Increased Genital Shedding of Herpes Simplex Virus Type 2 in HIV-Seropositive Women

Methods

From September 1991 to May 1994, 106 HIV-seropositive and HSV-2-seropositive women were enrolled as part of the Women's AIDS Cohort Study, a prospective study of the natural history of HIV infection in women. Women were recruited from two HIV primary care clinics at a large inner-city public hospital. Women who had a history of AIDS-defining illness [7] or who were pregnant were not eligible for enrollment. All women eligible for enrollment were asked to participate, and approximately 90% agreed. Seventy sexually active HIV-seronegative and HSV-2-seropositive controls from the same community were enrolled as part of the Heterosexual AIDS Transmission Study, a prospective evaluation of the behavioral and biologic risk factors for HIV acquisition [8].

Serum samples from both groups were screened for the presence of antibodies to HIV-1 using the Recombigen enzyme immunoassay (Cambridge Biotech Corporation, Worcester, Massachusetts). Samples positive by enzyme immunoassay were tested by Western blot analysis (Dupont Company, Wilmington, Delaware). A specimen was considered positive if any two of three (p24, gp41, gp120/160) bands were detected.

Information pertaining to basic demographics, socioeconomic status, history of alcohol or drug use, and sexual practices was obtained with a standard questionnaire. A medical history and a physical examination were done. Lymphocyte phenotyping was done for HIV-seropositive women by flow cytometry using monoclonal antibodies (Becton-Dickinson, San Jose, California).

For all HIV-seropositive and HIV-seronegative women, serologic evidence of previous infection with HSV-2 was determined by Western blot. Herpes simplex virus type 2 lysate from standard HSV-2 strains was separated using gel electrophoresis and transferred to nitrocellulose strips. Strips were incubated with the patients' serum and then developed using a biotinylated horseradish peroxidase system. We assayed positive and negative controls for HSV-1, HSV-2, and combined HSV-1 and HSV-2 infection on each run. Serologic diagnosis of HSV-2 was made by the identification of the 92-kd gG glycoprotein band pattern, which previously has been shown to be specific for HSV-2 [9].

All women had gynecologic examinations. At enrollment, we collected cervical and vulvar specimens for HSV culture from all of the HIV-seropositive women, regardless of HSV-2 serostatus or the presence of symptoms consistent with HSV genital infection. Dacron and plastic swabs (S/P Sterile Swab, Baxter Healthcare Corporation, McGaw Park, Illinois) from both sites were placed together into a single vial of viral transport medium (2.0 mL of Hanks balanced salt solution, 0.5% bovine serum albumin, 20 mM HEPES, gentamicin, vancomycin, and nystatin). After specimens were vortexed for 15 seconds, 0.2 mL of material was inoculated into two tubes each of A549 and human neonatal kidney cells (BioWhittaker Incorporated, Walkersville, Maryland). We observed monolayers for 2 weeks for the development of cytopathic effect. We identified isolates by immunofluorescence using type-specific monoclonal antibodies (Syva Corporation, Palo Alto, California). If a genital ulcer was noted during the examination, a separate specimen for viral culture was obtained from that site. All cultures were done on the day of collection.

We obtained genital specimens for HSV-2 culture from 56 (80%) of the 70 HIV-seronegative and HSV-2-seropositive women during a follow-up visit. At the time of culture, none of these women was noted to have a genital lesion or symptoms consistent with HSV infection, and none of the women in either group was receiving therapy with agents active against HSV.

Differences in categorical variables were tested for statistical significance using the generalized Fisher exact test. Associations by CD4 grouping were evaluated by the exact trend test and two-sided mid-P values (StatXact, Cytel Software Corporation, Cambridge, Massachusetts).

Results

Demographic characteristics of study participants are shown in Table 1. There was a greater proportion of black women among the HIV-seronegative controls than among the HIV-seropositive women. Although race differed between HIV-seropositive and HIV-seronegative patients, the prevalence of recovery of HSV-2 infection by race did not differ (P = 0.75), making it unnecessary to control for race when comparing HSV prevalence by HIV serostatus.

Fourteen (13.2%) of the HIV-seropositive and HSV-2-seropositive women were positive for HSV-2 infection by culture from genital specimens, compared with 2 of 56 HIV-seronegative and HSV-2-seropositive women (P = 0.04; odds ratio, 4.1 [95% CI, 1.0 to 27.4]). Three of these HIV-seropositive women (21.4%) had lesions consistent with HSV genital infection at the time of their evaluation. Eleven (78.6%) were asymptomatic. All of the isolates recovered were HSV-2.

The rate of HSV-2 genital shedding increased as the CD4 cell count decreased (P < 0.025; Table 2). This trend persisted when only asymptomatic shedding was considered, but the finding did not reach statistical significance (P < 0.086).

Discussion

The prevalence of combined symptomatic and asymptomatic genital HSV-2 shedding was significantly greater in HIV-seropositive women than in HIV-seronegative controls drawn from the same community and was associated with declining CD4 cell counts. The rate of asymptomatic HSV-2 shedding by CD4 cell count did not reach statistical significance.

Asymptomatic shedding of HSV-2 in women is common. Almost all women who are seropositive for HSV-2 antibody will shed the virus asymptomatically at one time or another [1]. The isolation rate of virus varies in proportion to the frequency of sampling [1-310, 11]. Studies conducted with women who are presumably not infected with HIV and using traditional culture methods have shown an isolation rate of virus of approximately 1% of all specimens taken. In our study, the rate of isolation of HSV-2 from asymptomatic HIV-seropositive women (11%) was 11 times greater than that of these historical controls, but it was only three times greater than that of our own HIV-seronegative controls (11% compared with 3.6%).

The most conspicuous shortcoming of our analysis is its reliance on cross-sectional data. Shedding of HSV-2 is a dynamic process, which may transiently occur and abate. Measuring its occurrence at a single point in time does not necessarily reflect overall shedding rates. Many more patients may have shed the virus on days preceding or following their clinical evaluation.

These findings are important. Sexual transmission of HSV-2 occurs during periods of symptomatic and asymptomatic shedding [12-15]. Although viral shedding is quantitatively lower in asymptomatic periods, the attributable risk for transmission of HSV-2 is probably much higher because sexual activity is less likely to be restricted in asymptomatic women. Concomitant HIV and HSV infection may substantially increase the risk for HSV transmission through increased instances of symptomatic, and more importantly, asymptomatic shedding. The risk for neonatal acquisition of HSV infection from pregnant HIV-seropositive women may likewise be higher than that from HIV-seronegative women [16-18]. Given these concerns, the incidence of these phenomena should be examined further.

Presented in part at the 33rd Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, Louisiana, October 1993.

Dr. Feldman: Box 43, Clarkson Avenue, Brooklyn, NY 11203.

Dr. Chirgwin: Box 77, 450 Clarkson Avenue, Brooklyn, NY 11203.

Dr. Zenilman: Johns Hopkins University School of Medicine, Ross 1159, 720 Rutland Avenue, Baltimore, MD 21205.

Dr. Clarke: Box 37, 450 Clarkson Avenue, Brooklyn, NY 11203.

Dr. DeHovitz: Box 43, 450 Clarkson Avenue, Brooklyn, NY 11203.

Dr. Landesman: Box 122, 450 Clarkson Avenue, Brooklyn, NY 11203.

Dr. Minkoff: Box 24, 450 Clarkson Avenue, Brooklyn, NY 11203.