Herpes simplex virus (HSV) infection is common in persons coinfected with human immunodeficiency virus (HIV). In a prospective study, daily viral cultures of the mouth, genitals, and rectum were collected from 68 HIV-positive and 13 HIV-negative men who have sex with men. Subjects completed a median of 57 days of follow-up. Anogenital HSV-2 cultures were positive on 405 (9.7%) of 4167 days for HIV-positive men and on 24 (3.1%) of 766 days for HIV-negative men. Most reactivations were perirectal and subclinical. Risk factors for increased HSV-2 shedding among HIV-positive men were low CD4 cell count (odds ratio, 2.5; 95% confidence interval, 1.2–5.4) and antibodies to both HSV-1 and HSV-2 versus HSV-2 only (odds ratio, 1.9; 95% confidence interval, 1.0–3.7). Three isolates obtained from 3 separate subjects were resistant to acyclovir. Thus, subclinical HSV-2 reactivation is an important opportunistic infection in persons with HIV infection. Further studies are necessary to determine the impact of subclinical HSV-2 reactivation on the natural history of HIV infection.
Persistent anogenital lesions due to herpes simplex virus (HSV) were one of the first opportunistic infections to be described in persons with AIDS . While recent reports indicate that 22% of the general population is infected with HSV-2 , seroepidemiologic studies have established that HSV is among the most common viral coinfection in persons with AIDS, with 95% of men who have sex with men (MSM) and 40%–60% of injection drug users demonstrating serum antibodies to either HSV-1, HSV-2, or both [3–8]. Despite these high seroprevalence data, few studies of the clinical course of HSV infections among human immunodeficiency virus (HIV)-positive persons have been done to document the frequency and anatomic sites of infection as well as duration of clinical versus subclinical HSV-2 reactivation. This is important, as HSV-2 has been linked to both the acquisition and transmission of HIV-1 [8–14] and, in vitro, has been shown to up-regulate HIV [15–20]. To resolve some of these issues, we instituted a prospective study to evaluate the frequency, location, and duration of subclinical and clinical HSV reactivation among HIV-positive men.
Study Participants and Design
HSV-2-seropositive MSM (both HIV-1-seropositive and HIV-1-seronegative) were recruited into this study. Subjects were recruited from local physicians and through advertisements. Eligibility criteria included HSV-2 serum antibodies demonstrated by Western blot and age >18 years. A history of anogenital herpes was not required. There were no restrictions due to sex, race, or stage of HIV-1 infection. While we made no sex restrictions, only 2 HIV-positive women were enrolled, reflecting the epidemiology of HIV-1 infection and AIDS in Washington State at the time the study was completed. As such, we have limited our analyses to men. At entry, each subject completed a standardized interview to record history of prior HSV recurrences, frequency and type of past antiviral therapy for both HSV and HIV-1, and staging of HIV-1 disease (using the 1993 CDC criteria).
Collection of Daily Home HSV Cultures
The study protocol involved the daily collection of viral cultures from oral and genital sites for 60 consecutive days. This time period was selected to optimize compliance and provide a reasonable assessment of the true rate of HSV-2 shedding. The methods used in the daily collection of swabs at home have been described [21–25]. Each person collected 4 specimens daily from each anatomic site (oropharynx, penile shaft, urethra, and rectum) using separate Dacron swabs for each culture and placing them in separate transport containers. Participants were instructed to obtain cultures at the same time each day, preferably on awakening. Oral or pharyngeal cultures were obtained by inserting a Dacron swab into the mouth and vigorously rubbing it along the gum line and over the palate. Urethral cultures were obtained by rubbing a Dacron swab over the urethral meatus. Penile skin cultures were obtained by rubbing a Dacron swab along the entire ventral and dorsal shaft of the penis. Perianal or rectal cultures were obtained by inserting the swab ∼2–4 cm into the anus and gently rotating it. Buttock cultures were obtained in people identified (at entry) as having a history of HSV recurrence on the buttocks. Each Dacron swab was placed into a separate vial containing virus transport media and stored in the refrigerator. Cultures were picked up by a courier within 36 h of collection, transported to the virology laboratory of the University of Washington, and immediately set up in tissue culture.
Subjects also filled out a daily diary card that recorded site-specific presence or absence of symptoms (pain, tingling, numbness, or itching) and/or presence of lesions. All diary cards were collected monthly and reviewed by the study clinician.
Serologic assays and T cell analysis
HIV-1 seropositivity was confirmed by EIA and Western blot. Serologic testing for HSV antibodies was done on entry sera by Western blot analysis . CD4 subset analysis was done by the hematopathology laboratory of the University of Washington, which is certified by the AIDS Clinical Trials Group for this assay.
Isolation procedures for HSV have been previously described . Initially, each culture was inoculated into duplicate tubes of diploid fibroblasts and triplicate wells of 48-well microtiter plates. Because similar results were obtained with individual tubes and with microtiter plates, after the first 10 subjects were enrolled, cultures were inoculated only onto the 48-well plates. Each culture was maintained for 28 days and examined three times per week. All wells exhibiting cytopathic effects were confirmed and typed by use of monoclonal antibodies.
A random sample of culture-positive virus isolates was tested for in vitro susceptibility to acyclovir by the dye-uptake assay [27, 28]. This is a cell protection, semiautomated, colorimetric method based on the amount of dye uptake measured in live cells. This technique measures a quantitative reduction in cell viability due to viral cytopathic effect. In a 96-well plate format containing serial dilutions of acyclovir, virus isolates are inoculated into Vero cells. Neutral red dye is added, and the amount of dye taken up by cells is measured. The plates are read at 540 nm. A dose-response curve is generated with HSV-uninfected cell controls (100% protection) and HSV-infected cells with no acyclovir added (no protection). IC50 is the concentration of acyclovir that inhibited >50% of viral replication. An IC50 ≥3.0 μg/mL was used as indicative of potential acyclovir resistance.
Study End Points and Definitions of Terms
Total, subclinical, and clinical HSV shedding rates were determined for each anatomic site and overall. The definitions used are similar to those described [21–26, 29]. Total HSV shedding rate was defined as total days that HSV was isolated divided by total days that cultures were obtained. Subclinical HSV shedding rate was defined as days on which the subject did not report a lesion at the site of a positive culture divided by total culture days with no reported lesions. Clinical HSV shedding was defined as total days of reported lesions from a site of positive culture divided by total culture days that a lesion was reported from that site. We defined a recurrence as consecutive days of lesions and an episode as consecutive days of viral shedding (independent of presence or absence of a lesion). If an episode of shedding was interrupted by ≥1 day (because either the patient did not perform a culture that day, a culture result was lost or contaminated, or the culture was negative), it was scored as two separate episodes. In addition to reporting culture results from separate anatomic sites, we combined sites in the anogenital area (penis, urethra, rectum, or buttock) to define overall anogenital shedding. If the culture from any anogenital site was positive on a given day, it was scored as a day of anogenital shedding. Clinical anogenital shedding required that a positive culture be accompanied by a lesion at some anogenital site.
Censoring for Days of Acyclovir Use
Patients were allowed to take acyclovir for episodes of HSV reactivation as prescribed by their regular physician. Episodic acyclovir use was uncommon and was taken by only 11 (14%) of 81 patients on 2% of study days. None of the patients used daily acyclovir therapy during the study period. Censoring days on which episodic acyclovir was taken had no effect on total, subclinical, or clinical HSV-2 shedding rate. However, duration of recurrences and shedding episodes would be underestimated if these days were omitted; therefore, days of acyclovir use were included in the analysis.
Characteristics of HIV-positive and HIV-negative men were compared with Fisher's exact and Wilcoxon rank sum tests. Two-sided tests at the 5% level of significance were used. Risk factors for HSV shedding were assessed using odds ratios (ORs) from logistic regression for grouped data , with the cultures for each subject grouped together. Since culture results for a given subject tend to be positively correlated, the variability among subjects is greater than assumed in the classical logistic regression model. A scale factor was estimated to account for this in assessing the significance of predictors and computing confidence intervals (CIs).
Study population and compliance to the protocol
We enrolled 81 HSV-2-seropositive men, of which 68 were HIV-positive and 13 were HIV-negative. All 13 HIV-negative men were MSM and described themselves as at risk for HIV-1 infection. There were no significant differences in the demographic characteristics, HSV history, or follow-up and compliance with study protocol between these 2 groups (table 1). Most (94%) were Caucasian, and among HIV-positive subjects, the principal risk factor for HIV-1 acquisition was sexual contact with another male. The median CD4 cell count at entry was 351/mm3 for HIV-positive men and 1022/mm3 for HIV-negative men. Most of the HIV-positive men were asymptomatic at entry, with 50 (73%) classified as CDC stage A1, A2, B1, or B2 and the remaining 18 (27%) as having AIDS (CDC stage A3, B3, or C1–3). The median age at entry for all 81 subjects was 37, and overall, 59 (73%) subjects reported a clinical history of anogenital HSV infection and 36 (44%) reported a history of oral HSV infection.
This study was completed before the widespread use of plasma HIV RNA testing as a means of monitoring HIV-1 disease progression. As such, plasma samples stored in appropriate buffer were available for only 22 of 68 HIV-positive men. Among these 22, the median plasma HIV-1 RNA (measured by the branched DNA assay; Chiron, Emeryville, CA) was 14,000 copies/mL. Overall, 32 (47%) of 68 HIV-positive men were receiving antiretroviral therapy, 34 (50%) were not receiving antiretroviral therapy, and 2 (3%) lacked therapy data. Of those, 22 (32%) of 68 were receiving monotherapy with either zidovudine (17/22) or didanosine or zalcitabine. Only 10 (15%) were receiving combination therapy, all with reverse transcriptase inhibitors.
The median number of days on which HSV cultures were obtained was 55, with HIV-positive men completing a total of 4171 culture-days and the 13 HIV-negative men completing a total of 766 culture-days. All men obtained cultures on more than half of the days on study, and 63 (78%) obtained cultures on >190% of days.
Frequency and duration of HSV-2 reactivation
A total of 31 of the 68 HIV-positive men (46%) and 4 (31%) of the 13 HIV-negative men reported anogenital lesions over follow-up. Overall, 52 separate recurrences (consecutive days of anogenital herpes lesions) were recorded among the HIV-positive men and 9 among HIV-negative men. The average monthly recurrence rate was slightly, though not significantly, higher among the HIV-positive men (0.34 recurrences/month, vs. 0.23 in the HIV-negative group). The median length of a recurrence was 6 days for both groups.
HSV-2 shedding episodes (consecutive days of positive culture) were much more frequent, though shorter, than recurrences, a feature noted in both HIV-positive and HIV-negative men. HSV-2 shedding rates were higher in the HIV-positive men (figure 1).
Overall, 51 of the 68 HIV-positive versus 7 of the 13 HIV-negative men exhibited at least one episode of HSV-2 shedding. More men (46 HIV-positive and 7 HIV-negative) experienced subclinical HSV-2 shedding than had lesions. Among the HIV-positive men who shed HSV-2, 29 shed only subclinically, 17 had both clinical and subclinical shedding, and 5 shed only clinically. The corresponding numbers for the HIV-negative men were 5, 2, and 0, respectively. There were 166 separate episodes of anogenital HSV-2 shedding among the HIV-positive men, 112 (67%) of which were subclinical. Among the episodes of subclinical shedding, 43 (38%) lasted >1 day. Among the HIV-negative men, there were 16 episodes of anogenital HSV-2 shedding, half subclinical, and all of the subclinical episodes lasted only 1 day.
The anatomic sites and frequency of HSV-2 shedding are shown in table 2. Among HIV-positive men, HSV-2 was isolated on 405 (9.7%) of the 4167 days from which cultures were obtained, and among HIV-negative men, on 24 (3.1%) of 766 days. The most frequent site of HSV-2 shedding among HIV-positive men was the rectal area, which accounted for 355 (79%) of the 449 positive cultures. Among HIV-negative men, the distribution of HSV reactivation was similar.
Multianatomic shedding of HSV in HIV-positive men
The median CD4 cell count of patients with multiple-site reactivation was 225 (range, 14–847). Of the 46 days with multiplesite reactivation, 27 (59%) involved anatomically distant sites (e.g., mouth and rectum, urethra and rectum), perhaps demonstrating the global immunosuppressive effects of HIV infection. Isolation of HSV from multiple anatomic sites on the same day was commonly seen among HIV-positive men (table 3). Fifteen (22%) HIV-positive men and 1 (8%) HIV-negative man shed HSV from multiple anatomic sites on the same day. Forty-six of the 405 days of HSV-2 shedding involved more than one anatomic site of infection among HIV-positive men. Of the 46 days, 43 (94%) were associated with rectal HSV-2 shedding and recovery on the same day: HSV-1 from the oropharynx (6 days), HSV-2 from the oropharynx (3 days), HSV-2 from the genitals (10 days), HSV-2 from the buttocks (18 days), and HSV-2 from two other anatomic sites on 6 days. In the HIV-negative man, all 5 days of multianatomic shedding included subclinical HSV-2 shedding from the rectum and clinical shedding (associated with a lesion) from the buttocks.
Frequency of HSV-1 shedding
HSV-1 was not recovered from the genitals or perirectal area in any of the men. HSV-1 was shed from the oropharynx on 59 (2.1%) of 2772 days in HIV-positive and 3 (0.9%) of 319 in HIV-negative men (table 2). There was no difference in the rate of clinical shedding (9.1% of days with oral lesions) between HIV-positive men and HIV-negative men; however, the rate of subclinical shedding was higher in HIV-positive men, who had 1.8% of days without lesions positive for HSV-1, compared with 0.3% in HIV-negative men. These differences were not, however, statistically significant.
Factors associated with an increased risk of HSV-2 shedding
HIV-positive men had significantly higher frequencies of both clinical and subclinical shedding. The relative risk of total anogenital HSV-2 shedding for HIV-positive versus HIV-negative men was 3.3 (95% CI, 1.1–9.9). For subclinical anogenital shedding (i.e., shedding on days when a lesion was not present at any anogenital site), the OR was 6.9 (95% CI, 1.3–36.8).
Among the 68 HIV-positive subjects, CD4 cell count, HSV antibody status, history of anogenital HSV, and stage of HIV-1 infection were evaluated as predictors of HSV-2 anogenital shedding (table 4). A low CD4 cell count was significantly associated with both total (figure 2) and subclinical anogenital HSV-2 shedding in univariate and multivariate analyses. The relative risk of total anogenital HSV-2 shedding for men with CD4 cell counts <200/mm3 compared with men with CD4 cell counts 1500 was 2.5 (95% CI, 1.2–5.4). The corresponding OR for subclinical shedding was 2.6 (95% CI, 1.1–6.2). In addition to low versus high CD4 cell count, persons with antibodies to both HSV-1 and HSV-2 were more likely to shed HSV-2 than were persons with only HSV-2 antibodies (OR, 1.9; 95% CI, 1.0–3.7). Past history of anogenital HSV was not associated with an increased risk of anogenital HSV-2 shedding.
In vitro susceptibility to acyclovir
A total of 160 HSV isolates from 25 patients were randomly selected to test for resistance to acyclovir by the dye-uptake method. Of these, 25 (16%) isolates were HSV-1 and 135 (84%) were HSV-2. Of the isolates subjected to sensitivity testing, 121 (76%) were from the rectum, 6 (4%) were from the genitals, and 33 (21%) were from the oropharynx. The mean IC50 for these 160 isolates was 1.37 μg/mL (SD, 0.639; range, 0.26–5.1). There were no significant differences in the mean IC50 for isolates obtained from the oropharynx, genitals, or rectum (mean IC50, 1.32, 1.32, and 1.40 μg/mL, respectively) or for isolates that were HSV-1 compared with those that were HSV-2 (mean IC50, 1.28 and 1.40, respectively; table 5). Overall, 3 (1.9%) of 160 isolates, obtained from 3 different patients, were found to have reduced sensitivity to acyclovir in vitro (IC50, ≥3 μg/mL). Two of the patients had used acyclovir during the study (1 patient on 15 of 56 days of follow-up and the other on 9 of 61 days of follow-up). Acyclovir-resistant HSV-1 (IC50, 5.1 μg/mL) was isolated from the oropharynx of a patient with antibodies to both HSV-1 and HSV-2 whose CD4 cell count was 510/mm3. Acyclovir-sensitive HSV-1 was isolated from the same site 6 days before and 11 days after the resistant isolate was obtained (IC50, 1.0 and 1.3 μg/mL, respectively). Acyclovir-resistant HSV-2 was isolated from the rectum of 2 different patients (IC50, 3.0 and 4.1 μg/mL) with CD4 cell counts of 46 and 608/mm3, respectively. In both patients, acyclovir-sensitive HSV-2 was isolated from the same site both before and after isolation of the acyclovir-resistant HSV-2.
This observational cohort study demonstrates several issues of relevance to the care of HIV-positive men with HSV-2 infection. Even with a relatively short duration of follow-up, most HIV- and HSV-2—positive men reactivate HSV-2. Overall, HSV-2 was isolated from the anogenital area of HIV-positive men on 9.7% of days, compared with 3.1% of days in HIV-negative men. While there are few published data about HSV-2 shedding rates in immunocompetent men, HSV-2 shedding rates in our HIV-negative control group of MSM were comparable to those found in a group of 96 immunocompetent women with HSV-2, who reactivated HSV-2 subclinically on 2% of days .
In HIV-positive subjects, HSV-2 shedding was subclinical on 2 of 3 days. This is the reverse of the result in the HIV-negative men and in previously studied immunocompetent women , among whom 2 of the 3 days of shedding were associated with lesions. Most subclinical shedding was rectal rather than from other anogenital sites. One explanation for the higher rate of subclinical shedding in HIV-positive men is the anatomic site of reactivation. Perirectal reactivation was the most common site of HSV-2 isolation. Some of these reactivations are likely to be associated with unrecognized lesions, given the difficulty of self-examination of the perirectal area. However, not all such subclinical reactivations are likely to be due to unrecognized genital lesions. For example, more days of subclinical HSV-2 shedding than shedding on days with lesions occurred on penile skin (an area that is easy to examine).
Our study indicates that, like HIV-1 RNA levels, there is a wide range in rates of HSV-2 shedding among persons with similar CD4 cell counts. However, despite the interperson variability, both total HSV-2 shedding and subclinical HSV-2 shedding were more frequent in persons with low CD4 cell counts. Compared with persons with CD4 cell counts 1500, men with CD4 cell counts <200/mm3 were more than twice as likely to shed HSV-2 both overall and subclinically. This suggests that persons with advanced disease are an important group to target for suppressive antiviral strategies. However, as emphasized in figure 2, there is wide overlap between rates of HSV-2 shedding and CD4 cell count; many persons with relatively high CD4 cell counts had rates of shedding similar to those of men with low counts, and some persons with <200 CD4 cells/mm3 had relatively low shedding rates. This likely relates to host-virus interactions. Our group has recently associated a low frequency of CD8 cytotoxic T lymphocytes to HSV as a major predictor of HSV-2 reactivation in HIV-positive men.
One interesting aspect of HSV-2 reactivation in HIV-positive men was the wide distribution; >11% of days of HSV shedding involved multiple anatomic sites (46 days in 15 HIV-positive subjects). Shedding from multiple sites is quite uncommon in immunocompetent persons. Shedding at multiple sites was not restricted to subjects with low CD4 cell counts or advanced stage of HIV-1 disease. Although half of the patients with shedding at multiple sites had CD4 cell counts below 225/mm3, the other half had counts ranging up to 847 cells/mm3.
A random sampling of isolates demonstrated that 1.9% of HSV-2 isolates had reduced sensitivity to acyclovir. This is similar to that seen in immunocompetent persons with genital HSV-2 in Seattle 15 years ago. The 3 isolates that were acyclovir-resistant were from patients with both advanced and early disease and from anatomic sites that had acyclovir-sensitive HSV-2 isolated from them before and after isolation of the resistant strain. This observation supports the hypothesis that isolation of acyclovir-resistant viruses is related to duration of HSV replication in epithelial cells, with spontaneous generation of mutants that are thymidine kinase—deficient or thymidine kinase—altered. While acyclovir was used on only 2% of days of follow-up, prior use of acyclovir for episodic and suppressive therapy was commonly reported in this group.
In summary, our cohort is, to date, the largest observational study of HSV-2 in HIV-positive men. We performed nearly 20,000 HSV-2 cultures with >5000 days of follow-up and have demonstrated several unique findings. Perirectal HSV-2 shedding was common, particularly on days on which the subject did not report a lesion. HSV-2 shedding was significantly associated with HIV serostatus and CD4 cell count <200/mm3. Surprisingly, we did not detect an increased frequency of acyclovir resistance among this group of people at high risk for resistance with a history of frequent acyclovir use. Taken together, these data demonstrate that in addition to its role in the transmission and acquisition of HIV-1, HSV-2 is an important opportunistic infection in people with HIV-1 infection and that unrecognized subclinical infection is common. Recent reports have suggested an association between clinical HSV-2 reactivation and increased levels of HIV-1 plasma viremia ; however, the clinical meaning of this observation is unknown. Whether subclinical HSV shedding is associated with increased rates of HIV-1 replication is also unknown.
We thank Stacy Selke for her outstanding contribution in data management; the staff of the University of Washington Virology Laboratory, who performed nearly 20,000 HSV cultures; Mary Shaughnessy, Gail Barnum, and the patients who donated a significant amount of time and energy to performing daily home cultures; and Anna Wald for careful review of the manuscript.