Clinical Reactivation of Herpes Simplex Virus Type 2 Infection in Seropositive Pregnant Women with No History of Genital Herpes

Methods

From November 1985 to July 1989, all women who received obstetrical care at three West Los Angeles private obstetrics practices were questioned about genital HSV infections, and 1355 reporting no history of such infection agreed to participate in our study. After informed consent was obtained in accordance with the guidelines of the University of California, Los Angeles, Human Subject Protection Committee, 7 mL of blood was collected from each participant, and the women were given questionnaires (with separate stamped, self-addressed envelopes) for themselves and their husband or partner. The questionnaires solicited detailed demographic and sexual history data.

Between 4 and 8 weeks after enrollment, the women were informed of their herpes simplex virus type 1 (HSV-1) and HSV-2 serologic status; if a woman was HSV-2 seropositive, she was informed of her infection as well as the risk for asymptomatic shedding of virus from the genital tract or lesions. Seropositive women were counseled by their obstetricians to arrange for an examination and culture if papules, vesicles, or pruritus developed on the vulva, buttocks, or thighs, or if dysuria occurred.

From November 1985 to June 1988, cultures for HSV were obtained from any lesions and from the vulva and cervix of seropositive women weekly from 34 or 36 weeks gestation through early labor. Beginning in June 1988, after the American College of Obstetrics and Gynecology recommended abandoning weekly screening cultures [12], cultures for HSV were obtained from lesions and from the vulva and cervix only during early labor. The frequency of visits and the mode of delivery was left to the discretion of the patient and her obstetrician.

Antibodies for HSV-1 and -2 were determined by Western blot analysis previously characterized in our laboratories [13].

Cultures were obtained using a dacron or cotton swab, which was vigorously passed over the cervix and vulva, or lesion, and was then placed in 2 mL of transport media. An aliquot of each sample was placed onto WI-38 cells for viral isolation using standard tissue-culture techniques. When a cytopathic effect was observed, HSV was confirmed and typed by immunofluorescence staining (Microtrak HSV-1/HSV-2 Kit, Syva Company, Palo Alto, California).

The individual data items from the confidential questionnaire were tabulated. Descriptive statistics and CIs were calculated [14]. Statistical tests used included the two-tailed chi-square test, the Student t-test, and stepwise logistic regression [14]. To evaluate the model, we classified each woman in our data set as either HSV-2 seropositive or seronegative using various cutpoints for the predicted probability to calculate the sensitivity and specificity.

Results

Herpes Simplex Virus Serologic Testing

Serologic test results are shown in Table 1. Of the 1355 study participants, 53 (4%) were excluded from further analysis because of an indeterminate serologic test result or missing serum specimens. Nearly one fourth of enrolled women were HSV negative, and nearly one third were HSV-2 seropositive. The 619 (46%) women who returned their questionnaires did not differ in seropositivity rate from the larger population: One hundred ninety (32%) were HSV-2 seropositive, of whom 119 (61%) were also HSV-1 seropositive. Two hundred fifty-eight (42%) of the subgroup had only HSV-1 antibody, and 165 (27%) were HSV seronegative.

Demographic Characteristics

Most of the women enrolled in the study were white college graduates and professionals and were married and of upper-middle socioeconomic status. Among the 619 women who returned questionnaires, the rates of HSV-2 seropositivity and seronegativity did not differ greatly by age (mean age ±SD, 31.8 ± 4.3 years compared with 31.3 ± 4.5 years, P > 0.2); marital status (88% compared with 93% were married, P = 0.02); ethnicity (91% compared with 86% were white, P = 0.11); place of residence before Los Angeles (P > 0.2); education (88% compared with 91% completed college or professional school, P = 0.2); occupation (56% compared with 66% were employed in managerial or professional positions, P = 0.05); income (72% compared with 79% had family incomes greater than $40 000/year, P = 0.18); or religion (P > 0.2 for comparison of Catholic, Protestant, Jewish, and “other”).

Reproductive and Sexual Histories

Data on the sexual and reproductive histories of our study participants and their husbands or partners are shown in Table 2. Compared with HSV-2-seronegative women, HSV-2-seropositive women had their first sexual intercourse at a significantly earlier age and had a substantially greater total lifetime number of sexual partners. Seropositive women reported having slightly more pregnancies (2.6 ± 1.3 compared with 2.3 ± 1.2, P = 0.0005) and therapeutic abortions (0.84 ± 1.0 compared with 0.52 ± 0.83, P = 0.0001) than did seronegative women. No difference in the number of live births was found between the two groups (data not shown). A history of vaginal “yeast infection” was noted to occur significantly more frequently in seropositive women, as was a history of “any” sexually transmitted disease. Of the sexually transmitted diseases, gonorrhea (6.2% compared with 2.2%, P = 0.01) and Trichomonas infection (15% compared with 9.6%, P = 0.05) were more frequent in the seropositive than in the seronegative group. No intergroup differences were found, however, regarding history of papillomavirus infection, chlamydial infection, syphilis, or “other sexually transmitted diseases not including herpes simplex virus” (data not shown). Seropositive women were much more likely to have or to have had a sexual partner with genital herpes.

Sexual Histories of Partners

Six hundred ten husbands or other partners of the pregnant women returned questionnaires. Partners of seropositive women were slightly older than those of seronegative women (mean age, 35.3 ± 6.1 years compared with 34.0 ± 6 years, P = 0.02). This group also reported having had more sexual partners during their lifetime, first sexual intercourse at an earlier age, a greater number of sexually transmitted diseases other than genital herpes, and an increased prevalence of genital herpes. Sexual histories are summarized in Table 2.

Logistic Regression Model

Three variables related to sexual history were selected as statistically significant for inclusion in the logistic model: These included total (lifetime) number of sexual partners, past or present partner with a history of genital herpes, and a history of yeast infections. Codes for number of sexual partners were as follows: 1 to 9 for reported values 1 to 9; 15 for reported values 10 to 19; 25 for values 20 to 29; and 35 for values 30 to 160. Scoring for the other two variables was as follows: no = 0 and yes = 1. Of the 619 women returning questionnaires, 568 had sufficient data for the logistic model analysis. The logistic model coefficients (± SE) and the estimated odds ratios are given in Table 3. Plots of the sensitivity and specificity of the logistic model for various cutpoints of the predicted probability of seropositivity are presented in Figure 1.

Compared with the use of single risk factors, the multivariate logistic model improved both the sensitivity and the specificity of predicting seropositivity. Using a cutpoint of 0.25 for predicted probability in the logistic model, we found that 68% of seropositive women and 62% of seronegative women were correctly identified (see Figure 1).

Asymptomatic Shedding of Herpes Simplex Virus

Viral cultures obtained from the genital tract for detection of asymptomatic HSV infection were available from 264 women late in gestation and from 140 women at term. Of 1020 (0.5%) cultures obtained late in pregnancy, 5 (obtained in five different women) were positive for HSV, and 0 of 140 cultures obtained during labor (within 48 hours of delivery) were positive, for an overall asymptomatic shedding rate of 0.43%.

Clinical Genital Herpes

Of 264 seropositive women followed to term, 43 (16%) recognized genital lesions (subsequently confirmed as herpes infection by their obstetrician) for the first time during pregnancy. Twenty-four of 32 lesions cultured grew HSV-2. Twenty-four of 43 women (56%) first recognized HSV lesions during the third trimester. Sixteen of these women delivered their babies by cesarean section because of herpes infection.

Primary Genital Herpes Simplex Virus Infection during Pregnancy

A history of genital herpes infection was reported by 96 (15.7%) partners. Of the 53 corresponding pregnant women followed to term, 47 (85%) were seropositive for HSV-2 at enrollment. No symptoms of primary or first-episode genital herpes, positive viral cultures, or seroconversion to HSV-2 occurred in the eight initially seronegative female partners.

Discussion

The seroprevalence of HSV-2 infection in diverse North American populations of reproductive age has ranged between 7% and 65% [1-7]. Our study was done in a large cohort of pregnant women who were of upper-middle socioeconomic status and who had no history of genital herpes. Seroprevalence of HSV-2 was high (32%) in this group, but asymptomatic HSV shedding was infrequent (0.43%). A number of women (16%), many of whom were near term, recognized herpetic lesions for the first time during their pregnancy. As a result of careful assessment of the genitalia for herpetic lesions, 16 clinical outbreaks of herpes were recognized near term; the women involved all delivered their babies by cesarean section.

The risk factors for infection in our population were similar to those identified in a population of women who were lower in socioeconomic status and of different racial and ethnic mix: These risk factors included younger age at first sexual intercourse, a greater number of sexual partners, and a history of sexually transmitted diseases. In addition, our study participants reported that they had a history of yeast infections and a past or present sexual partner with genital herpes. Even so, the logistic regression model only predicted seropositivity in 68% of the participants and seronegativity in 62%. Low sensitivity and specificity of risk factors has previously been reported [15].

Only 3% to 22% of HSV-2-seropositive persons report a history of genital herpes [1-7, 16]. Our study is the first to show that asymptomatic, seropositive women can recognize lesions for the first time during pregnancy. Although the frequency of recurrences (16%) was unexpected, the actual rate of reactivation in this population may have been underestimated because the women's understanding of symptoms associated with genital herpes varied. Asymptomatic, HSV-2-seropositive women have been “taught” by others to recognize symptoms of genital herpes [15]. Detailed education and close follow-up during a mean of 5 months enabled 42% of a small group of nonpregnant, seropositive but previously asymptomatic women in Seattle to recognize clinical reactivation of genital herpes [15].

Reactivation at delivery is important because of the potential for causing neonatal herpes. Exposed infants have a 3% to 8% chance of developing disease [17, 18], and 40% to 70% of infected infants are born to women with reactivation herpes [17, 18]. Recent studies have assessed the likelihood of neonatal herpes due to newly acquired or reactivation herpes during pregnancy [17, 19]. Of 10 cases of neonatal herpes in Seattle, 2 followed maternal primary, 4 followed first-episode nonprimary, and 4 followed reactivation infections with HSV type 2 [17].

Specific serologic testing during pregnancy could identify women at risk for transmitting neonatal herpes, namely, HSV-2-seropositive women and seronegative women with seropositive husbands or partners. To reduce the chance of neonatal herpes, women at risk could be educated to recognize genital herpes. In addition, fetal-scalp monitoring could be avoided, and cesarean section could be implemented were lesions detected. The use of condoms might also be advocated to prevent transmission of herpes to seronegative, pregnant women who are at risk [20].

Although currently available serologic tests do not accurately differentiate type 1 from 2 herpes simplex virus antibodies [21, 22], several sensitive, specific, and economical tests have been developed and should soon be available. Recently, a Western blot serologic assay similar to ours was found to be the most practical way to identify women with established subclinical HSV-2 infection; however, viral culture was superior in identifying those with newly acquired herpes infections [15]. Unfortunately, diagnostic tests rapid enough to be done during labor have been found insensitive [23]. Polymerase chain reaction, although highly sensitive [24], is not yet practical for screening during labor.

Until advances allow rapid and accurate diagnosis or prevention of herpes simplex virus at delivery, we recommend that all pregnant women be educated regarding the high prevalence of HSV-2 infection, the inability to diagnose women at risk by sexual risk factors or by licensed serologic tests, and the mild and varied symptomatology. Women should be encouraged to bring suspected lesions to the attention of their obstetrician.