Suppression of Subclinical Shedding of Herpes Simplex Virus Type 2 with Acyclovir

Methods

Study Participants and Design

Because natural history studies of subclinical HSV shedding have indicated that recent acquisition of genital herpes and infection with HSV-2 alone are risk factors for subclinical shedding [14, 15], only women with these clinical characteristics were considered for enrollment in our study. Other entry criteria included general good health without clinical or laboratory evidence of immunosuppression and without the use of immunosuppressive medication. Potential study participants were recruited by newspaper advertisements and through information distributed to local health care providers. We screened 103 women with the above-mentioned clinical characteristics and enrolled 34 (33%). Among the 69 women who were not enrolled, 29 (42%) were not eligible because serologic testing indicated that they had infection other than that with HSV-2 alone [16]. Of the 40 eligible women, 19 (28%) decided not to participate in the study because of the time that it required. Women were also not enrolled if they were planning to move from the area during the study period (6%) or were planning pregnancy (3%).

The study was a double-blind, placebo-controlled, crossover trial that was done at the University of Washington Virology Research Clinic and approved by the Human Subjects Review Committee. Participants gave written informed consent. The study was designed as two 70-day treatments with an intervening 14-day washout period. The randomization list was generated by a computer. Participants were randomly assigned to receive either acyclovir, 400 mg twice daily, or placebo for the first 70 days, followed by a 14-day washout period during which all patients received placebo capsules. Participants then received the alternate regimen for 70 days. The randomization scheme was stratified by the duration of genital herpes: One stratum consisted of women who had had herpes for less than 6 months, and the other consisted of women who had had herpes for more than 6 months but for less than 2 years. The reason for the stratification was the expected difference in shedding rates between these two groups [14, 15]. Both clinicians and participants remained blinded to the results of the cultures until the conclusion of the study. The virology laboratory technologists were blinded to the clinical status of the participants. Study medication was dispensed in 2-week cycles (80 capsules per cycle). At enrollment, a medical history was taken, and a physical examination and a pregnancy test were done.

Collection of Genital Swabs for Herpes Simplex Virus

The clinical staff instructed women in the collection of cervicovaginal, vulvar, and anal samples for viral cultures [17]. A separate Dacron swab was used to sample each site daily. For vulvar cultures, the swab was passed over the entire labia majora and minora. Cervicovaginal swabs were obtained by touching the exocervix and posterior vaginal fornix. Perianal samples were obtained by inserting the swab into the opening of the anal canal and rotating it. Swabs were placed into individual vials of viral transport medium and refrigerated until transport. Participants were instructed to collect the samples on awakening, if possible. Samples were either delivered to the laboratory by the participants themselves or were collected by courier at least three times per week. Previous studies have shown that specimens obtained by patients in this manner are as sensitive for detecting subclinical HSV shedding as those obtained by clinicians [15]. Each woman received a diary card for recording genital lesions, genital symptoms such as localized pain and irritation, and the number of pills taken daily. The use of open-label acyclovir or other topical or systemic antiviral agents was not permitted during the study. Participants were instructed to visit the clinic within 24 hours of the onset of recurrent genital lesions for genital examination and confirmation of an HSV recurrence. Patients also visited the clinic at 2-week intervals, at which time the diary cards were reviewed and collected, the study drug was refilled, and compliance was monitored.

Laboratory Studies

Each collected sample was inoculated daily into triplicate wells of diploid fibroblasts in 48-well microtiter plates [18]. We examined wells three times during the first week and twice during the second week. All wells that exhibited cytopathic effect were confirmed and typed using HSV-specific monoclonal antibodies as previously described [19]. At the completion of the trial, acyclovir sensitivity testing was done on selected isolates using previously reported methods [20-22].

Definition of Subclinical and Symptomatic Herpes Simplex Virus Shedding

We evaluated the effect of acyclovir therapy on the frequency and duration of subclinical shedding of HSV, the primary end point of our study, and on the frequency and duration of symptomatic shedding. We defined subclinical shedding as isolation of HSV in the absence of genital or perianal lesions as noted by patient or clinician. The subclinical shedding rate was calculated as the number of days without genital lesions but with positive viral cultures divided by the total number of days without lesions. The rate of symptomatic HSV shedding was defined as the number of days with reported or observed genital lesions and with positive viral cultures, divided by the total number of days with genital lesions.

Statistical Analyses

All data collection, entry, analyses, and interpretation were done at the University of Washington, independent of the company sponsor. We evaluated the treatment effect in two ways. The first was an intent-to-treat analysis that included all randomized participants for the initial treatment period and that used Mann-Whitney two-sample test and chi-square test to compare shedding rates between the women who received placebo and those who received acyclovir therapy. The time to subclinical shedding and the risk factors for subclinical shedding were estimated using Kaplan-Meier survival curves and the Cox proportional-hazards model.

A second analysis was done on the data from the 26 women who successfully complied with both sequences of the trial (successful compliance was defined as collection of cultures on at least 80% of the study days). This analysis included a paired analysis of treatment, period, and sequence effects using the Wilcoxon signed-rank test and the McNemar test [23]. Thus, we were able to assess the influence of the sequence of treatment (acyclovir followed by placebo compared with placebo followed by acyclovir) on the results of the study. The relation between the days on which women observed lesions while receiving acyclovir therapy and while receiving placebo was examined with the Spearman correlation coefficient.

Results

Thirty-four women were enrolled in the study between July 1992 and August 1993. Eighteen of these women had had genital herpes for less than 6 months (median time since acquisition, 3.5 months), and 16 had had genital herpes for more than 6 months but for less than 2 years (median time since acquisition, 18 months) (Table 1). The median age of the participants was 26 years, and 29 were white. Nine (50%) of the 18 women who had had genital herpes for less than 6 months and 8 (50%) of the 16 who had had genital herpes for more than 6 months were assigned to receive acyclovir first and placebo second; the 17 remaining women received the study drugs in the reverse order.

Thirty-three of the 34 women successfully completed the first arm of the study, and 26 women completed both the first and the second arm. The median number of days on study was 154 (range, 54 to 177 days), and the median number of days on which cultures were obtained was 146 (range, 53 to 158 days). The median number of days on which cultures were obtained was 65 for the acyclovir arm, 66 for the placebo arm, and 13.5 for the washout period. The median percentage of the prescribed drug dosage taken was 98% during the first arm of the study, 99% during the second arm of the study, and 96% during the washout period.

Of the eight women who failed to complete both portions of the study successfully, four were assigned to receive placebo initially and four were assigned to receive acyclovir initially. Six women withdrew from the study for administrative reasons (such as moving from the area), and two women completed the study but obtained cultures on only 47% and 65% of the study days, respectively. No patient withdrew from the study because of side effects or toxicity caused by the study drugs, nor did any participant drop out from either of the study arms to receive antiviral therapy.

Intent-to-Treat Analysis

Twenty-six of the 34 women had either a culture-confirmed clinical or subclinical episode of genital herpes during the study. Of the 17 women who were assigned to receive placebo initially, 15 had at least one episode of subclinical shedding compared with 3 of the 17 women who received acyclovir first (P < 0.001). Subclinical shedding of HSV-2 was detected on 64 of 928 (6.9%) days in patients who were assigned to receive placebo initially compared with 3 of 1057 (0.3%) days in patients who were assigned to receive acyclovir initially (P < 0.001). The time to first day of subclinical shedding in both groups is shown in Figure 1. The relative risk for subclinical shedding in women who received acyclovir initially compared with those who received placebo initially was 0.09 (95% CI, 0.03 to 0.35). The relative risk for subclinical shedding was higher among women who had had HSV-2 infection for less than 6 months than among those who had had HSV-2 infection for more than 6 months (relative risk, 2.6; CI, 1.0 to 7.2). However, the benefits of acyclovir were similar in both groups of women; for example, the frequency of subclinical shedding was 0.35% and 0.21% during acyclovir therapy and 8.5% and 5.3% during placebo administration for women who had had herpes for less than 6 months and for more than 6 months, respectively.

Comparison among Women Who Completed Both Arms of the Study

Subclinical Shedding

Twenty-six women, 13 of whom initially received acyclovir and 13 of whom initially received placebo, obtained cultures on more than 80% of the total study days. Nineteen of these 26 (73%) women shed subclinically while receiving placebo compared with 5 of 26 (19%) during acyclovir therapy (P < 0.001). Subclinical shedding of HSV-2 occurred on 83 of 1439 (5.8%) days with placebo compared with 6 of 1611 (0.37%) days with acyclovir (P < 0.001) (Table 2). The rates of positive cultures were reduced by 95%, 89%, and 97% in samples obtained from the cervicovaginal, vulvar, and rectal areas, respectively. Of the 89 days of subclinical shedding, 9 (7 days for the placebo arm and 2 days for the acyclovir arm) were days during which the participant was seen in the clinic and the absence of genital lesions was confirmed. The median number of days of subclinical shedding was 2 (range, 0 to 15 days) for the placebo arm compared with 0 (range, 0 to 2 days) for the acyclovir arm.

We reviewed the diary records for the 38 days on which HSV was isolated from vulvar skin swabs and no lesions were noted. Of these 38 days of vulvar subclinical shedding, 31 (82%) were days on which participants reported no genital symptoms. On 2 days (5%), subclinical vulvar shedding was associated with prodromal symptoms and the development of genital lesions within 24 hours. On 5 days (13%), subclinical shedding was associated with localized itching and soreness.

Seven positive subclinical cultures were obtained during acyclovir therapy. Four women shed subclinically for a single day during acyclovir therapy: on day 1, from the vulva (1 patient); on day 5, from the rectum (1 patient); on day 31, from the vulva (1 patient); and on day 45, from the vulva (1 patient). One woman had a 2-day episode of subclinical shedding in which HSV-2 was isolated on day 23 from the vulva and on day 24 from both the vulva and the cervix. Review of compliance records showed that 4 women took all of their medication during the 24 hours preceding shedding episodes; the fifth woman who had a single day of genital shedding had taken only 61% of the prescribed pills during the 2 weeks preceding the shedding episode.

Symptomatic Genital Herpes

Genital lesions occurred in 21 (81%) of the 26 women who completed both arms of the study. Subclinical shedding also occurred in 3 of the 5 women who did not have symptomatic genital recurrences. Sixty-seven episodes of symptomatic genital herpes were reported, of which 58 (87%) were confirmed in the clinic. Twenty (77%) women had recurrences while receiving placebo compared with 12 (46%) women during acyclovir therapy (P = 0.01); 43 episodes of symptomatic genital herpes occurred during the placebo portion of the study and 17 during acyclovir therapy (P = 0.002). Genital lesions were reported on 330 of 1833 days (18%) during the placebo portion of the study, and on 112 of 1839 (6.1%) days during acyclovir therapy (P = 0.001). During the placebo portion of the trial, the median number of days per month that participants had lesions was 4.2 (range, 0 to 17.5 days) compared with 0 (range, 0 to 8.3 days) during the acyclovir portion (Figure 2).

The rate of HSV isolation from genital lesions during episodes of symptomatic shedding was also reduced with the use of acyclovir. Thirty-one of the 67 episodes of symptomatic genital lesions were associated with a positive HSV culture: These included 23 of 43 episodes occurring during the placebo portion of the trial, 2 of 17 episodes occurring during the acyclovir portion, and 6 of 7 episodes occurring during the washout period. Overall, HSV-2 was isolated from the genital region on 90 (29%) of 315 days on which genital lesions were noted during the placebo portion of the study compared with 2 of 104 days (1.9%) during the acyclovir portion of the study (P = 0.008).

Five women had genital lesions on more than 10 days during acyclovir therapy. A positive HSV-2 culture was obtained for only 2 of these 11 episodes of genital lesions. Women who had frequent genital lesions during the placebo arm of the study tended to have genital lesions more frequently during the acyclovir period (P = 0.03).

Overall Effect on HSV-2 Shedding

Overall, HSV-2 was isolated from at least one anatomical area on 173 of 1754 (9.9%) days during placebo administration, on 23 of 350 (6.6%) days during the washout period, and on 8 of 1715 (0.5%) days during acyclovir therapy (P < 0.001) (Figure 3). Ninety-seven of the 204 (48%) culture-positive days were days during which patients had subclinical viral shedding. Between the two treatment periods, 6 women had reductions in the HSV shedding rate ranging from 62% to 95%, and 14 women had complete cessation of shedding during acyclovir therapy.

Period and Sequence Effects

To determine whether the differences in HSV shedding rates during the placebo and acyclovir administration reflected the treatment effect only or were related to the study design, we examined the data for period and sequence effects [23]. To test for period effect, we compared the subclinical shedding rate in women who received placebo during the first phase of the study with the shedding rate of the women who received placebo during the second phase. We made similar comparisons for the acyclovir arm of the study. The differences were not statistically significant (P = 0.74 and P = 0.55 for shedding rates during acyclovir and placebo administration, respectively), and we concluded that the period effect did not significantly alter the study results. To test for sequence effect, we considered whether the difference in the subclinical shedding rate during acyclovir and placebo administration depended on which study drug was received first. The results were not statistically significant (P = 0.66). Analysis of period and sequence effects were also done for the symptomatic shedding rate and the occurrence of lesions; the results were not significant.

Acyclovir Sensitivity Testing

We did acyclovir sensitivity testing on all initial viral isolates from each participant and on all isolates that were obtained during acyclovir therapy. Twenty-five women had a positive viral culture during the study: Twenty-four of the 25 HSV-2 isolates had an IC₅₀ (in vitro inhibitory concentration of acyclovir that protected 50% of cells from cytopathic effect) of less than 3.0 µg/mL. One woman had an isolate with IC₅₀ of 12.4 µg/mL. This isolate was recovered on day 12 of the study when the patient was receiving placebo. She had 2 subsequent days of HSV shedding while receiving placebo and 1 day of shedding while receiving acyclovir. The 3 HSV isolates in these latter instances showed that they were sensitive to acyclovir (IC₅₀ range, 0.64 to 0.80 µg/mL). None of the strains that were isolated during acyclovir therapy were resistant to acyclovir (IC₅₀ range, 0.37 to 1.55 µg/mL).

Discussion

Our trial illustrates several novel aspects about genital herpes and the role acyclovir may play in its therapy. The major therapeutic finding of the trial is that the standard suppressive dose of acyclovir (400 mg twice daily) decreased the frequency of subclinical shedding of HSV-2 by at least 90% at all sites. Because subclinical shedding occurred on 48% of the total days on which HSV-2 was isolated from the genital tract, this finding markedly extends the potential therapeutic benefit of acyclovir for persons with frequently recurring HSV-2.

Our study design, which required both daily sampling for virus and evaluation for genital lesions, also revealed several important clinical features of genital herpes. Overall, half of the culture-positive episodes were not associated with discernible lesions, even among a group of persons carefully taught to recognize them. The differences in the frequency of genital lesions in the two arms of the study and the similarity of the data during the washout period and placebo administration appear to confirm the ability of our study participants to recognize herpes-related genital lesions. Moreover, during nine episodes of subclinical shedding in which the patient had a genital examination in the clinic, the research clinicians observed no genital lesions. We believe that the high frequency of subclinical shedding that we observed in our study accurately reflects the true spectrum of HSV-2 infection in these women. In fact, we would expect to see higher subclinical shedding rates in persons less attuned to the signs and symptoms of genital herpes. In previous studies, we have also shown that the methods our study participants use for obtaining HSV cultures are as sensitive for detecting HSV as the technique used in the medical care setting by trained clinicians [15]. Therefore, we believe that the HSV isolation data that we observed are accurate.

We selected women who were expected to have a high rate of subclinical shedding because of recent acquisition of genital herpes and infection with HSV-2 only [14, 15]. Our selection criteria were intended to reduce the number of enrollees, which would allow us to monitor compliance and conduct a more cost-effective protocol (albeit we did more than 10 500 separate viral isolations in these 34 women). In addition, our original study design was predicated on detecting less dramatic differences (80%) between the acyclovir and placebo arms. The subclinical HSV shedding rate of 6.9% during the placebo phase of the study is the highest rate we have noted in a population of immunocompetent persons with HSV-2 infection. This rate of subclinical shedding is more than three times higher than the mean rate observed in a larger cohort of women that we studied using daily sampling techniques [15] and confirms that recent acquisition of HSV-2 infection is associated with high subclinical shedding rates [14].

The clinical and demographic manifestations of genital herpes in this cohort of women were typical of patients we have studied over the last decade [5, 24-26], as was the frequency of recurrences before enrollment and during the study. Moreover, the reduced frequency of shedding that we saw during acyclovir therapy was similar in women who shed frequently while receiving placebo (> 5% of days) and in those who shed less frequently (1% to 5% of days). This finding suggests that acyclovir is likely to have similar efficacy in women with a longer duration of HSV-2 and lower baseline shedding rates.

One of the interesting aspects of our trial is that oral acyclovir therapy had a greater effect on reducing shedding than on reducing genital lesions. Twelve of the 26 women who completed both arms of the trial continued to have genital lesions during therapy, but all except two of these episodes were culture-negative. The cause of these lesions is unclear; culture-negative lesions have been reported in other studies of recurrent genital herpes, both with and without acyclovir therapy [27, 28]. Our rates of culture positivity during episodes of genital lesions were comparable to those found in other studies [8]. These episodes may result from viral reactivation and local tissue destruction that is not accompanied by a detectable virus, even with daily sampling.

Although the effect that acyclovir therapy had on subclinical shedding was surprisingly great, it was not absolute. Four women experienced five episodes of “breakthrough” shedding, despite complete adherence to the medication schedule. Interestingly, the viral isolates from these episodes were sensitive to acyclovir in vitro, suggesting that the mechanism of this breakthrough shedding is not the emergence of resistant virus.

Acyclovir has been studied as suppressive daily therapy for genital herpes for more than 5 years and has not been associated with serious side effects [6, 7, 29]. Of interest, the one isolate in our study that had high-level in vitro resistance occurred on placebo and was followed by several sensitive isolates. These data suggest that many isolates with reduced sensitivity to acyclovir represent naturally occurring thymidine kinase-deficient variants.

That acyclovir does not have an absolute effect on shedding, especially early in therapy, may explain the findings of previous studies [11-13]. In our study, we used a greater sampling frequency, sampled a wider range of anatomic sites, monitored compliance more closely, and recruited a cohort of women that were intensively trained to identify clinical episodes of genital herpes. It should be noted that we observed subclinical HSV shedding in 5 of 26 (19%) women, which is not statistically different from the 7 of 48 (15%) women observed by Bowman and colleagues [12] or the 1 of 5 (20%) women observed by Haddad and colleagues [13]. In their study of acyclovir and HSV shedding, Straus and colleagues [11] also used daily home cultures for assessment of shedding rates and found little difference in the rate of asymptomatic shedding between women who received placebo (1.1%) and in those who received daily acyclovir (0.72%). In this trial, the dose of acyclovir was 600 mg daily and asymptomatic shedding was defined as the isolation of HSV on days without genital lesions or symptoms, whereas we defined subclinical shedding as the isolation of HSV on days without lesions only. One of our recent studies indicated that the specificity of reported genital symptoms such as itching and pain for defining genital herpes was poor [29]. This finding led to our definition of subclinical as days without lesions only. The difference in the definitions also explains the marked difference in subclinical shedding rates during placebo administration between the two studies. These differences in study design and in the acyclovir dose make direct comparisons between studies difficult.

Our study raises the question of whether suppressive acyclovir can be used to reduce the transmission of genital herpes. Several studies have identified subclinical shedding as the major mode of transmission of genital herpes to sexual partners and infants [1-4, 30-32]. In addition, transmission of infection to others is a major concern to patients with genital herpes [33]. Our data support the concept that the short-term use of acyclovir to reduce transmission is worth further study. However, we caution against assuming, on the basis of our data, that acyclovir will be effective in such situations. At present, we counsel all persons with HSV-2 infection, whether they are receiving acyclovir therapy or not, to use condoms at all times and abstain from sexual intercourse during recurrences. We know that HSV can be transmitted to neonates during culture-negative episodes, and HSV DNA has been detected by polymerase chain reaction during these episodes [34]. Preliminary observations by us indicate that HSV DNA may be detected even in patients who are receiving acyclovir therapy. Thus, low levels of viral replication may occur despite suppressive acyclovir and may be sufficient to result in HSV transmission. Therefore, although our findings are provocative and extend the therapeutic utility of acyclovir, current counseling guidelines for reducing the transmission of HSV should not change until data are available to evaluate the effect of acyclovir on subclinical shedding in men and on transmission [35].

Dr. Zeh: University of Washington, Statistics, Box 354322, Seattle, WA 98195-4322.

Ms. Barnum: Virology Research Clinic, 1001 Broadway, Suite 320, Seattle, WA 98122.

Dr. Davis: Glaxo Wellcome Company, 3030 Cornwallis Road, Research Triangle Park, NC 27709.

Dr. Corey: University of Washington, Virology Division, Room 9301, 1200 12th Avenue South, Seattle, WA 98144.