 and HIV may also facilitate the transmission of both viruses to others through increased reciprocal viral replication, increased viral shedding, and increased frequency of genital ulceration as a portal for viral excretion15-17.

As a biologic marker of high-risk sexual behavior, HSV-2 seropositivity can identify populations or sexual networks in which the potential for the spread of HIV is high.1,6,18,19 Therefore, new serologic tests that specifically identify HSV-2 infection have far-reaching public health applications. HSV-2 seropositivity provides an objective marker for sexual risk behavior in cross-sectional surveys. HSV-2 prevalence can be used to monitor trends in the interconnected STD and HIV epidemics, predict future transmission, and identify geographic areas where undetected HIV infection may be found. Population-level data on HSV-2 infection can be used to target HIV and STD screening, health education outreach, and other prevention activities. However, few studies have measured HSV-2 prevalence outside clinic-based or convenience-based samples.3,4

A recently completed survey of young women residing low-income neighborhoods of five Northern California counties (the Young Women's Survey) provided the opportunity to obtain population-based estimates of HSV-2 prevalence.20 This article reports the survey-weighted prevalence of HSV-2 and associated demographic characteristics, biologic correlates, and behavioral factors.

The Young Women's Survey was a single-stage, cluster sample, population-based, door-to-door, cross-sectional survey designed to measure the prevalence of HIV, STD, and related risk behavior. Five counties in Northern California participated: Alameda, Contra Costa, San Francisco, San Joaquin, and San Mateo. A detailed description of the Young Women's Survey study methods, study population, and primary outcomes is available upon request.20

The target population was young women residing in low-income neighborhoods. Eligible participants included women age 18 to 29 years who were fluent in English or Spanish and resided in the target area. The target area was defined as the 1990 US Census block groups below the tenth percentile for median household income. City blocks were randomly selected from the target area. All households within the selected blocks were enumerated. All women in enumerated households who met the eligibility criteria were recruited to participate. The study sought to enroll approximately 500 women per county and 2,500 women overall. A total of 24,113 inhabited dwellings were enumerated in 448 randomly selected street blocks within the five-county target area. Contact was made with a resident in 19,546 dwellings (81.1%). Of 3,560 eligible women identified, 2,547 (71.5%) agreed to the interview and to have blood drawn. Recruitment occurred from April 1996 to January 1998.

A 45-minute face-to-face structured interview was administered in private in English or Spanish to obtain information on sociodemographic characteristics, medical history, drug-use history, sexual behavior, and knowledge, attitudes, and beliefs about HIV transmission and prevention. Screening for biologic markers of infectious diseases varied by county: HIV and syphilis testing were conducted in all five counties; hepatitis C, HSV type 1 (HSV-1), and HSV-2 testing were done in four counties; and hepatitis B, gonorrhea, and chlamydia testing were done in three counties. In the four counties conducting HSV type 1 and 2 testing (Alameda, San Joaquin, San Mateo, and San Francisco), sufficient sera for testing were available for 1,657 of 2,081 eligible women (79.6%). Of these, 22 specimens (1.3%) had nonspecific HSV results and were not included in subsequent analysis (n = 1,635).

A strip-recombinant immunoblot assay (RIBA HSV Type 1/Type 2 SIA; Chiron, Emeryville, CA) was used to detect and differentiate HSV-1-specific and HSV-2-specific antibodies based on recombinant antigen bands for gG1, gB1, gG2, and gD2.21 The presence of HIV antibodies was demonstrated by enzyme immunoassay (Abbott Laboratories, Abbott Park, IL) and confirmed by immunofluorescent antibody (Neufeld, Vienna, Austria). Sera were tested for syphilis by rapid plasma reagent or VDRL; reactive specimens were confirmed by microhemagglutination-Treponema pallidum. Hepatitis B core antibody was detected by enzyme immunoassay (Abbott Laboratories) and surface antigen was detected by microparticle enzyme immunoassay (Abbott Laboratories). Ligase chain reaction (Abbott Laboratories) was used to detect gonococcal and chlamydial DNA in urine specimens.

Point prevalences, 95% CI, and odds ratios (OR) were constructed accounting for the single-stage, cluster-sample survey design using Stata software version 5.0 (Stata Corp., College Station, TX).22 Estimates were adjusted using probability weights for the four-county target population, and 95% CI were adjusted to account for homogeneity within the primary sample units (i.e., city blocks). Multiple logistic regression analysis adjusting for the survey design was used to identify independent predictors of HSV-2 seropositivity using a backward and forward stepwise process. Demographic and behavioral variables with statistically significant bivariate associations with HSV-2 were first entered into a model. Variables of a priori interest, such as HSV-1 seropositivity, were also entered to test specific hypotheses. A stepwise process removed nonsignificant variables that were not important confounders from the full model. The final model retained variables with a significance level of P < 0.05. A forward stepwise procedure using P to enter at less than 0.05 converged on the same set of predictor variables as the backward process. Interactions between age, race or ethnicity, and HSV-2 prevalence were explored through stratified analyses.

Adjusting for the survey design, the population represented by the four-county study sample was 38.3% black, 32.9% Latino, 15.3% white, and 13.5% Asian or Pacific Islander or other race. Most women (70.0%) were born in the United States; 17.5% were born in Mexico, and 12.5% were born in other countries. Fewer than one third of women (32.5%) had an education beyond high school. The median household income was between $500 and $1000 per month. The median age was 24 years. Nearly all women (96.6%) reported ever having sexual intercourse; 0.6% reported sex with women only, 9.4% reported sex with women and men. Among women reporting any sexual intercourse with men, 21.1% had more than 10 lifetime male sex partners, 38.6% had vaginal intercourse before the age of 16 years, and 11.5% reported exchanging sex for drugs, money, or other needs. Nearly one third of women (32.2%) had a history of STD (syphilis, chlamydia, gonorrhea, pelvic inflammatory disease, genital warts, hepatitis B, or trichomoniasis). Three percent of women reported ever injecting drugs, and 16.7% reported ever using cocaine.

The overall weighted prevalence of HSV-2 antibodies in the four-county target population was 34.8% (95% CI, 30.4-39.2). The crude prevalence was 32.7% (unweighted 95% CI, 30.4-35.0), based on 534 HSV-2 positive results of 1,635 women screened.

Table 1 presents HSV-2 prevalence by demographic characteristics. HSV-2 prevalence differed by county, with Alameda having the highest prevalence (51.4%). By race, black women had the highest prevalence of HSV-2 (58.6%); Latinos (18.3%) and Asian or other (16.8%) had lower prevalence of HSV-2 compared with whites. Women born outside of the United States had a lower prevalence of HSV-2 compared with US-born women. Markers of lower socioeconomic status, lower monthly income, and less education were associated with increased HSV-2 prevalence (Figure 1). HSV-2 prevalence increased with older age, from 27.8% among women 18 to 21 years to 45.8% among women 26 to 29 years. Higher HSV-2 prevalence was also associated with having any children.

TABLE 1. Prevalence of HSV-2-Specific Antibodies by Demographic Characteristics*

Fig. 1. HSV-2 and socioeconomic status, San Francisco Bay area young women.

Table 2 displays behavior associated with HSV-2 seropositivity. All women with HSV-2 reported having sex with men. HSV-2 antibodies were not detected among women reporting sex with women only or among women reporting no history of sexual intercourse. HSV-2 prevalence increased with increasing number of lifetime male sexual partners and with earlier age of sexual onset. The highest prevalence of HSV-2 (72.7%) for any group was found among women who exchanged sex for money, drugs, or other needs. Women who used cocaine and injected drugs were more likely to be HSV-2 positive than nonusers.

TABLE 2. Prevalence of HSV-2-Specific Antibodies by Risk Behavior*

Prevalence of HSV-2 was higher among women who recalled that a doctor or nurse had told them they had any STD. Of note, 71.9% of women who reported that a doctor or nurse told them they had genital herpes were HSV-2 positive. However, few women overall (3.1%) reported ever having had genital herpes-a finding in contrast to the high prevalence of HSV-2 antibodies in this population. Among women who tested HSV-2 seropositive, only 6.3% recalled having been previously diagnosed with genital herpes. Despite significantly higher prevalence of HSV-2, only 2.0% of black women compared with 8.4% of white women recalled being diagnosed with herpes. Of note, the association of increased HSV-2 prevalence and black race persisted when stratified by lifetime male sex partners (Figure 2).

Fig. 2. HSV-2 prevalence by ethnicity and number of lifetime male sexual partners, San Francisco Bay area young women.

Associations between HSV-2 and markers for other infectious diseases are shown in Table 3. Chlamydia, hepatitis B, and hepatitis C infection were all significantly associated with HSV-2 antibodies. No association was found between HSV-2 antibodies and syphilis or gonorrhea; however, the number of women testing positive for these infections was small (15 and 10, respectively). HSV-1 and HSV-2 antibodies were not associated in bivariate analysis. The one study participant with HIV tested positive for HSV-2.

TABLE 3. Prevalence of HSV-2-Specific Antibodies and Markers of Other Infectious Diseases*

Independent correlates of HSV-2 seropositivity in multiple logistic regression analysis are shown in Table 4. The association of HSV-2 infection and black race persisted after controlling for other demographic characteristics and risk behavior. The negative association of HSV-2 infection with Latino or Asian or other race did not. Other independent demographic correlates of HSV-2 prevalence were older age, lower monthly income, and having any children. Several markers of sexual risk were independently associated with HSV-2 infection, including lifetime number of male sex partners, earlier age at first vaginal intercourse, exchanging sex for money or drugs, and history of STD. Cocaine use also remained an independent predictor of HSV-2. Current infection with chlamydia was associated with HSV-2 in multivariate analyses; however, it is omitted from the final model presented here because women in San Joaquin county (27% of the sample) were not screened for chlamydia. The presence of HSV-1 antibodies was negatively associated with HSV-2 antibodies in multivariate analysis. Variables eligible for inclusion but not retained in the final model were county of residence, country of birth, lower education, marital status, injection drug use, reported genital ulcer diagnosis, and markers for syphilis, gonorrhea, hepatitis B, and hepatitis C.

TABLE 4. Independent Associations With HSV-2 Seropositivity*

Our findings support the use of HSV-2 serology as a biologic marker of lifetime sexual risk, especially in studies in which detailed behavioral data are not available. The high prevalence of HSV-2 and its close association with sexual risk behavior point to high risk for HIV and other STDs in the target population. In agreement with other studies, we confirm that HSV-2 seropositivity is associated with greater number of lifetime sexual partners, younger age onset of sexual activity, trading sex for money and drugs, and history of other STD.4,6,7 In contrast to a recent study among black college women,23 HSV-2 prevalence increased with number of lifetime sexual partners among the black women in our population. The absence of HSV-2 antibodies among women reporting no sexual intercourse supports the reliability of self-reported information regarding sexual behavior, and that HSV-2 is transmitted primarily by genital sexual intercourse and not by casual contact.2,3 Cocaine use, which has been linked in other studies to high-risk sexual activity and dissemination of STD,24,25 was also associated with HSV-2 seropositivity.

Few women reported ever being diagnosed with genital herpes despite the high prevalence of HSV-2 infection. The finding suggests a high level of subclinical infection, unrecognized signs and symptoms, barriers to accessing health care, underdiagnosis, or underreporting having been diagnosed with infection. Although black women were more than twice as likely to be HSV-2 positive than women of other races, white women were four times as likely to report a history of genital herpes. Barriers to health care and a lack of health education information may be particularly severe for black women in the target area.

Our study has several limitations. Despite achieving our goals of a sample size of 2,500 women, a greater than 80% household contact rate, and a greater than 70% participation rate, bias due to nonresponse is possible. Further, serologic specimens were not available for 17.8% of participating women, most of whom were from San Joaquin county. Omitting San Joaquin from the analysis did not substantially change the primary findings. Moreover, women with HSV test results did not differ from those without HSV test results with respect to age, race, income, lifetime number of male sex partners, or history of genital herpes. Another limitation is that the assay used to identify HSV-2 antibody (RIBA HSV Type 1/Type 2 SIA, Chiron) is no longer available for use or for further evaluation of its sensitivity and specificity against other assays. As a result, comparisons with other studies will be difficult if, for example, the assay has a higher false-positive rate. Nonetheless, we believe our estimates of HSV-2 prevalence and correlates of infection to be generally valid. Preliminary results found the assay's sensitivity to HSV-2 to be 98.2% and specificity to be 99.4% compared with Western blot analysis.21 Of note, weighted prevalence estimates and associated 95% CI did not differ greatly from crude estimates. This is likely the result of well-proportioned recruitment by county, a large overall sample size, a high degree of heterogeneity within sampled clusters (city blocks), and the large number of clusters sampled (n = 448).

Despite potential limitations, our data provide rare population-based information regarding HSV-2 prevalence and associated risk behavior. Geographically characterizing HSV-2 infection among young women provides multiple public health benefits. Population-based data among young women contribute to the development and planning of programs to counsel pregnant women on potential risks of HSV-2 to their newborns and prevention and treatment options. HSV-2 prevalence data can also be used to identify populations and locations where other STDs may be undetected. Identifying areas with high HSV-2 prevalence enables the concentration of HIV and STD prevention activities, such as targeted health education, community outreach, and mobile STD and HIV screening and treatment services. Population-based estimates of HSV-2 can be used to measure the impact of prevention activities and serve as a community-level biologic marker of sexual risk. HSV-2 prevalence data may be particularly useful in monitoring risk in populations where the incidence of HIV may be low, but increased transmission may be imminent.