Do antepartum herpes simplex virus cultures predict intrapartum shedding for pregnant women with recurrent disease?

OBJECTIVE: To examine antenatal screening as a predictor of intrapartum shedding of herpes simplex virus (HSV) and to determine its usefulness in guiding the appropriate route of delivery for patients with recurrent HSV in pregnancy. METHODS: A population of 198 pregnant women with a history of recurrent genital HSV were cultured in the last weeks of their pregnancy by specially-trained personnel and intrapartum by their delivering attendants. RESULTS: Of cultures from a total of 906 antenatal visits, 17% were culture positive, with an asymptomatic shedding rate of 3.4%. Asymptomatic shedding occurred in 12.6% of women. Over the 8-week antepartum period, viral culture-positivity rates for each visit ranged from 11% to 19.5%. This provided an expected delivery culture-positivity rate of 15.3%. However, actual intrapartum viral culture positivity occurred in only three of 191 women (1.5%; P < 0.001). Because previous studies have suggested antepartum culture positivity fails to predict intrapartum viral shedding, evaluations, including cultures, as well as predictive values for subsequent culture positivities, were determined under the supervision of an infectious disease specialist. Under these conditions, positive predictive values were 59% when the interval between visits was 2 days, but only 19% when days between visits were >2 (P < 0.0001). No cases of neonatal herpes were seen in this population, although cesarean deliveries were performed in 31% of the patient population, with genital herpes as the indication for 56% of those. CONCLUSIONS: Antepartum serial screening by viral culture is not predictive of an infant's risk of intrapartum viral exposure when conducted at weekly intervals. However, more frequent assessments of patients can be predictive of an infant's exposure risk to HSV; for patients with frequent recurrent disease near term or primary infection in pregnancy, frequent late antepartum screening may be appropriate.

livery was recommended to prevent transmission of virus to the infant in the setting of either clinical or virologic evidence of active infection. Arvin et al., however, demonstrated that weekly antepartum cultures in the third trimester were not actually predictive of asymptomatic intrapartum shedding. 7 This has led to revision of the guidelines, with the elimination of serial antepartum viral screening. 8 Further arguments against serial screening of symptomatic patients include the premise that most women with genital HSV are infected subclinically. Furthermore, Brown et al. have demonstrated that neonatal risk is highest in the setting of recent acquisition of infection. 9 In fact, Randolph et al. have recently suggested the elimination of cesarean delivery for active recurrent disease intrapartum. 1 Nevertheless, recommendations of these authors have not been widely accepted as yet, 1 and recurrent disease still accounts for up to 50% of neonatal cases. Hence, the current management of patients with recurrent genital herpes remains unsettled, the decision regarding cesarean delivery being largely left to clinical judgment regarding evident symptomatic lesions in labour. As a compromise, some investigators have chosen to suppress recurrences with oral acyclovir during the final 2-4 weeks of gestation, z,3 hoping to avoid cesarean deliveries.
During the mid-1980s, a group of patients considered at high risk for genital herpes were referred to a university teaching hospital clinic for prenatal assessment and follow up. Because of the recent changes in our understanding of the pathogenesis of neonatal infection, we analysed data collected from that period to extend our understanding of the natural history of the disease. We report here our experience of antenatal screening to predict intrapartum shedding and its value in guiding the appropriate route of delivery for patients with recurrent herpes in pregnancy.

Patient Population
The study population consisted of 198 pregnant women referred from May 1985 to July 1987 to the University of British Columbia Herpes Clinic, Vancouver, Canada, by their local medical practitioners because the women had a previous history of genital HSV, previous genital culture positive for HSV, or their partners had a history of HSV. After written consent was obtained, patients completed a questionnaire detailing patient demography, pregnancy and sexual history, and past sexually-transmitteddisease history with detailed questioning for oral and genital herpes (including a photographic questionnaire). Serum was obtained from each woman at initial assessment for HSV type-specific antibodies.
Clinical evaluation of patients occurred at weeks 32, 34, and 36 of gestation and weekly to delivery according to standard practice at that time. Evaluations were performed by an infectious disease specialist, and at each visit symptoms and signs of any recent genital HSV were documented. In addition, cultures were taken from any apparent clinical lesion(s), or, if no lesions were present, a general sweep from the external genital area (a Dacrontipped metal swab previously moistened in normal saline was wiped firmly over the pubic area, clitoris, periurethral area, labia minora and majora, fourchette, perineal and perianal areas), and a separate swab was also collected from the ectocervix.
At the conclusion of each visit, a recommendation was made to the attending physician as to the ideal mode of delivery. Where no clinical disease was evident, vaginal delivery was recommended, whereas for an active episode (symptomatic/ asymptomatic), cesarean delivery was considered the ideal delivery mode, if labour intervened. For those with an active episode antepartum, more frequent follow up of the patient was arranged at the Herpes Clinic, until a return to a culture-negative and asymptomatic clinical status was achieved. At birth, repeat cultures (as described above) were performed of the upper and lower genital tract by the attending obstetrician or physician. In addition, the fresh placenta, the infant (mouth, axilla, groin), and, in the case of those delivered by cesarean delivery, amniotic fluid, were also cultured for HSV. A research nurse collected individual patient details in labour and baby outcome at birth, as well as follow up at 6 weeks postpartum, with a telephone call to parents to inquire regarding the baby's health. Viral Culture After sampling, swabs were placed into viral transport medium which consisted of Medium 199 with Hank's balanced salt solution containing 1% inactivated fetal calf serum plus antibiotics (penicillin, streptomycin, and amphotericin B). Specimens were transported daily for inoculation in duplicate onto human foreskin fibroblast monolayers and processed by standard rnethods. 14 Monolayers were examined daily for development of cytopathic effects, when they were stained with HSV type-specific monoclonal antibodies (Syva, Palo Alto, CA) for confirmation and typing as previously described. 15 Viral Serology Antibodies to HSV (type specific) were detected by microneutralization and Western blot assay. 6

RESULTS
Demographic details of the patient population are outlined in Table 1. For the 8-week antepartum period during which the 198 patients were regularly assessed clinically and by culture, 906 total visits were made. For the total visits, 884 had completed data for both symptoms and signs, whereas for symptoms or for signs, 898 and 903, respectively ( Table 2). The median and mean number of visits were 4.0 (interquartile, 3-6) and 4.6 (standard deviation [SD], 2.6), respectively; the median interval between visits was 7.0 days (interquartile, 4-9), mean, 8.2 (SD, 9.9). Between the last antenatal visit to delivery, the median number .of .days was 4.0 (interquartile, 2-8), mean, 8.9 (SD, 27.8). Overall 88.7% of patients were positive by Western blot for HSV type-2 antibodies (with or without HSV type-1 antibodies); 62% were positive for HSV type-2 antibodies alone, and 8.7% were positive for HSV type-1 alone, a6 Of the 198 patients, 191 had at least one followup visit. Details of clinical lesions and of culture results for all the antepartum visits and of the ultimate visit prior to delivery are depicted in Table   2. It is evident that symptoms were reported in 20% and signs consistent with HSV evident in 28% of all visits, whilst 17% of all visits were positive by culture. Overall, 94 patients (48%) had HSV confirmed by culture (for 154 visits), and 97% of these isolates were HSV type 2. Patients were either symptomatic (as defined by prodrome, redness, blister, ulcer, or crust) or had signs consistent with HSV (macule, papule, vesicle, pustule, or ulcer) in 89% of these visits. Twenty-four women (12.6%) shed HSV asymptomatically. Six of these women shed on two occasions. Therefore, the asymptomatic shedding rate for the total antepartum visits  signs (mean, 12.6 days; range, 2-44), six (75%) were culture positive.
As depicted in Figure 1, the ability of an antepartum culture positive for HSV to predict a subsequent positive result (i.e., positive predictive value) was 59% (30/51) when the interval between visits was -<2 days, but only 19% (53/257) when days between visits were >2 (P < 0.0001). Clinical evidence of HSV was less predictive of a subsequent positive culture result. When the interval between visits was -<2 days, clinical prediction of a positive culture (i.e., positive predictive value) was 40% (27/67), whereas it was 17% (59/355) for >2 days (P < 0.0001).
At the final antenatal visit before labour, symptoms were reported in 13.6%, whilst signs of HSV were evident in 15.8%, and cultures were positive in 8.4% (Table 2). Intrapartum, a similar number of women reported symptoms (11.5%), and signs were evident on clinical examination in 9.5%. However, cultures taken by the patients' attending physician intrapartum were positive in only 1.5% (3/191) of cases (Table 3). All t.hree had previous signs and symptoms of HSV and were Western-blot-positive for HSV type 2 antibodies, and two were culture positive antepartum. In two of these cases, patients reported symptoms of herpes in labour, and active herpetic lesions were evident on exam. Both cases were delivered by cesarean delivery with a favourable outcome. The third case was asymptomatic at term and was delivered vaginally without adverse outcome. Of the 24 women who had asymptomatic excretion of HSV during pregnancy, three had signs or symptoms intrapartum, and none were culture positive.
The results of HSV culture positivity for the final six antepartum visits (as collected by the staff of the Herpes Clinic) together with the actual rate at delivery (as collected by the patients' own care provider) are shown in Table 3. As the intrapartum culture rate (1.5%) is at variance with the foregoing antepartum weekly rates (11.4-19.5%), an estimated intrapartum positivity was calculated based on these antepartum results. Two methods of estimation were performed. A crude estimate was made by taking an average percentage of the culture-positive rate of the previous six visits and was found to be 16.1%. The second estimate was a weighted average of the probability of culture positivity for each patient at delivery, based on the number of their previous visits. These probabilities were weighted by the number of visits for each patient, then the average was calculated. The anticipated delivery culture-positivity rate was calculated at 15.3% had the Herpes Clinic staff taken these specimens. Calculations were adjusted to eliminate any selection bias; i.e., assessment was only made for those patients initially negative by culture who became positive. Cases positive at the initial antepartum assessment were excluded. All analyses were carried out using a Statistical Analysis System. For antepartum visits that occurred within 7 days of delivery (130), viral shedding occurred in 13 (10%) ( Table 4). The asymptomatic shedding rate of these visits was 3%. Of those advised to have a cesarean delivery on the basis of active lesion or positive HSV cultures, 85% (11/13)   (P < 0.001) ( Table 4). There were no cases of neonatal herpes detected, and all baby and placental swabs taken at delivery were negative by viral culture.

DISCUSSION
In this patient population with histories of recurrent genital herpes, the observed culture-positivity rate for HSV for the 6-week period prior to delivery was consistently high, ranging from 11% to 19.5%.
An unexpected finding in this study was the low culture-positivity rate when nonspecialized personnel took genital swabs for culture. This unanticipated occurrence confounded the intrapartum cul- ture rate, which was at variance with the preceding antepartum shedding rate of herpes for the patient population. Only three patients were culture positive intrapartum and two of these in fact had clinical lesions and were delivered by cesarean delivery accordingly. The third patient shed HSV asymp, tomatically and delivered vaginally; her baby was culture negative and had a good outcome. Over the 8-week antepartum period, symptoms consistent with HSV occurred in 20% of visits, and these were culture positive in 55.7%. These findings are similar to those reported by Vontver, where 56% of recurrent clinical lesions were culture positive. 17 In our study group, 17% of anteparturn visits were culture positive and the asymptomatic shedding rate was 3.4% (external genital, 96.6% cervical, 3.4%). Overall, 12.6% of patients shed asymptomatically antepartum. This figure contrasts W'ith that of other similar studies in which asymptomatic excretion has been reported in antepartum patients at around 4% and at a shedding rate of 0.35-0.75%. 18 We believe this tenfold difference in shedding detection relates predominantly to expert personnel taking the clinical samples rather than others. This is further supported by the unexpectedly low intrapartum shedding rate (samples collected by delivery ward staff) found in our study, compared with that anticipated by antepartum rates (samples collected by personnel experienced in infectious disease). A further explanation for these findings may be the dilution effect on viral load by the flushing effect of amniotic fluid following rupture of the chorioamniotic membranes. The actual rate of shedding intraparturn of 1.5% for this study is similar to those previously reported (range, 0.35-2.4%). 7,18'9 No cases of neonatal herpes were seen in this study and no babies cultured positive at birth. The overall cesarean delivery rate was 31%, HSV being the indication in 56%. The recommendation to deliver by cesarean delivery, based on the last antepartum visit culture, correlated well with the actual delivery mode. Of those cultured within 7 days prior to delivery, 10% were positive (13/130), and 84.6% (11/13) who had been advised to have a cesarean delivery on these grounds complied.
Data from our study confirm the findings of Arvin et al. 7 that early preterm weekly cultures from patients are not helpful in predicting shedding at delivery. Those data were obtained by clinicians in the field, which, in light of our findings, may not have been a sensitive marker. Further, their finding of lack of predictability is not surprising and reflects inherent problems with HSV culture, which, at best, takes 48 hours to produce a result. Also, recurrent lesions shed for a short time (3-4 days); consequently, by the time a result is obtained from a laboratory, it poorly reflects the situation at that time. However, our study did show that antepartum exams were highly predictive of subsequent outcome where visits between prediction and outcome were 2 days but not predictive when the interval between visits was more than 2 days. Therefore, we believe that appropriately and carefully collected antepartum serial cultures may still have a role in selected patients, i.e., those suf-fering frequent recurrences near term. In this scenario, viral cultures every few days to demonstrate cessation of viral shedding could obviate the need for cesarean delivery.
The neonatal attack rate for women who deliver vaginally in the presence of recurrent genital HSV lesions is very low, on the order of <2%. z In contrast, women with primary genital HSV near or at delivery have a neonatal transmission rate 35-80%. 4'9'10'21 With such a very low rate of neonatal HSV from recurrent maternal infections, some advise covering genital lesions and allowing delivery vaginally as, the morbidity of cesarean delivery is greater, l,zz Yet, cervical shedding in association with recurrent external HSV lesions, although less frequent to primary infections (85%) is reported to occur in 12-16% of cases, z3 Moreover, as neonatal HSV carries the risk of significant morbidity and mortality, even with acyclovir treatment, we believe, as do others, z4 that cesarean delivery is the appropriate form of delivery when clinical maternal genital lesions are apparent. Asymptomatic shedding data from our study are consistent with the findings of Wald et al. in a nonpregnant population, zs They also demonstrated a much higher frequency of reactivation based on detection by polymerase chain reaction. Accordingly, wide inconsistency exists in any management strategy that requires recognition of symptoms. Some studies have adopted acyclovir suppression, now established to reduce both symptomatic and asymptomatic viral shedding rates outside of pregnancy, as an alternative strategy. Unfortunately, definitive answers regarding efficacy against both maternal shedding and neonatal infection are not yet available. Furthermore, while the safety of acyclovir in pregnancy is likely, it is unproven. In our view, there is little reason to doubt that acyclovir will predictably, but not incompletely, suppress viral shedding and symptomatic recurrent infection in pregnancy and, therefore, will likely reduce cesarean delivery rates for women with high recurrence frequency rates. This would appear to offer a strong cost benefit and would potentially obviate surgical morbidity while we continue to pursue neonatal infection risk factors. However, neonatal infection has occurred in the setting of inadequate maternal acyclovir suppression therapy. Clearly, properly conducted, placebo-controlled, randomised trials stratified accord-ing to type-specific immune status and frequency rates are warranted before new recommendations are made.
Finally, it is essential to recognise that current medical practice does not directly address the group at highest risk for neonatal transmission: HSV-susceptible women in the third trimester who are in HSV discordant relationships. 9 Only routine serologic screening of both partners can begin to address prevention in this highest-risk patient.