Skip Navigation

Oral Herpes Simplex Virus Type 2 Reactivation in HIV-Positive and -Negative Men

  1. H. Nina Kim1,
  2. Amalia Meier3,5,
  3. Meei-Li Huang4,
  4. Steve Kuntz5,
  5. Stacy Selke5,
  6. Connie Celum1,2,
  7. Lawrence Corey1,3,4 and
  8. Anna Wald1,2
  1. 1Department of Medicine, Division of Allergy and Infectious Diseases, and Departments of
  2. 2Epidemiology and
  3. 3Laboratory Medicine,
  4. 4Fred Hutchinson Cancer Research Center, and
  5. 5Virology Research Clinic, University of Washington, Seattle
  1. Reprints or correspondence: Dr. H. Nina Kim, University of Washington, 901 Boren Ave., Ste. 1300, Seattle, WA 98104 (hyangkim{at}u.washington.edu)

Abstract

BackgroundPrevious studies using viral cultures rarely reported herpes simplex virus type 2 (HSV-2) isolation from the mouth. We sought to characterize oral HSV-2 shedding as detected by HSV DNA polymerase chain reaction among HSV-2–seropositive men

MethodsParticipants collected daily swabs from oral and anogenital areas for HSV detection with a quantitative polymerase chain reaction assay

ResultsA total of 109 HSV-2–seropositive men (59 of whom were human immunodeficiency virus [HIV] negative, and 50 of whom were HIV positive) were sampled for a median of 64 consecutive days. Forty-four (40.4%) had HSV-2 detected from oral swabs on at least 1 day. Oral HSV-2 was detected on 148 (2.3%) of 6422 days, genital HSV-2 was detected on 1110 (17%) of 6505 days, oral HSV-1 was detected on 220 (5.5%) of 4018 days, and genital HSV-1 was detected on 88 (2.2%) of 4073 days. Oral HSV-2 shedding was never associated with an oral lesion, but it was often concurrent with genital HSV-2 shedding. Both oral and genital HSV-2 were detected on 90 (61%) of 148 days with oral HSV-2 shedding. Oral HSV-2 shedding occurred on 90 (8.2%) of 1110 days with genital HSV-2 shedding, versus 58 (1.1%) of 5316 days without genital HSV-2 shedding (P<.001). The HIV-positive men shed HSV-2 orally more frequently than did the HIV-negative men (odds ratio, 2.7 [95% confidence interval, 1.1–7.1])

ConclusionsOral HSV-2 reactivation was common (especially among HIV-positive men), was always asymptomatic, and often occurred on days of genital HSV-2 reactivation

Both herpes simplex virus type 1 (HSV-1) and HSV-2 can cause mucocutaneous infections in the oral-facial and anogenital regions and are clinically indistinguishable. However, these viral types display differences in their rates of reactivation at a specific site. Although anecdotal case reports have described the isolation of HSV-2 from the oropharynx [14], reactivation of oral HSV-2 has been infrequent, with rates of HSV-2 isolation by culture of 0.09% of days from oral swabs in a prospective cohort study of HSV-2–seropositive persons [5]. Among persons who acquired concomitant oral and genital HSV-1 or HSV-2 infection, both HSV-1 and HSV-2 reactivated from oral and genital sites, but persons with concomitant oral and genital HSV-2 infections were 330-fold more likely to experience genital rather than oral recurrences [6]. Conversely, oral HSV-1 infection recurred more frequently than oral HSV-2 infection. This suggests that both viral and site-specific factors are involved in the reactivation of HSV infection at oral and genital sites

To gain further understanding of oral HSV-2 reactivation, we characterized the frequency of oral HSV-2 shedding with detection of HSV DNA by polymerase chain reaction (PCR), the host factors associated with oral HSV-2 shedding, and the temporal pattern of oral HSV-2 reactivation in relation to genital HSV reactivation in a prospective cohort of men

Methods

Participants and settingHSV-2–seropositive men were enrolled in prospective observational studies conducted at the University of Washington Virology Research Clinic (Seattle) from 1994 to 2002. The participants were taught how to collect oral and genital swabs for HSV detection, as described elsewhere [7]. This analysis was restricted to men who collected samples on at least 30 consecutive days. Only men were included in this analysis, because HIV infection was an important variable of interest, and the participants who collected daily oral samples for this observational study included very few women with HIV infection. Written, informed consent was obtained from all study participants, and the study was approved by the University of Washington institutional review board

Data collectionDemographic and clinical data were collected on standardized forms. Participants were taught to recognize clinical herpes lesions by an experienced research clinician and to record these daily. Men were instructed to collect 1 specimen daily from each of 3 anatomic sites: oropharynx, penile skin, and perianal area. Oropharyngeal specimens were obtained by inserting a Dacron swab into the mouth and rubbing it vigorously along the gum line and over the palate. Male participants also swabbed the penile skin and perianal area, as described elsewhere [8, 9]. The reliability of the patient sampling method for detecting HSV DNA from mucosal and skin surfaces has been established in other studies and has been determined to be comparable or superior to clinician sampling [9, 10]. HIV-seropositive subjects were classified as receiving HAART if, on day 1 of their study participation, they were taking at least 3 drugs in 2 of the following categories: protease inhibitor, nucleoside analogue, or nonnucleoside analogue

Laboratory methodsSwabs for HSV DNA were placed in a 1× digestion buffer, were refrigerated, and were delivered to the laboratory every other week. Genital and oral samples were first evaluated for HSV DNA by a quantitative, real-time, fluorescence-based PCR assay that identifies both HSV-1 and HSV-2 [11, 12]. DNA specimens from all samples that were positive for HSV were then assayed with type-specific probes labeled with different fluorescent dyes, a method that has been shown to have a specificity of 99.5% for differentiating HSV-1 from HSV-2 [13]. We reported as positive those samples in which ⩾10 copies of HSV DNA per 20 μL reaction (or ⩾500 copies/mL) were detected

Serum samples were tested for HSV-1 and HSV-2 antibodies by Western blot [14]. HIV serostatus was confirmed using standard enzyme-linked immunoassay and Western blot. Absolute CD4+ cell counts were determined by flow cytometry. HIV-1 RNA levels were measured with the Roche Amplicor 1.0 assay (detection level, 500 HIV-1 RNA copies/mL), and, for those with <500 copies/mL, levels were measured with the Ultradirect assay (detection level, 50 HIV-1 RNA copies/mL) (Roche Diagnostic Systems)

Statistical analysisHSV shedding was the main outcome measure and was characterized as (1) a dichotomous variable (yes or no) based on the detection of HSV DNA by PCR on at least 1 day, (2) a shedding rate, defined by the number of days during which HSV was detected by PCR divided by the number of days tested, and (3) a quantity of shedding, defined by log10 copies of HSV DNA per mL of PCR buffer. These shedding outcomes were computed by anatomic site (oral or genital) and viral type (HSV-1 or HSV-2). Results for HSV-1 included only persons who were seropositive for both HSV-1 and HSV-2

Binomial regression using a logit link (for odds ratios [ORs]) and a correction factor for larger-than-expected variance (overdispersion) were used to calculate univariate and multivariate ORs. The main outcome of interest was oral HSV-2 shedding. The following baseline variables were tested: HIV status, age >50 years, HSV-1 and HSV-2 serostatus, sexual preference (men who have sex with men [MSM] or heterosexuals), duration of study follow-up ⩾80 days, history of genital herpes, history of oral herpes, and whether the participant shed genitally during the study. Univariate predictors with P values ⩽.1 were included in the multivariate model, in addition to other variables that were not significant in the univariate analysis but were thought to be biologically relevant

Generalized estimating equations with Gaussian distribution and robust SEs were used to compare day-level quantities of HSV-2 DNA by anatomic site. Intraperson correlation due to multiple observations per person were accounted for in this generalized estimating equation model. To assess whether oral HSV-2 and genital HSV-2 shedding outcomes were associated, day-level shedding data were compared for all subjects. The odds of shedding orally when genital shedding was present versus when it was not present were compared in a generalized estimating equation model. An OR >1 indicated that oral shedding was more likely to occur on days when genital shedding was present

Comparisons of person-level shedding rates by site (e.g., oral HSV-2 vs. genital HSV-2) or by HSV type (e.g., oral HSV-2 vs. oral HSV-1) were made by paired nonparametric testing with the Wilcoxon sign&amp;rank test. Comparisons of nonpaired person-level variables were made using the 2-sample Mann-Whitney U test. Normally distributed continuous variables were compared using t tests, and categorical variables were compared using the χ2 test. Additional analyses were conducted for HIV-seropositive participants, including CD4+ T lymphocyte count (evaluated as both a continuous and categorical variable) and plasma HIV RNA load (also assessed as both a continuous and categorical variable). These variables were tested for association with oral HSV-2 shedding outcomes as described above. Two-sided P values were calculated, and values <.05 were considered to be significant. All statistical calculations were performed using Stata software, version 8.2 (Stata)

Results

Characteristics of study participantsA total of 109 HSV-2 seropositive men were enrolled; their clinical and demographic characteristics are shown in table 1, according to HIV status. The median age was 42 years (range, 24–76 years) and most participants (95 [87%]) were white. MSMs comprised 83% of the study population (91 of 109 subjects). Of the 109 men, 68 (62%) were HSV-1 seropositive, 73 (67%) had a history of genital herpes, and 39 (36%) had a history of oral herpes

Among the 50 men who were HIV seropositive, 27 (54%) were receiving HAART. At study entry, the median CD4+ cell count was similar for untreated persons (327 cells/mm3) and persons receiving hart HAART (346 cells/mm3) (P=.49). Median plasma HIV RNA levels were 31,656 copies/mL for HIV-positive men not receiving HAART and 177 copies/mL for men receiving HAART (P<.001). Of the 27 men receiving HAART, 8 (30%) had HIV RNA loads <50 copies/mL

The study participants collected samples for a median of 64 days (range, 30–127 days). A total of 6505 genital swabs and 6422 oral swabs were collected and included in the analysis. Of 12,927 samples, 1566 days were positive for HSV: 1258 days (9.7%) were positive for HSV-2, and 308 days (2.4%) were positive for HSV-1. Twelve samples (6 oral and 6 genital) had HSV DNA detected but were not typeable

HSV shedding by anatomic site and viral typeThe overall frequency of genital HSV-2 detection was higher than that of oral HSV-2, and the frequency of oral HSV-1 detection was higher than that of genital HSV-1. Of 109 HSV-2–seropositive participants, 84 (77%) had genital HSV-2 detected on at least 1 day, and 44 (40.4%) had oral HSV-2 detected. Of the 68 HSV-1–seropositive men, 41 (60.3%) had oral HSV-1, and 21 (31%) had genital HSV-1 on at least 1 occasion. Figure 1 shows the percentage of HIV-positive and HIV-negative men with detectable HSV on at least 1 day, by site and HSV type

Figure 1

Percentage of men who had herpes simplex virus (HSV) detected on at least 1 day

Proportion of days with detectable HSV according to site, HSV type, HIV status, and sexual orientation are shown in table 2. Of total days observed for all men, the rate of oral HSV-2 shedding was 2.3%, compared with 17% for genital HSV-2 and 5.5% for oral HSV-1 (P<.001, for 3 pairwise comparisons of person-level shedding rates by site and HSV type). The rate of genital HSV-1 shedding (2.2% of days) was comparable to that of oral HSV-2 shedding

Figure 2

Person-level, site-specific variations of herpes simplex virus (HSV) shedding. Gray bars Shedding at the genital site. Black bars Shedding at both the oral and the genital sites. White bars Shedding at the oral site

Figure 3

Representative herpes simplex virus (HSV) shedding patterns of 5 participants. Participants 2 and 4 showed intercurrent reactivation with HSV-1 and HSV-2 occurring a few days apart. Two individuals (1 of whom is featured here as participant 1) demonstrated high-copy shedding of oral HSV-2 (>4 log10 copies/mL) for >1 week; both were HIV positive. The numbering of the days does not represent the actual study day; rather, it represents a consecutive series of days within the duration of study participation. Numbers in the colored boxes are the quantity of HSV DNA (log10 copies/mL) recorded on that day. Black boxes Days on which genital lesions were present

Figure 4

Quantity of herpes simplex virus (HSV) DNA (log10 copies/mL), by anatomic site and viral type in samples from persons with detectable HSV. The horizontal line represents the median quantity

Table 1

Characteristics of study participants

Table 2

Proportion of days on which herpes simplex virus (HSV) was detected among study participants

Characteristics of oral HSV-2 sheddingOral HSV-2 shedding was often concurrent with genital HSV-2 shedding; genital HSV-2 was detected on 90 (61%) of 148 days on which oral HSV-2 shedding occurred. Oral HSV-2 shedding occurred on 90 (8.2%) of 1110 days with genital HSV-2 shedding, compared with oral HSV-2 shedding occurring on 58 (1.1%) of 5316 days with no genital HSV-2 shedding (OR, 2.1 [95% CI, 1.1–3.1]; P<.001). The proportion of days with concurrent oral and genital HSV-2 shedding did not differ between HIV-positive and HIV-negative men

Oral HSV-2 shedding always occurred without oral lesions. In contrast, genital lesions were noted on 303 (27.3%) of 1110 days on which genital HSV-2 shedding occurred, oral lesions were noted on 21 (9.5%) of 220 days on which oral HSV-1 shedding occurred, and genital lesions were noted on 12 (13.6%) of 88 days on which genital HSV-1 shedding occurred (P<.001, for all pairwise comparisons of person-level shedding rates by site and HSV type). Two (12.5%) of the 16 men who were HSV-1 seronegative who shed oral HSV-2 reported a history of oral herpes, which suggests that oral HSV-2 infection can manifest infrequently as cold sores

Overall, only 5 (4.6%) of the 109 HSV-2–seropositive men had only oral and no genital HSV-2 shedding during follow-up (figure 2). In the subset of 36 HSV-2–seropositive men who had no history of genital infection and were identified solely on the basis of their positive HSV-2 serostatus, HSV-2 DNA was detected only in the genital area in 14 (39%), in the oral and genital area in 8 (22%), and in the oral area only in 3 (8%)

Quantity of HSV DNA by site and viral typeThe quantity of HSV-2 DNA detected by PCR in the mouth tended to be lower than that of HSV-2 detected in the genital area, HSV-1 in the mouth and HSV-1 in the genital area (figure 3 and 4). For the 148 days on which HSV-2 was detected in oral swabs, the median quantity was 3.4 log10 copies/mL (range, 2.6–6.1 log10 copies/mL), compared with 5.1 log10 copies/mL (range, 2.6–9.2 log10 copies/mL) for the 1110 days on which HSV-2 was detected in genital swabs. The median quantity was 4.5 log10 copies/mL (range, 2.6–8 log10 copies/mL) for the 220 days on which oral HSV-1 was detected and 5 log10 copies/mL (range, 2.6–8.1 log10 copies/mL) for the 88 days on which genital HSV-1 shedding occurred

For the 90 days when both oral and genital HSV-2 DNA were detected, the mean HSV-2 DNA quantity was 3.6 log10 copies/mL for oral HSV-2, versus 6.2 log10 copies/mL for genital HSV-2 (2.6 log10 copies/mL lower for oral HSV-2 [95% CI, 2.2–2.9]; P<.001). There was no statistically significant difference in oral HSV-2 DNA quantity between the days with and without concurrent genital HSV-2 shedding (3.6 vs. 3.8 log10 copies/mL; P=.36)

Quantity of oral HSV-2 did not differ substantially by HIV status; the mean quantity for the 45 days with oral HSV-2 shedding among HIV-negative individuals was 3.6 log10 copies/mL, compared with 3.7 log10 copies/mL for the 103 days with oral HSV-2 shedding among HIV-positive individuals (difference of +0.1 [95% CI, −0.2 to 0.5]; P=.43)

Risk factors for oral HSV-2 sheddingParticipants with HIV infection were more likely to have higher oral HSV-2 shedding rates (OR 2.7 [95% CI, 1.1–7.1]; P=.04) (table 3). They were also more likely to have higher genital HSV-2 shedding (OR 2.3; 95% CI, 1.4–3.8; P=.001). Age >50 years old, dual seropositivity for HSV-1 and HSV-2, history of genital herpes, history of oral herpes, a duration of follow-up ⩾80 days, and MSM status had no association with oral HSV-2 shedding rate. Inclusion of these variables in a multivariate model did not change the value of the OR for HIV infection or the P values

Table 3

Risk factors for oral herpes simplex virus type 2 (HSV-2) shedding

Among HIV-seropositive participants, we observed no correlation between CD4+ T lymphocyte counts or HIV RNA and oral HSV-2 shedding, rate, or quantity. In addition, the use of HAART did not affect the risk of oral HSV-2 shedding

Discussion

This study defines several new observations about shedding of HSV-2 in the oropharynx, which is more frequent than previously described in studies using viral isolation [5, 6]. The frequency of days with oral HSV-2 shedding (2.3%) was about half of that with oral HSV-1 shedding (5.5%). Oral HSV-2 shedding was always asymptomatic, occurred concurrently with genital HSV-2 shedding most of the time, and tended to be detected in lower quantities compared with oral HSV-1 and genital HSV-1 and HSV-2

Our findings suggest that HSV-2 can establish latency in the trigeminal ganglia, even in those also infected with HSV-1, because the frequency of oral HSV-2 detection did not differ in persons with HSV-2 antibody only, compared with persons who were HSV-1 and HSV-2 seropositive. Exogenous reinfection is thought to be less common than endogenous reactivation of latent virus [15, 16] and is unlikely to account entirely for the substantial proportion of men in our cohort who had detectable oral HSV-2 at repeated time points during follow-up. Detection of both viral types at the same anatomic site has been described in a case series for genital HSV-1 and HSV-2 [17] and in an autopsy study demonstrating both viral types in explanted human spinal ganglia from cervical to sacral regions [18]

Oral HSV-2 infection could originate from direct inoculation of virus locally—through oral-oral or oral-genital contact—or from viremic spread of primary anogenital infection. Similar oral HSV-2 shedding rates among MSMs and heterosexual men argue against a strong association between sexual practices and site of HSV-2 infection, although our sample of heterosexual men was small

The biological basis for the propensity of HSV-1 to reactivate in the oral area and of HSV-2 to reactivate in the genital area is not well understood. Differences in cell surface entry receptors do not explain the site specificity between these viral types, because both HSV-1 and HSV-2 have been demonstrated to infect oral epithelial cells [19], and HSV-2 can present in the mouth as well as in the genital area [1, 20]. Animal models suggest that HSV sequence differences in the latency-associated transcript may determine anatomic tropism [21, 22]. In the rabbit eye model, substitution of the HSV-2 latency-associated transcript region with that of HSV-1 resulted in a higher rate of ocular reactivation. This reversal in site-specific phenotype was also demonstrated in the guinea pig vaginal model of HSV-2, in which inoculation with this HSV-2 strain engineered to include the HSV-1 latency-associated transcript sequence resulted in a reduced frequency in genital recurrences; restoration of HSV-2 latency-associated transcript reestablished the higher rate of genital reactivation

Concurrent oral HSV-2 reactivation with genital HSV-2 reactivation, shown also in a culture-based study of oral HSV-2 shedding [5], suggests that systemic, rather than local, factors play a role in HSV reactivation. The triggers associated with the switch from viral latency to reactivation remain a subject of investigation. Local dermal triggers, such as ultraviolet radiation and trauma, as well as systemic stimuli, such as immunosuppression with dexamethasone or cyclophosphamide [23, 24] and fever [25], have all been shown to induce HSV reactivation in humans and animal models [26]. In addition to the potential influence of proinflammatory cytokines, ongoing immune surveillance—characterized in part by the presence of HSV-specific CD8+ T lymphocytes near latently infected neurons [2729]—may play an important role in regulating HSV reactivation. Our findings of concurrent reactivation at remote mucosal sites provide support that systemic events may mediate HSV reactivation

The disparity we observed between the quantities of oral HSV-2 DNA versus genital HSV-2 DNA or oral HSV-1 DNA is intriguing. Lower quantities of oral HSV-2 DNA may explain why prior studies that were based on viral culture observed such a low rate of oral HSV-2 reactivation, because the rate of HSV isolation is directly proportional to HSV DNA copy numbers [30]. The lower quantity of oral HSV-2 might also imply that local control of viral replication is more effective for oral HSV-2 than it is for the other sites and viral types

The HIV-positive men shed HSV-2 orally more frequently than the HIV-negative men did. Impaired immune function may increase the frequency of oral HSV-2 reactivation, but the use of highly active antiretroviral therapy did not appear to affect the frequency of oral HSV-2 shedding among HIV-infected men, which is similar to what has been reported previously for persons with genital HSV-2 infection [31]

Because our study population consisted largely of urban MSMs who agreed to take part in labor-intensive, daily home-sampling protocols, a limitation of this study may be lack of generalizability. Although it is doubtful that agreeing to participate in swabbing studies would influence oral HSV-2 shedding, it is possible that a lower-risk cohort with fewer sexual exposures and partners would not demonstrate oral HSV-2 reactivation to the same extent as this group of mostly MSMs. Erroneous detection of oral HSV-2 because of cross-contamination on genital shedding days is an unlikely source of misclassification, because quantities of HSV-2 DNA did not differ substantially between days with and without genital shedding. Falsely low oral HSV-2 DNA quantity because of inhibition of the real-time PCR assay in oral specimens is also unlikely, because we did not see lower copy numbers with oral HSV-1

Whether this lower quantity of oral HSV-2 translates to lower rates of oral transmission is not known. The relationship between viral load and probability of sexual transmission has not been established for HSV. However, seroepidemiologic studies show that HSV-2 is acquired after initiation of sexual activity, which suggests that genital transmission is, perhaps, more efficient for HSV-2. Importantly, our findings support that HSV-2 antibody in persons without a history of genital herpes most likely represents genital HSV-2, because most such men had HSV-2 detected in the genital area and only a few had oral HSV-2 in the absence of genital HSV-2 detection

Our data confirm the differences in anatomic tropism between HSV-1 and HSV-2, manifested in the lower frequency of reactivation of oral HSV-2 and genital HSV-1, but also suggest that oral HSV-2 infection may be more common than previously supposed. With the availability of type-specific assays, clinicians and counselors will need to convey a more nuanced message about the natural history of these 2 viruses. Specifically, individuals who have HSV-2 infection should be informed about the possibility of asymptomatic oral HSV-2 reactivation, particularly if they are coinfected with HIV

Acknowledgments

We thank Elizabeth Krantz for assistance with statistical analysis

Footnotes

  • Presented in part: 16th Biennial International Society for Sexually Transmitted Diseases meeting, Amsterdam, The Netherlands, July 2005 (abstract TP-021)

    Potential conflicts of interest: none reported

    Financial support: National Institutes of Health Herpes Program Project (grant AI-30731); National Institutes of Health STD/AIDS (research training grant T32 AI007140 to H.N.K.)

  • Received January 28, 2006.
  • Accepted March 30, 2006.

References

| Table of Contents