Genital Herpes Simplex Virus Type 2 Suppression With Valacyclovir Is Not Associated With Changes in Nugent Score or Absolute Abundance of Key Vaginal Bacteria

Abstract

Background

In women, genital herpes simplex virus type 2 (HSV-2) infection is associated with increased risk for recurrent bacterial vaginosis (BV), but causal relationships are unclear.

Methods

Women with a self-reported history of BV and HSV-2 seropositivity self-collected vaginal and anogenital swabs for 2 nonconsecutive 28-day periods, in the absence or presence of valacyclovir suppressive therapy (500 mg daily). HSV polymerase chain reaction was performed on anogenital swabs; vaginal swabs were used for assessment of BV by Nugent score and quantification of vaginal microbiota. Days with BV, defined by Nugent score ≥7, were compared during the observational period and valacyclovir treatment.

Results

Forty-one women collected swabs for a median of 28 days (range, 20–32 days) each study period. The HSV-2 shedding rate decreased from 109 of 1126 days (9.7%) presuppression to 6 of 1125 days (0.05%) during valacyclovir (rate ratio [RR], 0.06 [95% confidence interval {CI}, .02–.13]). BV occurred on 343 of 1103 days (31.1%) during observation and 302 of 1091 days (27.7%) during valacyclovir (RR, 0.90 [95% CI, .68–1.20]). The median per-person Nugent score was 3.8 during observation and 4.0 during valacyclovir. Average log₁₀ concentrations of vaginal bacterial species did not change significantly during valacyclovir treatment.

Conclusions

Short-term HSV-2 suppression with valacyclovir did not significantly affect the Nugent score or the vaginal microbiome despite potent suppression of HSV-2 shedding.

Short-term herpes simplex virus type 2 (HSV-2) suppression with valacyclovir did not significantly affect the Nugent score or the vaginal microbiome despite potent suppression of HSV-2 shedding in women with a self-reported history of bacterial vaginosis and HSV-2 seropositivity.

Reproductive medical conditions in women result in significant morbidity and mortality. Bacterial vaginosis (BV) and herpes simplex virus type 2 (HSV-2) are 2 highly prevalent infections affecting the female genital tract. BV, a polymicrobial dysbiosis, is the most common cause of vaginal discharge in women of childbearing potential, with an estimated prevalence of 21.2 million (29.2%) in women aged 14–49 years in the United States [1]. Molecular techniques have identified several novel bacterial species as being associated with BV, some of which have not been grown in culture, including BV-associated bacterium 1 (BVAB-1) and BV-associated bacterium 2 (BVAB-2) [2]. A high abundance of lactobacilli is considered optimal, while higher abundances of diverse anaerobes, including Gardnerella species, are associated with BV [3, 4]. BV-associated bacterium (BVAB-2) and Megasphaera are highly associated with BV, with a sensitivity of 96% and specificity of 94% for predicting BV [5], and have been included as bacterial targets in commercially available diagnostic tests [6]. HSV-2, the leading cause of genital ulcer disease, causes lifelong infection with a prevalence of >500 million infections worldwide and an estimated 11.9% prevalence rate in persons aged 14–49 years in the United States [7]. Genital HSV-2 infection is characterized by periodic genital shedding and lesions [8] and is associated with inflammation in the female genital tract [9, 10].

“Epidemiologic synergy” is a term used to describe interrelationships between human immunodeficiency virus (HIV) and sexually transmitted infections (STIs), where both pathogens increase the risk of acquisition of the other infection. For example, BV and HSV-2 are known risk factors for HIV, conferring a 1.6- to 3-fold increase of HIV acquisition, respectively [11, 12]. Furthermore, several studies have identified an association between BV and HSV-2. For instance, in a cohort of 293 HSV-2–seronegative women, prevalent or incident BV increased the risk of incident HSV-2 infection by 2.4-fold [13]. In another large cohort study of 670 HSV-2–seronegative women, BV was associated with a 2.1-fold increased risk for HSV-2 acquisition [14]. In other studies, BV and HSV-2 were often diagnosed simultaneously, and temporality was difficult to determine [15, 16]. Conversely, HSV-2 infection may increase the risk of BV. For instance, in a systematic review and meta-analysis, HSV-2–seropositive women had a 55% higher risk of BV [17]. The mechanistic associations between BV and HSV-2 have not been elucidated, and 1 hypothesis is that among HSV-2–seropositive women, inflammation associated with HSV-2 genital reactivation mediates the increased risk of BV through increased colonization with BV-associated bacteria.

To evaluate the relationship between BV and HSV-2 reactivation, we conducted a 1-way crossover study in women with a history of BV and HSV-2, evaluating the effect of daily valacyclovir for HSV-2 suppression on BV prevalence and on the vaginal microbiota, including measurement of vaginal concentrations of Lactobacillus crispatus, Lactobacillus iners, Gardnerella, Megasphaera, and BVAB-2. We hypothesized that antiviral suppression of HSV-2 would result in decreased prevalence of BV as assessed by a decrease in median Nugent score, the gold standard for BV diagnosis [18], and increased concentrations of lactobacilli that are associated with an optimal microbiota.

METHODS

Study Participants

Women with a history of BV within the past year, diagnosed by Amsel criteria or Nugent score (≥7), and HSV-2 infection were recruited into the study at the University of Washington Virology Research Clinic in Seattle, Washington, between 2014 and 2018. Eligible persons were >18 years of age, HSV-2 seropositive, and HIV seronegative and had a self-reported history of BV. Participants with gonorrhea, chlamydia, trichomoniasis, BV, or Candida at screening were treated and eligible for participation 1 week after completing treatment.

Patient Consent Statement

All participants provided signed written informed consent and the study protocol was approved by the University of Washington Human Subjects Division.

Study Design

The study was an open-label, 1-way crossover study in which participants did not receive any antiviral medication during the first 28-day period (observational period), initiated valacyclovir 500 mg daily for a 2-week lead-in period, and then continued valacyclovir 500 mg daily for an additional 28 days (valacyclovir suppression period) (Supplementary Figure 1). During the observational and valacyclovir suppression periods, participants obtained 2 self-collected vaginal swabs and 1 self-collected anogenital swab daily for 28 days, obtained cervicovaginal secretions using a menstrual cup (SoftCup) for 5–10 minutes, and completed a daily symptom diary [19]. Vaginal swabs were stored dry at –20°C in each participant’s freezer and transported to the clinic in ice pack containers provided by the study. Vaginal swabs were then stored at −80°C until processing.

Women with an intrauterine device (IUD) did not collect menstrual cup samples due to concern about removing the IUD with collection. Symptoms elicited in the diary included symptoms of BV, such as vaginal odor; vaginal discharge; abdominal pain; and symptoms of genital herpes recurrences, such as itching, burning, pain or soreness, or sores, blisters, ulcers, or crusts in the genital, buttock, vulvar, or perianal areas. Participants returned to clinic every 2 weeks during the study period for a cervicovaginal examination and to return samples and collect study medication. To measure medication adherence, pill counts were performed during the valacyclovir phase at study visits.

Laboratory Studies

At the screening visit, the University of Washington HSV Western blot was performed to detect antibodies to herpes simplex virus type 1 (HSV-1) and HSV-2 [20], and urine or cervical samples were obtained for Gonorrhoeae, Chlamydia, and Trichomonas nucleic acid amplification testing (Aptima). Wet mount and potassium hydroxide preparations of vaginal fluid were used to diagnose Candida vulvovaginitis. Gram stain was prepared from 1 of the daily vaginal swabs and microscopy was performed with calculation of the Nugent score by a single experienced laboratory technician who was blinded to treatment period. Nugent scores of ≥7 were defined as diagnostic of BV [18]. The other vaginal swab was stored at −80°C for DNA extraction. Similar approaches for longitudinal sampling, storage of vaginal swabs, BV diagnosis by Nugent score, and characterization of the microbiota using molecular approaches have been used by others [21–23].

Bacterium-specific quantitative polymerase chain reaction (qPCR) assays were performed on DNA extracted from the second vaginal swabs collected daily with primers to amplify the following bacteria: L crispatus [24], Lactobacillus jensenii [25], L iners [25], BVAB-2 [24], Megasphaera species [26], and Gardnerella species [24], as described in prior studies [24, 25, 27]. Bacterial targets for measurement of concentrations were selected for the following reasons: L crispatus and L jensenii have been consistently associated with health [4, 28, 29] and are considered optimal. Lactobacillus iners is typically present in women with and without BV [4, 30] and is not considered optimal. BVAB-2 and Megasphaera species are useful for BV diagnosis [5, 6], and presence of high abundance or concentrations of Gardnerella species is consistently associated with BV [6, 31]. PCR was performed on DNA extracted from anogenital swabs for the detection of HSV DNA [32].

Study Analysis

The primary outcome for this study compared the per-participant median Nugent score during observational and valacyclovir suppression periods. With a planned 36 participants, the study had 80% power to detect a 2.5-point change in the median Nugent score during therapy with valacyclovir, assuming a 0.2 within-person correlation of Nugent score and accounting for 10% loss to follow-up. Secondary analyses examined the presence and quantity of vaginal organisms associated with BV and optimal microbiota. Mixed modeling analyses were performed to compare within-participant Nugent scores during the observational period and during valacyclovir therapy. A 2-sided P < .05 was considered statistically significant. A post hoc sensitivity analysis was performed on only participants who had HSV-2 detected during the observational period to evaluate change in Nugent score and microbiota between the observational and treatment with valacyclovir phases. An additional post hoc analysis was performed including only the last 2 weeks of valacyclovir therapy for all participants to further evaluate the impact of longer valacyclovir exposure on BV.

RESULTS

Study Population

Sixty-five women who were HSV-2 seropositive were screened for the observational phase of the study. Fifty-four women completed the observation phase. Four participants withdrew from the study and 3 declined to participate in the valacyclovir suppression phase. Forty-seven women initiated valacyclovir. One participant withdrew, 2 experienced valacyclovir side effects, and 3 were lost to follow-up. A total of 41 participants completed both the observational and valacyclovir crossover phases of the study and were included in the data analysis (Supplementary Figure 2).

Of the 41 participants, the mean age was 38 years (range, 25–54 years). Twenty-one (51%) participants identified as White, 14 (34%) as Black, 1 (2%) as Asian, 5 (12%) as mixed race, and 1 (3%) as Hispanic/Latinx (Table 1). All participants were seropositive for HSV-2; 21 (51%) of participants were also HSV-1 seropositive. The women collected swabs on a median of 28 days (interquartile range [IQR], 23–30 days) during the observational period and 28 days (IQR, 20–32 days) during valacyclovir therapy. Of the 41 participants, 9 of 41 (22%) reported oral hormonal contraceptive therapy and 4 of 41 (10%) hormonal IUD placement.

Table 1.

Demographic and Clinical Characteristics of the Study Population (N = 41)

Characteristic	No. (%)
Female sex	41 (100)
Age, y, median (range)	38 (25–54)
HSV-1 and HSV-2 seropositive	21 (51)
Race/ethnicity
White	21 (51)
Black	14 (34)
Asian	1 (2)
Mixed race	5 (12)
Hispanic	1a (3)
History of genital herpes	33 (80)
Acquired HSV-2 ≥1 y ago	29/33 (88)
No. of HSV-2 recurrences past year, median (range)	2 (0–12)
No. of BV recurrences past year, median (range)	1 (1–12)
Hormonal contraceptive use at baseline	13 (32)

Abbreviations: BV, bacterial vaginosis; HSV-1, herpes simplex virus type 1; HSV-2, herpes simplex virus type 2.

^aOne participant did not self-report ethnicity.

HSV-2 Shedding

At least 1 day of HSV-2 shedding occurred in 18 (43.9%) participants during the observational period and 6 (14.6%) participants during valacyclovir suppression. During the observational period, HSV-2 shedding occurred on 109 of 1126 days (9.7%). During valacyclovir suppression, HSV-2 was detected on 6 of 1125 days (0.05%), a 94% reduction in viral shedding (rate ratio [RR], 0.06 [95% confidence interval {CI}, .02–.13]; P < .0001; Figure 1A). Valacyclovir adherence was high, with 36 of 41 (87.8%) participants taking >80% of the daily valacyclovir doses and 31 of 41 (75.6%) participants taking ≥90% of daily valacyclovir therapy.

Figure 1.

A, Herpes simplex virus type 2 (HSV-2) detection rate by quantitative polymerase chain reaction by study day during the observational period and valacyclovir suppression. The number of days with swabs collected is noted above each bar. B, Bacterial vaginosis positivity (Nugent score >7) rate during the observational period and valacyclovir suppression phase.

BV Status and Nugent Scores

During the observational period, 25 participants (60.1%) had a Nugent score ≥7 on at least 1 day; similarly, 25 (60.1%) participants had Nugent score ≥7 during valacyclovir suppression. BV was present on 343 of 1103 days (31.3%) during the observational period and 302 of 1091 days (27.7%) during valacyclovir suppression. The BV positivity RR during valacyclovir suppression compared to the observational period was 0.90 (95% CI, .68–1.20) (P = .47; Figure 1B).

A linear mixed model evaluated the average per-person change in median Nugent score during the valacyclovir period. The median per-person Nugent score was 3.8 during the observational period and 4.0 during valacyclovir suppression for a difference of 0.26 (95% CI, −.43 to .94) (P = .47; Figure 2). Variability was observed in individual participants for days of BV positivity during the observational period and during valacyclovir suppression (Figure 3Aand 3B). Figure 3Aand 3B suggests a trend toward a decrease in proportion of days with BV and median Nugent score during the valacyclovir treatment period. To explore this association, further analyses were conducted. In a sensitivity analysis only including the 18 persons with detected HSV-2 shedding at baseline, BV positivity was 26.2% (130/496) at baseline and 29.6% (142/480) while on valacyclovir. The BV positivity RR was 1.16 (95% CI, .80–1.69; P = .51). Additionally, an analysis was performed to evaluate whether the impact of valacyclovir treatment on BV occurred later in the treatment period; BV positivity was 31.1% (343/1103) at baseline and 25.8% (144/557) while on valacyclovir. The BV positivity RR was 0.83 (95% CI, .59–1.15; P = .26).

Figure 2.

Frequency of bacterial vaginosis (by Nugent scores) by number of participants and days during the observational period and valacyclovir suppression phase, with Nugent scores grouped as optimal (0–3), intermediate (4–6), or high (7–10).

Figure 3.

A, Proportion of days with bacterial vaginosis (BV) per participant during baseline and treatment phase with valacyclovir. Data for each participant are indicated by a single row, with proportion of days with BV during the observational period (orange or left of 0%) and during valacyclovir suppression (blue or right of 0%). Participants 33–41 did not have any days with BV during either the observational or valacyclovir suppression periods. B, Median Nugent score over all days for participants during the baseline and treatment phase with valacyclovir. Data for each participant are indicated by a single row, with median Nugent score during the observational period (orange or left of 0) and during valacyclovir suppression (blue or right of 0).

Frequency of Genital Symptoms and Lesions

Twenty-five (61.0%) participants during the observational period and 24 (58.5%) participants during valacyclovir suppression reported at least 1 genital symptom. HSV-2 shedding occurred on at least 1 day with reported symptoms in 8 (19.5%) participants during the observational period and 3 (7.3%) participants while on valacyclovir. Overall, participants reported symptoms on 21.0% of days during the observational phase and on 14.2% of days during valacyclovir treatment (RR, 0.67 [95% CI, .42–1.09]; P = .11).

Of the days available for symptom reporting, vaginal discharge was noted on 63 of 643 days (9.8%) with BV (≥7 Nugent score) and 35 of 552 days (6.3%) of Nugent score = 0.

Microbiome

Average log₁₀ concentrations of bacterial DNA (16S ribosomal RNA [rRNA] gene copies/swab) for L crispatus, L jensenii, L iners, BVAB-2, Megasphaera species, and Gardnerella species were not significantly changed comparing the observational period and during valacyclovir suppression (Table 2). Gardnerella species were present at an average log₁₀ concentration of 5.9 16S rRNA gene copies/swab during the observational period and 6.3 16S rRNA gene copies/swab during valacyclovir suppression. Megasphaera species were observed at an average log₁₀ 3.8 copies/mL during the observational phase and 3.4 copies/mL during the valacyclovir phase. BVAB-2 was observed at an average log₁₀ of 3.2 copies/mL during the observational period and 2.9 copies/mL on valacyclovir therapy. Lactobacillus crispatus and L jensenii both had an average log₁₀ concentration of 3.9 copies/mL during the observational period and 4.3 and 4.1 copies/mL, respectively, during valacyclovir suppression. Lactobacillus iners had an average log₁₀ 7.5 copies/mL during the observational period and 7.4 copies/mL during valacyclovir suppression. A second sensitivity analysis evaluating the change in microbiota was performed in only participants who shed HSV-2 (n = 18) during the baseline phase and showed no significant difference in microbiota (Table 3).

Table 2.

Per-Person Mean Concentrations of Bacterial Species DNA Measured as 16S Ribosomal RNA Gene Copies/Swab by Quantitative Polymerase Chain Reaction

Bacteria	Average Log10 Copies/Swab, Observational	Average Log10 Copies/Swab, on Valacyclovir	Change, Valacyclovir vs Baseline (95% CI)	P Value
Gardnerella spp	5.9	6.3	0.4 (−.1 to .9)	.13
Megasphaera spp	3.8	3.4	−0.4 (−1.0 to .2)	.21
BVAB-2	3.2	2.9	−0.2 (−.7 to .3)	.36
Lactobacillus crispatus	3.9	4.3	0.5 (−.1 to 1.0)	.10
Lactobacillus iners	7.5	7.4	−0.1 (−.4 to .3)	.85
Lactobacillus jensenii	3.9	4.1	0.2 (−.3 to .8)	.41

Abbreviations: BVAB-2, bacterial vaginosis–associated bacterium 2; CI, confidence interval.

Table 3.

Per-Person Mean Concentrations of Bacterial Species DNA Measured as 16S Ribosomal RNA Gene Copies/Swab by Quantitative Polymerase Chain Reaction for Participants With Herpes Simplex Virus Type 2 Shedding Detected at Baseline (n = 18)

Bacteria	Average Log10 Copies/Swab, Observational	Average Log10 Copies/Swab, on Valacyclovir	Change, Valacyclovir vs Baseline (95% CI)	P Value
Gardnerella spp	5.2	5.8	0.6 (−.4 to 1.6)	.23
Megasphaera spp	3.7	3.8	0.2 (−.5 to .9)	.62
BVAB-2	3.2	3.5	0.3 (−.4 to .9)	.40
Lactobacillus crispatus	3.8	3.9	0.1 (−.1 to .3)	.29
Lactobacillus iners	7.8	7.8	0.1 (−.4 to .6)	.25
Lactobacillus jensenii	4.0	3.8	−0.2 (−.7 to .2)	.33

Abbreviations: BVAB-2, bacterial vaginosis–associated bacterium 2; CI, confidence interval.

DISCUSSION

We evaluated the impact of HSV-2 viral suppression with 6 weeks of daily valacyclovir on the prevalence of BV and change in the vaginal microbiota, as measured by Nugent score and by bacterial taxon–specific qPCR, among women with HSV-2 infection and a self-reported history of BV. In our study, daily valacyclovir suppression did not significantly impact the median Nugent score or alter the concentrations of key vaginal bacteria indicative of health or BV.

We observed an HSV-2 shedding rate of 9.7% of days among this population during the observational period, which is comparable to prior studies [8, 33]. As anticipated, daily valacyclovir was highly effective at suppressing genital HSV-2 shedding, with 94% reduction in viral shedding during valacyclovir period. Nugent scores fluctuated in participants over days, consistent with prior studies [34].

We hypothesized that suppression of HSV-2 would be associated with a shift toward optimal vaginal microbiota. However, no significant change was observed in selected vaginal bacteria before and during valacyclovir treatment. These data suggest that short-term HSV-2 suppression with daily valacyclovir is not associated with alterations in the vaginal microbiota, using the indicator bacteria assessed in this study.

Several studies have shown an association between BV and HSV-2; however, causality has not been demonstrated. A meta-analysis of 3 prospective studies, with minimal heterogeneity, found that the relative risk for HSV-2 on BV was 1.55 (95% CI, 1.30–1.84) [17]. Other studies have suggested that women infected with HSV-2 have lower rates of lactobacilli that produce hydrogen peroxide and are less responsive to BV therapy than women without HSV-2 infection [16, 35]. Conversely, women infected with BV have a 2.1- to 2.4-fold increased risk for HSV-2 acquisition [13, 14]. Based on these epidemiologic data, we hypothesized that inflammation associated with HSV-2 shedding was a key driver in the observed association between and BV and HSV-2. However, the results of our study do not support the hypothesis that HSV-2 shedding leads to an increased risk of BV or increased concentrations of BV-associated bacteria, based on the finding that nearly eliminating HSV-2 shedding did not change the Nugent score or bacterial species that are indicative of BV.

Complicating our understanding is the shared risk factors for BV and HSV-2 including multiple sex partners, a new sex partner, prior STIs, Black race, and low socioeconomic status [36–38]. Both BV and HSV-2 impact inflammation at the mucosal level. Understanding if a causal relationship exists between BV and HSV-2 could offer different treatment approaches and improve prevention of acquisition for both infections. Potentially there is a causal relationship between BV and HSV-2; however, valacyclovir cannot sufficiently suppress the effects of HSV-2 on the vaginal microbiome. It is possible that the 6-week duration of valacyclovir suppression for this study was not sufficient to reverse the HSV-2–induced inflammation and that longer durations of valacyclovir therapy could yield different results. A sensitivity analysis including only the last 2 weeks of valacyclovir demonstrated an insignificant trend for a decrease in the BV positivity ratio, supporting the hypothesis that a longer duration of suppressive antiviral therapy may have a more significant impact on BV. In addition, this study was powered to detect a 2.5-point change in the median Nugent score during antiviral treatment. Although we recruited a population of women with a history of BV, during the observational period the median per-person Nugent score was 3.8, suggesting that either this measure was not the optimal endpoint or that we should have recruited people with active BV to see a clinically significant difference with antiviral suppressive therapy. Aside from antiviral therapy, other preventive strategies such as an HSV-2 vaccine could be effective in reducing the occurrence of BV. Specifically, prophylactic vaccines that prevent HSV-2 acquisition would likely avoid cervicovaginal inflammation and may prevent BV. To define the underlying associations between BV and HSV, well-planned comprehensive studies to evaluate temporality are needed. Interventional studies coupled with longitudinal shedding studies are critical to elucidate any causal relationship between BV and HSV-2. Future studies could include populations with active BV and could evaluate a broader vaginal microbiota including community state types not included in our study [39].

Our study included several strengths. To our knowledge, this is the first study evaluating the direct effect of HSV-2 on BV, via the suppression of HSV-2 infection. Participants were adherent with valacyclovir therapy as well as with genital and anogenital swab collection, and although adherence was measured by pill counts only, valacyclovir had the expected effect of HSV-2 suppression as documented by PCR. The crossover study design allowed participants to act as their own comparators. The study also had several additional limitations. It is possible that the short duration of the study did not capture the true nature of HSV-2 and BV interaction, and longer periods of HSV-2 suppression may be necessary to have a measurable impact on BV and the vaginal microbiome. In addition, only 43% of the participants had genital HSV shedding during the observational period, and therefore power of the study may have been limited. We did not adjust for other BV risk factors such as menstruation and sexual frequency in our analyses. The study population was small, and the diversity of participants did not represent the epidemiologic distribution of disease burden in the general population. In our study, a high proportion of participants identified as White, when the burden of BV and HSV-2 is disproportionally borne by those who identify as Black and Hispanic/Latinx [40, 41]. The 1-way crossover study design did not evaluate the effect of BV suppression on HSV shedding.

CONCLUSIONS

Our study demonstrated that the suppression of HSV-2 using daily valacyclovir for 42 days did not decrease BV as measured by Nugent score nor change the composition of the vaginal microbiota. Future large-scale, longer-term intervention studies may be useful to evaluate the potential “epidemiologic synergy” of HSV-2 and BV and to understand the underlying biologic mechanisms for these interactions in women, including how treatment or suppression of BV affects genital HSV shedding and acquisition of HSV and other STIs.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH) (grant number U19 AI113173, principal investigator: Anna Wald).